Oral antispastic drugs in nonprogressive neurologic diseases

Search strategy.

Studies were identified by various search strategies in MEDLINE, PubMed, and Cochrane Library. The search included combinations of the following MeSH terms: baclofen, tizanidine, dantrolene, benzodiazepines, diazepam, clonidine, muscle spasticity, and free text words as gabapentin. The last search was carried out in January 2004. Additional studies were identified from the reference list of retrieved reports. No contacts were made with pharmaceutical manufacturers and authors of articles for unpublished reports or additional information from published reports.

Selection criteria.

Randomized clinical trials (RCTs) of antispastic drugs in nonprogressive neurologic disease were sought. In order to limit clinical heterogeneity, the review focused on patients with nonprogressive neurologic disease. Language was restricted to English, French, German, and Spanish. Reports of clinical trials were considered if patients’ allocation was randomized, spasticity was due to nonprogressive neurologic diseases (spinal cord injury, cerebral palsy, and cerebrovascular disease), and the experimental drug was given by oral route. Only full publications were considered. Data from case reports, abstracts, and letters were excluded. Reports were also excluded if the study was not double-blind, when sample size was less than 10 patients, when patients with spasticity had MS, and when the experimental agent was administered by intrathecal or other parenteral routes (mainly baclofen and botulinum toxin A) (figure).

• Download figure
• Open in new tab
• Download powerpoint

Figure. Selection of randomized clinical trials (RCTs) for the systematic review.

Data extraction and study characteristics.

The following information was extracted from each report meeting the inclusion criteria:

1. Study characteristics: study design, sample size, duration of the observation period, outcome measures for spasticity, and efficacy and safety outcomes.

2. Methodologic quality: methodologic quality was assessed based on the items of the Jadad scale: randomization, double blinding, and reporting of withdrawals and dropouts.9 The scoring range is 0 to 5.

3. Patients’ characteristics: age, sex, and underlying neurologic disease.

4. Pharmacologic treatments: drug, time of administration, doses and titration, treatment period, number of treatment arms, control with placebo, and washout period.

5. Efficacy outcomes: scales for muscle tone (Ashworth scale),10 spasticity grade or accompanying symptoms, functional movements, scales assessing activities of daily living, and clinical evaluation by patients, their family, or health staff.

6. Tolerability outcomes: rates and types of adverse events.

7. Dropouts and withdrawals: rates and reasons.

Statistical analyses.

Descriptive statistical analyses (rates, means, and standard deviations) were carried out using the SPSS version 9.0 software package.

Trial flow.

A total of 325 articles were identified, of which 101 reported clinical trials. Eighty-nine reports were excluded because they described studies in patients with MS (38), the route of administration was not oral (31), the antispastic effect was not an outcome variable (8), less than 10 patients had been included in the trial (6), language other than those included (2), or they were not randomized or double-blind clinical trials (4), thus leaving 12 reports meeting the inclusion criteria (see the figure).11–22⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓⇓ The included reports were published between 1972 and 2001.

Clinical trials included.

A total of 469 patients were included in the 12 selected RCTs. Most of the trials (7, 58.3%) were crossover with a washout period of 7 to 11 days, except one where there was no washout.11 As shown in table 1, six RCTs (50%, 228 patients) assessed spasticity in patients with cerebrovascular disease or head trauma, three (25%, 174 patients) in patients with spinal cord injury, and three (25%, 67 patients) in children with cerebral palsy. Experimental drugs were tizanidine (four trials, 276 patients, 142 exposed to the drug), dantrolene (four trials, 103 patients, 93 exposed), baclofen (three trials, 70 patients, 55 exposed), diazepam (two trials, 127 patients, 76 exposed), and gabapentin (one trial, 28 patients, all exposed).

Ten RCTs compared an antispastic drug to placebo, and two trials compared tizanidine to other antispastic drugs (diazepam or baclofen). No trials with clonidine were found.

Sample size calculations were not reported in any trial (table 2). The randomization method was described in only one report.16 The method of blinding was described in eight reports, and the number of dropouts or withdrawals was specified in all of them. The overall Jadad score of the 12 trials was between 3 and 4.

Patients’ age and sex varied according to the underlying condition. Patients with cerebral palsy (67) were children (44.3% male), those with spinal cord injury (174) were middle aged (mean of 40 years old) and predominantly men (88%), and those with cerebrovascular lesions or stroke (228 patients) were older (mean of 55 years old) and also predominantly men (64.7%). The mean weighed duration of follow-up was 97.5 days (range, 17 days to 12 months). Treatment duration was between 2 and 6 weeks, except in one trial (30 patients) in which treatment was maintained during 30 weeks18 and another (28 patients) where the treatment period was only 48 hours.16 The mean treatment period (not including titration phase, tapering period, or washout period) for each drug was 84 days for baclofen, 80 days for tizanidine, 37 days for dantrolene, 25 days for diazepam, and 2 days for gabapentin. Most of the trials included a 2- to 3-week dose titration phase depending on the individual response.

Results on efficacy.

Details of the studies are included in tables 3 through 6⇓⇓⇓. The efficacy outcomes were heterogeneous. The mean number of efficacy outcome variables was 4.3 per trial (range 1 to 8). The Ashworth scale was used and reported in only six trials (50%, 154 patients in parallel group studies and 85 patients in crossover studies). Other common spasticity outcomes measured and scored were muscle tone, spasms, clonus, stretch reflexes, and muscle strength. In six studies (50%, 126 patients in parallel group studies and 101 in crossover studies) functional movement was assessed (mainly walking capacity). In seven studies (58.3%, 163 patients in parallel group studies and 38 in a crossover study) scales of daily living activities were used. In seven studies (58.3%, 259 patients in parallel group studies and 126 in crossover studies) the clinical outcome measure was overall evaluation by the patient, a relative, or a health professional (physician, physiotherapist, or nurse). Only in two trials was the analysis by intention to treat (see table 2).

Adverse effects, dropouts, and withdrawals.

All reports gave information on adverse events in the different treatment arms, and their rates, except two12,21⇓ (see tables 3 through 6⇑⇑⇑). Adverse effects were generally more frequent in the active treatment groups (range 25 to 91%) than in the placebo groups (range 0 to 53%). Adverse events associated with tizanidine (range 60 to 88% of patients treated) were mainly drowsiness and dry mouth, and less frequently increases in liver enzymes. Dantrolene was associated with a high incidence of adverse events (range 64 to 91%), mainly weakness, drowsiness, and malaise. Diazepam (range 27 to 72% of patients) was associated with drowsiness, fatigue, and muscle weakness. Baclofen (range of adverse events of 25 to 27% of patients) was associated with sedation, dizziness, and muscle weakness. Neurologic events seemed to be dose-related, and they tended to disappear when doses were reduced.

The mean rate of dropouts and withdrawals was 21.2% (range 11 to 38%), and in two trials no dropouts were reported. Of 106 withdrawals and dropouts, 56 (53%) were due to adverse events. No association was found between treatment duration and the rate of withdrawals due to adverse events.