We assessed upper gastrointestinal anatomy and function with contrast radiology and antroduodenal manometry in 51 children with chronic intestinal pseudo-obstruction (CIP) prior to entering these patients into an open-label outpatient trial of cisapride. The diagnosis of CIP was based on characteristic symptoms requiring special nutritional support (parenteral in 30, tube feeding in 12) or interfering with daily activities (documented by diary in nine). At a time the subjects were not acutely ill, antroduodenal pressures were recorded for > 4 h fasting and > 1 h after a complex liquid meal. Results were categorized by the most prominent manometric abnormality as myopathy (n = 6), absent migrating motor complex (MMC) (n = 27), failure to induce fed pattern (n = 7), MMC plus discrete abnormalities (n = 7), and postprandial duodenal hypomotility (n = 4). Patients in the first two categories did not have effective MMCs, but those in the last three categories did. Compared to children without MMCs, those with MMCs rarely required parenteral nutrition (p < 0.001). All children were treated with oral cisapride 0.2 mg/kg/dose t.i.d., and evaluated every 2 months for up to 1 yr. Of 49 evaluable subjects, the final global assessment was unchanged in 25, fair (improved symptom score) in 17, or excellent (change from TPN to tube feeding or tube feeding to oral feeding) in seven. Children with MMCs (13 of 18) responded more often to cisapride than those without MMCs (11 of 31), p < 0.02. All four subjects with postprandial duodenal hypomotility had excellent responses. Children with normal diameter bowel responded more often than those with dilated bowel, p < 0.004. To summarize, in children with CIP, absence of the MMC was associated with need for greater intensity of nutritional support and decreased response rate to cisapride. The response to cisapride was highly variable within the study group, but often could be predicted by the presence or absence of bowel dilation and MMCs.