Decreasing Incidence of Herpes Zoster in the Highly Active Antiretroviral Therapy Era

We evaluated the effect of highly active antiretroviral ther-apy (HAART) on the incidence of herpes zoster (HZ) inhuman immunodeficiency virus (HIV)–infected subjects.The annual incidence of HZ per 100 person-years decreasedsignificantly from 6.3 episodes in 1987 to 1.0 episode in2011, probably reflecting improved immune function.Keywords. HIV infection; herpes zoster; zoster vaccine;HAARTera; pre-HAARTera.The incidence of herpes zoster (HZ) is 10–20 times higher inpatients infected with human immunodeficiency virus (HIV)than in age-matched HIV-negative subjects [1,2].In the 1980s,HZ predicted the development of AIDS in a group of men whohad sex with men (MSM), with half of the AIDS cases arisingwithin 4 years of zoster [3].In the pre–highly active antiretrovi-ral therapy (HAART) era, several studies demonstrated inci-dence rates of 2.5–3.2 cases per 100 person-years in variouscohorts [4,5].The incidence, complications, and recurrences ofHZincreasewithlowerCD4T-cellcounts inHIV-infected sub-jects [4–6]. Although HZ incidence is higher with advancingage in the HIV-negative population, it might be lower in thecase of HIV-infected patients [5, 7]. In addition, HZ incidenceis likely to be higher in subjects who acquire HIV via sexualcontact vs intravenous drug abuse [8].There have been conflicting reports in the literature aboutthe incidence of HZ in HIV-infected subjects in the HAARTera compared to the pre-HAART era. The objective of ourstudy is to evaluate the incidence and risk factors for HZ in alarge cohort of HIV-infected persons in the pre-HAART andHAARTeras.METHODSTo evaluate the incidence of HZ in HIV-infected subjects, weincluded patients from the HIV Atlanta VA Cohort Study(HAVACS) who had an episode of HZ between 1982 and2011; this patient population is 99% male (seeSupplementaryData for full information on HAVACS). All patients with anepisode of HZ were included in the case series. HZ wasdefined as a clinical diagnosis made by the provider and sup-ported by the typical description of a painful unilateral rash,often with vesicular component. Only 5 cases of HZ wereconfirmed by polymerase chain reaction testing. Pertinentdata collected included year of diagnosis, race, HIV riskfactor, age, CD4 count, and antiretroviral treatment at thetime of diagnosis.The Clinical Case Registry of the Veterans Health Adminis-tration provided pertinent denominator data for this cohort in-cluding number of patients seen per year, age and risk factordistribution, distribution of CD4 by year, and number of pa-tients on antiretroviral therapy by year. Incidence data were de-veloped from the appropriate numerator and denominator orweighted data were developed for data, which changed per yearto produce standardized incidence rates (SIRs). We used SIRsfor our calculations as our demographics, including race andmedian age, prevalence of antiretroviral therapy, mean CD4counts, and mean HIV viral loads, have changed dramaticallyover the study period. This made it problematic to choose ap-propriate controls. Thus, we were not able to calculate oddsratios or perform multivariate analyses.RESULTSBetween January 1982 and December 2011, 650 episodes of HZwere documented in 3816 patients with an overall rate of 17.0cases per 100 patients. During this time period, 25 patients had2 episodes and 1 patient had 3 episodes of HZ. The overall inci-dence of HZ has decreased significantly from 28.6 cases per 100patients in the pre-HAARTera to 14.1 cases per 100 patients inthe HAARTera, reflecting a change in HZ risk factors (Supple-mentary Table 1). The date of the episodewas available for 94%