key: cord-0869455-hssuehm1 authors: Espenhain, L.; Tribler, S.; Joergensen, C. S.; Holm Hansen, C.; Wolf Sonksen, U.; Ethelberg, S. title: Prevalence of SARS-CoV-2 antibodies in Denmark: Nationwide, population-based seroepidemiological surveys date: 2021-04-09 journal: nan DOI: 10.1101/2021.04.07.21254703 sha: 487d1c551dc65fe80b2fdfd8359fcafe7d3ebcec doc_id: 869455 cord_uid: hssuehm1 Background. Seroprevalence studies have proven an important tool to monitor the progression of the coronavirus disease 2019 (COVID-19) epidemic. We present results of consecutive population-based seroprevalence surveys performed in Denmark in 2020. Methods. Invitation letters including a questionnaire covering symptoms were sent to representatively drawn samples of the population in spring, late summer and autumn/winter of 2020. Blood samples from participants taken at public test-centers were analyzed for total Ig and seroprevalence estimates per population segment calculated and compared to other surveillance parameters. Results. From 34,081 participating individuals (response rate 33%), we obtained seroprevalence estimates increasing from 1.1% (95%CI: 0.7%-1.7) in May to 4.0 % (95%CI: 3.4%-4.7%) in December 2020. By December 2020, 1.5% of the population 12 years and older had tested positive by PCR. Seroprevalence estimates were roughly 3 times higher in those aged 12-29 compared to 65+ and higher in metropolitan municipalities. Among seropositives, loss of taste/smell were the more specific symptoms, 32%-56% did not report any symptoms. In half of seroconverted families, we did not see evidence of transmission between generations. Infected individuals in older age groups were hospitalized several fold more often than in younger. Conclusions. Seroprevalence increased during 2020; younger age groups were primarily infected in the autumn/winter surge. Approximately half were asymptomatically infected. Denmark has a high per capita test rate; roughly two undiagnosed infections of COVID-19 were estimated to occur for each diagnosed case. The epidemic appears to have progressed relatively modestly during 2020 in Denmark. The degree to which the COVID-19 pandemic is spreading through different countries or regions may 2 be assessed through population-based seroprevalence studies, which aim to quantify the proportion of the 3 population that has developed antibodies against SARS-CoV-2. Such studies have to date been performed 4 in a number of countries [1] [2] [3] [4] [5] [6] [7] [8] [9] . 5 Similar to several other European countries, Denmark experienced increased transmission of SARS-6 CoV-2 infection in spring and late autumn 2020. A comprehensive lock-down was imposed in March 2020, 7 gradually lifted towards summer and again gradually reintroduced during autumn and winter [10] . The 8 Danish National Seroprevalence Survey of SARS-CoV-2 infection (DSS) was initiated in the spring of 2020, 9 following a parliamentary decision in April 2020, which called for a representative population study to be 10 performed. The study design was set by recommendation from a group of independently appointed 11 national experts in April 2020 [11] and seroprevalence subsequently determined at several time points 12 throughout 2020. Here we describe the set-up and results from DSS and relate the results to the national 13 surveillance of PCR diagnosed COVID-19-cases. 14 15 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 9, 2021. ; https://doi.org/10.1101/2021.04.07.21254703 doi: medRxiv preprint Design and study population 17 DSS is a nationwide population-based prevalence survey aiming to investigate seroprevalence for 18 SARS-CoV-2. The study was launched in the spring of 2020 and performed by Statens Serum Institut (SSI) 19 over three rounds: In May 2020 (DSS-I), August 2020 (DSS-II) and September-December 2020 (DSS-III). 20 For each survey round, a random population sample of Danish residents was drawn from the Danish 21 civil registry [12] . For DSS-I, adults aged 18 years and older living in one of 30 municipalities which had a 22 test facility (see below) at that time (n=5), or were neighboring a municipality with a test facility (n=25) 23 were eligible (approximately 45% of the population of Denmark). For DSS-II and DSS-III, people aged 12 24 years or older were selected by random sampling, with no restriction on municipality (n=98). Parents living 25 on the same address as invited children 12-17 years old, were also invited to participate. 26 27 We primarily invited participants using the Danish digitalized postal system covering 90% of the 29 Danish population [13] . Invitation letters (as pdf's) were sent via the secure, digital mailbox-system ("e-30 Boks"), using the civil registration number. Additionally, physical letters were sent by regular mail to 31 invitees below 18 years of age and to those without e-Boks. For DSS-I we invited 2,600 people on May 5 32 Denmark [16] . Using MiBa we identified antibody test results and previous COVID-19-PCR-test results 59 amongst the study participants. We used information on number of admitted and deceased by date, sex 60 and age group from the Danish surveillance system of COVID-19 [17, 18] . This involves daily registry linkage 61 to the National Patient Registry [19] for information about hospital admissions and to the Civil Registry and 62 The National Cause of Death Register [20] for information about deaths within 30 days for PCR-diagnosed 63 COVID-19 cases. 64 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 9, 2021. ; https://doi.org/10.1101/2021.04.07.21254703 doi: medRxiv preprint We included all persons with a conclusive antibody test result within ten weeks from the invitation. 66 We estimated the seroprevalence as the proportion of participating individuals with a positive antibody 67 test result. We adjusted the seroprevalence estimates for test sensitivity (0.97) and specificity (0.995) 68 (internal assessment at SSI) using the Rogan-Gladen estimator [21] and computed 95% confidence intervals 69 using Blaker's method [22] . We present seroprevalence by age group, sex, the five geographical regions of 70 Denmark and classification of municipality (capital, metropolitan, provincial, commuter, rural) as defined by 71 Statistics Denmark [23] . To evaluate whether the variation in response rate by age groups, sex and region 72 affected the estimated seroprevalence, we predicted the value (0 = negative and 1 = positive) of the 73 missing serology test by multiple imputation including sex, age group and region in the prediction model. 74 For the analyses, we defined four periods (May, August, October and December) as the time periods 75 the seroprevalence estimate reflected. We compared the estimates of infected individuals to the number 76 of PCR test positive, hospitalized and deceased in national surveillance. To do that, we subtracted 14 days 77 from the mean date of blood sampling in the four periods to compare with PCR-test positive from the 78 national surveillance system, and added 10 days to find the comparable date for hospital admission and 20 79 days for number of deaths. 80 81 The DSS was performed as a national disease surveillance project, registered with the Danish Data 83 Protection Agency and approved regarding legal, ethical and cyber-security issues by the SSI Compliance 84 department in conjunction with the Danish governmental law firm. Participation in the survey was 85 voluntary and invitees received information about the selection procedure, risks associated with 86 participation, data security issues, their legal rights, including the right to withdraw from the study, and the 87 use of their data and results in the letter of invitation. 88 89 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 9, 2021. ; https://doi.org/10.1101/2021.04.07.21254703 doi: medRxiv preprint Participation 91 The COVID-19 incidence and test intensity in Denmark in 2020 is depicted in Figure 1 . The three DSS 92 study rounds had 2,512 (48%), 7,015 (39%), and 18.161 (26%) participants, respectively ( Table 1 ). The 93 median dates of sampling for the four defined study periods were May 18 (defined as 'May'), September 19 94 ('August'), November 6 ('October'), and December 16 ('December') ( Figure 1 ). 95 Overall participation was lower in males and younger age groups (Table 1) . For DSS-II and DSS-III 96 respectively, the questionnaire was filled in by 2,737 (39%) and 10,358 (57%) of the participants. 97 98 Seroprevalence 99 The proportion of participants with detectable SARS-CoV-2 antibodies increased from 1.1% (95%CI: 100 0.7%-1.7%) in May 2020 to 4.0 % (95%CI: 3.4%-4.7%) in December 2020 (Table 2) . Restricting the analysis 101 to match the narrower geographic and age inclusion criteria for DSS-I, the estimated seroprevalence in 102 December was 5.1% (95%CI 4.0%-6.2%). When taking the non-response by age group, sex and region of 103 residence into account by imputation, the estimates increased with up to 0.4 percentage points (Table 2) . 104 Point estimates tended to be higher in the two younger age groups (12 to 17 years and 18 to 39 years of 105 age), lower in the 65 years and older age group (Figure 2 ), and higher in the Capital region than in the other 106 four regions. No difference was observed by sex (Table 2) . 107 108 A total of 1,244 families had a child and at least one parent tested. Among these, 79 (6.4%) families 110 were found to have at least one seropositive family member. These included 3.2% with a seropositive child 111 and 4.2% families with at least one seropositive parent. In 21 of the 79 seropositive families (27%), both 112 child and parent(s) were seropositive, in 19 families (24%) only the child was seropositive and in 39 (49%) 113 seropositive families the child was seronegative. 114 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 9, 2021. ; https://doi.org/10.1101/2021.04.07.21254703 doi: medRxiv preprint Of the 369 seropositive participants who filled in the questionnaire, 59% reported having 117 experienced at least one of five core symptoms (fever, cough, shortness of breath and/or loss of sense of 118 taste or smell) since February 2020, versus 28% among seronegative participants ( Table 3 ). Loss of smell or 119 taste, reported by 24% and 23% of the seropositives, respectively, were highly associated with 120 seropositivity (odds ratios of 16 and 14, respectively, Table 3 ). 121 In DSS-II 44% (95%CI: 32%-56%) of the seropositive participants reported no symptoms since 122 February and a further 9% reported symptoms not belonging to the five mentioned core symptoms. In DSS-123 III the equivalent figures were 27% and 11%. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 9, 2021. ; https://doi.org/10.1101/2021.04.07.21254703 doi: medRxiv preprint infected to PCR-diagnosed cases was three in December and six in May 2020 (Table 4 ). The estimated ratio 140 varied by age. It was higher in the 18-39 year age group in May and August and decreased during autumn. 141 No obvious pattern was seen for the 65-year and older age group, by region or sex during the period (Table 142 4). The infection fatality rate and rate of admitted out of the estimated number of infected increased 143 markedly with older age (Table 4) . 144 145 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 9, 2021. ; https://doi.org/10.1101/2021.04.07.21254703 doi: medRxiv preprint In this national representative seroprevalence study among Danish residents aged 12 years and 147 older, we found detectable SARS-CoV-2 antibodies in 4.0 % (95%CI: 3.4-4.7%) of the participants by the 148 beginning of December 2020. This was four times more than the estimate from May and twice the estimate 149 of August 2020. This study also provides information on the regional and demographic progression of the 150 epidemic and its results can be interpreted in the context of other surveillance parameters. 151 The seroprevalence varied by geography and age group, which is consistent with the general picture 152 from the national surveillance system. It appears that people from more densely populated urban areas 153 were infected in the early stage of the epidemic, and that the epidemic only gradually spread to the less 154 densely populated areas later. Our results are in line with the serological surveys of blood donors which has 155 also been carried out in Denmark [24] , although this group may not be representative of the general 156 population. 157 Between May and December 2020, PCR testing rose from 150 to more than 1000 tests per 10,000 158 inhabitants per week. Our results show that the proportion of PCR-confirmed COVID-19 cases out of all 159 estimated infections was 1:5 in the spring but just 1:2 by early December. There were relatively more 160 undiagnosed cases in the 18-39-year age group during the first six month of the epidemic, suggesting that 161 this age group may experience a less severe course of disease. It is well established that severity of illness 162 increases markedly with age [25] and the estimated infection fatality and infection admission rate also 163 increased markedly by age group in this study. 164 National representative seroprevalence surveys from other countries have shown quite variable 165 seroprevalence estimates [26] : Surveys in France [27] in May, and Spain [2] and Brazil [9] in June and the 166 Netherlands [7] in July 2020 estimated that respectively around 4.5%, 5%, 3.5% and 4% of the population 167 had been infected with SARS-CoV-2 at that point. All four surveys revealed substantial geographical 168 variability. In September a survey carried out in the US found that in 25 of 49 jurisdictions with sufficient 169 samples to estimate seroprevalence more than 5% of people had detectable SARS-CoV-2 antibodies [1] . 170 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 9, 2021. ; https://doi.org/10.1101/2021.04.07.21254703 doi: medRxiv preprint Our results are comparatively low and thus the epidemic appears to have affected Denmark only mildly in 171 2020. This may be corroborated by the cumulated mortality numbers, which by 31 Dec 2020 were 23 per 172 100,000 [17] , placing Denmark in the lower end of the European scale [28] . 173 Estimates from other studies of the share of asymptomatic infections out of the total number of 174 SARS-CoV-2 infections vary notably from a few percent to 41% with a pooled overall proportion of 17% 175 found in a recent meta-analysis [29] [30] [31] . In DSS-II carried out in the late summer, 44% of the seropositive 176 participants did not recall having had any symptoms of acute infection since February 2020 and an 177 additional 9% reported symptoms not typically associated with COVID-19. The percentage reporting no 178 symptoms since February 2020 fell to 27% in the DSS-III. However at this point other respiratory illnesses 179 may increasingly have affected the results. Thus, our best estimate is that around half of the seropositive 180 persons had an asymptomatic infection. We found that loss of smell or taste, experienced by almost ¼ with 181 SARS-CoV-2 antibodies, were by far the more specific symptoms of COVID-19 infection. 182 In more than two thirds of families with at least one seropositive family member, only the parent(s) 183 or the child had seroconverted, indicating that transmission between generations within households is the 184 exception rather than the rule. We were unable to disentangle the chain of transmission between 185 generations. 186 Though actual numbers were low, we found that <5% of previous PCR positive participants did not 187 have detectable SARS-CoV-2-antibodies 12 days after their first positive PCR test. This might be because of 188 waning immunity, or that the individuals did not elicit a detectable antibody response (possibly due to mild 189 or asymptomatic infection). The proportion does correspond to what has been reported from Iceland [32] 190 and a study in preprint from the UK [6] . 191 Denmark has a relatively high degree of IT penetration and frequently makes use of national registers 192 and public digital resources. Utilization hereof was among the strengths of this study. From the Danish 193 national civil register, it was possible to obtain a random sample of residents, and identify the parents of 194 those below 18 years of age. Individually referable national surveillance data allowed us to identify all 195 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 9, 2021. ; https://doi.org/10.1101/2021.04.07.21254703 doi: medRxiv preprint previous PCR tests amongst the participants and relate this to their antibody-status. Another strength of 196 the study was the use of the logistical set-up of the large state-driven test-system, TestCenter Denmark, 197 that was created as a parallel system to the existing clinical test system located at the hospitals. Using the 198 existing set up for PCR testing, meant that most people had an easy access to a test facility for antibody 199 testing. However, due to the geographical distribution of the test stations, some persons had quite long 200 driving distances to a test facility (up to 100 km), which may have affected their willingness to participate in 201 the study. 202 The study had several limitations. When interpreting the findings, the suboptimal participation rates 203 should be taken into account. Participation decreased from 48% in DSS-I, through 39% in DSS-II to 26% in 204 DSS-III, and even fewer replied to the questionnaire concerning symptoms. Even though the drawn sample 205 is representative of the population, participation may not be. We do not know if certain subgroups of the 206 population were underrepresented in the study, but may presume that hard to reach populations would be 207 so. The seroprevalence estimates were stable but slightly underestimated when taking the non-response by 208 age group, sex and region of residence into account. 209 In conclusion, our study provides estimates of SARS-CoV-2 dissemination in Denmark at four time 210 points, based on a representative sample of the population and relate it to the number of PCR-confirmed 211 COVID-19 cases in the national surveillance system. We found that the epidemic had predominantly 212 affected the capital and metropolitan areas and saw indications of a higher seroprevalence in young adults 213 throughout the epidemic, although children 12-17 years old were mainly affected in the second surge of 214 the epidemic. Overall, the estimated seroprevalence in Denmark throughout 2020 was low, compared to 215 other countries. The results seem to support that the measures introduced in Denmark in the spring of 216 2020 and onwards have been effective in keeping the epidemic from developing rapidly in the community, 217 however also indicate that the majority of the population is still at risk of contracting COVID-19. As more 218 than one in three infections seem to be asymptomatic, social distance measures and efforts to identify and 219 isolate new cases and their contacts are imperative for future epidemic control. 220 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Disclaimer. The funder asked for a 'population representative study design'. The funder had no role in data collection and analysis, writing of or decision to submit the manuscript for publication. Funding: This study was supported by an ad hoc grant from the Danish Government ( §16.11.73 on the National budget 2021). We thank everyone who participated in this study by giving blood. We thank all staff in TestCenter Denmark, including staff in test stations, the involved SSI departments and the expert group that advised on the design of the study. All authors (LE, ST, CSJ, CHH, UWS and SE) report no conflict of interest. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 9, 2021. ; https://doi.org/10.1101/2021.04.07.21254703 doi: medRxiv preprint Table 1. Number of invited persons and proportion who participated by DSS, age group, sex, and region DSS-I* DSS-II DSS-III Group Invited Participation Invited Participated Invited Participated . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 9, 2021. ; https://doi.org/10.1101/2021.04.07.21254703 doi: medRxiv preprint **Relevant symptoms includes one or more of the following: fever, cough, shortness of breath and/or loss of sense of taste or smell *** Muscle ache, eye pain, head ache, colored sputum, runny nose, sneezing, back pain, tiredness without one of the five relevant symptoms . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 9, 2021. ; https://doi.org/10.1101/2021.04.07.21254703 doi: medRxiv preprint Estimated SARS-CoV-2 Seroprevalence in the US as of Prevalence of SARS-CoV-2 in Spain (ENE-COVID): a nationwide, population-based seroepidemiological study Seroprevalence of anti-SARS-CoV-2 IgG antibodies in Geneva, Switzerland (SEROCoV-POP): a population-based study Seroprevalence of SARS-CoV-2 in the Norwegian population measured in residual sera -analyses collected in April Novel coronavirus epidemic in the Hungarian population, a cross-sectional nationwide survey to support the exit policy in Hungary Antibody prevalence for SARS-CoV-2 following the peak of the pandemic in England: REACT2 study in 100,000 adults Nationwide seroprevalence of SARS-CoV-2 and identification of risk factors in the general population of the Netherlands during the first epidemic wave Mass SARS-CoV-2 serological screening for the Principality of Andorra SARS-CoV-2 antibody prevalence in Brazil: results from two successive nationwide serological household surveys COVID-19 Transmission Within Danish Households: A Nationwide Study from Lockdown to Reopening The Danish Civil Registration System as a tool in epidemiology Paper-or Web-Based Questionnaire Invitations as a Method for Data Collection: Cross-Sectional Comparative Study of Differences in Response Rate, Completeness of Data, and Financial Cost Increased Risk of Hospitalisation Associated with Infection with SARS-CoV-2 Lineage B.1.1.7 in Denmark Assessment of protection against reinfection with SARS-CoV-2 among 4 million PCR-tested individuals in Denmark in 2020: a population-level observational study MiBa Board of R. The Danish Microbiology Database (MiBa) 2010 to 2013 Overvågningdata for covid-19 i Danmark og Europa Statens Serum Institut -covid-19 -Danmark (Kommunalt) The Danish National Patient Registry: a review of content, data quality, and research potential The Danish Register of Causes of Death Estimating prevalence from the results of a screening test Exact confidence limits for prevalence of a disease with an imperfect diagnostic test Municipality Groups v1 Characteristics and predictors of hospitalization and death in the first 11 122 cases with a positive RT-PCR test for SARS-CoV-2 in Denmark: a nationwide cohort SeroTracker: a global SARS-CoV-2 seroprevalence dashboard 5% of the population of metropolitan France had developed antibodies against SARS-CoV-2 The first results of the EpiCov national survey, 2020. 28. Control) EECfDPa. COVID-19 country overview Natural History of Asymptomatic SARS-CoV-2 Infection Clinical characteristics of asymptomatic and symptomatic patients with mild COVID-19 Estimating the extent of asymptomatic COVID-19 and its potential for community transmission: Systematic review and meta-analysis Humoral Immune Response to SARS-CoV-2 in Iceland *Includes only people 18 years and older living in one of 30 municipalities which had a test facility at the time (n=5) or