Background Training errors are considered as one of the main causes of running-related injuries, although few of them have been identified. Time-to-event analysis is a useful tool for researchers within sports injury prevention. In running, the time scale is commonly defined as kilometres or hours of practice. However, using calendar days as time scale could also be relevant.

Objective To determine the influence of the time scale on the association between cumulated running distance and injury risk.

Design Observational study with a 6-month follow-up.

Setting Leisure-time runners.

Participants Study participants (n=553) were required to report any running activity and injury using a dedicated internet platform.

Assessment of Risk Factors The main exposure was cumulated running distance over the week prior to a given session (time-dependent exposure) and was categorised as <20 km (reference group), between 20 and 30 km, and ≥30 km. Cox regression analyses were used to compare injury rates between exposure states, using either kilometres or calendar days as time scale.

Main Outcome Measurements First running-related injury based on a 1-day time loss definition.

Results A total of 93 of the runners reported at least one injury. When the time scale was kilometres at risk, the results revealed injury rate to be lower when the participants cumulated a higher running distance. Hazard ratios (HR) were respectively 0.52 (95% confidence intervals −95% CI=0.30–0.89) and 0.40 (95% CI=0.23–0.70) in the 20–30 km and ≥30 km exposure states compared with the reference. When the time scale was calendar days, these HR were respectively 1.38 (95% CI=0.80–2.38) and 1.83 (95% CI=1.12–3.00) for the same exposure states, which suggests that injury risk was higher when more kilometres were cumulated.

Conclusions The choice of time scale revealed highly contrasting results, and has to be carefully considered in studies examining risk factors for running-related injuries.