RESUMO
We aimed to perform a comprehensive systematic review of the existing ataxia scales. We described the disorders for which the instruments have been validated and used, the time spent in its application, its validated psychometric properties, and their use in studies of natural history and clinical trials. A search from 1997 onwards was performed in the MEDLINE, LILACS, and Cochrane databases. The web sites ClinicalTrials.gov and Orpha.net were also used to identify the endpoints used in ongoing randomized clinical trials. We identified and described the semiquantitative ataxia scales (ICARS, SARA, MICARS, BARS); semiquantitative ataxia and non-ataxia scales (UMSARS, FARS, NESSCA); a semiquantitative non-ataxia scale (INAS); quantitative ataxia scales (CATSYS 2000, AFCS, CCFS and CCFSw, and SCAFI); and the self-performed ataxia scale (FAIS). SARA and ICARS were the best studied and validated so far, and their reliability sustain their use. Ataxia and non-ataxia scores will probably provide a better view of the overall disability in long-term trials and studies of natural history. Up to now, no clear advantage has been disclosed for any of them; however, we recommend the use of specific measurements of gait since gait ataxia is the first significant manifestation in the majority of ataxia disorders and comment on the best scales to be used in specific ataxia forms. Quantitative ataxia scales will be needed to speed up evidence from phase II clinical trials, from trials focused on the early phase of diseases, and for secondary endpoints in phase III trials. Finally, it is worth remembering that estimation of the actual minimal clinically relevant difference is still lacking; this, together with changes in quality of life, will probably be the main endpoints to measure in future therapeutic studies.