Levodopa Slows Progression of Parkinson’s Disease. External Validation by Clinical Trial Simulation

Abstract

Purpose

To externally validate the model predictions of a DATATOP cohort analysis through application of clinical trial simulation with the study design of the ELLDOPA trial.

Methods

The stochastic pharmacokinetic-pharmacodynamic and disease progress model was developed from the large DATATOP cohort of patients followed for 8 years. ELLDOPA was designed to detect a difference between placebo and levodopa treated arms in the total Unified Parkinson’s Disease Rating Scale (UPDRS) taken at baseline and following 2 weeks levodopa washout after 40 weeks of treatment. The total UPDRS response was simulated with different assumptions on levodopa effect (symptomatic with/without disease modifying capability) and washout speed of symptomatic effect.

Results

The observed results of ELLDOPA were similar to the model predictions assuming levodopa slows disease progression and has a slow washout of symptomatic effect.

Conclusions

This simulation work confirmed the conclusion of the DATATOP analysis finding that levodopa slows disease progression. The simulation results also showed that a dose-related increased rate of progression in Parkinson’s disease, obscured by symptomatic benefit, is very unlikely. Finally, the simulation results also shown that 2 weeks washout period was not adequate to completely eliminate the symptomatic benefits of levodopa.

Appendix

This section described an alternative washout process for the symptomatic effect that we have examined.

Direct Effect Washout Process

We considered a washout process based directly on the time course of effect without the use of linking concentration-effect model. This process is the same as that used by Hauser and Holford (29). The symptomatic effect for the fast washout process (FWO) was assumed to be zero after the last levodopa dose. For the slow washout of symptomatic effect after the last levodopa dose, the time course was predicted by Eq. 14. The symptomatic effect at the time of the last levodopa dose (Tlastdose) was used as the initial value for this process (E_O(Tlastdose)). The change of total UPDRS during the 3 day levodopa dose step down washout period was predicted from the reduced dose and subsequent decrease in Ce and thus E_O(t). The change in total UPDRS over these 3 days is expected to be small because of the long value of TEQL.

$$\begin{aligned} & IF{\text{ }}{\left( {t > Tlastdose} \right)}{\text{ }}THEN \\ & {\text{ }}E_{O} {\left( t \right)} = {\text{ }}E_{O} (Tlastdose) \bullet e{\left( {^{{\frac{{ - Ln(2)}}{{TEQWO}} \bullet {\left( {t - Tlastdose} \right)}}} } \right)} \\ & ENDIF \\ \end{aligned} $$

(14)

The concentration (Ce) washout process is different from the direct effect washout process because the symptomatic effect is computed using the drug concentration at the effect site via a pharmacodynamic model whereas the direct effect washout process models the time course of effect without consideration of the effect site concentration. The direct effect washout process was based upon the method of Hauser and Holford (29) where the drug effect was simply observed after levodopa withdrawal.

In general, the predicted treatment effect size of the direct effect washout process was comparable with the Ce washout process in the low and medium dose arms. A smaller predicted change from baseline in total UPDRS at 42 weeks was seen with the direct effect washout process (Tables III and VI). The Ce washout process gave a closer prediction of the change from baseline at 42 weeks for the high dose arm to the observed change (Table V) than the direct effect washout process (Table VIII). The predicted size of symptomatic effect was also smaller for the direct effect washout process (Tables IV and VII).

Table VI Predicted Differences of Change From Baseline in Total UPDRS at 42 Weeks

Table VII Size of Treatment Effects After 2 Weeks Levodopa Withdrawal Under An Assumption of Slow Washout of Symptomatic Benefits

Table VIII Observed and Predicted Total UPDRS Change from Baseline For Placebo Treatment at 42 Weeks and Difference from Placebo or Levodopa Treatment Arms