Objective: To determine the efficacy and safety of octreotide for treatment of refractory, profuse diarrhea in patients with the acquired immunodeficiency syndrome (AIDS). Design: A prospective, open-label study. Setting: Inpatient metabolic units of four university medical centers. Patients: Fifty-one patients infected with human immunodeficiency virus (HIV) who had uncontrolled diarrhea (≥ 500-mL liquid stool per day) despite treatment with maximally tolerable doses of antidiarrheal medications. Intervention: After initial baseline studies, patients received octreotide, 50 μg every 8 hours for 48 hours. If stool volume was not reduced to < 250 mL/d, the dose of octreotide was increased stepwise to 100, 250, and 500 μg. Main Results: Fifty men and one woman (mean age, 36.3 ± 1.1 years) entered and completed the 28-day protocol (14 days of inpatient therapy and 14 days of outpatient therapy). Stool frequency and volume decreased significantly (6.5 ± 0.5 stools per day on day 0 compared with 3.8 ± 0.3 stools per day on day 21 [P < 0.001] and 1604 ± 180 mL/d on day 0 compared with 1084 ± 162 mL/d on day 14 [P < 0.001], respectively). Twenty-one patients (41.2%) were considered to be partial or complete responders (reduction in daily stool volume by ≥ 50% of initial collections or reduction to ≤ 250 mL/d). Of the 21 responders, 14 (67%) had no identifiable pathogens at initial screening compared with 9 of 30 (30%) nonresponders (P < 0.01). Conclusion: Patients with AIDS-associated refractory watery diarrhea, especially those without identifiable pathogens, may respond favorably to subcutaneously administered octreotide. This drug deserves further study in a randomized, placebo-controlled trial.