Placebo-controlled adjunctive trial of pramipexole in patients with bipolar disorder: targeting cognitive dysfunction

Abstract

Objective: Patients with bipolar disorder suffer from significant cognitive impairment that contributes directly to functional disability, yet few studies have targeted these symptoms for treatment, and the optimal study design remains unclear. We evaluated the effects of the dopamine D₂/D₃ receptor agonist pramipexole on cognition in bipolar disorder.

Method: Fifty stable outpatients with DSM-IV-diagnosed bipolar I or bipolar II disorder enrolled in an 8-week, double-blind, randomized, placebo-controlled cognitive enhancement trial between July 2006 and April 2010. Patients completed neurocognitive testing at baseline and at week 8, and the primary outcome measures were change scores calculated for each of the 11 tasks. Symptoms and side effects were monitored weekly.

Results: Forty-five patients completed the study (placebo, n = 24; pramipexole, n = 21), and groups were well matched on demographic and clinical features. Primary cognitive analyses indicated no compelling cognitive benefit of pramipexole versus placebo; however, secondary analyses highlight several important methodological issues for future trials and identify a subgroup of patients who might benefit more readily from cognitive enhancement strategies. This outcome suggests that the study design played a very important role in the results-implying a failed rather than altogether negative trial. Specifically, we found that even very subtle, subsyndromal mood symptoms at baseline had a significant influence on the degree of improvement due to active drug, with strictly euthymic patients faring best (multivariate analysis of variance, P = .03 in euthymic subgroup). In addition, the extent of baseline cognitive impairment also contributed to the likelihood of treatment response. Finally, concomitant medications may weaken, or in some cases enhance, response to cognitive treatment and should be accounted for in study design.

Conclusions: Although our results point toward a lack of clear effect of pramipexole on cognition in bipolar patients, our data revealed a potentially beneficial effect of pramipexole in a subgroup, providing some enthusiasm for pursuing this line of research in the future. Moreover, this study emphasizes the importance of rigorous subject selection for cognitive trials in bipolar illness. Future studies will be necessary to determine the possible clinical and functional implications of these results.

Trial registration: clinicaltrials.gov Identifier: NCT00597896.

Conflict of interest statement

Disclosures: Drs. Braga, Nnadi, and Stearns have no conflicts of interest to disclose. Yaniv Shaya has no conflicts of interest. Dr. Burdick has served on the speaker’s bureau for Schering Plough/Merck. Dr. Malhotra has received grant support from Eli Lilly; served as a consultant to Wyeth, PGx Health, and Eli Lilly; and has been on the speaker’s bureau for Schering Plough/Merck and Arbor Scientia (Speracor Inc).

© Copyright 2012 Physicians Postgraduate Press, Inc.

Figures

Figure 1

Subject Flow Diagram indicating subject enrollment and inclusion.

Figure 2. Effect Size of Neurocognitive Change by Treatment Group (All Completers)

All 46 completers are included in the primary analysis with results shown in this figure. The x-axis labels the individual neurocognitive measures included in the analyses. The Y-axis represents the effect size of the change from baseline to week 8 on each of these measures. Effect size is depicted as Cohen’s D. The pattern of change indicates some benefit of pramipexole over placebo but only one measure reaches modest statistical significance (Stroop Color; p=0.03).

Figure 3. Effect Size of Neurocognitive Change by Treatment Group (Euthymic Subgroup)

Only the euthymic subgroup is included in these results. Again the x-axis indicates specific cognitive tasks and the y-axis scale is Cohen’s D, estimating effect size changes in each treatment group. Here the overall MANCOVA achieves statistical significance (p=0.03), as do two specific neurocognitive measures (Stroop Color; p=0.02; and WAIS Digits Backward; p=0.01).

Figure 4. Enhanced Effect Size Change in the Euthymic Subgroup Treated with Pramipexole

The X-axis indicates individual neurocognitive measures and the Y-axis again depicts effect size in Cohen’s D. The solid red bars represent all patients who were assigned to pramipexole (n=21) and the striped red bars represent the euthymic subgroup who received pramipexole (n=16). The results indicate enhanced effect of pramipexole in the context of euthymia on 8/11 measures.