In this community-based study, unsafe alcohol intake was seen in 9.1%, and the problem was almost exclusively in males at both baseline and follow up. Among unsafe drinkers, the annual incidence of AFL was 7.7%. Only male gender was significantly associated with development of new onset AFL. Of those with AFL at baseline, and changed their drinking status from unsafe to safe or no alcohol, few had resolution of fatty liver after 7 years.
Unsafe alcohol consumption was almost exclusively seen among males in our cohort. Men are more likely to engage in alcohol use and are at a much greater risk of developing alcohol use disorder (AUD) than women throughout the world (10). In Asian countries, women are known to abstain from using alcohol due to social and cultural reasons (11). The unsafe alcohol consumption rate observed in the present study is slightly higher than in previous reports from urban community surveys in Sri Lanka (9.1 vs 5.2% and 6.2%) (12, 13), and estimates of the WHO global status report on alcohol and health (which used data from the WHO global survey on Alcohol & Health (2012) in addition to other surveys conducted in the respective countries) which estimated the 12-month prevalence of AUD among men in Sri Lanka to be 5.6% (14).
The community prevalence of AFL in the present study was 2.9%. The reported median prevalence of AFL from China is higher at 4.5% (15). We could find no other reports of community prevalence data on AFL. This is in contrast to the wealth of data available regarding the global and regional burden of non-alcoholic fatty liver disease (16).
We invited individual who had AFL at baseline to periodically (every 6 months) visit a community-based clinic for reinforcement of healthy lifestyle practices to abstain from unsafe alcohol use. However, those with AFL at baseline attended these clinics very infrequently.
Male gender was significantly associated with new onset ALF. We also found that the central obesity, over-weight state and raised ALT (> 2 x upper limit normal) were independently associated with AFL at baseline. However, the baseline over-weight state and central obesity were not associated with new onset AFL. Some observational studies have suggested that being overweight is an independent risk factor for the development of ARLD (17, 18). However, the Dionysos Study, a large cohort study on the prevalence of alcohol habits and chronic liver disease in the general population from Italy, did not report an association between body weight or body mass index and risk of ARLD (19).
We observed having metabolic features such as centrally obesity or being over-weight, or having DM and raised TG at baseline was significantly associated with AFL. However, out of the metabolic features, only central obesity or over-weight status was independently associated with AFL at baseline. The independent association of central and general obesity with AFL raises the possibility of their disease being NAFLD and AFL overlap. This in agreement with the new proposed definition of fatty liver disease. A consortium proposed metabolic (dysfunction) associated fatty liver disease (MAFLD) as more appropriate nomenclature of this disease and they created a simplified and easily applicable comprehensive proposal for redefining of fatty liver disease (20).
There have been no previous large, prospective community-based studies, from emerging economies, that report on new onset AFL. The strengths of the present study include the robust design and that over 70% of the relatively large baseline population presented for re-evaluation. One limitation is that information on alcohol consumption was obtained only by direct questioning of the participants. This may have led to under-reporting with consequent underestimation of the prevalence and incidence of AFL in both the inception and follow up cohorts. The initial cohort also comprised predominantly of Sinhalese (96.2%) and only a small proportion of other ethnicities (Tamils 1.3%, Muslims 1.3%, Burghers 1.3%) (13). Therefore, we could not analyse any ethnic variability in the occurrence of AFL. Only the ALT was measured among the participants and aspartate aminotransferase (AST) was not measured due to lack of funds. Having both AST and ALT would have been useful to assess alcohol induced liver injury among the ALF group. We also could not investigate liver-related outcomes of those detected to have AFL at baseline due to lack of resources.
In conclusion, in this prospective, community-based study, we observed unsafe drinking to be almost exclusively among males. The annual incidence of AFL among unsafe drinkers was 7.7% after seven years of follow-up. AFL at baseline was associated with male gender, over-weight state and central obesity. New onset-AFL was only associated with male gender.