A randomized, prospective trial in 90 de novo parkinsonian patients showed that 4 years' treatment with lisuride resulted in significantly fewer end-of-dose disturbances and peak-dose dyskinesias, but also less improvement in parkinsonian disability, than with levodopa. Early combination of lisuride and a low dose of levodopa, during a 4-year follow-up, resulted in a therapeutic response equal to that achieved with high-dose levodopa alone, but significantly fewer end-of-dose failures and dyskinesias. Thus it seems advisable that treatment should begin in the early phase of the disease with a dopamine agonist such as lisuride combined with a low dose of levodopa.