Background
COVID-19 vaccine effectiveness (VE) studies leveraging systematic surveillance in sub-Saharan Africa are limited. We aimed to assess BNT162b2 and Ad26.COV2.S VE against SARS-CoV-2-associated hospitalisation in South African individuals aged ≥ 18 years.
Methods
We conducted a test-negative case-control study to estimate VE against hospitalisation in individuals enrolled in pneumonia surveillance in South Africa. Inpatients with physician-diagnosed lower respiratory tract infection or suspected COVID-19, testing SARS-CoV-2 positive or negative from May 2021‒March 2022 were cases or controls, respectively. Receiving one Ad26.COV2.S dose or two BNT162b2 doses ≥ 14 days before enrolment was considered fully vaccinated. VE was estimated using multivariable logistic regression for the Delta- and Omicron BA.1/BA.2-predominant periods; stratified by age and HIV-status.
Results
A total of 985 cases and 1,963 controls were included. Thirty-eight (3.9%) cases and 186 (9.5%) controls were fully vaccinated with BNT162b2; 30 (3.0%) cases and 94 (4.8%) controls were fully vaccinated with Ad26.COV2.S. BNT162b2 VE against SARS-CoV-2-associated hospitalisation over Delta and Omicron BA.1/BA.2 periods was 77% (95% CI: 26%;93%) and 38% (-9%;64%), respectively. Ad26.COV2.S VE against hospitalisation over Delta and Omicron BA.1/BA.2 periods was 47% (-57%;82%), and − 19% (-128%;37%), respectively. BNT162b2 VE against hospitalisation over Delta period was 84% (37%;96%) and 76% (21%;93%) among adults aged ≥ 60 years and HIV-uninfected, respectively.
Conclusions
BNT162b2 vaccine was effective against SARS-CoV-2-associated hospitalisation during the Delta period for adults aged ≥ 18 years, those aged ≥ 60 years, and HIV-uninfected adults. VE for Ad26.COV2.S was non-significant potentially due to limited sample size or residual confounding. These findings highlight the utility of sentinel surveillance for estimating VE.