Growth curve modelling to determine distinct BMI trajectory groups in HIV-positive adults on antiretroviral therapy in South Africa

Objective: Obesity is a major long-term concern in HIV-positive patients due to the pathogenic link between obesity and noncommunicable chronic diseases (NCDs). We aim to characterize changes in BMI over time on antiretroviral therapy (ART) and investigate the association between weight gain and survival in South Africa. Design and methods: A prospective cohort study among HIV-positive adults on first-line ART between April 2004 and 2015 in Johannesburg, South Africa. We used latent-class growth modelling (adjusted for age, sex and CD4(+) cell count) to identify groups of individuals with similar patterns of change in BMI over time. Results: Eleven thousand, two hundred and sixty-three patients were included. The best fit model involved two linear and two quadratic trajectories. Thirty-five percent of patients were categorized into group one (mean BMI at ART initiation, 20.4 kg/m(2); mean BMI after 8 years of follow-up, 20.9 kg/m(2)), 38% into group two (24.5-26.2 kg/m(2)), 21% into group three (29.5-32.6 kg/m(2)) and 6% into group four (36.5-40.0 kg/m(2)). Over the 8 years of follow-up, 6% of our cohort went down in BMI standard category, while 45% went up. The largest increase occurred in the first 12 months on ART. In years 2 through 8, we saw a more gradual increase in BMI. Conclusion: The largest gain in BMI in HIV patients occurred in the first year on ART. During follow-up, over 50% of our population changed BMI categories putting them at an increased risk for NCDs. Consistent counselling on nutritional and lifestyle changes could help improve ART patients' long-term health outcomes.