Home-based SARS-CoV-2 lateral flow antigen testing in hospital workers

We read with interest the recent paper from Fowler et al. on the use of a reverse transcriptase loop mediated isothermal amplification (RT-LAMP) assay for rapid detection of SARS-CoV-2 infection¹ and the subsequent recommendation of its suitability for asymptomatic staff testing.² Lateral flow device (LFD) antigen tests have also been proposed as rapid point-of-care diagnostics for SARS-CoV-2 infection and have the advantage over the RT-LAMP assay in that they can be self-delivered by healthcare workers at home with immediate results. One LFD, Innova, had a specificity of 99.7% and sensitivity of 77% in a large national evaluation, with >90% of high viral load infections (cycle threshold, Ct ≤25.5) detected.³ However, sensitivity can be lower when LFDs are used by the general public and some have claimed these tests are not fit for purpose.⁴ In England from November 2020, twice weekly LFD testing has been offered to hospital workers. We present our experience at Oxford University Hospitals of deploying self-administered home-based LFD testing of anterior nasal swabs for our staff.

Between 23 November 2020 and 09 January 2021, 8657 healthcare workers and support staff registered and collected a box of 25 Innova LFD tests (Fig. 1 A). 46,503 test results were reported using an online form by 7567 asymptomatic individuals: median (IQR)[range] 6 (3–9) [1–23] tests/individual. 45,710 (98.3%) tests were negative, 465 (1.0%) invalid and 328 (0.7%; from 295 individuals) positive. 28 (9%) positive LFD results were in recently PCR-positive staff tested a median (IQR) [range] 3.5 (1-6) [1-13] days earlier. Of the remainder, 169 (52%) positive LFD results were followed by confirmatory PCR (Thermo Fisher TaqPath) at our hospital within 3 days (Fig. 1B), 161 were PCR-positive and 8 PCR-negative (positive predictive value, 189/197, 96% [95%CI 92-98%], false positive rate 8/27930, 0.03% [95%CI 0.01–0.06%], accounting for incomplete confirmatory testing by reducing the denominator proportionally to the percentage of confirmatory tests undertaken).

Fig. 1.

Lateral flow device collections and result submission (panel A) and confirmatory results within 3 days (panel B) and associated Ct values (panel C). Results are presented for 23 November 2020 to 09 January 2021 inclusive. PCR tests were performed using the Thermo Fisher TaqPath assay (targeting S and N genes, and ORF1ab, available mean cycle threshold (Ct) values for detected targets are shown). Panel C shows the PCR results from asymptomatic staff obtained within ±3 days of a positive LFD device, and PCR results from asymptomatic staff without a positive LFD. The box plots indicate median and interquartile ranges.

In contrast, more positive results were detected in staff asymptomatic screening using PCR of combined nasal/oropharyngeal swabs⁵ over the same time period, 127/8329 (1.5%). Viral loads, assessed using Ct values, were higher in those with positive LFD antigen results versus other PCR-positive asymptomatic staff, median (IQR) Ct 14 (12–18) versus 30 (22-33) (p<0.001; Fig. 1C).

1398/7567 (18.5%) staff reporting LFD results had not attended asymptomatic screening for ≥90 days, another 1128 (14.9%) never had. Of 295 staff who tested LFD positive, 116 (39%) had not attended for asymptomatic screening in the previous 30 days including 54 (18%) not in the last 90 days and another 40 (14%) had never attended. Staff reported the nasal swab used for LFD testing was preferable to the combined nose and throat swab taken for PCR, found testing kits easy to use, and the process acceptable even when done regularly.

Use of LFDs identified asymptomatic SARS-CoV-2 infections that would not otherwise have been detected. This enabled interventions to reduce staff-to-staff and staff-to-patient transmission, which were focused on staff with high viral loads, i.e. potentially those most infectious. Although individuals with low viral loads may be missed, with regular serial testing at least twice per week, those with early infection and rising viral loads are likely to be identified. Determining whether on-going PCR-based screening is required in addition to LFDs will require better estimates of the relative infectiousness of individuals by viral load. Kit failure rates and false positive results in this mixed population of healthcare workers and support staff were sufficiently low to support widespread use of LFDs in asymptomatic populations.

Ethics statement

Deidentified data were obtained from the Infections in Oxfordshire Research Database which has generic Research Ethics Committee, Health Research Authority and Confidentiality Advisory Group approvals (19/SC/0403, 19/CAG/0144).

Declaration of Competing Interests

DWE declares lecture fees from Gilead, outside the submitted work. No other author has a conflict of interest to declare.