Liver, cardiovascular and metabolic factors as predictors of all-cause mortality in a rural Ugandan Cohort

Markers of liver and metabolic disease have not been well described for African populations but could be crucial to inform population health interventions that reduce morbidity and mortality. We studied a population in rural Uganda to describe the baseline and longitudinal distributions of liver function tests, and to investigate how liver, metabolic and cardiovascular parameters are associated with all-cause mortality. Demographic and laboratory data were collected through the General Population Cohort (GPC) in South-Western Uganda. We summarised cohort characteristics at baseline using descriptive statistics, and used univariable and multivariable Cox proportional hazards models to investigate factors associated with hazards of death. Our dataset includes 7896 individuals, of whom 56.2% were female and 72.5% were aged under 45 years. Prevalence of hepatitis B virus (HBV) and human immunodeficiency virus (HIV) was 2.7% and 7.4%, respectively. Death was associated with older age (p<0.0001), male sex (aHR 1.56, 95% CI 1.3 to 1.86) and HIV positivity (aHR 1.67, 95% CI 1.25 to 2.22). Excess mortality was associated with alcohol intake, based on increasing hazard of death associated with increase in AST:ALT ratio and increased serum gamma glutamyl transferase. A 10mmHg increase in systolic blood pressure was associated with 7% increased hazards of death (aHR 1.07, 95% CI 1.03 to 1.11), and one unit increases in serum HbA1c were associated with 25% increased hazards of death (aHR 1.25, 95% CI 1.13 to 1.38). Under sensitivity analysis restricted to a subset of participants with longitudinal follow-up data (n=871), HBV infection was associated with death (aHR 5.38, 95% CI 2.01-14.42). Simple interventions to tackle hypertension, diabetes, alcohol intake, and to diagnose and treat HIV and HBV infection would have a significant impact on mortality in this setting.