Sensitivity of RT-PCR testing of upper respiratory tract samples for SARS-CoV-2 in hospitalised patients: a retrospective cohort study

  1. Thomas C. Williams
  2. Elizabeth Wastnedge
  3. Gina McAllister
  4. Ramya Bhatia
  5. Kate Cuschieri
  6. Kallirroi Kefala
  7. Fiona Hamilton
  8. Ingólfur Johannessen
  9. Ian F. Laurenson
  10. Jill Shepherd
  11. Alistair Stewart
  12. Donald Waters
  13. Helen Wise
  14. Kate E. Templeton

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Background: This study aimed to determine the sensitivity and specificity of reverse transcription PCR (RT-PCR) testing of upper respiratory tract (URT) samples from hospitalised patients with coronavirus disease 2019 (COVID-19), compared to the gold standard of a clinical diagnosis.

Methods: All URT RT-PCR testing for severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) in NHS Lothian, Scotland, United Kingdom between the 7 th of February and 19 th April 2020 (inclusive) was reviewed, and hospitalised patients were identified. All URT RT-PCR tests were analysed for each patient to determine the sequence of negative and positive results. For those who were tested twice or more but never received a positive result, case records were reviewed, and a clinical diagnosis of COVID-19 allocated based on clinical features, discharge diagnosis, and radiology and haematology results. For those who had a negative RT-PCR test but a clinical diagnosis of COVID-19, respiratory samples were retested using a multiplex respiratory panel, a second SARS-CoV-2 RT-PCR assay, and a human RNase P control.

Results: Compared to the gold standard of a clinical diagnosis of COVID-19, the sensitivity of a single upper respiratory tract RT-PCR for COVID-19 was 82.2% (95% confidence interval 79.0-85.1%).   The sensitivity of two upper respiratory tract RT-PCR tests increased sensitivity to 90.6% (CI 88.0-92.7%). A further 2.2% and 0.9% of patients who received a clinical diagnosis of COVID-19 were positive on a third and fourth test; this may be an underestimate of the value of further testing as the majority of patients 93.0% (2999/3226) only had one or two URT RT-PCR tests.

Conclusions: The sensitivity of a single RT-PCR test of URT samples in hospitalised patients is 82.2%. Sensitivity increases to 90.6% when patients are tested twice.  A proportion of cases with clinically defined COVID-19 never test positive on URT RT-PCR despite repeat testing.

Article activity feed

  1. Version published to 10.12688/wellcomeopenres.16342.1
  2. ScreenIT

    SciScore for 10.1101/2020.06.19.20135756: (What is this?)

    Please note, not all rigor criteria are appropriate for all manuscripts.

    Table 1: Rigor

    Institutional Review Board Statementnot detected.
    Randomizationnot detected.
    Blindingnot detected.
    Power Analysisnot detected.
    Sex as a biological variablenot detected.

    Table 2: Resources

    Software and Algorithms
    SentencesResources
    If WGS had been completed successfully for a sample, this was assumed to represent a true positive.
    WGS
    suggested: None
    For samples that tested positive using the SeeGene assay, Ct values for human RNase P were compared to negative results using a Welch two sample t-test in R [28] and plotted using GraphPad Prism version 6.04 for Windows (GraphPad Software, La Jolla California USA).
    GraphPad Prism
    suggested: (GraphPad Prism, RRID:SCR_002798)
    GraphPad
    suggested: (GraphPad Prism, RRID:SCR_002798)
    Convalescent serology samples (>14 days after onset of symptoms), if available, were analysed using the Abbott SARS-CoV- 2 IgG assay on the Abbott Architect platform [29].
    Abbott Architect
    suggested: (Abbott ARCHITECT i1000sr System, RRID:SCR_019328)

    Results from OddPub: We did not detect open data. We also did not detect open code. Researchers are encouraged to share open data when possible (see Nature blog).

    Results from LimitationRecognizer: We detected the following sentences addressing limitations in the study:
    As highlighted in the introduction, the included studies suffer from a number of limitations including reliance on RT-PCR itself as the diagnostic gold standard, which would lead to an increase in the estimated sensitivity. We are not aware of any studies which have used a clinical diagnosis of COVID-19 against which to assess the sensitivity of RT-PCR. Here we show that the sensitivity of an initial test is lower than reported in this meta-analysis, but that the chance of a false negative result (17.8%) is lower than the 29% estimated in a subsequent meta-analysis [3] using a subset of studies included in [2]. These widely varying estimates highlight the importance of more data to inform our understanding of the strengths and weaknesses of RT-PCR testing. Strengths and limitations: The strengths of the study include the large dataset of both COVID-19 positive and negative patients, and extensive further testing to rule out false negative RT-PCR results and alternative diagnoses in those patients given a clinical diagnosis of COVID-19. We also studied whether suboptimal sampling might be a possible explanation for false negatives. However in a cohort of 37 possible false negatives all samples had detectable RT-PCR for human RNase P, with no difference between this group and those that tested positive for SARS-CoV-2, showing that this was not a factor in determining the sensitivity of RT-PCR in this population. A limitation of the study is that the WHO/ECDC case definition of ...

    Results from TrialIdentifier: No clinical trial numbers were referenced.

    Results from Barzooka: We did not find any issues relating to the usage of bar graphs.

    Results from JetFighter: We did not find any issues relating to colormaps.

    Results from rtransparent:
    • Thank you for including a conflict of interest statement. Authors are encouraged to include this statement when submitting to a journal.
    • Thank you for including a funding statement. Authors are encouraged to include this statement when submitting to a journal.
    • No protocol registration statement was detected.

    About SciScore

    SciScore is an automated tool that is designed to assist expert reviewers by finding and presenting formulaic information scattered throughout a paper in a standard, easy to digest format. SciScore checks for the presence and correctness of RRIDs (research resource identifiers), and for rigor criteria such as sex and investigator blinding. For details on the theoretical underpinning of rigor criteria and the tools shown here, including references cited, please follow this link.

  3. Version published to 10.1101/2020.06.19.20135756v1 on medRxiv