Methods A systematic literature search of four databases was conducted up to November 2014. Prospective studies investigating the relationship between rotational traction and injury rate were included. Two researchers independently extracted outcome data and assessed the quality of included studies using a modified Downs and Black index. Effect sizes (OR+95% CIs) were calculated using RevMan software. Where possible, data were pooled using the fixed effect model.

Results Three prospective studies were included (4972 male athletes). The methodological quality was generally good with studies meeting 68–89% of the assessment criteria. All studies categorised athletes into low (lowest mean value 15 nm) or high traction groups (highest mean value 74 nm) based on standardised preseason testing. In all cases, injury reporting was undertaken prospectively over approximately three seasons, with verification from a medical practitioner. Injury data focused on: all lower limb injuries, ankle/knee injuries or ACL injury only. There was a clear relationship between rotational traction and injury and the direction and magnitude of effect sizes were consistent across studies. The pooled data from the three studies (OR=2.73, 95% CI 2.13 to 3.15; χ²=3.19, df=2, p=0.21; I²=36.5%) suggest that the odds of injury are approximately 2.5 times higher when higher levels of rotational traction are present at the shoe-surface interface.

Summary and conclusions Higher levels of rotational traction influence lower limb injury risk in American Football athletes. We conclude that this warrants considerable attention from clinicians and others interested in injury prevention across all football codes.