Group Based Trajectory Analysis of Cognitive Outcomes in Children with Perinatal HIV

Background Cognitive impairment is common in children with perinatally-acquired HIV (pHIV). It is not known whether exposure to HIV-related neuropathogenic mechanisms during vulnerable periods of neurodevelopment may produce distinct long-term cognitive phenotypes as children age. We used group based trajectory modeling to identify clusters of children with pHIV following a unique developmental trajectory across age and predictors of belonging to select cognitive trajectory groups. Methods Participants included children aged 1 to 17 enrolled in the PREDICT resilience study, a cohort study of children with pHIV in Thailand and Cambodia. Cognitive testing was conducted semi-annually over three years. Group based trajectory analyses determined subgroups of children with differing cognitive trajectories using maximum likelihood estimates and Bayesian statistics. Multiple logistic regression identified baseline factors associated with belonging to the lowest scoring trajectory group. Results Three distinct cognitive phenotypes were identified for each neurocognitive test categorized as high, medium and low scoring groups. A subgroup of children demonstrated normal developmental patterns for Color Trails Test 1 and 2. Children in the low trajectory group were more likely to present at an older age (>8 years, OR: 2.72; p 0.01) and report lower household income level (OR: 0.33-0.42; p<0.005). Neither CD4 nadir nor treatment arm was associated with cognitive trajectory status. Conclusion Our study reflects the benefit of using group based trajectory modeling to classify the heterogeneity in cognitive outcomes of children with pHIV. Children were described as belonging to three distinct subgroups determined at study onset alluding to the fact that cognitive outcomes are likely to be determined at an early age with little variability over time in children with pHIV. Demographic variables, including older age at presentation and household income, were associated with low scoring cognitive trajectories, whereas HIV related variables were not. These findings mirror other studies and demonstrate the impact of socioeconomic factors on cognitive development in children with pHIV.