Adverse reactions to the drugs employed in the National Cooperative Crohn's Disease Study were sought prospectively at each patient visit and by retrospective review of all patient charts. Prednisone caused evident side effects in over 50% of patients on high-dose suppressive therapy and in approximately one-third of patients on prophylactic dose. Thirty-two percent of patients on high-dose, and 26% on prophylactic-dose prednisone required dose reduction or withdrawal because of side effects. Comparable figures for sulfasalazine were 14% and 12%, and for azathioprine 32% and 20%. The incidence of nausea, vomiting, or anorexia among patients taking sulfasalazine was 46% and 34%, on high and low dose respectively; however, this incidence was no different than that observed among patients taking placebo. These symptoms occasioned withdrawal from the study of only 4% and 3% of patients on high and low doses of sulfasalazine, respectively. Azathioprine produced leukopenia at a dose of 2.5 mg/kg body weight in 15% of patients and the mean white cell count, lymphocyte count, granulocyte count, and hematocrit all fell significantly in patients on this dose. Pancreatitis occurred in 5% of patients taking azathioprine but in no other patients. Sulfasalazine proved to be the safest effective suppressive drug for Crohn's disease. Prednisone toxicity, though substantial, is acceptable in view of its demonstrated suppressive efficacy. Azathioprine was approximately as toxic as prednisone but no more effective than placebo in suppressing active disease. None of the drugs was effective prophylactically, and all showed appreciable long-term toxicity.