• Citer

Setting up a Network to Estimate Influenza Vaccine Effectiveness in Latin America, 2013-2017

par El Omeiri, Nathalie
Président du jury Vandenberg, Olivier
Promoteur Levêque, Alain
Publication Non publié, 2019-07-01

Thèse de doctorat

• Accès en ligne
• Détails
• Contenu
• Statistiques

Résumé :

As of 2018, 40 (77%) of 50 countries/territories in Latin America and the Caribbean have established policies for seasonal influenza vaccination. Despite high coverage levels among the high-risk groups targeted for vaccination, these countries have not routinely measured influenza vaccine effectiveness (VE). We aimed to estimate VE in preventing influenza-associated severe illness among young children and the elderly in Latin America to demonstrate the value of current vaccination strategies and sustain investments in influenza vaccines. We built on the existing severe acute respiratory infections (SARI) surveillance platforms, integrating data from immunization programs to conduct a multi-country case test-negative control study at 123 sentinel hospitals during 2013–2017 in Argentina, Brazil, Chile, Colombia, Costa-Rica, Honduras, El Salvador, Panama, and Paraguay. Surveillance staff identified children aged six months‒2 years and adults aged ≥60 years (both eligible for influenza vaccination provided free of charge by immunization programs) among patients hospitalized with SARI and collected a respiratory specimen from them within ten days of illness onset. Cases were SARI case-patients with influenza virus infection confirmed by reverse transcription polymerase chain reaction (RT-PCR); controls were SARI case-patients RT-PCR negative for influenza viruses. An individual was considered vaccinated if he/she had documented proof of vaccination during the most recent influenza vaccination campaign/season and at least 14 days before the onset of symptoms. We defined full vaccination as the receipt of one dose among adults, and children aged ≤2 years if previously vaccinated; two doses were required for full vaccination of vaccine naïve children aged ≤2 years. All countries used trivalent inactivated influenza vaccines. We used a two-stage random effects model to estimate pooled VE per target age group (children or older adults), adjusting for calendar time (month of illness onset), age, sex, and underlying medical conditions. When restricting the analysis to data from five consecutive seasons, in five countries that contributed consistently to the multicenter study, we included 9,197 SARI cases (3,160 children and 6,037 adults) in VE analyses. Overall VE against any influenza virus was 37% [95%CI: 33%─40%]. Among young children, the adjusted VE for covariates (VEadj) against any circulating influenza type/subtype was 32% [95%CI: 15%─45%] for children who received two doses and 11% [95%CI: -32%─40%] for those who received one dose. A similar pattern of higher VEadj point estimates were observed for A(H3N2) and B viruses. Among older adults, the VEadj against any circulating influenza virus was 40% [28%─51%]. A(H1N1)-specific Vadj was 44% [30%─55%], A(H3N2)-specific VEadj was 36% [24%─47%] and influenza B specific VE was 19% [-12%, 41%]. The VEadj against any circulating influenza type/subtype was 44% for those consecutively vaccinated in two seasons, 39% for those vaccinated in the current season only and 23% for those vaccinated in the prior season only. Trivalent inactivated influenza vaccines prevented more than one third (37%) of influenza-associated hospitalizations among children aged ≤2 years and adults≥60 years during 2013–2017. Sentinel surveillance networks in middle-income countries, such as some Latin American countries, could provide a simple platform to estimate regional influenza VE annually. Inter-institutional and multidisciplinary efforts between epidemiology, laboratory, and immunization programs was key to the successful implementation. Pooling data from multiple countries and seasons allowed us to obtain more robust estimates of VE per target group and type/subtype of the influenza virus.