Abstract
BACKGROUND/OBJECTIVES: To estimate (a) survival after SARS-CoV-2 infection by COVID-19 vaccination status, and (b) COVID-19 vaccine effectiveness in a middle-income country. METHODS: In this retrospective cohort study, secondary analysis of data from the national surveillance and vaccination databases was conducted. The primary outcome was COVID-19 death classified based on the WHO criteria. Data were analysed by vaccination status, age, sex, geographic region, and wave period. Kaplan-Meier curves were plotted; log-rank followed by multiple comparison tests were used to compare survival probabilities. Cox proportional-hazards models with time-varying covariates estimated hazard ratios (HR). Vaccine effectiveness was computed as (1-HR) × 100. RESULTS: A total of 55,299 COVID-19 cases were captured by the national surveillance system between 1 April and 31 December 2021. Of these, 45,774 (1581 vaccinated, 44,193 unvaccinated) were included in the analysis. After a follow-up of 327 days, there were 22 deaths (case fatality rate (CFR) 1.5%) among 1581 COVID-19 vaccinated cases and 1821 deaths (CFR 4.1%) among 44,193 unvaccinated cases. There was one COVID-19 death per 10,000 person days in vaccinated cases compared with 2.7 COVID-19 deaths per 10,000 person days in unvaccinated cases. After adjustment for age, sex, and geographic region, the effectiveness against COVID-19 death across all vaccine types (ChAdOx1 nCoV-19, BNT162b2, Ad26.COV2.S, or BBIBP-CorV) was 68% (95% CI: 51-79). Effectiveness was 75% (95% CI: 59-84) for ChAdOx1 nCoV-19. Vaccine effectiveness across all vaccine types was higher in younger cases, (82% (95% CI: 52-93), 18-64 years vs. 63% (95% CI: 41-77), ≥65 years), females (84% (95% CI: 63-93), females vs. 53% (95% CI: 24-71), males) and those vaccinated in the past 3 months (71% (95% CI: 47-85), past 0-3 months vs. 56% (95% CI: 23-75), 3-6 months). CONCLUSIONS: COVID-19 vaccines were effective in preventing COVID-19 death in a population with low vaccination coverage. Limitations of the analysis include the use of surveillance data (under-reporting of cases, missing data), exclusion of partially vaccinated cases, and insufficient data on important confounders (circulating variants and comorbidities).