The study included 19 parkinsonian patients presenting with fluctuations in motor performance after L-dopa therapy (9.9 years on average, range 6 months to 17 years) and 2 other patients not pretreated with L-dopa. In all patients of the first group, after 3 months on stable 'optimal' dosage schedule, the previous L-dopa treatment was abruptly replaced, dose for dose, from one day to another by Madopar HBS, a new controlled-release form of Madopar. The first clinical assessment was performed just before the beginning of the HBS treatment. Then, 4-6 weeks were allowed to optimize the dosage schedule of Madopar HBS before performing a second assessment. Long-term therapeutic effects were systematically evaluated after 6 and 12 months. Catamnestic evaluation took place at the last check-up for patients treated for more than 1 year. Of the 19 patients with fluctuations 1 dropped out after 7 days due to lack of cooperation. Thus, 18 of these patients were evaluable. At the end of the dose adaptation phase 12 patients (2/3) were better controlled by the new dosage form and were eligible for a long-term follow-up aiming to evaluate the maintenance of the benefit. It was concluded that in most of the patients the initial benefit was maintained during 1 year and even up to 2 years and 7 months for some of them. The best results were obtained in patients with mild to moderate forms of fluctuations (early stage). Positive results in de novo cases need confirmation.