We investigated differences in features between young-old and old-old patients beginning dialysis. Patients aged over 60 years who had newly started dialysis (121 cases) were studied. These were 71 men and 50 women. The age was 74 +/- 7 years (mean +/- standard deviation). The patients were divided into the young-old group (aged 60 to 74 years) and the old-old group (aged over 75 years), 64 patients were young-old and 57 were old-old. In every patient, the cause of renal failure (non-diabetes/diabetes), body indexes, comorbid conditions, laboratory data of nutritional status, ambulation, cognitive function, and psychosocial status (acceptance of dialysis therapy, marital status, younger cohabitants, and engagement in work) were surveyed. Data were assessed in each group. Intergroup comparisons were made using Student's t-test, the chi 2 test, and the Mann-Whitney's U-test. Diabetes was frequent in the young-old group. The results concerning body indexes, comorbid conditions, and laboratory data reflected age-related phenomena or skew distribution of diabetic patients. Though differences were not statistically significant, both the percentage of patients with inability to walk and the percentage of patients with impaired cognitive function were lower in the old-old. These results could not be interpreted as age-related phenomena or skew distribution of diabetic patients. From the data of psychosocial indexes, it would speculate that the young-old had some advantage over the old-old for initiation of dialysis. We look for the explanation to psychosocial aspects of dialysis therapy.