Efficacy of inactivated hepatitis A vaccine in HIV-infected patients: A hierarchical bayesian meta-analysis

Patients with human immunodeficiency virus-1 (HIV) have elevated risk for hepatitis A virus (HAV) coinfection. Sequelae from coinfection include increased risk of liver-related morbidity and mortality. This study synthesizes the results of trials measuring response to HAV vaccine in HIV-infected patients using Bayesian meta-analysis methodologies. PubMed, OhioLINK/Medline, and other sources were used to search for studies analyzing response to HAV vaccine in HIV-infected patients. Studies were evaluated for quality. A Bayesian hierarchical random-effects model was used to estimate overall response to vaccine. Between-study variance was calculated. Sensitivity analyses were conducted to determine the potential for bias. The literature search yielded eight studies for inclusion in the meta-analysis, with a combined total of 458 patients. Observed proportions of response in individual studies ranged from 50 to 95%. The intention-to-treat meta-analysis estimated a combined proportion of HlV+ patients responding to vaccine of 64% (95% CI 52-75%). Heterogeneity between studies was significant. The overall response rate to HAV vaccine in HIV-infected patients is lower than typically cited. Up to 1/2 of HIV-infected patients may be nonresponders. Future research will be required to better understand the correlates of response. (c) 2005 Elsevier Ltd. All rights reserved.