Background:
Investigating outcomes of patients hospitalized with COVID-19 throughout the pandemic is crucial to understand the effects of SARS-CoV-2 variants, previous immunity, and healthcare interventions. We compared 28-day in-hospital mortality in adults hospitalized with COVID-19 caused by Wild-type, Alpha, Delta, or Omicron variants. Whether the difference in risk by variant might vary by age was also evaluated.
Methods:
We conducted a multinational cohort study including patients hospitalized with COVID-19 from 9 countries (EuCARE hospitalized study). Patients >18 years, hospitalized any time from 2020-02-01 to 2022-10-15 with a SARS-CoV-2 positive test were included. Variant was classified based on sequenced viruses (if available) or from national public metadata. In-hospital mortality was compared using the cumulative incidence (CI) function and Fine-Gray sub-distribution hazard models adjusted for age, sex, and comorbidities which are risk factors for severe COVID-19. Results were shown age-stratified since there was evidence that age was an effect measure modifier.
Results:
We included 38,585 SARS CoV-2 infected hospitalized patients (16,754 females and 21,831 males): 19,763 Wild-type, 6,387 Alpha, 3,640 Delta, and 8,795 Omicron. For in-hospital mortality, an interaction between age group and variant was observed (P=0.03), driven by the youngest group for whom smaller differences in mortality risk by variant were seen (Figure). In the older groups, the largest differences were observed between Omicron and the other variants. Among patients aged >70 years, the aSHR for Delta vs. Omicron was 2.06 (95% confidence interval 1.75-2.43). This estimate was 2.72 (2.38-3.12) for Alpha vs. Omicron, and 2.94 (2.64-3.27) for Wild-type vs. Omicron. Among unvaccinated patients, the aSHR was 1.33 (1.03-1.74) for Delta vs. Omicron, 1.61 (1.32-1.98) for Alpha vs. Omicron, and 1.69 (1.42-2.02) for Wild-type vs. Omicron. When comparing Omicron sublineages, the aSHR for the BA.1 sublineage was 2.04 (154-2.70) compared to the BA.2 sublineage and 1.71 (1.26-2.31) compared to the BA.5 sublineage.
Conclusions:
Overall, the cumulative incidence and hazard of in-hospital mortality rates decreased throughout the study period, particularly during the Omicron period. However, contrasts varied by age, especially for those relative to the Alpha variant. BA.1 carried a higher risk of death than that seen with the more recently circulating BA.2 and BA.5 sublineages.
