Clinical evaluation of pramipexole in advanced Parkinson's disease: Results of a double-blind, placebo-controlled, parallel-group study

Pramipexole is an aminobenzathiazol-derived dopamine agonist with novel properties.¹ It acts as a full agonist on the D₂ receptor family with preferential affinity to the D₃ family.² These properties suggest its utility in Parkinson's disease (PD).

We studied pramipexole pharmacokinetics, efficacy, safety, and tolerability for 9 weeks in 55 patients with early PD who were not previously treated with levodopa.³ Pharmacokinetic data are reported separately. In the earlier study, the 28 de novo PD patients receiving pramipexole experienced a 140% improvement from baseline(p = 0.002) on the Unified Parkinson Disease Rating Scale (UPDRS) Activities of Daily Living (ADL) Scale and a 44% improvement (p = 0.10) on the UPDRS Motor Examination over the 27 patients receiving placebo.

Pramipexole's unique pharmacologic properties and its efficacy in patients with early PD who are not on levodopa mandated a larger study in patients with advanced PD experiencing motor fluctuations, the most common and difficult complication of levodopa treatment.

Methods. Design. This study was designed to compare the efficacy, safety, and tolerability of pramipexole with that of placebo in patients with advanced PD experiencing a suboptimal response to carbidopa/levodopa (including sustained-release preparations) characterized by motor fluctuations while receiving a stable dose of levodopa (end-of-dose or "wearing-off" phenomena). The study was conducted in two parts, a 32-week, double-blind, placebo-controlled, parallel-group study followed by an open-label extension trial. The results of the extension trial will not be presented here. Plasma pramipexole concentration-time data were used to assess mean pharmacokinetic values and their variance. These data will be reported elsewhere.

At visit 1 (screening visit), each prospective patient's consent was obtained in writing before any study-specific procedures were performed. At that time, every medication currently taken by the patient was recorded as"concomitant medication." Patients treated with anticholinergic agents, 1-deprenyl, and/or amantadine continued to take those agents. Prior to randomization, patients had to have been on a stable dose of carbidopa/levodopa for at least 30 days. Patient confidentiality was ensured by utilizing a patient identification code number, which corresponded to treatment data. At visit 2, patients who met the inclusion criteria were randomized in a 1:1 ratio to receive either pramipexole or matching placebo. The randomization schedule was computer-generated using a block size of four. All members of the monitoring team involved in the day-to-day supervision and/or monitoring of the study were blinded to the treatment assignments.

The 32-week, double-blind interval began with an ascending-dose phase, which lasted up to 7 weeks. Pramipexole or matching placebo was administered as an adjunct to levodopa in seven dosages upwardly titrated from 0.375 to 4.5 mg daily, each administered in divided doses as two tablets three times daily 2 hours after meals. The daily dosage was increased incrementally until in the judgment of the investigator the patient received the maximally tolerated dosage or the dosage was associated with stable improvement. The daily dosage of carbidopa/levodopa could be adjusted downward and subsequently increased if necessary, but not to a level exceeding the original daily dosage. The times of administration of carbidopa/levodopa and any ancillary medication remained constant throughout the trial. Following the ascending-dose phase, patients entered the maintenance-dose phase, which lasted up to 24 weeks. After completing the maintenance-dose phase (visit 18), patients entered a 1-week dose-reduction phase, during which study medication was withdrawn (visit 19). The number of tablets dispensed and returned at each visit and an estimate of the number of tablets the patient should have taken at each visit were recorded to estimate compliance.

Patient selection. The trial was conducted in 26 centers (22 in the United States and four in Canada), with each trial center expected to enroll about two patients per month and complete 12 to 20 patients by the conclusion of the tenth month. A completed patient was defined as one who had fulfilled at least one half of the visits during the maintenance-dose phase. Each treatment group was planned to enroll enough patients so that 150 subjects would complete the 32-week, double-blinded interval. These patients were required to be at least 30 years of age. Patients who withdrew from the trial for adverse events were considered as complete. If a patient withdrew, the reason for withdrawal was documented on the case report form. All withdrawn patients were asked to return to the clinic at the time that their visit-18 assessments would have been conducted. To be included in the study, patients of either sex must have been at least 30 years of age with advanced idiopathic PD in stage II to IV as measured by the Hoehn and Yahr Scale during an "on" period. Patients must have continued to experience motor fluctuations specifically characterized as an end-of-dose phenomenon or a"wearing-off" effect while receiving a stable dosage of carbidopa/levodopa for at least 30 days prior to entering the study. Patients must have been able to maintain an accurate daily record of times of "on" and "off" periods during waking hours. Family members, guardians, or nursing personnel were allowed to assist the patient.

Patients with atypical parkinsonian syndromes caused by drugs, encephalitis, progressive supranuclear palsy, or multiple system atrophy were excluded from the study. Also excluded were patients with cognitive impairment that could adversely affect the understanding of informed consent, compliance with medication, or maintenance of accurate diaries; patients with second- or third-degree atrioventricular block or sick sinus syndrome, resting heart rate below 50 beats per minute, congestive heart failure, myocardial infarction within 6 months of the study, clinically significant liver or renal disease, or active neoplastic disease; patients who within 180 days of the study had surgery that would negatively impact on their ability to participate or a history of stereotactic brain surgery; and patients with a supine systolic blood pressure less than 100 mm Hg or evidence of a 20-mm Hg decline in systolic blood pressure plus orthostatic symptoms 1 minute after standing compared with the previous supine systolic blood pressure. Females of childbearing potential were required to have a negative pregnancy test at baseline.

Efficacy evaluation. Efficacy was evaluated by two primary endpoints and 14 secondary endpoints.

The two primary endpoints, based on the UPDRS Parts II (parkinsonian symptoms in relation to activities of daily living) and III (parkinsonian motor signs), were the change from baseline to final maintenance visit of the average of the "on" and "off" ratings for UPDRS Part II and the change from baseline to final maintenance visit of UPDRS Part III. Consequently, each patient at each clinic visit was scored twice for Part II of the UPDRS but only during "on" periods for Part III. For scales such as the UPDRS, for which larger numbers mean more severe disease status, a negative change from baseline to final score represents improvement.

Secondary endpoints for the evaluation of efficacy were based on change from baseline to final maintenance visit for the following scales or measurements: UPDRS Part II for "off" periods; UPDRS Part II for "on" periods; average percentage of "off" time; average severity of "off" time; dosage of concomitant levodopa; Schwab-England Disability Scale to assess function for "off" periods; Schwab-England Disability Scale to assess function for "on" periods; Modified Hoehn and Yahr Scale for "off" periods; Modified Hoehn and Yahr Scale for "on" periods; UPDRS Parts I and IV; Parkinson Dyskinesia Scale to assess dyskinesia and dystonia during an "on" period; timed walking test to supplement the UPDRS motor portion; and individual components of UPDRS Parts II and III. Unlike the UPDRS, a positive change from baseline to final score represents improvement for scales such as the Modified Schwab-England Disability Scale. The area under the change-from-baseline curve over the maintenance period was defined as a secondary endpoint for the average "on" and "off" ratings for the UPDRS Parts II and III. Patient diaries were used to record the duration of "on-off" periods and the severity of "off" periods during waking hours. Two different endpoints were measured: change from baseline to final maintenance visit and area under the change-from-baseline curve over the maintenance period. The stage of Parkinson's disease was defined with the Hoehn and Yahr Scale for both "on" and "off" periods.

Safety evaluation. Adverse events were recorded by the investigator at each visit, including date of onset, date of termination, intensity, treatment required, investigator's assessment of the relationship to study drug, action taken with study drug, and outcome. Event intensity was recorded as mild, moderate, or severe.

An event was defined as serious if it was fatal, life-threatening, or permanently disabling; required prolonged hospitalization; was an overdose; or was serious in the investigator's opinion. The treatment required for each event was recorded as either none, drug change, other, or a combination of drug and other. The relationship of an event to the study drug was classified as either yes or no. The action taken with the study drug due to the event was classified as none, dosage reduction, dosage increase, drug discontinued, or drug discontinued and reintroduced. The outcome of an event was coded as recovered, not yet recovered, sequela, fatal, or unknown.

Any serious adverse event, whether or not considered related to the study drug, was reported immediately. All deaths, regardless of their relationship to the study drug, that occurred within 28 days of discontinuation were immediately reported as serious events.

Statistical analysis. The null hypothesis for all statistical tests was that of no difference between the two treatment groups, with the alternative hypothesis being that a difference exists. All statistical tests were two-sided. The protocol specified that the study would be considered successful if, using the LOCF analysis on the intent-to-treat data set, both primary endpoints were significant at the 0.05 level. The intent-to-treat data set was defined as all patients who were randomized to treatment who received at least one dose of medication and for whom at least one post-drug efficacy assessment was obtained. Since statistical significance required changes in both endpoints, no correction of the alpha level was done. The only covariate in the protocol was for the survival analysis of the timed walking test. For this analysis, the baseline was identified as a covariate. There were no windows around the visit dates that would cause data to be excluded from analysis, and no data were excluded because of visits deviating from the study schedule.

Sample size for this study was based on the change from baseline of the UPDRS Part III. Since only limited information was available, this study sample size was based on two-power tables. One table assumed a standard deviation of 5.0 for change from baseline to final visit, and the other table assumed a standard deviation of 10.0. The DATATOP trial suggested that this range was reasonable.⁴ These data, using an α significance level of 0.05 and a power of 90%, suggested that 150 patients per treatment group were adequate.

For this study, the standard deviation and the observed difference between the groups in change from baseline were close to those used in the sample size calculations. For the UPDRS Part III, a root mean square of 10.22 (a standard deviation measure) and a difference between the groups of 3.04 were obtained. These are compatible with the assumptions made in the sample size determination.

There were several subgroup analyses. One subgroup analysis involved pramipexole patients who were using anticholinergic medication or l-deprenyl(there were no differences). Another subgroup analysis included patients who were withdrawn and asked to return for visit-18 assessment. The purpose of this analysis was to determine how similar these patients' results were compared with those in the full intent-to-treat study (there were no differences). Still another subgroup analysis examined rater consistency. Because rating scales can be subjective, it was decided at the blinded report planning meeting to perform subgroup analyses of the two primary endpoints including only those patients in each analysis who were rated at baseline and final maintenance visit by the same rater (there were no differences).

Results. Study demographic data are summarized intables 1 and 2. Both groups were comparable. Of the 360 patients who were randomized into the study, 291 completed the protocol as planned, and 69 patients withdrew (table 2). Thirty (16.6%) of the withdrawing patients were from the pramipexole group, while 39 (21.8%) were from the placebo group. The placebo group had both more withdrawing patients and faster withdrawal than the pramipexole group. At baseline, 28% of pramipexole-treated patients were on antidepressants compared with 19% of placebo-treated patients. This difference was statistically significant.

Efficacy. The average of "off" and "on" data for the ADL UPDRS Part II is shown by visit in figure 1. The difference between pramipexole and placebo was statistically significant. Similarly, statistical significance can be shown if the data are plotted by visit for the ADL of UPDRS Part II separately for "off" and "on" periods. The data for the motor portion of UPDRS Part III "on" period are shown infigure 2. The difference between pramipexole and placebo was statistically significant.

Among the individual components of the UPDRS motor portion, the greatest percentage reduction for pramipexole versus placebo occurred in resting tremor (65% versus 46%, respectively); rigidity (98% versus 55%, respectively); finger tapping (89% versus 40%, respectively); hand movements (61% versus 21%, respectively); and rapid alternating hand movements (63% versus 41%, respectively). Because some of the individual components are composed of one score while others are composed of five independent scores averaged together, qualitative and not quantitative comparisons were made and statistical significance was not computed. The difference between pramipexole and placebo in the percent "off" time between the last visit and baseline as determined from patients' diaries is statistically significant (figure 3).

The severity of the "off" periods was rated by the patients on a numerical scale. Although this was not a primary measure of efficacy, it shows an overall statistically significant (see table 2) decrease in the severity of the "off" periods with pramipexole compared with placebo(figure 3). This represents one of the first attempts to measure both the duration and the severity of "off" periods. The levodopa dose was significantly reduced with pramipexole compared with placebo (seetable 3).

In addition to the primary efficacy measures, there was a statistically significant decrease with pramipexole compared with placebo in disability on the Schwab-England Disability Scale for the "off" periods and the "on" periods (see table 2). There was a statistically significant improvement in dyskinesias and "off" time with pramipexole compared with placebo as measured separately on the UPDRS Part IV (seetable 2). There were no changes in mentation as measured on the UPDRS Part I, dyskinesias as measured on the Parkinson Dyskinesia Scale, or timed walking (see table 2).

Safety. During the study, 171 (94.5%) pramipexole-treated patients and 158 (88.3%) placebo-treated patients experience drug-related adverse events based on the investigator's classifying the causal relationship to the study medication. The most commonly reported drug-related adverse events in the pramipexole-treated group, defined by an incidence of at least 10%, included dyskinesia (61.3%), asymptomatic orthostatic hypotension (48.1%), dizziness (36.5%), insomnia (22.7%), hallucinations(19.3%), nausea (17.7%), symptomatic orthostatic hypotension (16.0%), and confusion (11.0%). In the placebo-treated group, the most commonly reported drug-related adverse events, defined by an incidence of at least 10%, included asymptomatic orthostatic hypotension (48.0%), dyskinesia (40.8%), dizziness (31.8%), nausea (16.7%), insomnia (15.6%), tremor (12.3%), symptomatic orthostatic hypotension (11.2%), headache (10.6%), and pain(10.1%). A total of 54 patients (15%), 24 treated with pramipexole and 30 with placebo, had an adverse event that led to discontinuation of study medication.

There was no clinically significant difference between the treatment groups with respect to psychiatric disorders common in Parkinson's disease such as insomnia, confusion, agitation, paranoia, and depression; although more patients in the pramipexole group than in the placebo group were on antidepressants, they were not more depressed at the time of the study. However, 38 patients (21.0%) in the pramipexole group and 10 (5.6%) in the placebo group developed hallucinations (p ≤ 0.0001). These hallucinations tended to be visual, which is common for patients on dopamine agonists. There was no clinically significant difference between treatment groups for nausea and vomiting (common adverse events in patients on dopamine agonists), heart rate/rhythm disorders, or urinary frequency or incontinence.

Discussion. This study demonstrates that pramipexole, when administered concurrently with levodopa, improves patients with advanced PD and levodopa-induced motor fluctuations. Compared with placebo, pramipexole, administered at a maximal daily dosage of 4.5 mg, improved activities of daily living of patients with advanced PD, as determined by Part II of the UPDRS assessed in the "off" and "on" periods and by the average of the "off" and "on" periods. Compared with placebo, pramipexole patients were improved as determined by the UPDRS Motor Evaluation assessed in the "on" period. Pramipexole decreased the time patients spent in the "off" period and the severity of the "off" period; it decreased disability as determined by the Schwab-England Scale in the "off" and "on" periods; and it decreased PD severity as determined by the Hoehn and Yahr Scale in the "off" and "on" periods. Individual motor signs that improved most were resting tremor, rigidity, and limb bradykinesia. The addition of pramipexole resulted in a 27% reduction in the dosage of levodopa.

The magnitude of improvement in ADL (21%), motor evaluation (25%), and time in the "off" period (31%) compared favorably with that of other dopamine agonists such as bromocriptine, pergolide, and cabergoline in patients with advanced PD.^5-9 CNS adverse events were similar to those previously reported with bromocriptine, pergolide, and cabergoline, whereas gastrointestinal and cardiovascular adverse events were less frequent.

This study included more patients and centers and used more efficacy measures than previous studies that evaluated dopamine agonists.^6-8 In addition, this study, more than others, attempted to correct for differences among centers and observers.

Based on the results of this study and the previous study in de novo patients, pramipexole is an effective and safe treatment for all stages of PD from early to advanced, can be used alone or in combination with levodopa, and appears to be as effective as bromocriptine, pergolide, and cabergoline. The major difference between pramipexole and other dopamine agonists may reside in pramipexole's ability to bind with and activate D₃ receptors in the striatum and limbic region.^10-13 A mood disorder, distinct from depression and characterized by anxiety, decreased attention span, apathy, and/or anhedonia, affects many PD patients.¹⁴ Unlike PD patients with true clinical depression, patients with this mood disorder may not complain of guilt, remorse, or sadness, and they may not respond to antidepressants. In this study, many of the investigators commented on the decreased anxiety, enhanced attention span, and decreased apathy among patients treated with pramipexole. These observations may reflect investigators' optimism. However, these observations will have to be validated by further studies. The possibility that pramipexole may offer a window into the mood disorders of PD is provocative.

Appendix