Longitudinal Follow Up of Immune Responses to SARS-CoV-2 in Health Care Workers in Sweden With Several Different Commercial IgG-Assays, Measurement of Neutralizing Antibodies and CD4+ T-Cell Responses.

Background: The risk of SARS-CoV-2 infection among health care workers (HCWs) is a concern, but studies that conclusively determine whether HCWs are over-represented remain limited. Furthermore, methods used to confirm past infection vary and the immunological response after mild COVID-19 is still not well defined. Method: 314 HCWs were recruited from a Swedish Infectious Diseases clinic caring for COVID-19 patients. IgG antibodies were measured using two commercial assays (Abbot Architect nucleocapsid (N)-assay and YHLO iFlash-1800 N and spike (S)-assays) at five time-points, from March 2020 to January 2021, covering two pandemic waves. Seroprevalence was assessed in matched blood donors at three time-points. More extensive analyses were performed in 190 HCWs in September/October 2020, including two additional IgG-assays (DiaSorin LiaisonXL S1/S2 and Abbot Architect receptor-binding domain (RBD)-assays), neutralizing antibodies (NAbs), and CD4+ T-cell reactivity using an in-house developed in vitro whole-blood assay based on flow cytometric detection of activated cells after stimulation with Spike S1-subunit or Spike, Membrane and Nucleocapsid (SMN) overlapping peptide-pools. Findings: Seroprevalence was higher among HCWs compared to sex and age-matched blood donors at all time-points. Seropositivity increased from 6.4% to 16.3% among HCWs between May 2020 and January 2021, compared to 3.6% to 11.9% among blood donors. We found significant correlations and high levels of agreement between NAbs and all four commercial IgG-assays. At 200-300 days post PCR-verified infection, there was a wide variation in sensitivity between the commercial IgG-assays, ranging from 90% in the RBD-assay. There was only moderate agreement between NAbs and CD4+ T-cell reactivity to S1 or SMN. Pre-existing CD4+ T-cell reactivity was present in similar proportions among HCW who subsequently became infected and those that did not. Conclusions: HCWs in COVID-19 patient care in Sweden have been infected with SARS-CoV-2 at a higher rate compared to blood donors. We demonstrate substantial variation between different IgG-assays and propose that multiple serological targets should be used to verify past infection. Our data suggest that CD4+ T-cell reactivity is not a suitable measure of past infection and does not reliably indicate protection from infection in naive individuals.