Confirmation of symptomatic and disease modifying effects of levodopa using the ELLDOPA study
Bart Ploeger (1,2) and Nick Holford (3)
(1) Modelling & Simulation, DMPK iMed CNSP AstraZeneca R&D Södertälje, Sweden Division of Pharmacology. (2) Leiden/Amsterdam Center of Drug Research, Leiden University, Netherlands. (3) Department of Pharmacology and Clinical Pharmacology, University of Auckland, New Zealand
Objectives: The symptomatic relief of levodopa in Parkinson's disease has been well established, but controversy exists about possible disease modifying effects. In this study, we analyzed the ELLDOPA study using a disease progression model that describes the time course of the disease, placebo and drug effects.
Methods: In the ELLDOPA trial  361 patients with early PD received a carbidopa-levodopa combination 3 times daily at a final daily dose of 150 mg (n=92), 300 mg (n=88), 600 mg (n=91) or a matching placebo (n=90) for a period of 40 weeks followed by withdrawal of treatment for 2 or 4 weeks.
Data were analysed using a (mixed-effects) linear disease progression model . The probability of dropout was described using a hazard function. For the inactive treatment effects a mixture model was applied to distinguish patients who improved (placebo) or worsened (nocebo) and the onset and offset of these effects were described using a Bateman function. A delayed symptomatic effect, resulting in a transient change in the offset of the disease progression curve was combined with an immediate onset disease modifying effect (i.e. persistent change in the slope of the disease progression curve). Both types of drug effect were investigated using either a linear, Emax or dose-insensitive effect. Visual and numerical predictive checks were used for model evaluation.
Results: The analysis confirmed the combined symptomatic and disease modifying effects of levodopa (± 50% reduction in the offset and slope respectively), as described previously using the DATATOP cohort . Approximately 25% of the patients showed a nocebo effect (transient worsening) while the rest had a placebo effect (transient improvement), which is similar to the results of Ma et al.  after analyzing inactive treatment effects in 3 clinical studies (DATATOP, ELLDOPA and TEMPO). The disease progression rate was higher in the population that dropped out. This is consistent with the hypothesis that patients drop out due to relatively high disease progression and was confirmed by the dropout hazard model, showing a reduction in the hazard ratio associated with slowing of disease progression.
Conclusions: We believe that a model based analysis, in which assumptions are made explicit and tested when possible, allows more powerful inferences to be made about the type of treatment effect compared to change from baseline methods. This analysis confirms that levodopa has disease modifying effects.
 Fahn, S., D. Oakes, et al. (2004). "Levodopa and the progression of Parkinson's disease." N Engl J Med. 351(24): 2498-2508.
 Holford, N. H., P. L. Chan, et al. (2006). "Disease progression and pharmacodynamics in Parkinson disease - evidence for functional protection with levodopa and other treatments." J Pharmacokinet Pharmacodyn 33(3): 281-311.
 Ma, S.C. and Holford N. H. (2011). "Quantifying Disease Progress with Inactive Treatments in Multiple Parkinson's Disease Trials." PAGE PAGANZ 11 (2011)