Background and Objectives While combination antiretroviral therapy (cART) has dramatically increased the life expectancy of people with HIV (PWH), nearly 50% develop HIV-associated neurocognitive disorders. This may be due to previously uncontrolled HIV viral replication, immune activation maintained by residual viral replication or activation from other sources, or cART-associated neurotoxicity. The aim of this study was to determine the effect of cART on cognition and neuroimaging biomarkers in PWH before and after initiation of cART compared with that in HIV-negative controls (HCs) and HIV elite controllers (ECs) who remain untreated.
Methods We recruited 3 groups of participants from the University of Rochester, McGovern Medical School, and SUNY Upstate Medical University: (1) ART treatment–naive PWH; (2) age-matched HCs; and (3) ECs. Participants underwent brain MRI and clinical and neuropsychological assessments at baseline, 1 year, and 2 years. PWH were also assessed 12 weeks after initiating cART. Volumetric analysis and fractal dimensionality (FD) were calculated for cortical and subcortical regions. Mixed effect regressions examined the effect of group and imaging variables on cognition.
Results We enrolled 47 PWH, 58 HCs, and 10 ECs. At baseline, PWH had worse cognition and lower cortical volumes than HCs. Cognition improved after initiation of cART and remained stable over time. Greater cortical thickness was associated with better cognition at baseline; greater FD of parietal, temporal, and occipital lobes was associated with better cognition at baseline and longitudinally. At baseline, ECs had worse cognition, lower cortical thickness, and lower FD in all 4 lobes and caudate than PWH and HCs. Greater cortical thickness, hippocampal volumes, and FD of frontal, temporal, and occipital lobes were associated with better cognition longitudinally.
Discussion Initiation of cART in PWH is associated with improvement in brain structure and cognition. However, significant differences persist over time when compared with HCs. Similar trends in ECs suggest that results are due to HIV infection rather than treatment. Stronger associations between cognition and FD suggest this imaging metric may be a more sensitive marker of neuronal injury than cortical thickness and volumetric measures.