Evaluation of Dyskinesias in a Pilot, Randomized, Placebo-Controlled Trial of Remacemide in Advanced Parkinson Disease

Context  Long-term levodopa therapy for Parkinson disease commonly results in motor complications including "on-off" fluctuations and dyskinesias, but it is still unclear how best to assess treatment effects on dyskinesias in clinical trials.

Objective  To compare several methods of rating levodopa-induced dyskinesias to evaluate the effect of remacemide hydrochloride treatment in patients with advanced Parkinson disease.

Design  Two-week multicenter randomized, double-blind, placebo-controlled, parallel-group study.

Setting  Five academic sites of the Parkinson Study Group.

Patients  Thirty-nine subjects at least 30 years old with idiopathic Parkinson disease and disabling dyskinesias.

Interventions  Randomly received daily doses of 150 mg, 300 mg, or 600 mg of remacemide hydrochloride or matching placebo for 2 weeks.

Main Outcome Measures  The dyskinesia rating scales used were the Modified Goetz Dyskinesia Rating scale (MGDRS), a newly created Lang-Fahn Activities of Daily Living Dyskinesia scale (LFADLDS), and diary dyskinesia ratings.

Results  Patient and investigator diaries showed excellent agreement in dyskinesia ratings. The MGDRS score correlated with clinic diary ratings of the percentage of "on" time with dyskinesias, and the LFADLDS score correlated with home and clinic diary assessments of percentage of on time with severe dyskinesias. The MGDRS score did not correlate highly with the LFADLDS score. This pilot study also validated previous results demonstrating the safety and tolerability of remacemide treatment for advanced Parkinson disease but did not result in any demonstrable improvement or worsening in dyskinesia measures.

Conclusions  Diaries may provide a valid means of evaluating dyskinesias in clinical trials for Parkinson disease, but there remain other aspects of dyskinesias, as assessed by the MGDRS and LFADLDS, that are not reflected in diary ratings.

WITHIN 5 years of diagnosis approximately half of the patients with Parkinson disease (PD) treated with levodopa develop motor complications, including "on-off" fluctuations and dyskinesias.¹ Dyskinesias often exhibit a dose-dependent relationship with dopaminergic therapy, diminishing with dosage decreases and worsening with increases.² The pathophysiological origins of these phenomena have not been definitively established; however, the mechanism underlying the development of dyskinesias may be distinct from that of levodopa's antiparkinsonian properties.^3-8

Systemic administration of N-methyl D-aspartate (NMDA) antagonists can ameliorate levodopa-induced dyskinesias without a reduction in antiparkinsonian response in both primates and rodents.^6,9,10 Oh and colleagues¹¹ have reported that intermittent stimulation of normally functioning dopaminergic receptors activates intraneuronal signaling pathways in striatal spiny neurons that alters the phosphorylation state of coexpressed NMDA glutamate receptors. This causes an increased sensitivity of these glutamatergic receptors to cortical input to the striatum and consequent motor fluctuations and dyskinesias.^11,12 Thus, NMDA antagonists may reduce or prevent the development of these motor complications.

Remacemide hydrochloride is a noncompetitive NMDA channel antagonist with antiparkinsonian efficacy in rodent and primate models of PD when used in combination with levodopa, but not when administered alone.¹³ It is safe and well tolerated in patients with PD, and preliminary studies indicate that it may improve motor fluctuations in levodopa-treated patients.^14,15 Whether it has a significant effect on dyskinesias has not yet been established, but it has been suggested that an NMDA antagonist effect accounts for the efficacy of amantadine hydrochloride treatment in improving levodopa-induced dyskinesias.¹⁶

We planned this preliminary multicenter randomized, double-blind, placebo-controlled, parallel-group study to evaluate the effects of 3 different dosages of remacemide in patients with PD who have disabling dyskinesias. Since no standard criterion for the evaluation of dyskinesia severity within a clinical trial exists and the problem is multifaceted, we used several methods of assessment to determine which might be best suited for future studies. These included the Modified Goetz Dyskinesia Rating scale (MGDRS),¹⁷ the newly developed Lang-Fahn Activities of Daily Living Dyskinesia scale (LFADLDS) (Table 1), and dyskinesia diary rating cards (Figure 1).

Thirty-nine patients with idiopathic PD were enrolled from 5 sites of the Parkinson Study Group (Figure 2). Subjects were aged 30 years or older and had disabling dyskinesias that were defined by their presence historically greater than 25% of the average waking day (ie, score ≥2 on the Unified Parkinson's Disease Rating Scale [UPDRS], part IV [complications of therapy], item 32 [What proportion of the waking day are dyskinesias present?]) and interference with activities of daily living to at least a moderate extent (ie, score ≥2 on the UPDRS, part IV, item 33 [How disabling are the dyskinesias?], and a score of 2 or more on at least 2 of the 5 items in the LFADLDS [Table 1]). Subjects were treated with stable optimized dosages of levodopa for at least 2 to 4 weeks prior to enrollment, and could be taking stable dosages of selegiline hydrochloride, dopamine agonists, or catechol O-methyltransferase inhibitors. Patients were excluded if they were taking medications that could potentially alter parkinsonism, affect dyskinesias, or interfere with the metabolism or mechanism of action of the study medication, or if there was evidence of significant depression, dementia, psychosis, or other serious medical conditions.

After informed consent was obtained, screening medical and laboratory evaluations were performed with the patient in the "on" state (ie, with good function). Clinical assessments included the modified Hoehn and Yahr scale, UPDRS parts I to IV (ie, mentation, activities of daily living [ADL], motor function, and complications of therapy, respectively), Clinical Global Impressions scale, Schwab and England ADL scale, Mini-Mental State Examination, and the Beck Depression Inventory. Measures of dyskinesia severity were also obtained (described in the "Dyskinesia Outcome Measures" subsection of the "Patients and Methods" section), and the home dyskinesia diaries were introduced. To ensure reliable and accurate completion of these diaries by each patient, extensive instructions were given, an instructional videotape classifying dyskinesia severity was shown, and a training and practice scoring session was done with the investigator. During this session, the patient had to be seen in both "full off" and "on with severe dyskinesias" states (adjusting the dosage and timing of antiparkinsonian medications, if necessary, to do so), and complete agreement between the patient and the investigator on the classification of all observed clinical states had to be achieved. Patients were then given home dyskinesia diaries to fill out on each of 3 consecutive days immediately prior to the baseline visit.

Within 14 days, qualified patients returned for baseline evaluation and the 3-day home diaries were reviewed. At this time, pretreatment clinical assessments were also performed during an observation period of 6 hours. Patients were instructed to take their usual doses of antiparkinsonian medications with a low-protein breakfast prior to arrival at the clinic and to report to the research center for the study visit at a fixed time (9 AM). All antiparkinsonian medications were given on a predefined schedule identical to the patient's schedule at home, and a standardized lunch was provided at a set hour.

At the end of the baseline visit, subjects were randomly assigned to receive remacemide hydrochloride, 150 mg/d, 300 mg/d, or 600 mg/d, or matching placebo, in 2 divided doses 12 hours (±2 hours) apart. The computer-generated randomization plan included stratification by center and blocking to ensure approximate balance among the treatment groups within each center. Sites telephoned the Parkinson Study Group Coordination Center, Rochester, NY, to enroll patients, and patient identification numbers were assigned through an interactive computer module. Patients were then instructed to fill out another set of 3 home diaries immediately prior to the next visit.

Administration of medication was initiated on the evening of the baseline visit and titrated up to the assigned dosage over 5 days. This intervention continued for a total of 2 weeks. If a patient developed intolerance to the medication, a change from the twice daily regimen to a 4 times daily regimen, without changing the total daily dosage, was recommended.

Patients were reevaluated after 2 weeks during another 6-hour observation period with identical clinical assessments. At this visit, home dyskinesia diaries were also collected, medication compliance was checked, occurrence of adverse events was reviewed, and safety evaluation measures were obtained.

The primary objective of the study was to evaluate the relationships among several measures of dyskinesia severity. Evaluation of dyskinesias was done using the MGDRS, the LFADLDS, and dyskinesia diary rating cards at each study visit.

The MGDRS is a revision of a scale introduced by Obeso,¹⁸ and adapted by the Core Assessment Program for Intracerebral Transplantations.¹⁹ An investigator describes the phenomenology, severity, and functional disability associated with a patient's dyskinesias and rates these dyskinesias on the basis of direct observation of the patient as he or she performs specific tasks.¹⁷ In this study, the MGDRS was scored every 30 minutes over the 6-hour observation period (for a total of 13 assessments) at each visit. The investigator watched the patient walk, drink from a cup, put on a coat, and button clothing. Specific definitions for the severity rating codes (range, 0-4 for each task) were provided for reproducibility of results. A higher score indicates more severe impairment. Scores were averaged over the 13 assessments completed during that visit.

The LFADLDS is modified from the ADL section of the UPDRS (part II) and requires the patient to respond to questions assessing how dyskinesias, at maximal severity, influence his or her ability to write, eat, dress, attend to hygiene, and walk. It was scored at every visit, retrospectively assessing how the patient's worst dyskinesias affected ADL over the previous few days. Specific definitions for the severity rating codes (range, 0-4 for each task) were provided for reproducibility of results. A higher score indicates more severe impairment (Table 1).

The dyskinesia diary rating card can be completed by either an investigator or the patient. It requires notation of whether the patient is on state without dyskinesias, on state with mild dyskinesias, on state with severe dyskinesias, partially "off state" (ie, with poor function), fully off state, or asleep (Figure 1). Diary entries are made every 30 minutes throughout the day, or over a specified period. Diary-derived outcome variables include the percentage of time during waking hours that the patient spent in the on state, on state with dyskinesias, and on state with severe dyskinesias. The dyskinesia diaries were completed by each patient at home on 3 consecutive days immediately prior to each study visit (home diaries). The same rating cards were also simultaneously but independently scored every 30 minutes by the investigator and the patient during the 6-hour observation periods at the baseline and week 2 visits (clinic diaries).

The degree of agreement between the investigator and the patient with regard to clinic diary ratings of dyskinesia measures (percentage of on-state time with dyskinesias or percentage of on-state time with severe dyskinesias) was assessed using intraclass correlation coefficients. These were estimated using a 2-way random effects analysis of variance model.²⁰ The degrees of association among the various dyskinesia measures were quantified using Spearman rank correlation coefficients. These analyses were performed separately using data from the baseline and week 2 visits.

Changes from baseline to week 2 in the clinical assessments and dyskinesia measures were compared among the treatment groups using an analysis of covariance model with treatment group as the independent variable of interest, investigator as a stratification factor, and the baseline value of the outcome variable as a covariate. Using this model, the adjusted mean responses in the active treatment groups were compared separately with that in the placebo group using 2-tailed t tests. The significance level used for each test was .017, after Bonferroni correction for multiple group comparisons. Confidence intervals for each of the treatment effects were also computed using these models, including a Bonferroni-adjusted confidence coefficient of 98.3%. All statistical analyses were performed according to the intention-to-treat principle.

Baseline characteristics of the 4 treatment groups are given in Table 2. Overall, patients' ages averaged (±SD) 64.3 ± 8.4 years and they had a diagnosis of PD for 13.3 ± 5.8 years (mean ± SD). Patients spent 75.4% ± 16.1% (mean ± SD) of the waking day in the on state on average, including 59.5% ± 19.9% (mean ± SD) of the waking day with dyskinesias and 28.5% ± 25.7% (mean ± SD) with severe dyskinesias, according to the home diary. Some characteristics were not evenly distributed among the groups as a natural consequence of randomization. In particular, subjects in the 600–mg/d treatment group tended to be younger and have less severe PD than patients in the other 2 groups.

Near-perfect agreement was noted between the independent patient and investigator assessments of the percentage of on-state time with dyskinesias (intraclass correlation coefficients = 0.89 and 0.91 at baseline and at week 2, respectively) and the percentage of on-state time with severe dyskinesias (intraclass correlation coefficients = 0.99 and 0.98 at baseline and at week 2, respectively). Table 3 describes the overall excellent agreement between patients and investigators on all of the diary classifications using data from all of the half-hour periods (average, 12.3 per subject) during the baseline visit.

Spearman correlation coefficients describing the associations among the dyskinesia measures at the baseline and week 2 visits are provided in Table 4. Moderate to high correlations were observed between the home diary ratings and the clinic diary ratings. For the percentage of on-state time with dyskinesias, the correlations with home diary ratings at baseline were 0.47 for patient clinic ratings and 0.44 for investigator clinic ratings. These were similar to the correlations observed at the week 2 visit of 0.52 and 0.50, respectively. For the percentage of on-state time with severe dyskinesias, the correlations were 0.65 for both patient and investigator clinic ratings at baseline, and 0.78 and 0.77, respectively, at week 2.

Moderate correlations were seen between the MGDRS score and the home and clinic diary ratings of the percentage of on-state time with dyskinesias at the baseline and week 2 visits (Table 4). The associations were notably better between the MGDRS score and the clinic diary ratings compared with the home diary ratings. At week 2, moderate correlations were also observed between the MGDRS score and the clinic diary ratings of the percentage of on-state time with severe dyskinesias (Table 4).

Moderate correlations were demonstrated between the LFADLDS score and the home and clinic diary ratings of the percentage of on-state time with severe dyskinesias at the baseline and week 2 visits (Table 3). Only a weak correlation was observed with the home diary ratings of the percentage of on-state time with dyskinesias. The MGDRS and LFADLDS scores did not correlate with each other at baseline and correlated weakly at the week 2 visit (Table 4).

All 39 patients tolerated their assigned treatments for 2 weeks without adjustments in total daily dosage. Adverse events were similar to those seen in previous studies of remacemide treatment in patients with PD, the most common being dizziness and gastrointestinal discomfort (Table 5). These occurred in a dose-dependent fashion, with 4 patients each complaining of dizziness and gastrointestinal discomfort at the remacemide hydrochloride dose of 600 mg/d. No serious adverse events, changes in vital signs, or significant abnormal laboratory test results were noted, although prolonged QT intervals were seen in 2 patients (remacemide hydrochloride, 150 mg/d [n = 1], remacemide hydrochloride, 300 mg/d [n = 1]). Medication compliance was greater than 90% in all 4 treatment arms.

There were no significant differences between the placebo-treated and remacemide-treated groups for any of the dyskinesia measures (Table 6). Statistically significant improvements in the UPDRS motor score (P = .01) and the UPDRS ADL score (P = .004) in the off state were seen in patients receiving 150 mg/d of remacemide hydrochloride after adjustment for multiple comparisons. Trends toward increased percentage of on-state time in the home diaries (P = .04) and improved UPDRS ADL score in the off state (P = .04) were seen in patients receiving 300 mg/d of remacemide hydrochloride. Lastly, a trend toward improved UPDRS ADL score in the on state (P = .04) was seen in patients receiving 600 mg/d of remacemide hydrochloride (Table 6).

While there are standardized methods to assess changes in a patient's motor function and overall impairment from PD, there are none that satisfactorily assess treatment effects on dyskinesias.²¹ The UPDRS, the most widely used scale in the evaluation of PD, includes extensive assessments of motor and functional capacity.²² For evaluation of dyskinesias, however, it requests only a rough estimation of their duration (in 25% increments), association with pain, and resulting effect on function (ie, nil, mild, moderate, or severe effect). The Abnormal Involuntary Movement Scale, originally designed for studies of tardive dyskinesias, does not specifically reflect the disabilities, phenomenology, or distribution of involvement associated with dyskinesias in patients with PD.^4,17,23 Another scale, introduced by Fahn²⁴ and used by Kurlan et al²⁵ rated motor function on a continuum extending from akinesia to hyperkinesia, reporting these as being on opposite ends of the spectrum, and implying that these 2 phenomena are mutually exclusive.²⁶

We found consistently excellent agreement between the patient and investigator assessments using the clinic dyskinesia diary rating cards. Moderate to high correlations were also seen between clinic diary and home diary assessments. These validated the use of patient diaries in the documentation of the patient's functional state and motor fluctuations throughout the day in the home.

The LFADLDS score showed moderate agreement with home and clinic diary assessments of the percentage of on time with severe dyskinesias at the baseline and week 2 visits. This corroborated the usefulness of the scale in evaluating a patient's worst dyskinesias as these effect ADL, a measure for which this scale was specifically designed.

The MGDRS score showed better agreement with the home and clinic diary evaluations of the percentage of on-state time with dyskinesias. It did not correlate with diary assessments of the percentage of on-state time with severe dyskinesias until the week 2 visit. In addition, not surprisingly, associations with clinic diary ratings were consistently better than with home diary ratings. This emphasizes that the MGDRS score more accurately reflects a patient's status during the observation period, and only less so how the patient fares at home. Furthermore, the MGDRS is intrinsically a labor-intensive and time-consuming procedure, and because in practice only a limited number of assessments can be performed over any given period, it may not capture a representative sample of a patient's daily dyskinesias. The observation that it was only at the week 2 visit that more significant correlations were seen between this scale and both the percentage of on-state time with dyskinesias and percentage of on-state time with severe dyskinesias also suggests that there may be a learning curve for the investigators in the use of the MGDRS over time.

The MGDRS score did not correlate with the LFADLDS score at baseline, and the correlation was weak at the week 2 visit. Again, the week 2 results suggest that, indeed, there may be a change over time in the investigators' use of the scale, albeit that only a weak correlation was seen. The poor correlation between the historical information from the LFADLDS and clinical observation data from the MGDRS suggests that the dyskinesias evaluated in the office may not always be a reliable representation of the most severe dyskinesias experienced at home. This finding further reiterates the usefulness of a historical disability scale such as the LFADLDS in evaluation of dyskinesias.

Overall, we found that with proper training, dyskinesia diaries designed to assess the frequency of dyskinesias are a reliable measure of a patient's day-to-day state and may be a more convenient, less labor-intensive method of evaluating dyskinesias than an investigator-based assessment. Although they may not document the entire spectrum of frequency, severity, phenomenology, anatomical distribution, and functional impairment associated with a patient's dyskinesias, aspects that may be more accurately and individually assessed by other scales such as the MGDRS or the LFADLDS, they still provide a practical and reliable measure with potential utility for evaluating dyskinesias in clinical trials for PD.

In this study, we confirmed that remacemide hydrochloride is safe and well tolerated over a 2-week period at dosages up to 600 mg/d in patients with PD and dyskinesias who are treated with levodopa. There was improvement in parkinsonian impairment on some secondary outcome variables (eg, motor UPDRS scores, the percentage of on-state time, and UPDRS ADL scores in the on and off states), although others (eg, Clinical Global Impressions scale scores and Schwab and England ADL scores) failed to show any significant differences. We observed no significant improvement or worsening in dyskinesias with remacemide treatment as evaluated by the various scales used. However, this was a small pilot trial, and the 98.3% confidence intervals for the estimated treatment effects were wide. Thus, larger trials are required to establish the effects of remacemide treatment on dyskinesias and parkinsonian impairment.

Accepted for publication July 23, 2001.

Steering Committee

Principal Investigator: Ira Shoulson, MD, Rochester, NY. Co-principal Investigators: Anthony Lang, MD, Toronto, Ontario, and Stanley Fahn, MD, New York, NY. Chief Biostatistician: Michael McDermott, PhD, Rochester. Director, Clinical Trials Coordination Center: Karl Kieburtz, MD, Rochester. Medical Director: Steven Schwid, MD, Rochester. Project Coordinator: Elise Kayson, MS, RNC, Rochester. Steering Committee Members: Thomas Chase, MD, Bethesda, Md; Timothy Greenamyre, MD, Atlanta, Ga; and John Penney, MD, Boston, Mass.

Investigators and Coordinators

Columbia-Presbyterian Medical Center, New York, NY: Steven Frucht, MD; Hal Winfield, RN. Rush-Presbyterian–St Luke's Medical Center, Chicago, Ill: Christopher Goetz, MD; Lucia Blasucci, RN. Oregon Health Sciences University, Portland: John Nutt, MD; Julie Carter, RN, ANP; Sam Morehouse. University of Saskatchewan, Saskatoon: Ali Rajput, MD; Marianne Ewanishin, RN. Toronto Hospital, Western Division, Toronto: Lisa Johnston, RN, BSCN.

Parkinson Study Group Biostatistics and Coordination Centers

University of Rochester: Cindy Casaceli, MBA; Denni Day, RN, MSPH; Karen Hodgeman; Nirupama Laroia, MD; Joy Antonelle de Marcaida, MD*; Constance Orme; Arthur Watts; Rajeshwari Natarajan, PhD; Peng Huang, PhD.

Safety Monitoring Committee

University of Rochester: Pierre Tariot, MD, Chair; Christopher Cox, PhD, Carl Leventhal, MD.

*Dr de Marcaida is now with the Department of Neurology, Hartford Hospital, Hartford, Conn.

Corresponding author: J. Antonelle de Marcaida, MD, Department of Neurology, Hartford Hospital, 80 Seymour St, Hartford, CT 06102 (e-mail: jdemarc@harthosp.org).

This article was corrected 10/24/2001.