To study the possible relationship between central levodopa influx and short-term antiparkinsonian and dyskinetic responses to levodopa in patients with Parkinson disease.The clinical response to a single oral dose of standard and controlled-release levodopa/carbidopa was assessed in 12 patients with Parkinson disease complicated with motor fluctuations and dyskinesias. Plasma concentrations of levodopa and large neutral amino acids were determined, and the theoretical central levodopa influx was calculated using a model based on competitive inhibition of substrates to cross the blood-brain barrier.The mean (SD) central levodopa influxes at the onset of the antiparkinsonian clinical effect were 19.7 (10.9 x 10(-3) and 19.1 (7.4 x 10(-3) nmol min(-1) g(-)1 (P >0.1) for standard and controlled-release levodopa, respectively. The mean (SD) central levodopa influxes at the onset of choreic dyskinesias were 20.1 (8.2 x 10 (-3) and 19.9 (10.8 x 10(-3) nmol min (-1) g(-)1 (P 9 0.1) for standard and controlled-release levodopa, respectively. During the tests, choreic dyskinesias were associated with a central levodopa influx of 10 x 10(-3) nmol min(-1) g(-1) or greater, and foot dystonia occurred with a central levodopa influx less than 9 x 10(-3) nmol min(-1) g(-1).The clinical response to levodopa in patients with advanced Parkinson disease may be related to central levodopa influx. We found no differences in the central levodopa influx threshold for clinical improvement with different levodopa formulations. The central levodopa influxes at the onset of choreic dyskinesias and antiparkinsonian effect were similar. Choreic dyskinesias and foot dystonia were associated with high and low central levodopa influx, respectively.