GRADE guidance 37: rating imprecision in a body of evidence on test accuracy.

OBJECTIVES: To provide guidance on rating imprecision in a body of evidence assessing the accuracy of a single test. This guide will clarify when Grading of Recommendations Assessment, Development and Evaluation (GRADE) users should consider rating down the certainty of evidence by one or more levels for imprecision in test accuracy. STUDY DESIGN AND SETTING: A project group within the GRADE working group conducted iterative discussions and presentations at GRADE working group meetings to produce this guidance. RESULTS: Before rating the certainty of evidence, GRADE users should define the target of their certainty rating. GRADE recommends setting judgment thresholds defining what they consider a very accurate, accurate, inaccurate, and very inaccurate test. These thresholds should be set after considering consequences of testing and effects on people-important outcomes. GRADE's primary criterion for judging imprecision in test accuracy evidence is considering confidence intervals (i.e., CI approach) of absolute test accuracy results (true and false, positive, and negative results in a cohort of people). Based on the CI approach, when a CI appreciably crosses the predefined judgment threshold(s), one should consider rating down certainty of evidence by one or more levels, depending on the number of thresholds crossed. When the CI does not cross judgment threshold(s), GRADE suggests considering the sample size for an adequately powered test accuracy review (optimal or review information size [optimal information size (OIS)/review information size (RIS)]) in rating imprecision. If the combined sample size of the included studies in the review is smaller than the required OIS/RIS, one should consider rating down by one or more levels for imprecision. CONCLUSION: This paper extends previous GRADE guidance for rating imprecision in single test accuracy systematic reviews and guidelines, with a focus on the circumstances in which one should consider rating down one or more levels for imprecision.

Estimated causal effects of common respiratory infections on cardiovascular risk: a meta-analysis.

OBJECTIVE: Literature supports associations between common respiratory tract infections (RTIs) and risk of cardiovascular diseases, yet the importance of RTIs for cardiovascular risk management remains less understood. This systematic review and meta-analysis aimed to estimate the causal effects of RTIs on occurrence of cardiovascular diseases in the general population. METHODS: MEDLINE and EMBASE were systematically searched up to 4 November 2022. Eligible were all aetiological studies evaluating risk of cardiovascular outcomes after exposure to common RTIs within any follow-up duration. Evidence was pooled using random-effects models if data allowed. The ROBINS-E and GRADE approaches were used to rate risk of bias and certainty of evidence, respectively. All assessments were performed in duplicate. RESULTS: We included 34 studies (65 678 650 individuals). Most studies had a high risk of bias. COVID-19 likely increases relative risk (RR (95% CI)) of myocardial infarction (3.3 (1.0 to 11.0)), stroke (3.5 (1.2 to 10)), pulmonary embolism (24.6 (13.5 to 44.9)) and deep venous thrombosis (7.8 (4.3 to 14.4)) within 30 days after infection (GRADE: moderate) and about twofold within 1 year (GRADE: low to moderate). Other RTIs also likely increase the RR of myocardial infarction (2.9 (95% CI 1.8 to 4.9)) and stroke (2.6 (95% CI 1.1 to 6.4)) within 30 days (GRADE: moderate), and to a lesser extent with longer follow-up. CONCLUSIONS: RTIs likely increase the risk of cardiovascular diseases about 1.5-5 fold within 1 month after infection. RTIs may, therefore, have clinical relevance as target for cardiovascular risk management, especially in high-risk populations. PROSPERO REGISTRATION NUMBER: CRD42023416277.

How to Critically Appraise and Interpret Systematic Reviews and Meta-Analyses of Diagnostic Accuracy: A User Guide.

Systematic reviews of diagnostic accuracy studies can provide the best available evidence to inform decisions regarding the use of a diagnostic test. In this guide, the authors provide a practical approach for clinicians to appraise diagnostic accuracy systematic reviews and apply their results to patient care. The first step is to identify an appropriate systematic review with a research question matching the clinical scenario. The user should evaluate the rigor of the review methods to evaluate its credibility (Did the review use clearly defined eligibility criteria, a comprehensive search strategy, structured data collection, risk of bias and applicability appraisal, and appropriate meta-analysis methods?). If the review is credible, the next step is to decide whether the diagnostic performance is adequate for clinical use (Do sensitivity and specificity estimates exceed the threshold that makes them useful in clinical practice? Are these estimates sufficiently precise? Is variability in the estimates of diagnostic accuracy across studies explained?). Diagnostic accuracy systematic reviews that are judged to be credible and provide diagnostic accuracy estimates with sufficient certainty and relevance are the most useful to inform patient care. This review discusses comparative, noncomparative, and emerging approaches to systematic reviews of diagnostic accuracy using a clinical scenario and examples based on recent publications.

Symptom- and chest-radiography screening for active pulmonary tuberculosis in HIV-negative adults and adults with unknown HIV status.

BACKGROUND: Systematic screening in high-burden settings is recommended as a strategy for early detection of pulmonary tuberculosis disease, reducing mortality, morbidity and transmission, and improving equity in access to care. Questioning for symptoms and chest radiography (CXR) have historically been the most widely available tools to screen for tuberculosis disease. Their accuracy is important for the design of tuberculosis screening programmes and determines, in combination with the accuracy of confirmatory diagnostic tests, the yield of a screening programme and the burden on individuals and the health service. OBJECTIVES: To assess the sensitivity and specificity of questioning for the presence of one or more tuberculosis symptoms or symptom combinations, CXR, and combinations of these as screening tools for detecting bacteriologically confirmed pulmonary tuberculosis disease in HIV-negative adults and adults with unknown HIV status who are considered eligible for systematic screening for tuberculosis disease. Second, to investigate sources of heterogeneity, especially in relation to regional, epidemiological, and demographic characteristics of the study populations. SEARCH METHODS: We searched the MEDLINE, Embase, LILACS, and HTA (Health Technology Assessment) databases using pre-specified search terms and consulted experts for unpublished reports, for the period 1992 to 2018. The search date was 10 December 2018. This search was repeated on 2 July 2021. SELECTION CRITERIA: Studies were eligible if participants were screened for tuberculosis disease using symptom questions, or abnormalities on CXR, or both, and were offered confirmatory testing with a reference standard. We included studies if diagnostic two-by-two tables could be generated for one or more index tests, even if not all participants were subjected to a microbacteriological reference standard. We excluded studies evaluating self-reporting of symptoms. DATA COLLECTION AND ANALYSIS: We categorized symptom and CXR index tests according to commonly used definitions. We assessed the methodological quality of included studies using the QUADAS-2 instrument. We examined the forest plots and receiver operating characteristic plots visually for heterogeneity. We estimated summary sensitivities and specificities (and 95% confidence intervals (CI)) for each index test using bivariate random-effects methods. We analyzed potential sources of heterogeneity in a hierarchical mixed-model. MAIN RESULTS: The electronic database search identified 9473 titles and abstracts. Through expert consultation, we identified 31 reports on national tuberculosis prevalence surveys as eligible (of which eight were already captured in the search of the electronic databases), and we identified 957 potentially relevant articles through reference checking. After removal of duplicates, we assessed 10,415 titles and abstracts, of which we identified 430 (4%) for full text review, whereafter we excluded 364 articles. In total, 66 articles provided data on 59 studies. We assessed the 2 July 2021 search results; seven studies were potentially eligible but would make no material difference to the review findings or grading of the evidence, and were not added in this edition of the review. We judged most studies at high risk of bias in one or more domains, most commonly because of incorporation bias and verification bias. We judged applicability concerns low in more than 80% of studies in all three domains. The three most common symptom index tests, cough for two or more weeks (41 studies), any cough (21 studies), and any tuberculosis symptom (29 studies), showed a summary sensitivity of 42.1% (95% CI 36.6% to 47.7%), 51.3% (95% CI 42.8% to 59.7%), and 70.6% (95% CI 61.7% to 78.2%, all very low-certainty evidence), and a specificity of 94.4% (95% CI 92.6% to 95.8%, high-certainty evidence), 87.6% (95% CI 81.6% to 91.8%, low-certainty evidence), and 65.1% (95% CI 53.3% to 75.4%, low-certainty evidence), respectively. The data on symptom index tests were more heterogenous than those for CXR. The studies on any tuberculosis symptom were the most heterogeneous, but had the lowest number of variables explaining this variation. Symptom index tests also showed regional variation. The summary sensitivity of any CXR abnormality (23 studies) was 94.7% (95% CI 92.2% to 96.4%, very low-certainty evidence) and 84.8% (95% CI 76.7% to 90.4%, low-certainty evidence) for CXR abnormalities suggestive of tuberculosis (19 studies), and specificity was 89.1% (95% CI 85.6% to 91.8%, low-certainty evidence) and 95.6% (95% CI 92.6% to 97.4%, high-certainty evidence), respectively. Sensitivity was more heterogenous than specificity, and could be explained by regional variation. The addition of cough for two or more weeks, whether to any (pulmonary) CXR abnormality or to CXR abnormalities suggestive of tuberculosis, resulted in a summary sensitivity and specificity of 99.2% (95% CI 96.8% to 99.8%) and 84.9% (95% CI 81.2% to 88.1%) (15 studies; certainty of evidence not assessed). AUTHORS' CONCLUSIONS: The summary estimates of the symptom and CXR index tests may inform the choice of screening and diagnostic algorithms in any given setting or country where screening for tuberculosis is being implemented. The high sensitivity of CXR index tests, with or without symptom questions in parallel, suggests a high yield of persons with tuberculosis disease. However, additional considerations will determine the design of screening and diagnostic algorithms, such as the availability and accessibility of CXR facilities or the resources to fund them, and the need for more or fewer diagnostic tests to confirm the diagnosis (depending on screening test specificity), which also has resource implications. These review findings should be interpreted with caution due to methodological limitations in the included studies and regional variation in sensitivity and specificity. The sensitivity and specificity of an index test in a specific setting cannot be predicted with great precision due to heterogeneity. This should be borne in mind when planning for and implementing tuberculosis screening programmes.

QUADAS-C: A Tool for Assessing Risk of Bias in Comparative Diagnostic Accuracy Studies.

Comparative diagnostic test accuracy studies assess and compare the accuracy of 2 or more tests in the same study. Although these studies have the potential to yield reliable evidence regarding comparative accuracy, shortcomings in the design, conduct, and analysis may bias their results. The currently recommended quality assessment tool for diagnostic test accuracy studies, QUADAS-2 (Quality Assessment of Diagnostic Accuracy Studies-2), is not designed for the assessment of test comparisons. The QUADAS-C (Quality Assessment of Diagnostic Accuracy Studies-Comparative) tool was developed as an extension of QUADAS-2 to assess the risk of bias in comparative diagnostic test accuracy studies. Through a 4-round Delphi study involving 24 international experts in test evaluation and a face-to-face consensus meeting, an initial version of the tool was developed that was revised and finalized following a pilot study among potential users. The QUADAS-C tool retains the same 4-domain structure of QUADAS-2 (Patient Selection, Index Test, Reference Standard, and Flow and Timing) and comprises additional questions to each QUADAS-2 domain. A risk-of-bias judgment for comparative accuracy requires a risk-of-bias judgment for the accuracy of each test (resulting from QUADAS-2) and additional criteria specific to test comparisons. Examples of such additional criteria include whether participants either received all index tests or were randomly assigned to index tests, and whether index tests were interpreted with blinding to the results of other index tests. The QUADAS-C tool will be useful for systematic reviews of diagnostic test accuracy addressing comparative questions. Furthermore, researchers may use this tool to identify and avoid risk of bias when designing a comparative diagnostic test accuracy study.

Study designs for comparative diagnostic test accuracy: A methodological review and classification scheme.

OBJECTIVES: (1) To identify and classify comparative diagnostic test accuracy (DTA) study designs; (2) to describe study design labels used by authors of comparative DTA studies. METHODS: We performed a methodological review of 100 comparative DTA studies published between 2015 and 2017, randomly sampled from studies included in 238 comparative DTA systematic reviews indexed in MEDLINE in 2017. From each study report, we extracted six design elements characterizing participant flow and the labels used by authors. RESULTS: We identified a total of 46 unique combinations of study design features in our sample, based on six design elements characterizing participant flow. We classified the studies into five study design categories based on how participants were allocated to receive each index test: 'fully paired' (n=79), 'partially paired, random subset' (n=0), 'partially paired, nonrandom subset' (n=2), 'unpaired randomized' (n=1) and 'unpaired nonrandomized' (n=3). The allocation method used in 15 studies was unclear. Sixty-one studies reported, in total, 29 unique study design labels but only four labels referred to specific design features of comparative studies. CONCLUSION: Our classification scheme can help systematic review authors define study eligibility criteria, assess risk of bias, and communicate the strength of the evidence. A standardized labelling scheme could be developed to facilitate communication of specific design features.

GRADE Guidance: 31. Assessing the certainty across a body of evidence for comparative test accuracy.

OBJECTIVES: This article provides GRADE guidance on how authors of evidence syntheses and health decision makers, including guideline developers, can rate the certainty across a body of evidence for comparative test accuracy questions. STUDY DESIGN AND SETTING: This guidance extends the previously published GRADE guidance for assessing certainty of evidence for test accuracy to scenarios in which two or more index tests are compared. Through an iterative brainstorm-discussion-feedback process within the GRADE working group, we developed a guidance accompanied by practical examples. RESULTS: Rating the certainty of evidence for comparative test accuracy shares many concepts and ideas with the existing GRADE guidance for test accuracy. The rating in comparisons of test accuracy requires additional considerations, such as the selection of appropriate comparative study designs, additional criteria for judging risk of bias, and the consequences of using comparative measures of test accuracy. Distinct approaches to rating certainty are required for comparative test accuracy studies and between-study (indirect) comparisons. CONCLUSION: This GRADE guidance will support transparent assessment of the certainty for a body of comparative test accuracy evidence.

A randomized trial of an editorial intervention to reduce spin in the abstract's conclusion of manuscripts showed no significant effect.

OBJECTIVE: To estimate the effect of an intervention compared to the usual peer-review process on reducing spin in the abstract's conclusion of biomedical study reports. STUDY DESIGN AND SETTING: We conducted a two-arm, parallel-group RCT in a sample of primary research manuscripts submitted to BMJ Open. The authors received short instructions alongside the peer reviewers' comments in the intervention group. We assessed the presence of spin (primary outcome), types of spin, and wording change in the revised abstract's conclusion. Outcome assessors were blinded to the intervention assignment. RESULTS: Of the 184 manuscripts randomized, 108 (54 intervention, 54 control) were selected for revision and could be evaluated for the presence of spin. The proportion of manuscripts with spin was 6% lower (95% CI: 24% lower to 13% higher) in the intervention group (57%, 31/54) than in the control group (63%, 34/54). The wording of the revised abstract's conclusion was changed in 34/54 (63%) manuscripts in the intervention group and 26/54 (48%) in the control group. The four prespecified types of spin involved (i) selective reporting (12 in the intervention group vs. 8 in the control group), (ii) including information not supported by evidence (9 vs. 9), and (iii) interpretation not consistent with the study results (14 vs. 18), and (iv) unjustified recommendations for practice (5 vs. 11). CONCLUSION: These short instructions to authors did not have a statistically significant effect on reducing spin in revised abstract conclusions, and based on the confidence interval, the existence of a large effect can be excluded. Other interventions to reduce spin in reports of original research should be evaluated. STUDY REGISTRATION: osf.io/xnuyt.

Reporting of test comparisons in diagnostic accuracy studies: A literature review.

Comparative accuracy studies evaluate the relative performance of two or more diagnostic tests. As any other form of research, such studies should be reported in an informative manner, to allow replication and to be useful for decision-making. In this study we aimed to assess whether and how components of test comparisons were reported in comparative accuracy studies. We evaluated 100 comparative accuracy studies, published in 2015, 2016 or 2017, randomly sampled from 238 comparative accuracy systematic reviews. We extracted information on 20 reporting items, pertaining to the identification of the test comparison, its validity, and the actual results of the comparison. About a third of the studies (n = 36) did not report the comparison as a study objective or hypothesis. Although most studies (n = 86) reported how participants had been allocated to index tests, we could often not evaluate whether test interpreters had been blinded to the results of other index tests (n = 40; among 59 applicable studies), nor could we identify the sequence of index tests (n = 52; among 90 applicable studies) or the methods for comparing measures of accuracy (n = 59). Two-by-four table data (revealing the agreement between index tests) were only reported by 9 of 90 paired comparative studies. More than half of the studies (n = 64) did not provide measures of statistical imprecision for comparative accuracy. Our findings suggest that components of test comparisons are frequently missing or incompletely described in comparative accuracy studies included in systematic reviews. Explicit guidance for reporting comparative accuracy studies may facilitate the production of full and informative study reports.

Routine laboratory testing to determine if a patient has COVID-19.

BACKGROUND: Specific diagnostic tests to detect severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and resulting COVID-19 disease are not always available and take time to obtain results. Routine laboratory markers such as white blood cell count, measures of anticoagulation, C-reactive protein (CRP) and procalcitonin, are used to assess the clinical status of a patient. These laboratory tests may be useful for the triage of people with potential COVID-19 to prioritize them for different levels of treatment, especially in situations where time and resources are limited. OBJECTIVES: To assess the diagnostic accuracy of routine laboratory testing as a triage test to determine if a person has COVID-19. SEARCH METHODS: On 4 May 2020 we undertook electronic searches in the Cochrane COVID-19 Study Register and the COVID-19 Living Evidence Database from the University of Bern, which is updated daily with published articles from PubMed and Embase and with preprints from medRxiv and bioRxiv. In addition, we checked repositories of COVID-19 publications. We did not apply any language restrictions. SELECTION CRITERIA: We included both case-control designs and consecutive series of patients that assessed the diagnostic accuracy of routine laboratory testing as a triage test to determine if a person has COVID-19. The reference standard could be reverse transcriptase polymerase chain reaction (RT-PCR) alone; RT-PCR plus clinical expertise or and imaging; repeated RT-PCR several days apart or from different samples; WHO and other case definitions; and any other reference standard used by the study authors. DATA COLLECTION AND ANALYSIS: Two review authors independently extracted data from each included study. They also assessed the methodological quality of the studies, using QUADAS-2. We used the 'NLMIXED' procedure in SAS 9.4 for the hierarchical summary receiver operating characteristic (HSROC) meta-analyses of tests for which we included four or more studies. To facilitate interpretation of results, for each meta-analysis we estimated summary sensitivity at the points on the SROC curve that corresponded to the median and interquartile range boundaries of specificities in the included studies. MAIN RESULTS: We included 21 studies in this review, including 14,126 COVID-19 patients and 56,585 non-COVID-19 patients in total. Studies evaluated a total of 67 different laboratory tests. Although we were interested in the diagnotic accuracy of routine tests for COVID-19, the included studies used detection of SARS-CoV-2 infection through RT-PCR as reference standard. There was considerable heterogeneity between tests, threshold values and the settings in which they were applied. For some tests a positive result was defined as a decrease compared to normal vaues, for other tests a positive result was defined as an increase, and for some tests both increase and decrease may have indicated test positivity. None of the studies had either low risk of bias on all domains or low concerns for applicability for all domains. Only three of the tests evaluated had a summary sensitivity and specificity over 50%. These were: increase in interleukin-6, increase in C-reactive protein and lymphocyte count decrease. Blood count Eleven studies evaluated a decrease in white blood cell count, with a median specificity of 93% and a summary sensitivity of 25% (95% CI 8.0% to 27%; very low-certainty evidence). The 15 studies that evaluated an increase in white blood cell count had a lower median specificity and a lower corresponding sensitivity. Four studies evaluated a decrease in neutrophil count. Their median specificity was 93%, corresponding to a summary sensitivity of 10% (95% CI 1.0% to 56%; low-certainty evidence). The 11 studies that evaluated an increase in neutrophil count had a lower median specificity and a lower corresponding sensitivity. The summary sensitivity of an increase in neutrophil percentage (4 studies) was 59% (95% CI 1.0% to 100%) at median specificity (38%; very low-certainty evidence). The summary sensitivity of an increase in monocyte count (4 studies) was 13% (95% CI 6.0% to 26%) at median specificity (73%; very low-certainty evidence). The summary sensitivity of a decrease in lymphocyte count (13 studies) was 64% (95% CI 28% to 89%) at median specificity (53%; low-certainty evidence). Four studies that evaluated a decrease in lymphocyte percentage showed a lower median specificity and lower corresponding sensitivity. The summary sensitivity of a decrease in platelets (4 studies) was 19% (95% CI 10% to 32%) at median specificity (88%; low-certainty evidence). Liver function tests The summary sensitivity of an increase in alanine aminotransferase (9 studies) was 12% (95% CI 3% to 34%) at median specificity (92%; low-certainty evidence). The summary sensitivity of an increase in aspartate aminotransferase (7 studies) was 29% (95% CI 17% to 45%) at median specificity (81%) (low-certainty evidence). The summary sensitivity of a decrease in albumin (4 studies) was 21% (95% CI 3% to 67%) at median specificity (66%; low-certainty evidence). The summary sensitivity of an increase in total bilirubin (4 studies) was 12% (95% CI 3.0% to 34%) at median specificity (92%; very low-certainty evidence). Markers of inflammation The summary sensitivity of an increase in CRP (14 studies) was 66% (95% CI 55% to 75%) at median specificity (44%; very low-certainty evidence). The summary sensitivity of an increase in procalcitonin (6 studies) was 3% (95% CI 1% to 19%) at median specificity (86%; very low-certainty evidence). The summary sensitivity of an increase in IL-6 (four studies) was 73% (95% CI 36% to 93%) at median specificity (58%) (very low-certainty evidence). Other biomarkers The summary sensitivity of an increase in creatine kinase (5 studies) was 11% (95% CI 6% to 19%) at median specificity (94%) (low-certainty evidence). The summary sensitivity of an increase in serum creatinine (four studies) was 7% (95% CI 1% to 37%) at median specificity (91%; low-certainty evidence). The summary sensitivity of an increase in lactate dehydrogenase (4 studies) was 25% (95% CI 15% to 38%) at median specificity (72%; very low-certainty evidence). AUTHORS' CONCLUSIONS: Although these tests give an indication about the general health status of patients and some tests may be specific indicators for inflammatory processes, none of the tests we investigated are useful for accurately ruling in or ruling out COVID-19 on their own. Studies were done in specific hospitalized populations, and future studies should consider non-hospital settings to evaluate how these tests would perform in people with milder symptoms.

Risk of bias assessment of test comparisons was uncommon in comparative accuracy systematic reviews: an overview of reviews.

OBJECTIVES: Comparative diagnostic test accuracy systematic reviews (DTA reviews) assess the accuracy of two or more tests and compare their diagnostic performance. We investigated how comparative DTA reviews assessed the risk of bias (RoB) in primary studies that compared multiple index tests. STUDY DESIGN AND SETTING: This is an overview of comparative DTA reviews indexed in MEDLINE from January 1st to December 31st, 2017. Two assessors independently identified DTA reviews including at least two index tests and containing at least one statement in which the accuracy of the index tests was compared. Two assessors independently extracted data on the methods used to assess RoB in studies that directly compared the accuracy of multiple index tests. RESULTS: We included 238 comparative DTA reviews. Only two reviews (0.8%, 95% confidence interval 0.1 to 3.0%) conducted RoB assessment of test comparisons undertaken in primary studies; neither used an RoB tool specifically designed to assess bias in test comparisons. CONCLUSION: Assessment of RoB in test comparisons undertaken in primary studies was uncommon in comparative DTA reviews, possibly due to lack of existing guidance on and awareness of potential sources of bias. Based on our findings, guidance on how to assess and incorporate RoB in comparative DTA reviews is needed.

Nucleic acid and antigen detection tests for leptospirosis.

BACKGROUND: Early diagnosis of leptospirosis may contribute to the effectiveness of antimicrobial therapy and early outbreak recognition. Nucleic acid and antigen detection tests have the potential for early diagnosis of leptospirosis. With this systematic review, we assessed the sensitivity and specificity of nucleic acid and antigen detection tests. OBJECTIVES: To determine the diagnostic test accuracy of nucleic acid and antigen detection tests for the diagnosis of human symptomatic leptospirosis. SEARCH METHODS: We searched electronic databases including MEDLINE, Embase, the Cochrane Library, and regional databases from inception to 6 July 2018. We did not apply restrictions to language or time of publication. SELECTION CRITERIA: We included diagnostic cross-sectional studies and case-control studies of tests that made use of nucleic acid and antigen detection methods in people suspected of systemic leptospirosis. As reference standards, we considered the microscopic agglutination test alone (which detects antibodies against leptospirosis) or in a composite reference standard with culturing or other serological tests. Studies were excluded when the controls were healthy individuals or when there were insufficient data to calculate sensitivity and specificity. DATA COLLECTION AND ANALYSIS: At least two review authors independently extracted data from each study. We used the revised Quality Assessment of Diagnostic Accuracy Studies tool (QUADAS-2) to assess risk of bias. We calculated study-specific values for sensitivity and specificity with 95% confidence intervals (CI) and pooled the results in a meta-analysis when appropriate. We used the bivariate model for index tests with one positivity threshold, and we used the hierarchical summary receiver operating characteristic model for index tests with multiple positivity thresholds. As possible sources of heterogeneity, we explored: timing of index test, disease prevalence, blood sample type, primers or target genes, and the real-time polymerase chain reaction (PCR) visualisation method. These were added as covariates to the meta-regression models. MAIN RESULTS: We included 41 studies evaluating nine index tests (conventional PCR (in short: PCR), real-time PCR, nested PCR, PCR performed twice, loop-mediated isothermal amplification, enzyme-linked immunosorbent assay (ELISA), dot-ELISA, immunochromatography-based lateral flow assay, and dipstick assay) with 5981 participants (1834 with and 4147 without leptospirosis). Methodological quality criteria were often not reported, and the risk of bias of the reference standard was generally considered high. The applicability of findings was limited by the frequent use of frozen samples. We conducted meta-analyses for the PCR and the real-time PCR on blood products.The pooled sensitivity of the PCR was 70% (95% CI 37% to 90%) and the pooled specificity was 95% (95% CI 75% to 99%). When studies with a high risk of bias in the reference standard domain were excluded, the pooled sensitivity was 87% (95% CI 44% to 98%) and the pooled specificity was 97% (95% CI 60% to 100%). For the real-time PCR, we estimated a summary receiver operating characteristic curve. To illustrate, a point on the curve with 85% specificity had a sensitivity of 49% (95% CI 30% to 68%). Likewise, at 90% specificity, sensitivity was 40% (95% CI 24% to 59%) and at 95% specificity, sensitivity was 29% (95% CI 15% to 49%). The median specificity of real-time PCR on blood products was 92%. We did not formally compare the diagnostic test accuracy of PCR and real-time PCR, as direct comparison studies were lacking. Three of 15 studies analysing PCR on blood products reported the timing of sample collection in the studies included in the meta-analyses (range 1 to 7 days postonset of symptoms), and nine out of 16 studies analysing real-time PCR on blood products (range 1 to 19 days postonset of symptoms). In PCR studies, specificity was lower in settings with high leptospirosis prevalence. Other investigations of heterogeneity did not identify statistically significant associations. Two studies suggested that PCR and real-time PCR may be more sensitive on blood samples collected early in the disease stage. Results of other index tests were described narratively. AUTHORS' CONCLUSIONS: The validity of review findings are limited and should be interpreted with caution. There is a substantial between-study variability in the accuracy of PCR and real-time PCR, as well as a substantial variability in the prevalence of leptospirosis. Consequently, the position of PCR and real-time PCR in the clinical pathway depends on regional considerations such as disease prevalence, factors that are likely to influence accuracy, and downstream consequences of test results. There is insufficient evidence to conclude which of the nucleic acid and antigen detection tests is the most accurate. There is preliminary evidence that PCR and real-time PCR are more sensitive on blood samples collected early in the disease stage, but this needs to be confirmed in future studies.