Pacing Variation in Multistage Ultramarathons: Internet-Based Cross-Sectional Study

Background: Ultramarathon running is the most popular ultraendurance competition in terms of the number of races and runners competing annually worldwide; however, no study has compared pacing and performance over a long period. Objective: This study analyzes the pacing of successful finishers and nonfinishers in multistage ultramarathons worldwide. Methods: A total of 4079 athletes (men=3288; women=791) competing in 99 multistage ultramarathon events from 1983 to 2021 were analyzed, including the number of participants, age, gender, rank, and running speed of successful finishers. Results: The results showed a significant increase in the number of events (n=338) and a significant increase in the number of finishers and nonfinishers (n=5575) in the ultramarathons worldwide during this period. The general linear models (GLMs) of pacing variation showed nonsignificant effects for gender (F-1,F-36.2=2.5; P=.127;eta(2)(p)=0.063) and age group (F-10,F-10=0.6; P=.798;eta(2)(p)=0.367), but it showed a significant interaction (gender x age) effect (F-10,F-2689=2.3; P=.008;eta(2)(p)=0.009). Post hoc analyses showed that men have a higher pacing variation than women in the under 30 years (U30), U35, U45, and U50 groups. Additionally, the fastest women's age group (U35) had the lowest pacing variation. The GLM of pacing variation by gender and event distance showed significant effects for both gender (F-1,F-3=18.5; P<.001;eta(2)(p)=0.007) and distance (F-2,F-3=20.1; P<.001;eta(2)(p)=0.015). Post hoc analyses showed a growing pacing variation with increasing race distance for both men and women. In addition, men had a higher variation in long events. Furthermore, there was a significant main effect for both genders (F-1,F-3=33.7; P<.001;eta(2)(p)=0.012) and rank (F-1,F-3=136.6; P<.001;eta(2)(p)=0.048) on performance, with men being faster than women. Pacing varied greatly due to gender (F-1,F-3=4.0; P=.047;eta(2)(p)=0.001), with a lower (ie, more even) pacing variation for male athletes in the top 3 finishers. Male nonfinishers showed a higher performance than female nonfinishers (F-1,F-1340=25.6; P<.001), and no difference was identified for pacing variation (F-1,F-789=1.5; P=.228) based on gender. In addition, a weak but significant correlation (r=-0.130; P<.001) was identified between the average running speed and pacing variation for both female and male nonfinishers. Conclusions: In summary, multistage ultramarathon competitions showed an increasing number of competitors and a higher performance challenge. Men have a higher pacing (ie, less even) variation than women, especially observed in longer events. A higher pacing variation was associated with lower performance for men, women, and nonfinishers.