This prospective nonintervention single-center study was undertaken to investigate the role of protein-energy malnutrition (PEM) as a risk factor for morbidity in patients on long-term hemodialysis. Thirty-seven patients from the renal unit of Tygerberg Hospital, Tygerberg, South Africa, were studied for a mean period of 26 months. Morbidity was the main outcome and was defined as the number of hospitalizations and days of hospitalization per patient per year. Investigations included 4-monthly determinations of interdialytic protein catabolic rate (PCR), dietary intake of protein and energy, blood levels of albumin and urea, lymphocyte count, adequacy of dialysis (Kt/V), body weight, intradialytic weight loss, fat mass (FM), fat-free mass (FFM), body mass index (BMI), and bone-free arm muscle area (BF-AMA). A PEM composite score was derived from postdialysis serum albumin, BF-AMA, FM, FFM, and BMI. All-cause morbidity as defined by number of hospitalizations (see text for other definitions of morbidity) showed a significant correlation with the mean and baseline PEM score (P <.01), and a negative correlation with predialysis and postdialysis serum albumin (P <.05) and age (P <.05). There was no significant relationship with PCR, percentage intradialytic weight loss, Kt/V, reuse of dialyzer, period on maintenance hemodialysis, sex, race, and type of dialyzer membrane. When "only infection-related" morbidity was considered, the factors that showed a significant correlation were the mean (P <. 001) and baseline PEM score (P <.01), and percentage intradialytic weight loss (P <.01). There was no significant deterioration in the nutritional status of patients followed up for at least 24 months. It is concluded that infection-related morbidity was associated most strongly with the PEM score and the percentage intradialytic weight loss. The results suggest that PEM is one of the important contributing factors to morbidity, possibly via an effect on the immune system and infection.