BACKGROUND: International treatment guidelines recommend that HIV-positive patients be vaccinated for influenza annually. Evidence supporting this recommendation is limited. We assessed the efficacy of influenza vaccines in preventing influenza in HIV-positive patients through a systematic review and meta-analysis. METHODS: We searched 10 electronic databases independently, in duplicate (from inception to June 2007). We extracted data on study design, population characteristics and outcomes related to influenza symptoms and antibody titres. We pooled data using a random effects model and conducted sensitivity analyses to evaluate heterogeneity. RESULTS: We included four studies. Three studies were evaluable for meta-analysis and yielded a pooled relative risk reduction (RRR) of 66% [95% confidence interval (CI) 36-82%; I(2)=73%]. One case-control study yielded an odds ratio of 1.98 (95% CI 0.75-5.20). When we assessed heterogeneity according to study design, we found that the study of the highest quality, a randomized clinical trial (RCT), yielded the most conservative estimate (RRR 41%; 95% CI 2-64%). INTERPRETATION: Evidence supporting influenza vaccination of HIV-positive individuals is limited, poorly quantified and characterized by substantial methodological shortcomings. A reasonable estimate of influenza vaccination effectiveness in HIV-positive patients cannot be derived from these data. There is an urgent need for randomized trials to guide policy and clinical practice.
BACKGROUND: International treatment guidelines recommend that HIV-positivepatients be vaccinated for influenza annually. Evidence supporting this recommendation is limited. We assessed the efficacy of influenza vaccines in preventing influenza in HIV-positivepatients through a systematic review and meta-analysis. METHODS: We searched 10 electronic databases independently, in duplicate (from inception to June 2007). We extracted data on study design, population characteristics and outcomes related to influenza symptoms and antibody titres. We pooled data using a random effects model and conducted sensitivity analyses to evaluate heterogeneity. RESULTS: We included four studies. Three studies were evaluable for meta-analysis and yielded a pooled relative risk reduction (RRR) of 66% [95% confidence interval (CI) 36-82%; I(2)=73%]. One case-control study yielded an odds ratio of 1.98 (95% CI 0.75-5.20). When we assessed heterogeneity according to study design, we found that the study of the highest quality, a randomized clinical trial (RCT), yielded the most conservative estimate (RRR 41%; 95% CI 2-64%). INTERPRETATION: Evidence supporting influenza vaccination of HIV-positive individuals is limited, poorly quantified and characterized by substantial methodological shortcomings. A reasonable estimate of influenza vaccination effectiveness in HIV-positivepatients cannot be derived from these data. There is an urgent need for randomized trials to guide policy and clinical practice.