COVID-19 Vaccine Effectiveness Against Hospitalizations in Paraguay, May 2021-April 2022: A Test-Negative Design.

Background: Vaccine effectiveness (VE) estimates vary by population characteristics and circulating variants. North America and Europe have generated many COVID-19 VE estimates but relied heavily on mRNA vaccines. Fewer estimates are available for non-mRNA vaccines and from Latin America. We aimed to estimate the effectiveness of several COVID-19 vaccines in preventing SARS-CoV-2-associated severe acute respiratory infection (SARI) in Paraguay from May 2021 to April 2022. Methods: Using sentinel surveillance data from four hospitals in Paraguay, we conducted a test-negative case control study to estimate COVID-19 vaccine effectiveness against SARI by vaccine type/brand and period of SARS-CoV-2 variant predominance (Gamma, Delta, Omicron). We used multivariable logistic regression adjusting for month of symptom onset, age group, and presence of >= 1 comorbidity to estimate the odds of COVID-19 vaccination in SARS-CoV-2 test-positive SARI case-patients compared to SARS-CoV-2 test-negative SARI control-patients. Results: Of 4,229 SARI patients, 2,381 (56%) were SARS-CoV-2-positive case-patients and 1,848 (44%) were SARS-CoV-2-negative control-patients. A greater proportion of case-patients (73%; 95% CI: 71-75) than of control-patients (40%; 95% CI: 38-42) were unvaccinated. During the Gamma variant-predominant period, VE estimates for partial vaccination with mRNA vaccines and Oxford/AstraZeneca Vaxzevria were 90.4% (95% CI: 66.4-97.6) and 52.2% (95% CI: 25.0-69.0), respectively. During the Delta variant-predominant period, VE es-timates for complete vaccination with mRNA vaccines, Oxford/AstraZeneca Vaxzevria, or Gamaleya Sputnik V were 90.4% (95% CI: 74.3-97.3), 83.2% (95% CI: 67.8-91.9), and 82.9% (95% CI: 53.0-95.2), respectively. The effectiveness of all vaccines declined substantially during the Omicron variant-predominant period. Conclusions: This study contributes to our understanding of COVID-19 VE in Latin America and to global understanding of vaccines that have not been widely used in North America and Europe. VE estimates from Paraguay can parameterize models to estimate the impact of the national COVID-19 vaccination campaign in Paraguay and similar settings.