The association between self-reported alcohol intake spective assessment of alcohol intake at the time of and the risk of future liver disease was studied in a large diagnosis of liver damage, it has been observed that population-based prospective cohort with 12-year folthe risk of developing liver damage increased by the low-up. Alcohol intake was assessed in 13,285 men and amount consumed, and the minimum alcohol intake women aged 30 to 79 years by a self-administered quesassociated with a significant increase in risk ranged tionnaire. Diagnoses indicating alcohol-induced liver from 40 to 80 g daily in most studies, [3][4][5][6][7][8][9] but for methoddisease (n Å 261) or alcohol-induced cirrhosis (n Å 124) ological reasons the interpretation of these studies is were obtained from death certificates and the hospital difficult. 1 A prospective study based on an American discharge register, and data were analyzed by multiplihealth care program confirmed the dose-dependent ascative Poisson regression models. The total cumulated sociation, but both alcohol intake and occurrence of observation time was 130,558 person-years. The overall incidence rates of alcohol-induced cirrhosis were 0.2% liver disease may have been considerably underreper year in men and 0.03% per year in women. The nadir ported. 10 In another prospective study, heavy drinkers of the estimated relative risk of developing liver disease showed no further increase in risk of developing cirrhowas observed at an alcohol intake of 1 to 6 beverages sis by increasing alcohol intake. 11 per week, and above this level a steep increase in rela-Several studies have shown that women have more tive risk was observed. The risk function was indepenadvanced liver disease at time of diagnosis, run a more dent of age and stable over time. The level of alcohol severe course of the disease, and report a shorter hisintake above which the relative risk was significantly tory of alcohol abuse and a smaller intake of alcohol greater than 1 was observed at 7 to 13 beverages per than men. 5,[12][13][14][15][16][17][18][19] This has been taken as evidence for week for women and 14 to 27 beverages per week for women being more prone to develop alcohol-induced men. Women had a significantly higher relative risk of developing alcohol-related liver disease than men for liver disease than men. If we assume that men and any given level of alcohol intake. We observed a dosewomen have the same threshold for development of dependent increase in relative risk of developing alcoliver damage and that women drink less than men, hol-induced liver disease for both men and women, with then those women who develop alcohol-induced liver the steepest increase among women. In the general popdisease would on average have a lower alcohol intake ulation, self-reported current alcohol intake is a good than men with alcohol-induced liver disease. To estipredictor of the future risk of alcohol-induced liver dismate the risk at the individual level, the distribution ease. (HEPATOLOGY 1996;23:1025-1029.) of alcohol intake in the population must therefore be The association between alcohol intake and alcoholtaken into consideration. A few retrospective populainduced liver disease is unquestionable, but the dosetion studies, however, do suggest a steeper risk funceffect relation (the risk function) and its variation with tion in women, 5,9,12 but this is not supported by the sex remain controversial. 1,2 In studies based on retroprospective study by Klatsky and Armstrong. 10 Our aim was to assess the association between current alcohol intake and future risk of developing alco-Abbreviations: ICD, International Classification of Diseases.