Multiplex lateral flow test sensitivity and specificity in detecting influenza A, B and SARS-CoV-2 in adult patients in a UK emergency department.

Background: Rapid identification of individuals with acute respiratory infections is crucial for preventing nosocomial infections. For rapid diagnosis, especially in EDs, lateral flow devices (LFDs) are a convenient, inexpensive option with a rapid turnaround. Several 'multiplex' LFDs (M-LFDs) now exist, testing for multiple pathogens from a single swab sample. We evaluated the real-world performance of M-LFD versus PCR testing in detecting influenza A, B and SARS-CoV-2) in the ED setting.

Methods: After preliminary evaluation of an M-LFD (SureScreen) with laboratory-grown virus and PCR-negative clinical samples, it was evaluated in a real-world setting at the ED of St Thomas' Hospital (London, UK) from 1 December 2022 to 21 April 2023. Eligible participants were ≥18 years of age, admitted with respiratory symptoms and received concurrent M-LFD and PCR tests. Main endpoints were sensitivity to detect influenza A/B (primary) and SARS-CoV-2 (secondary) versus PCR. The probability of a true positive in relation to viral concentration (expressed as PCR cycle threshold (Ct)) was analysed using logistic regression.

Results: In total, 808 symptomatic participants were included (49.8% female; mean age 46.9 years). Test sensitivity (95% CI) was 67.0% (56.9% to 76.1%) for influenza A (n=100), 94.1% (71.3% to 99.9%) for influenza B (n=17) and 48.2% (39.7% to 56.8%) for SARS-CoV-2 (n=141). Sensitivity for SARS-CoV-2 was significantly lower than that for influenza A and B (p=0.0057 and p=0.00088, respectively). The probability of a true positive was 98% for Ct<25 for influenza A and SARS-CoV-2 (influenza B non-evaluable). No co-infections were identified by PCR or M-LFD.

Conclusion: The real-world performance of SureScreen M-LFD was consistent with laboratory evaluation and achieved a high sensitivity for individuals with high viral concentration, most likely to be infectious. Given the representative UK population sample, results could be generalised for use in other settings.