The substantial prevalence of malnutrition in the hospitalized patient population has only been recently recognized. Preoperative nutritional and immunological assessment was performed prospectively on admission in 64 consecutive surgical patients. Factors measured included weight loss, triceps skinfold, midarm muscle circumference, creatinine-height index, serum albumin level, serum transferrin level, total lymphocyte count, serum complement level, serum immunoelectrophoresis, lymphocyte T rosettes formation, neutrophil migration, and delayed hypersensitivity. Using these criteria for malnutrition, 97% of the patients had at least one abnormal measurement and 35% had at least three abnormal measurements. Patients were monitored for complications during their hospital course. Serum albumin level, serum transferrin level, and delayed hypersensitivity reactions were the only accurate prognostic indicators of postoperative morbidity and mortality. Substantial unrecognized malnutrition exists in the surgical patient population. An isolated indicator of malnutrition should be interpreted with caution. The visceral protein compartment (serum albumin and serum transferrin levels and delayed hypersensitivity) is the most accurate prognostic indicator of postoperative morbidity and mortality. Perioperative nutritional support may reduce operative morbidity and mortality in the malnourished operative candidate.