If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Efficacy of nabiximols oromucosal spray on spasticity in people with multiple sclerosis: Treatment effects on Spasticity Numeric Rating Scale, muscle spasm count, and spastic muscle tone in two randomized clinical trials
Clinician-rated Modified Ashworth Scale scores were improved with nabiximols.
•
Nabiximols has a larger effect on muscle groups more affected by spasticity.
Abstract
Background
To provide a comprehensive assessment of the treatment effects of nabiximols oromucosal spray on multiple sclerosis spasticity in two clinical trials, GWSP0604 and SAVANT.
Methods
Both studies enriched for responders before randomization, defined by a ≥20% improvement in Spasticity 0-10 numeric rating scale (NRS) score. Additionally, SAVANT used randomized re-titration following washout. Spasticity NRS outcomes, spasm count, and modified Ashworth scale (MAS) scores were analyzed.
Results
Mean change from baseline in average daily Spasticity NRS scores was significantly larger for nabiximols than placebo at all postbaseline timepoints, ranging from -0.36 to -0.89 in GWSP0604 and -0.52 to -1.96 in SAVANT. Percent reduction in geometric mean change from baseline in average daily spasm count for nabiximols ranged from 19–35% versus placebo. A treatment difference favoring nabiximols was observed in overall MAS scores during the randomized part of each study. Treatment effect was larger for combinations of lower limb muscle groups (ranging between -0.16 and -0.37).
Conclusions
Nabiximols leads to improvement in spasticity that was sustained over the 12-week treatment period as measured by average daily Spasticity NRS scores, daily spasm counts, and MAS scores for combinations of muscle groups, especially the combination of the 6 key muscle groups in the lower limbs in NRS responders to nabiximols treatment.
Global Burden of Disease Neurological Disorders Collaborator Group Global, regional, and national burden of neurological disorders during 1990-2015: a systematic analysis for the Global Burden of Disease Study 2015.
). Assessments of spasticity can include clinician-rated measures of muscle tone such as the Modified Ashworth Scale (MAS) and patient-reported outcomes (PROs) such as the Spasticity 0-10 numeric rating scale (NRS) or daily spasm count.
To administer the MAS, the clinician will rate the resistance in a joint on a scale from 0-4, with 0 indicating no increase in muscle tone and 4 indicating rigidity (
). Despite widespread use and satisfactory reliability across applications, the reliability and validity of the MAS has not been established specifically in PwMS (
). The measure also requires training and standardized technique, measures only resistance to passive movement, and is necessarily limited to the time and setting of the clinic visit, among other limitations (
The Spasticity NRS is a visual analog scale that ranges from 0 (no spasticity) to 10 (worst possible spasticity). As a PRO, the NRS allows the patient to provide feedback on their own experience at different times and places (
Validity, reliability, and clinical importance of change in a 0-10 numeric rating scale measure of spasticity: a post hoc analysis of a randomized, double-blind, placebo-controlled trial.
). An approximately 30% change in Spasticity NRS score is considered a clinically important difference, consistent with a “much improved” or better on the Patient Global Impression of Change scale (
Validity, reliability, and clinical importance of change in a 0-10 numeric rating scale measure of spasticity: a post hoc analysis of a randomized, double-blind, placebo-controlled trial.
). Additionally, a survey of North American Research Committee on Multiple Sclerosis (NARCOMS) registry participants showed that fewer than 50% (4518 of 10,353) of responders were satisfied or very satisfied with their spasticity treatment, highlighting a significant unmet need in MS spasticity symptom management (
Nabiximols, a cannabinoid-based oromucosal spray (GW 1000-02 [Sativex®]), was developed as a treatment for MS spasticity and is approved across EU Member States and in other regions for the treatment of MS spasticity; however, nabiximols is not approved for use in the US. Nabiximols is a complex botanical mixture containing delta-9-tetrahydrocannibinol (THC), cannabidiol (CBD), and other plant-derived cannabinoids and non-cannabinoid constituents. The efficacy of nabiximols in treating MS spasticity was studied in four randomized, placebo-controlled clinical trials (GWMS0106, GWCL0403, GWSP0604, and SAVANT) (
A randomized, double-blind, placebo-controlled, parallel-group, enriched-design study of nabiximols* (Sativex(R)), as add-on therapy, in subjects with refractory spasticity caused by multiple sclerosis.
Sativex(R) as add-on therapy vs. further optimized first-line ANTispastics (SAVANT) in resistant multiple sclerosis spasticity: a double-blind, placebo-controlled randomised clinical trial.
Studies GWSP0604 (EudraCT2015-004451-40) and SAVANT (NCT00681538) were enriched randomized, placebo-controlled, parallel-group studies that derived the primary endpoint from daily spasticity scores reported by patients using an 11-point Spasticity NRS, where a score of 0 equated to “no spasticity” and a score of 10 equated to “worst possible spasticity.” Both studies featured a nabiximols run-in trial period (Part A) to enrich for responders, as defined by a ≥ 20% improvement in Spasticity NRS score, but used distinct designs: GWSP0604 randomized trial period responders into Part B without a treatment discontinuation period (washout) between trial period and randomization, whereas SAVANT Part B included randomization and re-titration following a washout.
In addition to reanalyzing the available Spasticity NRS outcomes using updated statistical methods, this post hoc analysis of GWSP0604 and SAVANT data also includes an analysis of patient-reported daily spasm count and clinician-assessed MAS scores to provide a comprehensive assessment of the treatment effects of nabiximols. The analytical approach applies an equivalent statistical model across both studies, allowing for appropriate consideration of missing data and non-normal distribution of spasm count. Focusing on MAS scores for specified combinations of muscle groups, as well as efficacy in patient subsets with increasing baseline spasticity severity, also allows for a broader approach to the understanding of the clinical relevance of the treatment effect of nabiximols.
2. Materials and methods
2.1 Study design and patients
This post hoc analysis included patients from the pivotal randomized clinical trials GWSP0604 and SAVANT. The study designs and patient populations have been previously published (
A randomized, double-blind, placebo-controlled, parallel-group, enriched-design study of nabiximols* (Sativex(R)), as add-on therapy, in subjects with refractory spasticity caused by multiple sclerosis.
Sativex(R) as add-on therapy vs. further optimized first-line ANTispastics (SAVANT) in resistant multiple sclerosis spasticity: a double-blind, placebo-controlled randomised clinical trial.
). The studies were approved by the relevant Institutional Review Boards or Ethics Committees and were conducted according to Good Clinical Practice guidelines. All patients were aged 18 or over and provided written informed consent. GWSP0604 and SAVANT were both multicenter, randomized, placebo-controlled, parallel-group studies with enrichment for early response designs. Both studies used a 4-week treatment period of single-blind nabiximols (Part A) to identify, based on ≥20% improvement in Spasticity NRS scores, initial responders eligible for randomization into a double-blind placebo-controlled Part B.
In SAVANT only, a second enrichment step included nabiximols washout of up to 4 weeks between Part A (trial period) and Part B (randomization phase); Part A responders who also reported a reduction of at least 80% in their improvement in average daily Spasticity NRS score during washout were then eligible for entry into Part B. During Part B of SAVANT, patients were randomized to re-titrate nabiximols or placebo back to the optimal individual dose identified in Part A. In GWSP0604, there was no washout period, and Part A responders were randomized in Part B to either maintain the same dose of nabiximols attained at the end of Part A or switched to placebo.
2.2 Assessments
2.2.1 Numeric rating scale
Spasticity NRS scores were entered daily by the patient using an interactive voice response system (IVRS) (GWSP0604) or eDiary (SAVANT). Participants reported Spasticity NRS score (0–10) daily, and their 7-day average daily score was calculated. The treatment effect was measured as a change from Part B baseline score in 7-day average Spasticity NRS scores and was assessed weekly throughout the 12-week Part B treatment periods.
2.2.2 Spasm count
Spasm count data was entered daily by the patient using IVRS (GWSP0604) or eDiary (SAVANT). Patients reported a daily spasm count each day at bedtime from Visit 1 (baseline) through the end of Part B treatment (completion or withdrawal). Patients were asked how many spasms they had in the last 24 h. Patients were instructed to give a best estimate if a large number of spasms occurred and to enter ‘0’ if no spasms were experienced. The IVRS implemented in study GWSP0604 was limited to two digits for data collection. As such, the maximum spasm count allowed for a given day was 99. The eDiary in SAVANT allowed entry of up to three digits; therefore, the upper limit to the number of spasms was 999. For these post hoc analyses, the 7-day average baseline and 28-day average post-baseline daily spasm count were calculated. To account for non-normality, the treatment effect measured as a change from Part B baseline in the log of the average daily spasm count +1 (to avoid log 0) was assessed throughout the 12-week Part B treatment period. Since baseline spasm count may affect whether an improvement is observed, treatment effect analysis was conducted for different thresholds of 7-day average baseline spasm count: ≥0, ≥2, ≥4, and ≥6 spasms per day. Patients were grouped into threshold classifications based on the reported spasm count at the beginning of Part A in GWSP0604 (before the run-in trial period) and the beginning of Part B for SAVANT (after both the run-in trial period and the washout period).
2.2.3 Muscle tone MAS
Both GWSP0604 and SAVANT utilized the MAS as a clinician-rated measure of velocity-dependent muscle tone for 20 muscle groups across upper and lower limbs on both sides of the body. Average MAS (scale from 0 to 5) score was assessed at Part A baseline, Part B randomization, and the end of 12 weeks in GWSP0604, and every 4 weeks during the 12-week placebo-controlled treatment in SAVANT. Prior to statistical analysis, the rater-recorded MAS scores were transformed using the following algorithm: MAS untransformed [to MAS transformed] scores; 0[0], 1[1], 1+[2], 2[3], 3[4], and 4[5]. The transformation of the MAS scores enabled the 1+ scale score to be analyzed statistically. Because symptoms of spasticity predominantly affect muscle groups in the lower limbs of PwMS (
), an exploratory analysis also investigated muscle tone among six defined combinations of muscle groups derived based on the advice provided by several experts in MS clinical assessments and the MAS: total MAS (average of all 20 muscle groups); MAS-lower limb muscle tone (MAS-LLMT6) (average of six muscle groups: knee flexors, knee extensors, and plantar flexors on both right and left sides); MAS-LLMT4 (average of four muscle groups: knee flexors and knee extensors on both right and left sides); lower limbs (LL) (average of all eight LL muscle groups: hip adductors, knee flexors, knee extensors, and plantar flexors on both right and left sides); upper limbs (UL) (average of all 12 UL muscle groups: Elbow flexors, elbow extensors, elbow pronators, elbow supinators, wrist flexors, finger flexors on both right and left sides); and most affected limb (MAL) (average of all muscle groups in the MAL, defined as the limb [left leg, right leg, left arm, right arm] with the highest average MAS at baseline).
2.3 Statistical analysis
Differences in study design across the pivotal studies preclude pooling of data for integrative efficacy analysis, so outcomes were assessed separately for each study. The 7-day average daily Spasticity NRS scores were analyzed using the mixed model for repeated measures (MMRM). For the analyses of muscle tone and daily spasm count, both MMRM and analysis of covariance (ANCOVA) were used to analyze the data. A pattern mixture model with control-based imputation was also performed for each measure, where withdrawn patients from the nabiximols arm were assumed to follow the placebo mean trajectory post withdrawal. To address the non-normality of spasm count, analyses were conducted on a natural log-transformed scale, and a value of 1 was added to the average spasms for all patients to avoid log 0.
Percent change analyses were conducted with thresholds set at 20%, 30%, 40%, and 50% improvement to evaluate changes in the Spasticity NRS scores, MAS scores, and spasm count. A logistic regression model using a generalized estimating equation was fitted to binary responses over time for the 20% and 30% thresholds. To evaluate the change in spasticity measures after randomized withdrawal from nabiximols in GWSP0604, the odds of worsening by 20% or 30% was analyzed. An odds ratio (OR) greater than 1 indicates placebo patients had greater odds of manifesting 20% or 30% Spasticity NRS worsening following randomized withdrawal than patients continuing on nabiximols, suggesting a nabiximols benefit.
3. Results
3.1 Baseline characteristics and patient disposition
The demographic and clinical characteristics of patients enrolled in the GWSP0604 and SAVANT trials have been published previously (
A randomized, double-blind, placebo-controlled, parallel-group, enriched-design study of nabiximols* (Sativex(R)), as add-on therapy, in subjects with refractory spasticity caused by multiple sclerosis.
Sativex(R) as add-on therapy vs. further optimized first-line ANTispastics (SAVANT) in resistant multiple sclerosis spasticity: a double-blind, placebo-controlled randomised clinical trial.
). Selection criteria of both studies were similar, especially with regard to requiring Spasticity NRS greater than 4. Distinct from GWSP0604, SAVANT required at least two anti-spasticity medication treatment failures and current anti-spasticity treatment. Of the 572 patients entering Part A of GWSP0604, 241 patients qualified as responders and remained in the study to be randomized in the double-blind Part B portion (nabiximols n=124, placebo n=117) (
A randomized, double-blind, placebo-controlled, parallel-group, enriched-design study of nabiximols* (Sativex(R)), as add-on therapy, in subjects with refractory spasticity caused by multiple sclerosis.
Sativex(R) as add-on therapy vs. further optimized first-line ANTispastics (SAVANT) in resistant multiple sclerosis spasticity: a double-blind, placebo-controlled randomised clinical trial.
). During the 1- to 4-week washout period, 106 patients reported an increase in average daily Spasticity NRS score of at least 80% of the initial reduction in average daily NRS score during Part A and were randomized for re-titration in Part B (nabiximols n=53, placebo n=53).
The demographic and clinical characteristics of patients entering Part B of each trial were comparable between treatment groups (Table 1). There was a female predominance with similar mean age in both studies. At baseline, patients exhibited significant disability, with more than 50% of patients having a baseline mean Expanded Disability Status Scale score ≥6.0. The average time since diagnosis of MS for patients across both studies was at least 11 years before study entry, and the average time since onset of spasticity symptoms was at least 6 years. All SAVANT and 85% of GWSP0604 participants were taking concomitant anti-spasticity medication(s), most commonly baclofen.
Table 1Part B baseline demographics.
GWSP0604 randomized discontinuation
SAVANT randomized re-titration
Placebo + SOC (n=117)
Nabiximols + SOC (n=124)
Placebo + SOC (n=53)
Nabiximols + SOC (n= 53)
Female sex, n (%)
73 (62)
72 (58)
35 (66)
36 (68)
Mean age, years
48.1
49.2
50.1
52.0
Mean duration of MS, years
11.8
13.3
14.3
13.2
MS subtype, n (%)
Primary progressive
19 (16)
20 (16)
7 (13)
7 (13)
Secondary progressive
58 (50)
62 (50)
22 (42)
21 (40)
Relapsing/remitting
37 (32)
41 (33)
24 (45)
25 (47)
Progressive relapsing
3 (3)
1 (1)
0
0
Baseline EDSS score ≥6, n (%)
79 (67.5)
85 (68.5)
29 (55)
29 (55)
Mean duration of spasticity, years
6.7
8.6
8.6
7.9
Mean Part A baseline Spasticity NRS score
7.0
6.8
6.9
6.9
Currently taking anti-spasticity medication, n (%)
100 (85.5)
104 (83.9)
53 (100)
53 (100)
Baclofen
74 (63.2)
66 (53.2)
47 (88.7)
43 (81.1)
Tizanidine
23 (19.7)
24 (19.4)
16 (30.2)
17 (32.1)
Benzodiazepine derivative
24 (20.5)
19 (15.3)
9 (17.0)
11 (20.8)
EDSS: Expanded Disability Status Scale; MS: multiple sclerosis; NRS: numeric rating scale; SOC: standard of care.
In GWSP0604, mean Part B baseline Spasticity NRS scores, at the end of the Part A trial period when both groups were on nabiximols prior to the randomized withdrawal from treatment, was 3.9; the score did not differ between cohorts at randomization to placebo and nabiximols. Over the 12-week treatment period, the difference in mean change from baseline in average daily Spasticity NRS scores significantly favored nabiximols over placebo at each 1-week interval, ranging from -0.36 to -0.89 (Fig. 1A). Missing data had little impact on the findings, as the findings were similar for missing at random (MAR) or control-based imputation in MMRM.
Fig. 1Spasticity NRS mean change from baseline over time in (A) GWSP0604 Part B and (B) SAVANT Part B.
Notes: Analysis was performed using an MMRM approach, the model includes terms for baseline Spasticity NRS score, treatment, week number, treatment-by-week number interaction, and baseline-by-week number interaction as fixed effects. Week number (within-patient) was used as the repeated effect. Estimates were derived using the restricted maximum likelihood approach. The within-patient correlation was modelled using the unstructured covariance structure and the Kenward-Roger approximation to estimate the denominator degrees of freedom, P-value, estimate and the 95 % CI for the primary comparison (Week 12) between treatment groups.
CI: confidence interval; LS: least squares; MMRM: mixed models for repeated measures; NRS: Numerical Rating Scale.
Mean Spasticity NRS scores at SAVANT Part B baseline, after the washout period, were 6.9 for both the nabiximols and placebo groups. Over the 12-week treatment period, the difference in mean change from baseline in average daily Spasticity NRS scores significantly favored nabiximols over placebo at each 1-week interval (Fig. 1B), with differences ranging over time from -0.52 to -1.96. Missing data had little impact on the findings, as they were similar for MAR or control-based imputation in MMRM.
To evaluate the change in Spasticity NRS after randomized withdrawal from nabiximols in GWSP0604, the odds of worsening by 20% or 30% was analyzed. The ORs for 20% and 30% improvement thresholds across the 12 weeks of treatment generally showed a favorable pattern for nabiximols, supporting findings from the MMRM analysis (Fig. 2A).
Fig. 2Logistic regression analysis of Spasticity NRS responders in (A) GWSP0604 Part B and (B) SAVANT Part B.
Logistic regression analysis of Spasticity NRS scores for patients showing ≥20% and ≥30% improvement following randomized re-titration was performed in SAVANT. Adjusted OR for responders showing ≥20% and ≥30% improvement in change from baseline beginning Week 1 through Week 12 consistently favored nabiximols versus placebo. As shown by the estimates supported by the confidence intervals, the findings provide robust evidence of reduction in spasticity attributable to nabiximols (Fig. 2B).
3.3 Spasm count
The treatment difference as measured by the change from baseline of the log of the average daily spasm count +1 favored nabiximols in both GWSP0604 and SAVANT at all time points (Fig. 3). By week 12, the percent reduction compared with placebo for geometric mean change from baseline in average daily spasm count for the overall population (baseline spasm count ≥0 spasms per day) was 19% in GWSP0604 and 34% in SAVANT. Over both studies, in the subgroups with ≥2, ≥4, and ≥6 spasms per day at baseline, the percent reduction in daily spasm count in the nabiximols group, compared to the placebo group, increased slightly as the baseline threshold increased, ranging from 21% to 37% by Week 12. Missing data had little impact on the findings, as they were similar for MAR or control-based imputation in MMRM.
Fig. 3Percent difference in change in spasm count from baseline between
Logistic regression analysis of spasm count for patients showing ≥20% and ≥30% increase in GWSP0604 or reduction in SAVANT were conducted. In GWSP0604, placebo patients had greater odds of manifesting 20% or 30% worsening in average daily spasm count following randomized withdrawal than patients continuing on nabiximols. These GWSP0604 results suggest a benefit in spasm count attributable to nabiximols (Fig. 4A). In SAVANT, responder definitions for 20% or 30% reduction from baseline showed ORs in favor of nabiximols for Weeks 1 through 4, 5 through 8, and 9 through 12 (Fig. 4B). Findings were consistent when applying baseline thresholds ranging from 2 to 6 spasms per day. Overall, the data indicate a reduction over time in spasm count with nabiximols in SAVANT in NRS responders.
Fig. 4Logistic regression analysis of 20% and 30% spasm count responders in (A) GWSP0604 Part B and (B) SAVANT Part B.
Baseline MAS scores differed among the various muscle groups (Table 2), with LL muscle groups showing relatively greater pre-treatment scores (2.4 to 2.7) than the UL muscle groups (0.6 to 1.2). The designated MAL, defined as the limb with the highest average MAS score at baseline, was one of the legs in more than 96% of patients across both studies, and showed average baseline MAS scores ranging from 2.6 to 2.9.
Table 2Mean baseline scores for MAS and muscle groups.
GWSP0604 Part Aa
SAVANT Part B
Nabiximols
Placebo
Nabiximols
(n=241)
(n=53)
(n=53)
Overall MAS
1.4
1.7
1.4
MAS-LLMT6
2.6
2.6
2.5
MAS-LLMT4
2.7
2.7
2.5
LL
2.6
2.6
2.4
UL
0.6
1.2
0.8
MAL
2.9
2.8
2.6
aGWSP0604 Part A BL values shown because no washout before Part B.
Treatment difference for the average MAS scores from various muscle groups favored nabiximols over placebo (Fig. 5). While a treatment effect favoring nabiximols over placebo was observed for change from baseline in overall MAS score during Part B of each study (-0.09 to -0.24), the treatment effect was larger for various combinations of LL muscle groups (ranging between -0.16 and -0.37) compared with the average MAS score across all UL muscle groups at Week 12 (-0.05 and -0.19, respectively). Treatment differences were not statistically significant in GWSP0604. In SAVANT, conversely, nabiximols reached a statistically significant treatment difference compared with placebo at all time points and for all combinations of muscle groups. Missing data had little impact as the findings were similar for MAR or control-based imputation in MMRM.
Fig. 5Treatment differences (nabiximols – placebo) in MAS scores.
Since GWSP0604 was a randomized withdrawal study, logistic regression analysis of MAS-LLMT-6 was conducted for patients showing ≥20% and ≥30% increase (worsening) following randomization. For MAS-LLMT-6, the ≥20% and ≥30% findings favored nabiximols with confidence intervals supporting an effect attributable to nabiximols (Fig. 6A). The treatment effects were slightly greater among subsets of patients defined by higher MAS-LLMT-6 score at baseline. Logistic regression analysis of MAS-LLMT-6 for patients showing ≥20% and ≥30% reduction following randomized re-titration in SAVANT was performed. For MAS-LLMT-6, the findings for ≥20% and ≥30% reduction strongly favored nabiximols (Fig. 6B). The treatment effects were similar among subsets of patients defined by higher MAS-LLMT-6 score at baseline.
Fig. 6Logistic regression analysis of 20% and 30% MAS-LLMT6 responders in (A) GWSP0604 part B and (B) SAVANT Part B.
A randomized, double-blind, placebo-controlled, parallel-group, enriched-design study of nabiximols* (Sativex(R)), as add-on therapy, in subjects with refractory spasticity caused by multiple sclerosis.
Sativex(R) as add-on therapy vs. further optimized first-line ANTispastics (SAVANT) in resistant multiple sclerosis spasticity: a double-blind, placebo-controlled randomised clinical trial.
). Nabiximols was generally well tolerated in both studies, and the most common adverse events during the Part A trial period of nabiximols treatment included dizziness, fatigue, and somnolence. Incidence of adverse events was generally lower in the SAVANT study (Tables 3 and 4).
Table 3Part A AE summary.
PwMS, n (%)
GWSP0604 Part A – Nabiximols (n=572)
SAVANT Part A – Nabiximols (n=191)
With an AE
268 (46.9)
46 (24)
Withdrew due to an AE
35 (6.1)
5 (2.6)
AEs
Dizziness
80 (14.0)
4 (2.1)
Fatigue
34 (5.9)
3 (1.6)
Somnolence
29 (5.1)
3 (1.6)
Dry mouth
24 (4.2)
3 (1.6)
Nausea
23 (4.0)
4 (2.1)
Vertigo
21 (3.7)
14 (7.3)
With an SAE
9 (1.6)*
2 (1.0)
Withdrew due to an SAE
0 (0)
1 (0.5)
*Includes one patient with SAE reported post study.
AE: adverse event; PwMS: people with multiple sclerosis; SAE: serious adverse event.
This post hoc analysis of the GWSP0604 and SAVANT clinical trials assessed the efficacy of nabiximols in treating symptoms of spasticity in patients with MS with new methodology and exploratory endpoints based on original study findings, including additional methods for addressing missing data, log-transformed analyses to adjust for non-normality of data, exploratory MAS muscle groupings, logistic regression models for binary responder analysis, and subgroup analyses based on baseline symptom severity. The results of this analysis extend those from the original analysis of both studies, demonstrating that nabiximols led to a sustained treatment effect over 12 weeks on Spasticity NRS scores and average daily muscle spasm counts, as well as average MAS scores for various combinations of muscle groups at distinct time points in NRS responders to nabiximols treatment in SAVANT.
In this post hoc analysis of MAS scores for specified combinations of muscle groups, we found that certain muscle groups (eg, those in the LL, which are generally more affected at baseline, or those in the MAL at baseline) showed a larger nabiximols treatment effect than the overall MAS score. MAS-LLMT6 is highlighted as a representative MAS-derived outcome measure in MS spasticity. While it is not the only potential MAS muscle combination parameter for analysis, MAS-LLMT6 provides a clinician-rated measure of spasticity reflecting velocity-dependent increase in muscle tone in the clinically relevant muscle groups in the LLs. Although the term MAS-LLMT6 may be novel, it was coined to provide specificity and transparency around its MAS derivation. Other LL measures are highly correlated with MAS-LLMT6 and may also be suitable clinical trial endpoints. Compared with the MAS-LLMT4 endpoint, the MAS-LLMT6 endpoint has the advantage of including the ankle plantar flexors, which are clinically relevant, especially for gait, and are easily accessible to rating by a clinician. The change in MAS-LLMT6 score also showed less variability over time in patients assigned to placebo due to the larger number of muscle groups covered than MAS-LLMT4.
In addition to clinical assessments, patient-reported measures for spasticity are critical in the assessment of therapeutic interventions (
). The Spasticity NRS easily assesses the patients’ own daily experience, which is vital, especially for complex constructs such as spasticity. Overall, the Spasticity NRS is a global measure of patient-reported symptoms of spasticity and assesses changes in a patient's condition that may be undetectable by clinician-administered physical examination but are meaningful to the patient. Clinician-measured spasticity is usually assessed in the clinic (eg, at scheduled visits with MAS assessments); scores may be systematically impacted if visits are missed, delayed, or rescheduled for reasons related to spasticity itself (
). Since spasticity varies day-to-day, or even over the course of a day, clinician ratings are not well suited to capture the full range of the spasticity experience; assessment with the Spasticity NRS is a better reflection of daily patient experience, which can be averaged over multiple days to generate a reliable estimate of spasticity severity.
Muscle spasms are a spasticity-related symptom that may occur in 95% of patients with MS with spasticity; therefore patient-reported spasm count is an important element to assess when evaluating therapeutic interventions for spasticity (
). This post hoc analysis, which considered the percent change in average daily spasm count from baseline, utilized MAR for missing data, as opposed to the last observation carried forward assumption of the original analysis. The post hoc analysis model showed a statistically significant mean difference clearly favoring nabiximols over placebo for percent change in average daily spasm count at all time points in both studies. This finding was heightened in patient subsets with greater baseline daily spasm count. In published clinical trials of other anti-spasticity medications, effect on spasm frequency was commonly assessed with outcome measures other than daily spasm count (eg, Penn spasm frequency scale, which is a categorical measure) (
). Hence, the spasm count data from the nabiximols MS spasticity program may not be directly comparable with published data for baclofen, tizanidine, dantrolene, or gabapentin, which, moreover, are used to manage patients earlier on their spasticity.
In this analysis, nabiximols was shown to significantly improve spasticity across a range of outcome measures that reflect various approaches to the evaluation of spasticity in MS patients (
). Estimates for the minimum clinically important difference (MCID) and clinically important difference (CID) of Spasticity NRS scores in patients with MS were previously published based on clinical study data (
Validity, reliability, and clinical importance of change in a 0-10 numeric rating scale measure of spasticity: a post hoc analysis of a randomized, double-blind, placebo-controlled trial.
). Using an anchor-based method with a patient global impression of change scale, an 18.0% improvement in Spasticity NRS score was considered as the MCID, and a 29.5% improvement in Spasticity NRS score was considered as the CID. The -20% and -30% thresholds used for logistic regression analyses herein represent the rounded-up values of the MCID and CID for Spasticity NRS, respectively.
Although the MCID and CID definitions for Spasticity NRS were not specifically derived for worsening, the +20% and +30% thresholds for Spasticity NRS were used to define worsening in the analysis of data collected during the double-blind randomized withdrawal in GWSP0604 Part B. Treatment differences in GWSP0604 represent a combination of placebo worsening and continued nabiximols improvement during the double-blind randomized withdrawal phase. Overall, in Study GWSP0604, placebo patients had about twice the odds to manifest +20% or +30% Spasticity NRS score worsening following randomized withdrawal compared with patients continuing on nabiximols, a result favoring nabiximols, and with the confidence intervals suggesting reduction in spasticity attributable to nabiximols. Stronger findings were observed in SAVANT, with the odds of nabiximols patients being a -20% responder around six- to ten-times greater than for placebo patients, and the odds of being a -30% responder around four- to eight-times greater, with compelling statistical support. By these definitions, the Spasticity NRS benefit in both studies is clinically meaningful.
The 20% and 30% response definitions were also applied to changes from baseline in spasm count and muscle tone. These thresholds were carried forward as possible benchmarks for relevant changes in muscle tone and spasm count. Although clinically meaningful thresholds for changes in muscle tone and muscle spasms have not been formally established in the MS population, analyses of 20% and 30% change compared with placebo on LLMT-6 and spasm count clearly demonstrated the benefit of nabiximols. The changes in the overall MAS score are limited by the sensitivity to change of the 0-5 point converted scale and smaller changes in muscle tone may go undetected and are lost.
Study limitations include the post hoc nature of the analysis; the included studies were not specifically intended to be combined for analysis and had slightly different study designs. In addition, specific information regarding factors that could affect spasticity (eg, rates of urinary tract infection or recent relapses) was not captured in either of these studies. However, it is expected that randomization would ensure that these factors were evenly distributed between the treatment and placebo groups, which were comparable in terms of the demographics measured.
Both of the studies included in this post hoc analysis used a study population enriched for responders to nabiximols, which could limit generalizations to the larger population of people with MS spasticity. However, this was done to better reflect clinical use, since patients who do not respond to nabiximols are unlikely to remain on treatment, while those that do respond could experience substantial benefits (
A randomized, double-blind, placebo-controlled, parallel-group, enriched-design study of nabiximols* (Sativex(R)), as add-on therapy, in subjects with refractory spasticity caused by multiple sclerosis.
Sativex(R) as add-on therapy vs. further optimized first-line ANTispastics (SAVANT) in resistant multiple sclerosis spasticity: a double-blind, placebo-controlled randomised clinical trial.
). Examining an enriched population therefore allows for better estimation of the potential treatment effect of nabiximols in clinical practice.
This study, like the study of MS spasticity in general, is hampered by the lack of formal diagnostic and assessment criteria. The measures used in this analysis, the MAS and NRS, have both strengths and weaknesses. The MAS has been very widely used in research, and therefore allows for these results to be viewed in light of many previous studies. However, it has not been established as a reliable measure of MS spasticity specifically. The NRS has not been as widely used as the MAS but has demonstrated both reliability and validity, and is also in keeping with the current drive to improve the ability to capture the full patient experience of both a disease state and its treatment.
5. Conclusions
Overall, the findings detailed in this post hoc analysis demonstrate the positive outcome of using nabiximols compared with placebo in the treatment of spasticity in patients with MS based on well-controlled studies in which the prespecified analyses featured change in average daily Spasticity NRS score as the primary efficacy endpoint. Associated symptoms of spasticity such as pain, sleep disruption, or bladder dysfunction are not analyzed in this work. The robustness of the statistically significant effect observed per the original prespecified analysis plans was confirmed by post hoc analyses using MMRM analysis methods and treatment of missing data. The clinically meaningful Spasticity NRS results are accompanied by positive findings for change in spasm count in both GWSP0604 and SAVANT and improvements in spastic muscle tone in SAVANT.
Data availability statement
All relevant data are provided with the manuscript and supporting files.
Jacqueline Nicholas has received research grants from ADAMAS, Biogen Idec, Genentech, Novartis, and PCORI and consulting and/or speaking fees from Alexion, Bristol Myers Squib, EMD Serono, Genetech, Jazz Pharmaceuticals, Inc., Novartis, and Viela Bio.
Fred Lublin has consulted for, conducted studies funded by, or received honraria for services provided to Jazz Pharmaceuticals, Inc.
Sylvia Klineova has consulted for, conducted studies funded by, or received honraria for services provided to Biogen Idec, Alexion and Jazz pharmaceuticals, Inc.
Joris Berwaerts is an employee of Jazz Pharmaceuticals, Inc.
Robert Chinnapongse is an employee of Jazz Pharmaceuticals, Inc.
Daniel Checketts is an employee of Jazz Pharmaceuticals, Inc.
Joshua R. Steinerman is an employee of Jazz Pharmaceuticals, Inc.
Sajida Javaid has consulted for, conducted studies funded by, or received honraria for services provided to Jazz Pharmaceuticals, Inc.
Funding
This study was sponsored by GW Research Ltd (Cambridge, UK), now a part of Jazz Pharmaceuticals, Inc. Medical writing, under the direction of the authors, was provided by Luke Ward, PhD, and editing support by Dena McWain, both of Ashfield MedComms, an Inizio company, and funded by Jazz Pharmaceuticals, Inc.
References
Abboud H.
Macaron G.
Yu X.X.
et al.
Defining the spectrum of spasticity-associated involuntary movements.
Validity, reliability, and clinical importance of change in a 0-10 numeric rating scale measure of spasticity: a post hoc analysis of a randomized, double-blind, placebo-controlled trial.
Sativex(R) as add-on therapy vs. further optimized first-line ANTispastics (SAVANT) in resistant multiple sclerosis spasticity: a double-blind, placebo-controlled randomised clinical trial.
A randomized, double-blind, placebo-controlled, parallel-group, enriched-design study of nabiximols* (Sativex(R)), as add-on therapy, in subjects with refractory spasticity caused by multiple sclerosis.
00249-3&id=gr1.jpg){kind=link}
00249-3&id=gr2.jpg){kind=link}
00249-3&id=gr3.jpg){kind=link}
00249-3&id=gr4.jpg){kind=link}
00249-3&id=gr5.jpg){kind=link}
00249-3&id=gr6.jpg){kind=link}