This multicentre, randomized double-blind study was undertaken to assess the efficacy of corticosteroids as a palliative treatment of intestinal obstruction due to advanced and incurable cancer. Thirty-one French palliative care units agreed to participate in the study and 12 actually recruited at least one patient. To be included, patients had to have an advanced cancer with a surgically inoperable bowel obstruction and to have received no specific anticancer therapy within the preceding 28 days. They had to fulfil at least three of the following criteria: vomiting at least twice a day; colicky abdominal pain; no flatus for 12 h or more; no stool for at least 4 days, faecal impaction being excluded; intestinal distension; air-fluid levels or absence of gas in the colon on an abdominal radiograph. Patients were randomized in three groups to receive either a placebo for 3 days (group A), or methylprednisolone 240 mg daily for 3 days (group B) or methylprednisolone 40 mg daily for 3 days (group C). Symptoms were assessed daily but success or failure of the treatment was assessed on day 4, according to the disappearance or persistence of symptoms. Fifty-eight patients were randomized, of whom 52 were able to be evaluated. Details of symptoms and associated treatments are described below. Of 40 patients without a nasogastric tube, symptoms were relieved in 68% of cases versus 33% among placebo-treated patients (P = 0.047). In 12 patients who had a nasogastric tube already in place, the results are less significant (60% versus 33% with P = 0.080). Because of the small sample size, no conclusions can be reached about the relative efficacy of low versus high-dose treatment regimes.