RESUMEN
OBJECTIVES: We conducted an analysis to determine if differences in health-seeking behaviour can explain gender disparities in mortality among long-term survivors receiving antiretroviral therapy (ART) in rural Uganda. METHODS: From June 2012 to January 2014, we enrolled patients receiving a first-line ART regimen for at least 4 years without previous viral load (VL) testing in Jinja, Uganda. We measured HIV VL at study entry. We switched participants to second-line therapy, if VL was ≥ 1000 copies/mL on two measurements, and followed participants for 3 years. We collected clinical and behavioural data at enrolment and every 6 months after that. We used Poisson regression to examine factors associated with hospitalizations and Cox proportional hazards modelling to assess mortality to September 2016. RESULTS: We enrolled 616 participants (75.3% female), with a median age of 44 years and a median duration of ART use of 6 years. Of these, 113 (18.3%) had VLs ≥ 1000 copies/mL. Hospitalizations occurred in 101 participants (7% of men vs. 20% of women; P < 0.001). A total of 22 (3.6%) deaths occurred, 9% of men vs. 2% of women (P < 0.001). Multivariate modelling revealed that mortality was associated with age [adjusted hazard ratio (AHR) = 1.07 per year increase; 95% confidence interval (CI): 1.01-1.13], male gender (AHR = 2.57; 95% CI 1.06-6.23) and time-updated CD4 counts (AHR = 0.67 per 100 cell increment; 95% CI: 0.52-0.88). Virological failure was not associated with mortality (P = 0.762). CONCLUSION: Female patients receiving ART in rural Uganda were three times more likely to be hospitalized than men, but male mortality was nearly four times higher. Facilitating care for acute medical problems may help to improve survival among male ART patients.