Effect of levodopa‐carbidopa intestinal gel on dyskinesia in advanced Parkinson's disease patients

ABSTRACT Objective The purpose of this study was to assess the effect of levodopa‐carbidopa intestinal gel (carbidopa‐levodopa enteral suspension) in advanced Parkinson's disease patients with troublesome dyskinesia. Methods Post hoc analyses of patient data from a 12‐week, randomized, double‐blind study and a 54‐week open‐label study were performed. Efficacy was assessed in the subgroup of patients defined by ≥1 hour of “on” time with troublesome dyskinesia at baseline as recorded in Parkinson's disease symptom diaries (double blind: n = 11 levodopa‐carbidopa intestinal gel, n = 12 oral levodopa‐carbidopa; open label: n = 144 levodopa‐carbidopa intestinal gel). The changes in “off” time, “on” time with and without troublesome dyskinesia, and the overall safety and tolerability of levodopa‐carbidopa intestinal gel were analyzed. Results Although not significantly different from oral levodopa treatment (P > .05) in the double‐blind study, levodopa‐carbidopa intestinal gel treatment resulted in a reduction from baseline in “on” time with troublesome dyskinesia (mean [standard deviation] hours: baseline = 3.1 [1.7], change from baseline to final = −1.8 [1.8], P = .014), increase in “on” time without troublesome dyskinesia (baseline = 7.4 [2.2], change = 4.4 [3.6], P = .004), and decrease in “off” time (baseline = 5.5 [1.3], change = −2.7 [2.8], P = .015). Similar trends were found in the open‐label study. An increase in levodopa‐carbidopa intestinal gel dose was not significantly correlated with increased “on” time with troublesome dyskinesia in either study (double blind: r = −.073, P = .842; open label: r = −0.001, P = .992). Adverse events were usually mild to moderate in severity and related to the gastrointestinal procedure. Conclusion Our exploratory analyses suggest that optimizing levodopa delivery with levodopa‐carbidopa intestinal gel may reduce troublesome dyskinesia in advanced Parkinson's disease. © 2016 International Parkinson and Movement Disorder Society

A BSTRACT: Objective: The purpose of this study was to assess the effect of levodopa-carbidopa intestinal gel (carbidopa-levodopa enteral suspension) in advanced Parkinson's disease patients with troublesome dyskinesia. Methods: Post hoc analyses of patient data from a 12-week, randomized, double-blind study and a 54week open-label study were performed. Efficacy was assessed in the subgroup of patients defined by 1 hour of "on" time with troublesome dyskinesia at baseline as recorded in Parkinson's disease symptom diaries (double blind: n 5 11 levodopa-carbidopa intestinal gel, n 5 12 oral levodopa-carbidopa; open label: n 5 144 levodopa-carbidopa intestinal gel). The changes in "off" time, "on" time with and without troublesome dyskinesia, and the overall safety and tolerability of levodopacarbidopa intestinal gel were analyzed. Results: Although not significantly different from oral levodopa treatment (P >.05) in the double-blind study, levodopa-carbidopa intestinal gel treatment resulted in a reduction from baseline in "on" time with troublesome dyskinesia (mean [standard deviation] hours: baseline 5 3.1 [1.7], change from baseline to final 5 21.8 [1.8], P 5.014), increase in "on" time without troublesome dyskinesia (baseline 5 7.4 [2.2], change 5 4.4 [3.6], P 5.004), and decrease in "off" time (baseline 5 5.5 [1.3], change 5 22.7 [2.8], P 5.015). Similar trends were found in the open-label study. An increase in levodopa-carbidopa intestinal gel dose was not significantly correlated with increased "on" time with troublesome dyskinesia in either study (double blind: r 5 2.073, P 5.842; open label: r 5 20.001, P 5.992). Adverse events were usually mild to moderate in severity and related to the gastrointestinal procedure. C on c lus i o n: Our exploratory analyses suggest that optimizing levodopa delivery with levodopa-carbidopa intestinal gel may reduce troublesome dyskinesia in advanced Parkinson's disease.  Funding agencies AbbVie, Inc. funded these studies. AbbVie, Inc. participated in the study design, research, analysis, data collection, interpretation of data, reviewing, and approval of the publication.
Relevant conflicts of interests/financial disclosures: A.A. was a study investigator and served as a consultant for AbbVie, Inc. V.S.C.F. was a study investigator and has served on advisory boards for Abbvie, Inc. J.T.B. was a study investigator, received research support, and served as a consultant and scientific advisor for Abbvie, Inc. J.T.S. was a study investigator, received research support, and served as an advisor for AbbVie, Inc. C.H., S.E., K.C., and J.A.B. are employees of AbbVie, Inc. and hold stock or stock options.

Introduction
Parkinson's disease (PD) is a common neurodegenerative disorder 1 characterized by a progressive depletion of dopaminergic neurons in the nigrostriatal system 2 and motor dysfunctions such as bradykinesia, rigidity, tremor, and postural instability. 1 Immediate release oral levodopa-carbidopa (LC-IR) is the leading treatment for PD. 3,4 However, over time, LC-IRtreated PD patients develop complications of motor fluctuations and dyskinesia that are related in part to pulsatile stimulation and erratic gastric emptying. [5][6][7] These complications are often difficult to treat and negatively impact quality of life. 8 In particular, dyskinesia has a negative effect on measures of cognition, activities of daily living, social stigma, and bodily discomfort. 8 Currently there are no medications approved to treat dyskinesia, although some benefit has been shown with amantadine and continuous subcutaneous apomorphine. 9,10 Deep brain stimulation of the subthalamic nucleus and globus pallidus internus is efficacious, but there are significant restrictions as a result of age and cognitive behavioral state, which limit its clinical application. 11 Levodopa-carbidopa intestinal gel (LCIG) is a treatment option for advanced PD patients. LCIG is administered to the upper intestine via a percutaneous endoscopic gastrojejunostomy (PEG-J) tube and portable infusion pump. Previously published data demonstrated that switching from oral levodopa to LCIG significantly reduced "off" time, 12 with the benefits sustained for at least 12 months. 13,14 In this study, we performed a post hoc analysis of the efficacy of LCIG in a subgroup of patients with 1 hour of "on" time with troublesome dyskinesia (TSD) at baseline who were enrolled in a double-blind study, and we sought to confirm the findings using post hoc analyses from a larger, open-label study.

Study Designs
Data from two previously described studies 12,13 were used in this analysis. The study protocols for each study were approved by every participating institution's internal review board or ethics committee. All patients provided written informed consent. LCIG (designated in the United States as carbidopa-levodopa enteral suspension) was continuously infused during waking hours (approximately 16 hours). The total daily dose of LCIG per day was composed of 3 individually adjusted doses (morning, continuous maintenance, and extra). All patients were optimized to LC-IR prior to baseline measurements.
The primary study (NCT00357994/NCT00660387) was a 12-week, randomized, double-blind, double-dummy, parallel-group, phase 3 study. 12 All 71 enrolled patients received a PEG-J and were randomized to either LCIG (n 5 37) or LC-IR (n 5 34). The LCIG cohort received placebo tablets identical in size and color to LC-IR, and the placebo cohort received placebo gel. Patients were allowed to continue stable regimens of anti-PD medications with the exception of apomorphine. PD symptom diaries (see Efficacy section) were assessed at baseline and weeks 2, 3, 4, 6, 8, 10, and 12.
The second study (NCT00335153) was a 54-week, open-label, multicenter, phase 3 safety study (N 5 354) with patients who had PEG-J placed after titrating to an optimal dose of LCIG via a temporary nasojejunal tube. 13 Patients were not allowed to continue other anti-PD medications after the initiation of LCIG infusion and received LCIG monotherapy for at least the first 28 days of treatment. PD symptom diaries were assessed at baseline and weeks 4, 12, 24, 36, and 54.

Patients
Adults (30 years) with advanced PD consistent with UK Brain Bank criteria were enrolled in each study. 12,13 Patients were eligible if they were levodopa responsive, experienced motor fluctuations despite optimized therapy, and had at least 3 hours of "off" time on baseline diary assessment. Additional qualifying criteria for patients were previously reported. 12,13 Efficacy All patients with a baseline efficacy measurement as well as at least 1 postbaseline efficacy measurement were included in the analysis of efficacy. Patients recorded their motor status in 30-minute intervals as either "off" time, "on" time without dyskinesia, "on" time with non-TSD, "on" time with TSD, or asleep at home using the PD symptom diary. 15 The efficacy endpoints in each study included the mean change from baseline in the number of hours of "off" time, "on" time without TSD (dyskinesia and non-TSD), and "on" time with TSD as recorded in the PD symptom diary and the mean change from baseline to the final visit on the scores from the UPDRS dyskinesia questions (UPDRS part IV, questions 32-34 regarding dyskinesia).

Safety
All patients who received LCIG treatment were included in the analyses of safety. Evaluation of the safety and tolerability of LCIG included adverse event (AE) monitoring. AEs were coded using the Medical Dictionary for Regulatory Activities (MedDRA) version 14.0, and tabulated by MedDRA Preferred Term. Each event could be coded to more than one relevant preferred term. All AEs presented are treatmentemergent AEs, defined as those with an onset on or after the initial PEG-J (double-blind) or nasojejunal (open-label) tube placement and no more than 30 days after PEG-J removal. Study investigators rated each event as mild, moderate, or severe.

Troublesome Dyskinesia Subgroup and Statistical Analysis
Unless noted otherwise, all analyses reported were performed with a subgroup of patients with TSD, defined post hoc as those with an average of at least 1 hour of TSD per day during the 3 consecutive days prior to baseline (after optimization to LC-IR but prior to start of study drug) as recorded in the PD symptom diary. Data from the TSD subgroup were analyzed separately from each study, and no corrections were made for multiple testing in these post hoc analyses.
Values from the PD symptom diaries were normalized to a 16-hour waking day and averaged for the 3 consecutive days prior to each visit. "On" time without TSD was calculated as the sum of "on" time without dyskinesia and "on" time with non-TSD. However, "on" time without dyskinesia and "on" time with non-TSD were both included in the assessment of the distribution of time within the 16-hour waking day.
Baseline was postoptimization with LC-IR but prior to the start of the study drug. The final value for efficacy measurements was defined as the last nonmissing value assigned to the treatment period for the patient. Within-group change from baseline to each visit and to endpoint was assessed with a 1-sample t test. A repeated-measures model with the terms of treatment, country, baseline, and visit and the interaction terms treatment by visit and baseline by visit were used to compare the treatment group diary measures in the double-blind study. The change from baseline mean daily levodopa dose during the study was assessed with a 1-sample t test. The relationship between the change in daily levodopa dose and the change in "on" time with TSD was examined by scatterplot and Pearson correlation coefficient. AEs and serious AEs were summarized for each study's TSD subgroup.
In addition to the TSD subgroup analysis in each study, a supplemental analysis of the PD symptom diary outcomes was performed on data from subsets of patients in the open-label study with even higher amounts of TSD at baseline (1.5, 2, 2.5, and 3 hours).

Double-Blind Study
There were 11 (30%) LCIG-treated patients (n 5 37) and 12 (35%) LC-IR-treated patients (n 5 34) who had at least 1 hour of TSD at baseline. There were no clinically meaningful differences between treatment groups for baseline characteristics (Table 1). In these LCIG-and LC-IR-treated patients, respectively, the mean (standard deviation [SD]) daily dose of levodopa was 927 (287) mg and 963 (497) mg at baseline, and during the study it was 1019 (310) mg and 1238 (737) mg. "Off" time was significantly reduced for these patients in each treatment group when compared with baseline (mean [SD] change from baseline to Although not significantly different from LC-IR ("on" time with TSD, P 5 .328; "on" time without TSD, P 5 .491), LCIG-treated patients had a significant mean decrease from baseline to final visit in hours of "on" time with TSD (Fig. 1a), which was accompanied by a significant improvement in "on" time without TSD (Fig. 1b). The LC-IR-treated patients also reported a significant improvement in "on" time without TSD (Fig. 1b). LCIG-treated patients showed a trend of improvement in "on" time without TSD and "on" time with TSD from baseline to every time point, starting as early as week 2; however, these improvements were not significant when compared with LC-IR-treated patients (Fig. 2a).
The composition of time spent in the different motor states during the course of the 16-hour daily treatment period was also examined, further breaking down "on" time without TSD into "on" time without dyskinesia and "on" time with non-TSD (Fig. 3). For LCIG-treated patients, the decrease in "off" time at the final visit translated to an increase in "on" time without dyskinesia and a decrease in "on" time with TSD. "On" time with non-TSD did not change. The decrease in "off" time for LC-IR-treated patients was evenly distributed between an increase in "on" time without dyskinesia and "on" time with non-TSD, whereas the overall "on" time with TSD did not change. The percentage of "on" time without dyskinesia more than doubled in LCIG-treated patients, whereas it only increased by approximately a third in the LC-IR patients. An increase in LCIG dose was not The overall incidence of AEs and most frequently reported AEs were similar across the treatment groups ( Table 2). AEs were usually mild to moderate in severity and related to the gastrointestinal procedure.

Open-Label Study
Of the 354 patients, 144 (41%) had at least 1 hour of TSD at baseline. Baseline characteristics were similar to both treatment groups in the double-blind study ( Table 1). The mean (SD) daily dose of levodopa in these patients was 1090 (624) mg at baseline and 1489 (500) mg during the study. "Off" time for these patients decreased significantly (mean [SD] change from baseline to final 5 23.48 [2.68], P < .001).
Similar to the results for the LCIG-treated patients in the double-blind study, "on" time with TSD significantly improved from baseline to final visit (Fig. 1a), which was accompanied by a significant improvement in "on" time without TSD (Fig. 1b). In a supplemental analysis, an even greater improvement in "on" time with TSD was observed in patients with higher levels of TSD at baseline; patients with 3 hours of TSD at baseline (n 5 71) reported a mean (SD) decrease of 3.02 (2.98) hours of "on" time with TSD (Supplemental Table 1). An increase in LCIG dose was not significantly correlated with an increase in "on" time with TSD (r 5 2.001, P 5 .992; Supplemental Fig. 1). Patients had a significant improvement in "on" time without TSD and "on" time with TSD from baseline to week 4, which was sustained throughout the 54week study (Fig. 2b).
Similar to the results of the double-blind study, the decrease in "off" time translated to an increase in "on" time without dyskinesia and a decrease in "on" time with TSD at the final visit (Fig. 3). The percentage of "on" time without dyskinesia more than doubled, and "on" time with non-TSD increased slightly.
The overall incidence of AEs and most frequently reported AEs were similar to those in LCIG-treated patients in the double-blind study ( Table 2).

Discussion
Patients with at least 1 hour of "on" time with TSD at baseline were included in this analysis. This The studies differed in design with respect to dose initiation, concomitant antiparkinsonian medication usage, duration of treatment, and comparators, and neither study was powered to detect a change in measures of dyskinesia. The interpretation of the results is limited because they were exploratory and conducted post hoc. The strength of our findings is that, despite the differences in study designs, the results in the larger sample of patients in the open-label study confirmed the results in the double-blind study.
Our analysis of patients with dyskinesia showed that in the double-blind pivotal study, 12 patients randomized to LCIG showed significant improvements in the mean "on" time with TSD, mean "on" time without TSD, and mean "off" time; however, these improvements were not significant when compared with LC-IR. Notably, the decrease from baseline in time spent as "on" time with TSD converted mostly to increased "on" time without dyskinesia during the course of the study because there were minimal changes in "on" time with non-TSD. When further evaluating patients in the open-label study with more than 2 hours or 3 hours of baseline TSD, even greater reductions in TSD were observed.
Dyskinesia is difficult to treat in the advanced stage of PD because it becomes more prominent with longterm oral levodopa replacement therapy. 16 At this stage in disease, the "on" time of motor fluctuations and dyskinesias have previously been correlated with high levels of levodopa in the plasma or cerebrospinal fluid. [17][18][19][20] LCIG provides continuous exposure to levodopa during waking hours (16 hours) and has the potential to circumvent complications such as dyskinesia by attenuating peak levodopa levels. 21 Interestingly, the reduction in TSD was observed after the second week of treatment, and the mean daily levodopa dose during treatment in the double-blind study was approximately 90 mg higher than baseline and 400 mg higher in the longer, open-label study (patients in the open-label trial discontinued other antiparkinsonian therapies, hence the larger increase in levodopa dose). This suggests that an improvement in TSD may be related to the change in levodopa pharmacokinetics and continuous drug delivery. However, peripheral levodopa levels (including C max ) were not measured in these studies. Deep brain stimulation (DBS) of the subthalamic nucleus and globus pallidus internus is an alternative option for advanced PD patients, but patient selection is limitated. 11,22 In a randomized 6-month study, DBS reduced "on" time with TSD by 22.6 (95% confidence interval 5 23.3 to 22.0) hours per day,  23 Our findings suggest that patients with advanced PD and severe motor fluctuations that include TSD can derive benefits from LCIG of a similar magnitude to DBS, and both therapies should be considered. The data presented here indicate that LCIG may provide a similarly efficacious alternative, and for those patients older than 70 years of age, LCIG may be the only potential option. 22,24 The safety profile of LCIG treatment in the TSD subgroup was consistent with the safety profile in the total population, 12,13 with the majority of adverse events associated with the procedure being complication of device insertion, abdominal pain, and procedural pain. In both studies dyskinesia was reported as an AE by only a slightly higher percentage of patients in the TSD subgroup when compared with the entire treatment group: double-blind LCIG patients 5 18% versus 14% (TSD subgroup, entire treatment group), double-blind LC-IR patients 5 20% versus 12%, and open-label patients 5 11% versus 9%. Overall, the safety data suggest that there are no additional safety concerns for patients with higher baseline TSD that would preclude them from considering LCIG treatment.
Reductions in "off" time are often associated with improved "on" time without TSD. In patients with significant "on" time with TSD, this reduction is associated with improved "on" time without TSD. Therefore, LCIG can effectively reduce "off" time without worsening dyskinesia and decrease TSD without worsening "off" time. The mechanisms by which this effect is accomplished are primarily related to change in levodopa pharmacokinetics. Oral levodopa treatment, even in association with enzyme inhibitors, results in significant oscillations in plasma bioavailability. 25 This is also aggravated by erratic gastric emptying that makes the levodopa effect difficult to predict in advanced disease. 26 Our findings suggest that by modifying the pharmacokinetics of levodopa, even with a relative increase in its daily administered dose, changing the mode of its delivery to achieve more constant blood levels may enhance its efficacy, reduce motor response oscillations, and improve troublesome dyskinesia.