Twenty-one adult patients were randomised to receive ghrelin on days 1 and 8 and placebo on days 4 and 11 or vice versa, given intravenously over a 60-min period before lunch: 10 received 2 mg kg À1 (lower-dose) ghrelin; 11 received 8 mg kg À1 (upper-dose) ghrelin. Active and total ghrelin, growth hormone (GH), and insulin-like growth factor 1 levels were monitored at baseline (4 -5 days before day 1), during treatment days, and at end of study (day 17/18). Drug-related adverse events (assessed by NCI-CTC-toxicity criteria and cardiac examination) did not differ between ghrelin and placebo. No grade 3/4 toxicity or stimulation of tumour growth was observed. The peak increase of GH, a biological marker of ghrelin action, was 25 ng ml À1 with lower-dose and 42 ng ml À1 with upper-dose ghrelin. Morning fasting total ghrelin levels were higher (Po0.05) for upper-dose patients at end of study (3580 pg ml À1 ) than at baseline (990 pg ml À1 ). Insulin-like growth factor 1 levels did not change. At day 8, 81% of patients preferred ghrelin to placebo as against 63% at the end of study. Nutritional intake and eating-related symptoms, measured to explore preliminary efficacy, did not differ between ghrelin and placebo. Ghrelin is well tolerated and safe in patients with advanced cancer. For safety, tolerance, and patients' preference for treatment, no difference was observed between the lower-and upper-dose group.