OBJECTIVES:Nonalcoholic fatty liver disease is the most common chronic liver disease. Fatty pancreas has also been described but is diffi cult to assess. It is now possible to measure pancreatic and liver fat accurately with magnetic resonance imaging (MRI). We aimed to defi ne the normal range of pancreatic fat and identify factors associated with fatty pancreas. In addition, the effect of fatty liver and fatty pancreas on insulin resistance (IR) and pancreatic  -cell function was studied.
METHODS:Fat-water MRI and proton-magnetic resonance spectroscopy were performed on 685 healthy volunteers from the general population to measure pancreatic and liver fat, respectively. On the basis of fasting plasma glucose and insulin levels, the IR and  -cell function were assessed using the homeostasis model assessment (HOMA).
RESULTS:Among subjects without signifi cant alcohol consumption or any component of metabolic syndrome, 90 % had pancreatic fat between 1.8 and 10.4 % . Using the upper limit of normal of 10.4 % , 110 (16.1 % ; 95 % confi dence interval 13.3 -18.8 % ) subjects had fatty pancreas. On multivariable analysis, high serum ferritin, central obesity, and hypertriglyceridemia were independent factors associated with fatty pancreas. Subjects with both fatty pancreas and fatty liver had higher HOMA-IR than did those with either condition alone. Fatty pancreas was not associated with HOMA-β after adjusting for liver fat and body mass index.CONCLUSIONS: In all, 16.1 % of this community cohort of adult Hong Kong Chinese volunteers had a fatty pancreas by our defi nition. Central obesity, hypertriglyceridemia, and hyperferritinemia are associated with fatty pancreas. Individuals with fatty pancreas have increased IR.