CADTH Reimbursement Review

Foslevodopa-Foscarbidopa (Vyalev)

Sponsor: AbbVie Corporation

Therapeutic area: Parkinson disease

This multi-part report includes:

Clinical Review

Pharmacoeconomic Review

Stakeholder Input

Clinical Review

Executive Summary

An overview of the submission details for the drug under review is provided in Table 1.

Table 1: Background Information of Application Submitted for Review

NOC = Notice of Compliance.

Introduction

Parkinson disease (PD) is a progressive, neurologic disease^2,3 characterized by the dysfunction and loss of dopaminergic cells in the substantia nigra and other brain regions.² Patients with PD experience motor symptoms such as bradykinesia, tremor, rigidity, and postural instability, as well as nonmotor symptoms including cognitive impairment, mood disorders, and sleep problems.^4,5 It is the most common movement disorder, with estimated age-standardized incidence rates ranging from 108 to 212 per 100,000 among patients aged 65 years and older in North America.⁶ Approximately 10% to 20% of patients with PD do not achieve satisfactory control of their disease despite optimized oral treatment, indicating that their disease has progressed to advanced PD.^7,8 There is no universal consensus on a definition of advanced PD; the Delphi-based consensus recommendation⁹ and a simplified approach known as the “5-2-1” criteria¹⁰ are commonly used in clinical practice to guide identification of patients with advanced PD.

Oral therapy, and advanced device-aided therapies, including deep brain stimulation (DBS) and levodopa-carbidopa intestinal gel (LCIG), are currently available for the treatment of motor fluctuations in patients with levodopa-responsive advanced PD.¹¹ Most patients rely on optimization of oral therapy, which typically involves changes in dosage, dose frequency, or combinations of oral medications with different mechanisms of action, to control motor symptoms. While such adjustments can reduce medication response fluctuations, they increase the burden and complexity of oral medication use, and can require patients to take medication every few hours. LCIG, a continuous infusion of levodopa (LD) and carbidopa (CD) suspension into the small intestine, and DBS, which involves implanting electrodes within targeted areas of the brain that control movement, are effective in reducing “off” time (i.e., a period of uncontrolled PD symptoms despite treatment) and dyskinesia (i.e., uncontrolled and involuntary movements due to excess treatment effect) during “on” time (i.e., a period of controlled PD symptoms with treatment);¹¹ however, both are invasive treatments that require a specialized medical team to perform the procedure, monitor, and manage associated adverse events (AEs) and complications (e.g., device failure, infection risk).¹² DBS treatment is also only provided at specialized centres and is only appropriate in select patients without contraindications due to associated risks and potentially life-threatening AEs (e.g., intracerebral hemorrhage, stroke, infection, seizure induced by the surgery, and off-target stimulation effects such as changes in speech or freezing of gait).

The objective of this report is to review and critically appraise the evidence submitted by the sponsor on the beneficial and harmful effects of foslevodopa (240 mg/mL) and foscarbidopa (12 mg/mL) solution for subcutaneous infusion in the treatment of motor fluctuations in patients with advanced PD.

Stakeholder Perspectives

The information in this section is a summary of input provided by the patient and clinician groups who responded to CADTH’s call for input and from clinical expert(s) consulted by CADTH for the purpose of this review.

Patient Group Input

CADTH received 1 input from Parkinson Association of Alberta and 1 joint input from Parkinson Canada, Parkinson Society British Columbia, and Parkinson Quebec. Parkinson Association of Alberta conducted a survey of 26 patients with PD and care partners or family in Alberta. Parkinson Canada, Parkinson Society British Columbia, and Parkinson Quebec jointly gathered responses from 113 patients with PD and caregivers of patients with PD in Canada via a survey; the majority of respondents were from Ontario (72.6%).

According to both patient groups, “off” periods and motor fluctuations associated with PD substantially impacted quality of life and activities of daily living for patients, led to work absenteeism (and in some cases resulted in early retirement), and caused emotional and financial burden to the caregivers. Respondents from both patient groups noted that symptoms that are most important to control are changes in cognition and memory, fatigue and sleep issues, freezing and unpredictable “off” periods, changes in mood, rigidity, speech and swallowing issues, bladder and bowel issues, impaired balance, slowness, and tremors.

More than half of each patient group experienced side effects with taking oral medications, with fatigue, drowsiness, constipation, and bowel issues the most difficult to endure. More than half of the respondents to the joint input reported that high pill burden (up to 40 pills per day) impacted their lifestyle or quality of life. Difficulties related to medication adherence included difficulty with timing or remembering, swallowing, and storage of medications, and limited improvement of symptoms. Some patients also received some form of rehabilitation (physiotherapy, occupational/speech therapy, or exercise) as a treatment option, but respondents cited cost, lack of motivation, or lack of access as barriers, especially for patients in rural areas. No respondents from either patient group were receiving foslevodopa-foscarbidopa at the time of survey.

Respondents indicated that the most important unmet needs were treatment options that would not increase dyskinesia as time went on, medications that would treat cognitive issues, and longer-lasting medications that would reduce pill burden and OFF periods, eliminating the fluctuations and sleep interruptions caused by medications wearing off. The joint input indicated that a large proportion of patients were very reluctant about undergoing invasive treatment options such as DBS or LCIG, and the majority (65%) would be interested in an injection-based levodopa-carbidopa (LD-CD) treatment; however, only 1 (3.8%) respondent from Parkinson Alberta said they would consider it and 2 (7.7%) were unsure.

Clinician Input

Input From the Clinical Expert Consulted by CADTH

The clinical expert consulted by CADTH noted that there is an unmet need for treatment options that have less resource requirements such that treatment is accessible to patients, especially those residing in rural and remote areas, without the need to travel to major urban centres, as well as treatment options for patients who are ineligible for existing advanced therapies because of existing comorbidities. The clinical expert noted that foslevodopa-foscarbidopa could serve as a treatment option for patients with advanced PD and could fill a treatment gap for patients who cannot travel to access other advanced therapies or who have comorbidities or strong personal aversion to other options.

The clinical expert noted that patients with levodopa-responsive advanced PD would be considered eligible for foslevodopa-foscarbidopa treatment in clinical settings. The clinical expert noted that there is currently no universally agreed-upon definition for advanced PD, and that it would be appropriate to define advanced PD based on the Delphi-based consensus criteria or the “5-2-1” criteria or as “patients with PD who have motor fluctuations inadequately controlled by optimized oral therapy.” Patients with excessive “off” time or “on” time with bothersome dyskinesia are more likely to benefit from treatment, according to the clinical expert, while patients with levodopa-unresponsive symptoms are not expected to benefit from foslevodopa-foscarbidopa treatment as the system of delivery is dopamine precursor treatment. The clinical expert noted that a clinically meaningful response would include improvement in “on” and “off” time measurements and quality of life, which would typically be observed at 3 months after initiation. According to the clinical expert, clinically meaningfulness can be judged differently by treating neurologists and patients, for instance in the predictability of therapy or the flexibility patients have with longer continuous “on” periods; as such, a meaningful response may be best left to the discretion of the treating neurologist. The clinical expert also noted that treatment discontinuation could be considered when patients experience intolerable AEs or significant functional impairments that are not relieved by the treatment. The drug should be prescribed by neurologists who have experience in the treatment of patients with PD and are trained in the use of this drug, as per the clinical expert.

Clinician Group Input

Input was received from the National Movement Disorder Expert Group (11 clinicians), and the BC Movement Disorders Specialist Group (7 clinicians). Overall, the input from both clinician groups aligned with the input given by the clinical expert consulted by CADTH.

The clinician groups agreed about the unmet needs of patients with advanced PD. Patients receiving oral levodopa may have inadequate control of motor fluctuations despite increased dosing frequency over time, and may have contraindications, poor tolerance, or insufficient response to adjunctive medications. They described barriers to accessing advanced therapies for PD (i.e., DBS and LCIG treatments) due to limited numbers of specialists, uneven distribution of resources geographically, intense resource needs, medical contraindications, poor acceptance from patients because of the invasive nature and risks of the treatments, and the impact of PD itself on patients’ ability to travel long distances for DBS or LCIG treatment and to manage at-home aspects of LCIG treatment. The clinician groups also noted that no current treatments exist that address the underlying disease process of PD.

The clinician groups agreed that foslevodopa-foscarbidopa could serve as an additional treatment option for patients with advanced PD and could benefit patients experiencing bothersome end-of-dose “off” periods, unpredictable efficacy of oral therapies as a result of absorption delays, or complex oral medication schedules due to the subcutaneous delivery of foslevodopa-foscarbidopa.

The clinician groups indicated that, as with other existing advanced therapies, eligible patients would include those who have levodopa-responsive PD with bothersome motor and nonmotor fluctuations despite optimized oral therapies. They suggested that eligible patients have advanced PD, according to the “5-2-1” criteria, and that it would be reasonable to recommend first trying at least 1 monoamine oxidase type B (MAO-B) inhibitor and a catechol-O-methyltransferase (COMT) inhibitor, unless contraindicated. In cognitively intact patients aged 70 years or younger, the clinician groups also stated that it would be reasonable to recommend having tried at least 1 dopamine agonist and amantadine (if dyskinesia is bothersome), unless contraindicated. However, they suggested against requiring a previous trial of anticholinergics or apomorphine preparations for reimbursement of foslevodopa-foscarbidopa. The clinician groups agreed with the clinical expert that treatment response would be assessed based on “off” time, presence of disabling dyskinesia, and quality of life. They added that an association with easing of caregiver burden may be considered. The clinician groups agreed that discontinuation could be considered in patients with intolerable AEs (e.g., skin reactions or hallucinations) and those who are unable to use the pump correctly due to cognitive decline as a result of disease progression or lack of caregiver support. The clinician groups agreed with the clinical experts that movement disorder neurologists, general neurologists, and geriatricians with experience in treating PD could be comfortable and qualified to prescribe and maintain treatment with foslevodopa-foscarbidopa.

Drug Program Input

Input was obtained from the drug programs that participate in the CADTH reimbursement review process. The following were identified as key factors that could potentially impact the implementation of a CADTH recommendation for foslevodopa-foscarbidopa:

• relevant comparators

• consideration for initiation of therapy

• consideration for continuation or renewal of therapy

• consideration of discontinuation of therapy

• consideration for prescribing of therapy

• care provision issues

• system and economic issues.

The clinical experts consulted by CADTH provided advice on the potential implementation issues raised by the drug programs (refer to Table 5).

Clinical Evidence

Pivotal Studies and Randomized Controlled Trial Evidence

Description of Studies

One pivotal phase III, double-blind, double-dummy randomized controlled trial (RCT) (M15-736,¹³ N = 141) that assessed whether individualized foslevodopa-foscarbidopa continuous subcutaneous infusion (CSCI) increased change from baseline in average daily normalized “on” time without troublesome dyskinesia compared to oral LD-CD immediate-release (IR) tablet therapy, after 12 weeks in patients with PD who have motor fluctuations inadequately controlled by oral therapy, was included in the sponsor’s submission. Patients with prior DBS or LCIG treatment were excluded, and eligibility for DBS was not a consideration for enrolment. Study-defined key secondary end points included change from baseline in average daily normalized “off” time, Movement Disorder Society-Unified Parkinson’s Disease Rating Scale (MDS-UPDRS) Part II score, and presence of morning akinesia. Secondary end points included “on” time without dyskinesia and other measures of symptoms and health-related quality of life (HRQoL) (Parkinson’s Disease Questionnaire-39 items [PDQ-39], 5-Level EQ-5D [EQ-5D-5L], median and interquartile range [IQR] of bradykinesia and dyskinesia scores assessed using a Parkinson KinetiGraph/Personal KinetiGraph [PKG] device, and Parkinson’s Disease Sleep Scale-2 [PDSS-2]).

At baseline, the mean age of patients was 66.4 years (standard deviation [SD] = 9.5 years) and the majority were male and white. The mean time since PD diagnosis was 8.6 years (SD = 4.9 years). Mean time spent in “off” and “on” without troublesome dyskinesia motor states were 6.13 hours (SD = 2.097 hours) and 9.34 hours (SD = 2.514 hours), respectively.

Efficacy Results

The efficacy end points that were noted to be important to patients and clinicians based on stakeholder input are summarized in Table 2.

“On” Time Without Troublesome Dyskinesia

The least squares mean (LSM) difference between the foslevodopa-foscarbidopa arm and the oral LD-CD arm with respect to change from baseline to week 12 in average daily normalized “on” time without troublesome dyskinesia (primary end point) was 1.75 (95% confidence interval [CI], 0.46 to 3.05; P = 0.0083) hours, in favour of foslevodopa-foscarbidopa. Results of the sensitivity analyses assessing the impact of attrition and results of subgroup analyses of interest (age, duration of PD diagnosis, and levodopa dose intensity) were consistent with the primary analysis.

“Off” Time

The LSM difference between the foslevodopa-foscarbidopa arm and the oral LD-CD arm with respect to change from baseline to week 12 in average daily normalized “off” time (study-defined key secondary end point) was –1.79 (95% CI, –3.03 to –0.54; P = 0.0054) hours, in favour of foslevodopa-foscarbidopa. Results of the sensitivity analyses assessing the impact of attrition and subgroup analyses of interest were consistent with the primary analysis.

PDQ-39 Items (PD-Specific HRQoL Instrument)

The LSM difference between the foslevodopa-foscarbidopa arm and the oral LD-CD arm with respect to change from baseline to week 12 in PDQ-39 summary index (secondary end point) was –4.10 (95% CI, –8.14 to –0.05). The results for this outcome are at increased risk of type I error (false-positive results) because they were tested after failure of the statistical hierarchy.

MDS-UPDRS Part II Score (Motor Experiences of Daily Living)

The LSM difference between the foslevodopa-foscarbidopa arm and the oral LD-CD arm with respect to change from baseline to week 12 in MDS-UPDRS Part II score (study-defined key secondary outcome) was –1.58 (95% CI, –3.65 to 0.48; P = 0.13).

Parkinson’s Disease Sleep Scale-2

The LSM difference between the foslevodopa-foscarbidopa arm and the oral LD-CD arm in change from baseline to week 12 in PDSS-2 total score (secondary end point) was –5.40 (95% CI, –8.03 to –2.78). The results for this outcome are at increased risk of type I error (false-positive results) because they were tested after failure of the statistical hierarchy.

Table 2: Summary of Key Efficacy Results from the M15-736 Trial (FAS)

CD = carbidopa; CI = confidence interval; FAS = full analysis set; FOS-FOS = foslevodopa-foscarbidopa; HRQoL = health-related quality of life; LD = levodopa; LSM = least squares mean; MDS-UPDRS = Movement Disorder Society-Unified Parkinson’s Disease Rating Scale; PD = Parkinson disease; PDQ-39 = Parkinson’s Disease Questionnaire-39 items; PDSS-2 = Parkinson’s Disease Sleep Scale-2; SD = standard deviation.

Note: Outcomes summarized in this table includes the primary and secondary end points that were noted to be important to patients and clinicians based on input from patient groups, clinician groups, and the clinical expert consulted by CADTH.

^aThe analysis was conducted using a mixed model for repeated measures, adjusted for categorical fixed effects of treatment, country and visit, treatment-by visit and treatment-by-baseline interactions, and baseline measurement (continuous).

^bThis analysis was conducted using an analysis for covariate model, adjusted for categorical fixed effects of treatment and country, and baseline score (continuous).

^cAlthough the P value is ≤ 0.05, statistical significance cannot be claimed because the results for the second key secondary end point (MDS-UPDRS Part II), a prior end point in the testing hierarchy, were not statistically significant.

Source: M15-736 Clinical Study Report.¹⁴ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Harms

The key harms results from the M15-736 trial are summarized in Table 3.

Adverse Events

Treatment-emergent adverse events (TEAEs) were reported for 85.1% of patients in the foslevodopa-foscarbidopa arm and 62.7% of patients in the oral LD-CD arm. The most common TEAEs in the foslevodopa-foscarbidopa arm (at least 10%) were infusion-site erythema, pain, cellulitis, and edema, as well as dyskinesia, all of which were more commonly reported than in the oral LD-CD arm (infusion-site erythema and pain, 1.5% each). The frequency of falls was lower in the foslevodopa-foscarbidopa arm (8.1%) than in the oral LD-CD arm (17.9%).

Serious Adverse Events

Serious TEAE was reported for 6 (8.1%) patients in the foslevodopa-foscarbidopa arm and 4 (6.0%) patients in the oral LD-CD arm.

Withdrawals Due to AEs

Treatment discontinuation due to TEAEs was reported for 21.6% of patients in the foslevodopa-foscarbidopa arm and 1.5% of patients in the oral LD-CD arm. The most common TEAEs leading to treatment discontinuation in the foslevodopa-foscarbidopa arm were infusion-site cellulitis (5.4%), infusion-site pain (4.1%), infusion-site bruising, hemorrhage, and edema (2.7% each).

Mortality

No deaths were reported in the foslevodopa-foscarbidopa arm and 1 (1.5%) death was reported in the oral LD-CD arm.

Notable Harms

The frequencies of infusion-site reactions and infections were notably higher in the foslevodopa-foscarbidopa arm than in the oral LD-CD arm (infusion-site reactions: 62.2% versus 7.5%; infusion-site infections: 28.4% versus 3.0%).

The frequency of hallucination or psychosis was notably higher in the foslevodopa-foscarbidopa arm (14.9%) than in the oral LD-CD arm (3.0%). There were no reports of impulse-control disorder or impulsive behaviour in either treatment arm. There was no notable between-arm difference in the mean change from baseline in score for each impulse-control disorder and related behaviour parameters of the Questionnaire for Impulsive-Compulsive Disorders in Parkinson’s Disease – Rating Scale (QUIP-RS) across almost all time points. Based on the Columbia-Suicide Severity Rating Scale (C-SSRS) assessment, 5 (6.8%) patients in the foslevodopa-foscarbidopa arm and 2 (3.0%) patients in the oral LD-CD arm had suicidal behaviours or ideations. Depression was reported for no patients in the foslevodopa-foscarbidopa arm and 2 (3.0%) patients in the oral LD-CD arm.

Dizziness was reported for 3 patients in each treatment arm. Orthostatic hypotension by preferred term was reported for 1 (1.4%) patient in the foslevodopa-foscarbidopa arm and 2 (3.0%) patients in the oral LD-CD arm. Somnolence was reported for 1 (1.4%) patient in both treatment arms.

Table 3: Summary of Key Harms From the M15-736 Trial (SAS)

CD = carbidopa; LD = levodopa; SAS = safety analysis set; TEAE = treatment-emergent adverse event.

Note: Unless otherwise specified, safety data are reported using the Medical Dictionary for Regulatory Activities (MedDRA) preferred term.

^aAdverse event of special interest, including any adverse event in infusion site–related noninfection reactions company MedDRA query (CMQ).

^bAdverse event of special interest, including any adverse event in the hallucinations CMQ or psychosis and psychotic disorder standardized MedDRA query (SMQ).

^cAssessed using the Columbia-Suicide Severity Rating Scale.

Source: M15-736 Clinical Study Report.¹⁴ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Critical Appraisal

Results of an end-of-study survey aiming to assess the extent of unblinding suggested that the majority of patients were able to infer treatment assignment given the differences in treatment response. There is a risk of reporting bias in patient-reported outcomes (PROs), potentially in favour of foslevodopa-foscarbidopa, due to the subjective nature of these outcomes. However, the extent of bias is unclear. Further, while hierarchical testing procedure was in place to account for multiplicity, no definitive conclusion can be drawn with respect to end points other than the primary end point, and study-defined key secondary end points of “off” time, as well as MDS-UPDRS Part II score, due to a failure of statistical comparison in a prior end point in the testing hierarchy (i.e., MDS-UPDRS Part II score). No conclusion can be drawn on the prespecified subgroup analyses because of the lack of consideration for sample size and statistical power and control for multiplicity. As well, there was a risk of attrition bias in favour of foslevodopa-foscarbidopa due to higher attrition in the foslevodopa-foscarbidopa arm compared with the oral LD-CD arm; however, sensitivity analyses of the primary end point and the key secondary end point of “off” time, which assessed the impact of missing data, showed results consistent with the primary analysis, increasing certainty of the findings.

Patients with cognitive impairment and prior DBS or LCIG treatment were excluded from the study, which represents a gap in evidence; nonetheless, the clinical expert consulted by CADTH did not expect the exclusion of these patients to significantly impact the generalizability of the study population. With respect to outcomes, the clinical expert noted that the PD diary, MDS-UPDRS, and PDQ-39 are clinically relevant instruments that are used in clinical practice, while the relevance of bradykinesia and dyskinesia scores, EQ-5D-5L, and PDSS-2 are limited. Improved cognition was an unmet treatment need according to patients, and reduced caregiver burden is a treatment goal in advanced PD. No conclusion on these outcomes can be drawn from the study because the cognition was not assessed as a stand-alone end point (although it was captured as 1 of the items in MDS-UPDRS scale) and caregiver burden was not measured. The clinical expert noted that the duration of follow-up (12 weeks) was adequate for efficacy assessment, although longer follow-up is required to gain certainty on the maintenance of benefit and safety profile.

Long-Term Extension Studies

The M20-098 trial is an ongoing long-term open-label extension study of the pivotal RCT, M15-736, in which patients received individualized foslevodopa-foscarbidopa CSCI for 24 hours per day for up to 96 weeks. At the time of the submission, no patients had completed the trial, and data were available from fewer than 5 patients from week 24 and beyond for outcomes of interest. Data from M20-098 were therefore too immature to use to draw conclusions.

Indirect Comparisons

Description of Indirect Comparisons

One sponsor-conducted indirect treatment comparison (ITC), which indirectly compared foslevodopa-foscarbidopa with LCIG and best medical therapy (BMT) (oral therapy) with respect to change from baseline in mean “off” time, “on” time without troublesome dyskinesia, and PDSS-2 total score at week 12 in patients with advanced PD via a Bayesian network meta-analysis (NMA), was included in the sponsor’s submission.

Efficacy Results

In the Bayesian fixed-effect NMA, which was based on a total of 4 trials, foslevodopa-foscarbidopa, compared with BMT (oral therapy), was associated with |||||||| |||| |||||| |||| |||||||| at week 12 in average “on” time without troublesome dyskinesia (mean difference, ||||| ||| |||||||| |||||||| |||||||||| || ||||) hours, “off” time (mean difference, |||||| ||| |||| ||||| || |||||) hours, and PDSS-2 total score (mean difference, |||||| ||| |||| |||| || |||||). Compared with LCIG, foslevodopa-foscarbidopa was associated with |||| ||||||||||| |||| |||||| |||| |||||||| at week 12 in PDSS-2 total score (LCIG versus foslevodopa-foscarbidopa: mean difference, ||||| ||| |||| |||| || |||||) and || |||||||||| || “on” time without troublesome dyskinesia (mean difference, ||||| ||| |||| ||||| || ||||) hours, “off” time (mean difference, |||| ||| |||| ||||| || ||||) hours.

Critical Appraisal

The validity of the results of the NMA could not be determined because the key assumptions of the analysis, homogeneity and consistency, could not be determined based on insufficient reporting of study characteristics and a sparse linear network without a closed loop. Based on the available information, there was evidence of heterogeneity between the included studies based on study designs (i.e., blinding, dosing protocol for oral therapies, duration of follow-up), patient populations (i.e., presence of concurrent cognitive impairment and dyskinesia), and patient baseline characteristics (i.e., duration of PD diagnosis, “off” time) that were unaccounted for. These limitations result in uncertainty in the relative treatment effect estimates between foslevodopa-foscarbidopa versus BMT (oral therapy) and LCIG.

Studies Addressing Gaps in the Pivotal and RCT Evidence

M15-741 Trial

One supportive phase III, open-label, single-arm trial (M15-741, N = 244), with the aim of evaluating the safety and tolerability of foslevodopa-foscarbidopa in patients with advanced PD for 52 weeks, was included in the sponsor’s submission. Safety and efficacy were assessed as primary and secondary end points, respectively, most of which were consistent with the end points of M15-736. Baseline patient characteristics were in general similar to the M15-736 trial, although in the M15-741, mean time since PD diagnosis (12.3 years [SD = 5.3 years]) was longer, more patients were at advanced stages of PD (based on the Hoehn and Yahr Scale) and received, on average, more medications from different PD drug classes, suggestive of a patient population with more advanced disease than the patient population included in the M15-736 trial.

Efficacy Results

All efficacy results were not adjusted for multiplicity. Foslevodopa-foscarbidopa was associated with a statistically significant improvement from baseline in average daily normalized “on” time without troublesome dyskinesia ||||| ||| | ||||| ||||||, “off” time |||||| ||| | ||||| ||||||, “on” time without dyskinesia ||||| ||| | ||||| ||||||, PDQ-39 ||||| ||| | |||||||, MDS-UPDRS Part II score ||||| ||| | |||||| and IV score ||||| ||| | ||||||, PDSS-2 total score ||||| ||| | |||||||, and EQ-5D-5L summary index |||||| ||| | ||||||| at 52 weeks. Results did not suggest a difference in change from baseline in MDS-UPDRS Part I and III scores or median and IQR of bradykinesia scores and dyskinesia scores at week 52 with foslevodopa-foscarbidopa treatment.

Harms Results

TEAEs were reported in ||||| || ||||||||| ||| |||| |||||| ||||| ||| ||||| |||| |||| |||||||| |||| ||||||||| ||||||| ||| ||||||||||| ||||||| ||||| |||| |||||||| || ||||| || ||||||||| |||| ||| |||| |||||||||| |||||||| ||| ||||| |||||||| |||| |||||||||| ||||||| ||||||||| ||||||||||||||| ||| || |||| ||| |||||||| || ||||| || ||||||||| |||| |||||||| ||||||| || ||||||||||||| ||||||| |||||||| |||| |||||||| ||| |||||||||| ||||| |||||| ||| |||||||||| ||||||| ||||| |||||| ||||||||| |||| || ||||| ||| |||||||||| || || ||||||| || ||||| |||| || ||| ||| |||||||| The safety profile of foslevodopa-foscarbidopa in this trial was in general consistent with the M15-736 trial, with no new safety signal.

Critical Appraisal

The open-label study design could introduce reporting bias, potentially leading to inflated benefits of foslevodopa-foscarbidopa on PROs and less favourable harms results given the more subjective nature of these outcomes. The noncomparative design means that known and unknown confounding factors were not accounted for and no statistical adjustments were made in the analyses, making it impossible to be certain that the observed treatment benefits are attributable to foslevodopa-foscarbidopa alone. As well, a sizable proportion of patients ||||||| withdrew from study treatment, mostly due to AEs and consent withdrawal. As a result, attrition bias may explain the observed efficacy results as the patients remaining in the study were more likely to be those who experienced benefits and were better able to tolerate the treatment.

The inclusion and exclusion criteria overall aligned with the selection criteria for candidates for advanced therapies used in clinical practice. While patients included in this trial appeared to have more advanced PD than those in the pivotal M15-736 trial, the clinical expert consulted by CADTH noted that the patient population would fit within the spectrum of patients with advanced PD in Canada. The attrition rate was high |||||||, most commonly due to AEs. This could affect the generalizability of the results since patients remaining in the trial tend to be those who are better able to tolerate the AEs of foslevodopa-foscarbidopa. The difference in infusion systems used in the trial versus stated in the product monograph could introduce some uncertainty due to potential differences in treatment interruptions, adherence, and safety.

M15-737 Trial

Early results from the ongoing single-arm long-term open-label extension M15-737 trial were submitted by the sponsor and are summarized in this report. Patients who completed the M15-741 trial could enrol in the M15-737 trial. The objective of M15-737 is to assess the longer-term safety and tolerability of foslevodopa-foscarbidopa delivered by CSCI for 24 hours per day for up to an additional 96 weeks after the 52-week M15-741 trial. The primary outcomes are AEs and safety measures. Efficacy outcomes are also being collected as secondary end points. At the time of this submission, data were limited after 48 weeks, and no patients had completed the study.

Efficacy Results

|||||||| |||| ||||||| |||||||| |||| ||| |||||||| || |||||||||||||||||||||||| |||| |||||||||| |||| |||||||| || |||| || |||| ||||||| || ||||||| ||||| |||||||||| |||| |||| ||||||| ||||||||||| ||||||||||| |||| |||| |||| ||||||||||| ||||||||||| ||||| ||||| |||||||||| ||| |||| |||||| ||||| |||| ||||||| |||||||| ||||| ||| | ||||||||||||| ||||||||||| |||||||| || ||||| |||| |||||||| || |||| || || |||||||| || |||||| ||| ||||||||| |||||| ||||| || |||| || |||| |||||| ||||||| ||||||| |||||||| |||| || || || ||||||||| |||||| |||| ||| |||||| ||||| |||| ||| ||||| || |||| ||||||||||| |||| |||||| | | | || |||||||

Harms Results

In the M15-737 trial, AEs were reported || || |||||||| |||||||| ||| ||| |||||||| |||| |||||||||| ||| |||| || |||||||| || ||||||||| ||| |||| |||||||||| |||||||| ||| || |||| |||| |||||||| |||| ||||||||| |||||| ||| |||||||| |||| ||||||||||| |||| |||||||| |||||| ||| ||| |||| ||| || |||||||||||||||| ||| ||||| || ||||| |||| ||| |||||||||||||||||| ||| ||||||| || |||||| The investigator considered that none of the fatal events had a reasonable possibility of being related to the study drug. No new safety concerns were identified.

Critical Appraisal

The open-label study design could introduce reporting bias, potentially leading to inflated benefits of foslevodopa-foscarbidopa on PROs and less favourable harms results given the more subjective nature of these outcomes. The noncomparative design means that known and unknown confounding factors were not accounted for and no statistical adjustments were made in the analyses, making it impossible to be certain that the observed treatment benefits could be attributed to foslevodopa-foscarbidopa alone. Because patients could only enrol after completion of the parent study, there is a greater likelihood of selection bias given that patients who better tolerated the treatment or who perceived the treatment as benefiting them were more likely to enrol. Finally, the trial is ongoing, and data after week 48 are limited. At the available time points, sample sizes are small. No definitive conclusions could be drawn from the results of this study.

Conclusions

In the pivotal M15-736 trial, foslevodopa-foscarbidopa demonstrated a clinically meaningful improvement in “on” time without troublesome dyskinesia and “off” time compared with oral LD-CD therapy at 12 weeks in patients with advanced PD. Analyses of morning akinesia, HRQoL, bradykinesia, and sleep symptoms also favoured foslevodopa-foscarbidopa, although due to failure of a prior outcome in the statistical testing hierarchy, the results for these outcomes were considered supportive of benefit with foslevodopa-foscarbidopa treatment, but not conclusive. Results did not suggest a difference in motor experiences of daily living, though the MDS-UPDRS instrument may be limited in its utility in assessing certain aspects of motor functioning in patients receiving advanced therapies (e.g., consistent control of motor fluctuations and flexibility in performing daily activities). The pivotal study results were determined to be generalizable overall. The comparative effectiveness and safety of foslevodopa-foscarbidopa relative to comparators other than oral LD-CD could not be determined. There are no direct comparisons with LCIG, and the indirect comparison was inconclusive because of important limitations that prevented verifying whether the underlying assumptions of homogeneity and consistency were met. No direct or indirect comparisons between foslevodopa-foscarbidopa and DBS were submitted. Overall, the safety profile of foslevodopa-foscarbidopa was similar to oral LD-CD therapy, except that infusion-site reactions and infections were more frequent with foslevodopa-foscarbidopa; most reactions and infections were not serious, but some resulted in treatment discontinuation. No new serious safety concerns were identified in the longer-term safety studies.

Introduction

The objective of this report is to review and critically appraise the evidence submitted by the sponsor on the beneficial and harmful effects of foslevodopa (240 mg/mL) and foscarbidopa (12 mg/mL) solution for subcutaneous infusion in the treatment of motor fluctuations in patients with advanced PD.

Disease Background

Disease Overview

PD is an incurable, progressive, neurologic disease^2,3 characterized by the dysfunction and loss of dopaminergic cells in the substantia nigra and other brain regions.² Patients with PD experience motor symptoms such as bradykinesia, tremor, rigidity, and postural instability, as well as nonmotor symptoms including cognitive impairment, mood disorders, and sleep problems.^4,5 As PD progresses, patients experience declining mobility, increasingly debilitating symptoms, decreasing independence, and greater impact on their quality of life.^9,10,16 Advanced PD also impairs gastrointestinal absorption, decreasing the effectiveness of the standard-of-care medication for PD, oral levodopa, and commonly requiring an increase in the dosage and frequency of administration of this standard of care;^17,18 this can lead to complications such as severe dyskinesia and bradykinesia.¹¹

Approximately 10% to 20% of patients with PD do not achieve satisfactory control of their disease despite optimized oral treatment, indicating their disease has progressed to advanced PD.^7,8 While there is no universal consensus on a definition of advanced PD, a body of leading specialists from 10 European countries proposed the following 15 clinically important criteria based on a Delphi-panel approach:⁹

• Motor symptoms: moderate level of troublesome motor fluctuations; at least 2 hours of the waking day with “off” symptoms; at least 1 hour of the day with troublesome dyskinesia; moderate level of dyskinesia; troublesome dysphagia; daily oral levodopa doses “at least 5 times a day”

• Nonmotor symptoms: mild level of dementia; nontransitory troublesome hallucinations; moderate level of psychosis; nonmotor symptoms fluctuations; moderate level of nighttime sleep disturbances

• Functional impacts: repeated falls despite optimal treatment; needs help with activities of daily living at least some of the time; not able to perform complex tasks at least some of the time; moderate impaired mobility

The Delphi-based consensus recommendation was that patients exhibiting 1 of the 6 selected motor indicators or 1 of the 5 selected nonmotor symptoms be classified as having “suspected” advanced PD, depending on the severity of the symptoms.⁹ These criteria have been further simplified and operationalized as the “5-2-1” criteria, which propose that patients with 5 or more oral levodopa doses daily, 2 or more hours of “off” time (i.e., a period of uncontrolled PD symptoms despite treatment) daily, or 1 or more hours of troublesome dyskinesia (i.e., uncontrolled and involuntary movements due to excess treatment effect) daily should be classified as having suspected advanced PD.¹⁰

Incidence, Prevalence, and Mortality

PD is the most common movement disorder and the second most common neurodegenerative disorder globally.^2,19,20 According to a global burden of disease study, approximately 6.1 million individuals worldwide were estimated to have PD as of 2016, with approximately 103,903 in Canada (uncertainty interval, 78,532 to 126,685).²⁰ The prevalence of PD increases with age, and PD diagnosis is uncommon in people younger than 50 years.²⁰ PD is approximately 1.4 times more prevalent in men than in women.²⁰ In North America, age-standardized incidence rates were estimated to range from 108 to 212 per 100,000 among people aged 65 years and older, and from 47 to 77 per 100,000 among people aged 45 years or older.⁶

In 2020, 3,431 Canadians died from PD,²¹ and age-standardized mortality rate was 9 per 100,000 people.²² The mortality rate among people with PD was 1.2 to 2.4 times higher than in the general population.²³ The vast majority of disabilities and morbidities in PD are caused by recurrent falls and postural instability, leading to head trauma and hip fractures.^24-26 It is also common for patients with PD to have multiple comorbidities, which become significantly more likely to occur as disease progresses, including hypertension, constipation, heart disease, depression, anxiety, and dementia.^27,28

Diagnosis

There is no diagnostic test for PD or advanced PD. Diagnosis is made by clinicians with expertise in movement disorders, based on examination of clinical neurologic features.^11,29

Standards of Therapy

Oral therapy and advanced device-aided therapies, including DBS and LCIG, are currently available for the treatment of motor fluctuations in patients with levodopa-responsive advanced PD.¹¹

Levodopa, the cornerstone oral treatment for PD, is often prescribed with adjunctive dopamine agonist, a COMT inhibitor, MAO-B inhibitors, amantadine, or anticholinergics.^11,30 While oral treatments are effective in treating symptoms in early stages of PD, response to oral medication tends to become less stable and predictable as the disease progresses, requiring fine adjustments of drug regimen (e.g., dosage, frequency, timing of administration, combinations of oral medications with different mechanisms of action) to minimize medication response fluctuations, incomplete benefit, and dyskinesia. Patients with advanced PD might consider advanced therapies, including DBS or LCIG, if response to oral medication continues to be limited or becomes less predictable or consistent over time.

DBS involves surgically placing stimulating electrodes in the areas of the brain that control movement; it is typically used as an adjunctive treatment to oral therapy.¹² LCIG is a gel suspension of levodopa and carbidopa that is continuously infused into the small intestine through a percutaneous endoscopic gastrostomy-jejunostomy (PEG-J) tube connected to a portable pump. Most patients eliminate most or all of other PD medications while on LCIG, according to the clinical expert consulted by CADTH. While both treatments are effective in reducing “off” time and reducing dyskinesia during “on” time (i.e., a period of controlled PD symptoms with treatment),¹¹ each treatment has specific issues in terms of health care resource use and safety that need to be taken into account. Generally, DBS is only done at specialized centres in Canada because the procedure requires an experienced and well-trained multidisciplinary team to conduct presurgical assessments, the surgery, and ongoing device monitoring and programming. Indefinite follow-up to monitor for device failures is also required.³¹ DBS is associated with risks and potentially life-threatening AEs (e.g., intracerebral hemorrhage, stroke, infection, seizure induced by the surgery, and off-target stimulation effects such as changes in speech or freezing of gait). DBS is only appropriate in some patients, typically those aged 70 years or younger, with severe motor dysfunction and no history of psychiatric disorders, cognitive issues (such as depression), speech disorders, or suicide attempts.^11,12,32 The provision of LCIG requires a specialized medical team, including the treating neurologist, a gastroenterologist trained in inserting PEG-J tubes, and nursing care to help in monitoring, device programming, and assessing for device- and tube-related problems (e.g., infection, tube kinking, and dislocations).³³ LCIG is associated with AEs of weight loss, abdominal pain, peripheral neuropathy, and diphasic dyskinesia.³³

Rescue treatments such as subcutaneous and sublingual apomorphine are available for the acute intermittent treatment of severe motor complications in advanced PD.^11,30 Rehabilitative therapies (e.g., physiotherapy, occupational therapy, exercises) are recommended in patients with motor function symptoms and difficulties with activities of daily living.¹¹

Some nonmotor symptoms may respond to levodopa treatment (orally or LCIG) and can benefit from continuous levodopa therapy, as per the clinical expert CADTH consulted. Other treatments for nonmotor symptoms are also available, such as cholinesterase inhibitors for dementia; midodrine, fludrocortisone, and domperidone for orthostatic hypotension; botulinum toxin A for sialorrhea; quetiapine, clozapine, pimavanserin for psychosis.¹¹

According to the clinical expert, treatment choice for motor symptoms is individualized and based on patient and health system factors. Treatment goals include reducing disability from motor and nonmotor symptoms, allowing independent functioning, preserving HRQoL, and reducing caregiver burden.

Drug Under Review

Key characteristics of foslevodopa-foscarbidopa, LCIG, and oral therapies are summarized in Table 4.

Foslevodopa-foscarbidopa is a prodrug combination of levodopa monophosphate and carbidopa monophosphate that is converted in vivo to levodopa and carbidopa. Levodopa relieves symptoms of PD following decarboxylation to dopamine in the brain. Carbidopa, which does not cross the blood-brain barrier, inhibits the decarboxylation of levodopa to dopamine outside the brain, allowing a larger amount of levodopa be available for transportation to the brain and transformation into dopamine.¹

This is the first review for foslevodopa-foscarbidopa by CADTH. Foslevodopa-foscarbidopa was granted a Health Canada Notice of Compliance for the treatment of motor fluctuations in patients with advanced levodopa-responsive PD who do not have satisfactory control of severe, debilitating motor fluctuations and hyper-/dyskinesia despite optimized treatment with available combinations of medicinal products for PD. The sponsor is seeking reimbursement of foslevodopa-foscarbidopa for the treatment of motor fluctuations in patients with PD who are inadequately controlled on optimized oral therapies (advanced PD) and who are not candidates for DBS. The sponsor noted that, among other considerations, patients who are not candidates for DBS could be characterized based on 1 or several of the following reasons:³⁴

• The patient does not consent due to surgical risk, surgical wait times, hesitancy of undergoing a neurosurgical procedure.

• The physician considers the patient is inappropriate for DBS due to ethical concerns

• The physician considers that the risks of the procedure outweigh the benefits for the patient.

• The patient is older than 70 years of age.

• The patient has moderate to severe depression, cognitive decline, other neuropsychiatric disorders, or medical comorbidities that increase surgical risk.

• The physician determines that the wait time for access to the DBS procedure is too long for the patient based on their status.

The sponsor noted that, ultimately, physicians will determine which patients are not appropriate candidates for DBS based on their clinical judgment and in consultation with the patient and their caregiver, as appropriate.³⁴

Foslevodopa-foscarbidopa is administered as a continuous subcutaneous infusion, 24 hours per day, using an infusion pump based on an individualized dosing. The starting infusion rate is determined by calculating the levodopa equivalents from all levodopa-containing medications and COMT inhibitors taken during daytime hours, and then increasing it to account for a 24-hour administration. The dose may be adjusted to reach a clinical response that maximizes the functional “on” time and minimizes the number and duration of “off” episodes and “on” episodes with troublesome dyskinesia. The maximum recommended daily dose of foslevodopa is 6,000 mg (or 25 mL of Vyalev per day, equivalent to approximately 4,260 mg levodopa per day). Foslevodopa-foscarbidopa replaces levodopa-containing medications and COMT inhibitors. If required, other classes of medicinal products for PD can be taken concurrently. Foslevodopa-foscarbidopa should only be prescribed by neurologists who are experienced in the treatment of patients with PD.

Table 4: Key Characteristics of Foslevodopa-Foscarbidopa, LCIG, and Oral Therapies

COMT = catechol-O-methyltransferase; LCIG = levodopa-carbidopa intestinal gel; MAO-B = monoamine oxidase type B; NMDA = N-methyl-D-aspartate; PD = Parkinson disease; PEG-J = percutaneous endoscopic gastrostomy-jejunostomy; SC = subcutaneous.

Note: Deep brain stimulation is not summarized in Table 4, but it is also available for the treatment of motor fluctuations in advanced PD in Canada.

^aOral therapy includes the following: levodopa-decarboxylase inhibitor and adjunctive COMT inhibitor, dopamine agonists, MAO-B inhibitor, NMDA receptor antagonists, and anticholinergics.

^bHealth Canada–approved indication.

^cAs per the Health Canada Product Monograph.

Sources: Foslevodopa-foscarbidopa product monograph,¹ Duodopa product monograph,³⁵ Compendium of Pharmaceuticals and Specialties – Parkinson Disease.³⁰

Stakeholder Perspectives

Patient Group Input

CADTH received 1 input from Parkinson Association of Alberta and 1 joint input from Parkinson Canada, Parkinson Society British Columbia, and Parkinson Quebec. Parkinson Association of Alberta conducted a survey of 26 patients with PD and care partners or family in Alberta. Parkinson Canada, Parkinson Society British Columbia, and Parkinson Quebec jointly gathered responses from 113 patients with PD and caregivers of patients with PD in Canada via a survey; the majority of respondents were from Ontario (72.6%).

According to both inputs, off periods and motor fluctuations associated with PD substantially impacted quality of life and activities of daily living, led to work absenteeism (and in some cases, early retirement), and caused emotional and financial burden to caregivers. Respondents from both input groups noted that the symptoms that are most important to control were changes in cognition and memory, fatigue, and sleep, freezing and unpredictable “off” periods, changes in mood, rigidity, speech and swallowing issues, bladder and bowel issues, impaired balance, slowness, and tremors.

More than half of each patient group experienced side effects with oral medications, with fatigue, drowsiness, constipation, and bowel issues the most difficult to endure. More than half of the respondents to the joint input reported that high pill burden (up to 40 pills per day) impacted their lifestyle or quality of life. Difficulties related to medication adherence included difficulty to do with timing or remembering, swallowing, and storage of medications, and limited improvement of symptoms. Some patients also received some form of rehabilitation (physiotherapy, occupational/speech therapy, or exercise) as a treatment option, but respondents also cited cost, lack of motivation, or lack of access as barriers, especially for patients in rural areas. No respondents from either patient group were receiving foslevodopa-foscarbidopa at the time of survey.

Respondents indicated the most important unmet needs were treatment options that would not increase dyskinesia as time went on, medications that would treat cognitive issues, and longer-lasting medications that would reduce pill burden and “off” periods, eliminating the fluctuations and sleep interruptions caused by medications wearing off. The joint input indicated that a large proportion of patients were reluctant about undergoing invasive treatment options such as DBS or LCIG and that the majority (65%) would be interested in an injection-based LD-CD treatment; however, only 1 (3.8%) respondent from Parkinson Alberta said they would consider it and 2 (7.7%) were unsure.

Clinician Input

Input From Clinical Expert Consulted by CADTH

Unmet Needs

The clinical expert noted that access to currently available advanced PD therapies can be challenging to patients due to health system and geographical factors. First, the clinical expert noted there is limited access to there are fewer than 80 neurologists specializing in movement disorders in Canada, with most located in urban centres, for more than 100,000 patients with PD.³⁶ Second, the provision of DBS and LCIG treatments requires a complex system of care; DBS requires access to neurosurgical subspecialists and nurses with specialized expertise in device programming and monitoring, a system of care that includes the administration of treatment programs, and a hospital system with operating room staff and resources. As for LCIG, the clinical expert noted that in addition to neurologists and nurses with specialized training, a gastroenterologist trained to insert the PEG-J tube and a system of care that involves support in monitoring of the PEG-J tube and management of the infusion device, pump assistance by the care partner, and in some cases, wound or stoma care for PEG-J insertion site. Because of these resource needs, the clinical expert reported that LCIG and DBS treatments are typically provided in major urban treatment centres, which is a barrier to treatment access considering the inherent difficulties patients with advanced PD have in travelling to these sites. The clinical expert noted that there is an unmet need for treatment options that require less in the way of resources or the need to travel to major urban centres, so that treatment is more easily available, especially for those residing in rural and remote areas.

The clinical expert also noted an unmet need for treatment options for patients who are ineligible for advanced therapies because of comorbidities. DBS is generally not considered for patients with significant nonmotor symptoms (which rarely improve with DBS treatment), severe cognitive impairment, medical contraindications (e.g., long-term anticoagulation), other intracranial lesion, or high surgical risk due to medical comorbidities. Certain medical (e.g., long-term anticoagulation) or gastrointestinal comorbidities preclude eligibility for LCIG.

Place in Therapy

The clinical expert noted that foslevodopa-foscarbidopa and LCIG share a similar mechanism of action (via dopamine precursors) but have different mechanisms of delivery. The clinical expert noted that subcutaneous drug delivery of foslevodopa-foscarbidopa is associated with a lower risk of surgical complications, which addresses a concern expressed by many patients who were reluctant to receive LCIG or DBS because of the risks associated with insertion of PEG-J tube (for LCIG administration) or neurosurgery (for DBS treatment). According to the clinical expert, foslevodopa-foscarbidopa is expected to fill a treatment gap for patients with advanced PD who cannot travel to access other advanced therapy options (e.g., those who reside in rural and remote areas) or who have comorbidities or strong personal aversion to other options (or, specifically, those options that require surgical procedures). The clinical expert noted that foslevodopa-foscarbidopa eliminates the need to coordinate care between movement disorders specialists and a second specialist (e.g., gastroenterologist or neurosurgeon) upon treatment initiation. With these simplifications to advanced therapy for patients with advanced PD, the clinical expert noted that foslevodopa-foscarbidopa could lead to a shift in the current treatment paradigm for advanced PD by improving access to device-aided therapy.

The clinical expert expected that foslevodopa-foscarbidopa would serve as a treatment option in patients with advanced PD. In general, concomitant oral medications are be expected to be stopped at the time of titration of foslevodopa-foscarbidopa, as per the clinical expert.

Patient Population

The clinical expert noted that patients with levodopa-responsive advanced PD would be considered eligible for treatment with foslevodopa-foscarbidopa in clinical settings. Although the lack of a universally agreed-upon definition for advanced PD was noted, “5-2-1” criteria or the Delphi-based consensus criteria are commonly used in clinical practice to assess patients with suspected advanced PD. The clinical expert added that it is accurate to define patients with advanced PD as those who have motor fluctuations that are inadequately controlled with optimized oral therapy; however, this definition hinges on clinicians’ ability to assess motor fluctuations and their familiarity with oral therapy optimization approaches, which would require limiting prescribing to experts in treatment of movement disorders to ensure the treatment is used in appropriate patients. With respect to the definition of “optimized oral therapy,” the clinical expert noted that an individualized optimization approach is typically used; therefore, the decision of what constitutes “optimized oral therapy” is – in their opinion – best left to the judgment of the treating neurologist.

The clinical expert noted that patients with excessive “off” time or “on” time with bothersome dyskinesia are more likely to benefit from treatment with foslevodopa-foscarbidopa, while patients with levodopa-unresponsive symptoms (e.g., axial symptoms such as freezing of gait, aspects of cognitive impairment, or symptoms in a limb that are also affected by neuromuscular weakness or stroke) are not expected to benefit from foslevodopa-foscarbidopa treatment. The clinical expert noted that patients with prior DBS, or LCIG treatment and patients with cognitive impairment could potentially benefit from foslevodopa-foscarbidopa.

Assessing the Response Treatment

The clinical expert noted that in clinical practice, response assessments typically involve subjective discussions of disease management; depending on the clinic model, the target symptoms and the need for advanced therapies, symptom diaries to assess “on” and “off” times, and quality-of-life scales like PDQ-39 may also be conducted intermittently. In the clinical expert’s opinion, a clinically meaningful response would include improvement in “on” and “off” time measurements and quality of life, although the definition of clinical meaningfulness could vary among clinicians. The clinical expert noted that treatment response is typically determined at 3 months after initiation.

Discontinuing Treatment

The clinical expert noted that treatment discontinuation could be considered when patients experience significant functional impairments that are not relieved by the treatment (e.g., becoming dependent for transfers and daily activities) or intolerable AEs (e.g., cutaneous infusion reactions). Functional impairment could occur due to progression of levodopa-unresponsive symptoms, nonmotor symptoms, or other causes (e.g., stroke or other serious medical conditions).

Prescribing Considerations

The clinical expert noted that while specifying prescribing by movement disorders specialists would be appropriate, it was their opinion that some latitude should be given to allow prescribing by neurologists (nonmovement disorders specialists) sufficiently experienced, qualified, and trained to administer and monitor foslevodopa-foscarbidopa treatment in rural or remote areas, where access to movement disorders specialists can be limited. The clinical expert also noted that some community neurologists could be practising in rural or smaller urban settings without a multidisciplinary subspecialty clinic; therefore, prescribing should not be limited to neurologists practising in major urban centres. Taking these into consideration, the clinical expert’s opinion was that it would be appropriate to specify prescribing of foslevodopa-foscarbidopa by neurologists who have experience in the treatment of patients with advanced PD and are trained in the use of this therapy.

Additional Considerations

The clinical expert expressed concerns with limiting eligibility for foslevodopa-foscarbidopa to patients who are not candidates for DBS since this criterion could be interpreted by the jurisdictions to require a neurosurgical consultation to determine DBS eligibility before reimbursement of foslevodopa-foscarbidopa. The clinical expert noted that this would require considering a more invasive treatment option (i.e., DBS) before a less invasive option when the comparative effectiveness of the options has not been determined yet. This criterion, according to the clinical expert, would also remove the advantage of foslevodopa-foscarbidopa for patients living in areas underserved with respect to DBS, particularly rural and remote areas. The clinical expert also noted that patients have long wait times for DBS surgical consultation (e.g., at least 4 years in the clinical expert’s jurisdiction), which is a barrier to timely treatment access in a progressive and debilitating disease. The clinical expert highlighted the fact that not all patients assessed for DBS treatment are eligible to receive it, which also results in significant delays in treatment.

Clinician Group Input

Input was received from the National Movement Disorder Expert Group (11 clinicians), and the BC Movement Disorders Specialist Group (7 clinicians). Overall, the inputs of both clinician groups aligned with that of the clinical expert consulted by CADTH.

The clinician groups concurred on the unmet needs of patients with advanced PD. Patients receiving oral levodopa have inadequate control of motor fluctuations despite increased dosing frequency over time, and may have contraindications, poor tolerance, or insufficient response to adjunctive medications. They described barriers to accessing advanced therapies for PD (i.e., DBS and LCIG treatments) that vary geographically because of limited specialists, uneven distribution of resources geographically, intense resource needs, medical contraindications, poor acceptance by patients because of the invasive nature and risks of the treatments, and the impact of PD itself on patients’ ability to travel long distances for DBS or LCIG treatment and to manage at-home aspects of LCIG treatment. In addition, the clinician groups noted that no current treatments address the underlying disease process of PD.

The clinician groups were aligned in considering that foslevodopa-foscarbidopa could serve as an additional treatment option for patients with advanced PD, and could benefit patients experiencing bothersome end-of-dose “off” periods, unpredictable efficacy of oral therapies because of absorption delays, and/or excessively complex oral medication schedules due to the subcutaneous delivery of foslevodopa-foscarbidopa.

The clinician groups indicated that, as with other existing advanced therapies, eligible patients would include those who have levodopa-responsive PD with bothersome motor and nonmotor fluctuations despite optimized oral therapies. The inputs suggested that eligible patients have advanced PD, identified by the “5-2-1” criteria. They also suggested that it would be reasonable to recommend first trying at least 1 MAO-B inhibitor and a COMT inhibitor, unless contraindicated. It would also be reasonable to recommend that cognitively intact patients aged 70 years or younger first try at least 1 dopamine agonist and amantadine (if dyskinesia is bothersome), unless contraindicated. However, the inputs suggested against requiring a previous trial of anticholinergics or apomorphine preparations for reimbursement of foslevodopa-foscarbidopa. The clinician groups inputs agreed with the clinical expert that treatment response would be assessed based on “off” time, presence of disabling dyskinesia, and quality of life, and added that an association with easing of caregiver burden may be considered. The clinician groups agreed that discontinuation could be considered in patients with intolerable AEs (e.g., skin reactions or hallucinations) and those who were unable to use the pump correctly due to cognitive decline as a result of disease progression or who lack caregiver support. The clinician groups agreed with the clinical expert that movement disorder neurologists, general neurologists, and geriatricians with experience in the treatment of PD could be comfortable and qualified to prescribe and maintain treatment with foslevodopa-foscarbidopa.

Drug Program Input

The drug programs provide input on each drug being reviewed through CADTH’s reimbursement review processes by identifying issues that may impact their ability to implement a recommendation. The implementation questions and corresponding responses from the clinical experts consulted by CADTH are summarized in Table 5.

Table 5: Summary of Drug Plan Input and Clinical Expert Response

CDEC = CADTH Canadian Drug Expert Committee; COMT = catechol-O-methyltransferase; DBS = deep brain stimulation; IR = immediate release; LCIG = levodopa-carbidopa intestinal gel; MAO-B = monoamine oxidase type B; MMSE = Mini-Mental State Examination; pCPA = pan-Canadian Pharmaceutical Alliance; PD = Parkinson disease; PEG-J = percutaneous endoscopic gastrostomy-jejunostomy; SC = subcutaneous; vs. = versus.

Clinical Evidence

The objective of CADTH’s Clinical Review Report is to review and critically appraise the clinical evidence submitted by the sponsor on the beneficial and harmful effects of foslevodopa (240 mg/mL) and foscarbidopa (12 mg/mL) solution for subcutaneous infusion in the treatment of motor fluctuations in patients with advanced PD. The focus is on comparing foslevodopa-foscarbidopa to relevant comparators, and identifying gaps in the current evidence.

A summary of the clinical evidence included by the sponsor in the review of foslevodopa-foscarbidopa is presented in 4 sections, and CADTH’s critical appraisal of the evidence is included after each section. The sections include:

• the systematic review, which includes a description of the pivotal studies and RCTs selected according to the sponsor’s systematic review protocol (refer to the Pivotal Studies and RCT Evidence section)

• sponsor-submitted long-term extension studies (refer to the Long-Term Extension Studies section)

• indirect evidence submitted by the sponsor (refer to the Indirect Evidence section)

• additional studies that, according to the sponsor, addressed important gaps in the pivotal and RCT evidence (refer to the Studies Addressing Gaps in the Pivotal and RCT Evidence section).

Included Studies

Clinical evidence from the following is included in the CADTH review and appraised in this document:

• 1 pivotal phase III double-blind RCT (M15-736)

• 1 long-term extension study of the pivotal study (M20-098)

• 1 ITC

• 2 additional studies, including a phase III, open-label, single-arm study (M15-741), and its long-term extension study (M15-737).

Pivotal Studies and RCT Evidence

Description of Studies

Two phase III trials (M15-736¹³ and M15-741³⁷) met the inclusion criteria for the systematic review conducted by the sponsor. The pivotal M15-736 trial is presented in this section and its characteristics are summarized in Table 6. Since Health Canada considers the single-arm M15-741 trial to be a supportive trial for the approval of foslevodopa-foscarbidopa,³⁸ M15-741 and its long-term extension trial (M15-737) are summarized in the Studies Addressing Gaps in Pivotal and RCT Evidence section. The M15-741 and M15-737 trials potentially address gaps in evidence by assessing the longer-term efficacy and safety of foslevodopa-foscarbidopa.

The M15-736 trial was a phase III, randomized, multicentre, double-blind, double-dummy, active-controlled trial with the aim of evaluating the safety and efficacy of foslevodopa-foscarbidopa to demonstrate its superiority over oral LD-CD IR tablets for the treatment of motor fluctuations in patients with advanced PD over 12 weeks of treatment (N = 141). The study was conducted in 76 sites in Australia and the US. Patients were enrolled between October 19, 2020, and September 3, 2021. The study is now complete. It consisted of the following study periods:

• Screening period (6 to 60 days): Study eligibility assessments.

• Oral LD-CD stabilization period (14 to 21 days): All patients received oral LD-CD IR tablets at an individualized dose, and adjustments of dose and schedule were performed over 7 to 14 days to achieve optimal control of motor symptoms. This was followed by a 7-day period where no further adjustment to oral LD-CD IR were allowed.

• Double-blind treatment period (12 weeks; day 1 to day 85): Patients were randomized in a 1:1 ratio using interactive response technology to receive either foslevodopa-foscarbidopa plus oral placebo capsules for LD-CD IR (referred to hereinafter as the “foslevodopa-foscarbidopa arm”) or placebo infusion for foslevodopa-foscarbidopa plus oral LD-CD IR tablets (referred to hereinafter as the “oral LD-CD arm”). Randomization was stratified by study site. Between day 1 and day 29 (CSCI optimization phase), the CSCI rate of the study drug solution was adjusted based on clinical response while the same regimen of oral study drug was maintained. After day 29 (maintenance phase), all study drug regimens remained unchanged until the end of the study.

Efficacy and safety results of the double-blind treatment of M15-736 are summarized in this section. For results of the long-term extension study (M20-098),³⁹ refer to the Long-Term Extension Studies section.

Table 6: Details of Pivotal Studies and RCT Evidence Identified by the Sponsor

AE = adverse event; CD = carbidopa; CSCI = continuous subcutaneous infusion; C-SSRS = Columbia-Suicide Severity Rating Scale; DBS = deep brain stimulation; ECG = electrocardiogram; EQ-5D-5L = 5-Level EQ-5D; HRQoL = health-related quality of life; IQR = interquartile range; IR = immediate release; LD = levodopa; MDS-UPDRS = Movement Disorder Society-Unified Parkinson’s Disease Rating Scale; PD = Parkinson disease; PDQ-39 = Parkinson’s Disease Questionnaire-39 items; PDSS-2 = Parkinson’s Disease Sleep Scale-2; QUIP-RS = Questionnaire for Impulsive-Compulsive Disorders in Parkinson’s Disease – Rating Scale; RCT = randomized controlled trial; VAS = visual analogue scale.

^aAs per investigator’s assessment.

^bAssessed using the PD diary.

^cDefined as a Mini-Mental State Examination score of at least 24.

^dAssessed using the Parkinson KinetiGraph/Personal KinetiGraph wearable device.

Source: M15-736 Clinical Study Report.¹⁴ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Populations

Patients eligible for inclusion were aged 30 years or older, with a diagnosis of levodopa-responsive idiopathic PD. They were taking a minimum of 400 mg/day levodopa equivalents, were judged by the investigator to have motor symptoms inadequately controlled by current therapy, with recognizable or identifiable “off” and “on” states (motor fluctuations), and with an average “off” time of at least 2.5 hours per day over 3 consecutive PD diary days for a minimum of 2 hours each day. Patients who had previously received DBS, LD-CD enteral suspension, or any other PD medication as a continuous daily infusion were excluded.

Interventions

Oral LD-CD Optimization Period (14 to 21 Days)

All patients received an individualized regimen of oral LD-CD IR tablets that was based on the levodopa dose equivalent calculated from all levodopa-containing medications and COMT inhibitors established before the study using conversion rates published by Tomlinson et al. (2010)⁴⁰ and guidance from Espay et al. (2017).⁴¹ The dose and schedule of oral LD-CD were adjusted over the first 7 to 14 days by the investigator to optimize control of motor symptoms. An optimal clinical response was defined as maximizing functional “on” time and minimizing the number of “off” episodes during the day and “on” time with troublesome dyskinesia. Once stabilized, the dosing regimen remained unchanged for at least 7 days before the start of the double-blind treatment period.

In both oral LD-CD optimization and double-blind treatment periods, the following concomitant PD medications were allowed: nonergolinic dopamine agonists, selective MAO-B inhibitors, amantadine, safinamide, and istradefylline. These medications, if taken, were maintained at the already optimal and stabilized prerandomization dose and schedule for the remainder of the study. No changes were allowed unless, in the investigator’s considered opinion, they were medically necessary (e.g., management of dyskinesia, safety).

Prohibited medication included apomorphine, LCIG, LD-CD enteral suspension, dopamine-depleting agents, MAO-A inhibitor, nonselective MAO-B inhibitors, ergot dopamine agonists, dopamine antagonist or partial agonist, first generation antipsychotics, or antiemetic medications that interact with brain dopamine receptors, oral and/or inhaled medications containing levodopa (other than study drugs), and COMT inhibitors.

Double-Blind Treatment Period (Day 1 to Day 85 [Week 12])

In the CSCI optimization phase (day 1 to day 29), all patients first received 1 loading dose of each of oral LD-CD (100 mg/25 mg) tablets and foslevodopa (240 mg/mL) and foscarbidopa (12 mg/mL) on day 1 at a dose (i.e., number of tablets and volume of infusion) based on the first dose of oral LD-CD of the day at the end of the oral LD-CD stabilization period.

• Foslevodopa-foscarbidopa arm: After the dual loading doses, patients started receiving CSCI (24 hours per day) of foslevodopa-foscarbidopa solution via a portable infusion pump plus oral placebo capsules for LD-CD IR. The initial infusion rate of foslevodopa-foscarbidopa was calculated based on the patient’s stabilized oral LD-CD IR therapy at the end of the oral LD-CD stabilization period and a conversion algorithm based on the pharmacokinetic characteristics of foslevodopa-foscarbidopa established from phase I studies. The oral placebo capsules administered matched, in number of tablets and dosing frequency, the therapeutic schedule achieved during the end of the oral LD-CD stabilization period.

• Oral LD-CD arm: Patients received CSCI of placebo solution (0.9% weight/volume NaCl solution) in addition to oral LD-CD tablets and the infusion rate, and dose and frequency of oral administration were based on the (conversion of the) dosing of oral LD-CD tablets achieved at the end of the oral LD-CD stabilization period.

In the CSCI optimization phase, only changes to the CSCI rate were permitted in both treatment arms. The investigator adjusted infusion rates after an assessment of PD symptoms and clinical response during each visit to optimize clinical response. The allowable infusion rate ranged from 0.15 mL per hour to 1.04 mL per hour.

In the maintenance phase (day 29 to day 85), patients continued with the same treatment regimen for both CSCI and oral therapies until the end of study, with no dose adjustments allowed.

Training on the correct use of the infusion pump was provided in the screening and oral LD-CD stabilization periods, and all patients had to demonstrate familiarity with the use of the infusion pump to be enrolled in the trial. The Phillips-Medisize Parkinson’s Disease Subcutaneous Pump Delivery System was used (synonymous with the Vyafuser infusion pump approved for use with foslevodopa-foscarbidopa in Canada).

Open-label rescue treatment with oral LD-CD (100 mg/25 mg) IR tablets could be administered as needed in case of sudden deterioration of clinical condition (e.g., motor symptoms) in the double-blind treatment period. The suggested dosing or maximum dose was not stated in the submission.

All patients, caregivers, and investigators were blinded to treatment allocation. LD-CD IR tablets were over-encapsulated to look identical to the placebo capsule and identical packaging was used. Since foslevodopa-foscarbidopa and the placebo solution might not appear identical, and given the nature of the study design, several measures were undertaken to help maintain the blind and prevent functional unblinding. These measures included providing identical packaging of both study drug solutions, requiring separation of site personnel roles, including the use of a separate blinded rater (different from the treating investigator) to perform all in-person efficacy assessments, and disabling the available pump features (extra dose, and high and low infusion rates) designed for the intended commercial pump. In addition, the loading dose was only available during dual loading on day 1; it was then disabled by the investigator for the remainder of the study.

Outcomes

A list of efficacy end points assessed in this clinical review report are provided in Table 7. These end points are further summarized in the following. Summarized end points are based on those included in the sponsor’s Summary of Clinical Evidence¹⁵ as well as any identified as important to this review according to stakeholders, for example, the clinical expert, clinician groups, or patient groups.

Table 7: Summary of Outcome Measures in the M15-736 Trial

AE = adverse event; AESI = adverse event of special interest; C-SSRS = Columbia-Suicide Severity Rating Scale; ECG = electrocardiogram; EQ-5D-5L = 5-Level EQ-5D; HRQoL = health-related quality of life; IQR = interquartile range; MDS-UPDRS = Movement Disorder Society-Unified Parkinson’s Disease Rating Scale; PD = Parkinson disease; PDQ-39 = Parkinson’s Disease Questionnaire-39 items; PDSS-2 = Parkinson’s Disease Sleep Scale-2; QoL = quality of life; QUIP-RS = Questionnaire for Impulsive-Compulsive Disorders in Parkinson’s Disease – Rating Scale; SAE = serious adverse event.

^aThese end points were included in the multiple testing hierarchy.

^bMorning akinesia was defined as reporting “off” status as the first morning symptom upon awakening.

Source: M15-736 Clinical Study Report.¹⁴ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Efficacy Outcomes

Baseline for PD diary variables was defined as the average of the 3 valid diary entries completed before day –1. Baseline for all other efficacy measures was defined as the last nonmissing assessment before the initiation of double-blinded study drug (i.e., visit 5). Study visits were scheduled at baseline, days 1, 2, 8, 15, 22, 29, 57, and 85 (week 12).

“On” and “Off” Times Outcomes (PD Diary)

Change from baseline at week 12 in average daily normalized “on” time without troublesome dyskinesia (primary end point), “on” time without dyskinesia (secondary), and “off” time (key secondary) were assessed using data collected in the 3 most recent valid PD diary days before each visit (not measured at the day 2 visit). The recorded time spent in each motor state was normalized to a typical waking day of 16 hours (e.g., by dividing the absolute “off” time by the daily awake time, then multiplying by 16 hours). A valid PD diary day was defined as 1 within 7 days before a clinical visit with no more than 2 hours of missing data for the 24-hour diary.

Presence of morning akinesia was a key secondary end point, and morning akinesia was defined as reporting “off” status as the first morning symptom upon awakening and assessed based on the PD diary.

At screening, all patients received a PD diary training to learn about proper completion of the diary. An assessment was conducted to ensure minimum 75% concordance between the patient’s PD diary and the PD diary completed by the investigator before the patient was enrolled in the study. Patients were also required to complete the PD diary for 3 consecutive days before the second screening visit and all subsequent study visits. All patients were reminded to complete their diary entries before each visit to reinforce the importance of PD diary completion. Patients who incorrectly completed their PD diary received retraining.

For a description of the PD diary, as well as its psychometric properties, and minimally important difference (MID) estimates, refer to Table 8.

Movement Disorder Society-Unified Parkinson’s Disease Rating Scale

Changes from baseline to week 12 in MDS-UPDRS Part II score (key secondary end point) and Part I, III, and IV scores (exploratory end points) were assessed. The instrument was administered during the patient’s best “on” time and, if possible, at the same time of day at each study visit, except at day 1 visit. A separate rater (i.e., other than the site personnel who provide the study drug or the treating investigator) was responsible for performing MDS-UPDRS assessments. A description of the MDS-UPDRS instrument, as well as its psychometric properties, and MID estimates are given in Table 8.

Parkinson’s KinetiGraph and Personal KinetiGraph

The secondary outcomes of median and IQR of bradykinesia and dyskinesia scores were measured using the PKG. The PKG is a movement-recording device that the patients wore on their wrists continuously during the double-blind treatment period to collect data during the activities of daily living in the home environment. The PKG is used to assess PD symptoms such as tremor, bradykinesia, and dyskinesia.

For each patient, the PKG collected data continuously, and an algorithm calculated a bradykinesia score every 2 minutes between 9 a.m. and 6 p.m. across multiple days, with the median bradykinesia score calculated at a certain visit for each patient. There was no prespecified range of scores. A higher median score indicated worse bradykinesia. A higher bradykinesia IQR score (calculated as the difference between the third quartile and first quartile bradykinesia scores) indicated a higher degree of variability in bradykinesia score. The scoring of median and IQR of dyskinesia scores was the same.⁴²

Parkinson’s Disease Sleep Scale-2

Change from baseline to week 12 in PDSS-2 total score was a key secondary outcome, and measurements were made at the baseline and day 85 (weeks 12) visits.

The PDSS-2 scale was administered by central raters who did not have access to results of other study assessments or patients’ medical records, and did not participate in the care or management of patients. The instrument is designed to characterize different aspects of nocturnal sleep problems in patients with PD. It consists of 15 items that assesses 3 domains of sleep symptoms: motor symptoms at night, PD symptoms at night, and disturbed sleep. It assesses overall sleep quality, insomnia, sleep fragmentation, restless leg syndrome, hallucinations, sleep apnea, akinesia, pain, nocturnal immobility, and rapid eye movement behaviour disorder. The total score ranged from 0 to 60, where a higher total score indicates more severe nocturnal sleep problems.

Parkinson’s Disease Questionnaire-39 Items

Change from baseline to final visit (week 12) in PDQ-39 summary index was a secondary outcome, and measurements were made at the baseline and day 85 (weeks 12) visits. A description of the PDQ-39 instrument, as well as its psychometric properties, and MID estimates are presented in Table 8.

5-Level EQ-5D

Change from baseline to final visit (week 12) in EQ-5D-5L summary index was a key secondary outcome. The EQ-5D-5L instrument was measured at the baseline and day 85 (weeks 12) visits.

EQ-5D-5L is a generic PRO instrument for measuring HRQoL that can be used in a wide range of health conditions and treatments (including PD). The instrument measures patients’ overall health status by generating an index-based summary score on 5 health dimensions (mobility, self-care, daily activities, pain, and anxiety) based upon societal preference weights. In the study, the health status was converted to an index value that ranged from a worst score of –0.109 to a best score of 1 using the US-specific weighted scoring algorithm.

Harms Outcomes

Safety end points (secondary end points) included local and systemic tolerability based on AEs (at all visits), the Infusion Site Evaluation Scale (at all visits except for baseline and day 1), clinical and laboratory values (at baseline and day 85), electrocardiograms (at baseline, days 1, 15, 29, and 85), C-SSRS scores (all visits), and QUIP-RS scores (at baseline, days 15, 29, and 85).

Infusion Site Evaluation Scale

An investigator or qualified designee evaluated the infusion site using a 2-part (numeric and letter grading) scale to assess irritation. Numeric grades ranged from 0 (no evidence of irritation) to 7 (strong reaction spreading beyond the test site), and letter grading from A (no finding) to G (small petechial erosions and/or scabs). Any score above 2 or C was recorded as an AE.

Columbia-Suicide Severity Rating Scale

The C-SSRS, a patient-administered instrument, was used to assess suicidal behaviours and ideation, track and assess all suicidal events, and assess lethality of attempts. Additional features assessed include frequency, duration, controllability, reason for ideation, and deterrents. Patients were asked to indicate if they had suicidal ideation with plan within the last year by answering yes to questions in the suicidal ideation portion of the instrument or in a clinical interview.

Questionnaire for Impulsive-Compulsive Disorders in Parkinson’s Disease — Rating Scale

The QUIP-RS was administered by an investigator to assess severity of impulse-control disorders and related behaviours reported to occur in PD. The questionnaire uses a 5-point scale for respondents to rate the severity of each symptom based on its frequency, from 0 (never) to 4 (always). The QUIP-RS has 4 primary questions. A higher score indicated greater severity.

Table 8: Summary of Select Outcome Measures and Their Measurement Properties

ADL = activities of daily living; HRQoL = health-related quality of life; ICC = intraclass correlation; IR = immediate release; MDS = Movement Disorder Society; MDS-UPDRS = Movement Disorder Society-Sponsored Revision of the Unified Parkinson’s Disease Rating Scale; MID = minimally important difference; PD = Parkinson disease; PDQ-39 = Parkinson’s Disease Questionnaire-39 items; SF-36 = 36-Item Short Form Survey; SLR = systematic literature review; UPDRS = Unified Parkinson’s Disease Rating Scale.

^aIn the M15-736 trial, the questionnaire was administered by a blinded central rater via a telephone interview.

Statistical Analysis

Sample Size and Power Calculation

A sample size calculation determined that a total of 130 randomized patients was required to demonstrate statistically significant difference between the foslevodopa-foscarbidopa arm and oral LD-CD arm with respect to the change from baseline to week 12 in average daily normalized “on” time without troublesome dyskinesia (primary end point) at a 2-sided significance level of 0.05 with a power of 90%, assuming a treatment difference between foslevodopa-foscarbidopa and oral LD-CD of 1.86 hours and an SD of 2.9 hours, and 20% of treatment discontinuation. This sample size also has approximately 90% power for key secondary end points of change from baseline in average daily normalized “off” time, MDS-UPDRS Part II score, and presence of morning akinesia at week 12.

Primary Efficacy End Point: Statistical Test and Model

In the primary analysis, the difference in the mean change from baseline to week 12 in average daily normalized “on” time without troublesome dyskinesia between treatment arms in the full analysis set was tested at a 2-sided significance level of 0.05 using a mixed model for repeated measures (MMRM). The model was stratified by categorical fixed effects of treatment, country, and visit, treatment-by visit and treatment-by-baseline interactions, and the continuous fixed covariate of baseline measurement, assuming an unstructured variance covariance matrix.

Primary Efficacy End Point: Subgroup and Sensitivity Analyses

Prespecified subgroup analyses were performed by age, sex, race, country, duration of PD, concomitant dopamine agonist use, and levodopa dose category. Subgroup analyses by age (younger than 65 years versus at least 65 years), duration of PD (less than 10 years versus at least 10 years), and dose category (low versus high levodopa dose) are of interest to this review based on input from the clinical expert. The analyses were performed on the full analysis set using an analysis of covariance (ANCOVA) model adjusted for the effect of treatment, subgroup variable, the treatment-by-subgroup variable interaction, and baseline measurement as a covariate. No adjustment for type I error was made. The treatment-by-subgroup interaction was tested at the significance level of 0.10.

Two sensitivity analyses were performed to account for missing data:

• “Jump-to-reference” analytic approach – All missing data in the oral LD-CD arm and nonmonotonic missing data in the foslevodopa-foscarbidopa arm were assumed to be missing at random. For patients in the foslevodopa-foscarbidopa arm with monotonic missing values for reasons other than COVID-19 infections or logistical restrictions, missing values were assumed to be missing not at random and have a profile for visits after discontinuation of treatment that equals the profile of the oral LD-CD arm.

• Last available value approach – Missing week 12 data were imputed based on the patient’s last available value in estimating the change from baseline to week 12 based on an ANCOVA model with the categorical fixed effects of treatment and country, and baseline score as a covariate.

Primary Efficacy End Point: Handling of Intercurrent Events and Missing Data

Handling of missing items in efficacy instrument are summarized in Table 9. Intercurrent events related to discontinuation of study drug were handled using the hypothetical strategy, where all data collected after the last dose of blinded study drug were not used in the analyses. Intercurrent events related to rescue medication use were handled using treatment policy strategy, where collected data were used regardless of recue medication use. Missing data were implicitly imputed using a MMRM based on the assumption that data were missing at random.

Secondary Efficacy End Point: Multiple Testing Procedure

The key secondary and other secondary outcomes were included in multiplicity adjustment of the type I error to control the family-wise error rate at a 2-sided significance level of 0.05 in the hierarchal order, as outlined in Table 9.

Secondary Efficacy End Point: Statistical Test and Model

Change from baseline in “off” and “on” times end points, MDS-UPDRS Part II score, and median and IQR of bradykinesia and dyskinesia scores were analyzed using the same statistical approach as the primary analysis of the primary end point. The presence of morning akinesia at week 12 was analyzed using a generalized linear mixed model adjusted for categorical fixed effects of treatment, country, visit, treatment-by-visit interaction, and baseline first morning status upon awakening. The change from baseline to final visit in PDSS-2, PDQ-9, EQ-5D-5L scores were analyzed using an ANCOVA model adjusted for categorical fixed effects of treatment and country, and baseline score. The methods of handling intercurrent events and missing data were the same as the primary end-point analysis.

For all “on” and “off” times outcomes, subgroup and sensitivity analyses were conducted using the same approaches as for the primary end point. For all other key secondary and other secondary end point, only subgroup analyses by dose category were conducted.

Safety End Points

Safety outcomes were compared between treatment arms using descriptive statistics and no hypothesis testing was performed.

Table 9: Statistical Analysis of Efficacy End Points in the M15-736 Trial

ANCOVA = analysis of covariance; EQ-5D-5L = 5-Level EQ-5D; GLMM = generalized linear mixed model; IQR = interquartile range; MDS-UPDRS = Movement Disorder Society-Unified Parkinson’s Disease Rating Scale; MMRM = mixed model for repeated measures; PDQ-39 = Parkinson’s Disease Questionnaire-39 items; PDSS-2 = Parkinson’s Disease Sleep Scale-2; VAS = visual analogue scale.

^aThe results for this outcome should not be interpreted as statistically significant if the P value for the comparison is < 0.05 due to failure of a prior end point in the statistical testing hierarchy (MDS-UPDRS Part II score).

Source: M15-736 Clinical Study Report.¹⁴ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Analysis Populations

Analysis populations of M15-736 are summarized in Table 10.

Table 10: Analysis Populations in the M15-736 Trial

FAS = full analysis set; SAS = safety analysis set.

Note: The M15-736 trial also included an oral levodopa-carbidopa analysis set, but it was not of interest to this review and was not summarized.

Source: M15-736 Clinical Study Report.¹⁴ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Results

Patient Disposition

A summary of patient disposition in the M15-736 trial is shown in Table 11. Of 270 screened patients, 96 (35.6%) were screen failures, primarily because of failure to meet eligibility criteria (18.5%) and consent withdrawal (13.3%). Both the full and the safety analysis sets included 141 patients who were randomized to the foslevodopa-foscarbidopa arm (N = 74; 52.5%) and the oral LD-CD arm (N = 67; 47.5%) in the double-blind treatment period. Of the patients who discontinued study treatment, 26 (35.1%) were in the foslevodopa-foscarbidopa arm and 5 (7.5%) patients in the oral LD-CD arm. The most common reason for study treatment discontinuation in the foslevodopa-foscarbidopa arm was AEs (18.9%).

Table 11: Summary of Patient Disposition in the M15-736 Trial (SAS)

CD = carbidopa; FAS = full analysis set; LD = levodopa; SAS = safety analysis set.

Source: M15-736 Clinical Study Report.¹⁴ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Baseline Characteristics

A summary of baseline patient demographics, disease characteristics, and medication history in M15-736 trial is shown in Table 12. The baseline characteristics outlined in the table are limited to the ones that are most relevant to this review or that were considered to most likely impact the outcomes or interpretation of the study results.

The mean age of the study population was 66.4 years (SD = 9.5 years). The majority of patients were male (70.2%) and white (92.9%). Mean time since diagnosis was 8.6 years (SD = 4.9 years), and mean time since onset of motor fluctuation was 5.6 years (SD = 4.0 years). The majority of patients had a history of levodopa-induced dyskinesia (62.4%). The mean time spent in “off” states and “on” without troublesome dyskinesia motor states were 6.1 hours (SD = 2.1 hours) and 9.3 hours (SD = 2.5 hours), respectively. Mean dosing frequency was 6.2 times per day (SD = 2.6 times per day). The baseline characteristics were similar in the treatment arms, except that more patients in the foslevodopa-foscarbidopa arm had levodopa response of more than 5 years (77.0%) and received prior dopamine agonist treatment (56.8%) compared with the oral LD-CD arm (68.7% and 46.3%, respectively).

Table 12: Summary of Baseline Characteristics in the M15-736 Trial (SAS)

CD = carbidopa; COMT = catechol-O-methyltransferase; DBS = deep brain stimulation; IR = immediate release; LD = levodopa; MAO-B = monoamine oxidase type B; MDS-UPDRS = Movement Disorder Society-Unified Parkinson’s Disease Rating Scale; NR = not reported; PD = Parkinson disease; SAS = safety analysis set; SD = standard deviation.

^aBased on the PD diary.

^bRefers to PD medications other than oral LD-CD IR tablets.

Source: M15-736 Clinical Study Report.¹⁴ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Exposure to Study Treatments

Treatment exposure from the M15-736 trial is summarized in Table 13. Mean average total daily levodopa dose was 1,667 mg per day (SD = 741 mg per day) in the foslevodopa-foscarbidopa arm and 1,105 mg per day (SD = 583 mg per day) in the oral LD-CD arm. Mean duration of study drug exposure was 62.4 days (SD = 31.2 days) in the foslevodopa-foscarbidopa arm and 81.5 days (SD = 19.0 days) in the oral LD-CD arm. Mean treatment adherence for drug solution and drug capsule was a least 85% in both treatment arms.

Concomitant Medications

Concomitant PD medication use in the M15-736 trial is summarized in Table 14. In the trial, 74.3% of patients in the foslevodopa-foscarbidopa arm and 67.2% of patients received at least 1 concomitant PD medications. The most common class of PD medications (in at least 30% of patients in either treatment arm) were dopamine agonist and MAO-B inhibitors. A higher proportion of patients receiving dopamine agonist in the foslevodopa-foscarbidopa arm (45.9%) than in the oral LD-CD arm (37.3%). No notable between-arm differences were observed for the other PD medications.

Table 13: Treatment Exposure in the M15-736 Trial (SAS)

CD = carbidopa; LD = levodopa; SAS = safety analysis set; SD = standard deviation.

^aStudy drug solution adherence (%) = hours of infusion divided by 24, then multiply by 100.

^bstudy drug capsules adherence (%) = number of capsules taken divided by number of capsules prescribed, the multiply by 100.

Source: M15-736 Clinical Study Report.¹⁴ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Table 14: Concomitant PD Treatment in the M15-736 Trial (SAS)

CD = carbidopa; LD = levodopa; MAO-B = monoamine oxidase type B; PD = Parkinson disease; SAS = safety analysis set.

^aNote that patients could have received more than 1 concomitant medication during the treatment period.

Source: M15-736 Clinical Study Report.¹⁴ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Rescue Medications

Oral LD-CD was used as a rescue medication. A summary of its use is shown in Table 15. Oral LD-CD tablet as a rescue medication was required by 68.9% of patients in the foslevodopa-foscarbidopa arm and 56.7% oral LD-CD arm. Mean number of doses of rescue medication administered was 44.8 (SD = 52.5) in the foslevodopa-foscarbidopa arm and 77.5 (SD = 125.8) in the oral LD-CD arm. Mean total daily levodopa dose from rescue medications was numerically higher in the oral LD-CD arm than foslevodopa-foscarbidopa arm at most time points.

Table 15: Rescue Medication Use in the M15-736 Trial (SAS)

CD = carbidopa; IR = immediate release; LD = levodopa; SAS = safety analysis set; SD = standard deviation.

Source: M15-736 Clinical Study Report.¹⁴ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Protocol Deviations

Protocol deviations in the M15-736 trial are summarized in Table 16. The most common category of protocol deviations was related to study procedure; this occurred in 45.9% of patients in the foslevodopa-foscarbidopa arm and 40.3% of patients in the oral LD-CD arm. Study procedure deviations were most commonly related to extended use of single-use syringe (beyond 24 hours per day) and incomplete recording of PD diary, questionnaire, and rescue dose, although the incidences were not specified.

The sponsor noted that none of the protocol deviations recorded affected the study outcome or interpretation of the study results or conclusions.

Table 16: Protocol Deviations in the M15-736 Trial

CD = carbidopa; LD = levodopa.

^aIn the M15-736 trial, protocol compliance-study procedure deviations included, in decreasing frequency (exact frequencies not stated): single syringe use greater than 24 hours per dosing diary entry, insufficient days completed or incomplete Parkinson disease (PD) diary entries, missed or late questionnaire completion, rescue dose not recorded in dosing diary, study visit completed remotely not due to COVID-19 or before approval of protocol version 4.0, and incorrect loading dose.

^bOf the 14 patients who received wrong treatment or incorrect dose in the M15-736 trial, 12 received an incorrect infusion loading dose of 0.5 mL rather than 0.6 mL on day 1 as a result of an Interactive Randomization Technology programming error.

^cOf the 8 patients who entered the M15-736 trial but did not satisfy entry criteria, 1 patient’s total daily dose of LD equivalents from LD-containing medications and catechol-O-methyltransferase inhibitors did not meet the minimum 400 mg requirement, 1 patient did not meet the requirement of at least 44 of 48 PD diary entries on 1 of the 3 diary days and also did not have at least 2 hours of “Off” time on 2 of the 3 diary days at baseline, and the other 6 patients did not meet the requirement of at least 44 of 48 PD diary entries on 1 of the 3 diary days at baseline.

Source: M15-736 Clinical Study Report.¹⁴ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Efficacy

Unless otherwise specified, efficacy end points from the M15-736 trial that were noted to be important to patients and clinicians based on stakeholder input are summarized in Table 17.

“On” Time Outcomes

“On” Time Without Troublesome Dyskinesia

Change from baseline to week 12 in average daily normalized “on” time without troublesome dyskinesia was the primary end point. In the primary analysis, the LSM difference between the foslevodopa-foscarbidopa arm and the oral LD-CD arm was 1.75 (95% CI, 0.46 to 3.05; P = 0.0083) hours, in favour of foslevodopa-foscarbidopa. Results of the sensitivity analyses (jump-to-reference analytic approach and last available value approach) (refer to Table 46 in Appendix 1) and subgroup analyses of interest (age, duration of PD diagnosis, and levodopa dose intensity) (refer to Figure 4 in Appendix 1) were consistent with the primary analysis.

Results from the analyses of change from baseline in absolute hours and percent change from baseline in normalized hours (Table 47 and Table 48 in Appendix 1) were consistent with the primary analysis of normalized change from baseline.

“On” Time Without Dyskinesia

Change from baseline to week 12 in average daily normalized “on” time without dyskinesia was a secondary end point and is summarized in Table 49 in Appendix 1. The results for this outcome are at increased risk of type I error (false-positive results) because they were tested after failure of the statistical hierarchy. The LSM difference between the foslevodopa-foscarbidopa arm and the oral LD-CD arm was 1.81 (95% CI, 0.46 to 3.16) hours.

“Off” time

“Off” Time

Change from baseline to week 12 in average daily normalized “off” time without dyskinesia was a study-defined key secondary end point and was adjusted for multiplicity. The LSM difference between the foslevodopa-foscarbidopa arm and the oral LD-CD arm was –1.79 (95% CI, –3.03 to –0.54, P = 0.0054) hours, in favour of foslevodopa-foscarbidopa. Results of the sensitivity analyses (jump-to-reference analytic approach and the last available value approach, Table 46 in Appendix 1) and subgroup analyses of interest (age, duration of PD diagnosis, and levodopa dose intensity, Figure 5 in Appendix 1) were consistent with the primary analysis.

Results from the analyses of change from baseline in absolute hours and percent change from baseline in normalized hours (Table 47 and Table 48 in Appendix 1) were consistent with the primary analysis of normalized change from baseline.

Presence of Morning Akinesia

Presence of morning akinesia at week 12 was a study-defined key secondary end point and is summarized in Table 50 in Appendix 1. The results for this outcome are at increased risk of type I error because they were tested after failure of the statistical hierarchy. At baseline, morning akinesia was present in 78.9% of patients in the foslevodopa-foscarbidopa arm, and 76.1% of patients in the oral LD-CD arm. At week 12, morning akinesia was present in 17.0% of patients in the foslevodopa-foscarbidopa arm and 63.3% of patients in the oral LD-CD arm, with an odds ratio of 0.12 (95% CI, 0.04 to 0.31).

Results of subgroup analyses (by levodopa dose intensity) were not reported.

Health-Related Quality of Life

Parkinson’s Disease Questionnaire-39 Items

Change from baseline to week 12 in PDQ-39 summary index total score was a secondary end point. The results for this end point are at increased risk of type I error because they were tested after failure of the statistical hierarchy. The LSM difference between the foslevodopa-foscarbidopa arm and the oral LD-CD arm was –4.10 (95% CI, –8.14 to –0.05).

5-Level EQ-5D

Change from baseline to week 12 in EQ-5D-5L summary index score was a secondary end point. The results for this end point are at increased risk of type I error because they were tested after failure of the statistical hierarchy. Results are summarized in Table 49 in Appendix 1. The LSM difference between the foslevodopa-foscarbidopa arm and the oral LD-CD arm was 0.049 (95% CI, –0.001 to 0.100).

PD Symptoms

Movement Disorder Society-Unified Parkinson’s Disease Rating Scale

Change from baseline to week 12 in MDS-UPDRS Part I (nonmotor experiences of daily living), Part II (motor experiences of daily living), Part III (motor examination), and Part IV (motor complications) were measured.

Part II score was a study-defined key secondary end point and was adjusted for multiplicity. The LSM difference between the foslevodopa-foscarbidopa arm and the oral LD-CD arm in change from baseline to week 12 in MDS-UPDRS Part II score was –1.58 (95% CI, –3.65 to 0.48; P = 0.13).

Part I, III, and IV scores were exploratory end points and were not adjusted for multiplicity. The LSM difference between the foslevodopa-foscarbidopa arm and the oral LD-CD arm in change from baseline to week 12 in Part I, III, and IV scores were 0.00 (95% CI, –1.95 to 1.95), 0.32 (95% CI, –3.62 to 4.26), and –2.22 (95% CI, –3.36 to –1.08), respectively.

Bradykinesia and Dyskinesia Scores Assessed Using the PKG

Change from baseline to week 12 in median bradykinesia score, bradykinesia score IQR (difference between third and first quartile bradykinesia scores), median dyskinesia score, and dyskinesia score IQR, assessed using the PKG wearable device, were secondary end points. The results for these outcomes are at increased risk of type I error (false-positive results) because they were tested after failure of the statistical hierarchy. Results of these analyses are summarized in Table 49 in Appendix 1.

The LSM difference between the foslevodopa-foscarbidopa arm and the oral LD-CD arm in change from baseline to week 12 in median bradykinesia, bradykinesia IQR, median dyskinesia, and dyskinesia IQR scores were 1.72 (95% CI, 0.30 to 3.15), 0.18 (95% CI, –1.20 to 1.55), –2.73 (95% CI, –6.61 to 1.15), and –5.49 (95% CI, –12.71 to 1.73), respectively.

PDSS-2 (Sleep Symptoms)

Change from baseline to week 12 in PDSS-2 total score was a secondary end point and the results for this end point are at increased risk of type I error because they were tested after failure of the statistical hierarchy. The LSM difference between the foslevodopa-foscarbidopa arm and the oral LD-CD arm was –5.40 (95% CI, –8.03 to –2.78).

Table 17: Summary of Key Efficacy Results From the M15-736 Trial (FAS)

CD = carbidopa; CI = confidence interval; FAS = full analysis set; FOS-FOS = foslevodopa-foscarbidopa; LD = levodopa; LSM = least squares mean; MDS-UPDRS = Movement Disorder Society-Unified Parkinson’s Disease Rating Scale; PD = Parkinson disease; PDQ-39 = Parkinson’s Disease Questionnaire-39 items; PDSS-2 = Parkinson’s Disease Sleep Scale-2; QoL = quality of life; SD = standard deviation; SE = standard error.

Note: Outcomes summarized in this table were noted to be important to patients and clinicians based on input received from patient groups, clinician groups, and the clinical expert consulted by CADTH.

^aThe analysis was conducted using a mixed model for repeated measures, adjusted for categorical fixed effects of treatment, country and visit, treatment-by-visit and treatment-by-baseline interactions, and baseline measurement (continuous).

^bThis analysis was conducted using an analysis for covariate model, adjusted for categorical fixed effects of treatment and country, and baseline score (continuous).

^cAlthough the P value is < 0.05, statistical significance cannot be claimed because the results for the second key secondary end point (MDS-UPDRS Part II), a prior end point in the testing hierarchy, were not statistically significant.

^dNot adjusted for multiplicity.

Source: M15-736 Clinical Study Report.¹⁴ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Harms

A summary of harms in the M15-736 trial is shown in Table 18. Results of the M15-736 trial referred to findings in the double-blind treatment period.

Adverse Events

TEAEs were reported for 85.1% of patients in the foslevodopa-foscarbidopa arm and 62.7% of patients in the oral LD-CD arm. The most common TEAEs in the foslevodopa-foscarbidopa arm (reported in at least 10% of patients) were infusion-site erythema, pain, cellulitis, and edema, as well as dyskinesia, all of which were more commonly reported than in the oral LD-CD arm (infusion-site erythema and pain, 1.5% each). The incidence of falls was lower in the foslevodopa-foscarbidopa arm (8.1%) than in the oral LD-CD arm (17.9%).

Serious Adverse Events

Serious TEAE was reported for 6 (8.1%) patients in the foslevodopa-foscarbidopa arm (1 each: infusion-site cellulitis, catheter-site cellulitis, dehydration, migraine, psychiatric disorder, prostatomegaly) and 4 (6.0%) patients in the oral LD-CD arm (1 each: COVID-19 pneumonia, cellulitis, contusion or fall, catheter-site cellulitis).

Withdrawals Due to AEs

Treatment discontinuation due to TEAEs was reported for 21.6% of patients in the foslevodopa-foscarbidopa arm and 1.5% of patients in the oral LD-CD arm. The most common TEAEs leading to treatment discontinuation in the foslevodopa-foscarbidopa arm were infusion-site cellulitis (5.4%), infusion-site pain (4.1%), infusion-site bruising, hemorrhage, and edema (2.7% each). One patient in the oral LD-CD arm discontinued treatment due to cellulitis.

Mortality

No deaths were reported in the foslevodopa-foscarbidopa arm, and 1 (1.5%) death was reported in the oral LD-CD arm. The death was related to a serious TEAE of acute respiratory failure that was, according to the sponsor, deemed unrelated to the study drug.

Notable Harms

Infusion-Site Reactions and Infusion-Site Infections

The incidences of infusion-site reactions and infusion-site infections were notably higher in the foslevodopa-foscarbidopa arm than in the oral LD-CD arm (infusion-site reactions: 62.2% versus 7.5%; infusion-site infections: 28.4% versus 3.0%). Six patients (8.1%) in the foslevodopa-foscarbidopa arm and no patients in the oral LD-CD arm had at least 1 incidence of numeric grade at least 5 and letter grade at least D on the Infusion Site Evaluation Scale.

Hallucination or Psychosis, Impulse-Control Disorder, and Suicidality

The incidence of hallucination or psychosis was notably higher in the foslevodopa-foscarbidopa arm (14.9%) than in the oral LD-CD arm (3.0%). There was no report of impulse-control disorder or impulsive behaviour in either treatment arm. There was no notable between-arm difference in the mean change from baseline in score for each impulse-control disorder and related behaviour parameters of the QUIP-RS across almost all time points. Based on the C-SSRS assessment, 5 (6.8%) patients in the foslevodopa-foscarbidopa arm and 2 (3.0%) patients in the oral LD-CD arm had suicidal behaviours or ideations.

Dizziness and Orthostatic Hypotension

Dizziness was reported in 3 patients in each treatment arm. Orthostatic hypotension by preferred term was reported in 1 (1.4%) patient in the foslevodopa-foscarbidopa arm, and 2 (3.0%) patients in the oral LD-CD arm.

Somnolence

Somnolence was reported in 1 (1.4%) patient in both treatment arms.

Depression

Depression was reported in 0 patients in the foslevodopa-foscarbidopa arm, and 2 (3.0%) patients in the oral LD-CD arm.

Table 18: Summary of Harms — Pivotal and RCT Evidence (M15-736) (SAS)

CD = carbidopa; LD = levodopa; SAS = safety analysis set; TEAE = treatment-emergent adverse event.

Note: Unless otherwise specified, all TEAEs were summarized using Medical Dictionary for Regulatory Activities (MedDRA) preferred terms.

^a≥ 10% of patients.

^b≥ 2% of patients.

^cAdverse event of special interest, including any adverse event in infusion site–related noninfection reactions company MedDRA query (CMQ).

^dAdverse event of special interest, including any adverse event in the hallucinations CMQ or psychosis and psychotic disorder standardized MedDRA query (SMQ).

^eAssessed using the Columbia-Suicide Severity Rating Scale.

Source: M15-736 Clinical Study Report.¹⁴ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Critical Appraisal

Systematic Literature Review

A systematic literature review (SLR) was conducted by the sponsor to identify studies for inclusion into the submission. The prespecified selection criteria (population, intervention, comparison, outcomes, and study [PICOS]) were appropriate for identifying relevant studies. The criteria were broad with respect to comparator and study designs, including comparators irrelevant to the Canadian context (e.g., subcutaneous apomorphine injections, levodopa-entacapone-carbidopa intestinal gel) and a variety of study designs (phase II, III, and IV clinical trials and observation studies), although narrower criteria were subsequently applied to tailor the review to include phase II, III, and IV clinical trials of foslevodopa-foscarbidopa with results or data available. It is unclear if these narrower criteria were prespecified, but the risk of missing relevant studies due to these additional criteria appears to be low.

The literature search was comprehensive, including multiple databases (i.e., MEDLINE [In-Process], Embase, and Cochrane Library) and clinical trial registries (ClinicalTrials.gov, International Clinical Trials Registry Platform, Health Canada’s Clinical Trials Database, and European Union Clinical Trials Register) to identify literature published for all time until 6 months before submission to CADTH. Only publications in English were included, for which no justification was provided. (It is recommended that search strategies not be limited by language to minimize publication or language bias, and if such restriction is applied, justification should be provided.^68,69) Appropriate study selection and data extraction methods were used. Studies were selected independently by 2 reviewers with a third reviewer to resolve discrepancies. Study selection was appropriately presented using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) approach, and reasons for article exclusion were documented. Data extraction was performed by 1 reviewer and independently reviewed by a second reviewer, followed by a quality check by a third reviewer. A risk of bias evaluation for the studies included in the SLR was conducted using the Cochrane risk of bias tool. A high-level summary of the risk of bias of included studies was provided, but results for individual included studies were not given.

Pivotal and RCT Evidence

Internal Validity

The M15-736 trial was a randomized, double-blind, double-dummy, active-controlled trial. The methods of randomization, which involved stratification by study site, and interactive response technology for concealment of the randomized assignment, were appropriate. There was no notable difference between treatment arms for most baseline characteristics. The foslevodopa-foscarbidopa arm had a higher proportion of patients who had levodopa response for more than 5 years and had prior dopamine agonist use compared with the oral LD-CD arm, although the clinical expert consulted by CADTH noted that these differences would not be clinically significant in the studied population and were unlikely to have confounded the study results.

As foslevodopa-foscarbidopa and oral LD-CD require a different route of administration, a double-dummy design was used to preserve blinding for patients and investigators. The blinding procedures were in general appropriate; however, an end-of-study survey aiming to identify the extent of unblinding showed that the majority of patients (74.3% in the foslevodopa-foscarbidopa arm; 71.9% in the oral LD-CD arm) were aware of treatment assignment, most commonly for reasons related to change (or absence of change) in PD symptoms. MDS-UPDRS, PDSS-2, and PDQ-39 assessments were conducted by blinded raters who neither participated in the medical management of patients nor had access to study assessment results or medical patient records; however, due to the subjective nature of these outcomes, potential reporting bias in favour of foslevodopa-foscarbidopa might be involved in patients with knowledge of treatment assignment. The same also applies to PD diary outcomes (i.e., “on” and “off” times, presence of morning akinesia) as they were based on patients’ self-report of symptoms. Reporting of harms could also have been biased, potentially in favour of the oral LD-CD arm due to a higher tendency to report harms among unblinded patients in the foslevodopa-foscarbidopa arm. However, the extent of these potential biases is unclear.

Statistical analyses were in general appropriate for the outcomes evaluated. The study was powered to detect a treatment difference in the primary end point between treatment arms and the enrolled sample size was adequate. The sponsor noted that the study was also powered for the key secondary end points; however, the assumptions to support this claim are unclear. A hierarchical testing procedure was appropriately used to account for multiplicity in the key secondary and other secondary outcomes. No conclusion can be drawn on the prespecified subgroup analyses because of the lack of consideration for sample size and statistical power, and control of multiplicity.

A notably higher proportion of patients in the foslevodopa-foscarbidopa arm (35.1%) than in the oral LD-CD arm (7.5%) discontinued from the study. Missing data for the primary and key continuous secondary outcomes were implicitly imputed using an MMRM, which relies on the missing-at-random assumption. This assumption is unlikely to hold as withdrawals in the foslevodopa-foscarbidopa arm were largely driven by AEs. Two prespecified sensitivity analyses were conducted for PD diary-based outcomes to assess the impact of missing data, and their results were consistent with the primary analysis. The analysis based on the “jump-to-reference” approach, which assumed missing values in the foslevodopa-foscarbidopa arm to be missing not at random and allowed missing values after treatment withdrawal to be imputed with values equal to those of the oral LD-CD arm, was considered conservative and had increased certainty of the primary analysis. The analysis using the last available value approach was less informative given that the method relies on the missing-at-random assumption, similar to the primary analysis. The impact of attrition on outcomes other than PD-based outcomes was unclear in the absence of sensitivity analyses.

The trial included outcomes based on PD diary, MDS-UPDRS, and PDQ-39, disease-specific PRO instruments commonly used to assess function and symptoms in clinical trials of PD. There is evidence supporting these instruments’ validity and reliability and their responsiveness in assessing patients with PD. MID estimates for these instruments have been established in patients with PD, although for most estimates (except “off” time assessed based on PD diary), it is unclear if they were derived from patients with advanced disease. Procedures were in place to ensure proper completion of PD diary entries, including trainings on PD diary completion, requiring patients to have at least 75% concordance with the investigators at screening, and retraining on PD diary completion available for patients who incorrectly completed the diary. All patients were also reminded to complete the diary before each visit to ensure adherence.

Protocol deviations related to study procedure adherence was frequently reported (45.9% in the foslevodopa-foscarbidopa arm, and 40.3% in the oral LD-CD arm), and included improper use of syringes for drug administration, incomplete PD diary, missed or late questionnaire completion, missed recording of rescue dose, remote completion of study visits, and incorrect loading doses. According to the sponsor, none of the recorded protocol deviations could have affected the study outcome or study results or conclusion interpretation. The CADTH review team noted that deviations related to PD diary and questionnaire in particular could be a reason for concern for bias; however, results of the sensitivity analyses on PD diary-based outcomes provided some reassurance that the impact of missing data on these outcomes is possibly low.

External Validity

The clinical expert commented that the inclusion and exclusion criteria in general align with the selection criteria for candidates for foslevodopa-foscarbidopa treatment, although patients with cognitive impairment and prior DBS or LCIG treatment would not necessarily be excluded from treatment in clinical practice. Nonetheless, the clinical expert did not expect the exclusion of these patients to significantly impact the generalizability of the patient population in this study. The baseline patient characteristics were also generally reflective of the Canadian patient population with advanced PD, as per the clinical expert.

The infusion pump used in the trial (i.e., Phillips-Medisize Parkinson’s Disease Subcutaneous Pump Delivery System) is the same as the pump intended for use in Canada (brand name Vyafuser for the same device). The clinical expert considered oral therapy to be an appropriate comparator given that most patients with advanced PD currently remain on oral medications. The use of concomitant PD medications was consistent with practice in Canada, as per the clinical expert.

The efficacy outcomes measured in the study were of clinical importance to patients and clinicians, including motor and nonmotor symptoms, functioning, and HRQoL. However, the clinical expert noted that the PKG device used in the study for measuring bradykinesia and dyskinesia scores is currently not available in Canada, and the EQ-5D-5L and PDSS-2 instruments are not used in clinical practice. The clinical expert considered the PD diary, MDS-UPDRS, and PDQ-39 to be clinically relevant as these instruments are sometimes used in clinical practice to assess treatment response. Patients highlighted an unmet need for treatments that can improve cognition. Although this outcome was captured in the MDS-UPDRS Part I (nonmotor aspects of experiences of daily living) scale, the results were pooled with other item scores of nonmotor symptoms, and thus the effect of foslevodopa-foscarbidopa on cognition alone is unclear. The clinical expert noted that a goal of treatment of advanced PD is to reduce caregiver burden; however, the impact on caregiver burden is unclear as it was not a measured outcome. The clinical expert noted that the duration of follow-up (12 weeks in the double-blind treatment period) was adequate for efficacy assessment although longer follow-up is required to gain certainty on the safety profile.

It should be noted that this trial is the only phase III RCT providing direct comparative evidence between foslevodopa-foscarbidopa and standard of care for advanced PD. The clinical expert consulted by CADTH noted that DBS and LCIG therapies are also available for the treatment of advanced PD, although there are unique considerations that guide treatment decisions (e.g., the need for a specialized medical team for DBS and LCIG therapies, provision of DBS at specialized centres only, and selection of patients without contraindications). In patients with advanced PD who are potential candidates for any of the available advanced therapies, the absence of head-to-head evidence between foslevodopa-foscarbidopa versus DBS and LCIG represents an evidence gap.

Long-Term Extension Studies

Description of M20-098 Study

Characteristics of the M20-098 trial are summarized in Table 19. M20-098³⁹ is an ongoing single-arm, multicentre, open-label extension study of the previously described M15-736 trial, with up to 96 weeks of treatment. The objective is to assess the long-term safety and tolerability of foslevodopa-foscarbidopa delivered by CSCI for 24 hours per day. The primary outcomes are AEs and related measures. Efficacy outcomes are also being collected as secondary end points. Patients in M20-098 were previously randomized to either CSCI or oral medications in the parent study M15-736, and as such, the M20-098 long-term extension population had 2 different treatment pathways. Limited follow-up was available for the M20-098 trial at the time of this review.

Table 19: Details of the M20-098 Trial

AE = adverse event; AESI = adverse event of special interest; CSCI = continuous subcutaneous infusion; ECG = electrocardiogram; EQ-5D-5L = EuroQol 5-dimensions questionnaire; HRQoL = health-related quality of life; MDS-UPDRS = Movement Disorder Society-Unified Parkinson’s Disease Rating Scale; MMSE = Mini-Mental State Examination; NA = not applicable; PD = Parkinson disease; PDQ-39 = Parkinson’s Disease Questionnaire-39 items; PDSS-2 = Parkinson’s Disease Sleep Scale-2; SAE = serious adverse event.

Source: M20-098 Clinical Study Report.³⁹ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Populations

M20-098 included patients with PD who, at completion of the parent study (M15-736), remained on the study drug and met additional criteria (consent, demographic and laboratory assessments, study participant history, and contraception). Patients had to be able to understand the nature of the study and have had the opportunity to have questions answered by the investigator. Patients with decision-making capacity were allowed to sign and date an informed consent form. Patients also had to be willing to comply with procedures required in this study and be considered by the investigator to be a suitable candidate to continue foslevodopa-foscarbidopa. Patients also could not be exhibiting significant suicidal behaviours at the time of the final visit of the parent study. Patients with childbearing potential must have a negative urine pregnancy test, and both female and male patients must have agreed to certain protocol-specified contraception. Patient characteristics at M20-098 study baseline are summarized in Table 20.

Table 20: Summary of Baseline Characteristics in the M20-098 Trial

CD = carbidopa; COMT = catechol-O-methyltransferase; DBS = deep brain stimulation; LD = levodopa; MAO-B = monoamine oxidase type B; MDS-UPDRS = Movement Disorder Society-Unified Parkinson’s Disease Rating Scale; PD = Parkinson disease; SD = standard deviation.

Source: M20-098 Clinical Study Report.³⁹ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Interventions

Patients enrolled in M20-098 received an individualized CSCI (24 hours per day) of foslevodopa-foscarbidopa via an infusion set connected to a pump for up to 96 weeks. Use of concomitant medications reflected the protocol of the parent study.

Outcomes

Efficacy variables were derived from the PD diary, the MDS-UPDRS, the PDSS-2, the PDQ-39, and the EQ-5D-5L and cognitive impairment was assessed using the Mini-Mental State Examination (MMSE). Safety evaluations were the primary objective of this study and included safety and tolerability as evaluated by AEs, the Infusion Site Evaluation Scale, the C-SSRS, and the QUIP-RS.

Statistical Analysis

In M20-098, the full analysis set consisted of all patients who received foslevodopa-foscarbidopa infusion and have a baseline and post-baseline observation for at least 1 efficacy end point and was used for all efficacy analyses. The safety analysis set consisted of all patients who received any foslevodopa-foscarbidopa infusion in the M20-098 study and was used for all safety analyses. Efficacy analyses were conducted by treatment sequence and were evaluated in groups: “oral LD-CD à foslevodopa-foscarbidopa,” “foslevodopa-foscarbidopa à foslevodopa-foscarbidopa,” and overall. Outcomes were summarized descriptively including the number of observations, mean, SD, minimum, median, and maximum; hypothesis testing was not performed. No adjustments were made for missing visit data. For PD diary entries, if no data were available for a visit, the average daily normalized “off” or “on” times values were missing for that visit. The MDS-UPDRS total score and score of each part was calculated as long as no more than 15% of the answers were missing for that assessment. The missing item was imputed as the average of the nonmissing items from the same MDS-UPDRS assessment.

Results

Patient Disposition

Patient disposition is summarized in Table 21. M20-098 was ongoing at the time of submission. No patients had completed the study at the time of the data cut submitted by the sponsors.

The majority of patients who completed the parent study went on to enrol in the long term extension study. Notably more patients discontinued from foslevodopa-foscarbidopa in the M20-098 long-term extension after having received oral LD-CD in the parent study, when compared to patients who were already receiving foslevodopa-foscarbidopa in the parent study.

Exposure to Study Treatments

Prescribed dose by infusion rate and total daily dose, as well as the rates of adherence, are presented in Table 22. The study is ongoing. There were no signals of substantial imbalance between study arms in M20-098. The rates of adherence were high at the time of the available data cut-off.

Table 21: Patient Disposition in the M20-098 Trial

AE = adverse event; CD = carbidopa; FAS = full analysis set; LD = levodopa; LTE = long-term extension; NR = not reported; SAS = safety analysis set.

Source: M20-098 Clinical Study Report.³⁹ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Table 22: Summary of Patient Exposure in the M20-098 Trial

CD = carbidopa; IQR = interquartile range; LD = levodopa; SD = standard deviation.

^aEnd of optimization phase.

^bTotal dose calculated as “base” (mL/hour) × 24 hours per day × 170.75 mg LD per mL.

^cStudy drug solution adherence: (hours of infusion / 24) × 100.

Source: M20-098 Clinical Study Report.³⁹ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Efficacy

At the time of this review, M20-098 is still under way and efficacy results were limited. Each outcome is summarized briefly, but no conclusions can be drawn.

Mean Average Daily Normalized “On” Time Without Troublesome Dyskinesia, “On” Time With Troublesome Dyskinesia, and “Off” Time Based on the PD Diary

||||| ||| |||||| |||||| |||| ||| ||||||| ||||| |||||||| || ||||| |||| ||| || |||| ||||| ||| ||| || |||||||| |||| || ||||| ||||||||||| || |||| | || ||| ||||| ||| ||| |||||||| |||||||||||||||||||||||| || ||| |||||| |||||| |||||||||||| || ||||| |||| ||| || |||| ||||| |||| |||||||| ||| ||| || |||||||| |||| || ||||| ||||||||||| || |||| | || ||| ||||| ||| ||| |||||||| |||| ||||| || ||| |||||| ||||| ||| |||||||||||| || |||||||||||||||||||||||| || ||| ||||||||| |||||| ||| |||||||||| ||||| |||| |||| || ||||| |||| || || ||||| |||||||| |||| ||||||||| |||| || ||||||| || ||| |||| || |||| |||||||

Movement Disorder Society-Unified Parkinson Disease Rating Scale

||| ||||||||| |||| || |||||| || |||| | |||||| |||||| |||||||||| || |||||||| ||| ||| |||||||| |||||||||||||||||||||||| || ||| |||||| |||||| ||| |||| ||||||||||| || |||||||| ||| ||| |||||||| |||| ||||| || ||| |||||| |||||| |||| |||||| ||||||||| |||||| |||| | ||| ||||||||| || ||| |||| || |||| |||||||

Change From Baseline to Each Planned Visit in HRQoL (EQ-5D-5L)

EQ-5D-5L results for the M20-098 trial were not yet available at the time of this review.

Change in Baseline to Each Planned Visit in Cognitive Impairment Measured Using the MMSE

MMSE results for the M20-098 trial were not yet available at the time of this review.

Harms

All AEs presented were treatment emergent and analyzed in the safety analysis set. || ||| ||||||| |||||| ||| |||| |||||||| ||| || |||||||| ||||||| || ||| |||| || ||| |||| |||||||| ||| ||| |||||||| |||| |||||||||| ||| |||| || |||||||| || ||||||||| |||| |||| |||||||| ||| | |||||||| ||||||| || |||||||| ||| |||||

Critical Appraisal

Internal Validity

The main limitation is the absence of comparator arm and the potential biases related to the nonrandomized design. The open-label study design inherently increases risk of bias in the reporting of any subjective outcomes, including harms outcomes, because patients and their clinicians are aware of the treatment received. In addition, because patients could only enrol after completing the parent trial, there is a risk of selection bias given that patients who better tolerated the treatment or perceived the treatment as benefiting them were more likely to enrol in the extension studies.

External Validity

As the extension study consisted of patients who took part in the parent study, it is reasonable to expect that the same strengths and limitations related to generalizability apply to the extension studies, with the additional caveat of potential selection bias due to the nature of the open-label extension study design.

However, little follow-up was available for M20-098 at the time of this review, preventing interpretation of long-term safety or efficacy outcomes.

Indirect Evidence

Objective for the Summary of Indirect Evidence

One ITC comparing foslevodopa-foscarbidopa with alternative treatments for advanced PD was submitted by the sponsor and is included in this report. An ITC was required to address a gap in pivotal and RCT evidence in the absence of studies directly comparing foslevodopa-foscarbidopa with treatments for advanced PD other than oral medications.

Description of Sponsor-Submitted ITC

The sponsor first conducted an SLR to identify evidence for inclusion in the ITC. The relative efficacy of foslevodopa-foscarbidopa in the M15-736 trial was indirectly compared with other treatments for patients with PD who are inadequately controlled with current therapy, via Bayesian NMA. Comparators of interest for the sponsor-submitted NMA included LCIG and BMT (including oral therapy). Outcomes of interest included “off” time, “on” time without troublesome dyskinesia, and sleep symptoms assessed using the PDSS-2 total score.

Methods of Sponsor-Submitted ITC

Objectives

The objective of this study was to estimate the comparative efficacy of foslevodopa-foscarbidopa versus LCIG and BMT, both of which are relevant comparators for foslevodopa-foscarbidopa in Canada, using an NMA.

Study Selection Methods

A SLR was conducted using MEDLINE (In-Process), Embase, and Cochrane Library to identify clinical and observational studies published for all time till the date of search (i.e., June 1, 2021). Two updated searches were conducted in January and June 2022.

Article screening was performed independently by 2 reviewers in 2 stages (titles and abstracts, and then the full texts), with a third reviewer resolving disagreements. Reasons for exclusion were documented. Data extraction was performed by 1 reviewer, with quality check on all data by a second reviewer, and an additional check on 10% of data by a third reviewer. Quality assessment of the selected studies, with the exception of foslevodopa-foscarbidopa clinical studies (not published when the SLR was being conducted), was conducted by 1 reviewer to assess the risk of bias using the Cochrane risk of bias tool.

The initial article search included a broader set of comparators and study designs. More restricted selection criteria, as shown in Table 23, were subsequently applied at the full-text review stage to select trial for inclusion into the NMA.

Table 23: Study Selection Criteria and Methods in ITCs Submitted by the Sponsor

ITC = indirect treatment comparison; LCIG = levodopa-carbidopa intestinal gel; PD = Parkinson disease; PDSS-2 = Parkinson’s Disease Sleep Scale-2; RCT = randomized controlled trial.

Source: Sponsor-submitted Network Meta-Analysis Technical Report.⁷⁰ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

ITC Analysis Methods

A summary of the analysis methods for the NMA is shown in Table 24. All analyses were conducted using a Bayesian analysis framework. Fixed effects (FE) and random effects (RE) models were evaluated for the base-case analyses, and selection was determined based on the deviance information criterion (DIC) model fit statistic. Vague prior distributions on model parameters were used so that model outcomes would be determined primarily by the clinical trial data. These were selected using the recommended priors in the NICE DSU Technical Support Document 2.⁷¹ Posterior outcome distributions were based on at least 50,000 simulations after a burn-in of at least 50,000. Adequate convergence was assessed by visual inspection of the autocorrelation and history plots.

Statistical heterogeneity was evaluated based on the I² statistic derived from direct head-to-head meta-analysis of those treatment comparisons in each network that were reported by more than 1 study.

“On” time without troublesome dyskinesia and “off” time measures are continuous data based on sample means and standard errors that were analyzed using the normal likelihood and the normal link function.⁷¹ For studies where measures of uncertainty were not reported, the SD was input using the sample size-weighted average of the available SDs. In this analysis, the majority of studies across all treatments of interest reported a mean change from baseline for the outcomes and not the LSM change. As a result, a decision was made to use mean change as a unit of measure. The mean difference of “on” time without troublesome dyskinesia and “off” time measures for each treatment versus oral standard of care was the NMA output.

The base-case analysis included all 3 comparators for both outcomes: “off” time and “on” time without troublesome dyskinesia. The base-case analysis for “off” time without troublesome dyskinesia included all RCTs. For both of these outcomes, FE and RE models were fitted and compared on DIC to determine the better fitting model (lower DIC values indicate better fit to the data). When DIC differences are small (i.e., less than 3 to 5 points) across different fitted models, common practice is to choose the simplest model because the additional complexity does not result in a better fitted model.⁷¹ In addition, while there may be cross-trial heterogeneities in treatment contrasts, the small number of trials made it infeasible to quantitatively examine the level of such heterogeneities. As such, FE models were selected.

An NMA was also run for the outcome PDSS-2 total score. Only 2 RCTs were considered appropriate for potential inclusion into this analysis, and therefore an FE model was fitted. Each treatment was considered a separate node in the evidence network. All analyses assessed the change from baseline at week 12.

No sensitivity or subgroup analysis was conducted.

Assessment for consistency was not possible due to the lack of closed loops.

Table 24: Sponsor-Submitted NMA Analysis Methods

DIC = deviance information criterion; NMA = network meta-analysis; PDSS-2 = Parkinson’s Disease Sleep Scale-2.

Source: Sponsor-submitted Network Meta-Analysis Technical Report.⁷⁰ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Results of Sponsor-Submitted NMA

Summary of Included Studies

Five total studies were considered for inclusion in the NMA, including 2 foslevodopa-foscarbidopa studies: M15-736 and its extension M20-098. M20-098 could not be included because it was a single-arm trial where all patients received foslevodopa-foscarbidopa. The remaining 3 trials assessed the comparative efficacy between LCIG and oral therapies: Olanow et al. (2014),⁷² DYSCOVER,⁷³ and INSIGHTS.⁷⁴

An overview of the assessment of homogeneity of the 4 studies included in the ITC is presented in Table 25. All trials were phase III RCTs; M15-736 and Olanow et al. (2014) were double-blind, while DYSCOVER and INSIGHTS were open-label. All studies were multicentre trials, with sample sizes of between 61 and 141 patients. Year of study was from 2009 to 2022.

All trials included patients with advanced PD with motor fluctuations that were not controlled with optimized oral therapy, although for other inclusion criteria, including exclusion of patients with prior DBS treatment, inclusion of patients with dyskinesia, minimum daily levodopa-equivalent dose requirement, they were only included in some trials. |||| ||| || |||||||| || | ||||| ||||||||| ||| |||||| |||| ||||| ||| | ||||| ||||| ||| || ||||| ||| | ||||| ||||| |||| |||||||||| || ||||||| |||||| ||| ||||| ||||||||| |||| |||||| |||| ||| || |||| Reporting of certain baseline disease characteristics was not consistent across trials and there were variabilities with respect to the duration of PD diagnosis (reported in 3 trials, |||||| |||| |||| ||||| || ||||| |||||| |||| ||||| ||||| |||| ||||||||| || | ||||||| |||||| |||| |||| || |||| ||||| ||| ||||| ||| |||||| |||||| ||||||||| || | |||||||| |||||| ||||||| ||||| || ||||| ||||||), across studies. All treatments were initially dose based on the total levodopa dose established before the studies. Titration to clinical response was allowed for foslevodopa-foscarbidopa and LCIG across trials but was not consistently allowed for oral therapies. No notable differences in the definitions of end points were identified between trials. End points were assessed up to week 12 in all trials, except for INSIGHTS, which measured results up to week 26. The incidence of study withdrawal |||||| ||||||| ||| | || ||||||

The sponsor did not consider the between-trial differences to be clinically meaningful or that they could affect the comparability of the studies’ treatment effect within the NMA, and thus, no adjustments for heterogeneity were made.

Results of the quality assessment of included studies were not reported.

Table 25: Assessment of Homogeneity for Sponsor-Submitted NMA

LCIG = levodopa-carbidopa intestinal gel; NMA = network meta-analysis; PD = Parkinson disease; PDSS-2 = Parkinson’s Disease Sleep Scale-2; RCT = randomized controlled trial.

Sources: Sponsor-submitted Network Meta-Analysis Technical Report,⁷⁰ Olanow et al. (2014),⁷² DYSCOVER,⁷³ and INSIGHT.⁷⁴ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Results

Three studies were included in the “off” time and “on” time without troublesome dyskinesia networks (Figure 1). Two studies (M15-736, comparing foslevodopa-foscarbidopa versus BMT; INSIGHTS, comparing LCIG versus BMT) were included in the PDSS-2 network. The FE model was selected as the base case for “off” and “on” time without troublesome dyskinesia outcomes, based on the lowest reported DIC ||| |||||| ||||| |||||| ||| || |||||| ||||||. The FE model was selected as the base case for PDSS-2 analysis since only 2 studies were included. Results of the FE model and RE model are summarized in Table 26.

Figure 1: Network Diagram of Studies Included in “Off” Time and “On” Time Without Troublesome Dyskinesia Analyses in Sponsor-Submitted NMA

BMT = best medical therapy; LCIG = levodopa-carbidopa intestinal gel; NMA = network meta-analysis.

Note: This figure has been redacted at the request of the sponsor.

Source: Sponsor-submitted Network Meta-Analysis Technical Report.⁷⁰ (Note: Details have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

“Off” Time

For change from baseline in “off” time at week 12, |||||||||||||||||||||||| ||| |||||||| |||| ||| ||||| ||||||||||| ||||| |||| |||| ||||| || ||||||| in the base-case analysis (FE model) but there was insufficient evidence to show a difference in the RE model and the 95% credible interval (CrI) was wide. The evidence was also insufficient to show a difference between LCIG versus foslevodopa-foscarbidopa ||||| ||||||||||| ||| |||| |||| ||||| || ||||||, with a wide 95% CrI in the base-case FE analysis (RE model not reported).

For the pairwise comparisons informed by at least 2 studies (i.e., LCIG and BMT), the I² statistic was 0%.

“On” Time Without Troublesome Dyskinesia

For change from baseline in “on” time without troublesome dyskinesia at week 12, |||||||||||||||||||||||| ||| |||||||| |||| ||| ||||| ||||||||||| |||| |||| |||| |||| || |||||| in the base-case analysis (FE model) but there was insufficient evidence to show a difference in the RE model and the 95% CrI was wide. The evidence was also insufficient to show a difference between LCIG versus foslevodopa-foscarbidopa ||||| ||||||||||| |||| |||| |||| ||||| || ||||||, with a wide 95% CrI in the base-case FE analysis (RE model not reported).

The I² statistic was 0% for the pairwise comparisons between LCIG and BMT.

Parkinson’s Disease Sleep Scale-2

For change from baseline in PDSS-2 total score at week 12, |||||||||||||||||||||||| ||| |||||||| |||| ||| (foslevodopa-foscarbidopa versus BMT | |||| ||||||||||| ||||| |||| |||| ||||| || |||||) and LCIG (LCIG versus foslevodopa-foscarbidopa: |||| ||||||||||| |||| |||||||| |||| || |||||).

Table 26: Outcomes of Sponsor-Submitted NMA

BMT = best medical therapy; CrI = credible interval; FE = fixed effects; LCIG = levodopa-carbidopa intestinal gel; NMA = network meta-analysis; NR = not reported; PDSS-2 = Parkinson’s Disease Sleep Scale-2; RE = random effects; vs. = versus.

^aThis represents a comparison of treatment effects between foslevodopa-foscarbidopa and BMT. A mean difference (95% CrI) of less than 0 indicates results in favour of foslevodopa-foscarbidopa.

^bThis represents a comparison of treatment effects between LCIG and foslevodopa-foscarbidopa. A mean difference (95% CrI) of more than 0 indicates results in favour of foslevodopa-foscarbidopa.

Source: Sponsor-submitted Network Meta-Analysis Technical Report.⁷⁰ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Critical Appraisal of Sponsor-Submitted NMA

The sponsor-submitted NMA was informed by an SLR using the same methods as described in the Pivotal Studies and RCT Evidence section, and as such, much of the appraisal of the SLR noted previously with respect to the literature search, study selection, and data extraction methods, and risk of bias assessment applies to the SLR informing this NMA. Narrower selection criteria were applied to further distill studies and tailor to the objective of the NMA, although there is no mention of whether the NMA-specific inclusion criteria were determined a priori. Studies that assessed DBS, a relevant comparator of foslevodopa-foscarbidopa as per the clinical expert, was not eligible for inclusion into the NMA. As such, the comparative efficacy of foslevodopa-foscarbidopa and DBS remains unknown. The included outcomes, “on” time without troublesome dyskinesia and “off” time in particular, were relevant to patients and clinicians, but less so for PDSS-2 given that the instrument is not routinely assessed in clinical practice. Other outcomes that are of interest to this review, including MDS-UPDRS and PDQ-39 scores, were not assessed.

The CADTH review team’s assessment of the degree of heterogeneity between the included studies was complicated by the limited reporting of study design and patient characteristics in the sponsor-provided technical report. Aside from description of interventions, sample size, country, duration of follow-up, there is no reporting of other trial characteristics (e.g., the inclusion and exclusion criteria, treatment withdrawal frequency, handling of missing data, and end point definition, and so on) within the report. The sponsor concluded that there was no clinically meaningful difference that may affect the comparability of the studies’ treatment effect within the NMA based on a comparison of select baseline characteristics (i.e., age, duration of PD, baseline “off” time and “on” time without troublesome dyskinesia) between studies; however, the CADTH review team considered there to be notable between-trial differences with respect to duration of PD diagnosis and “off” time. Considering that these factors are potential treatment effect modifiers as per the clinical expert, it is possible that the heterogeneity in these baseline characteristics could result in changing relative treatment effects. No adjustment was made in the modelling of the NMA to account for these differences. Upon reviewing each included trial, the CADTH review team also noted some differences in eligibility criteria across trials. While all trials included patients with advanced PD who have motor fluctuations not controlled on optimized oral therapy, some trials excluded patients with cognitive impairment (M15-736, DYSCOVER), required enrolled patients to have dyskinesia (DYSCOVER), and a minimum duration of “off” per day, assessed using the PD diary (M15-736, Olanow et al., 2014). As well, there were differences in study design with respect to blinding, where 2 studies were open-label and 2 were double-blind. Differences in dosing protocol for oral therapies, follow-up duration, and treatment withdrawal frequency were also noted. Although these difference in study design could not be adjusted for in the analysis, these inherent differences suggests that transitivity assumption of the NMA likely have not been met.

Heterogeneity was statistically assessed based on I² values. Although the analyses suggest low heterogeneity (I² = 0) between included studies for “on” time without troublesome dyskinesia and “off” time outcomes, given the small number of studies in the network, the I² is subject to bias.

The NMA was informed by 4 studies; 3 were included in “on” time without troublesome dyskinesia and “off” time analyses; 2 were included in the PDSS-2 analysis. The resulting networks were sparse with no closed loop; as such, it was not possible to assess consistency of results between direct and indirect comparisons. Results of the NMA were informed solely by indirect evidence, and thus are associated with increased uncertainty.

The NMA was conducted using a Bayesian framework. Fixed-effect models were chosen for the base-case analyses for all assessed outcomes. The decision for “on” time without troublesome dyskinesia and “off” time analyses was based on a lower DIC value, while the decision for the PDSS-2 analysis was determined a priori considering the small number of studies. The fixed-effect model relies on the assumption that there was no between-trial heterogeneity, which unlikely holds true as previously discussed. In the absence of sensitivity analyses that assess the potential impact of trial heterogeneity, there is considerable uncertainty in the relative treatment effect estimates.

Studies Addressing Gaps in the Pivotal and RCT Evidence

The M15-741 trial,³⁷ which Health Canada considers to be a supportive trial for the approval of foslevodopa-foscarbidopa, and its ongoing long-term extension study, M15-737,⁷⁵ are summarized in this section. The aim of these studies was to assess the long-term safety and efficacy of foslevodopa-foscarbidopa for up to 148 weeks.

Description of the M15-741 Trial

Characteristics of the M15-741 trial are summarized in Table 27. The M15-741 trial was a phase III, open-label, single-arm trial that aimed to evaluate the safety and tolerability of foslevodopa-foscarbidopa in patients with advanced PD for up to 52 weeks of treatment (N = 244). The study was conducted at 60 sites in 13 countries (including 2 sites in Canada). Patients were enrolled between April 29, 2019, and November 5, 2021. The study is now complete. It consisted of a screening period of 10 days to 42 days to assess study eligibility, followed by a 52-week treatment period (day 1 to week 52). In the first 4 weeks (optimization phase), the dose of foslevodopa-foscarbidopa was adjusted to achieve optimal clinical response, and this dose remained unchanged for the remainder of the study (maintenance phase).

Table 27: Details of the M15-741 Trial

AE = adverse event; AESI = adverse event of special interest; CSCI = continuous subcutaneous infusion; ECG = electrocardiogram; EQ-5D-5L = 5-Level EQ-5D; HRQoL = health-related quality of life; IQR = interquartile range; MDS-UPDRS = Movement Disorder Society-Unified Parkinson’s Disease Rating Scale; NA = not applicable; PD = Parkinson disease; PDQ-39 = Parkinson’s Disease Questionnaire-39 items; PDSS-2 = Parkinson’s Disease Sleep Scale-2; SAE = serious adverse event.

^aAs per investigator’s assessment.

^bPatients who had received deep brain stimulation were eligible provided they were stable and still levodopa-responsive.

^cAssessed using the PD diary.

^dDefined by a Mini-Mental State Examination score of at least 24.

^eAssessed using the Parkinson’s KinetiGraph/Personal KinetiGraph wearable device.

Source: M15-741 Clinical Study Report.³⁷ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Populations

The key inclusion criteria were the same as those in the pivotal M15-736 study, except that there was no requirement for minimum daily levodopa-equivalent dose, the criterion for “off” time was based solely on a daily minimum of 2.5 hours each day, and patients who had received DBS therapy were eligible for inclusion if they were stable and still levodopa-responsive.

Interventions

Optimization Phase (Day 1 to Week 4)

Patients received 1 loading dose of IR oral LD tablet in combination with dopa decarboxylase inhibitor (e.g., CD or benserazide), followed by an CSCI (24 hours per day) of foslevodopa (240 mg/mL) and foscarbidopa (12 mg/mL) at an initial infusion rate calculated based on the patient’s oral LD therapy over the 16-hour period prior using the same conversion method, based on levodopa-equivalent dose, as described in the M15-736 trial. The infusion rate was adjusted at the investigator’s discretion to optimize clinical response. The allowable infusion rate ranged from 0.17 to 1.04 mL per hour. The Crono PAR Series III pump was used to deliver the infusion solution. Adjustments of concomitant PD medications were allowed.

Maintenance Phase (Week 5 to Week 52)

Patients continued CSCI of foslevodopa-foscarbidopa at the optimal therapeutic dose and dose adjustments were allowed. Patients maintained a stable regimen of all concomitant PD medications unless changes were considered medically necessary, as per the investigator.

The list of allowed concomitant PD medications was the same as the list in the M15-736 trial. Oral LD (100 mg) in combination with dopa decarboxylase inhibitor, as well as levodopa inhalation powder (84 mg) were allowed as rescue medications. The prohibited medication list was almost identical to that in the M15-736 trial, with the exception that second-generation antipsychotics were included on the prohibited medication list, while LCIG and LD-CD enteral suspension were not.

Outcomes

A list of efficacy end points assessed in this clinical review report are provided in Table 28. These end points are further summarized below. Summarized end points are based on those included in the sponsor’s Summary of Clinical Evidence¹⁵ as well as any identified as important to this review according to stakeholders, that is, the clinical expert consulted by CADTH, clinician groups, or patient groups.

Table 28: Summary of Outcomes of the M15-741 Trial

AE = adverse event; AESI = adverse event of special interest; BK = bradykinesia score; DK = dyskinesia; ECG = electrocardiogram; EQ-5D-5L = 5-Level EQ-5D; HRQoL = health-related quality of life; IQR = interquartile range; MDS-UPDRS = Movement Disorder Society-Unified Parkinson’s Disease Rating Scale; PD = Parkinson disease; PDQ-39 = Parkinson’s Disease Questionnaire-39 items; PDSS-2 = Parkinson’s Disease Sleep Scale-2; QoL = quality of life; SAE = serious adverse event.

^aThis end point was measured as the proportion of patients with numeric grade ≥ 5 and letter grade ≥ D on the Infusion Site Evaluation Scale in the M15-741 study.

^bThis end point was measured as change from baseline to end of study.

Source: M15-741 Clinical Study Report.³⁷ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Efficacy Outcomes

Efficacy end points were secondary in this study; they included the change from baseline to the end of study (week 52) for average normalized daily “off” and “on” times, assessed based on the PD diary; PD symptoms, assessed using the MDS-UPDRS Parts I to IV; sleep symptoms, assessed using the PDSS-2 total score; quality of life, assessed using the PDQ-39; and HRQoL, assessed based on the EQ-5D-5L. The measurement scales were the same as in the M15-736 trial.

Baseline was defined as the last nonmissing assessment before the initiation of foslevodopa-foscarbidopa. Study visits occurred at days 1 and 2 and weeks 1, 2, 3, 4, 6, 13, 26, 39, and 52. All end points were assessed on day 1 and all visits in the maintenance phase. Additional visits were scheduled in the optimization phase for assessing PD diary (week 1) and MDS-UPDRS (all visits) end points.

Harms Outcomes

The harms outcomes were primary end points of this study and included the proportion of patients with AEs, SAEs, adverse event of special interest, numeric grade equal to or higher than 5 and letter grade equal to or higher than D on the Infusion Site Evaluation Scale, and change in clinical laboratory test, vital signs, and ECGs from baseline to end of study.

Statistical Analysis

A sample size calculation determined that approximately 240 enrolled patients were required to obtain exposure data from at least 100 patients treated with foslevodopa-foscarbidopa for at least 12 months and to meet country requirements. With 240 patients receiving foslevodopa-foscarbidopa, the probability of observing an AE with an annual incidence rate of 0.005, 0.01, and 0.02 was 70%, 91%, and 99%, respectively.

All efficacy end points were presented using descriptive statistics. A 2-sided paired-sample t test was performed to assess within-group changes from baseline. No adjustments for multiple statistical testing were performed. The handling of missing items in efficacy instrument was the same as the M15-741 trial but no imputation was performed for missing visit data. Predefined subgroup analyses were performed for all efficacy analyses, with the same subgroup categories as in the M15-736 trial. A sensitivity analysis was conducted for all “on” and “off” times outcomes by modifying the definition of valid PD diary day as a PD diary recording day with no more than 2 hours of missing data (≤ 4 missing entries) for the entire 24-hour diary.

Safety outcomes were compared between treatment arms using descriptive statistics; no hypothesis testing was performed.

Analysis Populations

Analysis populations of the M15-741 are summarized in Table 29.

Table 29: Analysis Populations in the M15-741 Trial

FAS = full analysis set; SAS = safety analysis set.

Note: The M15-741 trial also included a treatment-naive analysis set but it was not of interest to this review and was not summarized.

Source: M15-741 Clinical Study Report.³⁷ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Results

Patient Disposition

As summarized in Table 30, ||| |||||||| |||| |||||||| ||| || ||||||| were screen failures. Both the full and the safety analysis sets included 244 patients who entered the 52-week treatment period. Study treatment discontinuation occurred in ||| ||||||| patients. The most common reason for treatment discontinuation was AEs |||||||, similar to the M15-736 trial.

Table 30: Summary of Patient Disposition in the M15-741 Trial (SAS)

AE = adverse event; CD = carbidopa; FAS = full analysis set; LD = levodopa; NA = not applicable; SAS = safety analysis set.

Source: M15-741 Clinical Study Report.³⁷ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Baseline Characteristics

A summary of baseline patient demographics, disease characteristics, and medication history in the M15-741 trial is shown in Table 31. The baseline characteristics outlined are limited to those which are most relevant to this review, or were felt to impact the outcomes or interpretation of the study results.

Patients in the M15-741 trial, compared with those in the M15-736 trial, had a longer mean duration of PD since diagnosis (|||||||| | |||| ||||| || ||| ||| | |||| ||||||, a higher proportion of patients with history of levodopa-induced dyskinesia |||||| || |||||| and of levodopa response of more than 5 years |||||| || ||||||. In addition, the trial had higher proportions of patients with prior DBS procedure ||||| || ||||| and prior oral PD medication other than dopa preparations and higher proportions of patients who had received at least 3 PD medication classes at baseline |||||| || ||||||.

Table 31: Summary of Baseline Characteristics in the M15-741 Trial (SAS)

CD = carbidopa; COMT = catechol-O-methyltransferase; DBS = deep brain stimulation; IR = immediate release; LD = levodopa; MAO-B = monoamine oxidase type B; MDS-UPDRS = Movement Disorder Society-Unified Parkinson’s Disease Rating Scale; NR = not reported; PD = Parkinson disease; SAS = safety analysis set; SD = standard deviation.

^aBased on the PD diary.

^bRefers to PD medications other than oral levodopa-carbidopa immediate release.

Source: M15-741 Clinical Study Report.³⁷ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Exposure to Study Treatments

Treatment exposure in the M15-741 trial is summarized in Table 32. Mean average total daily levodopa dose was not reported. Mean average levodopa dose (presented by week) ranged from 1,641 mg/day (SD = 667 mg/day) to 1,878 mg/day (SD = 723 mg/day). Mean duration of study drug exposure was ||||| ||| | |||||| days. Mean treatment adherence for drug solution was ||||| ||| | |||||.

Table 32: Treatment Exposure in the M15-741 Trial (SAS)

SAS = safety analysis set; SD = standard deviation.

^aStudy drug solution adherence (%) = hours of infusion divided by 24, then multiplied by 100.

Source: M15-741 Clinical Study Report.³⁷ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Concomitant Medications

Concomitant medication use in the M15-741 trial is summarized in Table 33. In the trial, ||||| of patients receive concomitant PD medications. The most common concomitant PD medication classes (in at least 30% of patients) were |||||||| ||||||||| |||| ||| |||| |||||||||||| ||| ||||| ||||||||||.

Table 33: Concomitant PD Treatment in the M15-741 Trial (SAS)

MAO-B = monoamine oxidase type B; PD = Parkinson disease; SAS = safety analysis set.

^aNote that patients could have received more than 1 concomitant medication during the treatment period.

Source: M15-741 Clinical Study Report.³⁷ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Rescue Medications

Mean total daily levodopa dose from rescue medications (by week) ranged between ||||| ||| | ||||| || ||| ||| || ||||| ||| | ||||| || ||| |||.

Protocol Deviations

As summarized in Table 34, protocol deviations were most commonly related to enrolment of patients who did not satisfy the inclusion or exclusion criteria ||||||, and study procedure compliance ||||||. The sponsor noted that none of the protocol deviations recorded could have affected the study outcome or interpretation of the study results or conclusions.

Table 34: Protocol Deviations in the M15-741 Trial

^aTwenty-three patients did not fully meet eligibility criteria: 5 patients had a low vitamin B12 levels; 7 patients did not have a recognizable/identifiable “off” and “on” state (motor fluctuations) confirmed by the concordance test performed during the screening period, at baseline; 4 patients did not have any valid Parkinson disease (PD) diaries; and 10 patients had PD diaries but at least 1 PD diary day contained < 2.5 hours of “off” time.

Source: M15-741 Clinical Study Report.³⁷ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Efficacy

Unless otherwise specified, the key efficacy results of the M15-741 trial are summarized in Table 35.

“On” Time Outcomes

“On” Time Without Troublesome Dyskinesia

Change from baseline to week 52 in the average daily normalized “on” time without troublesome dyskinesia was a secondary end point and was not adjusted for multiplicity. The mean change from baseline at week 52 was 3.58 (SD = 3.03) hours. Results of the sensitivity analysis were consistent with the primary analysis (Table 46).

“On” Time Without Dyskinesia

|||||| |||| |||||||| || |||| || || ||| ||||||| ||||| |||||||||| |||| |||| ||||||| |||||||||| ||| | ||||||||| |||||||| ||| ||| ||| |||||||| ||| ||||||||||||| ||| |||| |||||| |||| |||||||| || |||| || ||| |||| ||| | ||||| ||||||

“Off” time

Change from baseline to week 52 in the average daily normalized “off” time was a secondary end point and was not adjusted for multiplicity. The mean change from baseline at week 52 was –3.39 (SD = 3.08) hours. Results of the sensitivity analysis were consistent with the primary analysis (Table 46).

Health-Related Quality of Life

Parkinson’s Disease Questionnaire-39 Items

Change from baseline to week 52 in PDQ-39 summary index total score was a secondary end point and was not adjusted for multiplicity. ||| |||| |||||| |||| |||||||| || |||| || ||| |||| ||| | |||||||

5-Level EQ-5D

Change from baseline to week 52 in EQ-5D-5L summary index score was a secondary end point and was not adjusted for multiplicity. Results are summarized in Table 49 in Appendix 1. ||| |||| |||||| |||| |||||||| || |||| || ||| ||||| ||| | |||||||

PD Symptoms

Changes from baseline to week 52 in MDS-UPDRS Part I, Part II, Part III, and Part IV were secondary end points and were not adjusted for multiplicity. ||| |||| |||||| |||| |||||||| || |||| || || |||| || ||| |||| ||| || |||||| |||| |||| ||| | |||||| |||| ||| | |||||| ||| ||| | ||||||| ||| |||| ||| | |||||| |||||||||||||

Bradykinesia and Dyskinesia Scores Assessed Using the PKG

Change from baseline to week 52 in median bradykinesia, bradykinesia IQR, median dyskinesia, and dyskinesia IQR scores were exploratory end points and were not adjusted for multiplicity. Results of these analyses are summarized in Table 49. ||| |||| |||||| |||| |||||||| || |||| || || ||||| |||||||| ||||| ||| ||||||| |||||| ||| ||| | |||||| |||| ||| | |||||| ||| ||| | |||||| ||| ||| ||| | ||||||| ||||||||||||| Results are summarized in Table 49 in Appendix 1.

PDSS-2 (Sleep Symptoms)

Change from baseline to week 52 in PDSS-2 total score was a secondary end point and was not adjusted for multiplicity. ||| |||| |||||| |||| |||||||| || |||| || ||| |||| ||| | |||||||

Table 35: Summary of Key Efficacy Results From the M15-741 Trial (FAS)

CI = confidence interval; FAS = full analysis set; FOS-FOS = foslevodopa-foscarbidopa; MDS-UPDRS = Movement Disorder Society-Unified Parkinson Disease Rating Scale; NR = not reported; PD = Parkinson disease; PDQ-39 = Parkinson Disease Questionnaire-39 items; PDSS-2 = Parkinson Disease Sleep Scale-2; QoL = quality of life; SD = standard deviation.

^aTwo-sided paired sample t test. End point was not adjusted for multiplicity.

Source: M15-741 Clinical Study Report.³⁷ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Harms

A summary of harms in the M15-741 trial are shown in Table 36. Results of the M15-741 trial referred to the findings in the entire study period (CSCI optimization phase and maintenance phase combined).

|||| ||| |||||||| || ||||| || ||||||||| || |||||||| || ||| |||||| ||||| |||||||| || ||| ||||||| ||||| |||||| |||||||| |||| ||||||||| ||||| |||||||||| ||| ||||||| |||||||| |||| ||||||| |||| ||| ||||||||||||| |||| |||| |||||||| || || ||||| ||| || |||||||| || ||| ||||||| ||||||

Serious Adverse Events

||||||| |||| ||| |||||||| || |||| | || ||||||||| |||| ||| |||| |||||| ||| ||||| |||||||| |||| |||||||||| ||||||| |||||||| || |||||||| |||| ||||||| ||||||| ||||||||||||| ||||||| || ||||||| ||| ||||||||| |||||||| |||||||

Withdrawals Due to Adverse Events

||||||||| ||||||||||||||| ||| || |||| ||| |||||||| || ||||| || ||||||||| |||| |||||||| ||||||| || ||||||||||||| ||||||| |||||||| |||| |||||||| ||| |||||||||| ||||| |||||| ||| |||||||||| |||||||

Mortality

||||| |||||| |||||||| ||| || ||||| ||||||||||||||||||| ||||||| |||||||| |||| |||||| ||| |||||||| ||||||||| ||| ||||||||||||||| |||||||||| ||| |||||| |||||||| |||||| || |||| ||||| ||| |||| |||||||||||||||||||||||| |||||||| ||| || ||||||||| ||||||||| ||||| ||||||||||| |||||||| ||| |||||||||| |||| || ||| |||||| ||| |||||||||| || || ||||||| || ||||| |||| || ||| ||| ||||||||

Notable Harms

Infusion-Site Reactions and Infusion-Site Infections

|||||||| |||| ||||||||| ||| |||||||| |||| |||||||||| |||| |||||||| || ||||| ||| ||||| || ||||||||| ||||||||||||| |||||||||||| |||||||| ||| || ||||| | ||||||||| || ||||||| ||||| || || ||||| | ||| | |||||| ||||| || || ||||| | || ||| |||||||| |||| |||||||||| ||||||

Hallucination or Psychosis, Impulse-Control Disorder, and Suicidality

||||||||||||| || ||||||||| ||| |||||||| || ||||| || ||||||||| ||||||||||||||||| |||||||| ||| ||||||||| ||||||||| |||| |||||||| || | ||||||| ||| | ||||||||||| ||||||||||||| ||| ||||||| |||||| || |||| |||| ||||| |||| ||| ||| |||| ||||||||| ||||||| |||||||| ||| ||||||| |||||||| ||||||||| |||| |||||| ||||| |||| ||||||||| |||| ||||||||| |||||| |||||||||| |||||| |||| || ||||||| |||||||| ||| |||||||| |||||||||| || ||||||||||

Dizziness and Orthostatic Hypotension

||| ||||||||| || ||||||||| ||| ||||||||||| ||||||||||| |||| |||||| ||| ||||| |||||||||||||

Somnolence

|||||||||| ||| |||||||| || || |||||| |||||||||

Depression

|||||||||| ||| |||||||| || | |||||| |||||||||

Table 36: Summary of Harms From the M15-741 Trial (SAS)

CSCI = continuous subcutaneous infusion; SAS = safety analysis set; TEAE = treatment-emergent adverse event.

Note: Unless otherwise specified, all TEAEs were summarized using Medical Dictionary for Regulatory Activities (MedDRA) preferred terms.

^a≥ 15% of patients.

^b≥ 2% of patients.

^cAdverse event of special interest, including any adverse event in infusion site–related noninfection reactions company MedDRA query (CMQ).

^dAdverse event of special interest, including any AE in the hallucinations CMQ or psychosis and psychotic disorder standardized MedDRA query (SMQ).

^eAssessed using the Columbia-Suicide Severity Rating Scale.

Source: M15-741 Clinical Study Report.³⁷ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Critical Appraisal

Internal Validity

The open-label study design could introduce reporting bias, potentially leading to inflated benefits of foslevodopa-foscarbidopa on PROs and less favourable harms results given the more subjective nature of these outcomes. The noncomparative design means that known and unknown confounding factors were not accounted for and no statistical adjustments were made in the analyses, making it impossible to be certain that the observed treatment benefits could be attributed to foslevodopa-foscarbidopa alone. As well, a sizable proportion of patients ||||||| withdrew from study treatment, mostly due to AEs and consent withdrawal. As a result, attrition bias may explain the observed efficacy results as patients remaining in the study were more likely to be those who experienced benefits and able to tolerate the treatment better.

External Validity

The inclusion and exclusion criteria generally align with the selection criteria for candidates for advanced therapies used in clinical practice. While the participants in this trial appeared to have more advanced PD than those in the pivotal M15-736 trial, the clinical expert noted that the patient population would fit within the spectrum of patients with advanced PD in Canada. The infusion pump used in the trial (i.e., Crono PAR Series III pump) was different from the one intended for use in Canada (i.e., Vyafuser or the Phillips-Medisize Parkinson’s Disease Subcutaneous Pump Delivery System), and it is unclear if there is any functional difference between them. Much of the appraisal of the M15-736 trial with respect to appropriate use of concomitant PD medications and relevance of study outcomes applies to the M15-741 study given that the study designs were similar. The clinical expert commented that the duration of follow-up of 52 weeks was adequate for assessing the safety of foslevodopa-foscarbidopa.

Description of the M15-737 Trial

Characteristics of the M15-737 trial are summarized in Table 37. M15-737 is an ongoing single-arm, multicentre, open-label extension study of M15-741, with up to 96 weeks of treatment. The objective is to assess the long-term safety and tolerability of foslevodopa-foscarbidopa delivered by CSCI for 24 hours per day. The primary outcomes are AEs and related measures. Efficacy outcomes are also being collected as secondary end points.

Table 37: Details of the M15-737 Trial

AE = adverse event; AESI = adverse event of special interest; CSCI = continuous subcutaneous infusion; ECG = electrocardiogram; EQ-5D-5L = 5-Level EQ-5D; HRQoL = health-related quality of life; MDS-UPDRS = Movement Disorder Society-Unified Parkinson’s Disease Rating Scale; MMSE = Mini-Mental State Examination; NA = not applicable; PD = Parkinson disease; PDQ-39 = Parkinson’s Disease Questionnaire-39 items; PDSS-2 = Parkinson’s Disease Sleep Scale-2; SAE = serious adverse event.

Source: M15-737 Clinical Study Report.⁷⁵ (Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Populations

M15-737 included patients with PD who, at completion of the parent study, remained on the study drug and met additional criteria (consent, demographic and laboratory assessments, study participant history, and contraception). Patients had to be able to understand the nature of the study and have had the opportunity to have their questions answered by the investigator. Patients who had decision-making capacity were allowed to sign and date an informed consent form. Patients also had to be willing to comply with the procedures required in this study and not be considered by the investigator to be an unsuitable candidate for foslevodopa-foscarbidopa. Patients also could not be exhibiting significant suicidal behaviours at the time of the final visit of the parent study. Patients of childbearing potential must have had a negative urine pregnancy test, and both female and male patients must have agreed to certain protocol-specified contraception. Patient characteristics at baseline are summarized in Table 38.

Table 38: Summary of Baseline Characteristics in the M15-737 Trial

COMT = catechol-O-methyltransferase; DBS = deep brain stimulation; MAO-B = monoamine oxidase type B; MDS-UPDRS = Movement Disorder Society-Unified Parkinson’s Disease Rating Scale; NR = not reported; PD = Parkinson disease; SD = standard deviation.

Source: M15-737 Clinical Study Report.⁷⁵ (Note: Details in the table have been taken from the sponsor’s Summary of Clinical Evidence.¹⁵)

Interventions

Patients enrolled in M15-737 received an individualized continuous subcutaneous infusion (24 hours per day) of foslevodopa-foscarbidopa via an infusion set connected to a pump for up to 96 weeks. Use of concomitant medications reflected the protocol of the parent studies.

Outcomes

Efficacy variables were measurements of average normalized daily “off” time and “on” time, assessed using the PD diary; PD symptoms, assessed using the MDS-UPDRS Parts I to IV; quality of life, assessed using the PDQ-39 and EQ-5D-5L; and cognitive impairment, assessed using the MMSE. Safety evaluations were the primary end points and included AEs (including serious adverse events and adverse events of special interest), Infusion Site Evaluation Scale scores, vital signs, ECGs, C-SSRS scores, and QUIP-RS scores.

Statistical Analysis

In M15-737, efficacy analyses were performed in the full analysis set using data collected no more than 1 day after the end of the infusion of foslevodopa-foscarbidopa, unless otherwise stated. Two-sided paired-sample t tests were performed to test the change from baseline. Safety analyses were performed in the safety analysis set. All safety variables (except AEs and infusion site evaluations) were evaluated using data collected no more than 1 day after the end of the infusion of foslevodopa-foscarbidopa. For categorical safety outcomes, the number of and percentage of each category were summarized by visit. Hypothesis testing was not performed. Unless otherwise noted, the baseline value for efficacy variables and continuous safety outcomes was collected from the last assessment in the parent study. No adjustments were made for missing visit data. For PD diary entries, if no data were available for a visit, the average daily normalized “off” or “on” times values were missing for that visit. For the MDS-UPDRS, MMSE, and PDQ-39, the score was calculated as long as no more than 15% of the answers were missing for that assessment, and the missing item was imputed as the average of the nonmissing items from the same assessment. EQ-5D-5L summary index was only calculated if answers were provided for all questions.

Results

Patient Disposition

Patient disposition is summarized in Table 39. M15-737 is an ongoing study. No patients had completed the study at the time of the data cuts submitted by the sponsors. The majority of patients who completed the parent study went on to enrol in the long term extension study.

Table 39: Patient Disposition in the M15-737 Trial