key: cord-0056744-3m41r5ou authors: nan title: The European Union One Health 2019 Zoonoses Report date: 2021-02-27 journal: EFSA J DOI: 10.2903/j.efsa.2021.6406 sha: 5049399fa1c5eb9308dda4c2e34a16364170db31 doc_id: 56744 cord_uid: 3m41r5ou This report of the EFSA and the European Centre for Disease Prevention and Control presents the results of zoonoses monitoring activities carried out in 2019 in 36 European countries (28 Member States (MS) and eight non‐MS). The first and second most reported zoonoses in humans were campylobacteriosis and salmonellosis, respectively. The EU trend for confirmed human cases of these two diseases was stable (flat) during 2015–2019. The proportion of human salmonellosis cases due to Salmonella Enteritidis acquired in the EU was similar to that in 2017–2018. Of the 26 MS reporting on Salmonella control programmes in poultry, 18 met the reduction targets, whereas eight failed to meet at least one. The EU prevalence of Salmonella target serovar‐positive flocks has been stable since 2015 for breeding hens, laying hens, broilers and fattening turkeys, with fluctuations for breeding turkey flocks. Salmonella results from competent authorities for pig carcases and for poultry tested through national control programmes were more frequently positive than those from food business operators. Shiga toxin‐producing Escherichia coli (STEC) infection was the third most reported zoonosis in humans and increased from 2015 to 2019. Yersiniosis was the fourth most reported zoonosis in humans in 2019 with a stable trend in 2015–2019. The EU trend of confirmed listeriosis cases remained stable in 2015–2019 after a long period of increase. Listeria rarely exceeded the EU food safety limit tested in ready‐to‐eat food. In total, 5,175 food‐borne outbreaks were reported. Salmonella remained the most detected agent but the number of outbreaks due to S. Enteritidis decreased. Norovirus in fish and fishery products was the agent/food pair causing the highest number of strong‐evidence outbreaks. The report provides further updates on bovine tuberculosis, Brucella, Trichinella, Echinococcus, Toxoplasma, rabies, West Nile virus, Coxiella burnetii (Q fever) and tularaemia. List of Tables Table 1 Categorisation of data used in EUOHZ 2019 (adapted from Boelaert et al., 2016) ............... 16 Table 2 Reported hospitalisations and case fatalities due to zoonoses in confirmed human cases in the EU, 2019 Table 12 Comparisons Table 14 Comparisons Table 42 Reported human cases of trichinellosis and notification rates per 100,000 population in the The (European Union) EU system for monitoring and collection of information on zoonoses is based on the Zoonoses Directive 2003/99/EC 1 , which obliges EU Member States (MS) to collect relevant and, when applicable, comparable data on zoonoses, zoonotic agents, antimicrobial resistance and foodborne outbreaks. In addition, MS shall assess trends and sources of these agents, as well as outbreaks in their territory, submitting an annual report each year by the end of May to the European Commission covering the data collected. The European Commission should subsequently forward these reports to the European Food Safety Authority (EFSA). EFSA is assigned the tasks of examining these data and publishing the EU Annual Summary Reports. In 2004, the European Commission entrusted EFSA with the task of setting up an electronic reporting system and database for monitoring zoonoses (EFSA Mandate No 2004-0178) . Data collection on human diseases from MS is conducted in accordance with Decision 1082/2013/EU 2 on serious cross-border threats to health. This Decision replaced Decision 2119/98/EC on setting up a network for the epidemiological surveillance and control of communicable diseases in the EU in October 2013. The case definitions to be followed when reporting data on infectious diseases to the European Centre for Disease Prevention and Control (ECDC) are described in Decision 2018/945/EU 3 . ECDC has provided data on zoonotic infections in humans, as well as their analyses, for the EU Summary Reports since 2005. Since 2008, data on human cases have been received via The European Surveillance System (TESSy), maintained by ECDC. According to List A of the Annex I of the Zoonoses Directive 2003/99/EC data on animals, food and feed must be reported on a mandatory basis for the following eight zoonotic agents: Salmonella, Campylobacter, Listeria monocytogenes, Shiga toxin-producing Escherichia coli (STEC), Mycobacterium bovis, Brucella, Trichinella and Echinococcus. In addition and based on the epidemiological situations in the MS, data must be reported on the following agents and zoonoses (List B of the Annex I of the Zoonoses Directive): (i) viral zoonoses: calicivirus, hepatitis A virus, influenza virus, rabies, viruses transmitted by arthropods; (ii) bacterial zoonoses: borreliosis and agents thereof, botulism and agents thereof, leptospirosis and agents thereof, psittacosis and agents thereof, tuberculosis due to agents other than M. bovis, vibriosis and agents thereof, yersiniosis and agents thereof; (iii) parasitic zoonoses: anisakiasis and agents thereof, cryptosporidiosis and agents thereof, cysticercosis and agents thereof, toxoplasmosis and agents thereof; and (iv) other zoonoses and zoonotic agents such as Francisella, Cysticercus and Sarcocystis. Furthermore, MS provided data on certain other microbiological contaminants in foods: histamine, staphylococcal enterotoxins and Cronobacter sakazakii for which food safety criteria are set down in the EU legislation. The general rules on monitoring of zoonoses and zoonotic agents in animals, food and feed are laid down in Article 4 of Chapter II 'Monitoring of zoonoses and zoonotic agents' of the Directive. Specific rules for coordinated monitoring programmes and for food business operators are, respectively, in Articles 5 and 6 of Chapter II. Specific rules for monitoring of antimicrobial resistance are in Article 7 of Chapter III 'Antimicrobial resistance', whereas rules for epidemiological investigation of food-borne outbreaks are in Article 8 of Chapter IV 'Food-borne outbreaks '. According to Article 9 of Chapter V 'Exchange of information' of the Directive, MS shall assess trends and sources of zoonoses, zoonotic agents and antimicrobial resistance in their territory and each MS shall send to the European Commission every year by the end of May a report on trends and sources of zoonoses, zoonotic agents and antimicrobial resistance, covering the data collected under Articles 4, 7 and 8 during the previous year. Reports, and any summaries of these, shall be made publicly available. The requirements for those MS-specific reports are described in Parts A-D of Annex IV as regards the 1 OJ L 293, 5.11.2013, pp. 1-15. 3 Commission Implementing Decision 2018/945/EU on the communicable diseases and related special health issues to be covered by epidemiological surveillance as well as relevant case definitions. OJ L 170, 6.7.2018, pp. 1-74. monitoring of zoonoses, zoonotic agents and antimicrobial resistance carried out in accordance with Article 4 or 7, and in Part E of Annex IV as regards the monitoring of food-borne outbreaks carried out in accordance with Article 8. In accordance with Article 9 of Directive 2003/99/EC, EFSA shall examine the submitted national reports and data of the EU MS 2019 zoonoses monitoring activities as described above and publish an EU Summary Report on the trends and sources of zoonoses, zoonotic agents and antimicrobial resistance in the EU. The 2019 data on antimicrobial resistance in zoonotic agents submitted and validated by the MS are published in a separate EU Summary Report. Since 2019, the annual EU Summary Reports on zoonoses, zoonotic agents and food-borne outbreaks have been renamed the 'EU One Health Zoonoses summary report ' (EUOHZ) , which is jointly drafted and co-authored by EFSA and ECDC. The MS, other reporting countries, the European Commission, members of EFSA's Scientific Panels on Biological Hazards (BIOHAZ) and Animal Health and Welfare (AHAW) and the relevant European Union Reference Laboratories (EURLs) were consulted while preparing the present EU One Health Zoonoses 2019 report. The efforts made by MS, the reporting non-MS and the European Commission in the reporting of zoonoses data and in the preparation of this report are gratefully acknowledged. The present EU One Health Zoonoses summary report focuses on the most relevant information on zoonoses and food-borne outbreaks within the EU in 2019. If substantial changes compared with the previous year were observed, they have been reported. It is noteworthy that EFSA and ECDC were informed on the incompleteness of certain data provision by a few MS due to the COVID-19 pandemic. The latter impacted on national resources allocated to zoonoses and food-borne outbreaks data collection leading to a delay in reports from regional to national levels. Such incompleteness has been mentioned in a few chapters. When the UK data were collected, the UK was an EU MS but as of 31 January 2020, it has become a third country. The analyses of data from infections in humans in the EU Summary Report for 2019 were prepared by the Food-and Waterborne Diseases and Zoonoses (FWD) programme (brucellosis, campylobacteriosis, congenital toxoplasmosis, echinococcosis, listeriosis, salmonellosis, STEC infection, trichinellosis, yersiniosis), Emerging and Vector-borne Diseases (EVD) programme (Q fever, rabies, tularaemia, West Nile virus (WNV) infection) and tuberculosis (TB) programme (TB due to Mycobacterium bovis and M. caprae) at the ECDC. Data were based on the data submitted via The European Surveillance System (TESSy), hosted at ECDC. Please note, as explained above, that the numbers presented in the report may differ from national reports due to differences in case definitions used at EU and national level or to different dates of data submission and extraction. The latter may also result in some divergence in case numbers presented in different ECDC reports. TESSy is a software platform that has been operational since April 2008 and in which data on 56 diseases and special health issues are collected. Both aggregated and case-based data were reported to TESSy. Although aggregated data did not include individual case-based information, both reporting formats were included when possible to calculate number of cases and country-specific notification rates. Human data used in the report were extracted from TESSy as of 7 September 2020 for FWD, as of 9 October 2020 for EVD (except for rabies as of 29 October) and as of 5 October 2020 for TB due to M. bovis and M. caprae. The denominators used for the calculation of the notification rates were the human population data from Eurostat 1 January 2020 update. Data on human zoonoses cases were received from 28 MS and from two non-MS (Iceland and Norway). Switzerland reported its data on human cases directly to EFSA. The human data for Switzerland include data from Liechtenstein. Interpretation of the data should consider data quality issues and differences between MS surveillance systems, and therefore, comparisons between countries should be undertaken with caution. For the year 2019, 28 MS submitted data and national zoonoses reports on monitoring results in food, animals, feed and food-borne outbreaks. In addition, data and reports were submitted by four non-MS and European Free Trade Association (EFTA) countries: Iceland, Norway, Switzerland and Liechtenstein. 4 For some food, animal and feed matrices and food-borne outbreaks, EFSA received data and reports from pre-accession countries Albania, Bosnia and Herzegovina, the Republic of North Macedonia, Montenegro and Serbia. Data were submitted electronically to the EFSA zoonoses database, through EFSA's Data Collection Framework (DCF) . MS could also update data from previous years, before 2019. The deadline for data submission was 31 May 2020. Two data validation procedures were implemented, by 12 June 2020 and by 15 July 2020. Validated data on food, animals and feed used in the report were extracted from the EFSA zoonoses database on 27 July 2020. The draft EU One Health Zoonoses Report was sent to MS for consultation on 7 December 2020 and comments were collected by 23 December 2020. The utmost effort was made to incorporate comments and data amendments within the available time frame. The report was finalised by 22 January 2021 and published online by EFSA and ECDC on 25 February 2021. The detailed description of the terms used in the report is available in the EFSA's manuals for reporting on zoonoses (EFSA, 2020a,b) . The national zoonoses reports submitted in accordance with Directive 2003/99/EC are published on the EFSA website together with the EU One Health Zoonoses Report. They are available online at http://www.efsa.europa.eu/en/biological-hazards-data/reports. Comparability and quality of the data For data on human infections, please note that the numbers presented in this report may differ from national zoonoses reports due to differences in case definitions used at EU and national level or because of different dates of data submission and extraction. Results are generally not directly comparable between MS and sometimes not even between different years in one country. For data on food, animals and feed please note that the numbers presented in this report may differ from national zoonoses reports due to different dates of data submission and extraction. The data obtained in the EFSA DCF can vary according to the level of data quality and harmonisation. Therefore, the type of data analyses suggested by EFSA for each zoonosis and matrix (food, animals, feed or food-borne outbreaks) sampling results strongly depended on this level of harmonisation and can either be a descriptive summary of submitted data, or the following up of trends (trend watching) or the (quantitative) analysis of trends. EFSA carried out data analyses according to Table 1 as adapted from Boelaert et al. (2016) : food, animal, feed and food-borne outbreaks data can be classified into three categories according to the zoonotic agent monitored and the design of the monitoring or surveillance carried out. It follows that these three distinct categories condition which type of data analyses can be implemented. Following the rationale of listing of zoonoses in Annex I of the Directive 2003/99/EC, of the mandatory reporting on food-borne outbreaks and of the above-mentioned categorisation of food, animal and feed data (Table 1) A chapter on food-borne outbreaks constitutes the second section of the EUOHZ. The data submitted to ECDC and to EFSA for List B zoonoses are rather unbalanced (varying numbers of reporting countries and varying data volumes across years) and are collected without harmonised sampling design. Therefore, these zoonoses only supported a simplified chapter structure underpinned The EUOHZ 2019 presents a harmonised structure for each chapter, starting with the key facts. In addition, a section explains the monitoring and surveillance in the EU for the specific disease or for food-borne outbreaks. A results section summarises the major findings of 2019 as regards trends and sources. A summary table displaying the data of the last 5 years (2015) (2016) (2017) (2018) (2019) for human cases and for major animal and food matrices is presented. Each chapter also contains a discussion and ends with a list of related projects and links with useful information for the specific disease. For each chapter, overview tables present reported data by any reporting country. However, for the tables summarising MS-specific results and providing EU-level results, unless stated otherwise, data from industry own check programmes and hazard analysis and critical control point (HACCP) sampling as well as data from suspect sampling, selective sampling and outbreak or clinical investigations are excluded. Moreover, regional data reported by countries without statistics at the national level were also excluded from these summary tables. Statistical trend analyses in humans were carried out to evaluate the significance of temporal variations in the EU and the specifications of these analyses are explained in each separate chapter. The number of confirmed cases for the EU/EEA by month is presented as a trend figure. All countries that consistently reported casesor reported zero cases over the whole reporting periodwere included. The trend figure also shows a centred 12-month moving average, illustrating the overall trend by smoothing seasonal and random variations. Also, in humans, the implemented general-use statistical tests must be viewed as hypotheses generating, not as confirmatory, tests. Analyses other than trend analyses in humans are carried out for confirmed EU cases only (EEA cases were excluded). Spatial trends in food and animals were visualised using R software (www.r-project.org); package ggplot2 as well as ArcGIS from the Economic and Social Research Institute (ESRI) were used to map the data. Choropleth maps with graduated colours over a continuous scale of values were used to map the proportion of positive sample units across the EU and other reporting countries. Statistical trend analysis of food-borne outbreaks was performed to evaluate the significance of temporal variations at the single MS level over the period 2010-2019, as described in the food-borne outbreaks chapter. All summary tables and figures used to produce this report, and that are not displayed, are published as supporting information to this report and are available as downloadable files from the EFSA knowledge junction at the general-purpose open-access repository zenodo at https://doi.org/ 10.5281/zenodo.4298993. All validated country-specific data on food, animals, feed and food-borne outbreaks are also available at the mentioned URL. The numbers of confirmed human cases of 13 zoonoses presented in this report are summarised in Figure 1 . In 2019, campylobacteriosis was the most commonly reported zoonosis, as it has been since 2005, representing 50% of all the reported cases. Campylobacteriosis was followed by other bacterial diseases; salmonellosis, STEC infections and yersiniosis in being the most frequently reported. Severity of the diseases was analysed based on hospitalisation and outcome of the reported cases (Table 2) . Based on data on severity, listeriosis and West Nile virus infection were the two most severe diseases with the highest case fatality and the highest hospitalisation, respectively. Almost all confirmed cases with data available on hospitalisation for these two diseases were hospitalised. About one out of every fifth and one out of 10 confirmed listeriosis and WNV cases, respectively, with known data were fatal. • Seven MS reported monitoring results from official control samples collected in the context of the Campylobacter process hygiene criterion in force for food business operators. Of the 3,346 neck skin samples from chilled broiler carcases, 1,365 (41%) were Campylobacter-positive and 506 (15%) exceeded the limit of 1,000 CFU/g. Seven MS reported such monitoring data based on sampling results collected from the food business operators. Of the 15,323 neck skin samples, 2,038 (13%) tested positive and 1,033 (7%) exceeded the limit of 1,000 CFU/g. • The proportion of Campylobacter-positive samples within the categories 'ready-to-eat' and 'non ready-to-eat' food was 0.2% and 20.6% respectively. In 3,691 'ready-to-eat' food sampling units reported by eight MS, six Campylobacter-positive units were detected; two from raw milk, two from 'fruits, vegetables and juices', one from salads and one from 'other processed food products and prepared dishes'. From 'non ready-to-eat' food, 16 MS reported data and 'meat and meat products' was the most contaminated food category followed by 'milk and milk products' and 'fruits, vegetables and juices', with 23.0%, 2.0% and 0.2% positive sampling units, respectively. Campylobacter was isolated from all fresh meat categories, with the highest percentage of Campylobacter-positive sampling units being reported from fresh meat from turkeys and broilers; 33.0% and 29.6%, respectively. • Sixteen MS reported 2019 sampling results on Campylobacter in animals, mainly from broilers and bovine animals: the highest overall proportion of positives was observed in broilers (13%). Less samples were reported for pigs with a proportion of positives of 59%. Surveillance and monitoring of Campylobacter in the EU The notification of campylobacteriosis is mandatory in 21 EU MS, Iceland, Norway and Switzerland. In six MS, the notification is based on a voluntary system (Belgium, France, Greece, Italy, Luxembourg and the Netherlands) and in one country on another, unspecified system (the United Kingdom). Greece started to report campylobacteriosis data in 2018. The surveillance systems for campylobacteriosis cover the whole population in all MS except in four (France, Italy, the Netherlands and Spain). The estimated coverage of the surveillance system is 20% in France and 52% in the Netherlands. These estimated proportions of population coverage were used in the calculation of notification rates for these two MS. No estimates of population coverage in Italy and Spain were provided, so notification rates were not calculated for these two MS. Tables and figures that are not presented in this chapter are published as supporting information to this report and are available as downloadable files from the EFSA knowledge junction at zenodo https://doi.org/ 10.5281/zenodo.4298993. Summary statistics of human surveillance data with downloadable files are retrievable using ECDC's Surveillance Atlas of Infectious Diseases at http://atlas.ecdc.europa.eu/public/index.a spx , Spain did not receive data from all regions due to COVID-19, so the number of reported cases was lower than expected. The drop in cases in Luxembourg in 2019 is a surveillance artefact caused by a change to non-culture methods in private laboratories, resulting in reduced numbers of isolates sent to the national reference laboratory. From March 2020, an electronic laboratory notification system has been in place in Luxembourg and the campylobacteriosis notifications are expected to increase as a result. All countries reported case-based data except Belgium, Bulgaria and Greece, which reported aggregated data. Both reporting formats were included to calculate annual numbers of cases and notification rates. Diagnosis of human infection is generally based on culture from human stool samples and both culture and non-culture methods (polymerase chain reaction (PCR)) are used for confirmation. Biochemical tests or molecular methods are used for species determination of isolates submitted to the National Public Health Reference Laboratories (NPHRL). Monitoring of Campylobacter along the food chain is conducted during the primary production stage (farm animals), during harvest/slaughter and processing and at retail stages. A regulatory limit (microbiological process hygiene criterion (PHC)) for Campylobacter has been set for broiler carcases in Regulation (EC) No 2073/2005 (point 2.1.9 of Chapter 2 of Annex I). The Campylobacter PHC evaluates the counts above 1,000 CFU/g of Campylobacter on neck skins from broiler carcases after chilling, considering a set of 50 (pooled) samples derived from 10 consecutive sampling sessions. This criterion aims to stimulate action to lower the counts of Campylobacter on broiler carcases and to reduce the number of human campylobacteriosis cases due to the consumption or handling of chicken/broiler meat. This PHC has been in force since 1 January 2018. Food business operators (FBOp) shall use the criterion to validate and verify the correct functioning of their food safety management procedures based on HACCP principles and Good Manufacturing Practices (GMPs). FBOp must carry out corrective actions if the criterion target is exceeded. Official samples taken by the Competent Authorities (CA) serve the purpose of auditing the FBOp actions and ensure that the FBOp complies with regulatory requirements. Since 14 December 2019, the Commission Implementing Regulation (EU) 2019/627 5 entered into force to harmonise the sampling within official control. Also, reporting of results became mandatory. According to this legislation, the CA has to verify whether the FBOp correctly implements and checks the PHC conducted on broiler carcases by choosing between two approaches: implementing ad hoc official samplings 6 or collecting all information on the total number and the number of Campylobacter samples with more than 1,000 CFU/g taken by FBOp in accordance with Article 5 of Regulation (EC) No 2073/2005. These harmonised official control results, which became compulsory to report, will allow better trend watching and trend analyses than before (Table 1) . Official control results from tests for Campylobacter on chilled broiler carcases had the following specified options for the different data elements: sampler: 'official sampling ' and/or 'industry sampling' and 'HACCP and own check' (self-monitoring) ; sampling context: 'surveillance, based on Regulation (EC) No 2073/2005'; sampling unit type: 'single'; sampling strategy: 'objective sampling' and sampling stage: 'slaughterhouse'. Campylobacter monitoring data at slaughter from poultry caeca as part of the annual antimicrobial resistance monitoring are collected in a more harmonised way. Other monitoring data on Campylobacter from food and animals and submitted to EFSA according to Chapter II 'Monitoring of zoonoses and zoonotic agents ' collected without harmonised design. These data have other specified options for the different data elements (including sampling context other than based on Regulation (EC) No 2073/2005) and allow for descriptive summaries at EU level to be made, but they do not support EU-level trend analyses and trend watching (Table 1 ). In 2019, general data on food and animals reported to EFSA by MS and non-MS derived mainly from official sampling, industry sampling and from HACCP and own checks, in the context of national monitoring and surveillance and/or organised surveys. In addition, for animal data, other reported samples were from clinical investigations by private veterinarians and industry (artificial insemination centres). The reported occurrence of Campylobacter in the most important food categories for the year 2019 and for the 4-year period 2015-2018 was descriptively summarised, making a distinction between RTE and non-RTE food. Data sets were extracted with 'Objective sampling' being specified as sampler strategy, which means that the reporting MS collected the samples according to a planned strategy based on the selection of a random sample, which is statistically representative of the population to be analysed. Detection of Campylobacter in food and animals is generally based on culture and both biochemical and molecular methods (such as PCR) as well as matrix-assisted laser desorption/ionisation, time-offlight mass spectrometry (MALDI-TOF MS) are used for confirmation and species identification. Table 3 summarises EU-level statistics on human campylobacteriosis, and on the occurrence and prevalence of Campylobacter in food and animals, respectively, during 2015-2019. Food data of interest reported were classified into the major categories 'meat and meat products' and 'milk and milk products' and aggregated by year to obtain an annual overview of the volume of data submitted. The number of sampling units reported for 2019 for 'meat and meat products' increased substantially compared with 2018, which is likely due Commission Implementing Regulation (EU) 2019/627 prescribing compulsory reporting of PHC monitoring data (see above). A more detailed description of the food-borne outbreaks statistics is in the chapter on food-borne outbreaks. When the UK data were collected the UK was an EU MS but as of 31 January 2020 it has become a third country. For 2019, 220 ,682 confirmed cases of human campylobacteriosis were reported by 28 EU MS, corresponding to an EU notification rate of 59.7 cases per 100,000 population (Table 4 ). This is a decrease by 6.9% compared with 2018 (64.1 cases per 100,000 population). The highest country-specific notification rates in 2019 were observed in Czechia (215.0 cases per 100,000), Slovakia (141.1), Denmark (93.0) and the United Kingdom (88.1). The lowest rates in 2019 were observed in Bulgaria, Cyprus, Greece, Latvia, Poland, Portugal and Romania (≤ 8.6 per 100,000). Most (94.4%) of the campylobacteriosis cases reported with known origin were infected in the EU (Table 3 ). The highest proportions of domestic cases (> 97%) were reported in Czechia, Hungary, Latvia, Malta, Poland, Portugal, Romania and Slovakia. The highest proportions of travel-associated cases were reported by the Nordic countries: Finland (77.8%), Denmark (44.1%), Sweden (56.3%), Iceland (57.0%) and Norway (54.8%). Among 14,501 travel-associated cases with known country of infection in the MS, almost half of the cases (48.1%) were linked to travel within the EU, with most of the cases having acquired infections in Spain, Greece and Italy (13.9%, 4.1% and 3.6%, respectively). Turkey, Thailand and Morocco were the most often reported probable countries of infection outside the EU (8.2%, 7.8% and 4.9%, respectively). (a): The summary statistics, referring to MS, were obtained by summing all sampling units (single, batch, slaughter batch), sampling stage (farm, packing centre, automatic distribution system for raw milk, processing plant, cutting plant, slaughterhouse, catering, hospital or medical care facility, restaurant or cafe or pub or bar or hotel or catering service, retail, wholesale, unspecified), sampling strategies (census, convenience sampling, objective sampling and unspecified) and sampler (official sampling, official and industry sampling, private sampling, unspecified, not applicable). (b): Meat and meat products refer to carcases and fresh meat/ready-to-eat (RTE), cooked and fermented products. (c): Milk and milk products refer to raw and pasteurised milk and all dairy products including cheeses. Between 2015 and 2019, there was a clear seasonality in the number of confirmed campylobacteriosis cases reported in the EU/EEA, with peaks in the summer months. Annual winter peaks, albeit with lower numbers compared with summer, were also observed in January annually from 2012 to 2019. The EU/EEA trend was stable (flat) during 2015-2019 ( Figure 2) . Hungary was the only MS reporting decreasing (p < 0.01) trend, in the period 2015-2019. Four MS (Italy, Latvia, Portugal and Romania) reported increasing trends in the same time period. Information on hospitalisation status was provided for 29.1% of all campylobacteriosis cases by 16 MS in 2019. Of cases with known hospitalisation status, 31.8% were hospitalised. The highest hospitalisation rates were reported in Cyprus, Latvia, Lithuania, Poland, Romania and the United Kingdom, where most reported cases were hospitalised. The outcome was reported for 78.0% of all cases by 17 MS. Forty-seven deaths due to campylobacteriosis were reported in 2019, resulting in an EU case fatality of 0.03%. This was similar to the average percentage of fatal outcome observed over the last 5 years. Campylobacter species information was provided by 24 MS for 55.2% of confirmed cases reported in the EU, which was at the same level as in 2018. Of these, 83.1% were Campylobacter jejuni, 10.8% Campylobacter coli, 0.1% Campylobacter lari, 0.1% Campylobacter fetus and 0.1% Campylobacter upsaliensis. 'Other' Campylobacter species accounted for 5.8%, but the large majority of those cases were reported at the national level as 'C. jejuni/C. coli/C. lari not differentiated'. Overall, for the year 2019, 94.5% of the number of reported human campylobacteriosis cases who acquired the infection in the EU (109,930; Table 3 ) were domestic (acquired within the home country) infections and 5.5% were acquired through travel in EU. Campylobacter was the third most frequently reported causative agent for food-borne outbreaks at the EU level, by 18 MS, with 319 outbreaks communicated to EFSA, 1,254 cases of illness, 125 hospitalisations and no deaths. Comparing the food-borne outbreak cases (1, 254) , reported to EFSA, and cases of human campylobacteriosis acquired in the EU (109,930) considering also the proportion Source(s): Austria, Cyprus, Czechia, Denmark, Estonia, Finland, France, Germany, Hungary, Iceland, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, the Netherlands, Norway, Poland, Romania, Slovakia, Slovenia, Sweden and the United Kingdom. Belgium, Bulgaria, Croatia, Greece, Portugal and Spain did not report data to the level of detail required for the analysis. (Table 3) , reported to ECDC, could suggest that overall in the EU, in 2019, only 0.6% of human campylobacteriosis cases would be reported through food-borne outbreaks investigation. It is important to clarify that the case classification for reporting is different between these two databases. In TESSy, the cases reported are classified based on the EU case definition. All these cases visited a doctor and are either confirmed by a laboratory test (confirmed case) or not (probable case and classification is based on the clinical symptoms and epidemiological link). Cases that never visited a doctor are not reported to TESSy. Moreover, there may be missing probable cases in TESSy, as these data are not analysed or published and there is no incentive for reporting such cases. Information on which cases are linked to an outbreak and which not is also not systematically collected. In practice, the cases reported to TESSy are considered to be mostly sporadic cases. In food-borne outbreaks, the human cases are the people involved in the outbreak as defined by the investigators (case definition), and cases must be linked, or probably linked, to the same food source (Directive 2003/99/EC). This can include both ill people (whether confirmed microbiologically or not) and people with confirmed asymptomatic infections . Cases can be classified as confirmed or probable outbreak cases, but currently these specific classification data are not collected by EFSA. C. jejuni and C. coli were identified in 72 and 7 outbreaks, respectively. However, most campylobacteriosis food-borne outbreaks were reported without speciation information (240 outbreaks: 75.2%). Eighteen campylobacteriosis outbreaks were reported with strong-evidence and 301 with weak evidence. Of the former outbreaks, eight were caused by broiler meat and three by milk. During 2010-2018, these were also the food vehicles causing most strong-evidence campylobacteriosis food-borne outbreaks. Further details and statistics on the campylobacteriosis foodborne outbreaks for 2019 are in the food-borne outbreaks chapter. Campylobacter data in the context of Regulation (EC) No 2073 In total, seven MS (Bulgaria, Croatia, Cyprus, Estonia, Latvia, Romania and Spain) reported 2019 ad hoc official sampling results collected in the context of the Campylobacter PHC, which are quantitative data relating to neck skins from broiler carcases sampled at slaughterhouses. Of the 3,346 neck skin samples from chilled broiler carcases, 1,365 (41%) tested positive and 506 (15%) exceeded the limit of 1,000 CFU/g. However, the MS-specific percentage of quantified results exceeding that limit varied widely and ranged from zero to 34%. Seven MS (Denmark, Estonia, Germany, Ireland, Latvia, Romania and Sweden) reported 2019 Campylobacter PHC monitoring results collected from the FBOp. Of the 15,323 neck skin samples from chilled broiler carcases, 2,038 (13%) tested positive and 1,033 (7%) exceeded the limit of 1,000 CFU/ g. The MS-specific percentage of quantified results exceeding that limit varied from zero to 14%. Other food monitoring data Table 5 summarises the reported occurrence of Campylobacter in the most important food categories for the year 2019 and for the 4-year period 2015-2018. Distinction is made between RTE, and non-RTE food, and fresh meat. The proportion of Campylobacter-positive samples within the RTE and non-RTE categories was 0.2% and 20.6% respectively. , most results from the 3,691 RTE food sampling units reported by eight MS originated from 'fruits, vegetables and juices ' (27.3%) , followed by 'other processed food products and prepared dishes ' (27.1%) , 'milk and milk products ' (22.2%) and 'meat and meat products ' (8.9% ). In total, Campylobacter was detected in six RTE food samples: two from raw milk, two from 'fruits, vegetables and juices', one from salads and one from 'other processed food products and prepared dishes '. During 2015 '. During -2018 , in the RTE food category, 27 Campylobacter-positive sampling units were reported from 'meat and meat products', in particular from broiler meat and broiler meat products, six from raw milk, two from cheeses and one from 'fruits, vegetables and juices '. Results reported by 16 MS for non-RTE food show that 'meat and meat products' was the most contaminated food category as compared with 'milk and milk products ' and 'fruits, vegetables and juices', in 2019. This was also the case for the years 2015-2018. Fifteen MS reported for 2019 results for fresh meat categories and all had some positive samples but the percentages of Campylobacterpositive sampling units for fresh meat from broilers and turkeys were very high. This was also the case for the years 2015-2018. In 2019, in total, 16 MS and four non-MS reported monitoring data on Campylobacter in animals. Most samples originated from broilers and from bovine animals, and all proportions (%) of positive sampling units are displayed in Table 6 . Campylobacteriosis has been the most commonly reported zoonosis in humans in the EU since 2005. Despite comprehensive surveillance and national coverage in most MS, reported cases represent only a small proportion of Campylobacter infections occurring in the EU (Teunis et al., 2013) . There has been a significantly increasing trend in the number of cases at the EU level and at country level in half of the MS between 2009 and 2018. In the last 5 years from 2015 to 2019, the EU trend of confirmed cases has stabilised. In 2019, in two-thirds of the MS, the number of confirmed campylobacteriosis cases decreased and the EU notification rate decreased by 6.9% compared with the rate in 2018. Despite this reduction, only one MS had a significant decreasing trend in the last 5 years. Four MS reported increasing trends, whereas most MS had stable, flat trends in 2015-2019. One MS notified that the reported number of campylobacteriosis cases is lower than expected as data were not received from all regions due to the COVID-19 situation in 2020. It is not clear if, and to what extent, the pandemic situation had an effect on the decrease of notifications noted in several MS in 2019. In previous years, there has been a steady annual increase in reported cases in several countries. This may not only reflect changes in exposure but also improvements in surveillance systems, a better coverage of routine diagnostics across the country, requirement for medical laboratories to report positive test results and better knowledge and awareness among physicians. Almost half of the MS reported having the capacity to perform whole genome sequencing (WGS) on Campylobacter isolates (ECDC survey, 2020, data not published). Campylobacter has a characteristic seasonality with a sharp increase of cases in the summer. Campylobacter tends to be more prevalent in humans during warmer seasons of the year; however, a smaller but distinct winter peak has become apparent in the past 8 years in the EU, including in 2019. Disease onsets of cases that were notified during winter peaks occurred predominantly in the threefirst calendar weeks of the year. This points towards exposures around Christmas and New Year. Winter peaks have been observed in Austria, Belgium, Finland, Germany, Luxembourg, the Netherlands, Switzerland and Sweden. Increased travel during the holiday season might be another explanation for the increase in many countries. In some countries with an observed winter peak, the consumption of meat fondue or table-top grilling is popular during the festive season and could promote Campylobacter transmission (Bless et al., 2017) . In the EU, over 20,000 campylobacteriosis cases were hospitalised in 2019. This is the highest number of hospitalisations compared with all other food-borne infections. The proportion of hospitalised campylobacteriosis cases was higher than expected in some MS, where all or most of the confirmed cases were hospitalised. These MS also reported the lowest notification rates, indicating that the surveillance is focusing mainly on hospitalised, i.e. severe cases. Hospitalisation status is ascertained and reported by hospitals, while for cases reported from other sources, e.g. laboratories, hospitalisation status is often missing. This can result in an overestimation of the proportion of hospitalised cases in some countries. Broiler meat is considered the main source of human campylobacteriosis (EFSA BIOHAZ Panel, 2010) . In 2011, EFSA published an opinion on 'Campylobacter in broiler meat production: control options and performance objectives and/or targets at different stages of the food chain , which suggested the introduction of a microbiological criterion for Campylobacter on broiler carcases at the slaughterhouse. EFSA estimated that the public health risk from Campylobacter could be reduced by > 50% if no batches would exceed a critical limit of 1,000 CFU/g on neck and breast skin. This process hygiene criterion (PHC) has been in force for food business operators since 1 January 2018. Moreover, a 2012 EFSA opinion on the public health hazards to be covered by inspection of poultry meat identified the need to address Campylobacter as a high priority (EFSA BIOHAZ, CONTAM and AHAW Panels, 2012) . In line with the high priority set by this EFSA opinion on poultry meat inspection, competent authorities ought to sample themselves for Campylobacter or carefully verify the implementation of the process hygiene criterion by the operator. Official samples taken by the competent authorities serve the purpose of auditing the food business operators' actions and ensure that the food business operators comply with regulatory requirements. Since 14 December 2019, the Commission Implementing Regulation (EU) 2019/627 entered into force to harmonise the sampling within official control. Also, reporting of results became mandatory. Seven MS reported 2019 official control monitoring data from neck skin samples from chilled broiler carcases collected in the context of the Campylobacter PHC. Overall, one in six samples exceeded the limit of 1,000 CFU/g. Six MS reported such monitoring data based on sampling results collected from the food business operators and these data showed that one in 14 samples exceeded the limit of 1,000 CFU/g. Better populated EU summary tables with more complete data sets from all MS will in future allow better trend watching and trend analyses. Other monitoring data on Campylobacter from food were submitted to EFSA according to Chapter II 'Monitoring of zoonoses and zoonotic agents' of the Zoonoses Directive 2003/99/EC. These data are collected without harmonised design between the MS. Eight MS reported monitoring data for RTE food and overall a few Campylobacter-positive units were detected; in raw milk, 'fruits, vegetables and juices', salads and 'other processed food products and prepared dishes'. Monitoring data considered were collected according to an 'objective' sampling strategy. Also considering the fact that for certain food categories, such as RTE milk, the overall sampling effort was small (five MS reporting 204 sample results) the finding of Campylobacter-contaminated RTE food is of concern because it poses a direct risk to the consumer. No Campylobacter-positive RTE meat and meat products were reported for 2019; however, the overall sampling effort was small (six MS, 328 sampling units). During 2015-2018, one in 40 RTE meat and meat products sampling units was reported positive, and for RTE meat and meat products from broilers, one in five was positive, albeit based on a small sample size (three MS, 117 samples). Quantitative data (counts) of Campylobacter are currently only collected in the context of the aforementioned PHC. Monitoring data for non-RTE 'meat and meat products' showed that one in five samples were positive, for 'milk and milk products' one in 50 and for 'fruits, vegetables and juices' one in 500. Fifteen MS reported results for fresh meat categories and the overall percentage of Campylobacter-positive sampling units for fresh meat from broilers and turkeys were very high, 32.10% and 33.04%, respectively. In 2020, EFSA experts updated the 2011 scientific opinion (EFSA BIOHAZ Panel, 2011) using more recent scientific data and reviewed on-farm control options for Campylobacter in broilers (EFSA BIOHAZ Panel, 2020a). The relative risk reduction in EU human campylobacteriosis attributable to broiler meat was estimated for on-farm control options using population attributable fractions for interventions that reduce Campylobacter flock prevalence, updating the modelling approach for interventions that reduce caecal concentrations and reviewing scientific literature. The updated model resulted in lower estimates of impact of interventions (control options) than the model used in the 2011 opinion. A 3-log 10 reduction in broiler caecal concentrations was estimated to reduce the relative EU risk of human campylobacteriosis attributable to broiler meat by 58% compared with an estimate larger than 90% in the previous opinion. • The trend for salmonellosis in humans has been stable (flat) over the last 5 years after a long period of a declining trend. • 'eggs and egg products', followed by 'bakery products', 'pig meat and products thereof' and 'mixed food', as in previous years. • Official control samples verifying compliance with food safety criteria according to Regulation (EC) No 2073/2005 found the highest percentages of Salmonella-positive samples in poultry meat, including fresh meat (3.5%), minced meat and meat preparations intended to be eaten cooked (8.3%) and meat products intended to be eaten cooked (6.4%). • For 2019, 66,113 'ready-to-eat' and 191,181 and 'non ready-to-eat' food sampling units were reported from 21 and 25 MS with 0.3% and 1.5% positive samples, respectively. Within the category of 'ready-to-eat' food samples, positive samples were from divers food products; 'meat and meat products', 'milk and milk products', 'fruits, vegetables and juices', 'fish and fishery products', 'spices and herbs ', 'salads', ' other processed food products and prepared dishes', 'cereals and nuts', 'infant formulae and follow-on formulae', 'other food' and 'cocoa and cocoa preparations, coffee and tea'. Within the category of 'non ready-to-eat' food samples, positive samples originated also from divers food products and were mostly from 'meat and meat products', notably from fresh meat from broilers and from turkeys. • Significantly lower percentages of Salmonella-positive pig carcases were reported, based on food business operators self-monitoring data, compared with official control data from the competent authorities. The same observations were made for 2018 and 2017 data. • Eighteen of the 26 Member States reporting on Salmonella control programmes in poultry populations met all the reduction targets, compared to 14 in 2018. The number of MS that did not meet the Salmonella reduction targets was five in breeding flocks of Gallus gallus, four in laying hen flocks, one in broilers flocks, zero in flocks with breeding turkeys and one in fattening turkey flocks. • Among the target Salmonella serovars in the context of national control programmes in poultry, the reported flock prevalence was highest for S. Enteritidis in breeding flocks of Gallus gallus and laying hens. For broilers, the flock prevalence of S. Enteritidis and of S. Typhimurium were comparable, whereas for turkeys (both breeding and fattening flocks), the flock prevalence of S. Typhimurium was highest. • In the context of national control programmes in poultry, proportions of Salmonella target serovars-positive broiler and fattening turkey flocks reported by food business operators was significantly lower than those reported by competent authorities. • A significant increase was noted in estimated Salmonella prevalence in breeding flocks of Gallus gallus, laying hens and breeding turkeys over the last 4-6 years. The trends in the prevalence of Salmonella target serovar-positive flocks were, in contrast, quite stable (flat) since 2015 for all animal categories, with some fluctuations for breeding turkey flocks. • Of all serotyped Salmonella isolates reported by MS from food and animal sources, 70% originated from the broiler source, 12% from the pig source, while the laying hen and turkey sources accounted each for about 7% and isolates from the cattle source made up about 1%. The top five serovars responsible for human infections were distributed as follows among the serotyped isolates (17,176) from these food-animal sources: S. Infantis accounted for 29.7% of them, S. Enteritidis 6.9%, monophasic variant of S. Typhimurium 4.5%, S. Typhimurium 3.9% and S. Derby 3.7%. On the left side of the infographic are shown: (a) Map of the salmonellosis notification rates per 100,000 population in the EU/EFTA; (b) the single Member States' prevalence in the context of national control programmes (NCP) in poultry compared with the European reduction target for laying hens (2%) and other poultry populations (1%); (c) the trends of the prevalence of poultry flocks positive for Salmonella target serovars in the context of NCP; (d) the comparisons between the results of the competent authorities (CA) and food business operators (FBOp) data in the context of the NCP; on the right side; (e) the distribution of the human top five Salmonella serovars coming from serotyped isolates from food and animal matrices reported by reporting MS, and (f) the distribution of human top five Salmonella serovars isolates according to different food and animal matrices. Figure 3 summarises the main data reported in the Salmonella chapter and the major findings. It is a 'graphical abstract' presenting a global overview of the data mentioned in the Key facts section. Surveillance and monitoring of Salmonella in the EU The notification of non-typhoidal salmonellosis in humans is mandatory in 22 MS, Iceland, Norway and Switzerland, whereas in five MS reporting is based on a voluntary system (Belgium, France, Luxembourg and the Netherlands) or other systems (the United Kingdom). In the United Kingdom, although the reporting of food poisoning is mandatory, isolation and species identification of the causative organism is voluntary. The surveillance systems for salmonellosis cover the whole population in all MS except in France, the Netherlands and Spain. The estimated coverage of the surveillance system is 48% in France and 64% in the Netherlands. These proportions of populations were used in the calculation of country-specific and EU-level notification rates. No estimation for population coverage in Spain was provided, so the notification rate was not calculated. For 2019, Spain did not receive data from all regions that are normally reporting due to COVID-19, and therefore, the case numbers are lower than expected. All countries reported case-based data except Bulgaria, which reported aggregated data. Both reporting formats were included to calculate annual numbers of cases and notification rates. Diagnosis of human Salmonella infections is generally carried out by culture from human stool samples. All countries, except Bulgaria perform serotyping of isolates. Monitoring of Salmonella along the food chain is conducted during preharvest (farm animals and their feed), processing (slaughterhouses and cutting plants) and post-processing (wholesale, retail and catering) stages. Regulatory limits (microbiological criteria) for Salmonella have been set for food specified in Regulation (EC) No 2073/2005 (Figure 4 ), which lays down Salmonella food safety criteria (FSC) and Salmonella PHC. Compliance with these criteria ought to be legally verified by the individual food business operator in the context of their own HACCP programmes, through self-monitoring when implementing the general and specific hygiene measures of Regulation (EC) No 852/2002. Respect of the criteria should be guaranteed by the FBOp by preventive approaches (e.g. implementing good hygiene practices, GMPs and the application of risk management procedures based on HACCP). The collection of these data is not fully harmonised across MS, because the sampling objectives, the place of sampling and the applied sampling frequency vary or are interpreted differently between MS. The competent authority (CA), through official sampling or oversight of data, ensures that the food business operator (FBOp) complies with the regulatory requirements. The Salmonella FSC prescribe that Salmonella is not detected in 25 or 10 g of different products (from 5 to 30 sampling units for the specified food categories) when they are on the market, during their shelf-life. Moreover, according to Regulation (EC) No 1086/2011, in fresh poultry meat, the FSC prescribes that target serovars for poultry populations (S. Enteritidis and S. Typhimurium including monophasic S. Typhimurium) are 'not detected in 25 g'. Salmonella PHC are regulated for carcases of pigs, cattle, sheep, goats, horses, broilers and turkeys, and evaluate the presence of Salmonella on a specific area of a tested carcass, or on a pooled sample of neck skin from broilers and turkeys, considering a set of 50 samples derived from 10 consecutive sampling sessions. Salmonella isolates collected from broilers and turkeys must be serotyped for the identification of S. Enteritidis and S. Typhimurium. Since 14 December 2019, the Commission Implementing Regulation (EU) 2019/627 6 entered into force to harmonise the sampling within official control. Also, reporting of results became mandatory. According to this legislation, the CA has to verify whether the FBOp correctly implements and checks the PHC conducted on carcases (points 2.1.3, 2.1.4 and 2.1.5 of Chapter 2 of Annex I of Regulation (EC) No 2073/2005) by choosing between different approaches: implementing ad hoc official samplings 7 and/or collecting all information on Salmonella-positive samples from own checks by the FBOp and/or collecting information on Salmonella-positive samples as part of national control programmes in the MS with special guarantees (Regulation (EC) No 853/2004). These harmonised official control results, which became compulsory to report, will allow better trend watching and trend analyses than before (Table 1) . Official control results from Salmonella had the following specified options for the different data elements; sampler: 'official sampling', except for pig carcases for which the sampler has to be labelled as 'official, based on Regulation No 854/2004' and/or 'industry sampling' and 'HACCP and own check' (self-monitoring) , for the PHC; sampling context: 'surveillance, based on Regulation (EC) No 2073/ 2005'; sampling unit type: 'single'; sampling strategy: 'objective sampling'; and sampling stage: sampling units collected at the processing phase (e.g. slaughterhouse and cutting plant), or at the retail stage, identified as 'catering', 'hospital or medical care facility', 'restaurant or cafe or pub or bar or hotel or catering service ' and 'wholesale'. Monitoring data for compliance with the Salmonella national control programmes in poultry According to EU Regulation (EC) No 2160/2003 and its following amendments, MS have to set up Salmonella national control programmes (NCP) aimed at reducing the prevalence of Salmonella serovars that are considered relevant for public health (from this point forward termed target serovars), in certain animal populations. An overview of NCP for the poultry populations, relative targets to reach and serovars to be targeted is shown in Table 7 . It is compulsory for MS to annually report results for Salmonella NCP and, in addition for broiler flocks and breeding and fattening turkey flocks, it is compulsory to report separate results for samplings conducted by CA and by FBOp. These NCP data allow data analyses such as assessing spatial and temporal trends at the EU level. They also allow for descriptive summaries at the EU level to be made and allow EU trends to be monitored (Table 1) . Other monitoring data for foods, animals and feed Food, animal and feed monitoring data other from those described above are not collected in a harmonised way, because there are no requirements for sampling strategies, sampling methods, Figure 4 . There are also no harmonised rules for reporting these data. These data are summarised only and do not serve the purpose of trend watching or trend analyses ( Table 1) . The reported occurrence of Salmonella in the most important food categories for the year 2019 and for the 4-year period 2015-2018 was descriptively summarised making a distinction between RTE and non-RTE food. Data sets were extracted with 'objective sampling' being specified as sampler strategy, which means that the reporting MS collected the samples according a planned strategy based on the selection of a random sample, which is statistically representative of the population to be analysed. Reported Salmonella serovar data are also viewed as part of this category. MS are obliged to report the target serovars as part of the NCP in poultry populations, whereas for the other animal populations, serotyping is not mandatory and if it is performed, reporting of the serovar data is not mandatory either. Also, for the food sector, the FSC are the absence of Salmonella, except for fresh poultry meat, for which the criterion is limited to the absence of the target serovars. Therefore, some MS may decide to not report the presence of non-target serovars, which would lead to a possible reporting bias for target serovars in poultry populations and for fresh poultry meat. Hence, the compulsory reporting of target serovars in the context of NCP and in the context of the FSC for fresh poultry meat guarantees the consistency of such data over many years and among MS but could result in an overestimation of these target serovars compared with the other serovars. For the remaining matrices, the serovar data collected could be strongly biased by what each MS serotyped and reported. Also, in this context, detection of Salmonella serovars other than those covered by the reduction targets does not in any way equate with a 'Salmonella free' finding. The surveillance and monitoring of Salmonella in food, food-producing animals and feed according to the sampling stage, the sampler, the objective of the sampling, the quality of data and the degree of harmonisation Comparison of Salmonella results from CA and FBOp in the context of NCP for those programmes requiring separate reporting (NCP for broilers, fattening turkeys and breeding turkeys) as well as Salmonella PHC monitoring data from carcases (pigs), was carried out. The significance of differences was verified by the one-tailed Fisher's exact probability test, in cases in which the expected values in any of the cells of a contingency table were below 5; otherwise the z-statistic one-tailed test was calculated. A p-value < 0.10 7 was considered significant to consider every possible evidence of differences between FBOp and CA. Differences in official control sampling results by CA and selfmonitoring results by FBOp were expressed by exact binomial confidence interval (95% level). R software (www.r-project.org) was used to conduct the above-mentioned analyses. Statistical trend analyses were carried out with the objectives of evaluating the significance of temporal variations in the EU-level flock prevalence of Salmonella and target Salmonella serovars in poultry since the start of the implementation of the NCP. The tested flocks were either positive or negative for target serovars and Salmonella, and so the status of the flocks is a dichotomous outcome variable. Therefore, the binomial probability distribution for the response variable was assumed and the logit link function was computed in the model for the trend analysis. The logit is defined as the logarithm of p/(1p), where p/(1p) is the odds of being positive for Salmonella. According to the temporal flock prevalence trends in the MS, polynomial or B-spline basic models (in case of a supposed high degree of polynomial trend) for the logit of the probability of flocks being positive were fitted for the different poultry categories over the entire period of NCP implementation. Moreover, attention has been paid to the period after the achievement of the minimum prevalence reported to date, to capture any evidence of a significant increase in Salmonella prevalence. Marginal and conditional generalised linear models for repeated measures were used to perform these trend analyses. Details about the estimated parameters of the models, odds ratios, prevalence and graphical analyses (conditional and marginal) are reported in the supporting information to this report. To investigate the EU-level prevalence considering the relevant heterogeneity among MS for flock prevalence of Salmonella and target serovars over time, the results obtained using the conditional generalised mixed model for longitudinal binary data were summarised and discussed in the report, for all poultry categories covered by the NCP. To take account of the different levels (baselines) of risk of MS having positive flocks, but similar patterns over time, a random MS-specific intercept effect was included in the model. To consider the trend over time, the variable 'time' was included in the model as a fixed effect. The correlation among repeated observations in the same MS in subsequent years was considered using a first autoregressive or exchangeable structure of the correlation matrix for the residuals. To evaluate the significance of the overall effect of fixed factors specified in the model, Type III F-tests were applied, whereas the receiver operating characteristic (ROC) curve was used to assess the goodness of fit of the model. A p-value < 0.10 was considered to be significant for both random and fixed effects. GLIMMIX and SGPLOT procedures in SAS 9.4 software were used to fit the models and to produce the graphical outputs, respectively. With the aim of evaluating the distribution of Salmonella serovars along the food chain and identifying the potential sources for human infections, descriptive analyses were made from serovar data on food and food-producing animals for the most commonly reported Salmonella serovars from 7 Chapter 11.2 of Statistical Models in Epidemiology, Clayton D and Hills M (2013). human cases acquired within the EU (domestically or during travel within the EU). For animal categories covered by the NCP, only serovar data reported in the context of these programmes were presented. For cattle, meat-producing animals were considered, whereas for pigs, data from fattening animals were used. To interpret serovar data, it must be kept in mind that for NCP, mandatory reporting is limited to target serovars only and this could lead to a possible bias towards the reporting of these regulated serovars to the detriment of non-regulated ones. For all the other animal speciesfood matrices the reporting of serovar data is carried out on a voluntary basis by the MS. Apart from possible reporting bias as regards serovars, the reporting on animal or food categories could also be unbalanced and specific sources (e.g. cattle) may be underrepresented. Sankey diagrams were provided to show the most reported Salmonella serovars from humans in relation to their likely food and animal sources and in relation to the MS reporting them (geographical provenance). Stacked bar plots for each of the serovars of interest were prepared to show for each source the frequency of reporting in animal and food sources. Both graphical representations were performed using R software (www.r-project.org). The infographic, showing the most relevant data about Salmonella, was produced using Adobe Illustrator and InDesign. Table 8 summarises EU-level statistics on human salmonellosis and on Salmonella in food and animals, respectively, during 2015-2019. Food data of interest reported were classified into the major categories and aggregated by year to obtain an annual overview of the volume of data submitted. More detailed descriptions of these statistics are in the results section of this chapter and in the chapter on FBOs. When the UK data were collected, the UK was an EU MS but as of 31 January 2020, it has become a third country. In 2019, the number of reported human salmonellosis cases acquired in the EU (i.e. by domestic infection and through travel within the EU) was at the same level as in 2018. The number of outbreakrelated cases and the total number of food-borne salmonellosis outbreaks was lower in 2019 compared with 2018 and at a lower level compared with 2017 and previous years. The number of sampling units reported in 2019 for the different food categories was higher compared with 2018 and, in general, a constant increase was seen over the years (2015) (2016) (2017) (2018) (2019) . The only exception was for 'fish and fish products', for which the number of sampling units reported in 2019 decreased, compared with 2018, although it was higher than in the previous years (2015) (2016) (2017) . The number of reporting MS has been fairly stable over the years. The number of sampling units related to animal categories fluctuated over the years, except for 'Gallus gallus (chicken)' for which the reported number of sampling units increased over the period 2015-2019. This fluctuation was very important for the category 'bovine', for which the number of sample units reported in 2019 was higher than 2018, but lower than the number of sample units reported especially in 2017, but also in 2015. For 'pigs', in the last year, there was an increase in the number of reported sample units, but it was comparable with 2017 and lower than in the two previous years (2015) (2016) . For the category 'ducks and geese', the number of flocks with monitoring data submitted to EFSA decreased compared with 2018 (even though the number of reporting countries increased), but it remained higher than the number of flocks reported in the previous years (2015) (2016) (2017) In total, 90,105 human salmonellosis cases were reported by 28 EU MS in 2019. Of these, 87,923 were confirmed cases resulting in an EU notification rate of 20.0 cases per 100,000 population (Table 9 ). This was at the same level as in 2018 (20.1 cases per 100,000 population). As in the previous year, the highest notification rates in 2019 were reported by Czechia (122.2 cases per 100,000 population) and Slovakia (91.6 cases per 100,000 population), while the lowest rates were reported by Cyprus, Greece, Ireland, Italy, Portugal and Romania (≤ 7.1 cases per 100,000 population). The proportion of domestic vs. travel-associated cases varied markedly between countries, but most of the confirmed salmonellosis cases were acquired in the EU (66.3%), whereas 7.2% reported travel outside EU and 26.5% of infections were of unknown origin (Table 8) . Considering all cases, the highest proportions of domestic cases over 95% were reported by Czechia, Hungary, Latvia, Lithuania, Malta, Portugal, Poland, Slovakia and Spain. The highest proportions of travel-related cases were reported by five Nordic countries: Finland (78.5%), Denmark (64.2%), Sweden (60.9%), Iceland (66.7%) and Norway (76.1%). Among 7,900 travel-associated cases with known information on probable country of infection, 80.3% of the cases represented travel outside EU. Turkey, Egypt, Thailand and India were the most frequently reported travel destinations outside EU (15.3%, 10.5%, 10.4% and 6.0%, respectively). In the EU, Spain and Greece were the most common travel destinations. A seasonal trend was observed for confirmed salmonellosis cases in the EU/EEA in 2010-2019, with more cases reported during summer months ( Figure 5 ). The overall EU/EEA trend for salmonellosis was stable (flat) in 2015- 2019. Finland was the only MS reporting a significantly decreasing trend (p < 0.01) in the last 5 years (2015) (2016) (2017) (2018) (2019) . An increasing trend was not observed in any MS in 2015-2019. In total, 15 MS provided information on hospitalisation. The proportion of confirmed cases with known hospitalisation information was 44.5% at the EU level. Among these, the proportion of hospitalised cases was 42.5%, which was about at the same level as in 2018. The highest proportions of hospitalised cases were reported, as in previous years, in Cyprus, Greece, Lithuania, Poland and the United Kingdom, where most of the cases were hospitalised. The high proportion of hospitalised cases is probably due to surveillance focus on severe illnesses that require hospital care. Two of these countries also reported the lowest notification rates of salmonellosis, which indicates that the surveillance systems in these countries primarily capture the more severe cases. Overall, 17 MS provided data on the outcome of salmonellosis and, among these, 11 MS reported 140 fatal cases resulting in an EU case fatality of 0.22%. Here, 46 fatal cases (32.9%) were reported by the United Kingdom. Human serovar data are described in Section 2.4.6. In total, 87,923 confirmed human salmonellosis cases were reported to TESSy in 2019. Overall, 97.3% of the number of reported human salmonellosis cases who acquired the infection in the EU (58,271; Table 9 ) were domestic (acquired within the home country) infections and 2.7% were acquired through travel in EU. Salmonella was identified overall by 23 MS in 926 FBOs that together affected 9,169 people in EU, with 1,915 hospitalised and seven deaths, as reported to EFSA. The vast majority (72.4%) of the salmonellosis FBOs were caused by S. Enteritidis. Comparing the FBOs outbreak cases (9,169) and confirmed cases human salmonellosis acquired in the EU (58,271) and also considering the estimated Source: Austria, Belgium, Cyprus, Czechia, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Lithuania, Luxembourg, Latvia, Malta, the Netherlands, Norway, Poland, Portugal, Romania, Slovakia, Slovenia, Sweden and the United Kingdom. Bulgaria, Croatia and Spain did not report data to the level of detail required for the analysis. cases with unknown travel data (0.901 9 23,309) ( Table 8 ) could suggest that overall in the EU in 2019 11.6% (9,169/79,292 9 100) of human salmonellosis cases would be reported through FBOs investigation. It is important to clarify that the case classification for reporting is different between these two databases. In TESSy, the cases reported are classified based on the EU case definition. All these cases visited a doctor and are either confirmed by a laboratory test (confirmed case) or not (probable case and classification is based on the clinical symptoms and epidemiological link). Cases that never visited a doctor are not reported to TESSy. Moreover, there may be missing probable cases in TESSy, as these data are not analysed or published and there is no incentive for reporting such cases. Information on which cases are linked to an outbreak and which not is also not systematically collected. In practice, the cases reported to TESSy are considered to be mostly sporadic cases. In food-borne outbreaks, the human cases are the people involved in the outbreak as defined by the investigators (case definition), and cases must be linked, or probably linked, to the same food source (Directive 2003/99/EC). This can include both ill people (whether confirmed microbiologically or not) and people with confirmed asymptomatic infections . Cases can be classified as confirmed or probable outbreak cases, but currently these specific classification data are not collected by EFSA. For the 265 strong-evidence outbreaks in EU in 2019 caused by Salmonella, 37.0% were caused by 'eggs and egg products ', 11 .7% by 'bakery products ', 9 .8% by 'pig meat and products thereof' and 8.7% by mixed food. Further details and statistics on the salmonellosis food-borne outbreaks for 2019 are in the FBOs chapter. Data collected in the context of Regulation (EC) No 2073/2005 on microbiological criteria Considering data for samples collected at the retail stage, Salmonella-positive samples from official controls were reported for 'minced meat and meat preparations from poultry intended to be eaten cooked ' (8.3% , 60 out of 725), 'meat products from poultry intended to be eaten cooked' (6.4%, 14 out of 218), 'fresh poultry meat' (3.5%, 89 out of 2,533), 'live bivalve molluscs and live echinoderms, tunicates and gastropods' (2.3%, 4 out of 176), 'RTE pre-cut fruits and vegetables' (2.2%, 10 out of 461), 'ice cream' (2.1%, 8 out 384), 'dried infant formulae, dried diet foods for medical purpose, and dried follow-on formulae' (1.4%, 10 out of 718), 'meat products intended to be eaten raw' (1.3%, 6 out of 466), 'minced meat and meat preparation from species other than poultry intended to be eaten cooked' (1.0%, 29 out of 2,944), 'RTE sprouted seeds' (0.8%, 1 out of 133), 'minced meat and meat preparation intended to be eaten raw' (0.6%, 1 out of 158) and 'cooked crustaceans and molluscan shellfish' (0.3%, 1 out of 330). The percentage of positive samples for category 'dried infant formulae, dried diet foods for medical purpose, and dried follow-on formulae' was strongly influenced by the subcategory 'dried infant formulae' with 10 out of 502 (1.99%) positive samples (all notified by Spain). Overviewing all poultry meat categories, there was an increase in the occurrence of Salmonella-positive samples from 0.9% reported in 2018 to 8.3% in 2019 for 'minced meat and meat preparations intended to be eaten cooked', from 0% in 2018 to 6.4% in 2019 for 'meat products from poultry intended to be eaten cooked' and from 1.8% in 2018 to 3.5% in 2019 for 'fresh poultry meat'. As defined by EU Regulation (EC) No 2073/2005, the microbiological criteria for fresh poultry meat targets positive samples for S. Enteritidis and S. Typhimurium and, according to these criteria, 29 out of the 1,345 samples (4.2%) were non-compliant for the presence of these target serovars, according to data reported by four MS. Considering data collected at production level (e.g. cutting and processing plants), 'meat products from poultry meat intended to be eaten cooked' category had a percentage of Salmonella-positive samples of 27.8% (10 out of 36). Positive samples were also reported from 'mechanically separated meat' (9.2%, 6 out of 65), 'fresh poultry meat' (2.5%, 6 out of 292), 'minced meat and meat preparations from poultry intended to be eaten cooked ' (2.1%, 16 out of 759), 'cheese, butter and cream made from raw or low-heat treated milk' (0.7%, 8 out of 1,114), 'meat products intended to be eaten raw' (0.6%, 3 out of 482), 'minced meat and meat preparations from species other than poultry intended to be eaten cooked ' (0.4%, 15 out of 3,399) . As also pointed out in previous years, data collected in the context of Regulation (EC) No 2073/2005 on microbiological criteria, which could serve the purpose of trend observation, were scarce and unrepresentative of the EU situation, because few data were reported for the specified food categories and regardless of the sampling stage. Results are summarised in Figure 6 . As regards Salmonella PHC monitoring data from pig carcases collected at the slaughterhouse before chilling, 19 MS provided data. Four MS (Cyprus, Ireland, Slovakia and the United Kingdom) reported official control data only; seven MS (Austria, Denmark, Germany, France, Latvia, Portugal and Slovenia) self-monitoring data only, from FBOp (Table 10 ) and eight MS (Belgium, Bulgaria, Estonia, Italy, Malta, the Netherlands, Poland and Spain) both samplers' data. Considering pig carcass data sent by the latter eight MS, the percentage of Salmonella-positive single samples from carcases was 3.88% (N = 15,745) for samples collected by CA and 1.11% (N = 35, 765) for samples collected by FBOp. For Belgium, Estonia, Italy, the Netherlands, Poland and Spain, the percentage of positives based on official controls was significantly higher than that from self-monitoring. Considering all Salmonella PHC monitoring data from pig carcases sent by the 19 MS, the percentage of Salmonella-positive samples from carcases based on official controls was 3.15% ( N = 22,271) and was significantly higher than that based on self-monitoring (1.51%, N = 111,939) . Comparing these data with the data collected in 2018, increase in prevalence was reported for samples collected by CA ( Finland, Sweden and Norway, which are countries with special guarantees in relation to Salmonella on pig carcasses (according to Regulation (EU) No 853/2004), reported the following monitoring results: Finland five positive samples out of 6,507 tested (food business operator sampling), Sweden one positive out of 5,935 official control samples and Norway zero positive out of 3,314 official control samples tested. As regards official control Salmonella PHC monitoring data from other animals than pigs, results from chilled carcases of broilers were reported by six MS (Bulgaria, Cyprus, Estonia, Latvia, Spain and Sweden) and only one single MS (Spain) provided data from chilled turkey carcases. The overall proportion of Salmonella-positive broiler carcase samples was 9.8% (99 positive samples out of 1,012 tested carcases), whereas for turkey carcases it was 22% (11 positive samples out of 50 tested carcases). Additionally, Sweden, which has special guarantees in relation to Salmonella on broiler carcasses, reported zero positive out of 1,866 official control samples tested. Finland, Sweden and Norway, which are countries with special guarantees in relation to Salmonella on pig carcasses (according to Regulation (EU) No 853/2004), reported the following monitoring results: Finland five positive samples out of 6,507 tested (food business operator sampling), Sweden one positive out of 5,935 official control samples and Norway zero positive out of 3,314 official control samples tested. Monitoring data reported from food samples, which do not fit with the criteria described in the previous paragraphs, are described by merging investigations from all the monitoring and surveillance activities, from all the sampling stages (retail, slaughterhouse, processing, border inspection activities and unspecified) and from all the sampling units (single and batch). Table 11 summarises the reported occurrence of Salmonella in the most important food categories for the year 2019 and for the four year-period 2015-2018. A distinction is made between RTE and non-RTE food including fresh meat. For 2019, 66,113 RTE and 191 ,181 non-RTE food sampling units were reported from 21 and 25 MS with 0.27% and 1.52% positive samples, respectively. Within the category of RTE food samples, positive samples were from divers food products; 'meat and meat products', 'milk and milk products', 'fruits, vegetables and juices', 'fish and fishery products', 'spices and herbs ', 'salads', ' other processed food products and prepared dishes', 'cereals and nuts', 'infant formulae and follow-on formulae', 'other food' and 'cocoa and cocoa preparations, coffee and tea', with the percentages of positive samples ranging from 0.04% in 'fish and fishery products' to 1.63% in 'infant formulae and follow-on formulae'. Within the category of non-RTE food the highest percentage of positive samples was reported for 'fresh meat from broilers' (7.66%), 'fresh meat from turkeys' (3.62%), 'infant formulae' (1.78%) and 'other fresh meat ' (1.60%) . The number of sampling results for 'meat and meat products' was high both for RTE and no-RTE food with 0.55% and 1.66% positive samples, respectively. In the following descriptive analyses, food categories include RTE food and non-RTE food. A summary of results from the major meat and meat product categories and the sampling points is in Figure 7 , considering all sampling units (single and batch). Considering the entire production chain for meat and meat products, the highest percentages of Salmonella-positive samples were found for 'Fresh broiler meat' and 'Fresh turkey meat ' (respectively, 7 .66 and 5.38%). Salmonella-positive samples of 'Fresh broiler meat' were collected mainly at the slaughterhouse, while for 'Fresh turkey meat', positive samples were both from slaughterhouses and cutting plants. For the other categories, 2.59% of the 'Fresh poultry meat other than broiler and turkey' samples were Salmonella-positive, and these samples were reported mainly at the processing plants, as for 'RTE minced meat, meat preparations and meat products from pig meat'. Austria, Bulgaria, Denmark, Germany, Italy, Poland, Portugal and Slovakia reported monitoring results for a total of 4,493 tested table egg sampling units and 6 (0.13%) were Salmonella-positive: Austria, Germany and Italy found two positives each. As regards egg products, the same MS and Croatia, Cyprus, Lithuania and Spain reported data and overall two (0.16%) of the 1,246 sampling units collected were Salmonella-positive and reported by Austria and Croatia. Results of 663 live bivalve molluscs sampling units were reported. Two (0.3%) were positive for Salmonella. Of the 7,462 units of fruit and vegetables tested, 0.10% were Salmonella-positive. Poultry monitoring data according to the Salmonella national control programmes Breeding flocks of Gallus gallus In total, 24 MS and three non-MS reported Salmonella NCP data from breeding flocks of Gallus gallus. Luxembourg and Malta do not have such flocks, whereas Hungary and Lithuania have flocks, but did not report any data. In the EU in 2019, Salmonella was found in 340 ( Red vertical bars indicate the target to be reached, which was fixed at 1% for all poultry populations with the exception of laying hens for which it was 2% for all MS with the exception of Poland, for which it was 3.5%. Luxembourg met the target in laying hens (having less than 50 flocks with one positive for target serovars). (22) reported, additionally to the overall merged results, separate investigational results from the CA and the FBOp, from their broiler flocks. Four MS did not comply; France, Italy and the Netherlands only reported overall merged results and Croatia provided separate data for CA sampling only. Considering the data from the MS that reported separate results from both CA and FBOp, the prevalence of Salmonella target serovar-positive flocks was, respectively, 1.60% (5,013 tested flocks by the CA) and 0.06% (241,344 tested flocks by FBOp). At the EU level, the prevalence of Salmonella target serovar-positive broiler flocks obtained by the CA was significantly higher than that obtained from the FBOp' self-monitoring results. The same finding was also evident individually for Bulgaria, Czechia, Germany, Poland and Spain. For the remaining reporting MS, the differences between the results of both types of sampler were not significant or the sample sizes for one or both samplers were too low for analyses, or data were missing (Table 12) Breeding turkey flocks For breeding turkeys, 13 MS and two non-MS reported Salmonella NCP data. Although Hungary had breeding flocks of turkeys, they did not report such data. Salmonella was found in 85 (5.19%) of the 1,637 flocks tested, compared with 3.85% in 2018 and 2.63% in 2017. This increase is related to the marked increase of Salmonella-positive breeding turkey flocks reported by Spain (30.43% of positive flocks in 2019 and 8.73% in 2018) . In 2019, the prevalence of flocks positive for either of the two target Salmonella serovars was 0.30% (N = 5) compared with 0.47% and 0.50% in 2018 and 2017, respectively. The five target Salmonella serovar-positive flocks were all positive for S. Typhimurium ( Figure 16 ). Therefore, 5.9% (5 of 85) of reported Salmonella-positive breeding turkey flocks were positive for S. Typhimurium. All reporting MS met the reduction target. Salmonella NCP monitoring data for turkey breeding flocks must be reported separately for investigations performed by CA and by FBOp, in addition to the overall merged results. Three MS (Croatia, France and Italy) did not comply with this reporting requirement, whereas 10 MS did (Table 13 ). The prevalence of Salmonella target serovar-positive flocks based on official control samples and on self-monitoring conducted by the FBOp were 0% (N = 544) and 0.28% (N = 721), respectively. All samples collected by CA and FBOp were negative, except for two isolates collected by FBOp and reported by the United Kingdom. For fattening turkey flocks, 22 MS and three non-MS provided data. Hungary and Lithuania had flocks of fattening turkeys but did not report any data. In the EU in 2019, Salmonella was found in 2,241 or 5.84% of fattening turkey flocks compared with 6.32% in 2018. The EU prevalence of flocks positive for either of the two target Salmonella serovars was 0.24% (N = 93) (Figure 17 ), compared with 0.34% in 2018. Therefore, 4.1% (93 of 2,241) of reported Salmonella-positive fattening turkey flocks were positive for either of the two target serovars. In total, 11 MS and two non-MS reported no Salmonella target serovar-positive flocks. Only Belgium did not meet the reduction target ( Figure 8 ) of 1%. Belgium reported six S. Typhimurium-positive flocks in 2019, similar to 2018. The EU flock prevalence was higher for S. Typhimurium (0.18%) than for S. Enteritidis (0.06%), with 56.99% of positive flocks for both serovars being reported by France, similar to the previous years. Salmonella NCP monitoring data for turkey fattening flocks must be reported separately for investigations performed by CA and by FBOp, in addition to the overall merged results. Eighteen MS complied with the requirement, while four MS (Croatia, France, Italy and the Netherlands) did not send separate data from CA and FBOp. Considering all data sent, the percentages of target Salmonella-positive flocks were, respectively, 0.64% (corresponding to 787 tested flocks) by the CA and 0.09% (corresponding to 22,299 tested flocks) by FBOp. The EU prevalence of Salmonella target serovar-positive flocks based on official control samples (CA) was significantly higher than the FBOp' self-monitoring results. The same finding was also evident for data reported by Germany and Spain, like in 2018. In contrast, for the other MS that reported separate data from both CA and FBOp there were no significant differences between the two sampling categories (Table 14) . Comparing data collected in 2019 with those reported in 2018, the prevalence of fattening turkey flocks positive for target serovars based on official control samples in 2019 (0.64%) was lower than the prevalence in 2018 for the same monitoring approach (2.07%). The trends in the EU flock prevalence of Salmonella target serovars in poultry flocks since the implementation of the EU-wide NCP 2007-2019 are displayed in Figure 18 . In the supporting information to this report ('Salmonella poultry outcome trends analyses'), the EU percentages of positive flocks for Salmonella, target and non-target Salmonella serovars and S. Enteritidis over time are shown and compared for each poultry population covered by the NCP. Moreover, figures show the modelling of prevalence trends of Salmonella and target Salmonella serovars in poultry flocks. Detailed outputs of trend analyses (at subject level and at population level) are reported. The apparent discrepancy between the percentage of positive flocks (both for target Salmonella serovars and for Salmonella, described above) and the estimated prevalence shown below is due to the fact that the first value is the ratio between all positive over all tested flocks, whereas the estimated prevalence is obtained by modelling the ratio between positive and tested flocks of each country, taking into account the variability among MS. As observed during previous years, S. Enteritidis was by far the most common target serovar reported in 2019 in breeding flocks of Gallus gallus. Moreover, the temporal trend of S. Enteritidis in breeding Gallus gallus flocks was very similar to trends of the Salmonella target serovars, of Salmonella and of non-target serovars. The data used to model the trend in EU Salmonella flock prevalence for target serovars in breeding Gallus gallus for the period 2007-2019 were from 26 MS. Two MS (Estonia and Latvia) reported no single flock positive for target serovars during this entire period of implementation of NCP. Since the beginning of the NCP, there has been an overall decreasing trend for the prevalence of breeding Gallus gallus flocks positive for target serovars (Figures 18 and 20 [1.23; 2.53] in 2019. This latter estimated prevalence was not significantly different from those of the previous 2 years, but it was significantly higher than the minimum prevalence estimated in 2015 (p-value = 0.0701). Focusing the trend analysis modelling on the last 5 years confirmed that the estimated Salmonella flock prevalence in Gallus gallus breeding flocks has increased significantly and was significantly higher in 2019 compared with 2015 (p-value = 0.042). As observed during previous years in laying hen flocks, the temporal trends for S. Enteritidis, for target serovars, for non-target serovars and for Salmonella were similar, because of its dominance, even though the prevalence differed. Data used to model the trend in the EU Salmonella flock prevalence for target serovars in laying hen flocks over the period 2008-2019 were from all MS. No MS reported 0% prevalence for target serovars during this period. Since the beginning of the NCP, there has been a decreasing overall trend for the prevalence of flocks positive for target serovars (Figures 18 and 20) ; the prevalence estimated by modelling was 3.71% CI 95 [2.43; 5.61] The estimated EU Salmonella prevalence in laying hen flocks was 7.36% CI 95 [4.5; 11.83] in 2008 and decreased to 2.07% CI 95 [1.34; 3.19] in 2014, with a steep downturn. During the following years, it increased and reached 3.44% CI 95 [2.33; 5 .06] in 2019. In 2019, the estimated Salmonella prevalence in laying hen flocks was not significantly different compared with the previous 2 years, but it was different compared with 2014, when the estimated prevalence reached the minimum value seen to date (p-value = 0.0468). Focusing the trend analysis modelling on the last 6 years confirmed the prevalence of Salmonella in EU laying hen flocks has increased significantly, reaching a significantly higher prevalence in 2019 than in 2014 (p-value = 0.075). As observed during previous years, in broiler flocks, the temporal trend of S. Enteritidis mimics that of the target serovars, because of its dominance. Moreover, the temporal trends of Salmonella and non-target serovars are similar. The data from 27 MS were used to model the trend in the EU Salmonella flock prevalence for target serovars in broilers flocks for the period 2009-2019. Finland reported no broiler flocks positive for Salmonella target serovars during this entire period, whereas Estonia notified its first positive flock for target serovars in 2019. From the beginning of the NCP, the flock prevalence for target serovars estimated by the model steeply decreased in the first time interval (until 2011) and then further decreased (Figures 18 and 20 [1.06; 3.47] in 2019. This increase was probably related to the increased reporting of non-target serovars, in particular S. Infantis, the most frequently reported serovar from broiler flocks. Nevertheless, the estimated EU prevalence of Salmonella-positive broiler flocks in 2019 was not significantly different to that of the previous 2 years or in 2015, when the estimated prevalence reached the minimum value. Focusing the trend analysis modelling on the last 5 years, the prevalence of Salmonella in broiler flocks was confirmed as increasing. However, the estimated Salmonella prevalence in EU broiler flocks in 2019 was not significantly higher than in 2015. In breeding turkey flocks, the temporal trends of S. Enteritidis and target serovars were similar, although with different prevalence, whereas the trends of Salmonella and non-target serovars overlapped. The data used to model the trend in EU Salmonella flock prevalence for target serovars in breeding turkey flocks for the period 2010-2019 were from 15 MS. Six MS reported no breeding turkey flocks positive for target Salmonella serovars over this entire period. The remaining MS had, from time to time, some positive flocks. The prevalence of Salmonella target serovar-positive breeding turkey flocks fluctuated for the entire period around an estimated value of 0.35% CI 95 [0.28; 0.44]. After an initial fluctuation of the EU prevalence of Salmonella-positive breeding turkey flocks from 7.7% CI 95 [3.36; 16.72] in 2010 to 1.33% CI 95 [0.69; 2.54] in 2016, when the estimate prevalence reached the lowest value seen in the entire study period, the estimated prevalence increased over time to 5.02% CI 95 [2.10; 11 . 51] in 2019. This estimated prevalence in 2019 was not significantly different from the previous 2 years, but it was significantly higher than the estimated prevalence in 2016 (p-value = 0.0468). Focusing the trend analysis modelling on the last 4 years results confirmed that the prevalence of Salmonella in breeding turkey flocks increased significantly, so the prevalence in 2019 was significantly higher than in 2016 (p-value = 0.0249). This increase was probably related to the increased reporting of non-target serovars. In fattening turkey flocks, the temporal trends of S. Enteritidis and the target serovars were different. Conversely, the temporal trends of Salmonella and non-target serovars were very similar. The data used to model the trend in the EU Salmonella flock prevalence for target serovars in (Figures 18 and 20) . Nevertheless, there were no significant differences in the estimated prevalence of the Salmonella target serovars in EU fattening turkey flocks in the last three years. For this poultry category, after an initial fluctuation of the EU prevalence of Salmonella-positive flocks from 5.9% CI 95 [1.44; 3.64] in 2010 to 2.1% CI 95 [1.06; 1.44] in 2015. In this last year, the estimate prevalence reached the lowest value and then it increased to 3.17% CI 95 [1.58; 6 .23] in 2019. This increase was related to the increased reporting of non-target serovars. Nevertheless, the prevalence in 2019 was not significantly different from those of the previous 2 years or from the minimum estimated prevalence in 2015. Focusing the trend analysis modelling on the last 5 years, a significant increasing trend of Salmonella prevalence in fattening turkey flocks was confirmed. However, the estimated prevalence of Salmonella in fattening turkey flocks in 2019 was not significantly higher than in 2015. Six MS (Estonia, Italy, Lithuania, Latvia, Poland and Sweden) and one non-MS (Norway) reported monitoring data on Salmonella flock prevalence in ducks and geese for 2019. Of 8,343 flocks, 1.07% were positive for Salmonella, whereas 0.47% were positive for S. Enteritidis and/or S. Typhimurium. In total, 15 MS and two non-MS (Norway and Switzerland) reported data on Salmonella prevalence in pigs. Overall, 36.02% of the 66,624 reported sample units were positive for Salmonella. Among these, 72.3% (N = 48, 184) were collected at the slaughterhouse and 49.07% were positive. In cattle, based on data reported by 15 MS and four non-MS at the EU level, the overall prevalence of Salmonella-positive samples was 3.34% with 2,898 positive samples, whereas the prevalence of positive samples at the slaughterhouse was 7.76%. The overall prevalence of Salmonella-positive units in 'animal and vegetable derived feed' supplies in 2019 in the EU was 2.46% of 29,111 reported units. In compound feed (the finished feed for animals), the prevalence of Salmonella-positive units in 2019 was 1.64% of 15,812 tested samples for poultry, 0.92% of 3,124 tested samples for cattle and 1.23% of 5,032 tested samples for pigs. As for feedingstuffs for animals other than pigs, cattle and poultry, the prevalence of Salmonella-positive units in EU was 1.32% out of 9,686 tested samples. The prevalence of Salmonella-positive sampling units for pet foods was 9.4% out of 3,448 tested samples. For humans, information on Salmonella serovars was available for 90.2% of the total number of confirmed cases (79,300 cases out of 87,923) from 27 MS (Bulgaria did not report case-based serovar data), Iceland and Norway. Data include all cases reported with serovar information regardless of the travel status. As in previous years, the three most commonly reported Salmonella serovars in 2019 were S. Enteritidis (50.3%), S. Typhimurium (11.9%) and monophasic S. Typhimurium (1,4,[5] ,12:i:-) (8.2%), representing 70.3% of the 79,300 confirmed human cases with known serovar in 2019. The proportion of these three serovars was at the same level as in 2017 and 2018, as well as S. Infantis, which was the fourth most commonly reported serovar (Table 15 ). The fifth most common serovar S. Newport decreased by 20.0% compared with 2018. Serovar S. Mikawasima increased by 92.1% and 137.1% compared with 2018 and 2017, respectively. This serotype entered the top 20 list in 2019 and replaced serovar Brandenburg. To estimate the impact of the Salmonella infections acquired at the EU level, serovar data were analysed for domestic and travel-associated cases in which the probable country of infection was an EU MS. Information on Salmonella serovars with travel data was available from 24 MS, representing 74.8% of cases with known serovar data in 2019. Most cases (88.1%) with known data on serovar and travel were infected within the EU. Among the travel-related cases, the most frequently reported travel destinations were Spain (28.9%), Greece (14.5%), Poland (9.9%), Italy (7.5%) and Croatia (6.9%), as in 2017-2018. From reported cases of human salmonellosis acquired in the EU, S. Enteritidis dominated and almost two in three (61.6%) of the reported cases were infected by this serovar. Together with S. Typhimurium and monophasic S. Typhimurium 1, 4, [5] ,12:i:-, these three serovars represented 78.3% of the confirmed human cases acquired in the EU in 2019 (Table 16 ). S. Enteritidis cases were predominantly (93.1%) infected within EU. The proportion of S. Enteritidis was about at the same level as in 2017-2018. The proportion of S. Typhimurium and its monophasic variant strains 1,4,[5],12:i:slightly decreased and S. Infantis and S. Derby remained at the same level as in 2018. Among the cases acquired in the EU, S. Newport has alternated between fifth and sixth places among the top six serovars. A seasonal trend was observed for confirmed S. Enteritidis infections acquired in the EU in 2010-2019, with more cases reported during summer months. The trend from 2015 to 2019 was stable (flat) ( Figure 21) . Malta was the only MS reporting a significantly decreasing (p < 0.01) trend of S. Enteritidis infections acquired within the EU over the last 5 years (2015) (2016) (2017) (2018) (2019) . A significant increasing trend was not observed in any MS for the last 5 years. Descriptive analyses were made from food and animal data from 2019 for the five Salmonella serovars that were most frequently reported from cases of human salmonellosis acquired in the EU (Table 16) . These top five serovars were S. Enteritidis, S. Typhimurium, monophasic S. Typhimurium, S. Infantis and S. Derby. Only isolates related to food-producing animals and specific food matrices were aggregated into the following categories for further analyses: broiler flocksbroiler meat, laying hen flockseggs, fattening turkey flocksturkey meat, pigspig meat and cattlebovine meat. In total, 17,176 Salmonella serotyped isolates were reported that matched the mentioned inclusion criteria (Table 17) . Source(s): Austria, Czechia, Denmark, Estonia, Finland, Germany, Greece, Hungary, Ireland, Italy, Latvia, Malta, the Netherlands, Portugal, Slovakia, Spain, Sweden and the United Kingdom. Belgium, Bulgaria, Cyprus, Croatia, France, Lithuania, Luxembourg, Poland, Romania and Slovenia did not report data to the level of detail required for the analysis. Hence, more than 70% of these serotyped isolates were from broilers (both animals and food), pig sources accounted for about 12% of the serotyped isolates, laying hens and turkeys about 7% each (but for both species the vast majority of the isolates were from the animal sources), whereas serotyped isolates from cattle made up about 1% of the serotyped isolates. The top-five serovars responsible for human infections were distributed as follows among the serotyped isolates (17,176) from these food-animal sources: S. Infantis accounted for 29.7% of them, S. Enteritidis 6.9%, monophasic variants of S. Typhimurium 4.5%, S. Typhimurium 3.9% and S. Derby 3.7%. The Sankey diagram ( Figure 22 ) illustrates how the EU top five Salmonella serovars in human salmonellosis cases acquired in the EU are associated with the most important animal species. S. Enteritidis was primarily associated with broiler sources (67.8% of the S. Enteritidis isolates were from broiler flocks and meat) and secondly with layers (26.7%). S. Typhimurium was mainly associated with pig, broiler and layer sources, respectively, 42%, 34.8% and 13.5%. Monophasic S. Typhimurium was associated mainly with pig (72.1%) and secondly with broiler (17.1%) sources. S. Infantis was mostly related to broiler sources (93.1%). S. Derby was primarily associated with pig (72%) and secondly with turkey (19.8%) sources. To interpret these data, it is important to be aware that the distribution of the serotyped isolates among the different sources is very heterogeneous in terms of number of isolates per species, as detailed above. The Sankey diagram in Figure 23 illustrates how the EU top five Salmonella serovars in human salmonellosis acquired in the EU were proportionally reported by the reporting MS from specified foodanimal sources mentioned, in 2019. In this context too, the number of serotyped isolates reported by each MS is very heterogeneous which must be considered when interpreting the following data. Twenty-seven MS reported the top-five Salmonella serovars from the above sources. S. Enteritidis was widely reported by most MS, even though Poland accounted for the greatest percentage (49.6%) of the isolates, followed by France that reported 13.6% of the S. Enteritidis. Similarly, S. Typhimurium and monophasic S. Typhimurium isolates were reported by all MS, but the highest percentage of both serovars was reported by France, accounting for 29.4% and 27.8%, respectively. S. Infantis isolates were mostly reported by Italy (50.6%), whereas S. Derby was mostly reported, in decreasing order, by the United Kingdom (22.8%), Denmark (20.7%), Italy (13.5%) and France (11.8%). The left side of the diagram shows the five most reported Salmonella serovars from human salmonellosis cases acquired in the EU: S. Enteritidis (pink), S. Typhimurium (green), monophasic S. Typhimurium (yellow), S. Infantis (blue) and S. Derby (violet). Animal and food data from the same source were merged: 'broiler' includes isolates from broiler flocks and broiler meat, 'bovine' includes isolates from bovines for meat production and bovine meat, 'pig' includes isolates from fattening pigs and pig meat, 'turkey' includes isolates from fattening turkey flocks and turkey meat and 'layers' includes isolates from laying hen flocks and eggs. The right side shows the five sources considered (broiler, bovine, pig, turkey and layers). The width of the coloured bands linking sources and serovars is proportional to the percentage of isolates of each serovar from each source. Figure 24 shows the percentages of the EU top five Salmonella serovars in human salmonellosis acquired in the EU and reported from specified food and animal matrices, by food-animal category with isolates. Considering all poultry sources, S. Infantis was the most reported serovar, accounting for 5,043 of 14,905 (33.8%) serotyped isolates, followed by S. Enteritidis (1,156; 7.8%). S. Infantis was massively reported for broiler matrices, both from animals (36.3% of all serotyped isolates) and from food matrices (49.1%) . It was also present, but to a lesser extent, in turkey flocks (13.3% of all serotyped isolates), turkey meat (13.9%) and in layer flocks (10.2%) ( Figure 22 ). More than 50% of the S. Infantis isolated in 2019 from broilers was reported by Italy. Looking in detail at the serovar data from broiler flocks and focusing on the four MS that reported more than 75% of all serotyped isolates from this source (Italy 34.78%, France 19.4%, the United Kingdom 14.14% and the Netherlands 9.46%), the situation, in terms of reporting of S. Infantis, was very heterogeneous. Italy and the Netherlands reported 64.9% and 42.9%, respectively, of their serotyped isolates as The left side of the diagram shows the five most reported Salmonella serovars from human salmonellosis cases acquired in the EU: S. Enteritidis (pink), S. Typhimurium (green), monophasic S. Typhimurium (yellow), S. Infantis (blue) and S. Derby (violet). The right side shows the reporting MS. The width of the coloured bands linking MS and serovars is proportional to the percentage of isolates of each serovar reported by each MS. S. Infantis. In contrast, the United Kingdom and France reported, respectively, none and less than 1% of the isolates belonging to this serovar from broiler flocks, whereas these two countries frequently reported other serovars in broiler flocks (e.g. S. Livingstone, S. Montevideo, S. Mbandaka, S. Kedougou). Irrespective of the situation in broilers, for most of the reporting MS, S. Infantis was the most common serovar reported from broiler meat (about one in two isolates were from this source). S. Enteritidis accounted for 50% of all Salmonella isolates serotyped from eggs and 24.8% of the serotyped isolates from layer flocks. It also accounted for 25.2% of serotyped isolates from broiler meat. More than 50% of S. Enteritidis isolated in 2019 from these sources was reported by Poland. For the other sources, a very small number of S. Enteritidis isolates was reported. For S. Typhimurium and its monophasic variants, they showed similar patterns, with S. Typhimurium accounting for 12.7% and 14% of the serotyped isolates from pig herds and pig meat and its monophasic variants accounting for 28.8% and 26.6% of serotyped isolates from these matrices, respectively. For bovine meat, 31.8% and 13.6% of serotyped isolates were S. Typhimurium and its monophasic variants, respectively. Finally, S. Derby accounted for 24.1% of all the serotyped isolates from pigs and 21.3% of all serotyped pig meat isolates, while the percentages from turkey matrices were considerably lower (11.6% and 2.1% of all serotyped isolates from turkeys and turkey meat, respectively). Among the remaining animal/food categories, this serotype was rarely reported. The percentages were calculated based on the total number of isolates serotyped for each of the five animal/ food categories (bovine, broiler, layers, pig and turkey). The values at the top of each box are the numbers of Salmonella serovar isolates and the numbers in parentheses are the number of reporting MS, for animal matrices (grey) and food matrices (black). Each plot shows the percentage of isolates belonging to the reported serovar out of the total number of serotyped isolates. Figure 24 : Percentages of the EU top-five Salmonella serovars in human salmonellosis acquired in the EU and reported from specified food-animal categories, by food-animal category with isolates, EU, 2019 2.5. Salmonellosis remains the second most common zoonosis in humans in the EU after campylobacteriosis. The previous decreasing trend of confirmed cases has stabilised since 2014 and, in 2019, the number of reported confirmed human cases and the EU notification rate were at the same level as in 2018. In 2019, only one MS (Finland) reported a decreasing trend in the last 5 years, whereas all other MS reported stable, flat trends during 2015-2019. S. Enteritidis infections that were acquired within the EU also stabilised in 2015-2019, after several years of an increasing trend. S. Enteritidis infection is predominantly acquired in the EU, more frequently than other serovars. A large European multi-country outbreak of S. Enteritidis associated with contaminated eggs from Poland was confirmed in 14 EU/EEA countries in 2016. Poland implemented control measures and the cases declined in 2017 but started to increase again at the end of the same year. It is likely that this multi-country outbreak had already existed since 2012 and was still ongoing during 2019. Since 2016, the number of confirmed S. Enteritidis human cases has steadily increased and cases have been confirmed in 18 EU/EEA countries, with the most recent epidemiological update reported in February 2020 EFSA and ECDC, 2017a ,b,c, 2018a ,b, 2020a . In each year from 2016 to 2018, outbreak cases peaked in September, with large waves of cases reported between late spring and early autumn. Such a large seasonal increase was no longer observed in 2019. In this context, it is noteworthy that 54.7% of the S. Enteritidis-positive breeding flocks of Gallus gallus were reported by Poland. All MS except Bulgaria, Croatia, Ireland, Poland and Slovenia met the flock prevalence target of maximum 1%. In laying hens, 80.9% of S. Enteritidispositive flocks were reported by six MS (France, Germany, Italy, the Netherlands, Poland and Spain) and Bulgaria, Croatia, Poland and Spain did not meet their reduction target, which was 2% flocks remaining positive for all MS except for Poland for which it was 3.5%. The three most commonly reported serovars S. Enteritidis and S. Typhimurium (including monophasic variants) accounted for over 70% of human cases acquired in the EU. S. Infantis has been consistently the fourth most frequently reported serovar in the domestically acquired and travelassociated human infections. As in previous year, serovars S. Derby and S. Newport were reported in almost equal numbers, being the fifth and sixth most frequently reported serovars in 2019. The EU trends for these six serovars have been stable in the last 5 years between 2015 and 2019. Notification rates for salmonellosis in humans vary between MS, reflecting variations in, for example, quality, coverage and disease-severity focus of the surveillance systems, practices in sampling and testing, disease prevalence in the food-producing animal population, food and animal trade between MS and the proportion of travel-associated cases. The hospitalisation rate varied from 23.5% to 96%. Countries reporting the lowest notification rates for salmonellosis had the highest proportions of hospitalisation, suggesting that the surveillance systems in these countries are focused on the most severe cases and underlining the variation in national surveillance systems. Monitoring results for Salmonella contamination in food is in large part based on data collected in the context of Regulation (EC) No 2073/2005, which guarantees a certain level of harmonisation in terms of food categories considered, analytical methods used and sampling points. In this specific context, poultry meats (including fresh meat, minced meat, meat preparations and meat products) have been identified as the food categories for which Salmonella was most frequently reported, even though Salmonella national control programmes in poultry at the primary production level have been specifically implemented for several years (Antunes et al., 2016) . Moreover, looking at FBOs, as in the previous years, egg and eggs products ranked first of food vehicles causing strong-evidence salmonellosis FBOs. This matrix was implicated in 37% of such outbreaks. Monitoring results for Salmonella contamination in RTE and non-RTE food were also described for samples collected according to an 'objective' sampling strategy. The overall percentages of Salmonella-positive samples for RTE and non-RTE food were 0.27% and 1.52%, with 'meat and meat products' reported to have 0.55% and 1.66% positive samples, for the two categories, respectively. The findings of Salmonellacontaminated RTE food is of concern because it poses a direct risk to the consumer. Another food category reported both within RTE and non-RTE food was 'infant formulae' with 1.63% and 1.78% positive samples for RTE (N = 123) and non-RTE products (N = 562). These findings merit attention because this product is intended for young, susceptible children. Outbreaks due to contaminated infant formula are reported and during 2019 a multi-country outbreak associated with infant formula contaminated by Salmonella Poona involved three MS (France, Belgium and Luxembourg) affecting 32 infants and young children (EFSA and ECDC, 2019a) . Analytical evidence linked that outbreak to another S. Poona outbreak relating to the same facility in 2010-2011, indicating a persistent source of contamination (Jones et al., 2019) . Control programmes in poultry at primary production level focus on serovars of particular relevance for public health (i.e. S. Enteritidis and S. Typhimurium), whereas data collected from poultry food categories refer to the genus Salmonella, regardless of serovar (with the only exception being fresh poultry meat). Trends for the target Salmonella serovar-positive flocks have been quite constant (flat) over recent years for almost all poultry categories. The number of MS that did not meet the annual targets for the different poultry categories decreased in 2019 compared with 2018. Combining all these data, it seems that efforts aimed at control of S. Enteritidis and S. Typhimurium in poultry flocks have been partially effective. However, if we look at trends of Salmonella flock prevalence in poultry populations over the last 4-6 years, a significant increase was noted in breeding Gallus gallus, laying hens and breeding turkeys. These increasing trends for Salmonella can be partly explained by the emerging spread of certain clones in the different animal populations e.g. S. Infantis (EFSA BIOHAZ Panel, 2019). S. Infantis is overall by far the most frequently reported serovar in broilers and their derived carcases. When considering the four countries reporting more than 75% of all reported serotyped isolates from the broiler source (in decreasing order: Italy, France, UK and the Netherlands), their reports on S. Infantis are very diverse. For Italy and the Netherlands, most of reported serovars from broiler flocks were S. Infantis, (64.9% and 42.9%, respectively), while UK and France reported almost no S. Infantis, but reported mainly isolates of S. Montevideo and S. Livingstone (France) and S. Kedougou and S. Mbandaka (UK). Other countries reporting a proportion of S. Infantis higher than 50% of the serotyped isolates, from broiler flocks, were Austria (75.2%), Slovakia (62.5%), Spain (60%) Croatia (54.1%) and Romania (53.6%). Caution is warranted when interpreting these data because the reporting of this serovar, as the other non-target serovars, is not mandatory for broilers (reporting bias). Still, irrespective of MS-specific reports for broiler flocks, S. Infantis was the most common serovar reported from broiler meat (about one in two isolates from this source), for most reporting MS. The recent epidemiological success of this serovar can be associated with its ability to enter and persist along the poultry food chain and this represents a growing risk for public health . Moreover, the worldwide emergence of S. Infantis clones with enhanced epidemiological fitness has been attributed to the acquisition of a conjugative megaplasmid providing the bacteria with new resistance features, virulence-associated properties, high tolerance to disinfectants and resistance to heavy metals (Garc ıa-Soto et al., 2020). S. Kentucky is another serovar that has undergone emergent spread both in humans and in the food chain, especially some clones (e.g. ST 198) characterised by resistance to multiple antimicrobials including some critically important ones (e.g. fluoroquinolones) (EFSA and ECDC, 2020b). This scenario has led France to include S. Kentucky among the regulated serovars in poultry, at national level. It has been hypothesised that the recent spread of some serovars could be partly associated with to the regulatory policy addressing a limited selection of target serovars in the different poultry populations and that this surveillance approach could have allowed the expansion of other serovars that have found new niches in the poultry industry (EFSA BIOHAZ Panel, 2019) . As recently proposed by EFSA (EFSA BIOHAZ Panel, 2019) , an alternative approach based on an 'all serovars' target for breeding flocks could be more effective. Moreover, this extended approach could be valuable in limiting the spread of emerging or re-emerging serovars showing epidemic potential. Eventually this extended approach could have a direct effect in reducing the Salmonella prevalence in foodstuffs. However, this new extended target could be rather challenging for many MS and a good compromise could be a dual prevalence target for 'all serovars' and for 'the selected/high priority serovars' with different control measures and containment methods based on the identified serovar. Anyway, in 2019, S. Enteritidis remained the most common serovar in humans causing most FBOs. The flock prevalence of breeding Gallus gallus and laying hens was highest for S. Enteritidis, whereas for broilers the prevalence was at the same level as S. Typhimurium. These data indicate that it is important to prioritise attention on this serovar to avoid underestimating the risk posed by S. Enteritidis, especially in laying hens, where its true prevalence is likely to be substantially underestimated (EFSA BIOHAZ Panel, 2019) , as this would have a direct effect on the control of most Salmonella cases in humans (De Cesare, 2018) . Salmonella was found in 2.46% tested units of 'animal and vegetable derived feed' supplies and 1.64% of the compound feed for poultry. These data demonstrated that feed remains a putative source of infections for poultry populations and finally for humans, although target serovars are not common in feed, but unfortunately, as for many other categories, prevalence data from feed are not representative of the EU situation since the number of serotyped isolates are very limited and are reported from few countries that vary over the years. According to the legislation, the surveillance of Salmonella along the food chain is based on controls implemented by FBOp and CA. When there were data available to compare the Salmonella prevalence identified by the two systems, the percentage of Salmonella-positive units reported by official controls was generally higher than that reported in the context of own check controls by FBOp. These differences can be related to the fact that the CA generally focuses their samplings on the most problematic herds/slaughterhouses (risk-based approach). Anyway, this situation deserves attention as in the EU Salmonella surveillance at all levels of the food chain is primarily based on the controls conducted by FBOp. They are the cornerstone of the strategy and their control systems must be as effective as possible to guarantee proper surveillance of the pathogen. In light of this, comparative data collection (official controls vs. own checks) on sample sensitivity could also be considered for breeding and laying hen flocks. Integrated surveillance based on the 'One health' approach combined with effective containment measures along the entire food chain (based on the application of biosecurity measures, effective surveillance and vaccination at the farm level, good manufacturing and hygienic practices during slaughtering, food processing, at retail and in the consumer phase) within integrated systems implemented by FBOp under the control of CAs are essential to control the spread of Salmonella, especially the most important current and emergent epidemic clones (Antunes et al., 2016; Campos et al., 2019) . Related projects and Internet sources • The EU trend of confirmed listeriosis cases remained stable (flat) in 2015-2019 after a long period of an increasing trend. • Listeria infections were most commonly reported in the age group over 64 years and particularly in the age group over 84 years. • The overall EU case fatality was high (17.6%) and increased compared with 2018 and 2017 (13.6% and 15.6%, respectively). This makes listeriosis one of the most serious food-borne diseases under EU surveillance. • In 2019, the number of outbreaks caused by L. monocytogenes (n = 21) was 50% higher compared with 2018 (n = 14) and the related illnesses jumped from a total number of 748 cases reported at the EU level between 2010 and 2018 (83.4 annual cases on average) to 349 cases. This increase was mainly due to outbreaks in Spain, which reported 3 outbreaks, 225 cases, 131 hospitalisations and 3 deaths, compared with zero reported in 2018. • The occurrence of L. monocytogenes varied according to the RTE food category and the sampling stage. In all food categories covered by the Regulation (EC) No 2073/2005, the level of non-satisfactory results remained low at retail (0.0% for hard cheeses to 2.1% for products of meat origin, fermented sausages). At processing, this level is systematically higher for all categories. The highest level was found, as previous year, for fish, with 5.8% unsatisfactory single units. Surveillance and monitoring of Listeria monocytogenes in the EU Surveillance of listeriosis in humans in the EU is based on invasive forms of L. monocytogenes infection, mostly manifested as septicaemia, meningitis or spontaneous abortion. Diagnosis of Listeria infections in humans is generally carried out by culture from blood, cerebrospinal fluid and vaginal swabs. Notification of listeriosis in humans is mandatory in most EU MS, Iceland, Norway and Switzerland, except for three MS, where notification is based on a voluntary system (Luxembourg and the United Kingdom) and another, non-specified system (Belgium). The surveillance systems for listeriosis cover the whole population in all MS, except in Belgium and Spain. Since 2015, the coverage of the surveillance system is estimated to be 80% in Belgium and this proportion of populations was used in the calculation of notification rates. No estimate for the population coverage was provided for Spain, so the notification rate was not calculated. For 2019, Spain did not receive data from all regions due to COVID-19 so the case numbers might therefore not be complete. All countries reported case-based Tables and figures that are not presented in this chapter are published as supporting information to this report and are available as downloadable files from the EFSA knowledge junction at zenodo https://doi.org/ 10.5281/zenodo.4298993. Summary statistics of human surveillance data with downloadable files are retrievable using ECDC's Surveillance Atlas of Infectious Diseases at http://atlas.ecdc.europa.eu/public/index.a spx data except Bulgaria, which reported aggregated data. Both reporting formats were included to calculate numbers of cases and notification rates. Monitoring of L. monocytogenes is conducted along the food chain during preharvest (e.g. animals at the farm and their feed), processing (e.g. cutting plant, slaughterhouses) and post-processing (e.g. retail and catering). The public health risk of L. monocytogenes posed by RTE food also depends on the effectiveness of its control, which includes the implementation of Good Agricultural Practices (GAPs) at the farm level, the Good Manufacturing Practices (GMP) and HACCP programme during processing and retail in food business operators (FBOp). Regulation (EC) No 2073/2005 8 on microbiological criteria lays down the microbiological criteria and the implementing rules to be complied with by the FBOp when implementing the general and specific hygiene measures of Regulation (EC) No 852/2002. In this Regulation, RTE food is defined, as 'Food intended by the producer or the manufacturer for direct human consumption without the need for cooking or other processing effective to cut out or reduce to acceptable level microorganisms of concern'. The National CAs conduct investigations (official sampling) to verify whether the FBOp implement correctly the legal framework of own check programmes (compliance with FSC, including for L. monocytogenes) as well as the analyses as part of HACCP (industry monitoring) according to the General Food Law principles. The rationale for surveillance and monitoring of L. monocytogenes in animals, feed and food at the different stages along the food chain and the number of samples provided to EFSA for 2019 is shown in Figure 25 . In 2019, 25 MS reported 218,439 samples tested for L. monocytogenes on different RTE food categories at retail or processing stages and 13 MS reported 22,135 samples tested at primary production level. Most of the monitoring data on L. monocytogenes in animals and feed provided are generated by non-harmonised monitoring schemes across MS and for which mandatory reporting requirements do not exist. Among several transmission routes, listeriosis in animals can be acquired via the consumption of contaminated feed such as poor-quality silage. Data on L. monocytogenes occurrence in feed are only collected as part of clinical investigations in farm animals. Hence, monitoring data on L. monocytogenes in animal feed are rarely available. Reported data on L. monocytogenes in RTE food are, in the most part, food chain control data (official monitoring) and are collected by the CA conducting investigations to verify whether FBOp implement correctly the above-mentioned FSC, which have been in force since January 2006. Data provided to EFSA within that context only allow a descriptive summary at the EU level and are not harmonised. The reporting of food-borne outbreaks is mandatory according to Zoonoses Directive 2003/99/EC and the reported data represent the most comprehensive set of data available at the EU level for assessing their burdenincluding those caused by L. monocytogenes. More details can be found in the chapter on food-borne outbreaks. The following two data streams were distinguished for summarising the information on L. monocytogenes in RTE foods. The first stream of data is the official food chain control data; these data comprise samples collected by the CA as part of verification of the compliance of L. monocytogenes FSC listed in Regulation (EC) No 2073/2005 to verify whether FBOp implement correctly the legal framework of own check programmes as well as the analyses as part of HACCP according to the General Food Law principles. These data were filtered from the database using the criteria 'official sampling' for the sampler, 'single units' for the sampling unit and 'objective sampling' for the sampling strategy. L. monocytogenes FSC of the Regulation (EC) No 2073/2005, which are to be complied with by FBOp and which are batch based, are specified by RTE food category, by sampling stage and are underpinned by the results of either the detection (ISO, 2017a) or enumeration (ISO, 2017b) analytical methods (Table 18) . Data reported by MS were separated into the different categories of RTE food/sampling stages based on the assumptions described in the EU summary zoonoses and food-borne outbreaks report of 2016. 9 Briefly these assumptions are: all sampling units that were collected from 'cutting plants' and 'processing plants' were considered as units collected at the processing stage, while sampling units that were obtained from 'catering', 'hospital or medical care facility', 'retail', 'wholesale', 'restaurant or cafe or pub or bar or hotel or catering service', 'border inspection activities', 'packing centre' and 'automatic distribution system for raw milk' were considered as units collected at retail. When stage was 'not available', 'unspecified', data have also been considered as part of the retail stage. As no data on physicochemical parameters of the sampled foods such as pH, water activity (a w ), levels and types of preservatives are provided to EFSA, it was considered that all RTE foods are able to support the growth of L. monocytogenes. So, the criterion applied for samples collected at the processing stage within the context of Regulation (EC) No 2073/2005 was 'not detected in 25 g'. Two exceptions were applied for the 'hard cheeses' and 'fermented sausages', for which the criterion of '≤ 100 CFU/g' was applied. EFSA assumes that 'hard cheeses' and 'fermented sausages' belong to the category of foods that are unable to support the growth of L. monocytogenes, because foods classified under these two categories of RTE products undergo ripening/fermentation and are expected to have low pH and In a two-class attributes sampling plan defined by n = 10, c = 0 and a microbiological limit of 'not detected in 25 g', in order for the food batch to be considered acceptable, L. monocytogenes must not be detected in qualitative (detection) analyses of 25-g food test portions obtained from each one of 10 sample units taken from the batch. If even one of the sample units from the batch is found to contain L. monocytogenes (detected in 25 g), then the entire batch is deemed unacceptable. This criterion applies to products before they have left the immediate control of the producing food business operator, when he is not able to demonstrate, to the satisfaction of the competent authority, that the product will not exceed the limit of 100 CFU/g throughout the shelf-life. (d): This criterion applies if the manufacturer is able to demonstrate, to the satisfaction of the competent authority, that the product will not exceed the limit 100 CFU/g throughout the shelf-life. The operator may fix intermediate limits during the process that should be low enough to guarantee that the limit of 100 CFU/g is not exceeded at the end of the shelf-life. moderate a w values. More information on the impact of RTE food processing, like fermentation and drying on pathogen loads in the RTE food can be found elsewhere (EFSA BIOHAZ Panel, 2018a). The RTE foods that are considered able to support the growth of L. monocytogenes are expected to have near-neutral or moderately low pH and relatively high a w values or can be very heterogeneous in terms of their manufacturing technology and physicochemical characteristics. In assessing RTE food category 'other dairy products', EFSA is presenting the results in a conservative way by considering all 'other dairy products' as capable of supporting the growth of L. monocytogenes. The second subset of data includes all monitoring and surveillance activities results reported by MS and non-MS to assess the occurrence of L. monocytogenes in different RTE food categories. In this case, only the data retrieved using detection methods were used, as these have a higher sensitivity compared with the quantitative investigations (using L. monocytogenes enumeration methods). All levels of sampling unit (single and batches), sampling stage (processing and retail) and sampling context (surveillance, monitoring and surveillancebased on Regulation (EC) No 2073/2005) were considered. Data obtained from the sampling strategies 'census sampling', 'convenient sampling' and 'objective sampling' were used, excluding data reported from 'suspect sampling', 'selective sampling' and 'other' contexts. When the sampling strategy was not spelled out (either 'not reported', 'not available', not specified or 'import sampling'), the data were included assuming that these would not fall into the category of suspect or selective sampling. All samplers' data were included. Specific graphs were prepared to illustrate the occurrence in different RTE food categories during the 2016-2019 period. Each point of these graphs represents the overall observed occurrence and the 2.5th and 97.5th percentiles of the uncertainty distributions of these occurrences. Data used for calculating uncertainty levels were the total number of samples (n) and the number of positive samples (s) observed. The uncertainty distributions were calculated with beta distribution beta (s + 1, ns + 1) (Vose, 1998) . To describe the occurrence of L. monocytogenes in animals and feed, all the sampling strategies were included even data reported for 'suspect sampling' and 'selective sampling'. 3.4.1. Overview of key statistics along the food chain, EU, 2015-2019 Table 19 summarises EU-level statistics on human listeriosis and on samples from RTE food tested for L. monocytogenes during 2015-2019. Food data of interest reported were classified into the major categories and aggregated by year to obtain an annual overview of the volume of data submitted. The sampling effort of the MS in 2019 for L. monocytogenes in some major RTE food categories can be found in Appendix A (Table A.1) . In 2019, as in previous years, the most sampled RTE food categories for L. monocytogenes detection and/or enumeration were 'RTE meat and meat products' (29.6% from total RTE food samples) and 'RTE milk and milk products' (28.4%). 'RTE fish and fishery products' samples represent 6.1% of the total reported by MS. The total number of sample units tested by MS increased by 38% in 2019 compared with 2018. This result is explained by an increase of 12% of the sampling units tested for 'RTE meat and meat products' and of 204% for 'other RTE food products'. More specifically, a higher number of samples were tested for 'bakery products' (+75%), 'broiler meat and meat products thereof' (+304%) and fruit and vegetables (+79%). Romania contributed particularly to the increase for 'other RTE food products ' (with 51,192 sampling units tested in this category in 2019). Table A .1 in Appendix A contains the samples taken by country at processing and retail levels. 80% of 'RTE milk and milk products' data were provided in decreasing order by Italy, Poland, Bulgaria, Romania, Germany and the Netherlands. Similarly, 80% of 'RTE meat and meat products' were provided by Poland, Romania, Germany, Bulgaria and Belgium; 80% of 'fish and fishery products' were provided by Poland, Germany, Romania, France, Bulgaria, the Netherlands, Belgium and Italy. 'Other RTE products' were mainly reported by Romania (67% of the total reported in this category), Germany, Ireland and Spain. As previous years relatively few samples (0.8%) were reported for 'RTE foods intended for infants and for medical purposes'; samples were mainly provided by Slovakia, Belgium, Ireland, Germany, the Netherlands and Italy. When the UK data were collected, the UK was an EU MS but as of 31 January 2020, it has become a third country. In 2019, 28 MS reported 2,621 confirmed cases of invasive listeriosis in humans (Table 20) . The EU notification rate was 0.46 cases per 100,000 population, which was at the same level as in 2018 (0.47 cases per 100,000 population). The highest notification rates were observed for Estonia, Sweden, Denmark and Malta with 1.59, 1.10, 1.05 and 1.01 cases per 100,000 population, respectively. The lowest notification rates were reported by Bulgaria, Croatia, Cyprus and Romania (≤ 0.19 per 100,000). The majority (99.3%) of listeriosis cases with known origin of infection was reported to be acquired in the EU in 2019 (Table 19 ). Ten MS reported 28 travel-associated listeriosis cases with known travel destination, 14 cases were travelled outside the EU and 14 cases within EU. The proportion of reported listeriosis cases without data on travel status or with unknown country of infection was 30.2% of all confirmed cases in 2019 (Table 19) . In the period 2010-2019, a seasonal pattern was observed in the listeriosis cases reported in the EU/EEA, with high summer peaks followed by smaller winter peaks. Over the 5-year period during 2015-2019, the trend of confirmed listeriosis cases was stable (flat) ( Figure 26 Information on hospitalisation was provided by 19 MS for 51.1% of all confirmed cases in 2019. Among the cases with information on hospitalisation status, 92.1% were hospitalised. Listeriosis had the highest proportion of hospitalised cases of all zoonoses under EU surveillance. The outcome was reported for 1,707 confirmed cases (65.1%). Twenty-one MS reported 300 deaths with listeriosis in 2019. This represented a 31.0% increase compared with 2018 (229 deaths). There was a steady increase in the annual number of deaths between 2010 and 2019 (annual average: 217). The overall EU case fatality among cases with known outcome was 17.6% and increased from 13.6% and 15.6% in 2017 and 2018, respectively. France reported the highest number of fatal cases (56) followed by Spain (55) and Poland (54). Listeria infections were most commonly reported in the age group over 64 years. At the EU level, the proportion of listeriosis cases in this age group has steadily increased from 56.1% in 2008 to 64.5% in 2019 and especially in the age group over 84 years, with an increase from 7.3% to 14.3% in the same time period. The case fatality was 19.5% and 23.0% in the age group 64-84 years and over 84 years, respectively, in 2019. In total, 2,621 confirmed human listeriosis cases were reported to TESSy in 2019. Overall, there were 1,803 domestic (acquired within the home country) confirmed listeriosis cases reported to the TESSy, which was 99.3% of the number of reported human listeriosis cases infected in the EU (domestically or through travel within EU) during 2019 (Table 19) . Listeria monocytogenes was identified overall by 10 MS in nine strong-evidence and 12 weakevidence food-borne outbreaks that together affected 349 people in the EU (of which 207 in Spain), with 236 hospitalised and 31 deaths, as reported to EFSA. For nine strong-evidence food-borne outbreaks in the EU in 2019 caused by L. monocytogenes, three were caused by 'meat and meat products' (one reported with additional information 'cold cuts'), two by 'broiler meat and products thereof' (with additional information 'RTE meat products' and 'chicken mayo sandwich') and one by each of the categories 'bovine meat and products thereof' ('potted beef') , 'pig meat and products thereof' (no additional information), 'mixed food' ('hummus and salads prepared in a small establishment') and 'vegetables and juices and other products thereof' ('black ') . Previously, during 2010-2018, 'mixed food', 'fish and fish products' and 'vegetables and juices and products thereof' were the most frequently reported food matrices causing strong-evidence listeriosis food-borne outbreaks. Further details and statistics on the listeriosis foodborne outbreaks for 2019 are in the food-borne outbreaks chapter. Comparing the food-borne outbreak cases (349) and confirmed cases of human invasive listeriosis acquired in the EU (1, 817) and considering also the estimated cases with unknown travel data (0.993 9 792) ( Table 19 ) could suggest that overall in the EU in 2019 13.4% (349/2,604 9 100) of human listeriosis cases would be reported through food-borne outbreak investigation. It is important to clarify that the case classification for reporting is different between these two databases. In TESSy, the cases reported are classified based on the EU case definition. All these cases visited a doctor and are either confirmed by a laboratory test (confirmed case) or not (probable case and classification is based on the clinical symptoms and epidemiological link). Also, surveillance of listeriosis in humans in the EU is based on invasive forms of L. monocytogenes infection, mostly manifested as septicaemia, meningitis or spontaneous abortion. Cases that never visited a doctor are not reported to TESSy. Moreover, there may be missing probable cases in TESSy, as these data are not analysed or published and there is no incentive for reporting such cases. Information on which cases are linked to an outbreak and which not is also not systematically collected. In practice, the cases reported to TESSy are considered to be mostly sporadic cases. In food-borne outbreaks, the human cases are the people involved in the outbreak as defined by the investigators (case definition), and cases must be linked, or probably linked, to the same food source (Directive 2003/99/EC). This can include both ill people (whether confirmed microbiologically or not) and people with confirmed asymptomatic infections . Cases can be classified as confirmed or probable outbreak cases, but currently these specific classification data are not collected by EFSA. Data on L. monocytogenes on RTE foods in the context of the Food Safety Criteria laid down in Regulation (EC) No 2073/2005 In total, 14 MS (BE, BG, CY, DK, EE, ES, GR, HR, LU, LV, RO, SI, SK) reported data according to the specifications mentioned above (Section 3.3.1) for 11 RTE food categories (Table 21) . At retail, depending on the RTE food category, 0.0-2.1% of single samples from official sampling were positive for L. monocytogenes, whereas at processing results ranged from 0.0% to 5.8%. A lower overall proportion of positives was reported at retail level compared with processing stage for all RTE food categories. and 'processing plants'. (c) : Includes sampling units that were obtained from 'catering', 'hospital or medical care facility ', 'retail', 'wholesale', 'not available', 'unspecified', 'restaurant or cafe or pub or bar or hotel or catering service', 'automatic distribution system for raw milk', 'border inspection' and 'packing centre'. (d): The results from qualitative examinations using a detection method were used to assess the criterion of 'not detected in 25 g' and the results from quantitative analyses using an enumeration method were used to assess the criterion of '≤ 100 CFU/ g'. qualitative analyses or number of L. monocytogenes > 100 CFU/g for enumeration analyses) and in parenthesis the number of tested samples (single samples or batches) and the number of reporting MS. (g): Includes RTE fish that is 'cooked', 'gravad/slightly salted ', 'marinated' or 'smoked'. (h) : Includes crustaceans, molluscan shellfish, fishery products unspecified, surimi, fishery products from fish species associated with a high amount of histidine and fish canned. (i): Includes 'curd', 'fresh' and 'soft or semi-soft', cheeses made with milk from different species ('cows', 'goats', 'sheep', 'mixed' or 'unspecified or other animal'). (j): Includes 'hard' cheeses made with milk from different species ('cows', 'goats', 'sheep', 'mixed', 'unspecified' or from other animals'). (k): Includes 'unspecified' cheeses made with milk from different species ('cows', 'goats', 'sheep', 'mixed', 'unspecified' or from other animals'). (l): Includes 'butter', 'buttermilk', 'cheese analogue', 'cream', 'dairy desserts', 'dairy products, not specified', 'fermented dairy products', 'ice cream', 'milk-based drinks', 'milk powder and whey powder', 'sour milk ' and 'yoghurt'. (m) : Includes milk ('pasteurised', 'UHT', or 'raw , intended for direct human consumption') from 'cows' or 'sheep'. Raw milk and raw milk for the manufacture of raw and low heat-treated products are not included. (n): Includes fermented sausages made from meat of different animal species ('bovine animals', 'pig', 'mixed', or 'other animal species or unspecified') . (o): Includes 'meat products' ('cooked ham', 'cooked, RTE', 'heat treated, RTE', ' raw and intended to be eaten raw ', 'pât e', 'unspecified, RTE' or 'unspecified') and meat preparations ('intended to be eaten raw') from different animal species ('bovine animals', 'pigs', poultry ('broilers', 'duck', 'turkeys', 'unspecified') , 'mixed', 'farmed game-land mammals', or 'other animal species or not specified'). (p): Includes bakery products ('cakes', 'desserts', 'pastry') , beverages, non-alcoholic ('soft drinks') , fruits ('pre-cut', 'products') , fruits and vegetables ('pre-cut') , juice ('fruit juice', 'mixed juice', 'vegetable juice'), RTE salads (also those 'containing mayonnaise'), seeds, sprouted ('RTE') , soups ('RTE') , spices and herbs ('dried') , vegetables ('pre-cut', 'products') and other processed food products and prepared dishes ('unspecified', 'sandwiches', 'sushi') . (q): Includes data from Croatia that has only been reported as ≤ 100 CFU/g (and has not been reported as > 100 CFU/g although all negative). Details on the occurrence of L. monocytogenes in the main RTE food matrices in 2019 together with 2017 and 2018 results can be found in Appendix B (Table 1B) . Below text summarises the results for the major food categories for the 2016-2019 period, considering all levels of sampling unit, sampling stage and sampling context. Over the 2016-2019 period, 24 MS and four non-MS reported data on RTE fish and fishery products. A summary of the occurrence of L. monocytogenes-positive units in RTE fish and fishery products in the EU over the period 2016-2019 is presented in Figure 27 . , the overall occurrence of L. monocytogenes in RTE fish was 4.3% with Bulgaria, Germany, the Netherlands and Poland reporting more than 80% of the positive samples as in 2018. The overall occurrence of L. monocytogenes in RTE fishery products was 4.2% with Germany, Italy, Poland and Romania reporting more than 80% of positive samples. The occurrence by merging RTE fish and RTE fishery products was 4.3%, 2.7%, 5.3% and 4.7% for the period 2019-2016. 'Fish, RTE' includes data on 'fish' of the following types: 'chilled', 'cooked-chilled', 'gravad/slightly salted', 'Fishery products, RTE' includes the following types: 'crustaceansprawnscooked', 'crustaceanslobsterscooked', 'crustaceansunspecifiedcooked', 'crustaceansshrimpsshelled, shucked and cooked', 'crustaceansunspecifiedshelled, shucked and cooked', 'crustaceansshrimpscooked', 'fish -fishery products from fish species associated with a high amount of histidinenot enzyme maturated', 'fish -fishery products from fish species associated with a high amount of histidinewhich have undergone enzyme maturation treatment in brine', 'fishery products, unspecifiedcooked', 'fishery products, unspecified -RTEchilled', 'fishery products, unspecifiedsmoked', 'fishery products, unspecified -RTE', 'molluscan shellfishshelled, shucked and cooked', Cheeses. Sixteen MS (AT, BE, BG, CY, DK, EE, DE, HR, IT, NL, PL, PT, RO, SK, ES and UK) and two non-MS (ME and MK) reported 2019 data from L. monocytogenes detection in cheeses. Bulgaria, Germany, Italy, the Netherlands, Poland, Romania and Slovakia were the major contributor for all cheese samples tested (81.4%). Cheeses made from pasteurised cows' milk represent more than 41.2% of samples collected and reported. Overall, considering all milk origin (species) and all types of cheeses L. monocytogenes was detected in 0.7% of the 9,660 cheese samples tested. A summary of the proportion of L. monocytogenes-positive units for the various types of cheeses is presented in Figure 29 . Since data were mostly reported by a limited number of MS and are of a heterogeneous nature as these include various diverse subcategories, the findings presented in this figure may not be representative of the EU level or directly comparable across years. RTE pig meat products includes 'meat from pig, meat products' of the following types: 'cooked ham', 'cooked, RTE', 'fermented sausages', 'pât e', 'raw and intended to be eaten raw', 'raw ham', 'unspecified, ready-to-eat' and 'ready-to-eat' and 'meat from pigmeat preparation' of the following type 'intended to be eaten raw'. 'RTE turkey meat' includes turkey 'meat products' of the following types: 'cooked, RTE', 'ready-to-eat' and 'raw and intended to be eaten raw'. 'RTE broiler meat' includes broiler 'meat products' of the following types: 'cooked, RTE'. 'RTE bovine meat' includes 'meat from bovine animals, meat products' of the following types: 'cooked, RTE', 'fermented sausages', 'raw and intended to be eaten raw', 'pât e'; 'ready-to-eat'; and 'unspecified, RTE'; 'meat from bovine animals, meat preparation' of the following types: 'intended to be eaten raw' and 'meat from bovine animals, minced meat' of the following types: 'intended to be eaten raw'. The 2019 prevalence of soft and semi-soft cheeses (SSC) and hard cheeses (HC) made from rawlow heat treated (LHT) milk were comparable and ranged between 0.9 and 1.0%. The 2019 prevalence of SSC and HC made from pasteurised milk were 0.3% and 0.04%, respectively. In general, considering the 2016-2019 time period, a higher prevalence in raw-LHT cheeses (1.0% mean prevalence for HC and SSC) than in pasteurised cheeses (0.1% mean prevalence for HC and SSC) is observed. In 2019, results from other RTE food product categories, such as 'bakery products', 'fruit and vegetables', 'RTE salads', 'spices and herbs', 'sauces and dressings' and 'other processed food products and prepared dishes' were reported. For 'bakery products', samples testing using a detection method were reported by 11 MS. Overall, out of the 6,653 units of bakery products tested, 0.2% were found to be positive for L. monocytogenes, similar to 2018 results. Germany and Romania contributed to 80% of the samples taken in 2019. In 2019, 17 MS provided data from investigations of L. monocytogenes on 2,357 units of 'RTE fruit and vegetables' tested using a detection method. The overall occurrence was of 1.7% (compared with 1.8% in 1,257 units tested in 2018). Bulgaria, Germany, Italy, Romania, Spain and the UK mainly ' + 'unspecified') . 'Retail' corresponds to data obtained from catering, hospital or medical care facilities, retail, wholesale and restaurants or cafes or pubs or bars or hotels or catering services. For each sampling stage ('overall', 'retail' and 'processing') , data are pooled across both types of sampling units ('single' and 'batch') . 'Processing' corresponds to data obtained from packing centres, cutting plants and processing plants. Since data were mostly reported by a limited number of MS, the findings presented in this figure may not be presentative of the EU level. 'Hard cheeses pasteurised milk' and 'hard cheeses from raw or low heat-treated milk' includes cheeses made from cows' milk, sheep's milk, goats' milk, mixed milk from cows, sheep and/or goats and unspecified milk or other animal milk. 'Soft and semi-soft cheeses' includes both soft and semi-soft and fresh cheese made from cows' milk, sheep's milk, goats' milk, mixed milk from cows, sheep and/or goats and unspecified milk or other animal milk. contributed to the sampling effort with nearly 85% of the samples in 2019. The 'RTE fruit and vegetables' prevalence over the 2016-2019 period is presented in Figure 30 . For 'RTE salads', 3, 138 samples were analysed and 109 samples (3.5%) were found to be positive by a detection method, while for 'spices and herbs', 291 samples were analysed and two samples (0.7%) were found positive. For 'sauces and dressings', 369 samples were analysed and one sample (0.3%) tested positive. For 'egg products' and 'confectionery and pastes', respectively, 26 and 54 samples were analysed, and none was found positive by a detection method. In 'other processed food products and prepared dishes' (unspecified, sushi or ices and similar frozen desserts), 14 MS submitted data. Overall, L. monocytogenes was detected in 0.3% of the 42,925 units tested with Romania reporting more than 90% of the samples. In 2019, 12 MS and two non-MS reported data on several animal categories (food-producing, wild-, zoo-and pet animals, including birds) from different species. Reported data were mainly from animals (99%) compared with other sampling unit levels ('herd/flock' and 'holding') . In the EU, the major animal data for Listeria testing concerned cattle (82%), sheep (11%) and pigs (3%). The sample size, as well as the sampling strategy and the proportion of positive samples, varied considerably among the reporting countries and animal species. Most EU data at the animal level were reported by two MS, the Netherlands (51%) and Ireland (38%). In total, considering the three sampling units (animal, herd/flock and holding) together, MS reported 17,516 tested units for Listeria spp. and 246 (1.4%) were found to be positive. Among the positive units, 67 (27.2%) were reported as being positive for L. monocytogenes and only limited positive findings were reported as Listeria innocua (four units, 1.6%) and Listeria ivanovii (two units, 0.8%). As previous years, major positive findings (173 units, 70.3%) were reported as 'other' or 'unspecified species' for Listeria. In 2019, only one MS (HR) reported a negative sample in soya-derived feed material. EU surveillance of human listeriosis focuses on the severe, invasive form of the disease, which affects the following risk groups: elderly, immunocompromised people as well as pregnant women and infants. While still relatively rare with 2,621 confirmed cases in the EU (notification rate of 0.46 cases per 100,000 population) in 2019, it is one of the most serious food-borne diseases under EU surveillance causing hospitalisation, high morbidity and mortality, particularly among the elderly. Confirmed human cases of invasive listeriosis have shown a significant increasing trend since EU surveillance was initiated in 2008. This trend stabilised in the EU as a whole over the last 5-year period during 2015-2019 and in most MS, while three MS reported a significantly increasing trend. Most listeriosis caseswhen this information was knownhave been domestically acquired and few cases have been linked to travel, within or outside the EU. The number of cases acquired within the EU increased slowly in the last 5 years, as a smaller proportion of cases were reported with unknown information on travel status and country of infection in 2019. Since the beginning of EU-level surveillance, most listeriosis cases have been reported in people over 64 years of age. The number and proportion of cases reported for this age group have increased steadily from 2008 until 2017. Human cases almost doubled in the age group over 84 years in the same time period. The proportion of cases, however, slightly decreased in the age group over 64 years during the last 2 years in 2018-2019. This is particularly visible in the age group over 84 years. As in previous years, almost all reported listeriosis cases Àwith information on hospitalisation statusÀ were hospitalised. In 2019, the overall EU case fatality among cases with known outcome was 17.6% and the number of deaths increased by 31% compared with 2018. Listeriosis continues to cause the highest number of fatal cases among food-borne infections in the EU. The highest mortality was in age group over 84 years. The high incidence of Listeria infections in elderly may be partially explained by the ageing population in the EU and parallel increases in susceptibility due to underlying chronic diseases (EFSA BIOHAZ Panel, 2020b). As ageing of the populations will continue in most MS (EUROSTAT, 2020) in the coming years, it is important to raise awareness of listeriosis and the risk, especially to older people, associated with certain consumption habits and types of food (e.g. RTE fish products and frozen vegetables) ECDC, 2018a, 2019b; EFSA BIOHAZ Panel, 2018a . In 2019, the number of human cases reported as food-borne outbreak cases (349) was 13.4% of the estimated number of the acquired cases of invasive human listeriosis in the EU (2,604 cases). Overall, L. monocytogenes was identified by 10 MS in nine strong-evidence and 12 weak-evidence food-borne outbreaks that together affected 349 people in the EU, with 236 hospitalised and 31 deaths, as reported to EFSA. Outbreaks of listeriosis continue to occurfor strong-evidence outbreaksassociated with several food vehicles including 'meat and meat products' (three strong-evidence food-borne outbreaks), 'broiler meat and products thereof' (two strong-evidence food-borne outbreaks) and 'bovine meat and products thereof', 'pig meat and products thereof', 'mixed food' and 'vegetables and juices and other products thereof' (each one strong-evidence food-borne outbreak). In six of these nine outbreaks, the food was RTE whereas for the remaining three no additional food vehicle information was provided. Since 2016, MS continue to increase their sampling for most of the RTE food categories. The number of food samples tested was 38% higher in 2019 compared with 2018. This result is explained by an increase of 12% of the sampling units tested for 'RTE meat and meat products' and of 204% for 'other RTE food products'. More specifically, a higher number of samples were tested for 'bakery products' (+75%), 'broiler meat and meat products thereof' (+304%) and fruit and vegetables (+79%). Most food samples collected at processing and retail were from RTE products of animal origin. The number of samples tested for fruits and vegetables has increased since 2016 (+189% between 2017 and 2019). This could be a result of the awareness of the multi-country outbreak of L. monocytogenes ST6 over the period 2015-2018 caused by frozen vegetables. However, in 2019, this category still represents less than 2% of all food samples tested. EFSA published an opinion this year concluding that L. monocytogenes is the most relevant pathogen associated with blanched frozen vegetables. When these vegetables are consumed uncooked, the probability of illness per serving for the elderly (65-74 years old) population, is up to 3,600 times greater compared with those cooked, but still very likely lower than any of the evaluated RTE food categories. Routine monitoring programmes for L. monocytogenes should be designed following a risk-based approach and regularly revised based on trend analysis, being food processing environment monitoring a key activity in the frozen vegetable industry (EFSA BIOHAZ Panel, 2020b). The low number of data reported by MS in primary production (< 10% of the total reported data) reflects the absence of harmonised EU regulation in this sector. As previous years, in animals, an important proportion of isolates (70.3%) is reported as 'unspecified Listeria spp.' or 'Listeria spp. ' and were not identified at the species level. Listeriosis in animals is, however, known to be almost exclusively caused by L. monocytogenes and L. ivanovii (ANSES, 2011) . In 2019, the occurrence of L. monocytogenes varied according to the RTE food category and ranged from 0.04% for 'hard cheeses made from pasteurised milk' up to 4.3% for 'RTE fish'. Interpretation of trends for occurrence must be used with caution, since each year reporting data can vary according to the number of reporting MS, the food categories included in different contexts of the surveillance, the sampling efforts (sample size) and reporting attitude. Official sampling carried out by the CAs in the context of surveillance of the application of the FSC laid down in Regulation (EC) No 2073/2005 showed that the level of non-satisfactory results remains low at retail (from 0.0% to 2.1%). For previous years, this level was however systematically higher at the processing stage compared with the retail stage. New tools based on genotyping are now available to characterise isolates of L. monocytogenes. With these new developments in diagnostics and changes in the epidemiology of listeriosis outbreaks, the FAO/WHO JEMRA has launched in 2020 new work on L. monocytogenes in RTE foods. EFSA/ECDC surveillance data provide opportunities to validate the current risk assessment models for L. monocytogenes, assess their application to other food commodities and develop new management approaches to control L. monocytogenes. Combining such human, animal and food epidemiological data with molecular and genotyping data represents indeed an efficient tool to better understand the ecology of this pathogen among the different stages of the food chain and would improve the investigation of listeriosis outbreaks affecting one or several MS. Related projects and Internet sources Closing gaps for performing a risk assessment on L. monocytogenes in RTE foods: activity 1, an extensive literature search and study selection with data extraction on L. monocytogenes in a wide range of RTE food (external scientific report) https://www.efsa.europa.eu/en/supporting/ pub/1141e Closing gaps for performing a risk assessment on L. monocytogenes in RTE foods: activity 2, a quantitative risk characterisation on L. monocytogenes in RTE foods; starting from the retail stage (external scientific report) https://www.efsa.europa.eu/en/supporting/ pub/1252e Closing gaps for performing a risk assessment on L. monocytogenes in RTE foods: activity 3, the comparison of isolates from different compartments along the food chain and from humans using whole genome sequencing (WGS) analysis (external scientific report) https://www.efsa.europa.eu/en/supporting/ pub/1151e Evaluation of listeriosis risk related with the consumption of non pre-packaged RTE cooked meat products handled at retail stores in Greece (external scientific report) • The EU notification rate was 2.2 cases per 100,000 population, which was similar to 2018. • The highest notification rates were reported in Ireland, Malta, Denmark and Sweden. • The EU/EEA trend has been increasing from 2015 to 2019. • STEC was the third most frequent bacterial agent detected in food-borne outbreaks in the EU, with 42 outbreaks, 273 cases, 50 hospitalisations and 1 death reported in 2019. • The sources in the four strong-evidence STEC food-borne outbreaks during 2019 were 'bovine meat and products thereof' (two outbreaks), 'milk' and 'tap water, including well water' (one outbreak each). During 2010-2018, strong-evidence STEC outbreaks were mostly caused by 'bovine meat and products thereof' (18), 'tap water, including well water' (16), 'vegetables and juices and other products thereof' (10) and milk (8) • Sprouted seeds were tested by six MS with no positive STEC results from 331 official samples. An EU regulation with a microbiological criterion for the presence of STEC in this food commodity has been in force since 2013. • Overall, STEC was most commonly found in meat of different types derived from different animal species (4.1% STEC-positive), followed by 'milk and dairy products' (2.1%) while 'fruits and vegetables' was the least contaminated category (0.1%). • Sixteen MS tested 6,297 'ready-to-eat' food samples for STEC of which 37 (0.6%) were found to be STEC-positive, including 17 meat and meat product samples, 16 milk and milk product samples with 10 from cheese, two samples from spices and herbs, and one STEC-positive sample from salads and 'fruits, vegetables and juices' each. • Of the isolates from food with the reported information on the serogroup 21.6% belonged to the 'top-five' serogroups (O157, O26, O103, O111 and O145) in 2019 and more than half of all the remaining STEC belonged to the top-20 STEC serogroups reported in human infections to ECDC in 2015-2018. • Most of the virulotypes of STEC isolates from food and animal were also identified in severe STEC infections in humans. This identification, however, was only carried out on 52.9% of the food isolates for the stx gene typing (stx1 and stx2) and stx gene subtyping was only done for 6.1% of the food isolates, and even less for animal isolates. • Testing of animal samples was still not widely carried out in the EU with 2,588 animal samples tested for STEC by nine MS in 2019. Surveillance and monitoring of Shiga toxin-producing Escherichia coli in the EU The notification of STEC 10 infections is mandatory in most EU MS, Iceland, Norway and Switzerland, except for four MS, where notification is based on a voluntary system (France, Luxembourg) or another system (Italy and the United Kingdom). In the United Kingdom, although the reporting of food poisoning is mandatory, isolation and specification of the organism is voluntary. The surveillance systems for STEC infections cover the whole population in all EU MS except for three MS (France, Italy and Spain). The notification rates were not calculated in these three countries for the following reasons: (a) in France, the STEC surveillance in humans is based on paediatric haemolytic uraemic syndrome (HUS) cases; (b) in Italy, STEC surveillance is sentinel and primarily based on the HUS cases reported through the national registry of HUS; (c) no estimation for population coverage of STEC cases was provided by Spain (until 2018). In Belgium, full national coverage was set up in 2015 and rates before then are not displayed. For 2019, Croatia did not report data, and in Spain, not all regions reported data for 2019 due to COVID-19. Case numbers might therefore be lower than what could be expected. All countries report case-based data except Bulgaria, which reported aggregated data. Both reporting formats were included to calculate numbers of cases and notification rates. Diagnosis of human STEC infections is generally carried out by culture from stool samples and indirect diagnosis by the detection of antibodies against the O-lipopolysaccharides of E. coli in serum from HUS cases. In addition, diagnosis by direct detection of free faecal Shiga toxin/verocytotoxin or the identification of the presence of stx1/vtx1 or stx2/vtx2 genes in stools by PCR without strain isolation is increasing. Although the testing is intended to be mandatory, the sampling objectives and the sampling frequency applied varied or were interpreted differently between MS, making the data not fully harmonised. Data are also generated by the National CAs conducting inspections to verify whether the food business operators implement correctly the legal requirements (official monitoring data). The latter data are compliance checks and are not suitable for trend analyses, because a reference study population is mostly absent and because the sampling is risk based and so non-representative (Boelaert et al., 2016) . All the food and animal testing data, apart from those on sprouts testing produced in the context of the Regulation (EC) No 2073/2005, originate from the reporting obligations of MS under Directive 2003/99/EC (the zoonosis directive). Due to the absence in this Regulation of explicitly indicated sampling strategies, the data generated by MS are based on investigations with non-harmonised sampling. Moreover, mainly for animal samples, they are obtained with different analytical methods. Therefore, STEC monitoring data according to Directive 2003/99/EC are not comparable between MS and preclude subsequent data analysis such as assessing temporal and spatial trends at the EU level. In certain food categories, different sampling design and inaccuracies due to limited numbers of samples examined also preclude accurate prevalence estimation. Moreover, the use by MS of laboratory analytical methods that test only for E. coli O157 leads to inaccurate STEC prevalence estimations or STEC serogroup frequency distributions. While this problem affected less than 5% of food samples in 2019, these methods have been used to test more than 30% of the animal samples. Nevertheless, descriptive summaries of sample statistics at the EU level may be made and used to indicate the circulation of certain STEC types in food and animals, provided the mentioned relevant limitations of the data set are kept into consideration. To improve the quality of the EU data on STEC monitoring of food and animals, EFSA issued technical specifications for harmonised monitoring and reporting of STEC in animals and foodstuffs in 2009 (EFSA, 2009a) . With an additional Scientific Opinion, EFSA encouraged MS to extend the monitoring and report data on STEC serogroups (EFSA BIOHAZ Panel, 2013a). More recently, it has been recommended that the presence of the main virulence genes be reported, considering the most recent development in STEC testing and risk assessment (EFSA BIOHAZ Panel, 2020c; JEMRA FAO/WHO and NACMCF reports, see Section 4.6 for online reference of the last two reports). Finally, the latest published EFSA Scientific Opinion on the pathogenicity assessment of STEC presents important considerations related to the virulence of the different STEC types and underlines the importance of determining the virulence genes combinations (virulotypes) of the isolated STEC strains, with an emphasis on stx gene subtyping, which would facilitate a more precise assessment of the risk connected with different STEC isolates (EFSA BIOHAZ Panel, 2020c). The reporting of food-borne disease outbreaks of human STEC infections is mandatory according to Zoonoses Directive 2003/99/EC. Data validation and analyses of monitoring data from food and animals 4.3 The STEC monitoring data from food and animals reported for the year 2019 to EFSA were verified as regards their plausibility in line with the current knowledge. The following plausibility criteria focused on the level of completion and coherence of the information and on the consistency of the laboratory results with the analytical method reported: • Plausibility of reported occurrence values with respect to the STEC epidemiology based on the updated scientific literature. • Consistency of the reported laboratory results with the purposes of the STEC monitoring data collection. An example of data not consistent with the objective of the collection and for this reason excluded from the analysis, is the reporting of E. coli indicators or pathogenic E. coli other than STEC. • Consistency of the reported laboratory results with the analytical method used for the analysis. An example may be the reporting of STEC O26 or other non-O157 STEC serogroups for samples tested with the standard ISO 16654:2001 16654: (ISO, 2001 or equivalent methods, which can only detect serogroup O157. The monitoring data on sprouts as part of Regulation (EC) No 2073/2005 were aggregated and summarised for trend watching according the following specified data elements ('filters') ; Sampling context: 'surveillance, based on Regulation No 2073/2005'; Sampling unit type: 'single'; Sampling stage: as appropriate; Sampling strategy: 'objective sampling', and Sampler: 'official sampling'. For the description of the occurrence of STEC-positive samples in the different food categories, a subset of all validated monitoring data was used (N = 20,395). Data sets were extracted with 'objective sampling' being specified as sampler strategy, which means that the reporting MS collected the samples according a planned strategy based on the selection of random samples, which are statistically representative of the population to be analysed. Additionally, the data reported with a sampler 'HACCP and own checks' were excluded. For animal data (N = 1,802), the same filters applied. The full data set (N = 27,826) including also regionally only reported data (about 200 samples) was used instead for any other descriptive analysis of STEC findings in food and animals, primarily those on the serogroups and virulence genes' frequency distribution, with the aim to describe the full range of STEC isolated from food and animals. To descriptively analyse the reported STEC serogroups and virulence genes, the data were grouped according used test methods ( This disentanglement was necessary to minimise the impact of results based on E. coli O157specific methods, which do not allow identifying other STEC possibly present in the samples, on the analyses of the distribution of serogroups. Table 23 summarises EU-level statistics on human STEC infections and on samples from food and animals tested for STEC, during 2015-2019. Food and animal data were classified into the major categories and aggregated by year to obtain an annual overview of the volume of data submitted, considering the information reported for laboratory analytical methods (Section 4.3.2) and not considering the information reported on the sampling strategies/contexts. The proportion of human STEC cases infected domestically and through travel within the EU decreased since 2015 and increased slightly among the cases infected through travel outside the EU. For the year 2019, 22 MS provided results from analyses of 25,030 food units (batches or single samples). The most recent source attribution analysis available for STEC underlined that 'bovine meat and products thereof', 'milk and dairy products' and 'vegetables, fruit and products thereof' were the vehicles most frequently implicated in STEC infections in the EU in the period 2012-2017 (inclusive) (EFSA BIOHAZ Panel, 2020c), confirming the results of previous JEMRA FAO/WHO and NACMCF reports (see Section 4.6 for online reference of these two reports). These categories were those most commonly tested in 2019 in the EU and represented the 89% of the total food sample units tested by 21 MS. For the year 2019, 2,588 sampling units (single heads or herds or flocks) from animals were reported by nine MS. This number increased compared with the number of animals tested in 2018 (1, 631) . The proportion of animal samples tested for STEC and reported by EU MS in 2019 by the different analytical methods can be found in the supporting information to this report. When the UK data were collected, the UK was an EU MS but as of 31 January 2020, it has become a third country. In 2019, 7,894 cases of STEC infections, including 7,775 confirmed cases, were reported in the EU (Table 24) . Twenty-four MS reported at least one confirmed STEC case and three MS reported zero cases. The EU notification rate was 2.2 cases per 100,000 population, which was similar to the level in 2018 (2.3 cases per 100,000 population). The highest country-specific notification rates were observed in Ireland, Malta, Denmark and Sweden (16.3, 10.7, 10 .7 and 7.4 cases per 100,000 population, respectively). Seven countries (Bulgaria, Cyprus, Greece, Lithuania, Poland, Portugal and Slovakia) reported ≤ 0.1 cases per 100,000 population. Most STEC cases reported were infected in the EU (62.2% domestic cases or travel in the EU, 9.7% travel outside EU and 28.2% of unknown travel history or unknown country of infection) (Table 23) . Three Nordic countries -Finland, Sweden and Norway reported the highest proportion of travel-associated cases (52.1, 44.4 and 37.9%, respectively). Among 1,034 travel-associated cases with known probable country of infection, 72.5% of the cases involved travel outside the EU and 27.5% travel within the EU. Egypt was most frequently reported as the probable country of infection, followed by Turkey, Spain, Morocco, Italy and Thailand (14.7%, 13.4%, 5.0%, 3.3%, 3.1% and 3.1%, respectively). There was a clear seasonal trend in confirmed STEC cases in the EU/EEA between 2010 and 2019, with more cases reported during the summer months ( Figure 31 ). There was a significantly increasing trend (p < 0.01) for STEC in the EU/EEA in 2015-2019. Five MS (Austria, Denmark, Finland, Malta and Poland) reported significantly increasing trends (p < 0.01). One MS (the Netherlands) had a significantly decreasing (p < 0.01) trend over the same time period. This was due to a change in notification criteria in the Netherlands since 2016, where only acute infections with at least diarrhoea, vomiting and/or blood in stool have to be reported. Eighteen MS provided information on hospitalisation for 37.3% of all confirmed STEC cases in the EU in 2019. Out of the 2,903 cases with known hospitalisation status, 37.9% were hospitalised. The highest proportions of hospitalised cases (80.0-100%) were reported in Estonia, Greece, Italy, Poland and Slovakia. The number of HUS cases (394) was about the same level as in 2018. HUS cases were reported in all age groups with the highest proportion of patients in the youngest age groups from 0 to 4 years (272 cases; 69.4%) to 5-14 years (75 cases; 19.1%). The most common serogroups among HUS cases were O26 (38.7%), O157 (23.0%), O80 (9.0%) and O145 (8.0%); while 4.7% were untypable. In 2019, 10 deaths due to STEC infection were reported in the EU compared with 11 deaths in 2018. Six MS reported one to three fatal cases each and 14 MS reported no fatal cases. This resulted in an EU case fatality of 0.21% among the 4,739 confirmed cases with known outcome (61.0% of all reported confirmed cases). Deaths were reported in the age group 0-4 years (40%) and in all age groups over 25 years (60%). Half the deaths were associated with HUS. The serogroups and the stx gene subtypes associated with fatal cases were O157 (Stx2a), O145 (Stx1a, Stx2a) and O8 (Stx2d). For seven fatal cases, the serogroup was not specified. Overall, for the year 2019, 94.1% of the 4,835 reported STEC infections in humans who acquired the infection in the EU (Table 23) were domestic (acquired within the home country) infections and 5.9% were acquired through travel in EU. STEC was identified by 11 MS in 42 food-borne outbreaks that together affected 273 people in EU, with 50 hospitalised and one death, as reported to EFSA. Comparing the food-borne outbreaks cases (273), reported to EFSA, and cases of STEC infections in humans acquired in the EU (4, 835) considering also the proportion of unknown travel data (0.865 9 2,190) (Table 23) , reported to ECDC, could suggest that overall, in the EU in 2019 4.1% of human STEC cases could be reported through FBO investigations. It is important to clarify that the case classification for reporting is different between these two databases. In TESSy, the cases reported are classified based on the EU case definition. All these cases visited a doctor and are either confirmed by a laboratory test (confirmed Source: Austria, Cyprus, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, the Netherlands, Norway, Poland, Romania, Slovakia, Slovenia, Sweden and the United Kingdom. Belgium, Bulgaria, Czechia, Croatia, Portugal and Spain did not report data to the level of detail required for the analysis. case) or not (probable case and classification is based on the clinical symptoms and epidemiological link). Cases that never visited a doctor are not reported to TESSy. Moreover, there may be missing probable cases in TESSy, as these data are not analysed or published and there is no incentive for reporting such cases. Information on which cases are linked to an outbreak and which not is also not systematically collected. In practice, the cases reported to TESSy are considered to be mostly sporadic cases. In food-borne outbreaks, the human cases are the people involved in the outbreak as defined by the investigators (case definition), and cases must be linked, or probably linked, to the same food source (Directive 2003/99/EC). This can include both ill people (whether confirmed microbiologically or not) and people with confirmed asymptomatic infections . Cases can be classified as confirmed or probable outbreak cases, but currently these specific classification data are not collected by EFSA. The sources in the four strong-evidence STEC food-borne outbreaks during 2019 were 'bovine meat and products thereof' (two outbreaks) and 'milk' and 'tap water, including well water' (one outbreak each). During 2010-2018, strong-evidence STEC outbreaks were mostly caused by 'bovine meat and products thereof' (18), 'tap water, including well water' (16), 'vegetables and juices and other products thereof' (10) and milk (8) and cheese (8). Further details and statistics on the STEC food-borne outbreaks for 2019 are in the FBO chapter. Human serogroup and virulotype data are described in Section 4.4.5. Table 25 , those monitoring results are summarised and a distinction is made between RTE and non-RTE food including fresh meat. As regards RTE food, most of the results of the 6,297 RTE food sampling units reported by 16 MS originated from 'milk and milk products' notably cheeses (32.2%), followed by 'meat and meat products' (22.5%), 'fruits, vegetables and juices' (20.5%) and 'spices and herbs ' (10.9%) . In total, 37 RTE food samples were found to be positive for STEC: 17 in 'meat and meat products' (notably of bovine origin), 16 in 'milk and milk products' particularly in cheeses, two in 'spices and herbs' and one in each of the categories 'salads and fruits ' and 'vegetables and juices'. The analysis of RTE bovine meat products and meat preparations reported resulted in 1.48% positive samples out of 746 units tested by seven MS, while no positive samples were reported out of 113 units of RTE minced meat, meat preparations and meat products from pig meat, tested by five MS. Finally, 2.74% of the 146 RTE milk samples tested were STEC-positive. For the descriptive analysis of serogroups and virulence genes 6,757 sample units tested for STEC were available with 60 (0.9%) positive samples reported. The food categories concerned in this analysis included cheeses, sprouted seeds, spices and herbs, fruits and vegetables, meat products, fish and fishery products, raw milk and 'others'. Overall, the 0.7% of the samples proved positive with the most contaminated commodities being meat products (1.2% of the samples in this commodity) and cheeses (0.9% of the samples in this category, see above). Of all the STEC isolated from RTE food samples, only 16 were submitted with information on the serogroup. These included 11 different serogroups, with three STEC O157 isolated by one MS that used the ISO 16654:2001 method and three STEC O26. The characterisation of the virulence genes was carried out on 29 isolates for the stx genes (11 isolates with stx1, 12 with stx2 and six with stx1 and stx2) and on 18 for the presence of eae, while 16 isolates were provided with information on the type of stx gene and on the presence of the eae gene (four strains eae+; stx1+, two eae+; stx2+, three eae-; stx1+, four eae-; stx2+ and three eae-; stx1+ and stx2+). In the following descriptive analyses, food categories include RTE food and non-RTE food. Overall STEC contamination was detected in 494 (4.1%) out of 12,120 samples of meat and meat products reported by 16 MS. Bovine meat. In 2019, 5,794 units of fresh bovine meat were tested for STEC by 14 MS with 3.2% of these being positive (Table 25 ). Most of the units were sampled at the slaughterhouse (63.4%), followed by retail sampling stage (28.2%). The samples taken at the retail outlet were the most contaminated with 4.0% of the samples being found positive for STEC, whereas at the slaughterhouse level, there were 2.5% positive tests out of 3,682 samples. For the descriptive analysis of serogroups and virulence genes 198 isolates were available from 6,146 samples of bovine meat (fresh and other) tested by 15 MS. Information on the serogroup was reported by eight MS for 115 isolates (58.1%), which belonged to 39 different serogroups, among which the most frequently identified in 2019 were O13 (10 isolates) followed by O55 (eight isolates), O91 (seven isolates), O26, O174 and O113 (six isolates each) and others (Table 31 ). All the six most represented STEC serogroups identified in fresh bovine meat samples except the O13 serogroup were among the 20 most frequent serogroups reported in STEC from human disease in the EU in 2018 (EFSA and ECDC, 2019c). The analysis of the virulence genes of the isolated STEC included data reported by 12 MS and showed that 75.7% and 39.4% of the isolates were provided with information on the genes encoding the Shiga toxins (stx) and the intimin (eae), respectively, while 67 isolates were typed for both the genes. The latter isolates included 47 that were negative for the presence of the eae gene (12 with stx1, 19 with stx2 and 16 with both the stx1 and stx2) and 20 positives for eae (13 with stx1, four with stx2 and three with both stx1 and stx2). Ovine and goat meat. This food category is not widely tested at the EU level, reflecting the different eating habits of the different MS, particularly regarding goat meat. Conversely, small ruminants are important reservoir of STEC as reported in the literature (Persad and LeJeune, 2014) . In 2019, six MS reported the results of investigation of 816 sample units of fresh ovine meat with 11.6% of these being STEC-positive, whereas two MS reported on fresh goat meat with three STEC-positive sampling units out of the 18 tested (Table 25) . For the descriptive analysis of serogroups and virulence genes 93 isolates were available from ovine and goat meat (fresh and other). Information on the serogroup was available for 42 strains and the most represented was O146 (10 isolates), followed by O26, O157 and O15 (three strains each) among the STEC O-groups also represented amongst human isolates. The other isolates belonged to 17 other serogroups including some matching those isolated from human disease such as O113 and O91 (two and one isolate, respectively) (EFSA and ECDC, 2019c) (Table 32) . Seventy-five of the 93 STEC isolated from this food category in 2019 were provided with information on the presence of the Stxcoding genes. Thirty-five strains were stx1+, 18 and 22 isolates were stx2+ and stx1+ stx2+, respectively. In addition, 29 of these 75 isolates were provided with the information on the presence of the eae gene, which was present with stx1 in five isolates and with stx2 in four isolates. Meat from other ruminants. Only three MS provided information on the presence of STEC in fresh meat samples from deer. In total, 62 samples were taken and eight were found to be contaminated with STEC (12.9%). From monitoring results from fresh and other meat samples, six isolates were reported each belonging to a different serogroup, which included O91 and O146, both identified in STEC isolated from human disease (EFSA and ECDC, 2019c). Seven strains were reported with the information on the presence of the Stx-coding genes and were all positive for stx2 but one strain which possessed stx1 and stx2. The same set of isolates was also provided with the information on the presence of the eae gene and all were negative. Meat from other animal species. Four MS tested fresh pig meat in 2019 and reported data on 119 samples with eight of these being positive for the presence of STEC (6.7%) (Table 25) . From monitoring results from fresh and other meat samples five of the positive samples contained STEC O157, all isolated from one MS that used the ISO 16654:2001 method. The remaining three isolates were reported as STEC of unspecified serogroup. Five MS reported on the analyses carried out on 202 sample units of fresh meat from animal species other than bovine, ovine, goat, pigs and deer. These included samples taken from horses, rabbit, wild boars, poultry, wild and farmed game and unspecified meat. Fourteen samples were reported as STEC-positive (6.93%). One of the isolates was an STEC O103 and another was an STEC O157. The remaining isolates were not provided with serogroup information. STEC O157 was isolated by one MS, which reported testing six samples of poultry meat at retail, all were tested using the ISO 16654:2001 method. In 2019, one MS (Spain) presented data on the presence of STEC in fresh meat from broilers and turkeys. Thirty-seven samples from broilers and 14 from turkey meat were tested, all using the ISO 16654:2001 method, with four E. coli O157 reported in fresh meat from broilers. For the descriptive analysis of serogroups and virulence genes 2,560 sample results were available with 90 of them being positive for STEC. Information on the serogroup of the isolated STEC was provided for 17 isolates. Notably, 13 isolates were of the O157 serogroup, but all of these were from samples tested using the ISO 16654:2001 or equivalent methods. The remaining four belonged to different serogroups, which included STEC O103 and O174. Nine STEC isolates were reported with their stx genes profiles, four were stx2+, four were stx1+ and three stx1+; stx2+. The four isolates with the stx1-coding genes were also positive for the presence of the eae gene. Meat products and meat preparations. For the descriptive analysis of serogroups and virulence genes, 221 STEC isolates were available from 6,653 sample units (3.3%) of any meat products and meat preparations including those involving minced and mixed meats. The information on the serogroup was provided for 44 STEC strains, including 18 E. coli O157, seven of which had been detected using the ISO 16654:2001 method. The analysis of the presence of the stx and eae genes could be carried out on 78 and 52 isolates with this information reported, respectively (35.3% and 23,1%). Out of the 37 eae-positive isolates, six possessed the stx1 and four the stx2 genes, while, for the remaining 27, this information was not provided. Fourteen isolates were reported as being eae-negative and fell into three groups based on the stx genes profiles: stx1+ (two isolates), stx2+; (nine isolates) and stx1+; stx2+ (three isolates). Overall STEC was found in 61 (2.1%) out of 2,981 samples of milk and milk products reported by nine MS. In 2019, eight MS reported on the testing of 1,216 sample units of raw cows' milk with 48 positive units (3.9%). Information on the serogroup was provided for two isolates only (one STEC O26 and one O157). Three MS reported monitoring results of 27 sample units of raw goats' milk, while two MS reported only four samples of raw sheep milk. None of the samples tested was positive for STEC. One MS reported monitoring results of STEC in 102 samples of raw milk from other or unspecified animal species. Four positive samples were reported as STEC of unspecified serogroup. The presence of STEC in RTE dairy products other than milk and cheeses was reported by four MS, which tested 148 sample units of butter, cream, ice cream, whey, yoghurt and fermented dairy products. Overall found five isolates were found, one belonging to the O26 serogroup and the other four of a non-specified serogroup. For dairy products, in 2019, 2,696 cheese samples were tested for the presence of STEC, with 25 (0.9%) positive units. For the descriptive analysis of serogroups and virulence genes, 5,479 sample results were available, of which cheese accounted for 56.9%, with 91 positives (1.7%). Only six STEC were typed for the serogroup and four were STEC O26, one was O157 and the remainder belonged to the O181 serogroup. Characterisation of the stx and eae gene profiles also involved a small number of isolates with 22 and 12 strains reported with this information in the data set, respectively. Nine isolates were provided with the data on both the presence of stx and eae genes and included one stx1 and eae positive, and two eae and stx2 positive isolates. Six eae negative strains included two stx1, two stx2 and two stx1 and stx2 isolates. Overall STEC was found in two (0.1%) out of 2,171 samples of vegetables and fruits reported by nine MS. The positive records included two units of vegetables sampled at retail (leafy vegetables and pre-cut vegetable products), reported by two MS and both were contaminated with STEC of non-O157 serogroups. This category contains miscellaneous food commodities not included in the previously mentioned categories, which included cereals and meals, bakery products, non-alcoholic beverages, cereals and meals, juices, live bivalve molluscs, fish and fishery products, sauces and dressing, dried seeds and fresh and dried spices and herbs, infant formula, coconuts products, shrimps, water, honey and others. For the whole category, 1,704 samples were analysed by 10 MS with three (0.18%) positive samples reported from salads (one unit) and spice and herbs (two units) (see above RTE food). The STEC strains identified in spices and herbs included one STEC O88:H25 stx1+ stx2+ and one other STEC strain for which no further information on the serogroup and virulence genes was reported, whereas one STEC O11:H5 possessing the stx2 gene was reported in salads. Animals are tested much less than food in the EU. In 2019, 2,588 sample units from animals (animals or herds or flocks) were tested for STEC by nine MS. Overall, the presence of STEC was reported in 14.1% of them, considering the full data set. In total, 68.4% of the animal samples were tested using the ISO TS 13136:2012 method, while all the remaining samples were tested using methods targeting E. coli O157 only. As observed in previous years, the highest proportion of animal sampling units tested in 2019 was related to cattle, with 1,493 tested (62.4% of animal samples) with 17.1% positives. As for the other categories, 53.8% of the 104 sampling units from pigs proved positive for STEC, followed by the small ruminants with 61 sheep and goat sample units (14.8% positives) and the 270 deer samples with 11 positive units. The most relevant data reported on the animal categories are detailed below. Four MS reported the presence of STEC in 254 isolates (17%) out of 1,493 cattle sampling units. In total, 231 positive samples were detected out of 816 tested using the ISO TS 13136:2012 method or equivalent by three MS. Twenty-three positive samples out of 677 samples were obtained using the OIE method for E. coli O157 by two MS. The full data set (see Section 4.3.2, Data Analysis) included 276 STEC-positive sample results out of 1,615 samples tested from cattle. These included 13 additional STEC O157, nine of which were detected using the OIE method for the E. coli O157, one STEC O26, one isolate of serogroup O111, one STEC O145 and others. The remaining isolates were reported without information on the serogroup. Only about 12% of the cattle isolates were provided with information on the virulence genes. The STEC isolates with a more complete set of features determined included one O26 stx2+ eae+, one O145 stx2+ eae+, one O111 stx1+ eae+, four O157 stx2+ eae+, 24 O157 stx1+ stx2+ eae+ and one O168 stx1+; stx2+; eae-. Two MS reported the analysis of 15 samples taken at a goat farm, with six positive results (40%). By analysing the full data set, 61 samples from sheep and goats were reported from six MS. Nine positive samples yielded two STEC O157 and one STEC of O121 serogroup. The latter was also reported as possessing the stx2 and eae genes. Pigs were tested by two MS, which tested six single animals and 85 herds and reported for the latter 50 positive herds (58.8%). The full data set contained six additional isolates out of 104 units tested. These were one STEC O1, one STEC O2, one STEC O45 and three STEC of non-specified serogroup. The information on the stx genes was provided for 50 out of 56 strains and included 49 isolates positive for stx2 and one for stx1. The eae gene was not investigated in any of the pig isolates. In 2019, one MS (IT) reported the presence of STEC in 317 sample units of Cantabrian chamois, deer, Steinbock and wild boar with 17 (5.4%) positives. One MS (NL) reported on the testing of 377 broilers with one positive. Analysis of the STEC serogroups, conducted using the full data set, revealed 25 STEC isolates. For 20 of these, information on the serogroup was provided. In detail, five STEC were of seven were STEC O157, six belonged to O1, five to O2, one O103, one O24, the latter with the virulence genes stx2+; eae+. The remaining isolates did not have any virulence genes characterisation data. Humans Data on STEC serogroups (based on the O antigen) were reported in 2019 by 24 MS. Serogroup data were available for 57.9% of the human confirmed cases, which was a slight decrease compared with in 2018 (61.6%). As in previous years, the most commonly reported serogroup was O157, accounting for 26.6% of the cases in humans with a known serogroup. Its proportion has been decreasing to less than half from 54.9% in 2012 to 26.6% in 2019. The proportion of the second most common serogroup O26 slightly decreased compared with 2018 but has steadily increased from 11.6% in 2012 to 16.0% in 2019. These two serogroups represented less than half (42.6%) of the total number of confirmed human cases with known serogroups in 2019 (Table 26) . Serogroups O157 and O26 were followed by serogroups O146, O103, O91, O145 and O128 (the latter including variant O128ab). Three new serogroups (O27, O78 and O182) were added to and three serogroups (O5, O55 and O174) were dropped from the top 20 list in 2019. The proportion of serogroups other than O157 increased by 9.2% compared with 2018. The proportion of non-typable STEC isolates increased by 15.0% (75 cases) representing 12.7% of the reported cases with known serogroup in 2019. Data on virulotypes (based on Shiga toxin genes stx1, stx2 and the intimin-coding gene eae) were reported for 49.7% of confirmed STEC infections (7,775) in 2019 by 20 MS. This was a decrease compared with 2018 (62.3%). Virulence genes were reported for 51.4% of 1,853 severe STEC cases (hospitalised, bloody diarrhoea and/or HUS cases). Most isolates (91.2%) were subjected to subtyping of stx genes and 78.5% also had the information on the presence of the eae gene. The most commonly reported virulence gene combination was stx1-/stx2+/eae+, accounting for 42.1% (399/948) of the severe human cases with known virulotypes (Table 28 ). The proportion of the second most common virulotype stx1+/stx2+/eae+ accounted for 30.1% (285/948) of the cases. Together these two virulotypes represented 72.2% of the total number of severe human cases with known virulotypes in 2019. The most common stx gene subtypes were stx1a (261/865; 30.2%), stx2a (222/865; 25.7%), stx2c (182/865; 21.0%) and stx2a;stx2c (100/865; 11.6%), representing 88.5% of the total number of subtypes in severe human cases (Table 29) . This section includes the analysis of the data present in the full data set (Section 4.3.2, Data Analysis), which contained 25,238 sample units tested of which 2.5% (641) were STEC-positives. For analysis of the distribution of STEC serogroups 25 of these 641 isolates, reported by five MS from 1,284 samples, could however not be used because they were obtained using the analytical method ISO 16654:2001 or equivalent methods, which aimed at detecting the serogroup O157 only, so introducing a bias in the descriptive analysis. In total, 23,954 (94.9%) food sample units were reported with analytical method ISO TS 13136:2012 and equivalent methods, which aimed at detecting all STEC, and 616 (2.6%) were STEC-positive (Table 30) . Of these 616 isolates, 212 (34.4%) were provided with the information on the serogroup, which were the data used for the description of STEC serogroups in food. Of these 212 isolates 45 (21.2%) belonged to the top five serogroups (O157, O26, O103, O111 and O145) while the remaining 167 isolates (78.8%) belonged to 53 different O-groups (Table 31) . All the top 20 STEC serogroups isolated from human infections were also found in the STEC isolated from food in 2019 with the exception of serogroups O80, O5 and O76 only found in food (Tables 26 and 27 ). For 404 (65.6%), STEC isolates the only information reported was that the isolate did not belong to the O157 serogroup (88 isolates: 14.3%) or that the serogroup was unspecified. For the characterisation of the virulence genes of STEC strains from food, 641 isolates were available. These data reported from food were still fragmented, as observed in the previous year (EFSA and ECDC, 2019c). Information on stx1 and/or stx2 was provided for 339 (52.9%) STEC strains. Only 185 (28.9%) were reported to have been tested for the presence of the eae gene. The combination of the stx and eae genes was available for 138 isolates (21.5%) ( Table 28) . Thirty-nine STEC isolates (6.1%) were subjected to stx gene subtyping (Table 29 ) and for 11 (1.7%), the information on the presence of the eae genes was reported. Tables 28, 29 and 30 show the combinations of the virulence genes determined in the food isolates and their match with those found in the STEC isolated from severe human disease in the EU in 2012-2017, analysed in the latest pathogenicity assessment of STEC (EFSA BIOHAZ Panel, 2020c). Given the low number of food and animal isolates with the virulence genes characterised in 2019, the figures are displayed in terms of number of isolates instead of the relative frequency for each virulotype. NR: data present in the TESSy data set used, with less than 20 isolates. ND: Not detected. This section includes the analysis of the data present in the full data set (Section 4.3.2, Data Analysis), which contained 2,588 animal sample units tested of which 14.1% (366) were STECpositive. For the analysis of the distribution of STEC serogroups, 108 (29.5%) STEC isolates with information on the serogroups was available. However, 41 of these 108 isolates could not be used because they were obtained with the analytical method ISO 16654:2001 or equivalent methods, which aim at detecting the serogroup O157 only, so introducing a bias in the descriptive analysis. The remaining 67 STEC isolates (18.3%) were obtained by using the ISO TS 13136:2012 method or equivalent, targeting any STEC, which were the data for the description of STEC serogroups (Table 32) . Of these, eight (11.9%) belonged to the top five serogroups while the remaining 59 isolates (88%) belonged to 19 non-top five serogroups, including 10 of the top 20 serogroups isolated from human disease in 2018 (EFSA and ECDC, 2019c) (Table 32 ). For characterisation of the virulence genes of STEC strains from animals, 366 isolates were available but the virulence genes stx and eae were identified and typed only in a small proportion of the reported animal isolates. Out of the 366 STEC isolates reported, 92 (25.1%) were provided with the information on stx1 and/or stx2, but only 36 of these were reported together with the detection of the eae gene (Table 28 ). One MS also carried out stx gene subtyping and reported six STEC strains, four O157, one O26 and one O121, possessing the stx2c; stx2a combination, determined in the STEC O157 isolates and the stx2a subtype alone, found in the other two serogroups. All these isolates were also eae positive (Table 28) . All data provided by the reporting countries were used to generate an atlas of the STEC serogroups identified in the different food and animal categories for the years 2014-2019 ( Figure 1C ) and for 2019 ( Figure 2C ), and is shown in Appendix C. It must be emphasised that the differences in the sampling strategies and, to a lesser extent the analytical methods, applied by reporting countries did not allow confirmation of the existence of specific trends in the geographical distribution of STEC serogroups. 4.5. The number of cases and notification rate of human STEC infections increased notably in 2018, which made STEC the third most commonly reported zoonosis in EU. In 2019, the notification rate was at the same level as in 2018. The long-term trend for human STEC infections showed an increase since 2010, which was mainly due to a large STEC outbreak in 2011. The notification rate stayed at a markedly higher level after the outbreak than before the outbreak. The overall trend of reported cases stabilised after the outbreak but has shown an increase again in the last 5 years during 2015-2019. Part of the observed increase may be explained by improved general awareness of STEC detection following the reporting of large STEC outbreaks. Other contributing factors could probably be the changes in laboratory techniques such as increased diagnostic use of multiplexed molecular assays (PCR) and direct DNA extraction from a specimen followed by isolation and further strain characterisation. More than half of MS national public health laboratories reported having the capacity to perform whole genome sequencing (WGS) for STEC isolates (EFSA BIOHAZ Panel, 2020c). In 2019, 57.9% of the human confirmed cases have been reported with information on the serogroup. This was a slight decrease compared with 2018 when 61.6% of the human isolates had been serogrouped. As in previous years, the most commonly reported serogroup in human cases was O157, followed by O26. The proportion of serogroup O157, however, continued to decrease in 2019, whereas the proportion of non-O157 STEC serogroups has increased over several years. Increasing numbers of laboratories were testing for serogroups other than O157 and there has been a shift in diagnostic methods, with PCR being more commonly used for detection of STEC cases in several MS. Serogroup O26 was the most commonly reported among HUS cases instead of serogroup O157, as it has been since 2016. Over half of the HUS cases caused by this serogroup were reported by two countries (France and Italy), where the surveillance of STEC infections is mainly based on detection of HUS cases. The characterisation of the major virulence determinants such as the Shiga toxin-coding genes (stx) and, to a lesser extent, the intimin-coding eae gene has been indicated to have much more predictive power in terms of pathogenicity potential of STEC strains (EFSA BIOHAZ Panel, 2020c; and the JEMRA and NACMCF reports at Section 4.6 Internet sources) than the serogroups. In this respect, while the last pathogenicity assessment of STEC revolves around the statement that 'all strains are pathogenic to humans, causing at least diarrhoea', a deeper analysis of the virulence genes content, particularly the subtyping of the stx genes, allows identifying some virulence genes combinations (virulotypes), which have a higher frequency of association with severe disease in humans (EFSA BIOHAZ Panel, 2020c). About half of the STEC isolates from all human infections as well as severe human cases (hospitalised, bloody diarrhoea and HUS cases) were reported together with the information on the stx genes (stx1 or stx2) and for the presence of the intimin-coding gene eae. Despite the decrease compared with 2018, when the highest number of virulence gene typing data was reported to TESSy, there has been a steady increase of reporting of stx and eae genes since 2012 (EFSA BIOHAZ Panel, 2020c). Most (> 90%) of the severe human cases were reported with information on stx gene subtypes and 78.5% with data on the presence of the eae gene in 2019. Based on the analysis of the stx subtypes reported to TESSy from 2012 to 2017, all STEC subtypes may be associated with severe illness, albeit at different frequencies. Although stx2a previously showed the highest rates of severe outcome, the stx1a was most frequently associated with severe illness outcomes in 2019 followed by stx2a. Of the STEC cases with known hospitalisation status, more than one-third were hospitalised. Some countries reported very high proportions of hospitalised cases, but had notification rates that were among the lowest, indicating that the surveillance systems in these countries primarily captured the most severe cases. The age group most affected by STEC was infants and children up to 4 years of age, who accounted for two-thirds of the cases of HUS. Most cases of deaths (60%) were, however, reported in age groups > 25 years. Half the deaths were reported to be associated with HUS. In 2019, 22 EU MS reported monitoring results of STEC in 25,030 food samples. Not all reporting MS have tested all food categories equally. By aggregating the food samples into macro-categories in 2019, the number of MS testing and reporting data on the presence of STEC in food ranged from 20 MS reporting the testing for STEC in meat samples to 16 MS and 13 MS testing vegetables (including seeds) and milk and dairy products, respectively. Sprouted seeds were tested by 15 MS, considering the full data set, a number slightly higher than that observed in 2018 (13 MS). Despite the existing microbiological criterion for the presence of STEC in seeds (EU Regulation No 209/2013) , the sampling of this food category in the EU appears to be extremely low. The analytical procedures for food testing in the EU are substantially harmonised. Twenty-one out of the 22 MS reporting data have used the ISO TS 13136:2012 or equivalent methods. There was for 2019 still a residual amount of data being produced by some MS (five) for specific surveys using the ISO 16654:2001 or equivalent methods. These aimed at detecting the serogroup O157 only and do not give information on any other STEC serogroups possibly present in the sample. One MS reported food testing data only obtained with these methods. Additionally, the strategy that the methods for E. coli O157 are based on, revolves around the identification of the serogroup and does not include the determination of the stx gene or of the toxin produced. This laboratory analysis must be actively carried out by MS to confirm that isolated strains are actually STEC. This latter piece of information was missing in the 2019 data set for most O157 isolates. The general extent of contamination of food with STEC observed, 2.8%, was in line with what has been determined in previous years. Monitoring results for STEC contamination in RTE food were described for samples collected according an 'objective' sampling strategy. STEC-positive units were detected in the following RTE foods: in meat and meat products particularly in bovine meat, in milk and milk products notably cheeses, in spices and herbs, in salads and in fruits, vegetables and juices. Importantly, only one-third of the MS or less reported data for certain food categories with a limited sampling effort for certain foods (e.g. two MS reporting 285 sample results for RTE salads). Nevertheless, the finding of STEC-contaminated RTE food commodities is of concern as these are consumed without any treatment to reduce or eliminate the possible presence of the pathogen, posing a direct risk to the consumer. As observed in previous years, different frequencies of contamination with STEC were found in the different major food categories, RTE and non-RTE. The most contaminated food categories included commodities of animal origin, with fresh meat in particular. Small ruminants' meat, including meat from sheep, goats and deer, was the food commodity presenting the highest values (11.6%, 16.7% and 12.9%, respectively). These frequencies, however, may reflect the effect of the few samples tested. As observed in 2018 data (EFSA and ECDC, 2019c), raw cows' milk was the second food category with the highest STEC contamination frequency in 2019, with 3.9% STEC-positive samples. Finally, the 'vegetables and fruit' food category was the less contaminated with 0.1% of positive samples. The characterisation of the food STEC isolates is pivotal for the assessment of the risks for consumers posed by food. In this respect, determination of the serogroup is an important part of this process. Although the recent pathogenicity assessment of STEC (EFSA BIOHAZ Panel, 2020c) affirms that this feature is not an indication of pathogenicity, it still has some importance as an epidemiological marker, and it remains useful to observe the circulation of the different STEC types in food and human cases of disease. In 2019, 34.4% of the food isolates were provided with information on the serogroup, compared with 41.8% in 2018. Of these 21.6% belonged to the 'top-five' serogroups (O157, O26, O103, O111 and O145) whereas more than half of all the remaining STEC belonged to the top 20 STEC serogroups reported in human infections to ECDC in 2015-2018 (EFSA and ECDC, 2019c; Table 27 ). As for the animal monitoring results for 2019, overall, 14.1% of the samples were STEC-positive, compared with 7.6% in 2018. However, the number of animal sampling units tested has been very low in the last years, biasing the estimates. In 2019, this high prevalence may be explained by a very high value of 58.8% STEC-positive pig herds reported by one MS at the farm stage, but most of these are unlikely to involve zoonotic strains (Abubakar et al., 2017; Remfry et al., 2021) . A large increase in the STEC-positive samples, 17%, was also reported for cattle tested in 2019, compared with 3.1% in 2018, but only covered by data from four MS (three in 2018). The animal STEC strains were typed in a lower proportion than the food isolates, with 18.3% of the isolates obtained using the ISO TS 13136:2012 being serotyped. The analysis of the presence and subtypes of virulence genes is important for pathogenicity assessment. Unfortunately, this level of characterisation is still far from being comprehensive for food and animal isolates and only 52.9% of the STEC isolated from food in 2019 have been reported together with the information on the stx gene types (stx1 or stx2) and only 28.9% have been tested for the presence of the intimin-coding gene eae. These figures reduce dramatically to 6.1% and 1.7% when the information on the stx gene subtypes was considered, alone or together with the information on the presence of eae gene, respectively. As this typing and subtyping strategy represents the basis for molecular risk assessment of STEC circulating in the vehicles of infections, MS should be encouraged to expand the adoption of this approach. The analysis of the STEC isolated from food in 2019 showed that many of the virulotypes identified matched those associated with the STEC strains isolated from severe disease (HUS, hospitalisation or bloody diarrhoea) in the EU in the period 2012-2017 (EFSA BIOHAZ Panel, 2020c). Fewer animal STEC isolates were reported data on characterisation of the virulence genes as compared with food isolates. Only six animal isolates were subjected to stx gene subtyping by one MS. Nevertheless, also in this case, many of the virulotypes identified could find correspondence with the same feature of STEC isolated from human severe disease in the 2012-2017 time period (EFSA BIOHAZ Panel, 2020c). The methodologies for typing and subtyping the virulence genes of STEC are available, including those based on WGS, and are supported by external quality assurance (EQA) at the EU National Reference Laboratories level by the EURL for E. coli through its annual inter-laboratory studies scheme. For a greater adoption of the subtyping schemes for STEC to be achieved, it would be of fundamental importance that the cascade of methods distribution and EQA are disseminated down to the Official Laboratories level within the MS (EU Regulation 625/2017). This will provide a wider base of typing and subtyping data for food and animal STEC isolates enabling a deeper risk assessment of STEC in support of actions to be undertaken by the Competent Authorities to mitigate the impact of STEC on public health. Related projects and Internet sources • In 2019, the majority (69.4%) of M. bovis and M. caprae cases in humans was of EU origin (native cases and/or cases originating from other EU MS). Cases were more frequently reported by MS that were not officially bovine tuberculosis free (non-OTF) compared with MS that were officially bovine tuberculosis free in cattle (OTF). • No food-borne disease outbreak due to Mycobacterium spp. has ever been reported to EFSA since the start of the food-borne outbreaks data collection in 2004 and this was also the case for 2019. • Fourteen MS reported to have detected bovine tuberculosis for the year 2019. As in previous years, it was heterogeneous and much spatially clustered with herd prevalence ranging from absence to 11.7% within England, in the United Kingdom. • Seventeen MS were officially bovine tuberculosis free in cattle (OTF) during 2019 and of the 11 non-OTF MS four had OTF regions. • Overall, 143 (0.014%) bovine tuberculosis-infected herds were reported in the OTF regions of 21 MS, which was a rare event, as in previous years. The notification of tuberculosis in humans is mandatory in all EU MS, Iceland, Norway, Liechtenstein and Switzerland and covers the whole population. It has been possible to report M. caprae as a separate species since 2018. France did not report species-specific data within the Mycobacterium tuberculosis complex for the human tuberculosis cases reported in 2019; therefore, no human M. bovis or M. caprae data are available for France. In addition, Latvia did not report any Mycobacterium tuberculosis complex data for 2018 or 2019. Countries may update their data retroactively, and therefore, reported numbers are subject to change in the future or may vary from numbers reported in previous reports. The M. bovis and M. caprae notification rate was calculated using the combined population of the EU MS who reported data in 2019. The proportion of tuberculosis cases caused by M. bovis or M. caprae was calculated using the preliminary estimate of the total number of confirmed tuberculosis cases in 2019 among EU MS reporting species-specific data. As tuberculosis is a chronic disease with a long incubation period, it is not possible to assess travelassociated cases in the same way as for diseases with acute onset. Instead, the distinction is made between individuals with the disease originating from an EU MS (cases of EU origin) and those originating from outside the EU (case originating outside of EU). In the analysis, origin is mainly based on the reported birthplace, except for cases from Austria, Belgium, Greece, Hungary and Poland, whose origin is based on their reported nationality. The treatment outcome for tuberculosis due to M. bovis or M. caprae is assessed 1 year (12 months) after the case notification, as the shortest duration for treatment completion is 6 months according to the international treatment guidelines of tuberculosis. Bovine tuberculosis monitoring data from bovine animals originating from the National Control and Eradication Programmes and/or Officially Free status According to the Zoonoses Directive 2003/99/EC, MS must report bovine tuberculosis annual monitoring data. These data originate from national control and surveillance programmes implemented by the MS in accordance with EU legislation. The reports submitted by the MS are based on Council Directive 64/432/EEC and subsequent legislation and are essential for the assessment of the epidemiological situation in MS and MS regions, whether declared officially bovine tuberculosis free in cattle (OTF) or not yet declared OTF. Annual surveillance programmes are carried out in OTF regions to confirm freedom from bovine tuberculosis, whereas in all non-OTF regions control and eradication programmes for bovine tuberculosis are in place. These data are comparable across MS because the monitoring schemes are harmonised, and the data collected and reported to EFSA originate from the census as the sampling frame. In addition to trend analysis both at the EU level and at MS level and to trend watching and descriptive summaries, these data may also be used to assess the impact of control and eradication programmes (Table 1) . EU MS also need to notify outbreaks of bovine tuberculosis in terrestrial animals from OTF regions to the EU Animal Disease Notification System (ADNS) 11 and regular summaries are posted online. For bovine tuberculosis cases, all tuberculosis cases irrespective of their causative agent (i.e. also including those caused by M. caprae) are included in the statistics provided by MS, as opposed to the procedure for the above-mentioned statistics for humans, in which cases by M. bovis and M. caprae are separated. Based on the definition recommended by the bovine tuberculosis subgroup of the task force on monitoring animal disease eradication of the EU (SANCO/10200/2006), who made it explicit that all cases of tuberculosis in cattle due to a disease-causing member of the M. tuberculosis complex are to be considered as a case of bovine tuberculosis, all available information on the specific bacterial species part of the M. tuberculosis complex recovered from cattle was taken into account to summarise the EU situation on bovine tuberculosis. A distinction is made descriptively, whenever possible, of reporting by MS on Mycobacterium tuberculosis complex, M. bovis and M. caprae. Mycobacterium monitoring data from food and from animals other than bovine animals submitted to EFSA according the Zoonoses Directive 2003/99/EC and collected without harmonised design allow for descriptive summaries at the EU level to be made. They preclude trend analyses and trend watching at the EU level (Table 1) When the UK data were collected, the UK was an EU MS but as of 31 January 2020, it has become a third country. In 2019, there were 147 confirmed human cases of tuberculosis due to M. bovis or M. caprae reported by 26 EU MS (Table 34) . Of these cases, 136 were due to M. bovis and 11 were due to M. caprae. The 11 M. caprae cases were reported by Austria (n = 2), Germany (n = 3) and Spain (n = 6). Between 2015 and 2019, the number of M. caprae cases notified each year has ranged between nine (in 2017) and 13 (in 2018). Overall, M. bovis and M. caprae cases accounted for only 0.3% of total tuberculosis cases reported by the 26 EU MS with species-specific data within the Mycobacterium tuberculosis complex in 2019. Ten MS reported at least one confirmed case and 16 MS did not report any cases. The EU notification rate in 2019 was 0.03 cases per 100,000 population, which was slightly lower than the rate in the previous 4 years. The highest notification rate in 2019 was reported by Ireland (0.14 per 100,000), followed by Spain (0.07 per 100,000) (Table 34) . There were 17 EU MS that had OTF status (OTF, officially bovine tuberculosis free in cattle) in 2019, and, of these, 15 reported on species of the M. tuberculosis complex. The notification rate of human M. bovis and M. caprae cases among these 15 EU MS was 0.03 cases per 100,000 population. In the non-OTF EU MS, the notification rate was 0.04 cases per 100,000 population. Most cases, 69.4% (102/147), reported in 2019 were of EU origin (native cases and/or cases originating from other EU MS). The remaining cases originated from outside the EU (26.5%, n = 39), or had unknown origin (4.1%, n = 6) (Table 33 ). Cases were more likely to have originated from non-OTF EU MS (66.7%, n = 68) than from OTF EU MS (33.3%, n = 34). Treatment outcome after 12 months was reported for 90.1% (n = 164/181) of the human M. bovis and M. caprae cases reported in 2018. Among these cases, successful treatment was reported for 96 cases (58.5%), 22 cases (13.4%) died, five cases (3.0%) were still on treatment, one case (0.6%) was reported to have treatment failure and two cases (1.2%) were lost to follow-up. The treatment outcome was not evaluated for 38 cases (23.2%). Drug resistance to isoniazid and rifampicin among human M. bovis or M. caprae cases remained low in 2019; among 105 cases with test results reported for both isoniazid and rifampicin, only four were isoniazid-resistant (3.8%). No multidrug-resistant (resistance to rifampicin and isoniazid) cases were reported. Figure 32 shows, for the year 2019, the number of confirmed tuberculosis cases due to M. bovis and to M. caprae in individuals of EU origin overlaid with the national aggregated herd prevalence of bovine tuberculosis. No food-borne disease outbreak due to Mycobacterium spp. was reported for 2019 in EU and no single such food-borne outbreak has been reported to EFSA since the start of the food-borne outbreaks reporting, in 2004. No Mycobacterium monitoring data from food were submitted for the year 2019. The country status on 31 December 2019 of freedom from bovine tuberculosis (OTF) is presented in Figure 33 and in Norway and Switzerland were OTF, in accordance with EU legislation. Liechtenstein has the same status (OTF) as Switzerland. In Iceland, which has no special agreement on animal health status with the EU, the last outbreak of bovine tuberculosis was reported in 1959. Montenegro and the Republic of North Macedonia also reported data on bovine tuberculosis in their cattle. During 2019, the overall EU proportion of cattle herds infected with, or positive for, bovine tuberculosis remained very low (0.8%, which was 16,420 out of 1,961,990 herds). Fourteen MS reported no case of bovine tuberculosis in cattle; Belgium, Cyprus, Czechia, Denmark, Estonia, Finland, Latvia, Lithuania, Luxembourg, the Netherlands, Malta, Slovakia, Slovenia and Sweden (Table 35) . Bovine tuberculosis was reported by 14 MS and was heterogeneous and much spatially clustered with herd prevalence ranging from absence to 11.7% within the United Kingdom in England. In the OTF regions of the 21 MS with such regions, there were in total 1,059,412 cattle herds. Seven of these MS reported in total 143 (0.014% overall) bovine tuberculosis-infected herds (Table 35) No region of the MS is OTF. During 2019, the 11 non-OTF MS had in total 902,578 cattle herds in their non-OTF regions. Nine of these MS reported in total 16,277 (1.803% overall) bovine tuberculosis-positive herds (Table 35) In the EU, tuberculosis due to M. bovis or M. caprae is rare in humans because of decades of disease control in cattle and routine pasteurisation of cow's milk. In 2019, human M. bovis and M. caprae cases represented only a small proportion (0.3%) of all notified human tuberculosis cases in the 26 EU MS that reported on the causative species. The notification rate of M. bovis and M. caprae in humans was slightly higher for the non-OTF EU MS than in the OTF EU MS (0.04 vs. 0.03 per 100,000 population, respectively). During 2019, the overall EU proportion of cattle herds infected with, or positive for, bovine tuberculosis was 0.8%. Bovine tuberculosis was reported by 14 MS and was heterogeneous and much spatially clustered with herd prevalence ranging from absence to 11.7% within the United Kingdom in England. This demonstrates that the situation in Europe on bovine tuberculosis infection, detection and control remained heterogeneous, as substantiated by EFSA (EFSA AHAW Panel, 2014). Seventeen MS were OTF and in addition four non-OTF MS had OTF regions. Twelve of these 21 MS reported no case of bovine tuberculosis in cattle. In these OTF regions, the detection during 2019 of bovine tuberculosis-infected herds remained a rare event, as in the previous years. From 2010 to 2019, the overall annual number of infected cattle herds, the prevalence and the total number of cattle herds decreased. All 11 non-OTF MS except Cyprus and Malta detected bovine tuberculosis during 2019 in their non-OTF regions and overall, about one in 50 herds were positive. When comparing 2019 with 2018 data, the overall annual number of positive cattle herds, the prevalence and the total number of cattle herds all decreased in these non-OTF regions. When comparing 2010 to 2019 data, the overall annual number of reported positive cattle herds in these non-OTF regions proportionally decreased by 8.6%, whereas the prevalence increased by 72.1%. Concomitantly, the total number of cattle herds in those non-OTF regions was reduced to about half (decreased by 44.9%). This increase in prevalence can partly be explained by the increase in test-positive cattle herds being detected in these regions along with an important decrease in the total number of cattle herds due to the gradual declaration of OTF status in regions within non-OTF MS over this period. This overall increase can be further explained by specific trends in few non-OTF MS during recent years. In the United Kingdom, M. bovis is widespread in England and Wales and in Northern Ireland and an epidemic in cattle has been ongoing for many years. A summary presentation on the situation can be found online. 13 A major constraint to bovine tuberculosis eradication in cattle in those areas in which the infection is endemic in the Eurasian badger (Meles meles): this native wildlife species is a maintenance host of M. bovis. The challenge to successfully tackle bovine tuberculosis is also to address the reservoir of the disease in wildlife. Bovine tuberculosis remains one of the most serious and costly animal health problem for the UK cattle industry and taxpayer. Ireland also has for many years faced the challenge of containing the spread of bovine tuberculosis. It introduced a badger vaccination policy in 2018 and is also, among other control measures, reducing the badger population. A summary presentation on the situation in Ireland can be found online. 14 Stagnating or increasing trends in the prevalence of bovine tuberculosis-positive cattle herds demonstrate that control and eradication of this disease is a challenge, owing to the complex interactions between the pathogen, hosts and the local environments (EFSA AHAW Panel, 2014). MS-specific evaluations of status, trends and of the relevance of bovine tuberculosis as a source of disease for humans can be found in the 2019 Annual National Zoonoses Country Reports referenced in Section 5.5. In 2019, M. bovis was reported to be isolatedapart from bovine animalsfrom a wide range of animal species, both domestic and wild, reflecting that this causative agent of tuberculosis in cattle has a broad host range. M. caprae, recognised to cause bovine tuberculosis, was reported in cattle but also in farmed red deer. • There was a significantly declining EU/EEA trend in the number of confirmed brucellosis cases from 2015 to 2019. • Despite the declining trend, Greece reported the highest notification rate (0.61 cases per 100,000 population) of the domestic brucellosis cases in the EU followed by Portugal (0.32 cases per 100,000 population). • Most confirmed human cases (98 cases) were hospitalised and two deaths were reported in 2019. • One food-borne brucellosis outbreak was reported for 2019 in EU, due to raw milk. During 2005-2018, there were 16 food-borne outbreaks due to Brucella reported in EU, of which four were due to cheeses and 12 reported due to 'unknown' food. • Compared with 2018, the total number of Brucella-positive or -infected cattle herds and sheep and goat flocks in the not officially free regions further decreased by 14% and by 27%, respectively. • Brucellosis in cattle and in sheep and goats is still endemic in Italy, where the prevalence is highest in the southern region of Sicily, in Greece and in Portugal. In Italy and Portugal, the proportion of brucellosis-positive cattle herds and sheep and goat flocks in not officially free regions decreased during recent years. • Greece reported the highest notification rate of confirmed cases in humans, 10 times higher than the EU average, while at the same time reporting an enzootic situation in animals: 2.8% infected cattle herds and 3.3% infected sheep and goats herds on the Greek islands whereas from Continental Greece data were lacking. • Brucellosis is still an animal health problem with public health relevance in southern Europe/in countries that are not officially free of brucellosis. Surveillance and monitoring of Brucella in the EU Notification of brucellosis in humans is mandatory in 26 MS, Iceland, Norway and Switzerland. In Denmark, no surveillance system is in place for brucellosis and the disease is not notifiable nor reported at the EU level. Belgium has another (not specified) system. The surveillance systems for brucellosis cover the whole population in all MS reporting cases. For 2019, Spain did not receive data from all regions and rates are therefore not displayed for this year. All countries reported case-based data except Belgium and Bulgaria, which reported aggregated data. Both reporting formats were included to calculate numbers of cases, notification rates. Brucella monitoring data from bovine animals and sheep and goats originating from the National Control and Eradication Programmes and/or Officially Free status According to the Zoonoses Directive 2003/99/EC, MS must report bovine brucellosis and sheep and goat brucellosis annual monitoring data. These data originate from national control and surveillance programmes implemented by the MS in accordance with EU legislation. The reports submitted by the MS are based on Council Directive 64/432/EEC and subsequent legislation and are essential for the assessment of the epidemiological situation in MS and MS regions, whether declared officially brucellosis free in cattle (OBF) and/or officially B. melitensis free in sheep and goats (ObmF). Annual surveillance programmes are carried out in OBF regions to confirm freedom from bovine brucellosis and in ObmF regions freedom from B. melitensis in sheep and goats, whereas in all non-OBF and non-ObmF regions control and eradication programmes for brucellosis in cattle and in sheep and goats are in place. These data are comparable across MS because the monitoring schemes are harmonised, and the data collected and reported to EFSA originate from the census as sampling frame. In addition to trend analysis both at the EU level and at MS level and to trend watching and descriptive summaries, these data may also be used to assess the impact of control and eradication programmes (Table 1) . EU MS also need to notify outbreaks in terrestrial animals of bovine brucellosis and of caprine and ovine brucellosis (excluding Brucella ovis) in their OBF and/or ObmF regions to the EU ADNS 12 and regular summaries are posted online. Brucella monitoring data from food and from animals other than bovine animals and sheep and goats, submitted to EFSA according to the Zoonoses Directive 2003/99/EC and collected without harmonised design allow for descriptive summaries at the EU level to be made. They preclude trend analyses and trend watching at the EU level (Table 1 ). The reporting of food-borne brucellosis outbreaks in humans is mandatory according to the Zoonoses Directive 2003/99/EC. Table 36 summarises EU-level statistics on human and animal brucellosis and on food investigated for Brucella, during 2015-2019. A more detailed description of these statistics is in the results section of this chapter and in the food-borne outbreaks chapter. Reported food data of interest were categorised in the major category 'milk and milk products' and aggregated by year over the period 2015-2019 to obtain an overview, by year, of the amount of data sent. The numbers of sampling units reported, and the number of reporting MS are extremely low. The annual animal data statistics displayed in Table 36 When the UK data were collected, the UK was an EU MS but as of 31 January 2020, it has become a third country. In 2019, 27 MS provided data and information on brucellosis in humans. In total, 319 cases were reported in the EU. These included 310 confirmed cases, which was a decrease by 13.4% compared with 2018. The notification rate was 0.06 cases per 100,000 population (Table 37) , which represented The highest notification rates of brucellosis were reported in two MS that were non-OBF and/or non-ObmF (Table 37) : Greece and Portugal (0.61 and 0.32 cases per 100,000 population, respectively). The lowest notification rates were observed in OBF and ObmF MS where brucellosis cases were mainly travel-associated. Slovenia and Sweden which have the OBF/ObmF status and had a relatively high notification rate (0.29 and 0.14 cases per 100,000 population, respectively) reported that the majority (≥ 75.0%) of confirmed brucellosis cases was travel associated. Most brucellosis cases (71.6%) with known data on travel were reported as being infected in the EU (Table 36) . Among the 56 travel-associated cases with known travel destination, 50 (89.3%) travelled outside EU. The most common travel destinations of the imported cases outside the EU were Iraq 12 cases (21.4%), Turkey 10 cases (17.9%), Bosnia and Herzegovina five cases (8.9%) and Egypt three cases (5.4%), respectively. In the EU, three cases reported travel to Spain and one case reported travel to Romania during the incubation period. Eleven MS provided data on hospitalisation, accounting for 44.5% of confirmed cases in the EU. On average, 71.0% of the confirmed brucellosis cases with known status were hospitalised. In seven of the 11 countries reporting hospitalisation, the proportion of hospitalised cases ranged between 80% and 100%. Two deaths due to brucellosis were reported among 114 confirmed cases (36.8%) with outcome information by the 12 MS; one by the Netherlands and one by Spain in 2019. Brucella species information was missing for 63.8% of the 310 confirmed cases reported in the EU. Of the 111 cases with known species, 105 (94.6%) were infected by B. melitensis, three (2.7%) by B. abortus and one (0.9%) by B. suis. Figure 38 shows the number of domestically acquired (having not been outside the country of notification during the incubation period of the disease) confirmed brucellosis cases in humans overlaid with the national prevalence of Brucella-positive cattle herds and sheep and goat flocks in EU/EFTA in 2019. The map shows that Greece, Portugal and Spain (human data not reported in all regions in 2019) have a higher number of domestically acquired confirmed brucellosis cases in humans and a higher prevalence of Brucella-positive ruminant herds. Italy, which has reported a high number of human brucellosis cases over the years, did not report the origin of the infections in 2019. Source: Austria, Cyprus, Czechia, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Latvia, Lithuania, Malta, the Netherlands, Norway, Poland, Portugal, Romania, Slovakia, Slovenia and Sweden. Belgium, Bulgaria, Croatia, Luxembourg, Spain and the United Kingdom did not report data to the level of detail required for the analysis. Denmark does not have a surveillance system for this disease. Human brucellosis cases associated with food-borne outbreaks Table 38 summarises reported brucellosis outbreaks data during 2005-2019, by MS and by incriminated food vehicle. Austria reported for the year 2019 one food-borne outbreak due to Brucella melitensis in raw milk that was consumed in Turkey by the two affected persons. 15 During 2005-2019, overall 17 brucellosis food-borne outbreaks were reported, of which four with strong-evidence were due to cheese, one with strong-evidence due to raw milk and 12 with weak evidence due to 'unknown' food. Further details and statistics on the food-borne outbreaks for 2019 are in the food-borne outbreaks chapter. Very few 2019 Brucella monitoring data from food were submitted; in total from 586 milk and milk products sampling units, by Italy (78.8%, N = 462) and Portugal (21.2%, N = 124). In total, 15 Italian samples from a processing plant from 'milk from other animal species or unspecifiedpasteurised milk' tested positive for Brucella spp. with reported species: B. abortus biovar 3, B. melitensis biovar 3 and Brucella unspecified spp. The country status on 31 December 2019 of freedom from bovine brucellosis (OBF) is presented in Figure 39 and in Table 39 Four non-OBF MS had no OBF region: Bulgaria, Croatia, Greece and Hungary. Norway, Switzerland and Liechtenstein were OBF in accordance with EU legislation. Iceland, which has no special agreement on animal health (status) with the EU, has never reported brucellosis due to B. abortus, B. melitensis or B. suis. Montenegro and the Republic of North Macedonia also reported data on brucellosis in their cattle. During 2019, the overall EU proportion of cattle herds infected with, or positive for, bovine brucellosis remained a very rare event (0.025%, which was 489 out of 1,942,294 herds). Twenty-three MS reported no case of brucellosis in cattle. Bovine brucellosis was reported by five MS: Austria, Croatia, Greece, Italy and Portugal (Table 39) . In the OBF regions of the 24 MS with such regions, there were in total 1,650,343 cattle herds in 2019. Austria reported to have detected brucellosis due to B. melitensis in one cow in a herd in the context of a follow-up investigation of an outbreak in 2018. Italy reported three positives herds. Bovine brucellosis was not detected in 2019 in the non-MS: Iceland, Norway, Switzerland and Liechtenstein. During 2012-2019, there had been, respectively, nine, two, two, four, two, zero, three and four cattle herds reported infected in OBF regions in EU, meaning these were extremely rare events. In conclusion, in 2019, bovine brucellosis was mainly still present in a few MS, Greece, Italy and Portugal, in southern Europe. Sicily, in Italy, reported the highest regional prevalence in EU non-OBF regions, with 2.3% positive herds. From 2012 to 2019, the overall annual number of reported positive cattle herds in the non-OBF regions decreased by 58.9% from 1,181 to 485, whereas the prevalence increased by 63.6% from 0.10% to 0.17% (Figure 40 ). The latter is due to the drastic decrease in the total number of cattle herds from 1,162,978 to 291,951 during the same period, i.e. a decrease of 74.9%. When comparing 2019 with 2018 data, the annual number of positive cattle herds, the prevalence and the total number of cattle herds decreased by 13.9%, 7.8% and 6.5%, respectively. The country status on 31 December 2019 of freedom from ovine and caprine brucellosis by B. melitensis (ObmF) is presented in Figure 42 and in • in Italy: 13 regions and four provinces; • in Portugal: the Azores region (all nine islands); • in Spain: 13 autonomous communities and eight provinces. Four non-ObmF MS had no ObmF region: Bulgaria, Croatia, Greece and Malta. Norway, Switzerland and Liechtenstein were ObmF in accordance with EU legislation. Iceland, which has no special agreement on animal health (status) with the EU, has never reported brucellosis due to (Table 40) . In the ObmF regions of the 24 MS with such regions, there were in total 941,317 sheep and goat flocks in 2019 and one case of brucellosis was reported in these herds during 2019, by Italy. B. melitensis was not reported in 2019 by the non-MS: Iceland, Norway, Switzerland and Liechtenstein. During 2012-2019, there has been, respectively, 5, 4, 3, 10, 2, 7, 0 and 1 sheep and goat flocks reported infected in ObmF regions, meaning it was an extremely rare event. In conclusion, in 2019, B. melitensis brucellosis in sheep and goat flocks was mainly still present in a few MS, Greece, Italy and Portugal, in southern Europe. Sicily, in Italy, reported the highest regional prevalence in EU non-OBF regions, with 1.6% of positive herds. From 2012 to 2019, the overall annual number of reported positive sheep and goat flocks in the non-ObmF regions decreased by 73.4% from 1,693 to 451, whereas the prevalence decreased by 53.2% from 0.45% to 0.21% (Figure 43 ). The total number of sheep and goat flocks decreased by 43.1% from 377,690 to 214,782 during the same period. When comparing 2019 with 2018 data, the annual number of brucellosis-positive sheep and goat flocks, the prevalence and the total number of herds, respectively, decreased by 27.3%, 6.3% and 22.4%. Complementary to 2019 reports from cattle and from sheep and goats, Brucella species were reported from a wide range of animal species: Brucella unspecified species from 'farmed animals', dogs, pigs, rabbits and wild boars; B. suis from pigs and wild boars and notably biovar 2 from wild deer, wild hares, breeding pigs, pigs from mixed herds not raised under controlled housing conditions and wild boars; B. canis from dogs (pet) and B. pinnipedialis from wild seals. Brucellosis is a rare disease in the EU, although severe with most of the diagnosed human cases hospitalised. In 2019, the number of reported confirmed cases of brucellosis in humans and the EU notification rate was at the lowest level since the beginning of EU-level surveillance in 2007. During 2019, the highest notification rates and most of the domestic brucellosis cases were reported from two MS, Greece and Portugal, that are not officially brucellosis free in cattle, sheep or goats. These two countries accounted for 32% of all confirmed brucellosis cases in the EU and consistently reported the highest notification rates within the EU despite the declining trends in Greece since 2014 and in Portugal since 2009. Greece continued reporting a notification rate over 10 times higher and Portugal over five times higher than the EU average. An outbreak of B. melitensis from home-made fresh goat cheese, sold outside of the commercial circuit, was reported in the northern region of Portugal in Mendes et al. (2020) . In Italy, a general decrease of cases has been notified in all regions in the last 20 years and its notification rate was for the first time in 2019 similar to the EU average. Brucellosis remains, however, an important health problem particularly in southern part of Italy, reporting 89% of the annual cases (Facciol a et al., 2018) . Greece, Italy and Portugal were the southern European MS where bovine brucellosis and B. melitensis brucellosis in sheep and goat flocks were still present in 2019, with Sicily, in Italy, reporting the highest regional prevalence in bovine animals, and in sheep and goats. These findings underline that brucellosis is still an animal health problem with public health relevance in these southern European MS. Bovine brucellosis and ovine and caprine brucellosis have been eradicated by most EU MS. In MS and regions officially free of brucellosis, no infected herds were reported for the year 2019, except for one B. melitensis-infected cattle herd in Austria and four positive herds in Italy (three in cattle and one in small ruminants). Reported food-borne disease outbreaks due to Brucella have become rare in the EU. For the year 2019 one single food-borne outbreak due to B. melitensis was reported by Austria, due to unpasteurised milk consumed in Turkey. 16 As regards autochthonous human foodborne illnesses in MS that are officially free of brucellosis, the question is raised as to the origin of these infections. Food-borne exposure is normally limited to people consuming unpasteurised milk or dairy products from countries where brucellosis in animals is endemic. A recent study published by Jansen et al. (2019) based on samples from 2011 in Germany found Brucella-positive raw milk cheeses were available at German retail level, so putting consumers at risk without travel history to endemic countries. The authors hypothesised that, in Germany, which is officially free of Brucella in cattle, sheep and goat populations, there are uncontrolled imports of cheese (from endemic regions) that do not comply with food safety standards. The above outbreak in northern Portugal was a further episode adding to the concern of illegal trade of raw milk cheese and challenging food safety standards in EU. As a result of the eradication of animal brucellosis in most EU MS, human brucellosis has become quite rare in northern and western Europe, where most cases are associated with travel outside EU. In some northern European countries (Germany, France, Sweden and Norway) an increased disease incidence may occur in recently arrived migrants (Garofolo et al., 2016; Mailles et al., 2016; Norman et al., 2016; Georgi et al., 2017; Johansen et al., 2018) , a large part arriving from endemic countries (Africa, Middle East and Mediterranean countries). In France, a case report described the first case of brucellosis caused by an isolate whose genome is identical that of a frog isolate from Texas, demonstrating the zoonotic potential of amphibian-type Brucella inopinata (Rouzic et al., 2020) . This patient hospitalised with an altered general status, dyspnoea, night fever presented mediastinal lymphadenopathies, pulmonary condensations, emphysematous lesions and splenomegaly. Importantly, with such atypical Brucella, correct diagnosis cannot be performed using routine serological tests or identification methods. Some MS were not officially free of bovine brucellosis and/or brucellosis in sheep and goats, and both infections were still mainly present in 2019 in Greece, Italy and Portugal. The highest regional prevalence for both infections was reported for Sicily, in southern Italy, representing an ongoing public health threat as evidenced by the fact that 89% of human cases in Italy are reported in Sicily (Facciol a et al., 2018) . Greece and Portugal also reported the highest rates of confirmed human cases in 2019, respectively, 10 and 5 times higher than the EU average. At the same time, 2.8% cattle herds and 3.3% sheep and goat flocks were test-positive on the Greek islands, being from mostly unvaccinated herds. From mainland Greece, where vaccination programmes are run against both brucellosis in cattle (in mountainous areas) and sheep and goats (on the mainland and some bigger islands), no animal test data were reported. Non-food-borne transmission of brucellosis to humans also occurs, through direct contact with infected animals. People working with farm animals, including farmers, livestock breeders, butchers, abattoir workers and veterinarians, are known to be at increased risk of brucellosis in the endemic countries. The largest proportion of the human cases in EU MS occurred in workingage men, possibly indicating occupational exposure . This finding is in agreement with a recent study in Greece by Fouskis et al. (2018) , in which male patients were found to be related to high-risk jobs and animal contact, while brucellosis in women was related to recent consumption of dairy products. As compared with 2018, overall in the EU regions not officially free from bovine brucellosis the number of positive herds and the prevalence of bovine brucellosis decreased, respectively, by 14% and 8% in 2019, whereas in the regions not officially free from brucellosis in sheep and goats those proportions also decreased, respectively, by 27% and 6%. In Italy and Portugal, the prevalence of bovine, ovine and caprine brucellosis in not officially free regions has decreased in recent years. Croatia and Spain reported almost no positive herds during the last two years for these infections, meaning that in the coming years eradication of cattle and sheep and goat brucellosis is within reach. It is of note that compared with Spain, the situation is different for Croatia. In Croatia, cases in humans have been sporadic and low in prevalence and emerged only in animals and humans living close to the border of Bosnia and Herzegovina. In Bosnia and Herzegovina, the disease is enzootic. 18 These findings support the assumption that the illegal import of animals is the main source of brucellosis in the country (Duvnjak et al., 2018) . The most recent data on the incidence of brucellosis in humans in south-east Europe (Balkan countries) proved the persistence of brucellosis in the area. Bulgaria reported re-emergence of human brucellosis to the country, most probably related to import of infection from endemic areas in the near neighbouring countries, Greece and Macedonia (Karcheva et al., 2017) . In food, very few monitoring data were reported during these last years by the non-OBF/ObmF MS Italy and Portugal. Italy reported, for 2019, positive findings in pasteurised milk 'from other animal species or unspecified' at processing plants. Greece did not submit food monitoring results for Brucella. Related projects and Internet sources • The number of reported food-borne trichinellosis outbreaks was 5, compared with 10 in 2018, with 44 illnesses, 12 hospitalised people and no deaths. Most outbreaks were caused by pig meat and products thereof, as during previous years. • Trichinella spiralis was the only species that was reported from confirmed human cases to TESSy. Species reported to EFSA from food were T. spiralis from pig meat and products thereof in one food-borne outbreak in Croatia and one in Romania and T. britovi from other or mixed red meat and products thereof in one food-borne outbreak in Italy. The notification of Trichinella infections in humans is mandatory in all MS, Iceland, Norway and Switzerland, except in Belgium, France and the United Kingdom where surveillance systems are voluntary. No surveillance system for trichinellosis exists in Denmark. The surveillance systems for trichinellosis cover the whole population in all MS except in Belgium. All countries reported case-based data except Belgium, Bulgaria and the Netherlands, which reported aggregated data. Both reporting formats were included to calculate numbers of cases and notification rates. For 2019, Belgium did not report data and Spain did not receive data from all regions due to COVID-19. Rates are therefore not displayed for Spain for 2019. In humans, diagnosis of Trichinella infections is primarily based on clinical signs and symptoms and serology (indirect enzyme-linked immunosorbent assay (i-ELISA) and western blot). Histopathology on muscle biopsies is very rarely performed. Tables and figures that are not presented in this chapter are published as supporting information to this report and are available as downloadable files from the EFSA knowledge junction at zenodo https://doi.org/ 10.5281/zenodo.4298993. Summary statistics of human surveillance data with downloadable files are retrievable using ECDC's Surveillance Atlas of Infectious Diseases at http://atlas.ecdc.europa.eu/public/index. aspx Trichinella monitoring data from domestic pigs (both fattening and breeding animals), farmed wild boar and solipeds According to the Commission Implementing Regulation (EU) 2015/1375 19 , all Trichinella-susceptible animals intended for human consumption in the EU, i.e. domestic pigs (both fattening and breeding animals), farmed wild boar and solipeds, should be tested for the presence of Trichinella larvae in the muscles unless carcases have undergone a freezing treatment (freezing inactivates the parasite). It follows that data on Trichinella infections in these animals are comparable across MS because the monitoring schemes are harmonised and the data collected and reported to EFSA originate from census sampling (Table 41) . Domestic pigs, farmed and hunted wild boar and other wild animals (e.g. bears) that are not processed to be placed on the EU market (e.g. intended for own consumption) are exempted from the Commission Implementing Regulation (EU) 2015/1375 and their control falls under the national legislation. Commission Implementing Regulation (EU) 2015/1375 states that reporting of data for domestic pigs shall, at least, provide specific information related to the number of animals tested that were raised under controlled housing conditions as well as the number of breeding sows, boars and fattening pigs tested. Further, the regulation states that a negligible risk status for a country or region is no longer recognised. Trichinella monitoring data from animals other than domestic pigs, farmed wild boar and solipeds MS should monitor the circulation of these nematodes in the main natural reservoir hosts (carnivore and omnivore animals) to acquire information on the risk of transmission to domestic animals (and from these to humans) and on the introduction of new Trichinella species from non-EU countries. However, monitoring data provided by the MS to EFSA are generated by non-harmonised monitoring schemes across MS without mandatory reporting requirements. Wild animals are the main reservoir hosts of Trichinella, and their biology and ecology vary from one MS to another and from one region or habitat in the same MS to another due to the human and environmental impact on the ecosystems, resulting in different transmission patterns and prevalence of infection. Therefore, data from Trichinella in wild animals are not fully comparable between MS and the reported findings must be interpreted with caution. These data allow descriptive summaries at the EU level but preclude subsequent data analysis such as assessing temporal and spatial trends (Table 1) . Table 41 summarises EU-level statistics on human trichinellosis and on Trichinella in animals, during 2015-2019. Animal data of interest reported were classified into categories and aggregated by year to obtain an annual overview of the volume of data submitted. More detailed descriptions of these statistics are in the results section of this chapter and in the chapter on food-borne outbreaks. When the UK data were collected, the UK was an EU MS but as of 31 January 2020, it has become a third country. In 2019, 140 cases of trichinellosis, including 96 confirmed cases, were reported by 26 MS (Table 42) . There was about 50% increase in case numbers and the EU notification rate doubled from 0.01 cases per 100,000 population in 2018 to 0.02 cases per 100,000 population in 2019. Despite the increased number of cases in 2019 compared with 2018, the number of cases was below the 5-year average (117 cases). The increase was mainly due to the increased number of confirmed cases in three MS; Bulgaria (+10), Italy (+8) and Spain (+9). Together, these three countries accounted for 79.2% of all confirmed cases reported at the EU level in 2019. Bulgaria had the highest notification rate in the EU (0.79 cases per 100,000). Fourteen MS reported zero confirmed cases in 2019 including four MS (Cyprus, Finland, Luxembourg and Malta) that have never reported any trichinellosis cases. In 2019, 26 cases (27.1%) of trichinellosis cases with known travel status and with known country of infection were reported to be acquired in the EU (Table 41 ). Four MS reported five travel-associated trichinellosis cases of which two cases were infected outside the EU and one case infected within the EU. For 66 cases (68.7%), travel information was not reported. The EU/EEA trend in confirmed cases of trichinellosis has substantially been influenced by a number of smaller and larger outbreaks, often with peaks in January-February (Figure 45 ). The EU/EEA trend was significantly declining in 2015-2019. Romania reported a decreasing trend and none of the MS observed significantly increasing trend during the same time period. Bulgaria, which reported most of the cases and highest notification rate in the EU in 2015-2019 was not included in the EU trend calculations since monthly data were not available. Of the 12 MS reporting confirmed cases for 2019, five provided information on hospitalisation (16 cases, 16.7% of all confirmed cases reported in the EU). Among these, six cases (37.5%) were hospitalised, which was a decrease compared with 2018 (64.2%). Seven MS provided information on the outcome of their cases (24 cases, 25.0% of all confirmed cases). One death due to trichinellosis was reported in 2019 resulting in an EU case fatality of 4.2%. Species information was available for 22 (22.9%) of the reported confirmed cases from six MS. The only species reported to TESSy from confirmed human cases was T. spiralis. Species reported to EFSA from food were T. spiralis from pig meat and products thereof in one food-borne outbreak in Croatia and one in Romania and T. britovi from other or mixed red meat and products thereof in one foodborne outbreak in Italy (see below). Overall, for the year 2019, the number of reported human trichinellosis cases infected within the EU was 25, two cases contracted the infection outside EU and 68 cases were reported with unknown travel information (Table 41) . Overall Trichinella was identified by four MS in five outbreaks, which were all strong-evidence outbreaks and which together affected 44 people in`EU, with 12 hospitalised and no deaths, as reported to EFSA. Comparing the number of food-borne outbreak cases (44) reported to EFSA and the number of cases of human trichinellosis acquired in the EU (25) reported to ECDC, considering also the proportion of unknown travel data (0.926 9 68), reported to ECDC, could suggest that overall, in 2019, 50% of human trichinellosis cases in the EU would be reported through food-borne outbreak investigation. In this context, it is important to clarify that the case classification for reporting is different between these two databases. In TESSy, the cases reported are classified based on the EU case definition. All these cases visited a doctor and are either confirmed by a laboratory test (confirmed case) or not (probable case and classification is based on the clinical symptoms and epidemiological link). Cases who never visited a doctor are not reported to TESSy. Moreover, probable cases may be missing in TESSy, as these data are not analysed or published and there is no incentive for reporting such cases. Information on which cases are linked to an outbreak and which not is also not systematically collected. In practice, the cases reported to TESSy are considered to be mostly sporadic cases. In food-borne outbreaks, human cases are persons involved in the outbreak as defined by the investigators (case definition), and cases must be linked, or probably linked, to the same food source (Directive 2003/99/EC). This can include both ill people (whether confirmed microbiologically or not) and people with confirmed asymptomatic infections . Cases can be classified as confirmed or probable outbreak cases, but currently these specific classification data are not collected by EFSA. All five Trichinella food-borne outbreaks (Tables 41 and 43) were reported as strong-evidence outbreaks. They were reported by Bulgaria (two), Croatia (one), Italy (one) and Romania (one). Two food-borne outbreaks reported by Bulgaria involved, in total, 27 people from which only one person needed hospitalisation and these food-borne outbreaks were caused by unspecified Trichinella species. The two outbreaks reported by Croatia and Romania were caused by T. spiralis, involving three and five human cases, respectively, which needed hospitalisation. The food-borne outbreak reported by Italy was caused by T. britovi and three out of nine people were hospitalised; the vehicle was wild boar meat products. Two food-borne outbreaks reported by one non-MS (Serbia) involved 27 people from which eight people were hospitalised and were caused by an unspecified Trichinella species. Trichinellosis food-borne disease outbreaks were, during 2019, mostly caused by pig meat and products thereof ( Figure 46 and Table 43 ), as during previous years (2010) (2011) (2012) (2013) (2014) (2015) (2016) (2017) (2018) . Further details and statistics on the trichinellosis food-borne outbreaks for 2019 are in the food-borne outbreaks chapter. Countries that reported food-borne human trichinellosis cases are coloured according the food vehicle causing the outbreaks ('pig meat and products thereof', 'other or mixed red meat and products thereof' or 'unknown' food vehicle) (data reported to EFSA). The numbers without green box indicate the number of domestic trichinellosis human cases and the numbers in a green box indicate the number of travel-related trichinellosis human cases (data reported to ECDC except for Serbia (*) data reported to EFSA). Pig meat and products thereof Bulgaria (2) Croatia (1) Romania (1) 4 80.0 Romania (37) Lithuania (12) Croatia (5) Latvia (4) France (3) Belgium (1) Poland (1) Spain (1) 73 73 Other or mixed red meat and products thereof Italy (1) 1 20.0 Lithuania (6) Poland (6) Romania (3) Germany (1) Latvia (1) 18 18 Meat and meat products -(*) --Poland (5) Spain (2) Croatia (1) Germany ( Table 44 shows Trichinella summary monitoring results in domestic pigs and in farmed wild boar by housing conditions, for 2019. All pigs in mixed herds reported were not raised under controlled housing conditions. In 2019, 31 countries (all 28 MS and 3 non-MS) provided information on Trichinella in domestic animals (pigs and/or farmed wild boar). Six MS (Bulgaria, Croatia, France, Romania and Spain), as in 2018, reported positive findings in domestic pigs not raised under controlled housing conditions. No positive findings were found in farmed wild boars. Sixteen MS (Belgium, Bulgaria, Croatia, Denmark, Estonia, Finland, France, Ireland, Italy, Latvia, the Netherlands, Portugal, Romania, Spain, Sweden and the United Kingdom) and one non-MS (Iceland) reported data on breeding and fattening pigs raised under controlled housing conditions, no positive finding was reported. In 2019, 25 MS and two non-MS reported data on breeding pigs, fattening pigs, pigs from mixed herds or on farmed wild boar that were not raised under controlled housing conditions and six MS reported positive findings among breeding pigs, fattening pigs and pigs from mixed herds. In total, one breeding pig (< 0.01%), 195 (< 0.01%) fattening pigs and 23 (< 0.01%) pigs from mixed herds were positive. Spain accounted for most positive pigs followed by Romania, Poland, Croatia, Bulgaria and France. As during 2014-2018, these Trichinella infections were from free-range and backyard pigs reared in rural EU regions. All farmed wild boar (7,570) tested negative. Norway and Switzerland tested 3,907,231 fattening pigs not raised under controlled housing conditions and all tested negative. Two MS (Bulgaria and Croatia) reported data on food. Croatia reported eight positive units of meat from pig-meat products out of 13 tested. Bulgaria reported one positive fresh raw sausage made with wild boar meat. As shown in Figure In 2019, as in the previous 4-year period (2015-2018) , no positive finding was reported in domestic solipeds (156,815 animals and 2,236 slaughter animal batches tested) and reported by 22 MS (Austria, Belgium, Bulgaria, Czechia, Denmark, Estonia, Finland, France, Germany, Hungary, Ireland, Italy, Latvia, Luxembourg, Malta, the Netherlands, Portugal, Romania, Slovenia, Spain, Sweden and the United Kingdom) and in two non-MS (Iceland and Switzerland). Bulgaria reported two negative test results from fresh raw sausage made with horse meat. Summary data for wild animals are given in Table 45 . Seventeen MS (Austria, Bulgaria, Croatia, Czechia, Estonia, Finland, France, Germany, Hungary, Italy, Latvia, Poland, Romania, Slovakia, Slovenia, Spain and Sweden) and one non-MS (Republic of North Macedonia) reported positive findings in hunted wild boar (1,378 positive findings out of 1,767,487 animals tested (< 0.007%). In total, 10 MS and one non-MS reported data on Trichinella in red foxes (Vulpes vulpes) with, in total, 89 (1.3%) positive out of 6,697 tested animals. Eight MS reported data on Trichinella in brown bears (Ursus arctos) with 24 (3.08%) positive out of 779 tested in four MS. Six MS and one non-MS reported data on Trichinella in other wild animals. Positive findings were detected in eight species (lynx, otter, wolverine, wolf, raccoon dog, eagle, polecat and jackal) from four MS and one non-MS. The highest number of infected animals was observed in racoon dogs (41.2%) followed by wolverine (37.5%), lynxes (17.1%), wolves (14.6%), jackals (8.9%), polecats (11.1%), eagles (3.3%) and otter (1.6%). These distribution maps have been built based on data from reports (EFSA and ECDC, 2015a , 2015b , 2017a , 2019c . The EU/EEA trend for trichinellosis has been greatly affected by the number and size of food-borne outbreaks. The number of human cases and the EU notification rate have, however, been kept low in the last 5 years from 2015 to 2019 with the highest rate (0.03) reported in 2017 and 2015. In 2018, the lowest rate (0.01) was reported since the beginning of trichinellosis EU-level surveillance in 2007. Despite the increase in cases and notification rate (0.02) in 2019 compared with 2018, the 5-year trend from 2015 to 2019 was declining. The number of confirmed trichinellosis cases in 2019 was lower than the 5-year average in the EU. The increase in 2019 was mainly due to the increase of domestic trichinellosis cases in three MS (Bulgaria, Italy and Spain). The main reason for this increase was the higher consumption of various home-made pork products during winter as well as during the wild boar hunting season. Romania, which had experienced most Trichinella outbreaks in the previous years, reported fewer human cases in 2019 than in 2018. About one-third of the confirmed cases were hospitalised, with one fatal outcome. This represents fewer hospitalisations compared with previous years. In general, Trichinella infections in humans are linked to food-borne outbreaks. In 2019, fewer human food-borne cases (N = 44) were reported to EFSA (food-borne outbreaks database) than confirmed sporadic cases (N = 96) reported to TESSy managed by ECDC. However, in 2017 and 2018, the former number was higher than the latter. Such discrepancies result from different case classification for reporting between the two databases. Spain was unable to report 2019 data to TESSy from all autonomous regions, due to COVID-19. Given the pandemic emergency and related difficulties, exhaustive and accurate reportingto ECDC and to EFSAcould have been challenging for other MS. In 2019, five Trichinella outbreaks were reported by four MS (Bulgaria, Croatia, Italy and Romania, reporting rate < 0.01 outbreak per 100.000 population) and two outbreaks by one non-MS (Serbia). In total, 44 patients were affected in the EU and 27 in the non-EU MS of which almost half (20) were hospitalised. All outbreaks were reported with strong-evidence and associated with 'pig meat and products thereof', except one which was associated with 'other or mixed red meat and products thereof'. It is important to underline the reports of Trichinella-positive domestic pigs by Bulgaria, Croatia and Romania that also reported food-borne human cases (to EFSA food-borne outbreaks database), and the reports by Poland and Spain that reported confirmed domestic human cases (to ECDC TESSy). By contrast, in other MS during the last years, there was an increasing number of pigs raised under controlled housing conditions and increased control at slaughtering of pigs that are not raised under controlled housing conditions. These measures, in combination with activities raising awareness about trichinellosis and farmers' education, may have contributed to a reduction of the parasite biomass in the domestic habitat and the probability of acquiring an infection for humans ( Figure 47 ). In the EU, most pigs are subject to official meat inspection at slaughter in accordance with Regulation (EU) 2015/1375; only pigs slaughtered for own consumption are not covered by the Regulation. Around 218 million pigs were tested for Trichinella in MS and non-MS in 2019, out of about 246 million reared pigs in the EU (Marquer et al., 2014) , with only 219 positive animals, about 0.89 per million reared pigs. Only six out of 28 MS reported Trichinella in pigs in 2019, with an overall prevalence of 0.00001%. All positive findings were from pigs not raised under controlled housing conditions. In the EU, infected pigs are usually clustered in five MS (Bulgaria, Croatia, Poland, Romania and Spain) and sporadic infections are documented in other MS (Pozio, 2014) . In 2019, Spain accounted for the highest number of positive domestic pigs (113) followed by Romania (79) Poland (22), Croatia (3), Bulgaria (1) and France (1) . The reported number of Trichinella-positive domestic pigs is likely to be an underestimation of the true number, as most pigs at risk for this infection are slaughtered at home without any veterinary control and recording. EFSA has identified that noncontrolled housing condition is a main risk factor for Trichinella infections in domestic pigs and the risk of Trichinella infection in pigs from well managed officially recognised controlled housing conditions is considered negligible (EFSA BIOHAZ, CONTAM and AHAW Panels, 2011; EFSA and ECDC, 2011) . In addition to domestic pigs, hunted wild boar are an important source of trichinellosis infections for humans. However, the prevalence of Trichinella spp. infections in this animal species has declined over the years due to the increased control for these pathogens. From 2012 to 2016, the prevalence of infection was reduced threefold (from 0.13% in 2012 to 0.05% in 2016) but increased up to 0.09% in 2018 in the hunted wild boar population. In 2019, a new decrease of the prevalence to 0.08% was recorded in this animal species. Trichinella spp. were not detected in farmed wild boar; however, the number of tested farmed wild boar decreased during the last years. No positive finding was reported for solipeds in 2019. In the last 12 years, only four horses tested positive out of more than one million tested animals in 2008 (EFSA and ECDC, 2009 . This extremely low (< 0.001%) prevalence could be related to the effective control which, according to EFSA BIOHAZ Panel (2013b) , should be maintained as long as there is no full and reliable traceability system in place. Trichinella spp. circulate among wild animals in large parts of Europe. In 2019, seven MS and one non-MS reported positive findings in wild animals (brown bears and wild animals different from foxes and wild boar). The reporting of negative findings in MS could be explained by insufficient number of surveys, inadequate sample size or, investigations in regions in which environmental conditions that do not favour the transmission of these zoonotic nematodes among wildlife. Red foxes, having a large and widespread population, can be considered as the main natural reservoir of Trichinella in Europe. The prevalence decreased by twofold in the last 5 years (from 2% in 2013 to 1.1% in 2017) and then increased in 2018 (1.6%) and decreased again in 2019. In 2019, 10 MS and one non-MS monitored Trichinella spp. infection in 6,697 red foxes and positive animals were detected in five MS. The proportion of positive samples from wildlife was higher in raccoon dogs, wolverine, lynxes, wolves and jackals, but their population size and distribution in Europe are generally limited to a few countries. Data from Trichinella in wild animals are not fully comparable between MS as neither harmonised monitoring schemes nor mandatory reporting requirements are in place and the reported findings must therefore be interpreted with caution. These data allow descriptive summaries at the EU level but preclude subsequent data analysis such as assessing temporal and spatial trends. Identification of Trichinella larvae at the species level carried out in 2019 confirms that T. spiralis is more prevalent than T. britovi in pigs (Pozio et al., 2009 ). However, since T. spiralis is patchily distributed, T. britovi and Trichinella pseudospiralis were detected in pigs in some countries. Trichinella nativa has been documented in wild carnivores of Finland, Estonia and Sweden. T. pseudospiralis was documented in hunted wild boar, six lynxes and one eagle confirming its low prevalence in target animals (Pozio, 2016) . There is a relationship between unawareness and low-income of consumers, living in rural areas, inadequacy of local veterinary meat inspection services and the occurrence of Trichinella in domestic animals in the EU and non-EU countries (Pozio, 2014) . The increasing number of wild boar and red foxes and the spread of the raccoon dog population from eastern to western Europe and that of the jackal from southern-eastern to northern-western Europe may increase the prevalence of Trichinella circulating among wild animals (Alban et al., 2011; Sz ell et al., 2013) . notification of echinococcosis in humans is mandatory in most MS, Iceland and Norway, except for Belgium, France, the Netherlands and the United Kingdom, where reporting is based on a voluntary surveillance system. Denmark and Italy have no surveillance system for echinococcosis. In Switzerland, echinococcosis in humans is not notifiable. The surveillance systems for echinococcosis cover the whole population in those MS where surveillance systems are in place. For 2019, Spain did not receive data from all regions due to COVID-19 and the notification rate is therefore not displayed for this year. All countries reported case-based data except Belgium, Bulgaria and the Netherlands, which reported aggregated data. Both reporting formats were included to calculate numbers of cases and notification rates. An attempt to collect harmonised clinical data in the EU on a voluntary basis is currently undertaken by the European Register of Cystic Echinococcosis (ERCE) (Rossi et al., 2016 (Rossi et al., , 2020 ; http:// www.heracles-fp7.eu/erce.html) and in the past with the European (Alveolar) Echinococcosis Registry (EurEchinoReg) (Kern et al., 2003) . Estimates of the real burden of these diseases are extremely difficult to calculate because of the long incubation period (months or years) and the non-specific symptoms. A recent cross-sectional ultrasound-based survey, conducted in Romania and Bulgaria, estimated around 45,000 human CE infections in rural areas of these two endemic European countries (Tamarozzi et al., 2018) . Echinococcus multilocularis in Europe is mainly transmitted to humans by a sylvatic cycle that is wildlife based (Casulli et al., 2019a) . Intermediate hosts (IHs) for E. multilocularis are small rodents (microtine or arvicolid), while definitive hosts (DHs) are mainly red foxes and, to a lesser extent, other canids such as raccoon dogs, dogs, jackals and wolves. Echinococcus granulosus s.l. is a complex of species causing CE, in animals and humans. E. granulosus s.l. in Europe is mainly transmitted to humans by a pastoral cycle (Casulli et al., 2019b) . IHs for E. granulosus s.l. are mainly livestock species (mainly sheep, secondarily pigs but also cattle and goats), while DHs are shepherd dogs (rarely wild canids). People become infected with AE and CE through the ingestion of eggs of the tapeworm prevalent in these DHs. Surveillance for E. multilocularis in Europe is usually carried out on a voluntary basis, with the exception of the five reporting countries claiming to be free from this parasite according to the Commission Delegated Regulation (EU) 2018/772 supplementing Regulation (EU) No 576/2013 20 . Surveillance is carried out in the main European DHs, the red fox (Vulpes vulpes). Four MS (Finland, Ireland, Malta and the United Kingdom) have demonstrated the absence of E. multilocularis through the implementation of an annual surveillance programme required in accordance with Regulation (EU) 2018/772. One EEA State, mainland Norway (Svalbard archipelago excluded), also implements a surveillance programme in line with Regulation (EU) 2018/772. In all other MS, data on E. multilocularis rely on whether findings are notifiable and if monitoring is in place or if studies on E. multilocularis are performed. As data on E. multilocularis in animals vary geographically (also within countries) and over time, reported cases of E. multilocularis are difficult to compare within and between countries. According to a recent meta-analysis, based on studies published between 1900 and 2015, E. multilocularis has been documented in red foxes from 21 countries (Oksanen et al., 2016; Figure 48 ). Since 2015 and 2020, this parasite has been also found in foxes and golden jackals from Croatia and Hungary, respectively (Du sek et al., 2020; Balog et al., 2021) . Surveillance of E. granulosus s.l. is carried out in livestock IHs during slaughterhouse inspections. In particular, necropsy on sheep liver and lungs is used to detect the presence of parasitic cysts, while molecular PCR-based methods are used to confirm and to identify genotype/species belonging to the Echinococcus genus (Siles-Lucas et al., 2017) . Although Regulation (EU) 2018/772 is in force for E. multilocularis, no specific EU Regulation is in place for detecting E. granulosus s.l. in animals or humans, therefore surveillance for the latter parasite depends on national regulations. In 2019, 751 laboratory-confirmed echinococcosis cases were reported in the EU by 26 MS (Table 47) . Twenty-three MS reported at least one confirmed case and three MS reported zero cases. The EU notification rate was 0.18 cases per 100,000 population, which was the lowest notification rate in the last five years. The highest notification rates were observed in Lithuania with 2.90 cases per 100,000 population, followed by Bulgaria with 2.76 and Austria and Latvia with 0.41 and 0.31 cases per 100,000 population, respectively. Most echinococcosis cases (65.1%) were reported without travel-associated data, 23.0% were domestic or related to travel within the EU and 11.9% were associated travel outside the EU (Table 46) . Seven MS (Czechia, Estonia, Hungary, Latvia, Lithuania, Romania and Slovakia) of the 15 MS reporting information on imported cases in 2019 notified all Echinococcus spp. infections as being domestically acquired. The highest proportion of travel-related cases were reported by Finland (100%; eight cases), Luxembourg (100%; one case), Sweden (95.5%; 21 cases) and Norway (100%; seven cases). At a species level, E. multilocularis human infections were more often reported domestically acquired than E. granulosus s.l. human infections (85.1% vs. 34.9%). Among 112 travelassociated cases of Echinococcus spp. with known origin of infection, majority (79.5%) were reported as originating from outside the EU. Syria, Iraq and Turkey were the most frequently reported probable country of infection, representing half (50.0%) of the imported cases in 2019. In EU, Bulgaria and Romania were reported as probable country of infection for 12 (10.5%) and six cases (5.3%), respectively. Fourteen MS provided information on hospitalisation, covering 32.8% of all confirmed cases of echinococcosis in the EU in 2019. The overall hospitalisation rate was 44.3%, which represents a continuous decrease during the last 10 years from 100% in 2008, when only hospitalised cases were reported. In 2019, the highest proportions of hospitalised cases (60-100%) were reported in Czechia, Greece, Poland, Portugal, Romania, Slovakia and Slovenia. More than half (56.4%) of human AE cases were hospitalised compared with about one-third (36.4%) of human CE cases based on reporting by four and nine MS, respectively. Information on the outcome of the cases was provided by 14 MS. One fatal case due to the infection by E. granulosus s.l. and one fatal case due to infection by E. multilocularis was reported in Portugal and Poland, respectively. This resulted in an EU case fatality of 0.86% among the 232 cases for which this information was reported (30.9% of all confirmed cases) in 2019. Table 48 summarises the most relevant DH and IH species tested for E. multilocularis, such as foxes, raccoon dogs, dogs, jackals, wolves, cats, beaver, voles, wild boar, coypu, squirrel, mice and pigs and results reported by MS and adjacent countries in 2019. In accordance with the Regulation (EU) 2018/772, surveillance of E. multilocularis is mainly focused on red foxes as DH. In total, 13 MS and two non-MS (Norway and Switzerland) reported 2019 monitoring data on 6,326 and 621 foxes examined for E. multilocularis, respectively. Seven MS and one non-MS (Switzerland) reported a total of 12.9% positive samples: Czechia (21%), France (12.6%), Germany (15.2%), Hungary (4.8%), Luxembourg (19.1%), Poland (31.7%), Slovakia (16.1%) and Switzerland (39.7%). Czechia (N = 596) reported most infected foxes in Europe accounting for 68.6% of the positive findings. In addition to foxes, E. multilocularis has been reported in 18 dogs (two from France, three from Slovakia and 13 from Switzerland), two wolves from Switzerland, one cat from France, two jackals from Hungary, one coypu from France and two beavers and one mouse from Switzerland. Source: Austria, Czechia, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Latvia, Lithuania, Malta, Norway, Malta, Poland, Portugal, Romania, Slovakia, Slovenia, Spain and Sweden. Belgium, Bulgaria, Croatia, Cyprus, Denmark, Iceland, Italy Luxembourg, the Netherlands and the United Kingdom did not report data to the level of detail required for the analysis. In total, 19 MS and two non-MS reported data from 113,761,312 domestic and wild animals tested for E. granulosus s.l. of which > 99% were domestic animals (sheep, cattle, goats, pigs, horses, water buffalos, dogs and cats) (Table 49) . A large proportion of these data were obtained from domestic livestock during meat inspection at the slaughterhouse. Wild animals tested included deer, moose, mouflons, wild boar, other wild ruminants and wolves. Eleven MS reported in total 167,003 (0.15%) positive samples mainly from domestic animals. These positive samples reported by Bulgaria, Greece, Italy, Poland, Slovakia, Spain and the United Kingdom were mainly from small ruminants (sheep and goats; N = 131,850; 78.9%) ranging from 0.02% to 4.8% positives. There were 16,298 positive cattle (9.8% of animals positive for E. granulosus s.l.) reported by Bulgaria, Greece, Hungary, Italy, Romania, Slovakia, Spain and UK and 18,696 positive pigs (11.2% of animals positive for E. granulosus s.l.), of which 85.4% were reported by Poland. Belgium, Cyprus, Denmark, Estonia, Ireland, Malta, Slovenia and Sweden did not report any positive finding of E. multilocularis or E. granulosus s.l. Austria, Croatia, Lithuania, Netherlands and Portugal did not report any animal monitoring data for E. multilocularis or E. granulosus s.l. It should be emphasised that positive samples from dogs, cats, wolves and pigs without species specification were only mentioned in Table 48 and/or Table 49 for countries with known circulation of both E. granulosus s.l. and E. multilocularis. In fact, countries that are endemic for AE (i.e. Italy, Poland, Slovakia and Switzerland) reported 16,163 Echinococcus spp. positive pigs but the species identification was only reported by Switzerland, identifying E. multilocularis in seven pigs. The three mentioned MS endemic for AE (northern Italy, Poland and Slovakia) are also co-endemic for CE. Pigs are good hosts for E. granulosus s.l., while E. multilocularis metacestodes in pigs are abortive and their presence is often used as sentinel for the presence of this parasite as demonstrated in Switzerland (Meyer et al., 2020) . Hungary and Latvia reported 35 and 1 positive pigs, respectively, identifying the E. granulosus s.l. species. As shown in Figure 52 , Bulgaria, Greece, Italy, Poland, Romania, Spain and UK were the most endemic countries for Echinococcus granulosus s.l. in Europe during the period 2015-2019. Intermediate hosts included in map are cattle, deer, goats, horses, moose, mouflons, pigs, reindeer, sheep, water buffalos and wild boars. Legend: dark blue ≥ 500 positive cases; light blue < 500 cases; yellow = 0 cases reported; white = data not reported. Because of the co-endemicity with Echinococcus multilocularis, pigs were excluded from Latvia, Hungary, Poland, Germany, Slovakia and Switzerland when Echinococcus species information was not reported. The two parasitic diseases in humans, CE and AE, caused by E. granulosus s.l. and E. multilocularis, respectively, can be reported separately to ECDC TESSy database even though the EU case definition 'echinococcosis' does not differentiate between these two diseases. Most MS reported species information from 2008 to 2019. In addition, in 2018 and 2019, a few countries reported clinical presentation, which differentiates the two forms of the disease. Since the beginning of the surveillance of human echinococcosis in the EU in 2007, CE has been more frequently reported than AE, as expected by data reported in the scientific literature for Europe. The EU notification rate of confirmed human echinococcosis cases was stable and the trends for infections caused by E. granulosus s.l. and E. multilocularis did not show any significant increase or decrease in the last five years since 2015. In a few countries, the increase in the number of cases in the last few years could be explained by intensified surveillance and improved notification system for echinococcosis. The raised awareness of the disease among clinicians and immigration (people from endemic countries) may also have influenced the number of diagnosed cases in some countries (Richter et al., 2019) . Distinction between infection with E. granulosus s.l. and E. multilocularis is needed because the two diseases require different clinical management and strategies for control. It should be also emphasised that the true prevalence of these diseases is extremely difficult to estimate due to the long incubation period (AE and CE), the high proportion of asymptomatic or paucisymptomatic carriers who never seek medical attention (CE) and the underreporting/misdiagnosed cases (AE and CE), factors, which contribute to their neglected status (Casulli, 2020) . For these reasons, the patchy data reported by MS on the number of people with echinococcosis, currently represent the 'tip of the iceberg' of infections. The invisible portion includes asymptomatic carriers of CE and misdiagnosed cases of AE (Kern et al., 2017) . In animals, in 2019, 19 MS reported monitoring data on E. granulosus s.l., aetiological agent of CE and E. multilocularis, aetiological agent of AE. The highest number of animals infected with E. granulosus s.l. was reported in Bulgaria, Spain, Greece, Poland, Italy and UK and mainly observed in sheep. The highest number of animals (mainly foxes) infected with E. multilocularis was reported in Czechia, France, Germany, Hungary, Luxembourg, Poland, Slovakia and Switzerland. The surveillance of E. multilocularis in foxes is important to assess the prevalence in Europe, as the geographical distribution of E. multilocularis seems to have widened in the last decades. Whether the increased geographical distribution of E. multilocularis is due to an increased fox population in Europe (Oksanen et al., 2016) , or to the expansion of their habitat to urban areas (Deplazes et al., 2004) or whether it reflects an increased surveillance effort, is difficult to disentangle, as there is a general lack of baseline data and standardised detection methods. Also, in animals, notification is a requirement for reliable data and information on parasite speciation is very important for risk management efforts as E. granulosus s.l. and E. multilocularis have a different epidemiology and pose different health risks for humans (Casulli, 2020) . For E. granulosus s.l., a notification requirement would ensure that comparable data between MS are obtained from meat inspection of food-producing animals. For E. multilocularis, a general notification requirement for all MS can be questioned, but it is required in countries free from this parasite, according to EU Regulation (EU) 2018/772. In general, animal and human findings from 2019 seem similar to those of recent years. It should be emphasised that findings from most endemic countries fluctuated between years, but they reported positive findings in animals and humans in most years. Fluctuations in reported numbers of infected animals are probably associated with investigational efforts performed in a particular year, rather than reflecting a change in true prevalence. Moreover, it is unclear how the COVID-19 pandemic is impacting on the diagnosis and notification of these chronic parasitic diseases at European level. Related projects and Internet sources • The health impact of food-borne outbreaks in the EU was important in 2019 since 60 outbreakrelated deaths were reported; 20 more fatal cases than in 2018 (+50%). • A high number of deaths (N = 10) were registered in community settings such as 'residential institution (nursing home or prison or boarding school) ' . In addition, nearly 19% of cases involved in strong-evidence outbreaks (2,407 cases) were exposed to contaminated foods in schools or kindergartens. These findings highlighted the need for attention to the high risk of vulnerable populations to food-borne hazards. • The health burden of outbreaks caused by Listeria monocytogenes in the EU was remarkable since this agent was responsible for 349 cases of illness and more than 50% of total outbreak associated deaths (31 deaths; 10 deaths more than in 2018; 29 more than in 2017). Most of the deaths were due to the consumption of meat and meat products. The number of outbreaks, cases and hospitalisations associated with L. monocytogenes infection in the EU has continuously increased over the last four years. • At the EU level, the consumption of food of animal origin ('fish and fishery products', 'eggs and egg products', 'meat and meat products', 'milk and milk products') was associated with most of the food-borne strong-evidence outbreaks. • Outbreaks associated with the consumption of 'crustaceans, shellfish, molluscs and products thereof' increased markedly in the EU (by 80 outbreaks; 101.3% more than in 2018) even if this rise was entirely attributable to France which reported 129 outbreaks (81.1% of total outbreaks in the EU). Norovirus in 'fish and fishery products' was the agent/food pair causing the highest number of strong-evidence outbreaks in the EU. • Salmonella in 'mixed food', norovirus in 'fish and fishery products' and Salmonella in 'eggs and egg products' were the agent/food pairs that caused the highest number of cases. Pairs with Salmonella in different types of food ('eggs and egg products', 'mixed foods', 'meat and meat products', 'bakery products', 'buffet meals') caused the highest numbers of hospitalisations. • Mixed foods (i.e. food resulting from mixing together multiple ingredients in the same preparation) were the foodstuff most frequently consumed by outbreaks cases. These mixed foods were associated with a wide range of causative agents including bacteria, viruses, bacterial toxins and histamine. • The number of outbreaks implicating food of non-animal origin (FNAO) reported in 2019 was similar to those reported in recent years. Outbreaks by FNAO (mainly vegetables) were larger, on average, compared with outbreaks caused by food of animal origin and were associated with the widest variety of causative agents, mainly norovirus, Salmonella, Bacillus cereus and Cryptosporidium. According to Directive 2003/99/EC, reporting information on food-borne and waterborne outbreaks (FBOs) is mandatory for EU Member States (MS). EFSA is assigned the tasks of collecting, analysing and describing the data. The aim is to support characterising the epidemiology and the health impact of FBOs in the EU in the current year and the relative time trends. The main focus of the analysis is to provide a thorough description of the causative agents and the foodstuffs implicated in the FBOs, as well as to document the circumstances, the events and the potential risk factors that underlie the contamination of foodstuffs and the occurrence of the outbreaks. These data are collected annually by MS and reported to EFSA according to the standard defined by the EFSA Network for Zoonoses Monitoring Data and described in an updated technical document issued each year by EFSA (EFSA, 2020a,b). The current system is known as European Union Food-borne Reporting System (EU-FORS) and has been implemented since 2010. The data collection includes any outbreaks deemed to implicate the consumption of food (including water) contaminated by either bacteria, viruses, parasites, algae, fungi and their products, such as toxins and biological amines (e.g. histamine). The reporting is not limited to the causative agents whose transmission to humans occurs primarily through food (e.g. Salmonella, L. monocytogenes), but also includes agents for which the food-borne transmission is possible but usually accidental. Outbreaks caused by ingestion of drinking water are also deemed food-borne as drinking water is defined as a food in Regulation 178/2002/EC. Outbreaks are categorised as having 'strong-evidence' or 'weak evidence' based on the strength of evidence implicating a suspected food vehicle as the cause of the outbreak (EFSA, 2014). The strength of evidence is a qualitative measure of the level of uncertainty which affects the likelihood that a food item is the vehicle of the outbreak. For strong-evidence outbreaks, MS shall report a detailed data set describing the implicated food vehicle, contributory factors and source, whereas for weak-evidence outbreaks this reporting is not compulsory. The evaluation of the strength of evidence implicating a suspected food vehicle in FBOs is based on the assessment of all available types of evidence related to illness and exposure information (i.e. microbiological, epidemiological, descriptive, environmental, based on traceability (tracing back/forward) of the investigated foodstuffs) and according to the EU-FORS guidance and the last published manual for reporting on food-borne and waterborne outbreaks (EFSA, , 2020a . Although the data reporting rules follow the same EFSA standard specifications as described above in all MS, the surveillance activities of FBOs are not fully harmonised. Differences in sensitivity and type of outbreaks under surveillance exist. For this reason, the difference in the numbers and types of reported FBOs, as well as in the causative agents and the type of outbreaks may not necessarily reflect the level of food safety in the MS. A description of the system in place for outbreak surveillance and reporting in the reporting countries is in the national zoonoses reports submitted in accordance with Directive 2003/99/EC, which are published on the EFSA website together with the EU One Health Zoonoses Report and are available online at http://www.efsa.europa.eu/en/biological-hazards-data/ reports. Key summary statistics for all reported FBOs are summarised in figures and tables. The impact of FBOs on public health is described in terms of total number of outbreaks and reporting rate (per 100,000 population), number of cases (of illnesses), number of hospitalisations (% of hospitalisation), number of deaths (% deaths), mean outbreak size (cases per outbreak) and range of cases per outbreak (minimum and maximum). To limit the level of uncertainty, the description of food vehicles implicated in FBOs, the settings (places of exposure to contaminated food) and the risk factors refers to strong-evidence food-borne outbreaks only. However, the pattern of suspected food vehicles and settings is also summarised separately for weak-evidence FBOs, based on the detailed data set that MS can report also for this type of FBO. Causative agents, food vehicles and outbreak settings are summarised using multi-level hierarchical categorisation to optimise the description of the findings. A priority is given to the description of FBOs caused by agents included in Annex IA of the Dir. 99/2003/CE (Brucella, Salmonella, Campylobacter, Listeria monocytogenes, Shiga toxin-producing E. coli and Trichinella), as they are considered toppriority pathogens at the EU level. Causative agents listed under Annex IB of the same Directive, with major epidemiological relevance (Calicivirus, hepatitis A virus, botulism and agents thereof, Yersinia, Cryptosporidium) were also described distinctly. The other causative agents are described either separately, where possible, or in homogeneous groups by type of agent. In this latter case, the agents included in each group are listed in tables and figures footnotes. Unknown agents are described separately. Causative agents implicated in FBOs are grouped according either to taxonomy or the pathogenic mechanisms triggering illness in humans. In some circumstances (i.e. missing information) or to adjust the agents categorisation, further criteria have been applied as follows: any E. coli other than 'Shiga toxin-producing E. coli (STEC)', have been categorised into a single 'E. coli other than STEC' group; 'Bacillus cereus enterotoxins' and 'B. cereus' were grouped into 'B. cereus' group; 'Staphylococcus aureus '', 'Staphylococcus unspecified' and 'Staphylococcal enterotoxins' have been grouped into the 'S. aureus' group together; 'Clostridium unspecified' and 'C. perfringens' were grouped into the 'C. perfringens' group; histamine and scombrotoxin have been grouped; 'Calicivirus, unspecified', 'norovirus' and 'sapovirus' have been grouped into the 'norovirus and other Caliciviruses' group; 'hepatitis, unspecified' and 'hepatitis A' have been grouped into 'hepatitis A and other hepatitis virus, unspecified' group. Food vehicles have been grouped according to the general criteria adopted by EFSA for presenting data in this report. Places of exposures have been grouped according to the general characteristics and level of risk connected to the setting and the process behind food preparation. In tables and figures, sums and proportions (%) are the basic statistics used to describe the reported counts (numbers) of outbreaks. The mean annual rate of reported outbreaks per 100,000 population ('outbreak reporting rate') is calculated to compare MS independently on demographic size and its variations over time. Data on resident population from Eurostat were used for this purpose (updated on 1 January 2020). Populations of MS not providing data on FBOs were excluded from this calculation. Variations over time are described by comparison with different time frames. Data on food-borne and waterborne outbreaks for 2018 differ from those published in the European Union One Health 2018 Zoonoses Report, due to a delay in reporting from one MS (the Netherlands). Short-term variations are shown as absolute and relative (%) 2019/2018 difference. Long-term variations are also described using years 2010-2019 as the comparative period. Frequency distributions and trends are visualised at the EU level. However, trend analysis is only performed at the single MS level, according to the rationale described in Boelaert et al. (2016) for data quality. Time trends have been tested for statistical significance over the period 2010-2019 using the Cox-Stuart sign test, a non-parametric test appropriate for limited numbers of observations (10 years at the maximum). P value < 0.05 was considered to identify a statistically significant trend, beyond chance. However, the detection of significant trends at the country level should be interpreted with caution since changes in the reporting specifications for FBOs were introduced in 2014 (EFSA, 2014). Sankey diagrams, which are available as supporting documents from the EFSA knowledge junction at zenodo (see link in the beginning of this chapter), were produced using the free software R version 3.5.3 (GNU project r-project.org). When the UK data were collected, the UK was an EU MS but as of 31 January 2020, it has become a third country. Results and discussion 4.1. Overview of countries reporting food-borne outbreak data, 2019 During 2019, 27 MS reported 5,175 FBOs, 49,463 cases of illness, 3,859 hospitalisations and 60 deaths. In addition, 117 FBOs, 3,760 cases of illness and 158 hospitalisations were communicated by six non-MS (Iceland, Montenegro, Norway, Republic of North Macedonia, Serbia, Switzerland). Slovakia did not report data on FBOs. The total number of outbreaks reported by each MS in 2019 varied importantly, with a small number of MS reporting most of the outbreaks. Altogether, FBOs reported by five countries (Belgium, France, the Netherlands, Poland and Spain) accounted for more than three-quarters of total outbreaks (4,042 outbreaks; 78.1% of all outbreaks) and more than two-thirds of total cases observed in the EU in 2019 (32,883 cases; 66.5% of all cases). The breakdown of FBOs by countries and by strength of evidence is reported in Table 50 . In this table, the 'outbreak reporting rate' (per 100,000 population) describes how frequent was the reporting of FBOs in 2019, in EU/EFTA countries, regardless of the differently sized populations. The range of this value was huge, from 0.04 (Romania) to 9.12 (Malta) outbreaks (per 100,000 population) corresponding to a 253-fold difference. The 'mean outbreak size' (i.e. the mean number of cases per outbreak) and the range of cases per outbreak is shown to characterise the pattern of FBOs reported to EFSA by MS and non-MS. Altogether, these indicators provide evidence of the large variability among MS in the sensitivity of surveillance and the type of FBOs being monitored in each MS. As an example, household outbreaks (i.e. outbreaks in which all the human cases live in one single household) are usually small-sized outbreaks. As not all MS report household outbreaks to EFSA, this may influence the mean outbreak size as well as the number of outbreaks. Details on the type of FBOs reported to EFSA, by country, is visualised in Figure 53 . The overall distribution of FBOs and outbreak cases reported by MS during 2010-2019 are plotted in Figures 54 and 55 , respectively. For 2018, the numbers included in Figure 54 differs from those published in the European Union One Health 2018 Zoonoses Report, due to a delay in FBOs data reporting from one MS (the Netherlands). In 2019, the number of outbreaks reported in the EU was lower than in 2018 (727 outbreaks less; 12.3% less than in 2018). Cases of illness and hospitalisations also dropped, even if with different proportions. Cases decreased by 3.3% (1,708 cases less than in 2018) and hospitalisations by 20.0% (962 hospitalisations less than in 2018). The lack of 2019 FBOs data reporting from Slovakia may have substantially contributed to the reduction since this country had reported 522 outbreaks, 2,454 cases and 531 hospitalisations per year, on average in the five preceding years. The health impact of FBOs in 2019 was remarkable since 60 outbreak-related deaths were reported, 20 more fatal cases than in 2018 (50% more than in 2018). France and the United Kingdom reported each 15 deaths among outbreak cases which represents an important increase compared with the previous five years (3.8 and 4.2 mean deaths per year, in France and the United Kingdom, respectively). In France, 10 deaths were reported in outbreaks that occurred in a 'residential institution (nursing home or prison or boarding school)'. These data call for attention to the increased risk of vulnerable populations to food-borne hazards. In the United Kingdom, deaths were less clearly linked to specific settings. Most deaths were single cases involved in general dispersed outbreaks. However, seven deaths were reported from a single outbreak in hospital setting, which raises again the issue of the increased susceptibility to food-borne hazards of vulnerable patients. Spain also reported a high number of deaths (N = 9) among outbreaks cases. Three of them were linked to large nation-wide outbreaks by L. monocytogenes. In 2019, strong-evidence outbreaks (N = 716) were reported by 23 MS (all MS reporting data on FBO except Cyprus, Ireland and Malta) and accounted altogether for 13.8% of all outbreaks, which represents the highest proportion since 2010 (Table 50) . At country level this proportion varied widely. For eight MS (Finland, Greece, Hungary, Italy, Romania, Slovenia, Sweden, the United Kingdom) strong-evidence outbreaks accounted for more than a third of total reported FBOs. Interestingly, these MS also reported the smallest proportion of household outbreaks ( Figure 53 ). In six MS (Austria, Belgium, Germany, Lithuania, the Netherlands, Portugal), strong-evidence outbreaks did not exceed 10% of total FBOs. In addition, 57 strong-evidence outbreaks were reported by the non-MS countries which communicated to EFSA data on FBOs for 2019. The annual variations (%) in the outbreak reporting rate at the EU and MS level are plotted in Figure 56 . As the % variation is a relative measure of the increase or decrease in the frequency of FBOs reporting in 2019 compared with 2018, the figure allows a direct comparison between MS, regardless of the characteristics of FBOs surveillance. Seventeen MS (Austria, Bulgaria, Cyprus, Czechia, Estonia, France, Germany, Hungary, Italy, Latvia, Malta, the Netherlands, Poland, Portugal, Spain, Sweden, the United Kingdom) reported small variations with the outbreak reporting rate remaining relatively stable (i.e. below 20% increase). Eight MS reported large variations (≥ 20%) either increasing (Belgium, Croatia, Greece, Ireland, Lithuania) or decreasing (Denmark, Finland, Romania). Information provided by the reporting MS in their national zoonoses report 21 may help understand whether recent changes in the FBOs surveillance might have contributed to these variations. In some MS, improvements of procedures for outbreak detection and investigation and/or increased awareness (sensitivity) of consumers may be the likely reasons for such rise. Over the longer period (2010-2019), eight MS (Austria, Denmark, France, Hungary, Latvia, Lithuania, the Netherlands, the United Kingdom) and one non-MS (Norway) reported statistically significant variations in the rate of outbreak reporting (Figure 57 ). Although these trends should be interpreted with caution for the reasons explained above in Section 3, it is important to disclose the country-specific pattern of causative agents being monitored in outbreaks (Section 4.2) and their relative dynamics over time (Section 4.6) , to unravel the components underlying these trends In Austria, Hungary and Lithuania outbreak trends are mainly influenced by specific agents' variations over time, in particular Salmonella (section 4.6). For France and the United Kingdom, this is less evident. Trends observed for the Netherlands and to a lesser extent for Latvia seem to be driven by an increased reporting of small outbreaks of unknown aetiology. The trends in the number of outbreaks reported by MS were mostly consistent with trends in the number of cases reported during 2010-2019 (data not shown), except for Austria and the United For Slovenia the % variation cannot be calculated due to missing data reporting for 2018. Slovakia did not report data on outbreaks in 2019. Kingdom. In Austria, 571 more cases than in 2018 were counted, corresponding to a 2.6-fold increase. This rise was mainly due to two large general outbreaks caused by S. Enteritidis and norovirus in 'eggs and egg products', that each included more than 300 cases. In the United Kingdom, cases decreased over the years in parallel with the number of outbreaks, even if less markedly and with large yearly fluctuations. Repor ng rate (* 100,000) N. outbreaks In 2019, a causative agent was identified in 3,101 FBOs (59.9% of total outbreaks) causing 35,969 cases (72.7% of total cases), 3,290 hospitalisations (85.3% of total hospitalisations) and 54 deaths (90.0% of total deaths). Figure 58 shows the agents most frequently implicated in FBOs in the EU. For a high proportion of outbreaks (40.1%), the causative agent was 'unknown' or 'unspecified'. The Netherlands (693 outbreaks) , Belgium (554 outbreaks), France (288 outbreaks) and Spain (229 outbreaks) contributed most to this reporting (1,764 outbreaks altogether; 85.1% of outbreaks with 'unknown' or 'unspecified' causative agent) . Bacteria were reported to have caused most outbreaks (N = 1,364; 26.4%) followed by bacterial toxins (N = 997; 19.3%), viruses (N = 554; 10.7%), other causative agents (N = 155; 3.0%) and parasites (N = 31; 0.6%). Table 51 provides a detailed overview of the causative agents involved in FBOs and their overall impact on health in the EU in 2019. For each pathogens group and single causative agent, the proportion of hospitalisations and deaths among cases and the mean outbreak size facilitate description of the general characteristics of the FBOs and their impact on health. The highest proportion of hospitalisations and deaths were observed for outbreaks caused by bacteria. Salmonella was responsible for the highest number of hospitalisations (N = 1,915) and L. monocytogenes, alone, caused more than half of the fatal illnesses (N = 31). The number of deaths due to FBOs caused by L. monocytogenes doubled, compared with 2018 (10 deaths more than in 2018; 47.6% increase). Fatal cases also increased among outbreak cases caused by B. cereus (N = 7; 6 cases more than in 2018) mainly due to a single outbreak in France, with five fatal events reported among 17 cases. The breakdown of causative agents by countries is in Figure 59 . Sankey diagrams by type of agent are included in the supplementary information. 'Bacillus cereus' also includes FBOs whose causative agent was encoded as B. cereus enterotoxins. 'Clostridium perfringens' also includes FBOs whose causative agent was encoded Clostridium unspecified. 'Staphylococcus aureus' also includes FBOs whose causative agent was encoded as either Staphylococcus, unspecified or Staphylococcal enterotoxins. 'norovirus' also includes FBOs whose causative agent was encoded as Calicivirus, unspecified. 'Other causative agents' includes atropine and unspecified toxins. As the monitoring and the reporting of food-borne outbreaks among MS is poorly harmonised, the interpretation of pooled data at the EU level requires caution, as the situation at single MS level may differ importantly. The frequency distribution of the causative agents implicated in FBOs by MS is shown in Figure 60 . The size and colour of each sector are proportional to the number of outbreaks and cases associated with each causative agent. The graphic aims to emphasise the major differences between MS in the causative agents being reported in FBOs rather than providing details. A graphical visualisation of the contribution (weight) of each MS to the number of FBOs reported at the EU level, by type of agent, is provided as supporting documents from the EFSA knowledge junction at zenodo (see link in the beginning of this chapter). Information on the distribution of food vehicle implicated in the FBO by causative agent is presented in Section 4.3. Moreover, for the main causative agents, the ranking of food vehicles implicated in strong-evidence outbreaks is described in tables in the supplementary information. In 2019, Campylobacter was the fourth most reported causative agent for FBOs at the EU level, with 319 outbreaks communicated to EFSA (mostly weak-evidence outbreaks), 1,254 cases of illness and 125 hospitalisations. Campylobacter was the leading causative agent in FBOs in Austria (22 outbreaks) and Germany (166 outbreaks). Campylobacter jejuni and C. coli were identified in 72 and 7 outbreaks, respectively. However, most Campylobacter outbreaks were reported without speciation information (240 outbreaks: 75.2%). Three MS (Germany, France and Austria) accounted for most of Campylobacter FBO reporting (N = 250; 78.4% of all Campylobacter outbreaks) in the EU. Outbreaks were predominantly small-sized events of less than 10 cases (N = 298; 93.4%). However, single larger general outbreaks including up to 91 cases were reported by Denmark, France, Germany, Spain, Sweden and the United Kingdom. None of these were associated with C. coli. Outbreaks caused by Listeria monocytogenes in 2019 merit attention as they caused the highest burden in terms of deaths (N = 31; 51.7% of all outbreak associated deaths). In 2019, the number of outbreaks caused by L. monocytogenes (n = 21) was 50% higher compared with 2018 (n = 14) and the related illnesses jumped from a total number of 748 cases reported at the EU level between 2010 and 2018 (83.4 annual cases on average) to 349 cases. This increase was mainly due to outbreaks in Spain, which reported 3 outbreaks, 225 cases, 131 hospitalisations and 3 deaths, compared with zero reported in 2018. Most of the cases reported by Spain were associated with a community-wide outbreak that was considered one of the largest L. monocytogenes outbreaks has ever occurred in the EU and which was linked to the consumption of contaminated meat and meat products (see dedicated text box). The death toll linked to L. monocytogenes outbreaks was particularly high in the United Kingdom with 12 deaths among 17 outbreak related illnesses (seven deaths were reported from a single outbreak in hospital setting). Overall, in the EU the case fatality rate in L. monocytogenes outbreaks reported in 2019, 8.9% , was the highest among all causative agents implicated in FBOs. In the summer of 2019, a large community-wide outbreak caused by Listeria monocytogenes infection was detected in Andalusia (Spain). It started at the end of July 2019 with small household outbreaks in Andalusia and progressed up to involve 207 reported cases (189 confirmed cases with the human and food strains sharing the same sequences and 18 confirmed without human sequences available). Moreover, few cases were detected in other Spanish regions. Patients became infected through the consumption of chilled roasted pork meat contaminated with L. monocytogenes. Patients involved in the outbreak developed different clinical conditions, depending on the age, the health and pregnancy status and the presence of underlying conditions including involvement of the central nervous system, sepsis, stillbirth, abortion and preterm birth. Confirmed cases with a predominance of gastrointestinal symptoms presented an incubation period of 3 or less days, while in cases without gastroenteritis, the period of incubation was longer than 7 days. Cases needing hospitalisation were 131. Three deaths were reported among outbreak cases. Based on the information available in the alert published by the Spanish Ministry of Health, the epidemiological investigation, the food analysis and the food business operator inspections made it possible to trace back the origin of the contamination to a single manufacturer located in Andalusia (Spain). Joint comparative analysis of the sequences obtained from the clinical isolates of L. monocytogenes and from the strains isolated from the food products and from the environment (contact surface) at the manufacturing plant revealed a close genetic similarity, so confirming the evidence obtained from the tracing-back. The implicated products were withdrawn, and the manufacturing activity of the plant was suspended on August 14 and finally the products were recalled from the market on 16 August. Consumers were warned about the risk posed by the consumption of chilled roasted pork products and products of the implicated brand through a public communication campaign. The EU Commission and the other EU MS were also alerted through the Rapid Alert System for Food and Feed (RASFF). Other information systems at the EU and international level were used by the Spanish Competent Authority to deliver information about the ongoing outbreak including the Early Warning and Response System (EWRS) managed by the EU Commission, the Epidemic Intelligence Information System (EPIS) managed by ECDC and the INFOSAN managed jointly by FAO and WHO. Two outbreak cases were communicated from abroad. Both cases were both linked to the consumption of meat products purchased in Andalusia. The distribution of the contaminated products, even if on a large scale, was limited to Spain and this explains the national dimension of the outbreak. Nonetheless, the outbreak had a remarkable impact on media and public opinion also Shiga toxin-producing E. coli (STEC) Next to Salmonella and Campylobacter, Shiga toxin-producing E. coli (STEC) were the third most frequent bacterial agents detected in FBOs in the EU, with 42 outbreaks, 273 cases, 50 and 1 death reported in 2019. Only four of these were classified as strong-evidence outbreaks and for 17 outbreaks (40.5%) only, information on STEC serogroup was available. Although the STEC serogroup is no longer considered a valid predictor of the virulence, it plays an important role as a broad epidemiological marker. STEC O157, O26 and O145 were identified in nine, seven and one outbreaks in the EU, respectively. A single strong-evidence outbreak caused by STEC O26 was also reported by Iceland. Like in recent years, in 2019 STEC have been the leading agents of food-borne outbreaks in Ireland. Shigella was detected in 22 outbreaks, involving 106 cases and in 19 hospitalisations reported by MS (see Table 2 for details for EU reports), mostly small-sized events. In addition, two non-MS (Norway, Serbia) reported three outbreaks with 38 cases and four hospitalisations. Shigella sonnei was detected in four outbreaks (two of them were strong-evidence outbreaks) reported by three MS (France, Poland, Sweden) and Norway. Shigella flexneri was detected in three outbreaks reported by Sweden and Serbia (all strong-evidence outbreaks), respectively. Shigella flexneri serotype 3a was detected in a single medium-size general outbreak in Sweden with 12 cases and four hospitalisations. In 2019, outbreaks and illnesses by Yersinia (15 and 149, respectively) were reported by seven MS (Denmark, Finland, France, Germany, Lithuania, Poland, Sweden) in numbers close to recent years. Hospitalisations were reported for 14 cases. Yersinia enterocolitica was identified as the causative agent in all these outbreaks but one. Interestingly, two strong-evidence outbreaks caused by Y. enterocolitica biotype 4 were part of the same single multi-country outbreak linked to the consumption of food imported to both the Swedish and Danish markets. Among bacterial pathogens less frequently reported in food-borne outbreaks, Arcobacter butzleri, previously named Campylobacter butzleri, was detected in a single weak-evidence outbreak in Belgium involving 40 cases (no hospitalisations). Latvia, Spain and Sweden reported four outbreaks caused by enterotoxigenic E. coli (ETEC) which involved 199 cases and 7 hospitalisations. The largest event occurred in Sweden led to 130 notified cases. Another strong-evidence general outbreak by enteropathogenic E. coli (EPEC) with 38 cases was also reported by Sweden. Latvia and Norway reported one single outbreak caused by EPEC, each. Four outbreaks caused by E. coli 'unspecified' (including two strong-evidence outbreaks) were notified by Bulgaria, Spain and Serbia. Although the number of outbreaks caused by ETEC and EPEC is too small to draw conclusions on their trend over years, it is noteworthy that only four and three outbreaks caused by ETEC and EPEC, respectively, had been reported to EFSA by MS between 2010 and 2018. It is possible that the rise observed in 2019 may be linked to an improved capability to detect E. coli in food and to characterise E. coli pathogroups, even though no official methods exist for the detection of ETEC and EPEC in foodstuffs. Vibrio was identified in four small food-borne outbreaks reported by France and Italy. The outside Spain as it was considered one of the largest outbreaks of listeriosis that ever occurred in this country and in the EU. The importance of multisectoral collaboration and prompt sharing of information and the need for strengthening the control of L. monocytogenes at all stages in the food manufacturing and distribution are key points of the lessons learned from this outbreak. For more information on this outbreak: World Health Organisation (WHO) https://www.who.int/csr/don/16-September-2019-listeriosis-spain/en/ Spanish Ministry of Health https://www.mscbs.gob.es/profesionales/saludPublica/ccayes/alertasActual/listeriosis/home.htm agent was identified as V. parahaemolyticus in all French outbreaks, while no information was available for the others. Francisella tularensis, the causative agent of human tularaemia, a severe condition characterised by multiple clinical symptoms, was reported in two strong-evidence outbreaks in Norway and Serbia, causing 24 illnesses and 6 hospitalisations. Outbreaks caused by bacterial toxins represented an important proportion of all FBOs reported in the EU in 2019 (n = 997; 19 .3% of all outbreaks) and were mostly classified as weak-evidence outbreaks. These outbreaks caused a total of 10,555 cases, 361 hospitalisations and 14 deaths (Table 51 ). Outbreaks caused by bacterial toxins were mostly reported by France that communicated 876 outbreaks (87.9% of all outbreaks caused by bacterial toxins). In France, bacterial toxins were the leading cause of food-borne outbreaks. Toxins produced by B. cereus (155 outbreaks, 1,636 cases, 44 hospitalisations) were the agents most frequently reported at the EU level with a number of FBOs twice as much as the FBO numbers due to toxins produced by C. perfringens (75 outbreaks, 2,426 cases, 27 hospitalisations) or S. aureus (74 outbreaks, 1,400 cases, 141 hospitalisations). Fourteen deaths were reported among food-borne illnesses due to poisoning caused by bacterial toxins which correspond to a high proportion of all fatal cases reported in 2019 in FBOs (23.3%). Bacillus cereus was responsible for seven deaths. Five of them were associated with a single outbreak that occurred in a residential institution (nursing home or prison or boarding school) leading to 26 cases and 17 hospitalisations. Six deaths were caused by both C. perfringens and other undefined bacterial toxins. Clostridium botulinum (7 outbreaks, 17 cases and 15 hospitalisations in 2019) was responsible for one death. In 2019, 16 strong-evidence outbreaks, and 58 weak-evidence outbreaks caused by S. aureus enterotoxins poisoning were reported by 13 MS (Bulgaria, Croatia, Cyprus, Finland, France, Germany, Hungary, Italy, Poland, Portugal, Romania, Spain, Sweden) . Serbia and Norway also reported two strong-and one weak-evidence outbreaks, respectively. Among MS, most of these outbreaks were reported as general outbreaks (N = 54) and involved overall 1,324 cases (94,6% of all outbreak cases caused by S. aureus). Two large outbreaks were reported, causing 380 illnesses in Hungary and causing 300 cases of illnesses including one hospitalisation in France. The most severe outbreak was described in Italy where 44 out of 70 cases (62%) needed hospitalisations. No deaths due to S. aureus poisoning was reported. The number of outbreaks caused by S. aureus poisoning showed a 35.7% drop in 2019 compared with 2018, mainly due to fewer outbreaks in France and Spain (17 and 18 outbreaks less). In many food poisonings attributed to the intake of bacterial toxins, the implicated agent was not identified but generically classified as 'bacterial toxins, unspecified' (n = 686; 68.9%) . These events caused 5,076 illnesses and 134 hospitalisations. Such reporting was adopted by France only for the suspect cases identified on the basis of clinical signs, the median incubation time and types of consumed foods, when the pathogens and/or toxins were neither detected in human samples and/or food leftovers nor in food handling environment. A wide range of viruses were reported in FBOs in 2019, including adenovirus, flavivirus and Tick-borne encephalitis virus, hepatitis A virus, hepatitis E virus, norovirus, sapovirus, rotavirus. Overall, 554 outbreaks caused by food-borne viruses led to many illnesses (12,227 cases; 24.7% of all outbreaks cases). Nevertheless, no deaths were reported in FBOs caused by viruses and the number of hospitalisations (456 hospitalisations; 12% of the cases) was smaller, compared with FBOs caused by bacteria and other causative agents. In 2019, norovirus (and other Calicivirus) was the second most frequently reported causative agent in FBOs in the EU and was reported by 21 MS (Figure 58 ). In four of these (Denmark, Finland, Lithuania, the United Kingdom) and one non-MS (Norway) this agent was the leading cause of FBOs. Norovirus was associated with 457 outbreaks and, most importantly, with 11,125 related illnesses (22.5% of total cases) meaning one in five of all outbreak related illnesses in the EU. Norovirus was associated with large outbreaks (24.3 cases on average). In 2019, the number of outbreaks of medium size (involving between 10 and 100 cases) and large size (> 100 cases) were 204 and 14, respectively. Two very large outbreaks, reported by Greece and France, each involved more than 500 illnesses. Most norovirus outbreaks (N = 264; 57.8%) were general outbreaks; a proportion much higher than for other causative agents. In 2019, outbreaks caused by norovirus increased by 13.1% (53 outbreaks more than in 2018), with five countries contributing most to this rise, France (224 outbreaks more than in 2018), Lithuania (21 outbreaks more than in 2018), the Netherlands (17 outbreaks more than in 2018) and the United Kingdom (16 outbreaks more than in 2018). In total, 22 hepatitis A (including other Hepatitis virus, unspecified) outbreaks involving 135 cases were reported in 2019 by five MS (Germany, Italy, Poland, Spain, Sweden). In addition, the Republic of North Macedonia and Norway also reported three and one outbreaks, respectively. Compared with 2018, the number of notified hepatitis A (including other hepatitis virus, unspecified) outbreaks decreased in the EU (36 outbreaks less; 62.1% decrease), mainly due to reduced reporting by Poland. hepatitis A outbreaks were characterised by a high percentage of cases needing hospitalisation (99 cases, 73.3% of cases). Flaviviruses, including tick-borne encephalitis virus was associated with an even higher proportion of hospitalisations (80% of cases) and detected in three outbreaks and 15 cases. The number of FBOs caused by parasites reported in 2019 was limited compared with the other agents (31 outbreaks in MS and five outbreaks in non-MS) and fewer than in 2018. Among Trichinella outbreaks (N = 5) in the EU, which was half of the number compared with 2018, T. spiralis accounted for two events (six outbreaks less than in 2018) and T. britovi was identified in a single outbreak reported by Italy. No information was available for the remaining Trichinella outbreaks reported by one MS (N = 2) and one non-MS (N = 2). In 2019, Giardia caused most outbreaks (N = 14), involving parasites. Although there were five fewer outbreaks reported in 2019 compared with 2018, the total number of illnesses in 2019 increased fourfold, mainly due to a single large weak-evidence outbreak caused by G. intestinalis (lamblia) reported by Italy, which resulted in 199 illnesses. G. intestinalis (lamblia) was identified in four outbreaks while no details on the species was provided for the remaining outbreaks. Cryptosporidium (11 outbreaks and 468 cases in 2019) was the only agent among parasites that caused more outbreaks (2 outbreaks more) and cases (425 cases more; 988.4% increase) than in 2018. Seven outbreaks with in total 304 notified cases were reported by Sweden after no reported outbreaks of cryptosporidiosis during the two former years. Overall, C. parvum was implicated in eight outbreaks while no information on the species was available for the other outbreaks. This group of outbreaks includes mainly events caused by 'histamine', 'marine biotoxins' and a few other chemical agents of biological origin that accidentally may contaminate food or its ingredients. The reporting of outbreaks caused by other causative agents is the least harmonised among MS. These agents are not regularly covered by the national outbreak surveillance programmes that in many MS only target infectious agents. Consequently, outbreaks communicated to EFSA are sparse and the importance of this type of food poisoning is highly likely underestimated at the EU level. In 2019, 96 outbreaks caused by histamine were reported by 11 MS while only three MS reported 48 outbreaks caused by marine biotoxins (Table 51 and Figure 59 ). Histamine poisoning is usually associated with consumption of poor-quality raw materials preserved in inadequate conditions during storage and preparation. France and Spain are the MS which contribute more regularly to the reporting of outbreaks involving marine biotoxins. These biotoxins are mainly produced by algae or phytoplankton and accumulate in fish and filter-feeding molluscan shellfish. The toxins include also ciguatoxin, saxitoxin and its muscle-paralyzing toxin, okadaic acid. Ciguatoxin, the causative agent of Ciguatera fish poisoning is characterised by gastrointestinal, neurological and/or generalised disturbances and occurs most commonly in fish from Pacific, Caribbean and Indian Ocean regions. In 2019, France reported 19 outbreaks caused by Ciguatoxin. In Spain, the number of outbreaks caused by marine biotoxins (N = 13) was higher than in 2018 (8 outbreaks more; 160% increase). The United Kingdom reported in 2019 a single outbreak with 13 illnesses involving okadaic acid, a heat stable toxin that can be found in various species of shellfish. Only two outbreaks caused by okadaic acid had been previously reported to EFSA in 2012, although the contamination of various type of shellfish by okadaic acid has not rarely been signalled through the Rapid Alert System for Food and Feed (RASFF) system. Several reasons may explain the reporting of unknown/unspecified agents, including late reporting of illness, failure to detect causative agents in patients or in the food, unavailability of clinical or food samples (e.g. leftovers), delay in sample collection etc. For the same reasons, few outbreaks of unknown aetiology were classified as strong-evidence outbreaks. In 2019, 2,074 food-borne outbreaks of unknown aetiology accounted for 40.1% of total outbreaks and 27.3% of illnesses in the EU. At the country level, these proportions varied hugely. Outbreaks with unknown aetiology were mainly reported by Belgium and the Netherlands and these FBOs accounted for 1,274 outbreaks (60.1% of all outbreaks caused by unknown agents notified in the EU). They were mainly weak-evidence, smallsized (< 10 cases) events that included each, less than four cases, on average. In Belgium and in the Netherlands, this type of outbreak accounted for the majority the FBOs (Figure 60 ). These findings suggest that outbreaks caused by unknown agents occurred in confined contexts such as domestic settings or small groups, for which the identification of the link among cases was probably relatively easy. Conversely, 250 outbreaks with unknown aetiology involving each more than 10 cases (medium and large size outbreak) were reported by 15 MS. Not all MS, however, reported outbreaks of unknown aetiology to EFSA in 2019. The short-term relative variation (%) of the annual number of strong-evidence and weak-evidence outbreaks for specific causative agents and by MS are plotted in Figure 61 . Austria (1) Bulgaria (0) Cyprus (0) Denmark (3) Finland (0) Germany (23) Hungary (2) Italy (2) Lithuania (0) Malta (0) Poland (1) Romania (0) Spain (7) United Kingdom (2) Montenegro (0) Rep. of North Macedonia (0) Switzerland (0) S. Typhimurium and monophasic var. -50% 50% 114% 0% -120% -60% 0% 60% 120% Austria (1) Bulgaria (0) Cyprus (0) Denmark (1) Finland (2) Germany (5) Hungary (0) Italy (1) Lithuania (0) Malta (0) Poland (0) Romania (0) Spain (3) United Kingdom (3) Montenegro (0) Rep. of North Macedonia (0) Switzerland (0) L. monocytogenes 2019/2018 relaƟve variaƟon (%) Country (N of outbreaks reported i n 2019) -150% -100% -50% 0% 50% 100% Austria (17) Bulgaria (0) Cyprus (0) Denmark (9) Finland (1) Germany (127) Hungary (14) Italy (18) Lithuania (21) Malta (6) Poland (257) Romania (3) Spain (152) United Kingdom (15) Montenegro (0) Rep. of North Macedonia (1) Switzerland (0 (0) Denmark (1) Finland (0) Germany (72) Hungary (6) Italy (2) Lithuania (21) Malta (0) Poland (209) Romania (2) Spain (18) United Kingdom (7) Montenegro (0) Rep. of North Macedonia (0) Switzerland (0 Austria (0) Bulgaria (2) Cyprus (1) Denmark (0) Finland (1) Germany (3) Hungary (3) Italy (9) Lithuania (0) Malta (0) Poland (1) Romania (2) Spain (10) United Kingdom (0) Montenegro (0) Rep. of North Macedonia (0) Switzerland (0) S. aureus 2019/2018 relaƟve variaƟon (%) (0) Bulgaria (0) Cyprus (0) Denmark (0) Finland (1) Germany (3) Hungary (6) Italy (4) Lithuania (0) Malta (0) Poland (1) Romania (0) Spain (13) United Kingdom (0) Montenegro (0) Rep. of North Macedonia (0) Switzerland (0) B. cereus 2019/2018 relaƟve variaƟon (%) Country (N of outbreaks reported in 2019) Austria (2) Bulgaria (0) Cyprus (0) Denmark (1) Finland (0) Germany (7) Hungary (0) Italy (2) Lithuania (0) Malta (1) Poland (0) Romania (0) Spain (1) United Kingdom (6) Montenegro (0) Rep. of North Macedonia (0) Switzerland (0) Shigatoxin-producing E.coli 2019/2018 variaƟon (%) -150%-50% 50% 150% 250%350% 450% Austria (0) Bulgaria (0) Cyprus (0) Denmark (10) Finland (1) Germany (6) Hungary (2) Italy (3) Lithuania (0) Malta (0) Poland (0) Romania (0) Spain (11) United Kingdom (7) Montenegro (0) Rep. of North Macedonia (0) Switzerland (0) C. perfringens 2019/2018 relaƟve variaƟon (%) Country (N of outbreaks reported in 2019) Estonia (0) Finland (0) France (1) Germany (0) Greece (0) Hungary (0) Ireland (0) Italy (2) Latvia (0) Lithuania (0) Luxembourg (0) Malta (0) Netherlands (0) Poland (2) Portugal (0) Romania (1 ) Slovenia (0) Spain (1) Sweden (0) United Kingdom (0) MS (7) Iceland (0) Montenegro (0) Norway (0) -150% -100% -50% 0% 50% 100% Estonia (0) Finland (0) France (0) Germany (9) Greece (0) Hungary (0) Ireland (0) Italy (6) Spain (1) Sweden (1) United Kingdom (0) MS (22) Iceland (0) Montenegro (0) Norway (1) (13) Greece (1) Hungary (1) Ireland (3) Italy (8) Latvia (8) Lithuania (21) Luxembourg (0) Malta (1) Netherlands (17) Poland (42) Portugal (1) Romania (0) Slovenia (0) Spain (35) Sweden (11) United Kingdom (16) MS (458) Iceland (0) Austria (0) Belgium (1) Bulgaria (0) CroaƟa (2) Cyprus (0) Czechia (0) Denmark (0) Estonia (0) Finland (1) France (36) Germany (4) Greece (0) Hungary (0) Ireland (0) Italy (26) Latvia (1) Lithuania (0) Luxembourg (0) Malta (8) Netherlands (1) Poland (0) Portugal (0) Romania (0) Slovenia (0) Spain (9) Sweden (7) United Kingdom (0) MS (96) Iceland ( This section aims to describe the characteristics of food vehicles that in 2019 were implicated in outbreaks in the European countries. The description of the implicated food vehicles relies on strongevidence outbreaks, because only for these events, the link between the consumption of foods and the illnesses was proved with minimal uncertainty. Strong-evidence outbreaks represent a minority of all FBOs reported in 2019 (716 outbreaks, 13.8%). The overview of the food vehicles implicated in strong-evidence outbreaks and illnesses in the EU in 2019 is described in Table 52 . For a correct interpretation of the data, it is worth remembering that the pattern of food vehicles implicated in outbreaks at the EU level, is highly influenced by those countries which contributed the most to strong-evidence outbreaks data collection (Table 50 ). In 2019, these were France, Spain, Poland and Italy. Altogether these four MS provided information on almost three quarters of the total number of strong-evidence outbreaks (532 outbreaks, 74.3% of strong-evidence outbreaks), while the remaining outbreaks (184 outbreaks) were contributed by 19 MS altogether. (0) Spain (13) Sweden (0) United Kingdom (1) MS (48) Iceland (0) Montenegro (0) Norway (0) 138% -150% -100%-50% 0% 50%100%150% Austria (0) Belgium (554) Bulgaria (11) CroaƟa (6) Cyprus (0) Czechia (4) Denmark (0) Estonia (1) Finland (19) France (288) Germany (35) Greece (2) Hungary (5) Ireland (3) Italy (45) Latvia (0) Lithuania (2) Luxembourg (0) Malta (21) Netherlands (693) Poland (125) Portugal (6) Romania (0) Slovenia (0) Spain (229) Sweden (19) United Kingdom (6) MS (2074) Iceland (0) Montenegro (2) Norway (13) The consumption of food of animal origin ('fish and fishery products', 'eggs and egg products', 'meat and meat products', 'milk and milk products') was associated with most of the strong-evidence FBOs (469 outbreaks; 65.5%) and illnesses (5,709 cases; 41.7%) reported in 2019. Food of animal origin was mainly implicated in outbreaks caused by Salmonella (182 outbreaks; 38.8% of all FBOs by food of animal origin), norovirus and other Calicivirus (148 outbreaks; 31.6%), histamine (21 outbreaks; 4.5%), C. perfringens (20 outbreaks; 4.3%) and Campylobacter (14 outbreaks; 3.0%) (Figure 62) . The importance of 'crustaceans, shellfish, molluscs and products thereof' (159 strongevidence outbreaks) increased substantially in the EU in 2019 (80 outbreaks more; 101.3% more than in 2018) and in particular in France which reported 129 outbreaks (81.1% of total outbreaks in the EU) compared with 32 in 2018 (287.5% increase). Almost all outbreaks caused by 'crustaceans, shellfish, molluscs and products thereof' reported by France involved norovirus (124 outbreaks, 756 cases). No increase was observed in the other MS (Croatia, Denmark, Finland, the Netherlands, Spain, Sweden or the United Kingdom) which reported similar outbreaks. In Sweden, many cases (N = 208) became infected following consumption of 'crustaceans, shellfish, molluscs and products thereof', oysters, contaminated with norovirus GI and GII. Two more outbreaks involving norovirus in oysters causing 126 cases of illness were reported by Norway. The increase observed in France in 2019 was the only driver of the overall rise of outbreaks by 'fish and fishery products' group, which was the food most frequently implicated in strong-evidence outbreaks in the EU. 'Eggs and egg products', the next most frequently reported foodstuff, were implicated in 108 strong-evidence outbreaks reported by 11 M (Austria, Croatia, France, Germany, Hungary, Italy, Latvia, the Netherlands, Poland, Spain and the United Kingdom) . At the EU level, outbreaks linked to contaminated 'eggs and egg products' reduced by 24.5% in 2019 (35 outbreaks less than in 2018). The lack of data reporting by Slovakia may have contributed to this decrease. However, a significant drop was also observed for Italy, Poland and Spain. Germany and the United Kingdom were the only MS where outbreaks by 'eggs and egg products' increased. Consumption of contaminated 'eggs and egg products' has been often associated with very large EU-wide outbreaks, such as the extensive outbreaks that in 2017, 2018 and 2019 involved many EU countries. In 2019, 20 mediumsized outbreaks (including from 10 to 100 cases) and one single large outbreak (> 100 cases) linked to this food type caused 592 and 321 illnesses, respectively. At least six of these events were outbreaks dispersed in the EU, with cases scattered over large geographic areas including cross border zone. In these outbreaks, tracing of patients and the trace-back of batches of 'eggs and egg products' delivered to the marketplaces as well as typing of human and food isolates by WGS have been successfully applied. Outbreaks caused by 'meat and meat products' (151 outbreaks, Table 52 ) accounted for an important proportion of strong-evidence outbreaks in the EU. In this group, outbreaks by 'meat and meat products, unspecified' (41 outbreaks) , the item most frequently reported, had a twofold increase compared with 2018. This surge was mainly driven by Spain that reported 19 outbreaks in 2019 linked to this foodstuff (16 more than in 2018); bacterial toxins (3), Salmonella (3), L. monocytogenes (1) and ' Unknown' causative agent (12) . The number of outbreaks caused by 'pig meat' was stable in all the MS except France, which reported 19 outbreaks compared with five outbreaks in 2018. Sixteen of these were caused by Salmonella, including S. Typhimurium (12 outbreaks) and S. Infantis (two outbreaks). Outbreaks by 'other or mixed red meat and products thereof' also increased in France in 2019 (11 outbreaks, six more than in 2018). The implicated agent was mainly S. Typhimurium and its monophasic variants (eight outbreaks in total). The consumption of 'poultry meat' was associated with many illnesses (N = 870) in strongevidence outbreaks in 2019. This foodstuff was mostly identified in Salmonella outbreaks (19 outbreaks), Campylobacter (eight outbreaks) and bacterial toxins other than C. botulinum (9 outbreaks). Overall, the reported number of outbreaks caused by 'poultry meat ' (38 outbreaks) was rather stable at the EU level, even though seven MS (Denmark, Hungary, Finland, Latvia, the Netherlands, Spain and the United Kingdom) reported mild or moderate increases. Poland, in contrast, reported a reduction in outbreaks caused by 'poultry meat' (five outbreaks in 2019; eight less than in 2018) but this did not correspond to a parallel decrease in the number of illnesses. Denmark reported three outbreaks caused by C. jejuni in 'poultry meat' (115 cases involved) after 2 years with no reported outbreaks caused by 'poultry meat'. The contamination was traced back to a slaughterhouse. The same agent/food pair was also implicated in the only outbreak caused by 'poultry meat' reported by Finland. The number of strong-evidence outbreaks associated with the consumption of 'cheese' decreased markedly, at the EU-level and in all MS. There were 20 outbreaks in 2018 and four in 2019, which is the lowest number ever reported since the beginning of the FBOs data collection in the EU. Outbreaks caused by 'milk' were also less frequently reported in 2019 (nine outbreaks) mainly due to a remarkable decrease in milk-borne outbreaks of Campylobacter in Germany (three outbreaks in 2019; 19 outbreaks in 2018; 10 outbreaks in 2017). The number of outbreaks implicating other dairy products (four outbreaks) did not substantially change in 2019 compared with previous years in the MS. Iceland reported a general outbreak connected with dairy product (ice cream) contaminated with STEC O26. In 2019, 10 MS (Denmark, Finland, France, Germany, Italy, Latvia, the Netherlands, Poland, Spain, Sweden) reported 51 outbreaks associated with the consumption of FNAO. FNAO were mainly implicated in outbreaks caused by norovirus (14 outbreaks), Salmonella, (12 outbreaks), B. cereus (five outbreaks) and Cryptosporidium (four outbreaks). 'Vegetables (and juice) ' (30 outbreaks) were the most frequently reported food vehicle of this group. Interestingly, the mean size of outbreaks associated with this food (21.2 cases/outbreak) was approximately twofold larger than outbreaks linked to consumption of food of animal origin (12.2 cases/outbreak). Various types of leafy-green vegetables, olives, tomatoes, cucumbers and radish sprouts were the items described in this group. Vegetable-associated outbreaks increased markedly in Sweden and less importantly in Italy and Latvia. Sweden was the MS reporting the highest number of the outbreaks caused by 'vegetables (and juice)' (seven outbreaks). Four of these were associated with the consumption of kale or vegetable juice, contaminated by Cryptosporidium parvum, with 223 cases notified. Another 132 cases of illness were caused by two Salmonella outbreaks associated with the consumption of various types of tomatoes. In Italy and Latvia outbreaks by vegetables were less clearly associated with a specific causative agent even if norovirus, as in many other countries, was mostly identified. In Denmark and Sweden, a single outbreak caused by Y. enterocolitica biotype 4 associated with the consumption of imported fresh green spinach contaminated at the primary production level led to 20 and 37 cases, respectively. In Spain, 50 cases were involved in an outbreak caused by vegetables (not specified) contaminated with Enterotoxigenic E. coli. 'Fruits and juice', in particular 'frozen and fresh berries ', 'pre-cut melon' and 'dates', were implicated in outbreaks caused by norovirus (four outbreaks), hepatitis A (one outbreak), Salmonella (two outbreaks) and B. cereus toxins (one outbreak). 'Sweets and chocolate' were mainly identified in Salmonella outbreaks (five outbreaks) and 'cereal products including rice and seeds/pulses (nuts, almonds)' in outbreaks caused by Total ( bacterial toxins (three outbreaks). The only outbreak associated with 'herbs and spicy' was reported by Italy and was caused by norovirus. These foodstuffs include composite food resulting from the assembly of multiple ingredients or highly processed or manipulated foods. Interestingly, outbreaks associated with these foodstuffs were larger on average (29.8 cases/outbreak), than outbreaks associated with either food of animal origin (12.2 cases/outbreak) or FNAO (21.3 cases/outbreak). In 2019, the consumption of 'mixed foods' caused the highest number of cases of illness (N = 3,079, Table 52 ) in strong-evidence outbreaks. This foodstuff was associated with a wide range of causative agents including bacteria (Salmonella, Campylobacter, L. monocytogenes, Shigella), norovirus, bacterial toxins (B. cereus, C. botulinum, C. perfringens, S. aureus) and histamine. Outbreaks caused by 'mixed food' were mainly general outbreaks and were reported by 14 MS. In Hungary, the consumption of various types of 'mixed food' was associated with five outbreaks that altogether involved 946 illnesses. The largest event (575 cases) was associated with various types of contaminated mixed food, also involving crosscontamination, by S. Enteritidis. This was the largest outbreak by mixed food ever registered since the beginning of the surveillance of FBOs in the EU. Other seven large outbreaks (> 100 cases) linked to mixed food were reported by Belgium, Denmark, Hungary, Poland, Romania. Outbreaks caused by 'bakery products' (38 outbreaks) were mostly associated with Salmonella (31 outbreaks) and were frequently linked to domestic settings (19 outbreaks). This finding suggests that improper food handling and poor storage habits in households may contribute to this type of outbreaks. Outbreaks caused by 'other foods' (16 outbreaks) halved in 2019 compared with 2018, mainly due to decreased reporting from France (zero reporting). Few details on the type of food were provided by the seven MS (Germany, Hungary, Italy, Poland, Romania, Spain, Sweden) that reported this food vehicle. In Poland, 'other food, unspecified' was associated with two outbreaks in 'school or kindergarten'. In Germany, frozen Wakame algae was responsible for a community-wide dispersed outbreak, with 53 cases of illness. In Sweden, salad dressing basil oil contaminated with EPEC caused 38 cases of illness. 'Buffet meals' related outbreaks decreased importantly in 2019 among strong-evidence outbreaks and the reporting of this category (13 outbreaks) was quite sparse among MS. The causative agents associated with the consumption of different type of food implicated in strong-evidence FBOs are shown in the stacked bar chart in Figure 10 . Sankey diagrams by food groups are included in the supplemental information. Tables 53-56 show the top 10 pairs of causative agents and food vehicles among outbreaks having the highest health impact in 2019 in the EU in terms of total outbreaks, cases, hospitalisations and deaths, respectively. The number of MS that reported outbreaks implicating each food/agent pair is also included in the tables, to indicate how common these types of outbreaks were in the EU MS. Indeed, MS that contribute the most to the data collection may influence the rank position of the pairs. The same information for the 2010-2018 period is also shown in parallel, for trend watching purposes. 'Fish and fishery products' include 'crustaceans, shellfish, molluscs and products thereof', as well as 'fish and fish products'. 'Meat and meat products' include bovine meat and products thereof, broiler meat (Gallus gallus) and products thereof, other or mixed red meat and products thereof, other, mixed or unspecified poultry meat and products thereof, pig meat and products thereof, sheep meat and products thereof, turkey meat and products thereof. 'Food of non-animal origin' includes cereal products including rice and seeds/pulses (nuts, almonds), fruit, berries and juices and other products thereof, herbs and spices, sweets and chocolate, vegetables and juices and other products thereof. ' 'Milk and milk products' include cheese, dairy products (other than cheeses), milk. 'Other foods' includes canned food products and other foods, unspecified. 'Water' includes Tap water, including well water. France (124), Spain (7), Sweden (6), Finland (3), Denmark (2), United Kingdom (1) (9), Germany (6), United Kingdom (4), Netherlands (2), Hungary (2), Italy (2), Austria (1), Croatia (1), Latvia (1) 1 103.0 9.8 stable 3 Salmonella spp. Meat and meat products France (24), Poland (12), Spain (10), Hungary (7), Germany (5), United Kingdom (3), Denmark (2), Croatia (2), Latvia (2), Sweden (1), Netherlands (1), Czechia (1), Italy (1) Poland (12), Hungary (2), Spain (2), France (2), Germany (2), Belgium (1), Czechia (1) 'Fish and fishery products' include 'crustaceans, shellfish, molluscs and products thereof', as well as fish and fish products. 'Food of non-animal origin' includes cereal products including rice and seeds/pulses (nuts, almonds), fruit, berries and juices and other products thereof, herbs and spices, sweets and chocolate, vegetables and juices and other products thereof. 'Meat and meat products' include bovine meat and products thereof, broiler meat (Gallus gallus) and products thereof, other or mixed red meat and products thereof, other, mixed or unspecified poultry meat and products thereof, pig meat and products thereof, sheep meat and products thereof, turkey meat and products thereof. 'Fish and fishery products' include 'crustaceans, shellfish, molluscs and products thereof', as well as 'fish and fish products'. 'Food of non-animal origin' include cereal products including rice and seeds/pulses (nuts, almonds), fruit, berries and juices and other products thereof, herbs and spices, sweets and chocolate, vegetables and juices and other products thereof. 'Meat and meat products' include bovine meat and products thereof, broiler meat (Gallus gallus) and products thereof, other or mixed red meat and products thereof, other, mixed or unspecified poultry meat and products thereof, pig meat and products thereof, sheep meat and products thereof, turkey meat and products thereof. 'Food of non-animal origin' includes cereal products including rice and seeds/pulses (nuts, almonds), fruit, berries and juices and other products thereof, herbs and spices, sweets and chocolate, vegetables and juices and other products thereof. 'Meat and meat products' include bovine meat and products thereof, broiler meat (Gallus gallus) and products thereof, other or mixed red meat and products thereof, other, mixed or unspecified poultry meat and products thereof, pig meat and products thereof, sheep meat and products thereof, turkey meat and products thereof. ' Milk and milk products include cheese, dairy products (other than cheeses), milk. 'Other foods' includes canned food products and other foods, unspecified. The description of foodstuffs most frequently implicated in food-borne outbreaks provides useful indications about which sources at the primary production level or in the various sectors of food preparation should be targeted by control policies to reduce the public health impact of food-borne pathogens in humans. For each causative agent, the food vehicles implicated in outbreaks in 2019 are described in Figure 63 . In these figures, foodstuffs implicated in strong-evidence FBOs (dark coloured bars on left) are matched in parallel with suspect foods implicated in weak-evidence outbreaks (light coloured bars on the right). This visualisation allows presentation of the whole bulk of information provided by MS on food, but on the same time representing the different level of uncertainty affecting the findings. Data on foods implicated in weak-evidence FBOs must be interpreted with caution, given the high level of uncertainty affecting evidence from weak-evidence FBO. In 2019, 21 MS reported information to EFSA on the suspected food vehicle in 1,960 weakevidence outbreaks (37.9% of all outbreaks). The ranking of the importance of food very consistent with the grading based on strong-evidence outbreaks, for all the causative agents, with few exceptions. In outbreaks caused by S. Typhimurium and monophasic S. Typhimurium, 'eggs and egg products' was the foodstuff most frequently reported, followed by 'pig meat'. However, the link between the consumption of 'eggs and egg products' and the outbreaks was only supported by weak evidence (14 weak-evidence outbreaks; 22% of weak-evidence FBOs caused by S. Typhimurium and monophasic S. Typhimurium). This ranking was similar to that observed at the EU level for strongevidence outbreaks between 2010 and 2018. Discrepancies are also present in the ranking of items associated with outbreaks of STEC infection. Although 'meat and meat products' ranked first among strong-evidence outbreaks (two strong-evidence outbreak, four weak-evidence outbreaks), 'water' was the source most frequently suspected (one strong-evidence outbreak, 10 weak-evidence outbreaks). This finding deserves attention because waterborne outbreaks caused by STEC, even severe and large events, have been reported in the literature due to contamination of either public or private drinking water, recreational water, lake, rivers, wells (Vanden Esschert et al., 2020) . Interestingly, among all items described in STEC outbreaks in the last 10 years, water was the most frequently reported item (79 outbreaks between 2010 and 2018). Only a minority of these outbreaks (20 outbreaks; 25.3%) however were classified as strong-evidence outbreaks. The lack of standard methods for the detection of STEC in water and the analytical difficulties connected with this matrix could be a reason to explain the low proportion of STEC waterborne outbreaks supported by strong-evidence. The description of the settings of the outbreaks (places of exposure) characterises the stages of the food preparation chain where incidents leading to food contamination may have occurred and provides indications of where to plan risk mitigation strategies and control measures to prevent food-borne illnesses. Figure 64 describes strong-evidence FBOs' characteristics by place of exposure. The analysis of the settings implicated in FBOs in 2019 has been limited to strong-evidence outbreaks to avoid introducing the high-level of uncertainty that affected weak-evidence outbreaks reported by MS. This is evidenced by Figures 64 and 65 which show the ranking of the places of exposure implicated in strongand weak-evidence outbreaks, based on the number of outbreaks and cases of illness, respectively. In 2019, most of the strong-evidence FBOs were from 'domestic setting' (N = 296), similarly to previous years. This number is probably underestimated given that only 10 MS among those reporting strong-evidence outbreaks in 2019 (N = 23) communicated data on household outbreaks. Not surprisingly, most of the outbreaks in domestic settings are classified as 'household outbreak', meaning that all the human cases live in one single household (259 outbreaks; 87.5% of total outbreaks in domestic setting). In 'general outbreaks' (i.e. outbreaks involving cases of more than one household), (431 outbreaks; 60.2% of strong-evidence outbreaks), 'restaurant, pub, street vendors, take-away, etc.' were the settings most frequently described (202 outbreaks; 46.9% of strong-evidence general outbreaks), while 'canteen or catering to workspace, school, hospital, etc.' were the places where most cases became exposed to contaminated foods (4,899 cases; 39.3% of strong-evidence general outbreaks). Outbreaks linked to 'canteen or catering at workplace, school, hospital, etc' were on average much larger (mean cases: 52.7) than those in the 'restaurant, pub, street vendors, take-away, etc' (mean cases: 14.8 cases). In 2019, 12 large outbreaks connected to 'canteen or catering to workspace, school, hospital, etc' category were responsible altogether for 2,734 cases (20.0% of all cases involved in strong-evidence outbreaks). Eight of these large outbreaks occurred in 'school/kindergarten' and were mainly associated with S. Enteritidis in mixed foods (5 outbreaks), including one outbreak reported by Hungary that involved 575 cases and 80 hospitalisations. The three other large outbreaks were caused by B. cereus toxins, due to inadequate heat treatment, by norovirus due to food contamination by food handlers and by an unknown agent. (28), Farm (7), Others (243), Temporary mass catering (fairs or festivals) (9). N = number of outbreaks. Causative agents identified in strong-evidence outbreaks in the different settings are described in Figure 66 . The bar chart makes it possible to visualise the importance of causative agents in each group of settings. The findings refer to strong-evidence outbreaks only, to reduce the degree of uncertainty characterising weak-evidence outbreaks. Contributing factors in strong-evidence food-borne outbreaks Information on factors contributing to food contamination and outbreaks was available for a minority of food-borne outbreaks ( Figure 67 ). In household outbreaks the use of unprocessed contaminated ingredients was frequently reported (19 of 29 outbreaks with this information available). In general outbreaks, risk factors were documented in 167 strong-evidence outbreaks (38.7% of strong-evidence general outbreaks). Contamination by 'food handlers' was reported in 35 outbreaks in various settings and was mainly associated with norovirus (14 outbreaks; 16.9% of total strongevidence outbreaks caused by norovirus) and Salmonella (9 outbreaks; 6.3% of total strong-evidence general outbreaks caused by this agent). 'Cross-contamination' was identified in 39 outbreaks, mainly caused by Salmonella (15 outbreaks; 10.6% of total strong-evidence general outbreaks caused by this agent) as well as in six and four outbreaks caused by Campylobacter and L. monocytogenes, respectively (40% and 44% of total strong-evidence general outbreaks caused by these agents, each). 'Inadequate heat treatment' was identified in 45 outbreaks, mainly caused by C. perfringens toxins (14 strong-evidence outbreaks; 37.8% of total strong-evidence general outbreaks caused by this agent) and Salmonella (14 outbreaks; 9.9% of total strong-evidence general outbreaks caused by this agent). In 30 outbreaks, mainly associated with either C. perfringens toxins (12 outbreaks; 33.3% of total strong-evidence general outbreaks caused by this agent) or B. cereus, S. aureus and histamine, 'time/temperature storage abuse' was identified. 'Inadequate chilling' contributed to 24 outbreaks. Figure 68 shows the number of FBOs reported by MS during 2010-2019, by causative agent, including strong-evidence and weak-evidence FBOs. The two graphs allow demonstration of the importance of the causative agents at the EU level, in terms of absolute number of FBOs and visualising the major differences among them. It is important to remember that the variations over years in the frequency distribution of causative agents may not reflect the true epidemiological pattern at the EU level as the collection of outbreak data is not fully harmonised among MS. Figure 69 shows the distribution of Salmonella outbreaks, including strong-evidence and weakevidence ones, and the outbreak reporting rate (per 100,000) in MS and non-MS during 2010-2019. The trend analysis showed a statistically significant decrease in the number of Salmonella outbreaks for three MS (Austria, Germany, Lithuania) . The trend was primarily driven by S. Enteritidis outbreaks whose progressive decrease over the time period in question was also statistically significant in all the three MS, plus Hungary (Figure 70 ). Austria and Germany also reported significant decreasing trends for outbreaks caused by S. Typhimurium and monophasic S. Typhimurium. For Austria and Germany, the negative trend in Salmonella outbreaks matches with the corresponding significant negative time trend for the Salmonella outbreak cases (data not shown). In Lithuania outbreak illnesses also decreased, but by a lower proportion. For the other MS and non-MS, no significant trends were observed for outbreaks caused by Salmonella spp. (all serovars), S. Enteritidis or S. Typhimurium, including its monophasic variants. Note: The orange line (right axis) in the graphs represents the Salmonella outbreak reporting rate and was measured on the same scale for all MS, to allow a direct comparability among countries. The blue bars present the trend over years in terms of absolute numbers of Salmonella outbreaks, using for each country the most appropriate scale (left axis). * indicates countries with a statistically significant trend (p < 0.05) over several years. Other statistically significant trends in occurrence of FBOs by causative agents and MS are shown in Figure 70 . Given the lack of specific control programmes it is difficult to unravel the reasons underlying these trends. Campylobacter outbreaks in Austria dropped significantly in recent years. However, no information on implicated food vehicles was available for most of these outbreaks (444 of the 499 outbreaks reported between 2010 and 2019). Similarly, reasons underlying the trends for outbreaks caused by bacterial toxins and histamine could not be readily elucidated, mainly because the circumstances leading to intake of toxins or histamine through food vary importantly and are highly dependent on the conditions and practices of food preparation and preservation in the close proximity of consumers. The increasing trend for Hepatitis A in Germany refers to a small number of outbreaks and does not match with a parallel increase in the number of Hepatitis A cases. Reasons underlying increasing or decreasing trends of outbreaks caused by unknown agents might reflect progressive changes in the sensitivity of outbreak surveillance due to variations in the criteria for outbreak definition or improved citizens' engagement with FBOs surveillance. Figure 71 displays country-specific significant trends in the number of strong-evidence outbreaks for specific food vehicles, during 2010-2019. Decreasing trends were noted for 'eggs and egg products' in France and Poland, 'fish and fishery products' in the United Kingdom, 'meat and meat products' in the United Kingdom and 'mixed foods' in Belgium, Germany and Denmark. The decreasing trend for outbreaks by 'eggs and egg products' was mainly driven by S. Enteritidis in Poland and by S. Enteritidis and other serovars in France. In both countries, the number of Salmonella outbreaks decreased progressively, although with large yearly fluctuations, especially in recent years, suggesting that the trend is not stable. This is a reason of concern also considering that eggs from Poland in recent years have been repeatedly implicated in large prolonged multi-county outbreaks responsible for hundreds of cases reported in 18 MS. During 2010-2019, in Germany the reporting of outbreaks by 'milk and milk products' increased, even though in the most recent years a reverse trend was observed. This pattern was mainly guided by progressive variations in the number of Campylobacter outbreaks. Reasons explaining the trends for outbreaks for the other types of food are less evident. Note: only food vehicles and countries with statistically significant trends and more than five outbreaks reported per year, on average, are shown. 'Fish and fishery products' include 'crustaceans, shellfish, molluscs and products thereof', as well as 'fish and fish products'. 'Meat and meat products' include bovine meat and products thereof, broiler meat (Gallus gallus) and products thereof, other or mixed red meat and products thereof, other, mixed or unspecified poultry meat and products thereof, pig meat and products thereof, sheep meat and products thereof, turkey meat and products thereof. ' Milk and milk products include cheese, dairy products (other than cheeses), milk. 'Other foods' includes canned food products and other foods, unspecified. In 2019, ECDC and EFSA produced Joint Rapid Outbreak Assessments (ROA). These assessments concerned outbreaks caused by L. monocytogenes, Salmonella Poona and Salmonella Enteritidis. In all these outbreaks, clinical isolates were analysed using whole genome sequencing (WGS) allowing tracing of patients linked to the outbreak (and including them retrospectively) and assessing the extension of the outbreaks. The first outbreak was caused by L. monocytogenes sequence type (ST) 1247, clonal complex (CC) 8, and included 22 notified cases in five EU countries (Denmark, Estonia, Finland, France and Sweden). Cases had occurred between July 2014 and February 2019. Evidence from epidemiological, microbiological, environmental and trace-back investigations identified cold-smoked fished products (cold-smoked salmon and cold-smoked trout), manufactured by an Estonian processing company, as the suspected source of the outbreak. Control measures were taken in the Estonian processing company and only batches that complied the food safety criterion (absence of L. monocytogenes in 25 g in cold-smoked and salted product) were released onto the market (EFSA and ECDC, 2019b). A second prolonged outbreak caused by L. monocytogenes IVb sequence type ST 6 was responsible for 21 cases in two MS (the Netherlands and Belgium). Cases were identified between October 2017 and August 2019. The close genetic similarity of the strains and the temporal distribution suggested that the cases were part of a common-source food-borne outbreak. Following the national investigation, various RTE meat products, all manufactured by a Dutch company, were found to be contaminated with L. monocytogenes showing high similarity with the outbreak strain. The company suspended the activities and a withdrawal and recall of all the RTE meat-based products was implemented as control measure (EFSA and ECDC, 2019d) . Salmonella Poona was the causative agent of a multi-country outbreak with 32 cases (infants and young children) reported by three MS (France, Belgium and Luxembourg) between August 2018 and February 2019. The link between the cases was established by a core genome Multilocus Sequence Typing (cgMLST) cluster analysis. Epidemiological evidence obtained from the children's parents of 30 out of 32 confirmed cases identified various infant formula products based on rice proteins as the potential vehicles of infection. All were manufactured by the same Spanish company and marketed by a French company. Environmental investigation in the manufacturing plant and food testing were carried out without findings of the bacterium. Nevertheless, a precautionary recall and withdrawal of the infant formula products of the same brand implicated in this outbreak was implemented (EFSA and ECDC, 2019a). A multi-country outbreak of S. Enteritidis was responsible for 1,041 confirmed cases and 615 probable cases from 2012 to 2019. The last ROA update, published in February 2020, provided information on cases communicated since November 2018, totalling 248 new cases. Of these, 166 were confirmed cases (including 42 historical-confirmed cases) belonging to four distinct clusters identified by WGS and 82 (including 46 historical probable cases) were categorised as probable cases, belonging to six distinct MLVA profiles. This outbreak peaked between 2016 and 2018. Epidemiological, microbiological and food tracing investigations identified eggs from laying hen farms of a Polish consortium as the potential source of the outbreak. Control measures implemented at the farm and at the distribution level, including depopulation, cleaning and disinfection of the contaminated henhouses, however failed to limit the spread of the infection. The outbreak strain was persistently detected on the farms of the Polish consortium were positive during 2018-2019 for the outbreak strains, suggesting persistent contamination. To identify possible contamination at a higher level in the food chain, the feed supply chain as well as the origin of the animals were investigated up to parent stocks, but no significant information was obtained (EFSA and ECDC, 2020a). In 2019, the reporting of FBOs in the EU did not substantially change from previous years in terms of total outbreaks and illnesses. At country level, a large amount of variability was observed in the epidemiological indicators adopted to describe FBOs such as the reporting rate, the mean outbreak size, the type of outbreaks and their severity. This reflects both epidemiological differences and divergences in the approach and sensitivity of the surveillance of FBOs at the single country level. Overall, in 2019 fatal illnesses (N = 60) due to FBOs increased by 50% compared with 2018. Most deaths were reported in settings such as 'residential institution (nursing home or prison or boarding school) and hospital'. This finding calls for attention to the increased risk of vulnerable populations, including elderly and chronically ill patients to food-borne hazards. Another critical aspect emerging from the data analysis is the occurrence of FBOs outbreaks in schools and kindergartens. In 2019, almost 20% of cases involved in strong-evidence general outbreaks (2,407 cases; 1 in 5) became exposed to contaminated foods in a school or kindergarten. In Hungary, a single outbreak of S. Enteritidis involved 575 individuals who had consumed different types of foodstuffs in these settings. Food-borne outbreaks in schools/kindergartens are frequently reported in the literature and may be large or very large. In 2012, 11,000 people in Germany fell ill with gastroenteritis caused by norovirus, predominantly in schools and childcare facilities (Bernard et al. 2014 ). In 2019, school/kindergarten outbreaks were reported by 11 MS and involved wide range of causative agents. This suggests the need for strengthening the standard of hygiene and the procedures for food manufacturing and preparation, as well as the HACCP plans for such establishments. In 2019, although Salmonella was confirmed as the most identified agent in FBOs in the EU/EFTA and it was responsible for the highest number of hospitalisations, L. monocytogenes caused more than 50% of total outbreak associated deaths (31 deaths; 10 deaths more than in 2018; 29 more than in 2017). This is a critical finding as outbreak-associated cases and hospitalisations caused by L. monocytogenes have continuously increased over the last four years in the EU. A better tracing of patients, especially those affected by severe conditions and invasive listeriosis, resulting from the prompt application of WGS for the characterisation of L. monocytogenes clinical isolates may have contributed to improve the increase. Concern is also represented by the high epidemic potential of L. monocytogenes. In 2019, just after the end of the prolonged multi-country outbreak by frozen vegetables (EFSA and ECDC, 2018a), L. monocytogenes was identified as the causative agent of multiple prolonged multi-country outbreaks and was responsible in Spain of one of the largest outbreak ever recorded in the EU with 207 cases involved, 131 hospitalisations and 3 deaths. These findings need to be carefully considered, with particular attention to the large variety of food that may support the growth of L. monocytogenes and that have been implicated in community outbreaks in recent years (i.e. meat and meat products, cold meat, fish, cheese, vegetables). As in recent years, most of the outbreaks in 2019 took place in a 'domestic settings' (N = 296; 41.3%). This proportion is probably underestimated since this setting is usually associated with 'household' outbreak but not all MS report 'household' outbreaks to EFSA. Such a finding reinforces the importance to continuing deliver recommendations to consumers on the correct mode of food preparation, storage and consumption. Among public settings, 'restaurant, caf e, pub, bar, hotel catering service' and 'school or kindergarten' are the places associated with the highest number of FBOs and cases, respectively. The range of foodstuffs that have been identified or suspected in food-borne outbreaks reported in 2019 closely reflects the known epidemiology of the implicated causative agents. The consumption of foods of animal origin was associated with most of the reported outbreaks, especially those caused by Salmonella and Campylobacter. Salmonella outbreaks caused by the consumption of eggs, meat or meat products accounted for more than twice as many of the outbreaks as those associated with all other items altogether, suggesting the need for continuing implementation of control actions at the primary production level. Notwithstanding, most of the cases involved in Salmonella outbreaks became infected after the ingestion of mixed foods, as well as of other highly manipulated foods such as bakery products, sweets and chocolate, revealing that errors during the preparation and/or the preservation of foods occur frequently. Outbreaks linked to 'crustaceans, shellfish, molluscs and products thereof' increased markedly in 2019 due to outbreaks caused by 'norovirus and other Calicivirus' mainly reported from France. Highly manipulated foodstuffs such as mixed foods and buffet meals were also frequently implicated in norovirus outbreaks, in which the contamination of foodstuff by food handlers is very likely. 'Mixed food' is a miscellaneous group of foodstuffs which includes a large variety of multi-element and multi-origin ingredients. This heterogeneity makes it difficult to identify the primary source of contamination and the mechanisms leading to the contamination and/or cross-contamination of the final preparation. In 2019, mixed foods were responsible for the highest number of illnesses in strongevidence outbreaks in the EU. Incidents leading to contamination of mixed foods (e.g. unhygienic manipulation of the ingredients by food handler, cross-contamination, temperature abuse) frequently originate during the final preparation of the dishes close to the consumer in either restaurants, public settings or in the home. Preventive strategies in domestic settings require the engagement of citizens and media to deliver recommendations and promote education campaigns. In 2019, vegetables were associated with the widest range of causative agents of FBOs including bacteria (Escherichia coli, L. monocytogenes, Salmonella, Yersinia), bacterial toxins (B. cereus, C. botulinum), histamine, parasite (Cryptosporidium), virus (hepatitis A, norovirus and other Calicivirus), in spite of the relatively small number of outbreaks reported. Salmonella outbreaks linked to vegetables and fruits should not be overlooked since these events were frequently responsible for a large number of illnesses. In this type of outbreak, mechanisms leading to food contamination are complex and may originate at various levels of the food chain from the growers in the pre-harvest level up to the processing and retail chain. Water, especially irrigation water at the pre-harvest level, or wash-water, as well as wildlife, may play a critical role as a potential source of contamination by foodborne pathogenic agents especially bacteria and viruses. Others food/agent pairs that may merit special attention include enterotoxigenic E. coli (ETEC) and enteropathogenic E. coli (EPEC) in vegetables and 'other foods', respectively. Although the number of outbreaks caused by ETEC and EPEC reported in 2019 was limited, it is worth noting that the number of outbreaks implicating these agents was the highest since 2010. In conclusion, to correctly understand the pattern of food vehicles and sources implicated in FBOs, it is important to appreciate that annual variations in the incidence of food-borne illness depend not only on changes in the prevalence of foods contamination at the consumer level but also on the variations in the type of food being consumed and the consumption habits. Some food preparations, or novel foods or modes of consumption (e.g. delivered-food, take-away) may become progressively popular over the years leading to important changes in the pattern of exposure of consumers to foodborne hazards. Demographic changes and increased susceptibility of vulnerable populations (e.g. the elderly, patients with chronic or immunosuppressive conditions, long-term proton pump inhibitor users) should also be considered. Climate change may also play a part in increased exposure of foods to contamination, eating habits and multiplication of some bacterial pathogens in foodstuffs. Related projects and internet sources • Overall, 907 'ready-to-eat' food sampling units results were reported by four MS and 76 (8.4%) Yersinia enterocolitica-positive units were detected: 75 from 'meat and meat products' (8.3% positives) and one from 'other processed food products and prepared dishes' (one positive sample out of two tested). The positive meat and meat product samples were almost all (71) from mixed meat from bovine animals and pigs and a few (4) were from mixed meat of other animals. For 'non-ready-to-eat' food 5 MS reported 1,191 sampling unit results and 'meat and meat products' and 'milk and milk products' were the contaminated food categories, for 2019. Four MS reported on results of fresh meat categories and most positive samples reported were from pig meat (3.3% positives). Surveillance and monitoring of Yersinia in the EU An overview of the national surveillance systems for human yersiniosis in 2019 is available at https://www.ecdc.europa.eu/en/publications-data/yersiniosis-annual-epidemiological-report-2019 Although the reporting of Yersinia occurrence or prevalence in food and animals is not mandatory, MS can report monitoring data on Yersinia to the European Commission in accordance with the Zoonoses Directive 2003/99/EC. The Directive specifies that, in addition to the number of zoonoses and zoonotic agents, for which monitoring is mandatory, zoonoses such as yersiniosis and agents thereof shall also be monitored when the epidemiological situation so warrants. At present, there is no harmonised surveillance of Yersinia in the EU for food or animals and Yersinia food and animal monitoring data submitted by the MS to EFSA are collected without harmonised design. These data allow for descriptive summaries at the EU level to be made but they preclude trend analyses and trend watching at the EU level (Table 1) . A scientific report of EFSA suggested technical specifications for the harmonised monitoring and reporting of Y. enterocolitica in slaughter pigs in the EU (EFSA, 2009b) . The reported occurrence of Yersinia in major food categories for the year 2019 and for the fouryear period 2015-2018 was descriptively summarised making a distinction between RTE and non-RTE food. Data sets were extracted with 'objective sampling' being specified as sampler strategy, which means that the reporting MS collected the samples according a planned strategy based on the selection of a random sample, which is statistically representative of the population to be analysed. Biotype and serotype of Y. enterocolitica were rarely reported in 2019. Due to the relevance of certain pathoserotypes in the epidemiology of Y. enterocolitica, the access of typing information would be extremely important for a correct assessment of the public health significance and pathogenicity of Y. enterocolitica for humans. The reporting of food-borne yersiniosis disease outbreaks in humans is mandatory according to the Zoonoses Directive 2003/99/EC. When the UK data were collected the UK was an EU MS but as of 31 January 2020 it has become a third country. Summary of the submitted data 1.3 The human data are available at: https://www.ecdc.europa.eu/en/publications-data/yersiniosisannual-epidemiological-report-2019. Seven MS reported 15 yersiniosis food-borne outbreaks for the year 2019 (Denmark (1), Finland (2), France (3), Germany (4), Lithuania (1), Poland (2) and Sweden (2)), causing 149 illnesses, 14 hospitalisations and no deaths. These numbers were similar to recent years. Y. enterocolitica was identified as the causative agent in all these outbreaks but one. Three of these outbreaks were reported with strong-evidence, by Denmark and Sweden, caused by 'vegetables and juices and other products thereof' and by Finland, caused by 'buffet meals' (Table 58) . Interestingly, the two former strong-evidence outbreaks, caused by Y. enterocolitica biotype 4, were part of the same single multicountry outbreak linked to the consumption of food imported in both the Swedish and Danish markets. The food categories most reported to cause strong-evidence yersiniosis food-borne outbreaks during 2010-2019 were 'pig meat and products thereof' and 'vegetables and juices and other products thereof' (three each). Further details and statistics on the yersiniosis food-borne outbreaks for 2019 are in the food-borne outbreaks chapter. ' (99.3%) , whereas during 2015-2018, 14.5% of the sampling units were from 'milk and milk products' and 75.8% from 'meat and meat products '. In total 76 RTE food samples were found to be positive for Yersinia enterocolitica in 2019: 75 from 'meat and meat products' and one from 'other processed food products and prepared dishes'. The positive meat and meat product samples were almost all (71) from mixed meat from bovine animals and pigs and a few (4) were from mixed meat of other animals. During 2015-2018, six Yersinia enterocolitica-positive sampling units were reported for RTE food from 'meat and meat products' (five) and from salads (one). All five positive meat samples were from mixed meat of other animals. Monitoring data considered were collected according an 'objective' sampling strategy. Also considering that only few MS reported sampling results and that only a few results were reported for food categories other than meat and meat products, the finding of Yersinia enterocolitica-contaminated RTE food is of concern because it poses a direct risk to the consumer. Results reported by five MS for non-RTE food show that 'meat and Meat products' and 'milk and milk products' were the contaminated food categories, for 2019 and during 2015-2018, during which also a contaminated 'other processed food products and prepared dishes' sample was reported. Four (c): Badgerswild, birdswild, bison -zoo animals, camels -zoo animals, Cantabrian chamoiswild, cats, cats -pet animals, chinchillas -pet animal, deer, deerwild, deer -wild -fallow deer, deer -wild -roe deer, dogs -pet animals, ferretswild, foxeswild, guinea pigs -pet animals, hares, hareswild, hedgehogswild, martenwild, matrix, monkeys -zoo animal, mouflonswild, other animals -exotic pet animals, otterwild, parrots -pet animals, pigeons, rabbits -pet animals, raccoons, ratswild, rodentswild, squirrels, squirrelswild, starlings, Steinbockwild, turtleswild, water buffalos, wild boarswild, wolveswild, zoo animals. No EU Regulation exists with relation to the surveillance and monitoring of Toxoplasma gondii in animals. Therefore, the available and reported information is strictly determined by national legislation and whether the countries have a mandatory reporting system after the detection of Toxoplasma gondii. The main animal species tested are small ruminants (goats and sheep), cattle, pigs and pet animals (cats and dogs) using samples from aborted animals (ruminants) or clinically suspected animals. Mainly blood samples but also samples from tissue and organs are analysed with either indirect methods to detect antibodies (ELISA, LAT, complement fixation test (CFT) and immunofluorescence assay (IFA)) or direct methods (PCR and immunohistochemistry (IHC)). As the surveillance of Toxoplasma in animals is not harmonised, data on Toxoplasma only allow descriptive summaries to be made at the EU level (Table 1) . This is because the results submitted by different countries and from different regions within a country are mostly not directly comparable due to differences in sampling strategy, testing methods, as well as different sampling schemes. Both age of animals and production systems at farm level may influence the occurrence of Toxoplasma. The reporting of food-borne toxoplasmosis disease outbreaks in humans is mandatory according the Zoonoses Directive 2003/99/EC. When the UK data were collected the UK was an EU MS but as of 31 January 2020 it has become a third country. Summary of submitted data 2.3.1. Humans The human data are available at https://www.ecdc.europa.eu/en/publications-data/congenitaltoxoplasmosis-annual-epidemiological-report-2018. No food-borne disease outbreak due to Toxoplasma was reported for 2019 in the EU and no single such food-borne outbreak has been reported to EFSA since the start of the food-borne outbreaks reporting, in 2004. Available information discussed in the EFSA Scientific Opinion of food-borne parasites (EFSA BIOHAZ Panel, 2018b) suggests that food-borne transmission accounts for 40-60% of the T. gondii infections. Food-borne transmission of Toxoplasma gondii is possible via a range of routes, including consumption of undercooked meat or, to a lesser extent, unpasteurised milk, from an infected animal or via contamination with feline faeces. Although meat is considered to be the more usual source of food-borne infection in Europe, based on risk factor studies, the exact contribution of different foodborne routes is still a major research question. One MS, Italy, submitted monitoring results for Toxoplasma gondii in food, like the previous two years. In total, 386 samples were reported from non-RTE fish, meat products from pig, raw molluscan shellfish and from (RTE) honey and potable water. 22 Thirty-nine samples were positive (10.1%) and were from fish (nine), meat products from pig (25) and raw molluscan shellfish (five). Table 61 summarises statistics on Toxoplasma spp. occurrence in major animal species during 2015-2019. Animal data of interest reported were classified into the major categories and aggregated by year to obtain an annual overview of the volume of data submitted. 22 Additional information from Italian reference centre for Toxoplasma spp.: 'Drinkable water samples were collected from municipal water supplies to assess the potential remaining contamination with oocysts shed by infected cats after the water undergoes treatment. The same applied to RTE honey samples tested to rule out any potential faecal contamination during extraction and processing. ' Thirteen MS (Austria, Finland, Germany, Greece, Hungary, Ireland, Italy, Latvia, the Netherlands, Romania, Slovakia, Spain and the United Kingdom) and two non-MS (Norway and Switzerland) provided monitoring data on Toxoplasma in livestock (small ruminants, cattle, solipeds and pigs). In small ruminants (sheep and goats), 12 MS (Austria, Finland, Germany, Greece, Hungary, Ireland, Italy, Latvia, the Netherlands, Slovakia, Spain and the United Kingdom) and two non-MS (Norway, Switzerland) reported data. In total, 12,167 animals were tested and 1,648 were found to be positive (13.5%). In cattle, six MS (Austria, Germany, Ireland, Italy, Latvia and the United Kingdom) reported data on Toxoplasma-specific antibodies. At animal level, about 9.2% tested seropositive. From pigs, four MS (Austria, Germany, Italy and Slovakia) reported monitoring data: in total 1,108 animals were tested and 130 (11.7%) were detected as positive. In pet animals (cats and dogs), nine MS (Austria, Finland, Germany, Hungary, Italy, Latvia and the United Kingdom) and one non-MS (Switzerland) tested in total 3,169 animals (1,798 cats and 1,371 dogs) of which 323 were positive (10.2%) and obtained mainly from clinical investigations. Five MS (Austria, Finland, Germany, Italy and Slovakia) and one non-MS (Switzerland) reported on testing for Toxoplasma in wildlife. In total, 833 animals (mainly from Italy) were tested and 164 were positive (19.7%). The 2019 monitoring data reported by MS from animals show that Toxoplasma is present in most livestock species across the EU. The limitations of these surveillance data preclude any trend watching or trend analysis of prevalence in animals. The current surveillance system of Toxoplasma in animals of EU is strongly affected by several important limitations: (i) small amount of tested animals; (ii) the use of different indirect and direct detection methods, which, in most cases have been not validated by an independent body; (iii) unknown age of tested animals; and (iv) no information on type of breeding. Furthermore, there is no relationship between the presence of anti-Toxoplasma antibodies and infecting parasites in cattle and horses (Aroussi et al., 2015; Opsteegh et al., 2016) . For pigs, poultry and small ruminants, serological methods could be useful for the detection of high-risk animals/herds but not as an indicator of infection in individual animals, as the concordance between direct and indirect methods was estimated as low to moderate. All these limitations result in the lack of scientific value of data provided by MS and consequently of their use by the European Commission, MS and stakeholders. The data are mostly not directly comparable across MS. Table 62 below summarises rabies EU-level statistics in humans and in wild and domestic animals. For animals, the total number of samples taken from foxes, raccoon dogs, raccoon, dogs and bats, as well as the number of MS from which these samples originated, are shown. A significant reduction has been observed in the number of reported samples from foxes, the main reservoir, over the last 5 years at EU level. In 2019, the numbers of reported sampled foxes was halved compared with 2015. The human data are available at: https://www.ecdc.europa.eu/sites/default/files/documents/rabiesannual-epidemiological-report-2019.pdf Wildlife rabies In 2019, in total, 23,141 foxes (Vulpes vulpes) were investigated by 19 MS. More than half of the tested samples (67.3%) were taken by three MS: Romania, Poland and Czechia. In total, three cases of rabies in foxes were detected in the EU: one case in Poland and two in Romania. The geographical distribution and number of cases in foxes, as well as a choropleth map of the total number of foxes sampled per MS are shown in Figure 72 . Four non-EU countries (Norway, Republic of North Macedonia, Serbia, Switzerland) reported 1,274 tested foxes. None of these countries reported positive cases for rabies. (a): The number of tested animals includes national statistics submitted by MS and not regional data that were submitted without a national summary. In 2019, 1,542 raccoon dogs and six raccoons were reported and tested for rabies by nine MS (Austria, Czechia, Estonia, Finland, Latvia, Lithuania, Poland, Slovakia and Spain) . Most of these samples originated from raccoon dogs from three MS (Estonia, Finland and Latvia). All the samples tested were negative for rabies. Fifteen MS reported results for 2,393 terrestrial wild animals other than foxes, raccoon dogs and raccoons. Almost half of these samples (45.2%) were reported by Bulgaria, with 1,077 of these originating from jackals. Other most tested species were badgers (452), martens (390), wolves (98) and roe deer (81). Other species tested included bears, deer, red deer, ferrets, hares, hedgehogs, lynx, mice, minks, moles, moose, otters, polecats, rats, rodents, squirrels, wild boars and wolverine. In 2019, one rabies positive result was reported in Romania in a wild boar. In 2019, 18 MS and two non-MS reported surveillance data on bats. In total, 2,069 bats were investigated in EU ( Figure 73 ). Out of these, 39 samples tested positive in six MS: France (nine EBLV-1), Germany (eight unspecified virus species), the Netherlands (five EBLV-1), Poland (10 EBLV-1), Spain (three EBLV-1) and the United Kingdom (three EBLV-1 and one EBLV-2). Two non-MS, Norway and Switzerland, tested five and 18 bats, respectively, with all samples being negative. In conclusion, the results on bat rabies presented here (N = 39 positive cases) are in line with the previous years' findings and confirm bats to be a reservoir for rabies, reaffirming in this way the public recommendation to handle bats with utmost caution, if at all. The public health hazard of bat rabies in Europe ought to not be underestimated. Romania reported one case of rabies (wild strain) in a cow in 2019 and was the only MS reporting a case in a domestic animal, like in 2018. In total, 404 samples from farmed animals were tested by 17 MS (reports included mainly cattle, small ruminants and domestic solipeds). The number of samples taken from domestic farmed animals in 2019 was lower than the number taken in the last four years. No case of rabies was reported in 2019 in dogs and cats. Twenty-two MS reported in total more than 4,000 tested samples for dogs (1, 901) and cats (2, 389) . The numbers of samples reported for both species slightly decreased compared with 2018. These results indicate that, as in the previous years, rabies still occurs in domestic animals in Eastern Europe, indicating the persistence of an active wildlife reservoir there as evidenced by the above-mentioned results on cases of rabies in foxes (Poland and Romania) and wild boar (Romania). Overall, the results from the rabies surveillance carried out by MS in 2019 highlight once more the very low number of positives rabies cases detected in non-flying terrestrial animals in Europe (N = 5). Nonetheless, and as described in the report of the first meeting 25 of the Standing Group of Experts on Rabies in Europe, in 2019, under the umbrella of The Global Framework for the Progressive Control of Transboundary Animal Diseases (GF-TADs) cases can still appear in areas not far from the EU borders. Those experts also raised a concern in terms of rabies surveillance for in certain areas and strongly recommended an improvement of the surveillance in those areas. Appropriate surveillance is of paramount importance, particularly for MS countries close to rabies elimination (Cliquet et al., 2010 ; Table 63 summarises EU-level statistics on Q fever in humans and in major animal species, respectively, during 2015-2019. Animal data of interest were classified into the major categories and aggregated by year to obtain an annual overview of the volume of data submitted. When the UK data were collected the UK was an EU MS, but as of 31 January 2020, it has become a third country. In 2019, the number of Q fever cases in humans who acquired the infection in the EU increased compared with 2018 and is the highest in the past five years. This might partly be due to the decreasing proportion of cases with unknown travel status or unknown country of infection. In 2019, compared with the year 2018, the number of samples from animals submitted by EU MS from sheep and goats and from cattle decreased by 24.4% and by 41.1%, respectively. Since 2015, the number of submitted samples from animals has been decreasing, except for the year 2018 when samples collected increased. The overall proportions of positive samples ranged from 9.2% to 11.6% for sheep and goats and from 6.0% to 11.0% in cattle, during 2015-2019. Overall, 950 confirmed cases of Q fever were reported by 22 EU MS, eight cases were reported by Norway and 103 cases were reported by Switzerland (Table 64) . , Spain was the country that reported most confirmed cases (N = 332), followed by France and Germany (155 and 148 cases, respectively). The number of confirmed Q fever cases in 2019 was higher than in 2018. The EU notification rate was 0.19 per 100,000 population, which is higher than in 2018 but comparable with the notification rates from 2015 to 2017. For 2019, the highest notification rate (0.71 cases per 100,000 population) was observed in Spain, followed by Romania (0.56), Bulgaria (0.51) and Hungary (0.48). Six countries (Denmark, Estonia, Iceland, Latvia, Lithuania and Luxembourg) reported no human cases. A large majority (85.2%) of the Q fever cases were acquired in the EU (Table 64 ). In total, 14 travel-associated cases were reported in people who had travelled to Bosnia and Herzegovina, Brazil, Egypt, Guinea-Bissau, Kenya, Mauritius, Morocco, the Philippines, Senegal, Sri Lanka, Switzerland and Turkey. Between 2007 and 2010, the Netherlands experienced a large outbreak with more than 4,000 human cases (Schneeberger et al., 2014) . The number of cases in the Netherlands returned to the pre-outbreak level in 2013 and has remained low since then. Four deaths due to Q fever were reported for 2019 in the EU, all by Spain, resulting in an EU case fatality of 0.63% among the 639 confirmed cases with reported outcome. Poland and Italy tested together 96.9% of the holdings/flocks; Italy, Czechia, Switzerland, Norway and Slovakia accounted for 82.2% of the tested animals. Five MS and two non-MS reported data on animals other than sheep, goats and cattle. In total, 302 animals and 37 holdings/flocks were tested from different domestic and wild animal species (Alpaca 's, Alpine and Cantabrian chamois, antelopes, badgers, bears, bison, cats, deer, dogs, dolphin, dromedaries, foxes, hares, hedgehogs, horses, lamas, martens, mouflons, otter, parrots, pigeons, pigs, Steinbock, water buffalos, wild boars, wolves) . Among all holding/flocks tested, three (with several animal species) out of 20 tested were reported positive by Cyprus (15%). Two dogs were reported testpositive by Italy and one positive alpaca by Switzerland. Animal results were mainly submitted by Italy (n = 161; 27 different animal species), Slovakia (n = 60; hares and zoo animals) and Austria (n = 35; alpacas, Alpine chamois and pigs). Over the last five years (2015-2019), there was no statistically significant (p < 0.01) increase or decrease in confirmed Q fever cases in humans in the EU/EEA. While France and Germany reported most of the confirmed cases until 2016, Spain started to report the highest number of cases annually since 2017. The increase in the number of human cases reported by Spain is most likely explained by a change in their reporting system: from voluntary to mandatory. In 2019, Spain accounted for more than a third of the overall number of cases. Case fatality increased between 2016 and 2018 from 0.39% to 1.92% but decreased to 0.63% in 2019. The results obtained from animalsmainly from small ruminants and cattledo not allow following or analysing trends for Q fever at the EU level, because the results submitted by MS are mostly not directly comparable due to differences in sampling strategy, testing methods, coverage of the monitoring and sensitivity of the surveillance for C. burnetii. The regional variability within Europe highlights the importance of understanding risk factors that may operate at a local scale and may be subtle (Georgiev et al., 2013 • There was no significant increase or decrease over the last 5 years (2015-2019) for WNV infections in humans in the EU/EEA. • For the year 2019, 16 MS submitted WNV monitoring and surveillance data from birds and equids to EFSA. Italy and Spain submitted, respectively, 69.4% and 14.7% of these data for birds, while for equids it was Spain and Greece that, respectively, submitted 30.4% and 23.1% of the data. • Eight MS reported 153 WNV outbreaks in birds (53) and equids (100) to ADNS. Germany and Greece reported, respectively, 52 and 1 outbreaks in birds. Germany and Greece reported the highest number of outbreaks among MS in equids, accounting, respectively, for 32% and 21% of the total number of outbreaks. • ADNS outbreaks data and surveillance data submitted to EFSA indicated WNV circulation during 2019 in countries in Central and Eastern Europe and in the Mediterranean basin. WNV infections of humans and equids now regularly occur in those countries. West Nile fever, also known as 'West Nile virus disease', is an arboviral disease transmitted in natural conditions to humans and animals via infected mosquito bites (Diptera; Culicidae). The transmission period is typically between early or mid-summer until the end of October when mosquitoes (predominantly Culex spp.) are most active and more abundant. The mosquitoes, in which the WNV replicates, acquire infection by feeding on viraemic birds. WNV is maintained in a birdmosquito cycle, with birds acting as amplifying hosts. Apart from in humans, the virus can also emerge in equine species, which, as humans, are accidental hosts and which cannot in turn transmit the virus to the vectors. MS with areas that are typically prone to harbouring mosquitoes may be affected with both human cases and outbreaks in animals. Human WNV infections data are collected through two complementary data collection processes. During the period of high mosquito abundance and activity (June-November), the MS report human infections timely to TESSy at ECDC (ECDC, 2020). Complementary to this real-time data collection, an annual data collection is carried out. Countries who did not detect any infections during the year are asked to report 'zero cases'; all other countries are encouraged to report complementary data on detected infections if considered relevant. For 2019, 27 EU MS, Iceland and Norway reported information on WNV infections in humans to TESSy. The EU case definition was used by 26 countries. Germany did not specify which case definition was used and France and the United Kingdom used an alternative case definition. All reporting countries had a comprehensive surveillance system, except for Germany which did not specify the type of surveillance system. Reporting is compulsory in 26 EU/EEA countries, voluntary in France and the United Kingdom and not specified for Germany. Surveillance is passive, except in Czechia, Greece, Portugal, Slovakia and the United Kingdom. All countries have a national coverage of reporting and case-based reporting. According to Directive 2003/99/EC 27 , WNV infections in animals are not included in the zoonoses listed in Annex I, Part A of the Directive for which monitoring and surveillance activities as well as reporting are mandatory. Nevertheless, WNV is listed in Annex I, Part B (viruses transmitted by arthropods) to be monitored when to the epidemiological situation in a MS so warrants, in compliance 27 with Article 4.1 of the same Directive. EFSA so is being provided with annual WNV monitoring data by MS that regularly or recently experienced WNV outbreaks (in animals or humans), or that are at high risk and having so put in place a surveillance system for early detection of the disease in animals. In addition to EU MS, Switzerland and Serbia submit reports on surveillance and monitoring activities in animals to EFSA. The heterogeneity in study designs and the variety of analytical methods used, make the reported WNV data from different countries not directly comparable. These data allow descriptive summaries at the EU level to be made (Tables 65 and 67) . Proposals for harmonised schemes for monitoring and reporting of WNV in animals can be found in an External Scientific Report submitted to EFSA (Mannelli et al., 2012) . Nonetheless, according to Council Directive 82/894/EEC 28 , it is mandatory for MS to notify outbreaks 29 of WNF equine encephalomyelitis to the EU ADNS. 12 Every week, each officially confirmed outbreak should be notified by the Veterinary Authority of the MS where it occurred, to all other countries that are connected to the ADNS application. Report summaries and annual reports on disease outbreaks are available online on the ADNS website. 4 Moreover, animal WNF outbreak data reported to the World Organisation for Animal Health (OIE) are publicly available on the World Animal Health Information Database (WAHIS interface). Overview of key statistics, EU, 2015-2019 Table 65 summarises EU-level WNV infection statistics on humans and on birds and equids, during 2015-2019. More detailed descriptions of these statistics are in the results section of this chapter. There was no statistically significant (p < 0.01) increase or decrease over the last 5 years (2015-2019) for WNV infections in the EU/EEA (Figure 75) . At the country level, Greece reported a significantly (p < 0.01) increasing trend in the past 5 years (2015) (2016) (2017) (2018) (2019) . In 2018, a large number of human WNV infections were reported in the EU/EEA, far exceeding the annual totals for the previous years. The notification rate for locally acquired WNV infections in the EU/EEA was almost eight times higher in 2018 compared with 2017. Almost all countries in 2018 reported their highest number of cases ever. During the WNV transmission season, weekly epidemiological WNV updates including the geographical distribution of human cases in the EU/EEA and EU neighbouring countries are published on the ECDC website (ECDC, 2020). These updates include a summary of the WNV transmission season, data from the ECDC Surveillance Atlas and three maps: (1) human WNV infections; (2) WNV outbreaks among equids and/or birds; and (3) combined distribution of WNV infections among humans and outbreaks among equids and/or birds. The latter map is in Figure 76 . In relation to West Nile fever (WNF) in animals, there exist two sources of information mainly used for this report: the data of the annual surveillance and monitoring activities submitted to EFSA and the data of the outbreaks notified to the ADNS. 30 Table 66 includes for each MS the jointly displayed data from both data sources. In some cases, their comparison may be subjected to some discrepancies and the following points should be taken under consideration for the interpretation: (i) the data on the surveillance and monitoring activities, submitted to the EFSA, include all the units that have been analysed with different types of methods; (ii) the data reported in ADNS include only the outbreaks for which the disease has been confirmed clinically and/or laboratory, either by the detection of IgMspecific antibodies (indicator of recent infection with WNV) or by the detection of RNA particles via PCR-based methods, as a result of the surveillance and monitoring activities and the investigation of suspected cases; (iii) an outbreak can refer to more than one affected animal if they constitute a unique epidemiological unit or/and are identified at the same location; (iv) the positive results of the surveillance data refer to the positive results of ELISA to detect IgM antibodies, to the seroneutralisation and the positive results of PCR methods to detect the virus genome; and (v) some countries have not submitted data either to the ADNS or to EFSA. In 2019, according to the annual surveillance and monitoring data reported by 13 MS to EFSA, a total number of 14,922 samples from birds was tested for WNV, mostly wild birds but also fowl kept on farms (Tables 65 and 67 ). Two non-EU Countries (Serbia and Switzerland) also reported to EFSA the results of 590 samples of birds tested for WNV (Table 67 ). The analytical methods used to underpin Source: TESSy and ADNS. positive results in birds were mainly molecular methods based on PCR that detects the nucleic acid of WNV. In some cases, ELISA was the method used to detect immunoglobins IgG (Denmark and Romania) or IgM (Cyprus). Italy, in addition to PCR methods, reported positive results by seroneutralisation method. Bird species to be found positive were: doves, ducks, eagles, finches, flamingos, fowls (Gallus gallus), geese, gulls, hawks, herons, owls, pelicans, penguin, pheasants, pigeons, plovers, tits, birds of the family of Corvidae (e.g. crows, magpies, jays) and birds of the family of Psittacidae (e.g. parrots). Furthermore, 14 MS reported to EFSA the results of 5,563 samples from equids, almost all from horses (Tables 65 and 67 ). Two non-MS (Serbia and Switzerland) also reported to EFSA the results of 2,503 samples of equids tested for WNV. The analytical methods used to underpin positive results were mainly the IgM-capture ELISA and the real-time PCR. Czechia reported positivity to the seroneutralisation test. Positive animals to serological test were unvaccinated or had an unknown vaccination status. During 2019, 153 WNF outbreaks in animals, both in equids (100) and birds (53) were notified to the ADNS by the Veterinary Authorities of eight MS (Table 67 ). The geographical distribution of these outbreaks is visualised in Figure 77 . Based on the date of confirmation notified in the ADNS, the number of WNF outbreaks (all species) per month, aggregated for all the EU MS, has been calculated for each year and presented in Figure 78 . According the graphs in Figures 78 and 79 , the occurrence of WNF in animals is seasonal with the outbreaks mainly confirmed during the summer and autumn (July-October), while some sporadic outbreaks are confirmed during winter months (November, December, January). Out of the total number of the outbreaks in EU MS since 2013, 39% were confirmed in September, 23% in August, 26% in October, 6% in November and 4% in July. September looks like the month with the highest percentages of outbreaks for most of the MS: Germany (56.2%), France (50%), Hungary (43.97%), Italy (37.5%) and even in countries with very few outbreaks such as Austria (six out of eight), Croatia and Slovenia. In Spain and Portugal, respectively, 65% and 40% of the total amount of the outbreaks occurred in October. In Greece, it looks like most WNF outbreaks occurred earlier, with 25% of the outbreaks confirmed in July, 30.8% in August and 27.9% in September. More information on the evaluation of the status as regards WNV and trends are in the national zoonoses reports submitted in accordance with Directive 2003/99/EC, which are published on the EFSA website (available online http://www.efsa.europa.eu/en/biological-hazards-data/reports) together with the EU One Health zoonoses report. Specific information on WNV in some countries was extracted by the above-mentioned reports and are provided here below: Czechia . . . In total every year 783 horses are tested for antibodies against WNV by cELISA test with WNV antigen in the whole territory of Czechia. Virus neutralisation test is used to confirm the presence of antibodies against WNV. . . . In 2019, a total of 782 horses from entire Czechia were tested for the presence of antibodies against WNV. Samples that reacted positively in cELISA with WNV antigen were tested by virus neutralisation assay (VNT) for the presence of antibodies to WNV; 22 samples responded positively to VNT. . . . . . . In the last epidemic season, 51 positive mosquitoes pools were collected between July and September in Lombardy, Emilia Romagna, Veneto, Friuli Venezia Giulia and Piemonte regions. Genetic analyses of WNV strain confirmed the circulation of Lineage 2. . . . . . . During 2019, active surveillance activities were foreseen in animals owned by humans confirmed with West Nile fever. In 14 counties, samples were taken in 29 backyards from birds (hens, geese and ducks) and Equidae (horses, donkeys). Two ELISA tests were used (IgG ELISA for birds and IgM ELISA for Equidae). Animals from 13 backyards were positive for West Nile virus antibodies. . . . . . . West Nile Fever virus in horses was never isolated. Presence of virus was detected only serologically. In 2018 was one horse serologically positive for WNV and none in 2019. According Plan of veterinary prevention and protection of state territory monitoring of the epidemiological situation is carried out through monitoring of West Nile virus fever antibodies in horses. Detection of post-infection antibodies are performed within targeted intravital diagnostics in horses and the targeted intravital diagnosis of suspected CNS disease. In horse holdings the breeding stallions prior to and after the completion of a mating season, mares prior to mating, sport and production horses used for the breeding and animals with suspicion of the disease of CNS are tested. Diagnostic/analytical methods used: ELISA IgM, ELISA IgG, Real-time RT-PCR. . . . . . . In 2019, 26 horses were tested negative for WNV. In general horses should only be examined for WNV if they show neurological symptoms of unknown origin and if they were not vaccinated. In 2019 15 birds were tested for WNV using RT-qPCR at the National Reference Center for Poultry and Rabbit Diseases, University of Zurich. 62 FTA-cards which were placed in mosquito traps in the canton Ticino and in August and September 2019 were screened for Flavivirus and Alphavirus, all negative for WNV. The FTA-cards contain a sugar solution. If consumed by the mosquitoes, the saliva of the mosquitoes, which might contain virus, gets into the FTA-cards. The saliva contained virus is inactivated and fixed on the FTA-card. Up to date there were no autochthonous cases of WNF reported. However, it cannot be excluded that WNV is circulating in Switzerland, especially in wild birds and mosquito populations. . . . A large number of human WNV infections had been reported in the EU/EEA for 2018 (n = 1,615), exceeding, by far, the total number from the previous 4 years. , reported human WNV infections decreased again in most countries (n = 443), although in Greece the number remained at a relatively high level (n = 227). For 2019, Cyprus reported 23 locally acquired human WNV infections, after previously having only reported one human WNV infection in 2016 and 2018, each. During 2019 Slovakia and Germany reported the first mosquito-borne locally acquired human WNV infections. This was not unexpected as the presence of WNV among birds, equids and/or mosquitoes has been previously documented in those countries. All other human infections were reported in countries with known persistent transmission season in previous years. The case fatality among all locally acquired WNV infections, the case fatality among cases with WNND and the proportion of cases with WNND was slightly higher in 2019 compared with 2018. In 2019, 16 MS have submitted to the EFSA data on surveillance activities on animals, while 8 MS notified outbreaks in animals to the ADNS. As during previous years, the 2019 data indicate WNV circulation in Central and Eastern Europe and in the Mediterranean basin: Austria, Czechia, France, Greece, Hungary, Italy, Portugal and Spain. Germany reported outbreaks of WNV in animals to the ADNS, as it did for the first time in 2018. During the previous years, it identified seropositivity during the surveillance activities. These reported observations are consistent with the OIE's conclusion that the occurrence of WNF in humans and animals along with bird and mosquito surveillance for WNV activity demonstrates that the virus range has dramatically expanded including North, Central and South America as well as Europe and countries facing the Mediterranean Basin (OIE, 2018) . The risk of WNV transmission is complex and multifactorial; it concerns the virus, the vectors, the animal reservoirs, the environmental conditions, the human behaviour and the density of human and animal populations. Preventing or reducing mosquito-borne WNV transmission depends on successfully controlling the vector's abundance or interruption of human-vector contact. Human, animal and entomological WNF surveillance is crucial to allow the early detection of WNV infections in humans and take timely preventive measures. In horses, the development of WNV-associated diseases is preventable with proper vaccination and protection against mosquito bites. It is important to take into consideration that the absence of cases and outbreaks does not imply the absence of the virus in the environment. unknown food, four illnesses, no hospitalisations and no deaths). In EU, during 2005-2017, there were three food-borne outbreaks of tularaemia reported in EU, by Croatia (year 2015 , five illnesses, three hospitalisations and no deaths), Germany (year 2016, six illnesses, two hospitalisations and no deaths) and France (year 2012, three illnesses, no hospitalisations and no deaths). France reported that outbreak with strong-evidence as regards the incriminated food, which was 'other, mixed or unspecified poultry meat and products thereof. In 2019, two MS, Austria and Sweden, reported data on the occurrence of Francisella tularensis in hares (natural habitat) and overall, 67 out of the 211 were positive (31.7%). Sweden also reported data from 24 tested muskrats with eight positives. The Swedish reports on hares were regarding 128 European brown hares and 48 mountain hares, of which 27 European brown hares and 31 mountain hares tested positive for F. tularensis subsp. holarctica. 31 During the last two decades, in Sweden, the epidemiology of tularaemia has changed and the number of reported cases in animals, mainly European brown hares, infected south of the previous endemic region, has increased. In animals, outbreaks of tularaemia have in some countries been associated with rises in rodent and hare populations, but this has not been confirmed in Sweden. The epidemiological role of the hare as a possible carrier of F. tularensis remains unclear. A recent study from Hestvik et al. (2019) found that all predator and scavenger species included in the study (brown bear (Ursus arctos), Eurasian lynx (Lynx lynx), raccoon dog (Nyctereutes procyonoides), red fox (Vulpes vulpes), wild boar (Sus scrofa), wolf (Canis lupus) and wolverine (Gulo gulo)) may serve as sentinels for tularaemia in Sweden as they found seropositive animals in all the species studied. At the same time, the role of these species as reservoir stays unclear. Sweden has reported cases of tularaemia in humans and animals since 1931. Ever since the first Swedish tularaemia case was reported, endemic areas have been identified in northern and central Sweden. Switzerland also reported samples taken from wild species (hares, beavers, squirrels, hedgehogs, mice, deer, foxes and polecats) kept in zoos or from their natural habitat. The occurrence of Francisella tularensis in the tested hares was 87.1%. One pet cat was also found positive. None of the other tested animal species (N = 10) in Switzerland tested positive. Tularaemia has terrestrial and aquatic ecological cycles with an extensive host range among animals including vertebrates and invertebrates. Lagomorphs of the genus Lepus and small rodents are considered reservoirs, but antibodies against F. tularensis have been detected in other wild animals, such as red fox and wild boar and domestic animals such as cat and dog (Hestvik et al., 2015; Maurin and Gyuranecz, 2016) . As for humans, the animal species susceptible to tularaemia may be infected either through the terrestrial or the aquatic cycle. A study performed in the Netherlands during an outbreak in hares in 2015 to assess potential reservoirs and transmission routes of F. tularensis showed the importance of the environmental surveillance of water and its valuable use to monitor this pathogen (Janse et al., 2015) . Only Austria and Sweden reported data on hares obtained from passive surveillance. These data show that F. tularensis is still present in the wildlife and that hares (genus Lepus) are good indicator animals to monitor the occurrence. Wildlife may continue to play a role in the maintenance of F. tularensis in the ecological cycle and the occurrence of human cases. It is clear that Francisella spp. are widely present in the environment and a wide range of wild animals (such as hares), but also vectors (e.g. ticks as illustrated in the previous chapter) could be used to enforce passive surveillance in EU as they can be sources of infections in humans (WHO, 2007) . Greater efforts are needed to assess the extent of the true animal reservoir population of F. tularensis and to assess the occurrence of this zoonotic pathogen in the EU animal reservoir populations including the environment. Related projects and Internet sources O77 O77 O78 O78 O79 O79 O81 O81 O82 O82 O84 O84 O87 O87 O88 O88 O89 O89 O91 O91 O92 O92 O93 O93 O98 O98 O100 O100 O101 O101 O102 O102 O103 O103 O104 O104 O105 O105 O107 O107 O108 O108 O109 O109 O110 O110 O111 O111 O112 O112 O113 O113 O115 O115 O116 O116 O117 O117 O119 O119 O121 O121 O123 O123 O126 O126 O127 O127 O128 O128 O129 O129 O130 O130 O132 O132 O134 O134 O136 O136 O137 O137 O139 O139 O140 O140 O141 O141 O142 O142 O143 O143 O145 O145 O146 O146 O148 O148 O149 O149 O150 O150 O153 O153 O155 O155 O157 O157 O159 O159 O160 O160 O165 O165 O166 O166 O168 O168 O171 O171 O172 O172 O174 O174 O175 O175 O176 O176 O177 O177 O178 O178 O179 O179 O181 O181 O182 O182 O183 O183 O185 O185 O187 O187 O188 O188 O45 O50 O50 O55 O55 O78 O78 O84 O84 O88 O88 O91 O91 O98 O98 O100 O100 O102 O102 O103 O103 O104 O104 O105 O105 O108 O108 O109 O109 O111 O111 O113 O113 O115 O115 O116 O116 O117 O117 O121 O121 O123 O123 O127 O127 O128 O128 O130 O130 O136 O136 O145 O145 O146 O146 O148 O148 O149 O149 O150 O150 O153 O153 O155 O155 O157 O157 O159 O159 O160 O160 O168 O168 O171 O171 O174 O174 O176 O176 O177 O177 O178 O178 O179 O179 O181 O181 O183 O183 O185 O185 O187 O187 No of samples examined STEC serogroups Proportions of STEC serogroups: red boxes > 1%, orange boxes > 0.1% and ≤ 1%, yellow boxes > 0.0001% and ≤ 0.1% of positive samples. White boxes indicate absence of the serogroup. The food category 'other ruminants' meat' includes meat from deer; 'other meat' includes meat from animals other than ruminants; 'milk and dairy products' include any type of dairy product, cheese and milk other than raw milk; 'raw milk' includes raw milk from different species, but most tested samples were from cows; 'seeds' includes mostly sprouted seeds, but dry seeds are also included. Source: Twenty-two MS. The animal category 'other ruminants' includes deer; 'other animals' comprises pigeons, cats, chinchillas, dogs, ferrets, foxes, Gallus gallus, guinea pigs, hedgehogs, mice, rabbits, rats solipeds, water buffalos, weasels and wild boars. Source: nine MS. N=25) (N=26) (N=24) (N=27) (N=27) (N=28) (N=27) (N=27) (N=27 Repor ng rate Switzerland Strong-Evidence Outbreaks Weak-evidence outbreaks ReporƟng Rate Blue colour for both the trend line and the secondary Y-axis representing the FBO reporting rate was adopted for Latvia, Lithuania and Malta to highlight that the scale was different from the other countries Apart from this, the reduction in the number of S. Enteritidis outbreaks in 2019 is highly likely caused by the non-data reporting of FBO data by Slovakia, which reported 505 FBOs caused by S. Enteritidis in 2018. A drop in the total number of cases (1,199 cases less in 3), the serovar was not reported. The absence of this information introduces uncertainty in the identification of the most important sources of Salmonella at primary production level, given that the food vehicles implicated in Salmonella outbreaks differ importantly by serovar (Section 4.3). In addition, group B and group D Salmonella (Grimont and Weill, 2007), without full serotyping, were responsible for five and seven outbreaks, respectively. Campylobacter Causative agents are differently coloured. The size of each sector is proportional to the number of outbreaks (internal circle) and human cases (external circle) involved in outbreaks Meat and meat products (136) include: Poultry meat (52) Note: Data from 439 outbreaks are included: Austria (12) Other or mixed red meat and products thereof (2), Meat and meat products Food of non-animal origin (19) include: Cereal products and legumes (4), Fruits (and juice) (2), Sweets and chocolate (6), Vegetables (and juice Food of non-animal origin (11) include: Cereal products and legumes (1), Fruits (and juice) (1), Sweets and chocolate (5), Vegetables (and juice) (4). Milk and milk products (6) include: Cheese (1), Dairy products (other than cheeses Note: Data from 97 outbreaks are included: Austria (1) Other or mixed red meat and products thereof (10). Food of non-animal origin (2) include: Cereal products and legumes (1), Vegetables (and juice ) include: Crustaceans, shellfish, molluscs and products thereof (1), Fish and fish products Note: Data from 319 outbreaks are included: Austria (22) Other or mixed red meat and products thereof (5), Meat and meat products, unspecified (2). Milk and milk products (6) include: Cheese (1), Dairy products Note: Data from 21 outbreaks are included: Austria (1) Pig meat (1), Poultry meat (2), Meat and meat products Meat and meat products (5) include: Bovine meat (4) Note: Data from 155 outbreaks are included: Belgium (1) Food of non-animal origin (8) include: Cereal products and legumes (2), Fruits (and juice) (1), Vegetables (and juice) (5) Milk and milk products (3) include: Cheese (2), Dairy products (other than cheeses Note: Data from 75 outbreaks are included: Belgium (2) Other or mixed red meat and products thereof (2), Meat and meat products, unspecified (8). Food of non-animal origin (4) include: Cereal products and legumes (3), Vegetables (and juice ) include: Crustaceans, shellfish, molluscs and products thereof (1), Fish and fish products (2).and products thereof Note: Data from 74 outbreaks are included France (37) Poultry meat (2), Other or mixed red meat and products thereof (1), Meat and meat products Milk (3). Food of non-animal origin (2) include: Cereal products and legumes (1), Vegetables (and juice Note: Data from seven outbreaks are included: France (1) Note: Data from 96 outbreaks are included: Belgium (1) Denmark (19) Meat and meat products (8) include: Bovine meat (1), Pig meat (1), Poultry meat (2), Other or mixed red meat and products thereof Note: Data from 22 outbreaks are included Note: Data from 11 outbreaks are included: Denmark (1) Hepatitis A' includes also FBOs with causative agent encoded as 'hepatitis virus includes FBOs with causative agent encoded as B. cereus enterotoxins. 'C. perfringens' includes FBOs with causative agent encoded as Clostridium unspecified. 'S. aureus' includes FBOs with causative agent encoded as Staphylococcus the mean number of cases in strong-evidence waterborne outbreaks was 106 in MS and 304 in non-MS. In Norway, C. jejuni was responsible for more than 2,000 cases in a single outbreak. Six outbreaks caused by 'norovirus and other Calicivirus' in MS resulted in 984 cases. Most of the weak-evidence waterborne outbreaks reported by eight MS were caused by STEC (10 outbreaks), norovirus and other Calicivirus (six outbreaks), Giardia (three outbreaks), Cryptosporidium (one outbreak) An overview of the national surveillance systems for human rabies in 950 confirmed human cases of Q fever were reported in the EU. Spain reported the most cases (N = 332, more than one-third of all confirmed cases) for 2019 • The EU notification rate in humans was 0.19 per 100,000 population, which is slightly higher than in 2018 (0.16 per 100,000 population) • There was no statistically significant increase or decrease over the last 5 years • In animals, cattle and small ruminants are mostly sampled due to clinical investigations of animals suspected to be infected by C. burnetii. Because there is no compulsory harmonised monitoring or surveillance in animals in the EU, data reported to EFSA do not make it possible to follow or analyse trends for Q fever at the EU level or to compare national differences 18 MS and four non-MS reported 2019 data for C. burnetii from cattle, sheep and goats and several other domestic and wild animal species. The overall proportion of testpositive animals in EU was 8 Surveillance and monitoring of Coxiella burnetii in the EU/EFTA 27 EU MS, Iceland, Norway and Switzerland provided information on Q fever in humans. Twenty-three EU countries used the EU case definition, whereas Denmark, France, Germany and Italy used another case definition. Reporting is mandatory in 25 EU countries and voluntary in France and the UK. Disease surveillance is comprehensive 26 and mostly passive except in Czechia, Portugal and Slovakia Italy performed a systematic survey to estimate the national seroprevalence or to confirm the presence of C. burnetii in blood or organ/tissue samples from domestic and wild animals analysed mainly via ELISA. Because Q fever monitoring data reported by MS to EFSA are generated by non-harmonised monitoring schemes across MS with no mandatory reporting requirements, these data can only be used for descriptive summaries. Indeed, the results submitted by MS are mostly not directly comparable due to differences in sampling strategy, testing (laboratory analytical) methods, coverage of the monitoring and sensitivity of the surveillance for C. burnetii. They preclude additional data analyses such as following or assessing EU-level temporal and spatial trends. 26 (i) Comprehensive: All healthcare providers of at least one level of care in a defined geographical area cases occurred during the whole year but with a seasonal increase between April and September when more than 60% of the cases were reported. There was no statistically significant At the country level, Poland and Romania reported a significantly (p < 0.01) increasing trend and Germany and France a significantly decreasing trend in the past five years Sixteen MS and three non-MS provided data for sheep and goats, for 2019 793 animals were tested of which, respectively, 6.6% and 8.8% tested positive for C. burnetii. Samples at animal level were mainly taken by Italy (n = 2,670) Seventeen MS and four non-MS provided data for cattle for 2019. In total, 4,318 holdings/flocks and 19,035 animals were tested, of which, respectively, 10.2% and 5.3% tested positive and Slovakia. Slovakia reported locally acquired infections for the first time since 2015. Most locally acquired infections were reported by Greece, Romania and Italy, accounting, respectively, for 53%, 16% and 13% of the total number of reported infections in the EU. The overall EU notification rate per 100,000 population in 2019 was 0.09 compared with 0 • The EU notification rate for 2019 for human tularaemia cases was 0.25 cases per 100,000 population Tap water, including well water' was the incriminated food vehicle in both these strong-evidence outbreaks, causing 36 illnesses from whom six were hospitalised two MS (Austria and Sweden) reported data on the occurrence of Francisella tularensis in hares. Sweden also reported cases in muskrats. One non-MS (Switzerland) reported samples taken from wild species (hares, beavers, squirrels, hedgehogs, mice, deer, foxes and polecats Austria and Sweden) reported that 67 out of the 211 hares tested positive (31.7%) (17 An overview of the national surveillance systems for tularaemia in humans 01 01, but it is reportable to the OIE if a new disease event occurs in a country. However, the notification is mandatory by national law in the Netherlands, Sweden, Iceland and Switzerland. The monitoring data from animals on Francisella tularensis are voluntarily submitted by MS and EFTA countries to EFSA. The data are collected without harmonised design at the EU level and only allow for descriptive summaries and not for trend analyses The reporting of food-borne tularaemia disease outbreaks in humans is mandatory according the Zoonoses Directive 2003/99/EC. When the UK data were collected the UK was an EU MS but as of 31 Tap water, including well water' was the incriminated food vehicle in both these strong-evidence outbreaks, causing 36 illnesses from whom six were hospitalised, no deaths. Previously Norway also reported one tularaemia waterborne outbreak in 2016 (6 illnesses from whom one was hospitalised, no deaths) and one tularaemia food-borne outbreak in Disease Programme on Emerging, Food-and Vector-Borne Diseases 9789241547376_eng.pdf Animals Annual national zoonoses country reports (reports of reporting countries on national trends and sources of zoonoses Other zoonoses and zoonotic agents In 2019, among others, data on Bacillus, Chlamydia, Clostridium, Cysticercus, Enterococcus, hepatitis A virus, Klebsiella, Leptospira, marine biotoxins, norovirus, Proteus, Sarcocystis, Shigella, coagulase-positive Staphylococcus and tick-borne encephalitis virus were reported to EFSA. When the UK data were collected the UK was an EU MS but as of 31 Greece reported in total 2,079 monitoring results for Chlamydia (Chlamydia/ Chlamydophila psittaci) in animals. Overall 8.6% were positive and were from: birds, cattle, goats, Psittacidae, pigeons, pigs, sheep and wild ruminants Sampled foods were: bakery products, cereals and meals, cheeses made from cows' milk, crustaceans, dairy products (excluding cheeses), fats and oils (excluding butter), fishery products, unspecified, fruits, honey, juice, meat and meat products, other processed food products and prepared dishes, RTE salads, sauce and dressings, soups, vegetables and water. From animals Greece and the non-MS the Republic of North Macedonia submitted overall 136 samples and both countries reported positive domestic livestock Bulgaria was the only MS that reported data on non-pathogenic Enterococcus in 2019. None of the samples (potable water, N = 337 samples from own checks Romania and Slovenia) reported on the occurrence of norovirus in fruits and vegetables and other food of non-animal origin (N = 1,097). France reported five norovirus-positive samples from whole fruits intermedius and unspecified, excluding methicillin-resistant S. aureus and staphylococcal enterotoxins) in various animal (N = 6,058) and food (N = 11,110) products sampling units. Overall, from animals 18.9% and from food 9.6% were reported positive. Positive tested foods were; cheeses, made from unspecified milk or other animal milk, ice-cream, pre-cut fruits and vegetables, meat products from broilers, meat preparation and meat products from other animal species or not specified, meat products from other animal species or not specified, other processed food products and prepared dishes (amongst other pasta), sauce and dressings, pastry, soft and semi-soft cheeses made from cows' milk Slovenia reported test results for the presence of TBE of 20 batches of raw milk, from goats and from sheep and 30 batches of cheese from goat's milk and milk from sheep and none were positive Malta (N = 63,897 carcases from pigs, cattle, sheep and goats) and Sweden (3,005,930 carcases from pigs and cattle) reported no positive findings. Slovenia found eight (0.007% out of 116,495) positive cattle and no positive pigs. Bulgaria reported, respectively, < 0.001%, 0.47% and 0.06% positive pigs, sheep and goats and cattle out of 1,196,086, 235,286 and 29,274 examined. In Belgium, 1,075 out of the 840,654 cattle (0.13%) inspected at the slaughterhouse were positive. Luxembourg found 0.3% positive carcases from cattle out of 26,818 inspected. Spain provided data on cysticerci in various animal species: 74 (0.004%) out of 1,819,799 cattle, 0.004% out of 37,835,368 pigs, 5.2% out of 3,325,552 sheep and 2.9% out of 1,100,793 goats were positive for cysticerci. Finally, 38,917 wild boars and 127,264 deer were inspected at game handling establishment and one (0.003%) and one (0.001%) were positive for cysticerci, respectively. Examined carcases from 4,317 wild mouflons were all negative. Estonia did not submit data for 2019 but informed that no cases of cysticerci of Taenia saginata and Taenia solium were detected during visual post-mortem inspection at slaughterhouses of all slaughtered animals. Belgium reported for 2019 840,654 bovine carcases from slaughterhouse inspection for the threshold, it can lead to symptoms such as skin flushing, rash, gastrointestinal complaints and throbbing headache mg/kg) and 'Fishery products which have undergone enzyme maturation treatment in brine, manufactured from fish species associated with a high amount of histidine' (food category 1.26: n = 9 020) for histamine in 'fish, fishery products from fish species associated with a high amount of histidine' were reported at retail by four MS (Romania, Slovakia, Slovenia and Spain) and overall two (0.2%) were reported with quantified results exceeding 200 mg/kg, whereas three samples (0.3%) were exceeding 100 mg/kg but did not exceed 200 mg/kg and 12 (1.2%) were below or equal to 100 mg/kg. Data (N = 38) for histamine in 'Fishery products which have undergone enzyme maturation treatment in brine, manufactured from fish species associated with a high amount of histidine' were reported at retail by two Romania and Spain and none were reported with quantified results exceeding 400 mg/kg, whereas one sample (2.6%) exceeded 200 mg/kg but did not exceed 400 mg/kg and the other samples were negative Bacterial pathogens of pigs with particular reference to Escherichia coli: a systematic review and meta-analysis Towards a standardised surveillance for Trichinella in the European Union Data sheet on foodborne biological hazards -Listeria monocytogenes Salmonellosis: the role of poultry meat The occurrence of Echinococcus spp. in golden jackal (Canis aureus) in southwestern Hungary: should we need to rethink its expansion? EU-wide monitoring of biological hazards along the food chain: achievements, challenges and EFSA vision for the future Non-typhoidal Salmonella in the pig production chain: a comprehensive analysis of its impact on human health. Pathogens, 8, 19 Recognising the substantial burden of neglected pandemics cystic and alveolar echinococcosis. The Lancet. Global Health Echinococcus multilocularis Echinococcus granulosus sensu lato Statistical Models in Epidemiology Development of harmonised schemes for monitoring and reporting of rabies in animals in the European Union Salmonella in foods: a reemerging problem Wilderness in the city: the urbanization of Echinococcus multilocularis Nachweis von Escherichia coli O157 in Fleisch und Fleischerzeugnissen Molecular epidemiology of Brucella melitensis strains causing outbreaks in Croatia and Bosnia and Herzegovina Multi-country outbreak of Salmonella Enteritidis infections linked to Polish eggs -epidemiological update 2019. Brucellosis. In: ECDC. Annual epidemiological report for 2017 Scientific Report of EFSA on technical specifications for the monitoring and reporting of verotoxigenic Escherichia coli (STEC) on animals and food (STEC surveys on animals and food) Update of the technical specifications for harmonised reporting of food-borne outbreaks through the European Union reporting system in accordance with Directive Zoonoses, antimicrobial resistance and foodborne outbreaks guidance for reporting 2019 data. EFSA supporting publication 2020;EN-1792 Manual for reporting on zoonoses and zoonotic agents, within the framework of Directive 2003/99/EC, and on some other pathogenic microbiological agents for information derived from the year 2019. EFSA supporting publication 2020;EN-1791 Statement on a conceptual framework for bovine tuberculosis Scientific Opinion -update on oral vaccination of foxes and raccoon dogs against rabies Scientific Opinion on the public health hazards to be covered by inspection of meat (swine) Scientific Opinion on the public health hazards to be covered by inspection of meat (poultry) Scientific Opinion on Campylobacter in broiler meat production: control options and performance objectives and/or targets at different stages of the food chain Scientific Opinion on STEC-seropathotype and scientific criteria regarding pathogenicity assessment Scientific Opinion on the public health hazards to be covered by inspection of meat (solipeds) Scientific Opinion on Listeria monocytogenes contamination of ready-to-eat (RTE) foods and the risk for human health in the EU Scientific Opinion on the public health risks associated with food-borne parasites Pathogenicity assessment of Shiga toxin-producing Escherichia coli (STEC) and the public health risk posed by contamination of food with STEC Scientific Opinion on the Salmonella control in poultry flocks and its public health impact Update and review of control options for Campylobacter in broilers at primary production Scientific Opinion on the public health risk posed by Listeria monocytogenes in frozen fruit and vegetables including herbs, blanched during processing The Community summary report on trends and sources of zoonoses, zoonotic agents and food-borne outbreaks in the European Union in 2008 The European Union summary report on trends and sources of zoonoses, zoonotic agents and food-borne outbreaks in 2009 The European Union summary report on trends and sources of zoonoses, zoonotic agents and food-borne outbreaks in 2010 The European Union summary report on trends and sources of zoonoses, zoonotic agents and food-borne outbreaks in 2011 The European Union summary report on trends and sources of zoonoses, zoonotic agents and food-borne outbreaks in 2012 The European Union summary report on trends and sources of zoonoses, zoonotic agents and foodborne outbreaks in 2013 The European Union summary report on trends and sources of zoonoses, zoonotic agents and foodborne outbreaks in 2014 The European Union summary report on trends and sources of zoonoses, zoonotic agents and food-borne outbreaks in 2015 The European Union summary report on trends and sources of zoonoses, zoonotic agents and foodborne outbreaks in 2016 European Food Safety Authority and European Centre for Disease Prevention and Control) Multi-country outbreak of Listeria monocytogenes serogroup IVb, multi-locus sequence type 6, infections linked to frozen corn and possibly to other frozen vegetables -first update The European Union summary report on trends and sources of zoonoses, zoonotic agents and foodborne outbreaks in 2017 Multi-country outbreak Salmonella Poona infections linked to consumption of infant formula. EFSA supporting publication 2019;EN-1594 Multi-country outbreak of Listeria monocytogenes clonal complex 8 infections linked to consumption of cold-smoked fish products European Food Safety Authority and European Centre for Disease Prevention and Control), 2019c. The European Union One Health Multi-country outbreak of Listeria monocytogenes sequence type 6 infections linked to ready-to-eat meat products enteritidis-infections-linked-eggs EFSA and ECDC (European Food Safety Authority and European Centre for Disease Prevention and Control), 2020b. The European Union summary report on antimicrobial resistance in zoonotic and indicator bacteria from humans Brucellosis is a public health problem in southern Italy: burden and epidemiological trend of human and animal disease The epidemiology of brucellosis in Greece Emergence of multidrug-resistant Salmonella enterica subspecies enterica Serovar Infantis of multilocus sequence type 2283 in German broiler farms Cases of human brucellosis in Sweden linked to Middle East and Africa Whole genome sequencing of Brucella melitensis isolated from 57 patients in Germany reveals high diversity in strains from Middle East Q fever in humans and farm animals in four European countries Antigenic formulae of the Salmonella serovars, 9th Edition. WHO Collaborating Centre for Reference and Research on Salmonella, Institute Pasteur The status of tularemia in Europe in a one-health context: a review Francisella tularensis in Swedish predators and scavengers Real-time polymerase chain reaction (PCR)-based method for the detection of food-borne pathogens. Horizontal method for the detection of shiga toxin-producing Escherichia coli (STEC) and the determination of O157, O111, O26, O103 and O145 serogroups Environmental surveillance during an outbreak of tularaemia in hares, the Netherlands Brucella-positive raw milk cheese sold on the inner European market: a public health threat due to illegal import? Whole-genome sequencing and antimicrobial resistance in Brucella melitensis from a Norwegian perspective Outbreak of Salmonella enterica serotype Poona in infants linked to persistent Salmonella contamination in an infant formula manufacturing facility Human brucellosis -new public health problem in Bulgaria European echinococcosis registry: human alveolar echinococcosis The echinococcoses: diagnosis, clinical management and burden of disease Human brucellosis in France in the 21st century: results from national surveillance Inventory of available data and data sources and proposal for data collection on vector-borne zoonoses in animals Pig farming in the European Union: considerable variations from one Member State to another Tularaemia: clinical aspects in Europe Brucellosis: a rapid risk assessment by a regional outbreak team and its coordinated response with the Directorate-General for Food and Veterinary, North region of Portugal Combined crosssectional and case-control study on Echinococcus multilocularis infection in pigs in Switzerland Escherichia coli O157. Detection in food and feeding stuffs Imported brucellosis: a case series and literature review Manual of Diagnostic Tests and Vaccines for Terrestrial Animals The geographical distribution and prevalence of Echinococcus multilocularis in animals within the European Union and adjacent countries: a systematic review and metaanalysis Relationship between seroprevalence in the main livestock species and presence of Toxoplasma gondii in meat (GP/EFSA/BIOHAZ/2013/01). An extensive literature review. Final report Animal reservoirs of Shiga toxin-producing Escherichia coli Searching for Trichinella: not all pigs are created equal Trichinella pseudospiralis an elusive nematode Hosts and habitats of Trichinella spiralis and Trichinella britovi in Europe Shiga toxin-producing Escherichia coli in Feces of Finisher Pigs: isolation, identification, and public health implications of major and minor serogroups Cystic echinococcosis in unaccompanied minor refugees from Afghanistan and the Middle East to Germany The first meeting of the European Register of Cystic Echinococcosis (ERCE). Parasite Vectors The European Register of Cystic Echinococcosis, ERCE: state-of-the-art five years after its launch First Case of Brucellosis Caused by an Amphibian-type Brucella Q fever in the Netherlands-2007-2010: what we learned from the largest outbreak ever Echinococcus multilocularis and Trichinella spiralis in golden jackals (Canis aureus) of Hungary Prevalence of abdominal cystic echinococcosis in rural Bulgaria, Romania, and Turkey: a cross-sectional, ultrasound-based, population study from the HERACLES project Campylobacter seroconversion rates in selected countries in the European Union Outbreaks associated with untreated recreational water The application of quantitative risk assessment to microbial food safety Epidemic and pandemic alert and response 'Staphylococcus, unspecified' or 'Staphylococcal enterotoxins'. ' Meat and meat products' include bovine meat and products thereof, broiler meat (Gallus gallus) and products thereof, other or mixed red meat and products thereof, other, mixed or unspecified poultry meat and products thereof, pig meat and products thereof, sheep meat and products thereof, turkey meat and products thereof. ' This chapter has a simplified structure underpinned by descriptive summarisation of submitted data (see rationale p. 16 of Introduction). • Yersiniosis was the fourth most commonly reported zoonosis in humans in 2019 with 6,961 confirmed cases reported in the EU.• The trend of human yersiniosis cases was stable (flat) in 2015-2019. Toxoplasma gondii This chapter has a simplified structure underpinned by descriptive summarisation of submitted data (see rationale p. 16 of Introduction). Toxoplasma gondii is widely prevalent in humans and animals world-wide. Virtually all warmblooded animals can act as IHs, but the life cycle is only completed in the DHs: cats and other felines, including lynx which is present in Europe.Only congenital toxoplasmosis is reported to ECDC. There is two-year delay in data reporting and the most recent epidemiological data, which pertain to the year 2018, are available at https://www.ecdc. europa.eu/en/publications-data/congenital-toxoplasmosis-annual-epidemiological-report-2018• In 2018, 208 confirmed cases of congenital toxoplasmosis were reported in the EU/EEA, with France accounting for 72.6% of all confirmed cases due to the active screening of pregnant women.• No food-borne toxoplasmosis outbreak was reported in 2019 in EU and no such single foodborne outbreak has ever been reported to EFSA since the start of its food-borne outbreaks data collection in 2004.• In total, 13 MS and two non-MS reported 2019 monitoring data on Toxoplasma infections in animals. Most animals tested were sheep and goats that also showed the highest overall prevalence of Toxoplasma infections in animals (13.5%) as reported by 12 MS. Most samples were obtained from clinical investigations. It is not possible to make a good estimate of the prevalence of Toxoplasma infections in animals due to the use of different diagnostic methods (indirect methods detecting antibodies vs. direct methods), the different sampling schemes in the MS and the lack of information on the animals' age and rearing conditions. Surveillance and monitoring of Toxoplasma in the EU An overview of the national surveillance systems for human congenital toxoplasmosis is available at https://www.ecdc.europa.eu/en/publications-data/congenital-toxoplasmosis-annual-epidemiologicalreport-2018 This chapter has a simplified structure underpinned by descriptive summarisation of submitted data (see rationale p. 16 of Introduction). Key facts The aim of wildlife rabies surveillance is to demonstrate the absence of disease, or to identify its presence or distribution, to allow timely dissemination of information for integrated action among different sectors such as public health and veterinary sectors.According to Regulation (EU) No 652/2014 23 , multiannual programmes for eradication of rabies may be co-financed by the EU. In 2019, 12 MS (Bulgaria, Croatia, Estonia, Finland, Greece, Hungary, Latvia, Lithuania, Poland, Romania, Slovakia and Slovenia) had approved eradication, control and surveillance programmes for rabies. A wildlife oral rabies vaccination campaign (ORV) is currently ongoing in these MS, as well as in some of the EU-bordering countries. The surveillance of rabies is carried out by sampling and testing 'indicator animals'; these are animals that are found dead in their natural habitat and/or suspected animals from wildlife and domestic species (foxes, badgers, raccoon dogs, dogs, cattle, cats, sheep, equines, goats, rabbits, etc.), i.e. animals showing neurological clinical signs or abnormal behaviour compatible with rabies.The collection of healthy animals of the species targeted by oral vaccination (foxes, raccoon dogs and also golden jackals) is also valuable for monitoring the efficacy of the ORV campaign by determining the immunity and the oral vaccine bait uptake of animals.Imported or travel-related companion animals (mainly dogs and cats) from territories and non-EU countries not included in Annex II of Regulation (EC) No 577/2013 24 are currently tested for rabiesspecific antibodies.EU MS also need to notify outbreaks of infection with rabies virus in non-flying terrestrial animals to the EU ADNS. 12 In this report, the results of the surveillance activities for rabies are summarised for the indicator wild species such as foxes, raccoon dogs, raccoons (Procyon lotor) and other wild species (badgers, deer, marten, rodents, jackals, lynx, bears, hares, hedgehogs, minks, wolverine, wild boar, squirrels, ferrets, otter, polecat, etc.).Separate tables for rabies surveillance in domestic carnivores (dogs and cats) and farmed animals (cattle, small ruminants, solipeds, pigs, rabbits, ferrets) were also produced to summarise the surveillance activities in the different MS. These summary tables are in the supporting information to this report.All data were summarised (aggregated) at MS level; if MS reported data only at regional level or only for some regions, the total number of tested animals were not integrated in the summary tables or maps as it was not clear whether all regions in the MS were tested or not.When the UK data were collected the UK was an EU MS but as of 31 January 2020 it has become a third country.EFSA AHAW Panel, 2015) . Although a reduction in the number of samples taken from foxes was observed (0), caution must be taken when interpreting this decrease in the sample size. As those reported numbers include monitoring and surveillance strategies and are aggregated at a country level, the decrease in sample size could be the result of a smaller number of suspect cases throughout Europe due to a decrease in prevalence. Nonetheless, MS, especially those with a recent history of rabies, should ensure that a robust surveillance programme is in place capable of the early detection of any potential cases of rabies in their territories. Related projects and Internet sources This chapter has a simplified structure underpinned by descriptive summarisation of submitted data (see rationale p. 16 of Introduction).Tables and figures that are not presented in this chapter are published as supporting information to this report and are available as downloadable files from the EFSA knowledge junction at zenodo https://doi.org/ 10.5281/zenodo.4298993. The human epidemiological data for tularaemia for 2019 are available at https:// www.ecdc.europa.eu/en/publications-data/tularaemia-annual-epidemiological-report-2019. Summary statistics of human surveillance data with downloadable files are retrievable using ECDC's Surveillance Atlas of Infectious Diseases at http://atlas.ecdc.europa.eu/public/index.aspx samples from pets and domestic animals. For Shigella three food samples from Sweden were negative. For Vibrio, 326 food samples in total from Bulgaria, the Netherlands and Sweden, 32 were positive (9.8%). These positive results were from raw fish, from shrimps and from lobsters from third countries (border inspection activities). Out of the 535 samples tested (from fruits and vegetables) for hepatitis A virus (France, Romania and Sweden), no sample was positive. Out of a total of 136 cattle, sheep and goats tested for Klebsiella (Greece), one milk ewe was positive. Bulgaria reported monitoring results for marine biotoxins (N = 94) from raw molluscan shellfish with no positives.Microbiological contaminants subject to food safety criteria (Regulation (EC) No 2073 This chapter summarises the 2019 information provided by reporting countries on microbiological contaminants in foods: histamine, staphylococcal enterotoxins and Cronobacter sakazakii for which FSC are set down in the EU legislation (Regulation (EC) No 2073/2005).As for food categories subject to FSC, EFSA used the following specific testing data in the context of Regulation ( Appendix B -Occurrence of L. monocytogenes at retail and processing Appendix C -Atlases of STEC serogroups: food and animals, EU, 2019