Laboratory-based molecular test alternatives to RT-PCR for the diagnosis of SARS-CoV-2 infection
Ingrid Arevalo-Rodriguez,Miriam Mateos-Haro,Jacqueline Dinnes,Agustín Ciapponi,Clare Davenport,Diana Buitrago-Garcia,Tayeb Bennouna-Dalero,Marta Roqué-Figuls,Ann Van den Bruel,Karin J von Eije,Devy Emperador,Lotty Hooft,René Spijker,Mariska MG Leeflang,Yemisi Takwoingi,Jonathan J Deeks
DOI: https://doi.org/10.1002/14651858.cd015618
IF: 8.4
2024-10-15
Cochrane Database of Systematic Reviews
Abstract:Diagnosing people with a SARS‐CoV‐2 infection played a critical role in managing the COVID‐19 pandemic and remains a priority for the transition to long‐term management of COVID‐19. Initial shortages of extraction and reverse transcription polymerase chain reaction (RT‐PCR) reagents impaired the desired upscaling of testing in many countries, which led to the search for alternatives to RNA extraction/purification and RT‐PCR testing. Reference standard methods for diagnosing the presence of SARS‐CoV‐2 infection rely primarily on real‐time reverse transcription‐polymerase chain reaction (RT‐PCR). Alternatives to RT‐PCR could, if sufficiently accurate, have a positive impact by expanding the range of diagnostic tools available for the timely identification of people infected by SARS‐CoV‐2, access to testing and the use of resources. To assess the diagnostic accuracy of alternative (to RT‐PCR assays) laboratory‐based molecular tests for diagnosing SARS‐CoV‐2 infection. We searched the COVID‐19 Open Access Project living evidence database from the University of Bern until 30 September 2020 and the WHO COVID‐19 Research Database until 31 October 2022. We did not apply language restrictions. We included studies of people with suspected or known SARS‐CoV‐2 infection, or where tests were used to screen for infection, and studies evaluating commercially developed laboratory‐based molecular tests for the diagnosis of SARS‐CoV‐2 infection considered as alternatives to RT‐PCR testing. We also included all reference standards to define the presence or absence of SARS‐CoV‐2, including RT‐PCR tests and established clinical diagnostic criteria. Two authors independently screened studies and resolved disagreements by discussing them with a third author. Two authors independently extracted data and assessed the risk of bias and applicability of the studies using the QUADAS‐2 tool. We presented sensitivity and specificity, with 95% confidence intervals (CIs), for each test using paired forest plots and summarised results using average sensitivity and specificity using a bivariate random‐effects meta‐analysis. We illustrated the findings per index test category and assay brand compared to the WHO's acceptable sensitivity and specificity threshold for diagnosing SARS‐CoV‐2 infection using nucleic acid tests. We included data from 64 studies reporting 94 cohorts of participants and 105 index test evaluations, with 74,753 samples and 7517 confirmed SARS‐CoV‐2 cases. We did not identify any published or preprint reports of accuracy for a considerable number of commercially produced NAAT assays. Most cohorts were judged at unclear or high risk of bias in more than three QUADAS‐2 domains. Around half of the cohorts were considered at high risk of selection bias because of recruitment based on COVID status. Three quarters of 94 cohorts were at high risk of bias in the reference standard domain because of reliance on a single RT‐PCR result to determine the absence of SARS‐CoV‐2 infection or were at unclear risk of bias due to a lack of clarity about the time interval between the index test assessment and the reference standard, the number of missing results, or the absence of a participant flow diagram. For index tests categories with four or more evaluations and when summary estimations were possible, we found that: a) For RT‐PCR assays designed to omit/adapt RNA extraction/purification, the average sensitivity was 95.1% (95% CI 91.1% to 97.3%), and the average specificity was 99.7% (95% CI 98.5% to 99.9%; based on 27 evaluations, 2834 samples and 1178 SARS‐CoV‐2 cases); b) For RT‐LAMP assays, the average sensitivity was 88.4% (95% CI 83.1% to 92.2%), and the average specificity was 99.7% (95% CI 98.7% to 99.9%; 24 evaluations, 29,496 samples and 2255 SARS‐CoV‐2 cases); c) for TMA assays, the average sensitivity was 97.6% (95% CI 95.2% to 98.8%), and the average specificity was 99.4% (95% CI 94.9% to 99.9%; 14 evaluations, 2196 samples and 942 SARS‐CoV‐2 cases); d) for digital PCR assays, the average sensitivity was 98.5% (95% CI 95.2% to 99.5%), and the average specificity was 91.4% (95% CI 60.4% to 98.7%; five evaluations, 703 samples and 354 SARS‐CoV‐2 cases); e) for RT‐LAMP assays omitting/adapting RNA extraction, the average sensitivity was 73.1% (95% CI 58.4% to 84%), and the average specificity was 100% (95% CI 98% to 100%; 24 evaluations, 14,342 samples and 1502 SARS‐CoV‐2 cases). Only two index test categories fulfil the WHO‐acceptable sensitivity and specificity requirements for SARS‐CoV‐2 nucleic acid tests: RT‐PCR assays designed to omit/adapt RNA extraction/purification and TMA assays. In addition, WHO‐acceptable performance criteria were met for two assays out of 35 when tests were used according to manufacturer instructions. At 5% prevalence using a cohort of 1000 people suspected of SARS‐CoV‐2 infection, the positive predictive value of RT‐PCR assays omitting/ada -Abstract Truncated-
medicine, general & internal