ABSTRACT
BACKGROUND: Data on outcomes of non-alcoholic fatty liver disease (NAFLD) from South Asia are lacking. We compared mortality, among those with- and without-NAFLD, after 10-years follow-up among urban, adult Sri Lankans. METHOD: Participants (aged 35-64 years), selected by age-stratified random sampling, were screened by structured-interview in 2007. Anthropometric measurements, liver ultrasonography and biochemical/serological tests were done. NAFLD was diagnosed on ultrasound criteria, safe-alcohol consumption (Asian-standards) and absence of hepatitis B/C. Subjects without NAFLD were those without any ultrasound criteria of fatty liver, safe-alcohol consumption and absence of hepatitis B/C. The cohort was re-evaluated to assess mortality in 2017. Participants or their households were contacted by telephone/post, and deaths confirmed by home-visits and death certificate review. Cox-regression was used to determine predictors of all-cause mortality (ACM) and cardiovascular mortality (CVM) in those with- and without-NAFLD. RESULTS: 2724 (91.2%) of 2985 original participants were contacted (851-with NAFLD and 1072-without NAFLD). Overall there were 169 (6.2%) deaths [41-deaths among NAFLD (17-cardiovascular; 9-cancer-related; 4-liver-specific; 11-other) and 79-deaths among no-NAFLD (28-cardiovascular; 17-cancer-related; 1-liver-specific; 33-other)]. Metabolic syndrome (MetS), low-education level, higher age and male-gender independently predicted ACM. MetS, increasing age and male-gender independently predicted CVM. NAFLD did not predict either ACM or CVM. In those with NAFLD, MetS and age >55-years were independently associated with ACM, while MetS and male-gender were associated with CVM. CONCLUSION: In this community-based study, increasing age, male-gender and MetS, but not NAFLD, predicted 10-year ACM and CVM. Among those with NAFLD, only those metabolically abnormal were at a higher risk for mortality.