CMV co‐infection and risk of AIDS and non‐AIDS events in a large cohort of HIV‐infected patients

Abstract

In HIV‐positive patients (pts), CMV co‐infection has been proposed as a key factor in sustaining immune activation, which in turn could play a role in determining immune senescence. We evaluated the prevalence and predictors of CMV co‐infection in a cohort of HIV+ pts and assessed the impact of CMV co‐infection on the risk of AIDS and non‐AIDS events. We included pts in the ICONA study with<1 month follow‐up and<1 CMVIgG (CMV) test available without active CMV disease. Pts' characteristics at time of the first CMV test (baseline) were compared in those tested positive (CMV+) and negative (CMV‐) using X2/Wilcoxon tests. Factors associated with CMV+ were identified by logistic regression. A prospective analysis was also performed with endpoints AIDS/AIDS‐related death and severe non‐AIDS (SNA: cardio‐cerebrovascular, neurologic disease, renal failure, non‐AIDS tumours)/death due to SNA. Time to event was estimated by Kaplan‐Meier curves and Cox regression (multivariable model included: age, gender, ethnicity, risk factor for HIV, HCVAb and HBsAg, AIDS and CD4 at baseline, initiation of ART prior to baseline). 6,053 pts were included; 83.7% were tested CMV+ a median of 17 (IQR 6–45) months after enrolment. As compared to CMV‐, CMV+ were older (adjusted odds ratio (AOR) 1.03 per 1 year older [95% CI 1.02–1.04]), HIV infected by homosexual route (MSM) (AOR 1.39 [95% CI 1.06–1.82]), less frequently Caucasian (AOR 0.56 [95% CI 0.42–0.76]), with higher CD4 count at baseline (AOR per 1 cell higher 1.035 [95% CI 1.00–1.06] By 10 years from first CMV test, 402 (12.6% [95% CI 11.1–13.6]) CMV+ and 74 (10.1% [95% CI 7.7–12.5]) CMV‐ pts developed AIDS/AIDS‐related death (log‐rank p=0.43). After adjustment for potential confounders, CMV+was still not associated with the risk of AIDS/AIDS‐related death (adjusted hazard ratio (AHR) 1.23 [95% CI 0.96–1.60]). By 10 years, 339 (10.6% [95% CI 9.4–11.9]) CMV+ and 41 (6.4% [95% CI 6.1–6.6]) CMV‐ pts experienced a non‐AIDS event/non‐AIDS death (log‐rank p=0.0006): 151 cancers, 128 CVD, 33 neurological, 1 renal. The association was still significant after controlling for a number of potential confounders: AHR 1.77 [95% CI 1.25–2.51] p=0.001; Table). In our study population, CMV/HIV co‐infection was associated with the risk of non‐AIDS events/deaths independently of other prognostic factors, supporting a potential role of CMV infection in vascular/ degenerative organ disorders commonly associated with chronic immune activation and aging.