Traumatic spinal cord injury (SCI) is a devastating event with huge personal and societal costs. A limited number of treatments exist to ameliorate the progressive secondary damage that rapidly follows the primary mechanical impact. Mesenchymal stem or stromal cells (MSCs) have anti-inflammatory and neuroprotective effects and may thus reduce secondary damage after administration. We performed a systematic review with quantitative syntheses to assess the evidence of MSCs versus controls for locomotor recovery in rat models of traumatic SCI, and identified 83 eligible controlled studies comprising a total of 1,568 rats. Between-study heterogeneity was large. Fifty-three studies (64%) were reported as randomised, but only four reported adequate methodologies for randomisation. Forty-eight studies (58%) reported the use of a blinded outcome assessment. A random-effects meta-analysis yielded a difference in behavioural Basso-Beattie-Bresnahan (BBB) locomotor score means of 3.9 (95% confidence interval [CI] 3.2 to 4.7; P<0.001) in favour of MSCs. Trial sequential analysis confirmed the findings of the meta-analyses with the upper monitoring boundary for benefit being crossed by the cumulative Z-curve before reaching the diversity-adjusted required information size. Only time from intervention to last follow-up remained statistically significant after adjustment using multivariate random-effects meta-regression modelling. Lack of other demonstrable explanatory variables could be due to insufficient meta-analytic study power. MSCs would seem to demonstrate a substantial beneficial effect on locomotor recovery in a widely-used animal model of traumatic SCI. However, the animal results should be interpreted with caution concerning the internal and external validity of the studies in relation to the design of future clinical trials.