Identification of risk factors for falls in multiple sclerosis: a systematic review and meta-analysis.

BACKGROUND: Falls are a significant issue in people with multiple sclerosis (MS), with research demonstrating fall rates of more than 50%. PURPOSE: The purpose of this study was to evaluate the risk factors associated with falling in people with MS. DATA SOURCES: Mixed search methods were used, including computer-based and manual searches. Additionally, hand searches of reference lists and conference abstracts were performed. All literature published from the source's earliest date to January 2012 was included; only full-text English-language sources (or those where a translation was available) were included. STUDY SELECTION: Eligibility criteria specified articles evaluating any aspect of fall risk in adults with a confirmed MS diagnosis, where the incidence of falling as determined by prospective or retrospective participant report was included. DATA EXTRACTION: Data were extracted independently by 2 reviewers using a written protocol and standardized extraction documentation. Detailed assessment of each article was independently undertaken by both reviewers, including assessment of study quality using an adaptation of the Newcastle Ottawa Scale plus extraction of key data (participant characteristics, fall incidence, and outcomes). DATA SYNTHESIS: The final review comprised 8 articles with a total of 1,929 participants; 1,037 (53.75%) were classified as fallers. Eighteen different risk factors were assessed within the included studies. Meta-analysis demonstrated an increase in fall risk associated with impairments of balance and cognition, progressive MS, and use of a mobility aid. Narrative review of the qualitative articles and those factors where meta-analysis was not possible also was undertaken. LIMITATIONS: Variation in assessment, analysis, and reporting methods allowed meta-analysis for only 4 factors. CONCLUSION: There is limited evidence of the factors associated with fall risk in people with MS. Further methodologically robust studies are needed.