Parkinson Study Group . Pramipexole vs Levodopa as Initial Treatment for Parkinson DiseaseA Randomized Controlled Trial. JAMA. 2000;284(15):1931–1938. doi:10.1001/jama.284.15.1931
Context Pramipexole and levodopa both ameliorate the motor symptoms of early
Parkinson disease (PD), but no controlled studies have compared long-term
outcomes after initiating dopaminergic therapy with pramipexole vs levodopa.
Objective To compare the development of dopaminergic motor complications after
initial treatment of early PD with pramipexole vs levodopa.
Design Multicenter, parallel-group, double-blind, randomized controlled trial.
Setting Academic movement disorders clinics at 22 sites in the United States
Patients Three hundred one patients with early PD who required dopaminergic therapy
to treat emerging disability, enrolled between October 1996 and August 1997.
Interventions Subjects were randomly assigned to receive pramipexole, 0.5 mg 3 times
per day, with levodopa placebo (n = 151); or carbidopa/levodopa, 25/100 mg
3 times per day, with pramipexole placebo (n = 150). For patients with residual
disability, the dosage was escalated during the first 10 weeks. From week
11 to month 23.5, investigators were permitted to add open-label levodopa
to treat continuing or emerging disability.
Main Outcome Measures Time to the first occurrence of any of 3 dopaminergic complications:
wearing off, dyskinesias, or on-off motor fluctuations; changes in scores
on the Unified Parkinson's Disease Rating Scale (UPDRS), assessed at baseline
and follow-up evaluations; and, in a subgroup of 82 subjects evaluated at
baseline and 23.5 months, ratio of specific to nondisplaceable striatal iodine
123 2-β-carboxymethoxy-3-β-(4-iodophenyl)tropane (β-CIT) uptake
on single photon emission computed tomography imaging of the dopamine transporter.
Results Initial pramipexole treatment resulted in significantly less development
of wearing off, dyskinesias, or on-off motor fluctuations (28%) compared with
levodopa (51%) (hazard ratio, 0.45; 95% confidence interval [CI], 0.30-0.66; P<.001). The mean improvement in total UPDRS score from
baseline to 23.5 months was greater in the levodopa group than in the pramipexole
group (9.2 vs 4.5 points; P<.001). Somnolence
was more common in pramipexole-treated patients than in levodopa-treated patients
(32.4% vs 17.3%; P = .003), and the difference was
seen during the escalation phase of treatment. In the subgroup study, patients
treated initially with pramipexole (n = 39) showed a mean (SD) decline of
20.0% (14.2%) in striatal β-CIT uptake compared with a 24.8% (14.4%)
decline in subjects treated initially with levodopa (n = 39; P = .15).
Conclusions Fewer patients receiving initial treatment for PD with pramipexole developed
dopaminergic motor complications than with levodopa therapy. Despite supplementation
with open-label levodopa in both groups, the levodopa-treated group had a
greater improvement in total UPDRS compared with the pramipexole group.
Parkinson disease (PD) is a progressively disabling neurodegenerative
disorder treated most commonly by dopamine replacement with the precursor
levodopa, but agonists at dopamine-receptor sites have recently been proposed
as initial therapy in early stages of the disease.1
The rationale for initial dopamine agonist treatment derives from the belief
that early levodopa exposure adversely affects the course of disease and leads
to disabling dyskinesias and motor fluctuations (ie, dopaminergic complications).2 These dopaminergic complications are likely a consequence
of underlying nigrostriatal degeneration, elicited by exposure to dopaminergic
treatments, including levodopa and dopamine agonists.
Two reports have shown a reduced incidence of dyskinesias by initiating
treatment with dopamine agonists ropinirole or cabergoline compared with levodopa.3,4 The ropinirole trial found no statistical
differences in the occurrences of adverse events or in changes in the activities
of daily living (ADL) between the 2 groups, despite absolute differences in
favor of levodopa, leaving the relative benefits of these drugs in question.
Pramipexole is a nonergoline dopaminergic agonist5
that has been shown to be safe and effective compared with placebo in early
PD,6,7 but has not yet been compared
with levodopa. Accordingly, members of the Parkinson Study Group (PSG), an
independent academic consortium of investigators, conducted a multicenter
randomized clinical trial comparing initial treatment of early PD with pramipexole
vs levodopa. Given recent in vitro and animal data that have suggested that
pramipexole may be neuroprotective for dopamine neurons, we also explored,
in a subset of the subjects, the effects of these treatment strategies on
dopamine transporter density, a marker of the dopaminergic neuron terminal,
as measured by single photon emission computed tomography and iodine 123 [123I] 2-β-carboxymethoxy-3-β-(4-iodophenyl)tropane (β-CIT).8,9
A report detailing the methods of this trial has been published.10
This multicenter study was organized by the PSG in conjunction with
the sponsor, Pharmacia Corp (Peapack, NJ; formerly Pharmacia & Upjohn
Inc, Kalamazoo, Mich). Subjects were enrolled in the trial between October
1996 and August 1997 at 22 sites in the United States (17) and Canada (5).
The study was reviewed and approved by the institutional review board at each
of the participating sites, and all subjects gave written informed consent.
An independent safety monitoring committee was responsible for unblinded monitoring
of data concerning patient safety, with particular attention to patient death,
serious adverse events, and adverse events resulting in subject withdrawal
from the trial. There were no prespecified formal guidelines for recommending
modification or termination of the trial, and any decision regarding early
modification or termination would have been based on clinical judgment in
light of the results of significance tests.
Eligible subjects were adults aged 30 years or older who had idiopathic
PD for fewer than 7 years and who required dopaminergic antiparkinsonian therapy
at the time of enrollment. Patients who had taken levodopa or a dopaminergic
agonist in the 2 months prior to enrollment were excluded. Subjects were required
to be in Hoehn and Yahr stage I, II, or III, a scale that classifies PD into
5 clinical stages ranging from mild unilateral (stage I) to severe, bed-bound
illness (stage V).11 Subjects were excluded
if they had (1) history of a previous dopaminergic complication, (2) atypical
parkinsonian syndromes, (3) serious concurrent illness, (4) treatment with
methylphenidate, cinnarizine, reserpine, amphetamine, or monoamine oxidase
type A inhibitors in the past 3 months, (5) treatment with pramipexole in
the past 4 months, (6) treatment with neuroleptics, metoclopramide, alphamethyldopa,
or flunarizine in the past 6 months, or (7) an unstable dosage of selegiline,
amantadine, anticholinergic therapy, or other central nervous system active
therapies (eg, hypnotics, antidepressants, anxiolytics) in the past 2 months.
Eligible patients were randomized 1:1 to pramipexole or levodopa, in
combination with carbidopa, using a computer-generated randomization plan
that included stratification by investigator and blocking. A programmer at
the Pharmacia Corp generated a list of the subject identification numbers
and corresponding treatment assignments. The subject identification numbers
were sent to the PSG Biostatistics Center (Rochester, NY) and incorporated
in a computer interactive randomization module at the PSG Coordination Center
(Rochester, NY). Access to the randomization code was restricted to 2 programmers,
1 at the Pharmacia Corp and the other at the PSG Biostatistics Center. When
a patient was judged eligible and consented to be enrolled, a telephone call
was made to the Coordination Center, which provided a unique subject identification
number from the randomization module.
Pramipexole was taken as 0.25-mg, 0.5-mg, or 1.0-mg tablets or matching
placebo tablets, 3 times daily, which were identical in appearance, taste,
and smell. Carbidopa/levodopa was taken as 12.5/50-mg or 25/100-mg capsules
or matching placebo capsules 3 times a day. Treatment assignments included
active drug for one treatment and placebo for the other.
Subjects entered a 10-week dosage escalation period followed by a 21-month
maintenance period. All subjects were escalated initially to a daily dosage
of 1.5-mg pramipexole or 75/300-mg carbidopa/levodopa (level 1 dosage). Subjects
requiring additional therapy could escalate to 3.0-mg pramipexole or 112.5/450-mg
carbidopa/levodopa (level 2 dosage), or 4.5-mg pramipexole or 150/600-mg carbidopa/levodopa
(level 3 dosage). Therefore, all patients entered into the maintenance phase
(week 11) of the trial on level 1, 2, or 3 dosing. The pramipexole dosages
were determined from a previous dosage-ranging tolerability study in patients
with early PD.6 Levodopa and pramipexole dosages
were chosen as those commonly used in clinical practice and judged to be near
Throughout the maintenance period (week 11 through month 23.5), subjects
maintained on study dosage level achieved in the escalation phase. Subjects
with emerging disability were prescribed open-label carbidopa/levodopa as
needed.12 Sustained-release carbidopa/levodopa
preparations were not permitted.
Subjects were randomly assigned to the intervention groups at the baseline
visit and were evaluated at 4 and 10 weeks, and at 3, 6, 9, 12, 15, 18, 21,
and 23.5 months. The primary outcome variable was prespecified as the time
from randomization until the first occurrence of any of 3 specified dopaminergic
complications: wearing off, dyskinesias, or on-off fluctuations.
Dyskinesias were defined as an abnormal involuntary movement that includes
chorea, dystonia, myoclonus, or tics that could be either peak dose or end
of dose. Dyskinesias did not include early morning dystonia or other "off"
dystonias. Wearing-off was defined as a perception of loss of mobility or
dexterity, usually taking place gradually over minutes and usually bearing
close relationship to the timing of antiparkinsonian medications. On-off effects
were defined as an unpredictable and generally sudden (seconds to minutes)
shift between "on" (mobility) and "off" (immobility) not apparently related
to the timing of antiparkinsonian medications.10
One blinded investigator at each site made the judgment as to the occurrence
of a dopaminergic complication. Subjects reaching the primary end point continued
to be followed up throughout the 23.5 months of the trial.
Secondary outcome variables included changes in scores on the Unified
Parkinson's Disease Rating Scale (UPDRS),13
the Parkinson's Disease Quality of Life scale (PDQUALIF),14
the EuroQol,15 and the need for supplemental
carbidopa/levodopa. Measures of safety included the frequency and severity
of individual adverse experiences. The UPDRS is a standardized, reliable,
and valid instrument for assessing the severity of the clinical features of
PD.16 The PDQUALIF and EuroQol are disease-specific
and generic quality-of-life instruments, respectively. The PDQUALIF consists
of 32 items and is scored on a 100-point scale including 7 domains: social/role
function, self-image/sexuality, sleep, outlook, physical function, independence,
and urinary function.
A subset of subjects (n = 82) were enrolled from 17 of the 22 participating
study sites to undergo single photon emission computed tomography imaging
with [123I] β-CIT using methods reported previously.17 The imaging outcome measure was the ratio of specific
to nondisplaceable striatal [123I] β-CIT uptake. Subjects
were imaged before the baseline visit and just before the final 23.5-month
visit. All imaging evaluations took place at Yale University (New Haven, Conn).
The planned sample size of 300 subjects (150 per treatment group) was
chosen to provide 94% power to detect a 20% difference (70% vs 50%; hazard
ratio [HR], 0.57) and 77% power to detect a 15% difference (70% vs 55%; HR,
0.66) in the proportions of subjects reaching the primary end point between
the treatment groups. The assumptions underlying these calculations are detailed
The primary statistical analyses were performed by intention-to-treat.18 All statistical tests were 2-tailed and were performed
using a significance level of 5%. The analysis of the primary outcome variable
used the Cox proportional hazards regression model, with treatment group as
the factor of interest and stratified by the enrolling investigator. The HR
and 95% confidence interval (CI) comparing the 2 treatment groups were determined
from this model. The assumption of proportionality of hazards was examined
with the use of time-dependent covariates.19
Separate analyses of the time from baseline to the first occurrence of individual
dopaminergic complications and the need for supplemental levodopa were performed.
The cumulative probabilities of reaching the primary outcome and other end
points were estimated using Kaplan-Meier curves.
Mean changes in the total UPDRS score, as well as the mental, motor,
and ADL UPDRS scores, between randomization and 23.5 months were compared
among the treatment groups using analysis of covariance, with treatment group,
enrolling investigator, and the baseline UPDRS score included in the model.
A 95% CI was computed for the difference between the adjusted treatment group
means. Changes in UPDRS scores between baseline and the other visits were
analyzed similarly. These analyses also were used to examine change scores
in the quality-of-life measures. Interactions between treatment and enrolling
investigator were tested but not found. Two-tailed Fisher exact tests were
used to compare proportions of subjects experiencing adverse events between
the 2 groups. Changes in [123I] β-CIT uptake (striatum, caudate,
and putamen) were expressed as percentage changes from baseline, and means
were compared between the 2 groups using t tests.
For the analyses of continuous efficacy variables, if a subject was
missing a response at a particular visit, the last available observation for
that subject was carried forward and imputed for that visit. To determine
the impact of dropouts on the results, the analyses were repeated including
only subjects who had complete data for the response variable of interest.
The results of the latter analyses did not differ materially from the analyses
of the imputed data and hence are not reported here.
Of the 376 patients who were identified as potential participants, 52
were found to be ineligible and 23 declined for no specific reason (Figure 1). The remaining 301 patients were
randomized in the study, 82 of whom also enrolled in the [123I] β-CIT
substudy. No patients were lost to follow-up. Twenty-three subjects in the
pramipexole group (15.1%) withdrew prior to the planned 23.5 months of follow-up
compared with 19 subjects in the levodopa group (12.7%). In the pramipexole
group, 4 withdrew due to somnolence and nausea/vomiting, and 3 due to hallucinations
and edema. Six subjects in the levodopa group withdrew due to nausea. Two
deaths occurred in the levodopa-treated group and neither were judged to be
related to the study drug.
The 2 treatment groups were similar at baseline with regard to demographic
and clinical variables and the baseline characteristics for the 82 subjects
enrolled in the [123I] β-CIT substudy cohort were similar
to those for the entire study cohort (Table
The numbers of subjects at each dosage level were nearly identical in
the 2 treatment groups (Figure 1).
Subjects allocated to pramipexole took an average of 2.78 mg/d by the end
of the trial. Subjects allocated to levodopa took an average of 406 mg/d of
levodopa as experimental therapy. Fifty-three percent of subjects in the pramipexole
group required supplemental levodopa compared with 39% in the levodopa group
(HR, 1.54; 95% CI, 1.09-2.17; P = .02). At the end
of the trial, subjects in the pramipexole group who required supplemental
levodopa (n = 80) were taking a mean (SD) of 264 (245) mg/d of supplemental
levodopa compared with 252 (245) mg/d for subjects in the levodopa group requiring
supplementation (n = 58). Subjects in the levodopa group thus took an average
total daily dosage of 509 mg of levodopa (experimental plus supplemental).
Table 2 shows that 28% of
subjects assigned to pramipexole treatment reached the primary end point by
23.5 months compared with 51% in the levodopa group (HR, 0.45; 95% CI, 0.30-0.66; P<.001). The reduced risk was observed in each of the
four 6-month study periods (0-6 month HR, 0.46; 6-12 month HR, 0.27; 12-18
month HR, 0.56; 18-24 month HR, 0.65) and for specific dopaminergic complications
of wearing off and dyskinesias (Table 2; Figure 2).
Table 3 shows treatment
effects on dopaminergic end points vs timing of supplemental levodopa. The
absolute numbers of end points were larger in the levodopa group. Most of
the end points occurred after the use of supplemental levodopa, but in similar
proportions between the treatment groups. Of the 5 subjects taking pramipexole
who developed dyskinesias before the supplemental levodopa, 4 had no prior
The mean improvement in total, motor, and ADL UPDRS scores from baseline
to 23.5 months was greater in the levodopa group compared with the pramipexole
group (Table 4). The levodopa
group improved significantly from baseline to each follow-up visit relative
to the pramipexole group (P≤.002) in mean total,
motor, and ADL UPDRS scores (Figure 3).
Quality-of-life scores improved in both groups initially and then declined
over time (Figure 4). At 23.5 months
(102 weeks), the mean change scores were significantly different (P = .006) for the PDQUALIF with the scores higher (ie, better) for
those in the levodopa group. Mean change scores did not differ among the groups
at other time points. Analyses of the 7 PDQUALIF subscales revealed significant
differences at 23.5 months for 2 subscales in favor of the levodopa group:
sleep (P = .004) and self-image/sexuality (P = .02). Quality-of-life scores on the EuroQol scale showed
a similar divergence between the 2 groups at the 23.5-month visit (P = .06; Figure 4).
Significantly more patients in the pramipexole group experienced somnolence
(P = .003), hallucinations (P
= .03), and both generalized (P = .01) and peripheral
edema (P = .002) compared with those in the levodopa
group (Table 5). Of note, the
differences in somnolence and hallucinations between the 2 groups emerged
during the escalation phase of the trial, whereas the differences for edema
emerged during the maintenance phase of the trial.
Three subjects reported falling asleep while driving, 2 of whom had
been randomized to pramipexole and 1 to levodopa. None were taking open-label
levodopa. These events occurred while the subjects were receiving the level
2 dosing schedule at 2, 5, and 12 months after randomization. Two of these
events resulted in motor vehicle crashes, one in a subject randomized to levodopa
and the other in a subject randomized to pramipexole. Two additional subjects
complained of "abrupt" or "sudden onset" drowsiness unrelated to driving,
both were allocated to pramipexole and receiving level 3 dosing.
In the β-CIT substudy, 39 of the 40 subjects in the levodopa group
and 39 of the 42 subjects in the pramipexole group had a follow-up β-CIT
scan. The mean [123I] β-CIT uptake in the striatum, caudate,
and putamen at baseline was well below the uptake values reported for healthy
subjects (Table 1).17
The mean (SD) decline in β-CIT striatal uptake over the 23.5 months did
not iffer between the 2 treatment groups and was 20.0% (14.2%) in the pramipexole
group compared with 24.8% (14.4%) in the levodopa group (P = .15; Figure 5). Caudate
and putamen-specific β-CIT uptake during the 23.5-month observation period
also did not differ between the 2 treatment groups.
Our findings demonstrate that pramipexole, as initial therapy in patients
with early PD, reduced the risk of developing prespecified dopaminergic motor
complications by 55% compared with initiating therapy with levodopa over a
2-year period. The absolute risk reduction of 23% suggests that one would
need to treat 4 to 5 patients with pramipexole instead of levodopa over a
2-year period to prevent 1 additional dopaminergic complication from occurring.
Both pramipexole and levodopa improved parkinsonian features, as measured
by the UPDRS, but pramipexole was not as potent as levodopa in improving these
features. The UPDRS scores remained worse in the pramipexole group despite
the use of open-label levodopa for treating emerging or continuing disability.
Since the maximum benefit was seen during the 10-week escalation phase, research
subjects and investigators may have developed a clinical sense of a satisfactory
response and used this as a benchmark for measuring the adequacy of subsequent
treatment; that is, investigators added or adjusted supplemental levodopa
to maintain function rather than to improve it. The findings also suggest
that although UPDRS scores are not improved as much with pramipexole as with
levodopa, the subjects treated with pramipexole and the blinded investigators
judged their illness to be satisfactorily treated.
There are several potential explanations for why initial pramipexole
treatment reduced the risk of developing wearing off and dyskinesias compared
with initial levodopa treatment. First, the longer half-life of pramipexole
compared with levodopa (8 to 12 hours vs 1.5 to 2 hours) may reduce the pulsatile
stimulation of the striatal dopamine receptors thought to be important in
the development of dyskinesias and wearing off.20
Second, pramipexole and levodopa dosing may not have been equivalent and the
observed differences may in part be due to differences in dopaminergic potency
between the 2 groups. Finally, a neuroprotective effect of pramipexole, if
present, could also reduce the development of dopaminergic complications by
preventing the loss of dopamine neurons, although a significant effect in
[123I] β-CIT uptake was not seen in this study.
We did not detect significant differences in the quality-of-life scores
between the 2 treatment groups during the first 78 weeks of the trial, indicating
an initial equal satisfaction with both treatment options. We did detect a
significant group difference in the PDQUALIF total score, however, at the
end of the trial in favor of levodopa. It is difficult to judge the clinical
significance of this difference, which occurred primarily in the sleep subscale
of the instrument. A similar trend was seen in the visual analog component
of the EuroQol. Although generic and disease-specific quality-of-life scales
correlate with disease severity in patients with PD,21
the responsiveness of these quality-of-life scales to a clinically meaningful
change in function and quality of life remains unclear.
Pramipexole use was associated with a greater likelihood of somnolence,
hallucinations, and edema. The differences in somnolence and hallucinations
almost exclusively occurred when subjects started experimental therapy during
the escalation phase of the trial. Risk factors for these adverse effects
are not known, but they may be of particular concern for the elderly, those
with preexisting sleep disorders, or those receiving multiple concomitant
This study is the first, to our knowledge, to assess the rate of change
in striatal [123I] β-CIT uptake in a relatively large cohort
of patients with early PD over a 2-year period of observation. The rate of
decline in β-CIT uptake was similar to that found in smaller PD samples8 and was less in the pramipexole group than in the
levodopa group, but the group difference at 2 years was not significant. We
will continue to follow up this cohort to observe the course of neuroimaging
outcomes in the next 2 years.
Our findings extend the observations found in the randomized, double-blind
trials that compared initial treatment with ropinirole vs levodopa3 and cabergoline vs levodopa4
in early PD. The 268 patients in the ropinirole trial were slightly older
than our cohort (63 vs 61 years), but had similar UPDRS scores at baseline
and were followed up for 5 years. The data in our trial are similar to the
data found in the ropinirole trial in terms of dopaminergic and UPDRS outcomes.
However, in the ropinirole trial, only 2 comparisons were reported to be statistically
significant: occurrences of dyskinesia and mean changes in motor UPDRS scores.
The ropinirole trial also showed similar proportional increases in the occurrence
of somnolence (27.4% vs 19.1%) and hallucinations (17.3% vs 5.6%) in the ropinirole-treated
Our trial revealed significant group differences in the occurrences
of wearing off and dyskinesias in favor of pramipexole and in the occurrences
of somnolence and hallucinations in favor of levodopa. In addition, our trial
revealed significant group differences in the mean change in the ADL component
of the UPDRS. The smaller sample size (n = 268) and unbalanced allocation
ratio (2:1) in the ropinirole trial may have contributed to the differences
in statistically significant results seen between the 2 studies despite the
similarities in the magnitudes of the group differences.
These studies leave several questions unanswered. Does the trade off
between motor complications and efficacy as measured by the UPDRS favor levodopa
over agonists? What are the implications of the increased rates of somnolence
and hallucinations with agonist treatment? Further study should help address
these questions. Until longer-term data are available, the decision to initiate
treatment of early PD with pramipexole or levodopa should be made only after
considering the favorable dopaminergic motor complication profile associated
with pramipexole against the more potent antiparkinsonian effects associated