Prevalence, predictors, and effectiveness of the influenza vaccination in HIV-infected women: Nested studies in the Women's Interagency HIV study (WIHS)

Background. Since 1988, HIV-infected (HIV+) individuals have been recommended to receive the annual influenza vaccine. This study sought to determine prevalence, predictors, and effectiveness (VE) of the vaccine in HIV+ and similar HIV-uninfected (HIV-) women enrolled in the Women's Interagency HIV Study (WIHS).;Methods. Cross-sectional and prospective cohort study designs were nested in the WIHS surrounding the 2005-6 and 2006-7 influenza seasons. Predictors were investigated using Poisson regression models with robust variance to estimate univariate (PR) and adjusted (aPR) prevalence ratios and 95% confidence intervals ([ , ]) in the cross-sectional studies. The original prediction model was fit to data from the 2005-6 season, internally, prospectively validated in the 2006-7 season, and expanded with to include predictors of vaccination beliefs. Finally, VE estimates were calculated as 1 - hazard ratio (HR) of influenza illness in vaccinated compared to unvaccinated women. HRs were estimated in the cohort studies using Cox proportional hazards models with time varying vaccination status and adjusted (aHR) for confounders.;Results. Vaccination prevalence in HIV+ women estimated from the 2006-7 cohort study was 59.6%. Measures of expanded model discrimination suggested utility in the model's predictive ability (c index = 0.81; Somers' Dxy = 0.62). Predictors of influenza vaccination in HIV+ women included: receiving any other vaccination in the prior 6 months (aPR=1.48 [1.24, 1.76]); receiving the influenza vaccine in the prior year (aPR=1.58 [1.33, 1.88]); having a CD4+ T-lymphocyte count <200 cells/mm3 (aPR=0.72 [0.53, 0.97]); receiving highly active antiretroviral therapy (aPR=1.30 [1.04, 1.62]); belief the vaccination is protective (aPR=1.50 [1.16, 1.94]); and having a discussion with a health care provider about the vaccine (aPR=2.44, [2.04, 2.93]). VE estimates from the 2006-7 season using various definitions of self-reported influenza illness showed harmful effects in HIV+ women (VE=-40%, aHR=1.40 [0.83, 2.31] to VE=-68%, aHR=1.68 [1.12, 2.51]) and protective effects in HIV-women (VE=31%, aHR=0.69 [0.26, 1.80] to VE=48%, aHR=0.52 [0.18, 1.51]).;Conclusions. Compliance with influenza vaccination recommendations was low. The strongest predictors of vaccination measured an individual's access to health care and beliefs. These VE estimates support the need for better data collection tools for use in HIV+ populations.