We compared the tolerance of a diet providing 20 g/d lactose and a lactose-free diet in 14 patients with short-bowel syndrome with either the colon in continuity (group A, n = 8) or a terminal jejunostomy (group B, n = 6). Lactose tolerance was studied after a single 20-g lactose load in the fasting state, and during two 3-d periods during which the subjects consumed their usual diet plus either 20 g/d lactose, with no more than 4 g/d as milk, or no lactose. Records and measurements included symptoms, fecal weight, and during the 8 h after the lactose load, breath-hydrogen excretion (group A) or lactose and hexoses flow rates in stomal effluents (group B). Results are expressed as medians with ranges in parentheses. Lactose absorption was 61% (0-90) in group A and 53% (18-84) in group B, and no symptoms of intolerance were noticed. During the lactose-rich diet as compared to the lactose-free diet, no symptoms were noticed nor was there any worsening of diarrhea: 1534 g/d (240-4760) versus 1466 (1590-7030) in group A, and 4122 g/d (1730-6830) versus 3496 (1590-7030) in group B. We conclude that a diet providing 20 g/d lactose with no more than 4 g/d as milk is well tolerated in the majority of patients with short-bowel syndrome, and that a lactose-free diet has usually no benefit in these subjects.