Measures of step variability and body sway during gait have shown to correlate with clinical ataxia severity in several cross-sectional studies. However, to serve as a valid progression biomarker, these gait measures have to prove their sensitivity to robustly capture longitudinal change, ideally within short time-frames (e.g. one year). We present the first multi-center longitudinal gait analysis study in spinocerebellar ataxias (SCAs). We performed a combined cross-sectional (n=28) and longitudinal (1-year interval, n=17) analysis in SCA3 subjects (including 7 pre-ataxic mutation carriers). Longitudinal analysis revealed significant change in gait measures between baseline and 1-year follow-up, with high effect sizes (stride length variability: p=0.01, effect size rprb=0.66; lateral sway: p=0.007, rprb=0.73). Sample size estimation for lateral sway reveals a required cohort size of n=43 for detecting a 50% reduction of natural progression, compared to n=240 for the clinical ataxia score SARA. These measures thus present promising motor biomarkers for upcoming interventional studies.