• alcohol;
  • analysis of variance;
  • cirrhosis;
  • logistic regression;
  • reproducibility

Abstract: Our aims were to design a reproducible method of measuring life-time alcohol consumption in patients with cirrhosis, and to assess the risk of liver decompensation associated with alcohol intake using a case-control design and a multivariate analysis. We studied 439 patients (“cases”) with decompensated cirrhosis, and 233 with compensated cirrhosis (“controls”). Mean life-time daily amount and duration of alcohol intake were measured by a standardized questionnaire, whose reproducibility, assessed by interviewing 75 relatives, was 70% for daily alcohol intake and 84% for duration of intake. Better reproducibility was found by re-interviewing patients at discharge from hospital. Daily alcohol intake was significantly higher in males, younger patients and patients with liver decompensation. After stratification according to the average life-time daily alcohol intake, we found a significant increase in the risk of liver decompensation from 125 g ethanol intake per day onwards. No association was found between duration of alcohol intake and risk of liver decompensation. We conclude that alcohol intake can be reliably and reproducibly measured: in patients with cirrhosis, increased alcohol intake is associated with increased risk of liver decompensation, with a significant dose-effect above a daily intake of 125 g ethanol.