We compared the efficacy and safety of pergolide and bromocriptine in 57 patients with Parkinson's disease (PD) with a declining response to levodopa therapy in a single-blind, crossover study. Patients were placed randomly on the sequence bromocriptine-pergolide (12 + 12 weeks) or vice versa. Regular evaluations using the New York University Parkinson's Disease Scale were performed by a clinician blinded to treatment assignment. Patients' and clinicians' impressions also were recorded. The average daily dose of pergolide was 2.3 +/0- 0.8 mg, and that of bromocriptine was 24.2 +/- 8.4 mg. Significantly greater efficacy was demonstrated by both drugs as adjunctive therapy to levodopa compared with previous treatment of levodopa alone (pergolide, p = 0.0001; bromocriptine, p = 0.0005; Wilcoxon t test). Pergolide was more effective than bromocriptine in daily living scores (p = 0.020) and motor scores (p = 0.038). No difference in dyskinesias, dystonias, and psychosis was observed. Adverse events were more frequent in bromocriptine-treated patients. Most patients and physicians preferred pergolide to bromocriptine. Pergolide as adjunctive therapy to levodopa was more effective than bromocriptine in this short-term trial.