The incidence of malnutrition is widely held to be greater in the elderly, but this specific factor has not been extensively studied in elderly dialysis patients. In a 30-month follow-up prospective study, we evaluated the role of nutrition on the outcome of 290 stable hemodialysis (HD) outpatients aged older than 75 years followed up in 20 French HD centers (167 men, 123 women; age, 79.8 +/- 4.2 years; previous time on dialysis, 41 +/- 38 months). On the same day in January 1996, predialysis and postdialysis blood samples were collected according to recommended procedures for dialysis quantification. Normalized protein catabolic rate, dialysis adequacy parameters, and estimation of lean body mass (LBM; expressed as observed/expected LBM values [obs/exp LBM]) were computed from predialysis and postdialysis urea and creatinine levels. Overall survival rates were 80% and 65% after 1 and 2 years of follow-up, respectively, and were significantly less in patients with the lower quartile of obs/exp LBM. In univariate analysis using the Cox proportional hazards model, survival was significantly influenced by age, albumin level, prealbumin level, body mass index, and diabetes, but not by sex, Kt/V, duration of dialysis, cholesterol level, hemoglobin level, or obs/exp LBM. In multivariate analysis, no variable remained significant. Cardiovascular mortality accounted for 52.1% of the patient deaths. We conclude that in elderly HD patients, malnutrition influences overall survival despite adequate dialysis treatment.