Nutritional assessment and adequacy of spontaneous dietary intake was evaluated in thirty-seven clinically stable hospitalized patients with alcoholic liver cirrhosis. About two-thirds of the patients had ascites or oedema, or both, and, therefore, body weight could not be used for assessment of nutritional status. Lean body mass (LBM; measured by three consecutive 24 h creatinine excretions) was 62 (range 40–95)% of reference values, mid-arm-muscle area (MAMA) was 70 (range 43–115)% and triceps skinfold (TSF) was 45 (range 20–113)% of reference values (all median values). In patients without ascites or oedema, or both, there was a rectilinear correlation between body weight and LBM and between body weight and MAMA (r 0.93 and 0.85 respectively). In patients with ascites or oedema, or both, the correlation between body weight and LBM was poor as could be expected. We suggest that LBM is a useful measure of nutritional status when body weight is unreliable because of ascites or oedema, or both. Energy balance for the group was calculated from energy intake recorded by a 24 h dietary recall and energy expenditure calculated by the factorial method. Median intake was 102 (range 34–176)% of expenditure. N loss was calculated from the average of three 24 h urea excretions. Protein intake was calculated from the 24 h dietary recall. The N balance was positive in the patients as a group (median intake was 120 (range 26–183)% of output). The most malnourished patients tended to have the most positive N balance which was due to a significantly lower N excretion. The protein requirement for N balance was 0.83 (SE 0.05) g/kg per d and only at an intake above 1.20 g/kg per d were all patients in positive N balance. The median intakes of thiamin, folacin, vitamin D, vitamin E, Mg, and Zn were judged to be insufficient. It is concluded that impaired nutritional status is common among patients with liver cirrhosis, even in a stable clinical condition. It is suggested that nutritional status in these patients is evaluated by dietary recalls, in combination with measurement of body weight in patients without ascites or oedema, or both, or in combination with determination of LBM by three 24 h creatinine excretions in patients with ascites or oedema, or both. Criteria for selection of patients that might benefit from nutritional therapy are discussed.