Marathon run performance on daylight savings time transition days: results from a natural experiment.

Advancing clock times by 1 h in the spring to daylight savings time and setting clock times back 1 h in the autumn to standard time disrupts circadian timing, sleep and skilled motor behavior such as driving an automobile. It is unknown if endurance performance is impacted by daylight savings transition (DST). The natural experiment described here examined whether exposure to a DST in the 10 h prior to the start of a marathon race was associated with a different mean completion time compared to participants who ran the same course but were unexposed to a recent DST. The primary outcome was the average running time of finishers of United States marathons that were completed on either spring-DST or autumn-DST days in the years 2000-2018. Comparisons were made to results from the same marathon held in a different year that was not run on a DST day. Data were obtained from the public data base marathonguide.com/results. Analysis of the primary outcome used paired samples t-tests weighted by sample size. Spring and autumn data were analyzed separately. Eighteen spring and 29 autumn marathons met the inclusion criteria. Compared to control marathons, the weighted spring-DST performance was worse by 12.3 min (4.1%; P < .001) and equal to a moderate standardized effect size of 0.57 while autumn-DST was trivially worse by 1.4 min (0.5%), which was equivalent to an effect size of 0.13. Ambient temperatures for the DST and control races did not differ for either the spring (10.6 vs. 8.9℃; P = .212) or autumn marathons (7.6 vs. 9.3℃; P = .131). Within the limitations of a natural experiment research design, it is concluded that the findings support worse running performance in marathon races held in the spring on the day of transition to daylight savings time when there is a forced circadian change and sleep loss.