RESUMO
BackgroundWe conducted a systematic review and individual patient data (IPD) meta-analysis to evaluate the diagnostic accuracy of a commercial point-of-care test, the FebriDx lateral flow device (LFD), in adult patients with suspected COVID-19. The FebriDx LFD is designed to distinguish between viral and bacterial respiratory infection. MethodsWe searched MEDLINE, EMBASE, PubMed, Google Scholar, LitCovid, ClinicalTrials.gov and preprint servers on the 13th of January 2021 to identify studies reporting diagnostic accuracy of FebriDx (myxovirus resistance protein A component) versus real time reverse transcriptase polymerase chain reaction (RT-PCR) testing for SARS-CoV-2 in adult patients suspected of COVID-19. IPD were sought from studies meeting the eligibility criteria. Studies were screened for risk of bias using the QUADAS-2 tool. A bivariate linear mixed model was fitted to the data to obtain a pooled estimate of sensitivity and specificity with 95% confidence intervals (95% CIs). A summary receiver operating characteristic (SROC) curve of the model was constructed. A sub-group analysis was performed by meta-regression using the same modelling approach to compare pooled estimates of sensitivity and specificity between patients with a symptom duration of 0 to 7 days and >7 days, and patients aged between 16 to 73 years and >73 years. ResultsTen studies were screened, and three studies with a total of 1481 patients receiving hospital care were included. FebriDx produced an estimated pooled sensitivity of 0.911 (95% CI: 0.855-0.946) and specificity of 0.868 (95% CI: 0.802-0.915) compared to RT-PCR. There were no significant differences between the sub-groups of 0 to 7 days and >7 days in estimated pooled sensitivity (p = 0.473) or specificity (p = 0.853). There were also no significant differences between the sub-groups of 16 to 73 years of age and >73 years of age in estimated pooled sensitivity (p = 0.946) or specificity (p = 0.486). ConclusionsBased on the results of three studies, the FebriDx LFD had high diagnostic accuracy for COVID-19 in a hospital setting, however, the pooled estimates of sensitivity and specificity should be interpreted with caution due to the small number of studies included, risk of bias, and inconsistent reference standards. Further research is required to confirm these findings, and determine how FebriDx would perform in different healthcare settings and patient populations. Trial registrationThis study was conducted at pace as part of the COVID-19 National Diagnostic Research and Evaluation Platform (CONDOR) national test evaluation programme (https://www.condor-platform.org), and as a result, no protocol was developed, and the study was not registered. Lay summaryTests to diagnose COVID-19 are crucial to help control the spread of the disease and to guide treatment. Over the last few months, tests have been developed to diagnose COVID-19 either by detecting the presence of the virus or by detecting specific markers linked to the virus being active in the body. These tests use complex machines in laboratories accepting samples from large geographical areas. Sometimes it takes days for test results to come back. So, to reduce the wait for results, new portable tests are being developed. These point-of-care (POC) tests are designed to work close to where patients require assessment and care such as hospital emergency departments, GP surgeries or care homes. For these new POC tests to be useful, they should ideally be as good as standard laboratory tests. In this study we looked at published research into a new test called FebriDx. FebriDx is a POC test that detects the bodys response to infection, and is claimed to be able to detect the presence of any viral infection, including infections due to the SARS-CoV-2 virus which causes COVID-19, as well as bacterial infections which can have similar symptoms. The FebriDx result was compared with standard laboratory tests for COVID-19 performed on the same patients throat and nose swab sample. We were able to analyse data from three studies with a total of 1481 adult patients who were receiving hospital care with symptoms of COVID-19 during the UK pandemic. Approximately one fifth of the patients were diagnosed as positive for SARS-CoV-2 virus using standard laboratory tests for COVID-19. Our analysis demonstrated that FebriDx correctly identified 91 out of 100 patients who had COVID-19 according to the standard laboratory test. FebriDx also correctly identified 87 out of 100 patients who did not have COVID-19 according to the standard laboratory test. These results have important implications for how these tests could be used. As there were slightly fewer FebriDx false results when the results of the standard laboratory test were positive (9 out of 100) than when the results of the standard laboratory test were negative (13 out of 100), we can have slightly more confidence in a positive test result using FebriDx than a negative FebriDx result. Overall, we have shown that the FebriDx POC test performed well during the UK COVID-19 pandemic when compared with laboratory tests, especially when COVID-19 was indicated. For the future, this means that the FebriDx POC test might be helpful in making a quick clinical decision on whether to isolate a patient with COVID-19-like symptoms arriving in a busy emergency department. However, our results indicate it would not completely replace the need to conduct a laboratory test in certain cases to confirm COVID-19. There are limitations to our findings. For example, we do not know if FebriDx will work in a similar way with patients in different settings such as in the community or care homes. Similarly, we do not know whether other viral and bacterial infections which cause similar COVID-19 symptoms, and are more common in the autumn and winter months, could influence the FebriDx test accuracy. Our findings are also only based on three studies.
RESUMO
STRUCTURED SUMMARYO_ST_ABSBackgroundC_ST_ABSTests for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) viral ribonucleic acid (RNA), using reverse transcription polymerase chain reaction (RT-PCR) are pivotal to detecting current coronavirus disease (COVID-19) and duration of detectable virus indicating potential for infectivity. MethodsWe conducted an individual participant data (IPD) systematic review of longitudinal studies of RT-PCR test results in symptomatic SARS-CoV-2. We searched PubMed, LitCOVID, medRxiv and COVID-19 Living Evidence databases. We assessed risk of bias using a QUADAS- 2 adaptation. Outcomes were the percentage of positive test results by time and the duration of detectable virus, by anatomical sampling sites. FindingsOf 5078 studies screened, we included 32 studies with 1023 SARS-CoV-2 infected participants and 1619 test results, from -6 to 66 days post-symptom onset and hospitalisation. The highest percentage virus detection was from nasopharyngeal sampling between 0 to 4 days post-symptom onset at 89% (95% confidence interval (CI) 83 to 93) dropping to 54% (95% CI 47 to 61) after 10 to 14 days. On average, duration of detectable virus was longer with lower respiratory tract (LRT) sampling than upper respiratory tract (URT). Duration of faecal and respiratory tract virus detection varied greatly within individual participants. In some participants, virus was still detectable at 46 days post- symptom onset. InterpretationRT-PCR misses detection of people with SARS-CoV-2 infection; early sampling minimises false negative diagnoses. Beyond ten days post-symptom onset, lower RT or faecal testing may be preferred sampling sites. The included studies are open to substantial risk of bias so the positivity rates are probably overestimated. PANEL: RESEARCH IN CONTEXTO_ST_ABSEvidence before this studyC_ST_ABSThere are numerous reports of negative severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) reverse transcription polymerase chain reaction (RT-PCR) test results in participants with known SARS-CoV-2 infection, and increasing awareness that the ability of RT-PCR tests to detect virus depends on the timing of sample retrieval and anatomical sampling site. Individual studies suggest that positive test results from RT-PCR with nasopharyngeal sampling declines within a week of symptoms and that a positive test later in the disease course is more likely from sputum, bronchoalveolar lavage (BAL) or stool, but data are inconsistent. Added value of this studyWe searched 5078 titles and abstracts for longitudinal studies reporting individual participant data (IPD) for RT-PCR for participants with COVID-19 linked to either time since symptom onset or time since hospitalisation. Search included SARS-CoV-2 and RT-PCR keywords and MeSH terms. Each included study was subject to careful assessment of risk of bias. This IPD systematic review (SR) addresses RT-PCR test detection rates at different times since symptom onset and hospitalisation for different sampling sites, and summarises the duration of detectable virus. To our knowledge, this is the first rapid SR addressing this topic. We identified 32 studies available as published articles or pre-prints between January 1st and April 24th 2020, including participants sampled at 11 different sampling sites and some participants sampled at more than one site. At earlier time points, nasopharyngeal sampling had the highest virus detection, but the duration of shedding was shorter compared to lower respiratory tract sampling. At 10 to 14 days post-symptom onset, the percentage of positive nasopharyngeal test results was 54% compared to 89% at day 0 to 4. Presence and duration of faecal detection varied by participant, and in nearly half duration was shorter than respiratory sample detection. Virus detection varies for participants and can continue to be detected up to 46 days post-symptom onset or hospitalisation. The included studies were open to substantial risk of bias, so the detection rates are probably overestimates. There was also poor reporting of sampling methods and sparse data on sampling methods that are becoming more widely implemented, such as self-sampling and short nasal swab sampling (anterior nares/mid turbinate). Implications of all the available evidenceResults from this IPD SR of SARS-CoV-2 testing at different time points and using different anatomical sample sites are important to inform strategies of testing. For prevention of ongoing transmission of SARS-CoV-2, samples for RT-PCR testing need to be taken as soon as possible post-symptom onset, as we confirm that RT-PCR misses more people with infection if sampling is delayed. The percentage of positive RT-PCR tests is also highly dependent on the anatomical site sampled in infected people. Sampling at more than one anatomical site may be advisable as there is variation between individuals in the sites that are infected, as well as the timing of SARS-CoV-2 virus detection at an anatomical site. Testing ten days after symptom onset will lead to a higher frequency of negative tests, particularly if using only upper respiratory tract sampling. However, our estimates may considerably understate the frequency of negative RT-PCR results in people with SARS-CoV- 2 infection. Further investment in this IPD approach is recommended as the amount data available was small given the scale of the pandemic and the importance of the question. More studies, learning from our observations about risk of bias and strengths of example studies (Box 1, Box 2) are urgently needed to inform the optimal sampling strategy by including self-collected samples such as saliva and short nasal swabs. Better reporting of anatomical sampling sites with a detailed methodology on sample collection is also urgently needed.
