Spasticity, defined as abnormally increased muscular tone, is a common symptom in patients with multiple sclerosis (MS) patients.1 MS spasticity causes or worsens different MS symptoms, usually having a big impact on patients' quality of life.2 The most frequent symptoms associated with spasticity in patients with MS, beyond mobility worsening, are spasms, pain, poor sleep quality (linked to pain and nocturnal spasms) and urinary dysfunction.3 Classically recommended medications for generalised spasticity, including baclofen, tizanidine, dantrolene, benzodiazepines and clonazepam, have shown limited clinical benefit in systematic reviews.4 5 The approval of 9-δ-tetrahydocannabinol and cannabidiol (THC:CBD) oromucosal spray (Sativex), in a number of European countries, including Italy, provides a new opportunity as an add-on medication for the management of moderate to severe generalised spasticity and related symptoms in patients with MS resistant to common oral antispastic drugs.6 7 Sativex is an endocannabinoid system modulator containing THC and CBD in a near 1:1 ratio. THC interacts with cannabinoid human receptors that play a key role in the modulation of muscle tone, whereas CBD at higher than natural concentrations may limit the psychoactive effects of THC.8 The efficacy of Sativex oromucosal spray as an add-on therapy for symptoms improvement in patients with MS with moderate to severe MS spasticity has been demonstrated in several clinical trials.6 9–12 The largest pivotal phase III, enriched-design clinical trial included patients with MS with moderate–severe spasticity corresponding to a patient-reported numerical rating scale (NRS) score ≥4.

In this clinical trial, after a first 4 weeks single-blind THC:CBD trial period, 47% of patients showed an initial improvement in spasticity evaluated by the MS spasticity 0–10 NRS scale (defined as a ≥20% reduction vs baseline). During the second phase, consisting of a double-blinded randomisation of the initial responders to continue either with THC:CBD treatment or placebo, a significantly greater proportion of patients receiving THC:CBD achieved a clinically relevant response (defined as a ≥30% reduction in their baseline NRS score). Almost 47% of patients reported adverse events (AEs), with mild to moderate dizziness and fatigue being the most common treatment-related adverse event.12 Although clinical trials are needed to evaluate the efficacy and safety of a new compound to defend its approval, observational studies are essential to depict the effect of treatment in real-world conditions.13 Postapproval surveillance registries have been established to evaluate possible short-term and long-term risks associated with Sativex use.14 15 The MOVE 2 German observational study (n>300) of THC:CBD oromucosal spray showed results aligned with those from clinical trials, with 42% of patients being initial responders after 4 weeks.16 Another recent prospective observational study carried out in Spain (n>200) provided data about 2/3 of patients still on prescription after 6 months from treatment start.17 In these observational studies, patients tended to use lower doses compared with clinical trials (6–7 sprays/day vs >8) and the incidence of AEs was lower, with no evidence of addiction, abuse, misuse or memory impairment.17 Given the growing importance of data derived from real-world observational studies, we decided to collect information about Sativex use in Italy, using the Italian medicine agency (AIFA) prospective e-registry, mandatory for all patients receiving a Sativex prescription in Italy, as a main source document, considering that, after the large sample was collected, much larger than that in the available THC:CBD studies so far, it would bring a comprehensive picture of its effect, minimising biases. For further information, we also used complementarily the involved patients' medical charts retrospective review.

The main aim of our study was to describe Sativex effectiveness and AEs, including possible evidence of abuse or misuse in a large population of Italian patients with MS.

Data were analysed using the STATA V.11.0 software packages (StataCorp. 2011. Stata Statistical Software: Release 12. College Station TSL). Data cleaning was performed before the data analysis considering both range and consistence checks.

Quantitative variables were described using means and SDs. The difference between means and the difference between proportions was evaluated by the t test and the χ² test, respectively. A Shapiro-Wilk test was performed to assess the normal distribution of data. In case of a not normal distribution, appropriate non-parametric tests were performed. Unconditional logistic regression analysis was performed, and for each study variable we calculated ORs, 95% CIs and p values (two-tailed p≤0.05 significant level). Univariate logistic regressions were fit considering binomial ‘IR yes/no’ as a dependent variable and age, sex, disease duration, MS type, baseline EDSS and baseline NRS as independent variables. Whenever quantitative variables were dichotomised, the cut-offs were derived using the median. Multivariate analysis was performed to investigate the independent effect of a risk or protective factor after adjustment for one or several other factors or to adjust for confounding variables. Parameters associated with the outcome in the univariate analysis with a threshold of p=0.10 were included in the multivariate model.

In this large multicentre study, we found that 70.5% of patients reached the IR threshold and that 28.2% had already reached the CRR threshold at the end of the 4-week trial period. According to the physician's judgement, 65.4% of patients were considered responders after the first month of treatment, aligning with the proportion of patients reaching IR (70.5%), whereas 9.3% were considered partial responders, suitable to become responders in a longer follow-up.

We also found in patients reaching T1 (n=1432) a 22.6% mean NRS spasticity score reduction at T1 (first month) and a 35% mean NRS score reduction at T3 (6 months) compared with baseline, showing a decrease in NRS score from 7.5 at baseline to 5.8 at T1 and 4.8 at T3. Considering the IR population, we found a mean 30.3% reduction at T1 and a 33.3% reduction at T3. Considering the CRR group, we found a 40.8% reduction at T1 and 43.4% reduction at T3. These data suggest that a meaningful MS spasticity NRS reduction is obtained during the first month and that it is maintained or slightly increased over time (see the T3 6-month follow-up data). This is in line with previous observational studies.16 19 20 It is noteworthy that the mean number of puffs decreased slightly from 6.8 at T1 to 6.3 at T3, suggesting that effectiveness may be maintained with lower doses compared to clinical trials (mean number of 8.1 puffs/die in clinical trials),12 while ruling out the need to increase dose to maintain effect over time (tolerance) and even abuse/misuse trends.

A certainly higher proportion of patients in our study reached the IR ≥20% threshold in the first month (70.4%) when compared with patients from the German MOVE 2 study (41.7%). However, in the German study, clinicians reported observing clinical relief of resistant spasticity in 74.6% of patients in the first month of treatment according to their own judgement, in line with the results of our study (70.4%).16 In Italy The AIFA e-registry requires an IR threshold attained after the 4- week trial period to continue with the THC:CBD prescription, obtained through an automatic calculation and within a strict time frame. In other countries (Germany included), the NRS improvement is calculated by the clinician without the use of a compulsory web-based registry, thus possibly affecting the perceptions of the initial response. This is confirmed by another recent observational Italian study, showing that a proportion of 71.7% reached the IR with a mean NRS score reduction of 32%, in line with our results.20 The German MOVE 2 study IR patients' subgroup showed a mean reduction of NRS score of 40.2% at T1, while in our study a lower mean reduction, 30.3%, was observed. Considering the CRR ≥30% spasticity NRS change versus baseline threshold patients subgroup evolution, a somehow larger mean reduction of 49.9% was observed at T1 in Germany,16 while our results showed a mean reduction of 40.8% in this subgroup. In a phase 3 randomised placebo-controlled study, participants were also treated with THC:CBD as an add-on therapy in a single-blind manner for 4 weeks. After 4 weeks, 47.5% of patients reached IR, showing a MS spasticity mean NRS score reduction of 47.6% compared with baseline NRS, higher than the MOVE 2 observational study score of 41.7% but still lower than our 70.4% rate.12 Our results showed a higher proportion of IR responders but a smaller mean change in term of NRS between baseline and T1, not fully in accordance with previous studies, although in line with recent findings from an observational Italian study.20 This could in part be explained by the AIFA web registry features, not allowing an NRS score <4 at baseline and always requiring the IR threshold to be reached at the trial period follow-up visit, thus affecting the recording of the improvement. Indeed, our population seems to be more impaired, in terms of mean EDSS and NRS scores, and has a longer disease duration, when compared with other studies (see table 4). This could in part be explained by multivariate analysis findings showing that reaching the trial period IR threshold could be influenced by a baseline NRS score and disease course (progressive MS). The effectiveness of Sativex on patients with progressive MS spasticity has been investigated by a recent study evaluating clinical and neurophysiological effects of treatment. Although no effect on a neurophysiological measure (H reflex) was reported, the study confirmed the clinical benefit (as per the Modified Ashworth scale) in patients with progressive MS, affected by lower limb spasticity.21

This finding may in part be related to the higher percentage of SP and PP compared to relapsing remitting in our study population (80% PP+SP), suggesting that patients with progressive MS are more prone to have a spasticity not responding to common anti-spastic drugs, requiring Sativex treatment. Moreover, we found that patients with progressive MS are older, and have a longer disease duration and higher EDSS and NRS at baseline, suggesting that these differences might in part have biased our findings (see table 1). As in our cohort, it was recently reported that severe spasticity is more frequently observed in wheelchair users while patients able to walk 300–500 m reported a mild to moderate spasticity.7 On these grounds, it is conceivable that patients with a higher NRS score and those with progressive MS could be more satisfied with any improvement in their condition, while patients with lower EDSS scores and thus better walking ability are more demanding in terms of symptoms improvement. Indeed, we should bear in mind that the use of symptomatic treatment is logically mainly based on patients' reported outcomes, rather than on neurological signs examination. However, replication and further analyses are required to better clarify our findings.

About 40% of our population discontinued treatment, in particular 26.2% discontinued for lack of effectiveness and 18.7% for AEs. The most common AEs reported by patients were cognitive/psychiatric effects, fatigue and drowsiness, in line with the tolerability profile reported in other observational studies.14 16 19

Although this is the largest study reporting Sativex experience in daily clinical practice, our results may be affected by the observational nature of the study. This could in part be mitigated by the use of the AIFA prospective and mandatory e-registry as a source for spasticity NRS information and main patient features. Indeed, in the Italian e-registry, the evaluations are scheduled at 4 weeks from treatment start and then every 3 months, giving us the opportunity to collect data from every centre at the same time points. Another limitation is the lack of spasticity evaluation through the Ashworth healthcare professional impression categorical scale. Although the use of the Ashworth Scale for the assessment of spasticity could be not valid and reliable, results from the German MOVE 2 Study and a recent clinical trial indicated a small but significant improvement.16

Finally, the follow-up period duration (up to 6 months) may be insufficient to assess Sativex effectiveness, safety and tolerability long-term effects.

Besides the aforementioned limitations, this study is the largest multicentre observational study so far available evaluating Sativex effectiveness and tolerability profile in clinical practice. Our results confirm Sativex as an effective and safe option for patients with MS with moderate to severe spasticity resistant to common antispastic drugs. Longer and large multicentre follow-up studies are needed to evaluate treatment effectiveness and tolerability effect in the population with MS. Further exploitation of the AIFA Sativex e-registry data would be desirable.