OBJECTIVETo compare the predictive value of liver enzymes and alcohol consumption for determining risk of type 2 diabetes (T2DM).

METHODSA cross-sectional study was conducted in Zhengzhou with a total of 2, 693 men.Participants' height, weight, and histories of smoking and drinking were recorded. Levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), gamma-glutamyl transferase (GGT) and blood glucose, as well as related metabolic indexes were detected.

RESULTSModerate daily alcohol consumption (more than 35 g ethanol/week and less than 140 g ethanol/week) decreased the risk of type 2 diabetes (OR =0.376, 95% CI:0.306 -0.463, P less than 0.05) but increased risk for higher levels of GGT and ALT (OR GGT =3.012, 95% CI:2.357-3.849, Pless than 0.01; ORALT =1.473, 95% CI:1.043-2.081, Pless than 0.05). In joint analyses of alcohol consumption and liver enzymes, the group of nondrinkers/light drinkers (less than or equal to 35 g ethanol/week) in the fourth quartile of GGT levels had the highest risk for type 2 diabetes (OR =12.219, 95% CI:6.217-24.016, P less than 0.01). The relationship of ALT and daily alcohol consumption with the risk of type 2 diabetes was almost the same as that of GGT (nondrinkers/light drinkers in the fourth quartile of ALT levels (OR =5.357, 95% CI:3.070-9.350, P less than 0.0 1).

CONCLUSIONGGT, ALT and daily alcohol consumption were independently associated with risk of type 2 diabetes. Nondrinkers/light drinkers with the highest levels ofGGT orALT were at high risk of type 2 diabetes.