The association between physical activity and cognition in men with and without HIV infection

HIV-associated neurocognitive disorders are highly prevalent, and physical activity (PA) is a modifiable behaviour that may affect neurocognitive function. Our objective was to determine the association between PA and neurocognitive function and the effect of HIV on this association.

PA was assessed in the Multicenter AIDS Cohort Study with the International Physical Activity Questionnaire. A neuropsychological test battery assessed global impairment and domain-specific impairment (executive function, speed of processing, working memory, learning, memory, and motor function) every 2 years. Semiannually, the Symbol Digit Modalities Test and Trail Making Test Parts A and B were performed. Adjusted logistic regression models were used to assess the PA-neurocognitive function association. Using longitudinal data, we also assessed the PA category-decline of neurocognitive function association with multivariate simple regression.

Of 601 men, 44% were HIV-infected. Low, moderate, and high PA was reported in 27%, 25%, and 48% of the HIV-infected men vs. 19%, 32% and 49% of the HIV-uninfected men, respectively. High PA was associated with lower odds of impairment of learning, memory, and motor function [odds ratio (OR) ranging from 0.52 to 0.57; P < 0.05 for all]. The high PA-global impairment association OR was 0.63 [95% confidence interval (CI) 0.39, 1.02]. Among HIV-infected men only, across multiple domains, the high PA-impairment association was even more pronounced (OR from 0.27 to 0.49). Baseline high/moderate PA was not associated with decline of any domain score over time. HIV infection was marginally associated with a higher speed of decline in motor function.

A protective effect of high PA on impairment in neurocognitive domains was observed cross-sectionally. Longitudinal PA measurements are needed to elucidate the PA-neurocognitive function relationship over time.