OBJECTIVES: To assess the effects of chronic drinking on detection of low amplitude signals, and to determine the relation between late ventricular potentials (LVP) and liver biopsy findings. DESIGN: Prospective study. SETTING: General hospital. PATIENTS: 41 consecutive chronic alcoholics without known pre-existing heart disease. METHODS: About four days after each patient's last alcoholic drink, ECG, echocardiography, signal averaged electrocardiogram, liver biopsy, and blood tests were performed. RESULTS: Twenty eight per cent of patients had evidence of LVP. There was a correlation between the percentage of steatosis of the hepatic biopsy and the amplitude of the last 40 ms of average QRS (P = 0.04), the duration of the terminal low amplitude QRS signal (P = 0.05), and the number of positive criteria of late potentials (P = 0.02). CONCLUSIONS: Chronic drinking sufficient to cause steatosis is associated with positive findings on the signal averaged ECG.