Plasma cholesteryl ester transfer protein (CETP) activity was measured in 52 alcoholics and 38 controls and compared with conventional laboratory markers of alcoholism. Mean daily alcohol intake was 180 g/day among the alcoholics and 10 g/day among the controls. Plasma CETP activity was 26% lower in the alcoholics (P < 0.001) and was inversely correlated with daily alcohol intake (r = -0.288, P < 0.05). CETP activity detected 63% of the alcoholics, and its specificity was 82% if the cut-off point was set at the mean CETP activity of the controls -1 SD. The mean -2 SD gave a very low sensitivity for CETP (8%) and cannot be used as its cut-off point. The sensitivities and specificities of gamma glutamyltransferase, aspartate aminotransferase, alanine aminotransferase, mean corpuscular volume and high-density lipoprotein cholesterol were similar to those of CETP activity when the cut-off point for CETP was mean -1 SD. The results thus indicate that plasma CETP activity is not sufficient as a single marker of alcoholism but could be used as an additional method to detect alcohol misuse, although its wide variation in normal population and the elaborate analysis limit its usefulness.