Objective To examine whether physical measurements of the sensorimotor system predict concussion incurred by footballers.

Design Prospective cohort study.

Setting Australian professional sport setting.

Participants Fifty-three rugby union and rugby league players.

Assessment of risk factors Physical measurements including balance (sway across six test conditions), vestibular function (assessments of oculo-motor and vestibular ocular reflex) and spinal control (cervical joint position error and trunk muscle size and contraction ability) were performed during the preseason.

Outcome measures Concussion injuries during the season were diagnosed and recorded by club medical staff.

Results Age, height, weight, and years playing football were not associated with concussion in the season (p=0.43, 0.43, 0.19, 0.25). Preseason cross-sectional area of the multifidus muscle at the L5 vertebral level was 10% smaller in players who went on to sustain a season concussion compared to those who did not: mean difference (95% confidence interval) 1.0 cm² (0.1 cm² to 1.9 cm²), p=0.031. This difference in size was moderate (effect size 0.7) and exceeds the minimal detectable change of 0.4 cm². There were no significant differences observed for the other measures, p>0.05.

Conclusions Optimal spinal control is likely to be a key requirement in a physically demanding sport such as football. The multifidus is one of the muscles crucial for spinal control, and our findings report this measure to be predictive of footballers sustaining a concussion. This novel finding provides preliminary evidence that warrants further investigation as to the possible importance of spinal control in prevention programs

Competing interests None.