Prevalence of HIV-associated neurocognitive disorders (HAND) in virologically suppressed HIV-infected individuals in the ASSURE (EPZ113734) study versus non-HIV-infected control subjects

Purpose: Cognitive impairment prevalence and potential predictive variables in HIV-infected subjects were examined; cognitive impairment prevalence was compared to uninfected controls.

Methods: HIV-infected, ART-experienced subjects stably suppressed (HIV-1 RNA <75c/mL) with ≥6 months’ tenofovir/emtricitabine/atazanavir/ritonavir treatment had neurocognitive evaluations prior to clinical study randomization for ASSURE.  Z scores were obtained for 4 Cogstate tasks.  Poor cognition=2 of 4 component z scores ≤-1 or composite score ≤-2. Forty-three baseline variables underwent univariate and multivariate analyses to identify those associated with poor cognition. Using logistic regression and analysis of variance, results were compared to data from 700 uninfected controls (data supplied by Cogstate).

Results: 293 HIV-infected subjects were evaluated.  Baseline characteristics: median age 43 years (range:21-68); 79% male; African-American/White/Other race 34%/60%/5%; Hispanic/Latina ethnicity 26%; CDC Class C 18%; median CD4 count:492cells/mm³; median CD4 nadir:265cells/mm³; median days prior antiretroviral therapy:1000.  Control group (n=700); median age: 44 years (range:15-79); 50% male). Poor cognition was observed in 56% (164/293) of HIV-infected subjects compared to 17% (117/700) of uninfected controls (odds-ratio: 6.33 increase in risk for HIV-infected subjects). Significant differences were observed between HIV-infected subjects and uninfected controls (p<0.0001) and when adjusting for age (p<0.001) for each test component.  Poor cognition was associated with increasing age for both HIV-infected subjects and uninfected controls. In HIV-infected subjects, significant associations (p<0.05) were observed by univariate analysis between composite score and race, gender, ethnicity and homosexual contact, but not CD4 cell count or CD4 nadir. Multivariate analysis confirmed significant associations (p<0.05) between poor cognition and Hispanic/Latina ethnicity and between poor cognition and African-American race.

Conclusion: The HIV-infected ASSURE population had a 3-fold higher prevalence of greater cognitive impairment compared with non-HIV-infected controls (56% versus 17%) and a 6-fold higher odds ratio for developing cognitive impairment.  By multivariate analyses, Hispanic/Latina ethnicity and African-American race were associated with poorer cognition.