Evidence of significant alcohol-associated mortality in a large population cohort in Uganda

Background and aims: Liver function tests (LFTs) can be used as prognostic biomarkers of all-cause mortality, however few investigations have been undertaken in African populations despite a high burden of liver disease in these settings. Previous analysis of the General Population Cohort (GPC) in Uganda suggested high prevalence of alcohol-related liver disease, HIV infection, HBV and associated primary liver cancer. We aimed to investigate how LFTs associated with risk of all-cause mortality in the GPC. Method: The prospective GPC was established in 1989 in Kalungu District in rural south-western Uganda. In the 22nd survey round (years 2010/11), selected participants underwent measurement of biophysical and blood parameters to investigate disease risk factors andoutcomes.Outcomeswereascertainedviapresentationtohealth clinic. Descriptive statistics were used to summarise parameters and compare correlations between LFTs, blood lipids (including low density lipoprotein, LDL), AST:ALT ratio (values >2 indicate alcoholic hepatitis), Haemoglobin A1C (HbA1c). We investigated characteristics/biomarkers associated with hazards of death via Cox proportional hazards modelling. Results: In 7896 individuals with median age 30 years (IQR 17–46), 43.8% were male (n=3455). Minorities of individuals had confirmed chronic hepatitis B virus (HBV) (n=216, 2.7%) or human immunodeficiency virus (HIV) (n=582, 7.4%) infections throughout followup. Over 6.30 mean years (SD 3.26) of follow-up, 629 (7.8%) individuals died from any cause. Fisk factors for all-cause mortality included increasing age, male sex, HIV positivity (Figure 1). Nearly a quarter (23.3%, n=147) of participants who died had an AST:ALT value>2,ascomparedto9.8%(n=710)ofindividualswhosurvived.A one-unit increase in AST:ALT ratio was associated with 17% higher hazards of death, and a 10 unit increase in gamma-glutamyl transferase (GGT) with a 1% increase in hazards. Other cardiometabolic factors associated with death included HbA1c and LDL. Figure: Forest plot for factors associated with all-cause mortality in Ugandan General Population Cohort (GPC). † Per 10-unit increase in GGT Conclusion: Understanding of liver disease in the WHOAfrica region must urgently improve. Using AST:ALT ratio cutoff (established in Western populations) suggests an association of alcoholic hepatitis with mortality, which is substantiated by elevated GGT. Clinical and public health programmes should be informed by data suggesting high alcohol consumption is a significant contributor to mortality.