key: cord-0972856-rv5fsitx authors: Giacobbe, Daniele Roberto; Battaglini, Denise; Ball, Lorenzo; Brunetti, Iole; Bruzzone, Bianca; Codda, Giulia; Crea, Francesca; De Maria, Andrea; Dentone, Chiara; Di Biagio, Antonio; Icardi, Giancarlo; Magnasco, Laura; Marchese, Anna; Mikulska, Malgorzata; Orsi, Andrea; Patroniti, Nicolò; Robba, Chiara; Signori, Alessio; Taramasso, Lucia; Vena, Antonio; Pelosi, Paolo; Bassetti, Matteo title: Bloodstream infections in critically ill patients with COVID‐19 date: 2020-06-14 journal: Eur J Clin Invest DOI: 10.1111/eci.13319 sha: 7422125eea272180368a395033fa37f459eb65d2 doc_id: 972856 cord_uid: rv5fsitx BACKGROUND: Little is known about the incidence and risk of intensive care unit (ICU)‐acquired bloodstream infections (BSI) in critically ill patients with coronavirus disease 2019 (COVID‐19). MATERIAL AND METHODS: This retrospective, single‐centre study was conducted in Northern Italy. The primary study objectives were: (i) to assess the incidence rate of ICU‐acquired BSI; (ii) to assess the cumulative risk of developing ICU‐acquired BSI. RESULTS: Overall 78 critically ill patients with COVID‐19 were included in the study. Forty‐five episodes of ICU‐acquired BSI were registered in 31 patients, with an incidence rate of 47 episodes (95% confidence interval [CI] 35‐63) per 1000 patient‐days at risk. The estimated cumulative risk of developing at least one BSI episode was of almost 25% after 15 days at risk, and possibly surpassing 50% after 30 days at risk. In multivariable analysis, anti‐inflammatory treatment was independently associated with the development of BSI (cause‐specific hazard ratio [csHR] 1.07 with 95% CI 0.38‐3.04 for tocilizumab, csHR 3.95 with 95% CI 1.20‐13.03 for methylprednisolone, and csHR 10.69 with 95% CI 2.71‐42.17 for methylprednisolone plus tocilizumab, with no anti‐inflammatory treatment as the reference group; overall p for the dummy variable = 0.003). CONCLUSIONS: The incidence rate of BSI was high, and the cumulative risk of developing BSI increased with ICU stay. Further study will clarify if the increased risk of BSI we detected in COVID‐19 patients treated with anti‐inflammatory drugs is outweighed by the benefits of reducing any possible proinflammatory dysregulation induced by SARS‐CoV‐2. In a very few months, coronavirus disease 2019 (COVID-19) has become pandemic, and several countries worldwide are currently dealing with unprecedented epidemic foci of severe acute respiratory infection, a possible presentation of COVID-19 that may require intensive care unit (ICU) admission and carries a high case-fatality rate [1] [2] [3] [4] [5] . While the demographics, clinical characteristics, and overall survival of patients with COVID-19 admitted to ICU have been already extensively characterized by large reports from several parts of the word, little is still known about non-viral infectious complications such as bacterial or fungal bloodstream infections (BSI), that may participate in adversely influencing the outcome of any ICU-admitted patient 6, 7 . In the present study, we aimed to retrospectively assess the incidence rate, cumulative risk, predictors, and survival of ICU-acquired BSI in patients with COVID-19 admitted to two ICUs in a large teaching hospital in Northern Italy, one of the most affected areas in Europe to date 8 . This retrospective study was conducted in two ICUs (27 and 12 beds, respectively) at Ospedale Policlinico San Martino -IRCCS, a 1200-bed teaching hospital in Northern Italy. From February 20 th to April 10 th , 2020, all patients with COVID-19 admitted to the participating ICUs for >48 h were included in the study. The predefined primary study objectives were: (i) to assess the incidence rate of ICU-acquired BSI; (ii) to assess the cumulative risk of developing ICU-acquired BSI. Predefined secondary objectives were: (i) to describe the clinical characteristics of ICU-acquired BSI; (ii) to assess predictors of ICU-acquired BSI; (iii) to describe survival of ICU-acquired BSI. The collection of anonymized data for the present study was approved by the Ethics Committee of the Liguria Region (registry number 163/2020). Specific informed consent was waived due to This article is protected by copyright. All rights reserved the retrospective nature of the study. Reporting of the study conforms to broad EQUATOR guidelines 9 . Definitions COVID-19 was defined in the presence of a positive real-time polymerase-chain-reaction (RT-PCR) for SARS-CoV-2 on at least one respiratory specimen (nasopharyngeal swab, sputum, and/or lower respiratory tract specimens). ICU-acquired BSI was defined in presence of at least one positive blood culture for bacteria or fungi, drawn at >48 h after ICU admission. For coagulase-negative staphylococci and other common skin contaminants, at least two consecutive blood cultures positive for the same pathogen were necessary to define BSI 10 . In patients with multiple blood cultures positive for the same organism, novel BSI events were considered as independent if occurring at least 30 days after the last previous positive blood culture. Polymicrobial infections were considered as separate BSI events, one for each causative organism isolated from blood culture. The following data were collected from the patients' electronic medical records as baseline data at the time of ICU admission: age in years; gender; hypertension; diabetes mellitus; respiratory disease (defined as asthma or chronic obstructive pulmonary disease); end-stage renal disease (defined as estimated glomerular filtration rate <15 mL/min/1.73 m 2 ); moderate to severe liver disease (defined as compensated or decompensated liver cirrhosis); solid cancer; hematological malignancy; human immunodeficiency virus infection; sequential organ failure assessment (SOFA) score at ICU admission 11 ; antibiotic therapy (yes/no and type of administered antibiotic/s). Since they were constantly continued/started at ICU admission, possible off-label antiinflammatory treatments for COVID-19 (steroid treatment with intravenous methylprednisolone at 1 mg/kg once daily and/or intravenous tocilizumab at 8 mg/kg single administration or repeated once) were also recorded as dichotomic baseline variables (steroid yes/no; tocilizumab yes/no). The following data were collected related to the onset of BSI episodes (i.e., they were collected the day when the first positive blood culture was drawn): presence of fever This article is protected by copyright. All rights reserved (defined as temperature > 37.3°C); requirement of vasoactive agents; presence of acute kidney injury (defined as at least stage 1 of KDIGO [Kidney Disease: Improving Global Outcomes] classification of acute kidney injury 12 ); source of BSI (defined according to CDC/NHSN criteria 13 ); blood neutrophil count in cells^10 -3 /mm 3 ; blood platelet count in cells^10 -3 /mm 3 ; serum fibrinogen in g/L; serum lactate in mmol/L, serum C-reactive protein in mg/L; serum procalcitonin in ng/mL; causative agents of BSI and susceptibility test results (the Vitek-2 automated system, bioMérieux, Marcy l'Etoile, France, was used for isolate identification and antimicrobial susceptibility testing). Finally, the date of the following was collected, whichever came first: death in the ICU; discharge from the ICU. No sample size calculations were performed a priori for this exploratory, descriptive study. The incidence rate of ICU-acquired BSI in the study population was calculated as the number of events per 1000 patient-days at risk (defined as the cumulative days of stay elapsed from 48 h after ICU admission to death, discharge from ICU, or end of study period, whichever came first). The 95% confidence interval (CI) for the incidence rate estimate was obtained using the mid-p exact test 14 . The cumulative risk of ICU-acquired BSI was calculated using the Aalen-Johansen method, considering the first occurring BSI as the event of interest, death and discharge from the ICU as competing events, and length of ICU stay equal to 30 days or end of the study period (10 Apr 2020) as right- During the study period, 78 patients with COVID-19 admitted to ICU for >48 h were included in the study. Their median age was 66 years (IQR 57-70), and 77% were males This article is protected by copyright. All rights reserved anti-inflammatory treatment as the reference group; overall p for the dummy variable = 0.002). In multivariable analysis, only anti-inflammatory treatment (csHR 1.07 with 95% CI 0.38-3.04 for tocilizumab, csHR 3.95 with 95% CI 1.20-13.03 for methylprednisolone, and csHR 10.69 with 95% CI 2.71-42.17 for methylprednisolone plus tocilizumab; overall p for the dummy variable = 0.003) retained an independent association with development of BSI. As shown in figure 2 , the estimated survival of ICU-acquired BSI was of almost 75% after 24 days (no longer follow-up was available) from the first positive blood culture, although with the important limitation that follow-up after day 12 was available for less than one third of patients at the time of this report. In the present study, we estimated an incidence rate of 47 episodes of ICU-acquired BSI per 1000 patient-days at risk in critically ill patients with COVID-19, which is higher than that of 5-19 episodes per 1000 patient-days registered in other heterogeneous, non-COVID-19 ICU populations (even if adjusting for the fact that we considered only patientdays after 48 h from ICU admission, see methods) 6, [16] [17] [18] . It is also of note that our estimation of a cumulative 30-day risk of developing BSI of more than 50% is in contrast with the low prevalence of BSI or other bacterial/fungal infections (1-12%) reported in other epidemiological reports from China and the US 2,19,20 . In our opinion, this could be explained by different, non-mutually exclusive reasons. The first is that the number of patients under follow-up in the ICU for more than 20 days was limited in our analysis; thus, any generalization of our 30-day cumulative risk estimate should be made with due caution. On the other hand, the cumulative risk of almost 25% we estimated at day 15 (based on a larger portion of patients still under follow-up) is already higher than the overall prevalence of BSI registered in critically ill COVID-19 patients in earlier studies 2, 19, 20 . In this regard, a possible explanation may be the difficulty of diagnosing BSI in patients receiving anti-inflammatory drugs 21 , which prompted us to collect blood cultures in any case of worsening general conditions in COVID-19 patients, even in the absence of fever and increases in C-reactive protein serum levels. In support of this approach, in the present study, for example, fever was detected only in 38% of patients with ICU-acquired BSI previously treated with tocilizumab. Furthermore, serum This article is protected by copyright. All rights reserved C-reactive protein levels were frequently low and other classical inflammatory markers were usually uncharacteristic, making it rather difficult to clinically recognize a BSI event. Therefore, we feel the total number of BSI in COVID-19 patients may be underestimated wherever anti-inflammatory drugs are administered but collection of blood cultures remains based on classical clinical and laboratory indicators of BSI. Finally, and perhaps most importantly, we found an independent association between receipt of antiinflammatory agents and development of BSI. This effect seems to be mainly driven by steroids rather than tocilizumab, although it is worth noting that the highest instantaneous risk in Cox models was registered in patients receiving both steroid and tocilizumab. In Another different finding from previous reports is that we found a higher prevalence of Gram-positive than Gram-negative organisms as causative agents of ICUacquired BSI in COVID-19 patients 19 . Again, at least two different explanations may be considered. The first is that our sample of BSI events was larger than in previous reports, possibly depicting a more reliable estimate of the true distribution of the relative prevalence of the different causative agents. The second is that, borrowing from the observed microbiological epidemiology (higher prevalence of Gram-positive organisms) in patients with severe influenza 26 , and considering the frequent inability we had in rapidly differentiating between primary viral pneumonia and bacterial pulmonary superinfection in ICU-admitted COVID-19 patients, most COVID-19 patients in our center were treated with an anti-methicillin-resistant S. aureus cephalosporin (most frequently ceftaroline, see table 1) at ICU-admission. However, while on the one hand this may be in line with the unusual high relative prevalence of enterococcal BSI (because of the impaired activity of ceftaroline against Enterococcus spp., although possible in vitro activity against E. faecalis has been reported 27 ), on the other hand it is also true that This article is protected by copyright. All rights reserved coagulase-negative staphylococci and S. aureus were the other two most prevalent causative agents of BSI in our cohort. From this standpoint, conversely, the true relative prevalence of Gram-positive organisms could be even higher than registered in our cohort (because of the anti-staphylococcal and anti-pneumococcal activity of ceftaroline, that could have reduced the prevalence, absolute and relative, of staphylococci and pneumococci as causative agents of BSI). Finally, although we apparently observed an improved survival of BSI in comparison with previous large studies conducted in non-COVID-19 ICU patients 7 , it should be reminded that our sample size of BSI patients was limited (n = 31) and that follow-up was very short. Consequently, any related conclusion should be drawn cautiously. Other important limitations of our study are its retrospective nature (mainly because of possible information and selection biases), its single-center nature that may impact generalization of results, and the lack of adequate power for exploring possible predictors of improved survival in COVID-19 patients with ICU-acquired BSI. In conclusion, in critically ill patients with COVID-19, the incidence rate of ICU- HIV infection 0 (0) Hospital stay before ICU admission in days, median (IQR) 2 (0-5) This article is protected by copyright. All rights reserved This article is protected by copyright. All rights reserved Enterococcus faecium e 4 (9) 2 (13) 1 (8) 1 (14) 0 (0) Viridans group streptococci 3 (7) 0 (0) 1 (8) 0 (0) 2 (20) Pseudomonas aeruginosa g 2 (4) 0 (0) 0 (0) 1 (14) 1 (10) Enterobacter aerogenes h 4 (9) 0 (0) 1 (8) 1 (14) 2 (20) Escherichia coli h 1 (2) 1 (6) 0 (0) 0 (0) 0 (0) This article is protected by copyright. All rights reserved The novel Chinese coronavirus (2019-nCoV) infections: Challenges for fighting the storm Clinical Characteristics of Covid-19 in New York City Baseline Characteristics and Outcomes of 1591 Patients Infected With SARS-CoV-2 Admitted to ICUs of the Lombardy Region Characteristics of and Important Lessons From the Coronavirus Disease 2019 (COVID-19) Outbreak in China: Summary of a Report of 72314 Cases From the Chinese Center for Disease Control and Prevention COVID-19: focus on the lungs but do not forget the gastrointestinal tract Incidence and outcome of invasive candidiasis in intensive care units (ICUs) in Europe: results of the EUCANDICU project Characteristics and determinants of outcome of hospital-acquired bloodstream infections in intensive care units: the EUROBACT International Cohort Study European Centre for Disease Prevention and Control (ECDC) A catalogue of reporting guidelines for health research How to discriminate contamination from bloodstream infection due to coagulase-negative staphylococci: a prospective study with 654 patients The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine Diagnosis, evaluation, and management of acute kidney injury: a KDIGO summary (Part 1) CDC/NHSN surveillance definition of health careassociated infection and criteria for specific types of infections in the acute care setting Epidemiologic analysis with a programmable calculator Accepted Article This article is protected by copyright. All rights reserved 15 An Empirical Transition Matrix for Non-Homogeneous Markov Chains Based on Censored Observations Nosocomial bloodstream infection and clinical sepsis Population-based assessment of intensive care unit-acquired bloodstream infections in adults: Incidence, risk factors, and associated mortality rate Nosocomial bacteremia in a medical-surgical intensive care unit: epidemiologic characteristics and factors influencing mortality in 111 episodes Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia in Wuhan, China: a descriptive study Clinical features of patients infected with 2019 novel coronavirus in Wuhan C reactive protein may not be reliable as a marker of severe bacterial infection in patients receiving tocilizumab Common Infections in Patients Prescribed Systemic Glucocorticoids in Primary Care: A Population-Based Cohort Study Risk of infections in rheumatoid arthritis patients treated with tocilizumab Risk of infectious complications in patients taking glucocorticosteroids Balancing evidence and frontline experience in the early phases of the COVID-19 pandemic: current position of the Italian Society of Anti-Infective