We conducted a comprehensive search using MEDLINE/PubMed, Ovid, and CENTRAL (Cochrane Central Register of Controlled Trials) for English-language only literature published from 1999 to April 2009 using different combinations of the following MeSH and free text terms: cannabis, cannabinoid, THC, CBD, multiple sclerosis, spasticity, spasms. Reference lists from retrieved reports were reviewed for additional studies. Unpublished data were not sought and abstracts, letters, case reports, and review articles were excluded. (See Additional file 1 for a Quality of Reporting of Meta-analyses (QUOROM) statement checklist.)

Only randomized, placebo-controlled, human studies of shorter treatment periods (under 6 months) were included. Studies were evaluated for methodological quality using Jadad scores 24 and only studies with Jadad scores of 4 or higher were considered for inclusion. Relevant trials included those that had administered a combination THC and CBD extract, those in which clinically stable spasticity had been established prior to trial and those that reported objective measures of pre- and post-treatment spasticity. Studies that used active control groups were not excluded. Abstracts were reviewed for relevancy and full text versions of potentially relevant randomized controlled studies were reviewed. Reports not considered relevant were excluded and all included reports were read in entirety.

Data were extracted independently by the authors and any disagreements were resolved by consensus. The following information was extracted from each report: study type, study objective, sample size, controls, type and amount of cannabinoid used, treatment duration, objective and subjective outcome measures, and reported adverse events.

A qualitative summary of the data was completed to compare the various outcome measures used across the included studies. In addition, a quantitative analysis of the one common outcome measure (Ashworth scale) used by the included studies was performed in order to assess statistical heterogeneity.

Electronic searches found 38 studies that were potentially relevant to the present review. Of these, 33 did not meet the inclusion/exclusion criteria, including 27 reports that were not randomized, controlled trials. One study was excluded for focusing on spasticity-related pain and two were excluded for not assessing the effects of a combined extract of THC and CBD. Two studies were excluded for reporting long-term follow-up data (see Figure 1).

Six double-blind, randomized, placebo-controlled trials published between 2002 and 2007 were analyzed 15 16 17 18 19 20 . These studies included a total of 481 patients with MS who were administered a combined extract of both THC and CBD. Three trials used a crossover design. Three trials used a parallel design in which 339 patients were administered a placebo only. Trial periods ranged from 2 to 15 weeks. Objective spasticity measures were extracted when included in the assessment data from at least two studies. All six trials reported an adjusted mean change score in the Ashworth scale assessment. Other measures of spasticity included mean changes from a baseline score in the following assessments: Visual Analogue Scale (VAS), a rating scale to measure the severity of spasticity; walk time; Rivermead Mobility Index (RMI), a measure of disability related to mobility; and self-reported ratings of spasm frequency or severity. (See Table 1 for the specific characteristics of each reviewed study.)

Three of the studies 15 19 20 did not report adequate (mean and standard deviation) Ashworth scale data for inclusion in the quantitative meta-analysis. This left three studies for the calculation of the pooled mean difference in Ashworth scores. The chi-square test for heterogeneity showed evidence of significant variation between the three studies (χ² = 5.25, P = .07, l² = 62%). Given that only three of the six studies reported adequate Ashworth scale data 16 17 18 , of these three only one demonstrated statistically significant findings 18 , and the high level of heterogeneity, a quantitative analysis of the data was deemed inappropriate.

There were some limitations to the systematic review. First, this review did not include unpublished data. There may be ongoing clinical trials of combination THC and CBD therapy as it is a relatively recent therapy. There is also the possibility that other clinical reports using whole plant cannabis extracts may have been appropriate for review, but were not included without report of specific methodology. A meta-analytical review of the effects of cannabis on spasticity would be useful, but was not deemed appropriate for the present review because of the variation in assessment data.

The validity of the Ashworth scale as an outcome measure has been previously questioned 16 . However, we have shown that other objective measures of spasticity (i.e. RMI, walk time) may also fail to adequately support the improvements found in scores from more subjective measures (i.e. rating scales, diary entries). A long-term follow-up study 21 showed a significant improvement in the Ashworth scale, however, the change was still small and was found in the THC group only. Another concern is that participants of both active and placebo trials may not be entirely blind to their treatment status 16 , and this may affect subjective assessments. It remains that, without a validated, objective measure of spasticity, it will be difficult to accurately measure the effects of cannabis therapy on MS spasticity.

Adverse effects were reported in each trial in which patients received active treatment (including THC-only treatment). There is some evidence that combined extracts of THC and CBD may attenuate side effects of THC alone, and future studies are needed to compare the safety of combined cannabis extracts with traditional treatment. Dosage is another concern that should be considered in the context of side effects. Incidence of side effects varies greatly depending on the amount of cannabis needed to effectively limit spasticity. In one study 17 , it was noted that the initially permitted dosage level sometimes resulted in marked side effects, and the dosage was thereafter reduced. The careful monitoring of symptom relief and side effects is critical in reaching an individual's optimal dose. Finally, it should be noted that several adverse events were also reported in each trial in which patients received a placebo. In a long-term follow-up 22 of one of the reviewed studies 19 , it was determined that most of the reported adverse events were unrelated to cannabis treatment. Considering the distress and limitations spasticity brings to individuals with MS, it would be important to carefully weigh the potential for side effects with the potential for symptom relief, especially in view of the relief reported in subjective assessment.