SARS-CoV-2 Omicron Symptomatic Infections in Previously Infected or Vaccinated South African Healthcare Workers

Abstract

We investigated Omicron infections among healthcare workers (HCW) presenting with symptoms of SARS-CoV-2 infection and evaluated the protective effect of vaccination or prior infection. Between 24 November and 31 December 2021, HCW in Johannesburg, South Africa, were tested for SARS-CoV-2 infection by Nucleic Acid Amplification Test (NAAT). Blood samples collected either at the symptomatic visit or in the 3 months prior, were tested for spike protein immunoglobulin G (IgG). Overall, 433 symptomatic HCW were included in the analysis, with 190 (43.9%) having an Omicron infection; 69 (16.7%) were unvaccinated and 270 (62.4%) received a single dose of the Ad26.COV.2 vaccine. There was no difference in the odds of identifying Omicron between unvaccinated and Ad26.COV.2 vaccinated HCW (adjusted odds ratio (aOR) 0.81, 95% confidence interval (CI): 0.46, 1.43). One-hundred and fifty-four (35.3%) HCW had at least one SARS-CoV-2 NAAT-confirmed prior infection; these had lower odds of Omicron infection compared with those without past infection (aOR 0.55, 95%CI: 0.36, 0.84). Anti-spike IgG concentration of 1549 binding antibody unit/mL was suggestive of significant reduction in the risk of symptomatic Omicron infection. We found high reinfection and vaccine breakthrough infection rates with the Omicron variant among HCW. Prior infection and high anti-spike IgG concentration were protective against Omicron infection.

Keywords: COVID-19; Omicron; SARS-CoV-2; reinfection; vaccination.

Figure 1

Conditional inference of Omicron infection probability and anti-spike IgG levels by prior SARS-CoV-2 NAAT-confirmed infection. (A) Inferred significant splits in previous SARS-CoV-2 NAAT-confirmed cases and spike IgG levels impact on the probability of having an Omicron infection during the study period (indicated by the red bars). The tree was generated from a training set composed of 90% of all visits with a known serological result. The algorithms infection predictive power was measured to be 72% in the remaining 10% of the data, with 23% type I error. (B) Antibody density distributions for participants with either NAAT-confirmed Omicron infection (red line) or no infection (grey line) during the study period. The vertical black line corresponds to the threshold 1549 anti-spike IgG binding units that emerged from the analysis on panel A. (C) Anti-spike IgG levels by prior SARS-CoV-2 NAAT-confirmed infection and vaccination status. Kruskal–Wallis tests indicate significant differences in IgG concentrations between participants with different levels of prior infection (p = 0.00015), and with more doses of vaccination in those who were not previously exposed (p = 0.000057). A lower significance in differences in IgG concentration for groups with different vaccination status for those who had a prior confirmed SARS-CoV-2 infection (p = 0.038).