Our study found a NAFLD prevalence of nearly 18% despite this Indian Tamil ethnic population being from a rural estate area of Sri Lanka with wages earned mainly through manual labour, and where limited access to food, sanitation and transport still exists. Like in other studies, NAFLD was associated with constituent features of the metabolic syndrome. Although much lower than the 33% prevalence found in a community based study in a relatively sedentary, urban Sri Lankan population , the figure is well within the range of 16% to 23% identified by other population based studies that have used ultrasonagraphy for diagnosis of NAFLD [9, 10]. A prevalence of NAFLD similar to our study has been reported from Hong Kong, Korea and Malaysia . However, unlike in our population, access to excessive amounts of food and low activity levels were important features in these studies. Thus, our study strengthens the existing evidence for Asian Indians' increased predisposition to visceral fat accumulation; a feature that may be present from birth [11, 12].
Although liver biopsy remains the gold standard to confirm diagnosis of NAFLD, this is impractical in epidemiological studies. Non-invasiveness and low cost of ultrasonography makes it the most acceptable technique to diagnose fatty liver in field epidemiological studies. High sensitivity for diagnosing fatty liver has been reported using imaging techniques, including ultrasonography . The ultrsonographic criteria we used to detect fatty liver have an adequate threshold for detection of steatosis when more than 33% of hepatocytes contain fat on liver histology . Exclusion of those with evidence of hepatitis B and C infections and an alcohol intake above the safe limits are also prerequisites to diagnose NAFLD . One limitation of our study was obtaining information on alcohol consumption only by direct questioning of the subject. Another was that we were unable to assess inter-observer reliability between the sonographers before the commencement of the study.
Male sex was a significant independent predictor of NAFLD in our study. Though early studies in Western populations found that NAFLD was more prevalent in women, a more recent population-based study from the USA has shown that men have a higher prevalence of presumed NAFLD, based on elevated aminotransferases . Another recent study from Korea also showed a higher prevalence of NAFLD among men compared to women . Undisclosed alcohol abuse, which has been given as one possible explanation for the higher prevalence of NAFLD in men, may be relevant in this study too. 18% of men reporting weekly alcohol consumption above safe limit is one indication for this. However, it is also known that women in the estate sector in Sri Lanka are more physically active compared to the estate males; 65% of the plantation labour force is constituted of female tea pluckers . Abdominal obesity which is considered to be an important risk factor for metabolic syndrome and related diseases was also more prevalent in estate males than females, though not significantly. These are possible explanations for finding male sex as an independent risk factor for NAFLD in the present study.
There is increasing evidence for a genetic basis for the development of NAFLD. Schwimmer et al. report familial factors as a major determinant of NAFLD . Others have reported racial and ethnic differences in the prevalence of the condition [20, 21]. Petersen et al. have shown that polymorphisms C-482T and T-455C in APOC3 are associated with nonalcoholic fatty liver disease and insulin resistance in healthy Asian Indian men . Our finding of a high NAFLD prevalence in a rural, physically active, predominantly Tamil estate population from Sri Lanka supports results of studies reporting ethnic variations in susceptibility.