Access Document
Total Page:16
File Type:pdf, Size:1020Kb
BMJ Confidential: For Review Only Diagnostic accuracy of serological tests for COVID-19: a systematic review and meta-analysis Journal: BMJ Manuscript ID BMJ-2020-059052 Article Type: Research BMJ Journal: BMJ Date Submitted by the 30-May-2020 Author: Complete List of Authors: Bastos, Mayara; Research Institute of the McGill University Health Centre, Respiratory Epidemiology and Clinical Research Unit, Centre for Outcomes Research & Evaluation; Rio de Janeiro State University, Social Medicine Institute Tavaziva, Gamuchirai ; Research Institute of the McGill University Health Centre, Respiratory Epidemiology and Clinical Research Unit, Centre for Outcomes Research & Evaluation Abidi, Syed ; Research Institute of the McGill University Health Centre, Respiratory Epidemiology and Clinical Research Unit, Centre for Outcomes Research & Evaluation Campbell, Jonathon; McGill University Health Centre, Respiratory Epidemiology and Clinical Research Unit & Centre for Outcomes Research & Evaluation; McGill University, Departments of Epidemiology, Biostatistics & Occupational Health, and Medicine Haraoui, Louis-Patrick; University of Sherbrooke, 3- Department of Microbiology and Infectious Diseases, Faculty of Medicine and Health Sciences Johnston, James; Centre for Disease Control, Vancouver, British Columbia Lan, Zhiyi; Research Institute of the McGill University Health Centre, Respiratory Epidemiology and Clinical Research Unit, Centre for Outcomes Research & Evaluation Law, Stephanie; Harvard University, Department of Global Health & Social Medicine, Harvard Medical School Maclean, Emily; McGill University, Departments of Epidemiology, Biostatistics & Occupational Health, and Medicine Trajman, Anete; Research Institute of the McGill University Health Centre, Respiratory Epidemiology and Clinical Research Unit, Centre for Outcomes Research & Evaluation; State University of Rio de Janeiro Biomedical Centre, Social Medicine Institute Menzies, Dick; Research Institute of the McGill University Health Centre, Respiratory Epidemiology and Clinical Research Unit, Centre for Outcomes Research & Evaluation; McGill University, Departments of Epidemiology, Biostatistics & Occupational Health, and Medicine Benedetti, Andrea; McGill University, Departments of Epidemiology, Biostatistics & Occupational Health, and Medicine; Research Institute of the McGill University Health Centre, Respiratory Epidemiology and https://mc.manuscriptcentral.com/bmj Page 1 of 51 BMJ 1 2 3 Clinical Research Unit, Centre for Outcomes Research & Evaluation 4 Ahmad Khan, Faiz; Research Institute of the McGill University Health 5 Centre, Respiratory Epidemiology and Clinical Research Unit, Centre for 6 Outcomes Research & Evaluation; McGill University, Departments of 7 Medicine & Epidemiology, Biostatistics & Occupational Health 8 COVID-19, SARS-CoV2, diagnostic accuracy, serological tests, meta- 9 Keywords: analyses, systematic review, point of care 10 Confidential: For Review Only 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 2 of 51 1 2 3 Diagnostic accuracy of serological tests for COVID-19: a systematic review and meta-analysis 4 5 1,2 1 1 6 Mayara Lisboa Bastos, MD, Gamuchirai Tavaziva, MScPH, , Syed Kunal Abidi, BSc, Jonathon R. 7 Campbell, PhD,1,6 Louis-Patrick Haraoui,3 MDCM, James C. Johnston,4 MD, Zhiyi Lan, MSc,1 8 Stephanie Law, PhD,5 Emily MacLean, MSc,6 Anete Trajman, MD,1,2 Dick Menzies, MD,1,6 Andrea 9 Benedetti, PhD,1,6 Faiz Ahmad Khan, MDCM1, 6 10 11 Confidential: For Review Only 12 Affiliations 13 1- Respiratory Epidemiology and Clinical Research Unit, Centre for Outcomes Research & 14 Evaluation, Research Institute of the McGill University Health Centre, Montreal, Canada 15 16 2- Social Medicine Institute, State University of Rio de Janeiro, Rio de Janeiro, Brazil 17 3- Department of Microbiology and Infectious Diseases, Faculty of Medicine and Health 18 Sciences, Université de Sherbrooke, Sherbrooke, Québec, Canada 19 4- University of British Columbia, Vancouver, Canada 20 5- Department of Global Health & Social Medicine, Harvard Medical School, Boston, USA 21 22 6- Departments of Epidemiology, Biostatistics & Occupational Health, and Medicine, McGill 23 University, Montreal, Canada 24 25 Emails 26 Mayara Lisboa Bastos: [email protected] 27 28 Gamuchirai Tavaziva: [email protected] 29 Syed Kunal Abidi: [email protected] 30 Jonathon R. Campbell: [email protected] 31 Louis-Patrick Haraoui: [email protected] 32 James C. Johnston: [email protected] 33 Zhiyi Lan: [email protected] 34 Stephanie Law: [email protected] 35 Emily MacLean: [email protected] 36 37 Anete Trajman: [email protected] 38 Dick Menzies: [email protected] 39 Andrea Bendetti: [email protected] 40 41 Corresponding Author: 42 Faiz Ahmad Khan 43 44 5252 Blvd. de Maisonneuve, Office D3.60 45 Montreal, Quebec, Canada, H4A 3S5 46 [email protected] 47 514-934-1934, extension 32129. Mobile: 514-816-7651 48 49 50 Word count: 51 Manuscript (excluding references & footnotes): 3,311 52 Abstract: 400 words 53 54 55 Key words: COVID-19, Serological tests, diagnostic accuracy, point of care, systematic Review, 56 meta-analysis 57 58 1 59 60 https://mc.manuscriptcentral.com/bmj Page 3 of 51 BMJ 1 2 3 ABSTRACT (400 words) 4 5 6 Objective: To determine the diagnostic accuracy of serological tests (‘sero-diagnostics’) for 7 coronavirus disease-2019 (COVID-19). 8 9 Design: Systematic review and meta-analysis. 10 11 Confidential: For Review Only 12 Data sources: We searched MEDLINE, bioRxiv, and medRxiv, from 1 January to 30 April 2020, 13 using subject headings/subheadings combined with text words for the concepts of COVID-19 14 and sero-diagnostics. 15 16 17 Eligibility criteria and data analysis: Eligible studies measured sensitivity and/or specificity of a 18 COVID-19 sero-diagnostic compared to a reference standard of viral culture or reverse- 19 transcriptase polymerase chain reaction. We excluded studies with fewer than five participants 20 or samples. We assessed risk of bias using QUADAS-2. We used random-effects bivariate meta- 21 22 analyses to estimate pooled sensitivity and specificity. 23 24 Main outcome measure: The primary outcome was overall sensitivity and specificity, stratified 25 by sero-diagnostic method (enzyme linked immunosorbent assays, ELISA; lateral flow 26 27 immunoassays, LFIA; or chemiluminescence immunoassays, CLIA) and immunoglobulin class 28 (IgG, IgM, or both). Secondary outcomes were stratum-specific sensitivity and specificity within 29 sub-groups defined by study or participant characteristics, including time since symptom onset. 30 31 Results: We identified 5016 references and included 39 studies. We performed 48 risk of bias 32 33 assessments (one for each population and method evaluated). We found high risk of patient 34 selection bias in 47/48 (98%), and high or unclear risk of bias from performance or 35 interpretation of the sero-diagnostic in 35/48 (73%). Only 3/39 (8%) studies included 36 outpatients. No tests were evaluated at the point of care. For each sero-diagnostic method, 37 38 pooled sensitivity and specificity were not associated with the immunoglobulin class measured. 39 The pooled sensitivity of ELISAs measuring both immunoglobulins was 84% (95% confidence 40 interval 76% to 91%), of LFIAs, 68% (52% to 81%), and of CLIAs, 98% (46% to 100%). In all 41 analyses, pooled sensitivity was lower for LFIAs, the potential point-of-care method. Pooled 42 specificities were high for all three methods, ranging from 97% to 99%. However, 10165/12360 43 44 (82%) samples used for estimating specificity were from populations tested before the 45 epidemic, or not suspected to have COVID-19. Amongst LFIAs, pooled sensitivity of commercial 46 kits (67%, 51% to 80%) was lower than for non-commercial tests (88%, 83% to 91%). 47 Heterogeneity was seen in all analyses. In the first week of symptoms, pooled sensitivity ranged 48 49 from 13% to 56%, and after the third week from 70% to 98%. 50 51 Conclusion: There is an urgent need for higher quality clinical studies assessing the accuracy of 52 COVID-19 sero-diagnostics. Currently, available evidence does not support the continued use of 53 existing point-of-care sero-diagnostics. 54 55 56 Study registration: PROSPERO CRD42020179452. 57 58 2 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 4 of 51 1 2 3 4 5 6 Summary box 7 What is already known on this topic 8 Tests to detect antibodies (‘sero-diagnostics’) to severe acute respiratory syndrome 9 10 coronavirus-2 (SARS-CoV-2) could improve diagnosis of coronavirus disease 2019 (COVID- 11 19) andConfidential: be useful tools for epidemiologic For surveillance. Review Only 12 The number of sero-diagnostics has rapidly increased, and many are being marketed for 13 point-of-care use, but the evidence base supporting their diagnostic accuracy has not been 14 formally evaluated. 15 16 17 What this study adds 18 19 20 The available evidence on the accuracy of COVID-19 sero-diagnostics is characterized by 21 risks of bias, and heterogeneity, such that estimates of sensitivity and specificity are 22 unreliable and have limited generalizability. The evidence is particularly weak for point-of- 23 care diagnostics, none of which were studied at the point of care. 24 25 Caution is warranted if using COVID-19 sero-diagnostics for clinical decision-making or 26 epidemiologic surveillance. Evidence is insufficient to support the continued use of existing 27 point-of-care sero-diagnostics. 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 3 59 60 https://mc.manuscriptcentral.com/bmj Page 5 of 51 BMJ 1 2 3 INTRODUCTION 4 5 Accurate and rapid diagnostic tests will be critical for achieving control of coronavirus disease 6 2019 (COVID-19), a pandemic illness caused by severe acute respiratory syndrome-coronavirus 7 2 (SARS-CoV-2).