Body composition, postprandial symptoms, and social performance were assessed in 41 patients who were free of tumors 16 to 63 months (median, 41 months) after total gastrectomy with Roux-en-Y esophagojejunostomy (n = 15) or jejunal interposition (n = 26). There were no significant differences with respect to age, sex, initial tumor stage, interval since operation, and premorbid weight/height2 (body mass index). The lowest postoperative body mass index (BMI) was 72% +/- 3% of the preillness BMI in patients with Roux-en-Y reconstruction and 79% +/- 2% in patients with jejunal interposition (p less than 0.05). At the time of the study the relative BMI was 88% +/- 2% of the preillness BMI in patients with jejunal interposition but only 81% +/- 3% in patients with the Roux-en-Y reconstruction (p less than 0.01). Muscle mass and lean body mass estimated from anthropometric and bioelectric impedance measurements were correlated with sex (p = 0.0001) and with the mode of reconstruction (p = 0.02) independently, which was confirmed by multiple linear regression. The postprandial symptoms were not significantly associated with changes in body composition except for an inverse relationship between Sigstad's dumping score and the extracellular mass/body cell mass ratio (r = 0.553; p = 0.0002). Of the patients under 60 years of age, 10 of 15 patients with jejunal interposition and two of eight patients with Roux-en-Y reconstruction were back at work (p = 0.057). The persons who had resumed their work had a significantly higher relative BMI (90% +/- 2% vs 82% +/- 3%), lean body mass (53 +/- 3 kg vs 46 +/- 3 kg), and muscle mass (25 +/- 2 kg vs 21 +/- 1 kg) than persons in early retirement. We concluded that preserving the duodenal transit should be a main objective of gastric replacement after total gastrectomy.