Levodopa pharmacokinetic-pharmacodynamic modeling and 6-[F-18]levodopa positron emission tomography in patients with Parkinson's disease

Objective: Parameters of a pharmacokinetic-pharmacodynamic (PK-PD) model of levodopa have been claimed to reflect the magnitude of the dopaminergic deficit in patients with Parkinson's disease. The aim of this study was to correlate such parameters with positron emission tomography (PET) with levodopa tagged with 6-fluorine 18, an established imaging method for striatal dopaminergic neurons. Methods. Twenty-three patients in different disease stages (Hoehm. and Yahr stage 2.5-5 [Hoehn MM, Yahr MD. Parkinsonism: onset, progression and mortality. Neurology 1967;4:427-42]; median duration, 12 years) were studied. PK-PD modeling followed a single oral dose of levodopa/benserazide. The sum score of the Columbia Rating Scale (CURS Sigma) was used for clinical assessments. A nonparametric effect compartment approach assuming a sigmoidal E-max model was applied to the PK-PD analysis of plasma levodopa concentrations and corresponding CURS Sigma. Thereafter 6-[F-18]levodopa PET was performed, and the influx rate constants (k(c)) for the putamen and the caudatus region were correlated with the median effective concentration (EC50) and the equilibrium half-life (T-eq) of the PK-PD model. Results: (1) A significant correlation was observed between PK-PD parameters or with k(c) putamen as the dependent variable and the duration of the disease as the independent variable, which explains 33% of the variability of the EC50, 42% of the variability of T-eq, and 36% of the variability of k(c). (2) Significant correlations were observed between k(c) and either EC50 or T-eq, yielding the closest correlation for the putamen region (r = -0.47, P <.05; and r = 0.55, P <.01; respectively). Conclusions: Our findings show that key parameters of a PK-PD model of levodopa were in fairly close agreement with imaging of dopaminergic neurons by 6-[F-18]levodopa PET. However, although PK-PD modeling of levodopa has been proven as a useful investigation of approaches aimed to restore dopaminergic deficits or to monitor disease progression, this modeling cannot serve as a pathomorphologic surrogate for the loss of striatal dopaminergic neurons