Introduction
Few studies have examined the foot strike pattern of long-distance runners under competition conditions (11,17,18,20,29). The more recent studies document a high prevalence of rear-foot strike (RFS) pattern (11,17,20). Among an elite sample collected at a competitive half-marathon, an RFS pattern was observed in 74.9% of the runners (11). Among more diverse, mostly amateur samples collected from midsized city marathons, an RFS pattern was observed among 88.9–93.7% of the runners (17,20).
Rear-foot strike produces a significant impact transient, whereas forefoot strike (FFS) and midfoot strike (MFS) produce no impact transient (21). The impact transient, or sudden force of loading, is distributed across the runner's bones, muscles, and joints; and has been suggested to be a significant predictor of injury (1,2,10,25,32,33). Nevertheless, there is considerable controversy about the benefits of different foot strike patterns (19).
A theoretical mechanism by which long-distance runners may avoid the cumulative effect of impact transients is by adopting a FFS or MFS pattern (23,38). Marathon runners have not demonstrated this avoidance mechanism (20). Ultramarathon runners, however, are exposed to an even greater number of foot strikes, resultant impacts, and the subsequent cumulative effects. Thus, ultramarathon runners may be more likely to adopt a FFS or MFS pattern to reduce the cumulative adverse effects from impact during foot strike. However, fatigue may also influence foot strike pattern as RFS prevalence has been observed to increase in marathoners through race progression (20). This shift from a FFS or MFS pattern to an RFS pattern may represent the result of fatigue of the plantar flexor muscles, whether because of greater eccentric loading with FFS or MFS patterns (7,37) or from higher preactivation of the plantar flexor muscles (23). To date, no studies have examined foot strike pattern during an ultramarathon to provide insight into whether fatigue or avoidance of impact affects foot strike patterns.
Thus, the primary objectives of this study were to: (a) examine changes in foot strike pattern, stride rate, and stride length during a 161-km mountain ultramarathon; (b) determine if foot strike pattern is related to performance; and (c) ascertain if post-race blood creatine phosphokinase (CK) concentrations differ among those with a FFS or MFS pattern compared with an RFS pattern. We hypothesized that, in an effort to decrease the cumulative effect of impact transients, the RFS prevalence among ultramarathon runners in a 161-km mountain trail run will be less than previously documented for shorter road races. We further theorized that foot strike pattern will not be related to performance, but that post-race blood CK concentrations may be increased among those using a non-RFS pattern because of greater muscular demand from this foot strike pattern. If this work supports the latter theory, then an observation of the manner in which ultramarathon runners balance the need to minimize impact forces with that of minimizing muscular demands is of interest.
Methods
Experimental Approach to the Problem
To determine foot strike pattern, stride rate, and stride length and the possible relationship with performance and the post-race blood CK concentration in ultramarathon runners, we videotaped the 2012 Western States Endurance Run (June 23–24, 2012). High-frame rate video capture was obtained for each runner at 3 level sites and 1 downhill site during the 161-km ultramarathon. Post-race blood was collected from runners for analysis of post-race blood CK concentration. Overall RFS prevalence, stride rate, and stride length were determined by video analysis. Race performance and post-race blood CK concentration were compared with foot strike patterns to determine if either variable was associated with a specific foot strike pattern, and therefore, could be considered as a possible training modification to either improve performance or decrease muscle damage during a 161-km ultramarathon.
Subjects
In brief, the Western States Endurance Run is a 161-km point-to-point ultramarathon mostly run on single-track trails through the Sierra Nevada mountains. At the time of the study, it was the second largest 161-km ultramarathon run in North America averaging over 400 accepted entries each year. The race has a total climb of 5,500 m, total descent of 7,000 m, and altitude ranging from around 1900 to 2667 m. Nearby temperature ranged from 9 to 22° C during the 2012 event. Further details of the race are provided elsewhere (15,16). All race participants were experienced ultramarathon runners since the minimum requirement for race entry is prior completion of at least an 80-km ultramarathon in <11 hours. All registered race participants were sent a prerace email briefly describing the research being conducted at the race. The study was approved by the Institutional Review Board at the VA Northern California Health Care System with waiver of consent. None of the subjects were younger than 23 years.
Procedures
All race participants were videotaped at 3 level locations and 1 downhill location. The level sites were at 16.5, 90.3, and 161.1 km (250 m from the finish). The downhill site was at 90.7 km. The level sites were chosen so that data could be collected from near the early beginning, midpoint, and end of the race. The downhill site was chosen for its long consistent grade near the midpoint of the race. Considerations were also given to avoiding locations immediately before or after aid stations. The first site was 420 m before an aid station and the second site was 660 m after an aid station. Each site was located in an area where there was no significant deviation in terrain or direction of travel within 100 m before or after the site. In addition, each site was without significant terrain irregularity. Filming zones were 1 m wide and 4 m long and were identified by cones arranged to appropriately direct the runners. The final filming zone was on an all-weather synthetic track. Other filming zones were on hard-packed dirt that was prepared by removal of irregularities and compacted to ensure the surface was consistent and smooth. The level sites were verified to be level and the angle of the downhill was determined to be 9%. Approximately 100 m before each filming zone, signs were placed advising the runners to run single file through the filming zone. The dirt sites were lightly sprayed with water as necessary to minimize dust.
Video recording was performed with frame rate set at 240 frames per second (Casio EX-ZR200, Tokyo, Japan). It was identified that filming rate with this camera model decreased to 194 frames per second during the initial 30 seconds of each video recording following a consistent time course. Appropriate correction for this effect was made during data processing of the time-specific variable (stride rate). Cameras were oriented perpendicular to the filming zone 10 cm above the filming zone and offset 1.9 m from the middle of the filming zone. The appropriate offset was determined from prerace trials with the intent of capturing 3 foot strikes for the majority of runners. Video calibration was completed at each location intermittently during data collection with a 2.4 m calibration stick marked at 20-cm increments placed in the middle of the filming zone. Supplemental lighting was used as necessary during time periods with low lighting. Limits in the number of cameras prevented videotaping of the slowest runners at the downhill site.
Race numbers were sequentially recorded at each site to match video data with the specific runner. Video review and placement order at the nearest aid station or race finish were used to verify correct runner identification.
Race finishers were invited to provide blood samples immediately after finishing. The blood was collected in heparinized tubes within minutes of race finish by professional phlebotomists through venipuncture with the subject seated. Samples were centrifuged within 30 minutes and stored in an ice chest with cool packs until reaching a clinical laboratory where analyzed for CK concentration (Beckman Coulter LX20; Beckman Coulter, Brea, CA, USA).
Frame-by-frame video analysis of foot strike type was completed independently by 2 researchers. Foot strike type was defined as proposed by Lieberman et al. (21): RFS as the heel landing before the ball of the foot (heel-to-toe), FFS as the ball of the foot landing before the heel (toe-heel-toe), and MFS as simultaneous landing of the heel and ball of the foot. Foot strike type was determined for 3 foot strikes in 84.8% of cases, 2 foot strikes in 14.9% of cases, and 1 foot strike in 0.3% of cases. Based on these foot strike types, each runner was classified into one of the following foot strike “patterns” for each filming site: FFS, MFS, non-RFS (a mixture of FFS and MFS), RFS, mixed RFS/non-RFS (a mixture of RFS and either FFS or MFS), or unclassified. Unclassified was defined by either insufficient video capture (including runners captured on video who did not run through the filming zone) or absence of a double float phase, which has been defined as the demarcation between running and walking by Novachek (30). Runners who lacked a double float phase, and therefore had initial contact of 1 foot strike occurring before toe off of the preceding foot strike, are subsequently referred to as “shufflers.” In the event of disagreement on foot strike pattern between the 2 investigators, the video was reviewed by both researchers with an effort to reach consensus. When consensus was not achieved (only 2.6% of foot strike patterns, which was most often a disparity about a single foot strike), a third researcher reviewed the video to make the final determination.
Initial contact time and location were determined for 3 foot strikes (when possible) through the frame-by-frame video analysis. Initial contact time was defined as the frame of first ground contact. Location was based on the back of the shoe when this could be visualized for 3 consecutive foot strikes. Alternatively, the tip of the shoe was used if this allowed for visualization of 3 foot strikes. Stride rate and stride length were then calculated based on 3 foot strikes. In the event where only 2 foot strikes could be visualized, stride rate and length were estimated from the single step.
Runner sex, age, finish time, and finish place were obtained from official race results. These data were compiled with foot strike pattern, stride rate, and stride length for each runner at each filming location.
Statistical Analyses
To determine foot strike change pattern, foot strike patterns were compared between each pair of level sites and between the level and downhill sites at ∼90 km using McNemar's tests. Runner sex, age, finish time, post-race blood CK concentration, stride rate, and stride length were compared across the different foot strike patterns at each filming location. Runner sex, age, finish time, and post-race blood CK concentration were also compared among foot strike groups when collectively considering the level sites. For the collective analyses, the term foot strike group is used hereafter and is defined by the foot strike pattern across multiple sites. Specifically, 3 foot strike groups were composed: RFS, non-RFS (MFS, FFS, and non-RFS), and mixed striking. The mixed striking group for the collective analyses consisted of any combination of foot strike patterns across the sites being analyzed that included a mixed RFS/non-RFS pattern or any mixture of FFS, MFS, or non-RFS patterns with RFS pattern. Age, finish time, post-race blood CK concentration, stride rate, and stride length were analyzed with the D'Agnostino and Pearson omnibus normality tests. If data passed normality testing, further analysis was completed with 1-way analysis of variance (ANOVA) tests followed by Bonferroni's multiple comparison posttests. If data did not pass normality testing, an attempt at normalizing the data was made. If data could not be successfully normalized, further analysis was completed with Kruskal-Wallis tests followed by Dunn's multiple comparison posttests. Runner sex was analyzed with χ2 tests. Further analysis of post-race blood CK concentration at the level sites was completed with the Mann-Whitney test by merging the non-RFS and mixed striking groups to compare with the RFS group.
The relationship of foot strike pattern with finish place was examined for each level site with χ2 tests. In these analyses, the number of runners with each foot strike pattern was compared between top finishers and the remaining finishers in which usable data were obtained across all level sites.
Stride length and stride rate were analyzed across the 3 level sites with Kruskal-Wallis tests and Dunn's multiple comparison posttests as neither passed the D'Agnostino and Pearson omnibus normality test and could not be successfully normalized. These variables were compared between the level and downhill sites at ∼90 km with 2-tailed Wilcoxon matched pairs signed rank tests. Stride length and stride rate were also examined relative to finish place among those with usable data across the level sites (by groups of 20) with 2-way (finish place group by location) repeated measures ANOVA and Bonferroni posttests. Stride length and stride rate across sites were further examined among the top 20 runners by 1-way repeated measures ANOVA and Bonferroni posttests. Stride length and stride rate among finish places 21–240 were analyzed by Kruskal-Wallis tests and Dunn's multiple comparison posttests, as these data did not pass the D'Agnostino and Pearson omnibus normality test and could not be successfully normalized.
Chi-square analyses were used to compare foot strike patterns at each level site with prior published data from the 10 and 32 km sites of Larson et al. (20) and the 8 km site of Kasmer et al. (17) during road marathons. For these analyses, our foot strike patterns were reorganized into those with RFS pattern and those with FFS or MFS pattern. The “asymmetrical” category used in the prior publications was excluded from the analyses. Our foot strike data from the 16.5 km site were also compared with the findings at 15 km during a half-marathon reported by Hasegawa et al. (11) For this analysis, FFS, MFS, and RFS patterns could be directly compared between studies and our mixed RFS/non-RFS and non-RFS groups were excluded from the analysis. Statistical significance was set at p ≤ 0.05 for all analyses.
Results
There were 382 race starters (18.1% women) and 316 finishers (16.5% women). Finish times ranged from 14.77 to 29.97 hours. Of the 382 race starters, 373 were classified by foot strike pattern and 9 were unclassified (5 walkers, 3 shufflers, and 1 insufficient video) at the 16.5 km site. At 90.3 km, 282 were classified and 66 were unclassified (30 walkers, 26 shufflers, and 10 insufficient video). At 90.7 km, 191 were classified and 4 were unclassified (2 shufflers and 2 insufficient video). At 161.1 km, 280 were classified and 36 were unclassified (13 walkers, 19 shufflers, and 4 insufficient video).
The prevalence of foot strike pattern and sample size at each site is detailed in Table 1. Of the runners classified by foot strike pattern, an RFS pattern was used by 79.9, 89.0, and 83.9% of the runners at 16.5, 90.3, and 161.1 km, respectively. An RFS pattern was used by 84.8% of the runners at the downhill site.
;)
Table 1: Foot strike pattern and sample size at each site.*
Foot strike pattern changed significantly between the 16.5 and 90.3 km sites (n = 246, p = 0.04) and between the 90.3 and 161.1 km sites (n = 225, p = 0.03). Comparison of the 16.5 and 90.3 km sites revealed 23 runners had changed foot strike pattern between these sites: 17 runners (73.9%) were classified as non-RFS at 16.5 km and RFS at 90.3 km, whereas 6 runners (26.1%) were classified as RFS at 16.5 km and non-RFS at 90.3 km. Comparison of the 90.3 and 161.1 km sites revealed 13 runners had changed foot strike pattern between these sites: 11 runners (84.6%) were classified as RFS at 90.3 km and non-RFS at 161.1 km, whereas 2 runners (15.4%) were classified as non-RFS at 90.3 km and RFS at 161.1 km. Foot strike pattern did not differ significantly between the 16.5 and 161.1 km sites (n = 240, p = 0.5). Comparison of the 16.5 and 161.1 km sites revealed 22 runners had altered foot strike pattern between these sites: 9 runners (40.9%) were classified as RFS at 16.5 km and non-RFS at 161.1 km, whereas 13 runners (59.1%) were classified as non-RFS at 16.5 km and RFS at 161.1 km. Foot strike pattern was also not significantly different between the level and downhill sites at ∼90 km (n = 163, p = 0.8). Comparison of the 90.3 km level site and 90.7 km downhill site revealed 10 runners had altered foot strike pattern between these sites: 5 runners (50.0%) were classified as RFS at the level site and non-RFS at the downhill site, whereas 5 runners (50.0%) were classified as non-RFS at the level site and RFS at the downhill site.
Comparison of runner characteristics among the different foot strike patterns grouped by site is included in Table 2. There was no difference in proportion of men and women based on foot strike pattern or group when considering filming locations separately or collectively. Age did not differ based on foot strike pattern or group when considering the sites separately or collectively, except at the 16.5 km site where runners with non-RFS were older (p < 0.05) than those with RFS. Finish time did not differ relative to foot strike pattern.
Table 2: Characteristics of runners in each foot strike group at each site.*
Stride rate (p = 0.005) and stride length (p = 0.047) showed significant main effect differences across foot strike patterns at 16.5 km but not at other sites. Posttests revealed stride rate was significantly greater (p < 0.01) for non-RFS than RFS, but no pairwise comparisons were significant for stride length.
A significant difference in post-race blood CK was found across foot strike patterns at the 90.3 km site (p = 0.007) with posttest analysis revealing a greater post-race blood CK (p < 0.05) for non-RFS than RFS. A significant difference in post-race blood CK across foot strike groups was also present when the 90.3 km data were combined with data from either the 16.5 km site (p = 0.03) or the 161.1 km site (p = 0.03), but not when all 3 level sites were considered collectively. There was no significant difference in post-race blood CK across foot strike patterns at the 90.7 km downhill site.
Among the runners with usable foot strike data across all level sites (n = 243), a significant difference in foot strike pattern between the top 20 finishers and the remaining finishers existed at 161.1 km (p = 0.02) with non-RFS being more common among the top 20 finishers. Of the top 20 runners, a non-RFS pattern was used by 5 (25%) at 16.5 km, 1 (5%) at 90.3 km, and 6 (30%) at 161.1 km. Significance did not persist when comparing the top 40 finishers with the remaining finishers at this site (p = 0.5), nor was it demonstrated for the top 20 finishers compared with the remaining finishers at the 16.5 km (p = 0.4) or 90.3 km (p = 0.3) sites.
Stride rate and stride length for runners with usable stride parameter data across the 3 level sites (n = 240) demonstrated significant main effect differences across sites (p = 0.02 and p < 0.0001, respectively). Posttests revealed a greater stride length at the 16.5 km site compared with the 2 more distant sites (p < 0.0001 and p < 0.05, respectively). In addition, both stride rate and stride length were greater (p < 0.05 and p < 0.0001, respectively) at the 161.1 km site compared with the 90.3 km site. Comparison of the level and downhill sites at ∼90 km showed a greater stride length (n = 181, p < 0.0001) at the downhill site but no difference in stride rate between these sites.
Figure 1 demonstrates the changes in stride rate and stride length across sites and the effect of finish place. In the 2-way ANOVA considering groups of 20 based on sequential finish place, stride rate showed significant site (as described above) and group (p = 0.004) effects but no significant interaction effect (p = 0.8). For stride length, there were significant site (p < 0.0001) and group (p < 0.0001) effects and a significant interaction effect (p = 0.02). The top 20 finishers, however, demonstrated no change in stride rate (p = 0.2) or stride length (p = 0.7) across filming sites. It was the remainder of finishers who powered the overall pattern described above.
Figure 1: Changes in stride rate and stride length for top 20 finishers (1–20) and remainder of finishers (21–240) in which usable data were obtained across all 3 level sites. Remainder of finishers grouped collectively (21–240) represents the mean of sequential groups of 20 runners used in statistical analyses. Brackets represent 1 SD.
In comparing our findings with prior work, an RFS pattern was less common at the 16.5 km (p = 0.02) and 161.1 km (p = 0.04) sites compared with the 32 km site of Larson et al. (20). However, there was no statistically significant difference in foot strike pattern between our 90.3 km site and the 32 km site of Larson et al. (20) or between any of the 3 level sites and the 10 km site of this group or the 8 km site of Kasmer et al. (17) Comparison of our findings of foot strike pattern at 16.5 km with those of Hasegawa et al. (11) at 15 km revealed that RFS was more common (p < 0.0001) in this study.
Discussion
A key finding of this study is that 79.9% of runners were using an RFS pattern at the initial filming site (16.5 km). This RFS prevalence is significantly lower than previously observed by Larson et al. (20) at 32 km into a road marathon (93.0%) but did not reach statistical significance when compared with the findings of this group at 10 km into the marathon (87.8%) or Kasmer et al. (17) at 8 km into a marathon (93.7%). By approximately midway into the race (90.3 km), RFS prevalence increased to 89.0% and then decreased to 83.9% near the finish (161.1 km), which was not statistically different from that at the initial filming site. Thus, our hypothesis that runners in an ultramarathon would be more likely to adopt a non-RFS pattern to avoid the cumulative effect of impact transients seems partially correct. That is, when considering the foot strike pattern early in the race and in comparison with a group mostly consisting of amateur marathon runners, the present group of ultramarathon runners seemed to make less use of the RFS pattern.
In contrast to how the present findings compare with those of Larson et al. (20) and Kasmer et al. (17), RFS prevalence in this study was greater than that found by Hasegawa et al. (11) at the 15 km site among a more elite group of half-marathon runners (74.9%). It is unclear whether increased RFS prevalence in this study and in the study by Larson et al. (20) relates to filming technique affecting the accurate determination of foot strike pattern or if a lower RFS prevalence is an actual characteristic of faster runners, as observed by Hasegawa et al. (11). Hasegawa et al. (11) observed elite, and therefore, faster runners at a slower frame rate (120 frames per second). Thus, the lower RFS prevalence may reflect difficulty in correctly identifying foot strike pattern as compared with comparatively slower runners at a higher frame rate as in the present work and the work by Larson et al. (20).
The increase in RFS prevalence midway through the race is not surprising when considering the greater muscular demands required of the plantar flexor muscles associated with non-RFS patterns compared with an RFS pattern (7). Further explanation for this shift in foot strike pattern comes from the finding of the higher post-race blood CK concentrations among those runners who were observed to use a MFS or FFS pattern compared with those who consistently used an RFS pattern. Elevated post-race blood CK concentrations among runners have been demonstrated previously in ultramarathons (24) including the 2010 Western States Endurance Run (16), but the relationship of post-race blood CK concentration and “in-race” foot strike pattern has never been investigated. The relative elevation in post-race blood CK concentration among those runners observed to use a MFS or FFS pattern suggests that, by midway into the race, the balance had transitioned towards a need to protect the plantar flexor muscles from the eccentric loading demands by using an RFS pattern rather than limiting the impact transient with a non-RFS pattern. Interestingly, at the end of the race, non-RFS prevalence increased despite the softer synthetic track surface at 161.1 km, which may be expected to favor an increased RFS prevalence based on the theory of “impact-moderating behavior” initially proposed by Robbins et al. (34). The increased non-RFS prevalence at 161.1 km, especially among the top 20 runners, suggests that some runners who naturally prefer a non-RFS pattern may be able to resume this pattern at the end of a race despite fatigue or muscle pain, perhaps in favor of increased speed to finish the race. However, it did not seem that many runners using an RFS pattern earlier in the race adopted a non-RFS pattern at the end of the race. This suggests that most runners who naturally prefer an RFS pattern and may be more accustomed to the result of cumulative impact transients were not compelled to transition to a non-RFS pattern, which is associated with lesser impact transients, but greater muscular demands. It is possible that these runners preferentially altered factors other than foot strike pattern to reduce impact forces, such as increasing stride rate, which is supported by our finding, as initially observed in ultramarathoners by Morin et al. (26,28).
The fastest runners were more likely to use a non-RFS pattern at the 161.1 km site than the slower runners. Still, of the fastest 20 runners, a non-RFS pattern was only used by 5 (25%) at 16.5 km, 1 (5%) at 90.3 km, and 6 (30%) runners at 161.1 km. Furthermore, when considering all runners, those who used a non-RFS pattern were no faster than those who used other foot strike patterns. This finding of minimal relationship between foot strike pattern and performance in an ultramarathon is in accordance with the findings of Larson et al. (20) that half-marathon and full-marathon times were no different among those using different foot strike patterns. In contrast, Hasegawa et al. (11) and Kasmer et al. (17) demonstrated a greater non-RFS prevalence among faster runners in a half-marathon and full-marathon, respectively. It could be that the optimal balance of controlling eccentric loading demands relative to limiting impact transients is more dynamic in a longer race that involves the extent of climbs and descents and the associated high degree of muscular damage (16), as the Western States Endurance Run. Whereas, in events that place relatively less eccentric loading demands on the muscles, such as a road marathon, the beneficial effects on performance from a non-RFS pattern may be more likely to outweigh any negative effects.
Although there was minimal support for any relationship of performance level with foot strike pattern, there was a more apparent relationship of performance level with stride length. Stride length was significantly affected by finish place and site. The overall site effect (i.e., decreased stride length between 16.5 and 90.3 km) was similar to previous research in middle-distance runners (6,36), marathon runners (12), and endurance runners (31). However, this change in stride length was not characteristic of the fastest runners. Instead, the fastest runners had a greater stride length than the other runners that was maintained throughout the race. This effect is likely indicative of limited fatigue because of superior pacing strategies as a result of previous ultramarathon experience and the level of training related to become a top finisher in this ultramarathon. As a group, the runners finishing behind the top 20 were characterized by a decrease in stride length at 90.3 km, followed by an increase at 161.1 km. This effect is likely the result of fatigue at the 90.3 km site, followed by the ability to overcome fatigue for a short distance at the 161.1 km site as the finish line was approached.
Similar to the observed effect of performance level and site on stride length, stride rate was also significantly influenced by these factors. Considering all runners, stride rate was found to be greater at the 161.1 km site compared with the 90.3 km site but did not vary across sites among the top finishers. This is similar to previous studies in middle-distance runners that allowed a self-selected speed and demonstrated maintenance of stride rate despite a decrease in stride length over distance (6,36), and in endurance runners that demonstrated an increase in stride rate after completion of a mountain ultramarathon (28) or during a 24-hour treadmill run (26). Therefore, an increase in stride rate may represent an additional means by which runners can reduce impact forces. This theory is supported by previous studies demonstrating reduced lower extremity loading variables (13,14) and reduced incidence of impact-associated stress fractures (5,9) from an increased stride frequency.
Stride length and stride rate demonstrated significant main effect differences across foot strike patterns at the 16.5 km site (Table 2) with posttests revealing a trend toward shorter stride length and a significantly higher stride rate among non-RFS runners. This pattern of shorter stride length and higher stride rate has been previously demonstrated in comparisons of non-RFS and RFS running (3,4,35). This relationship, however, was less apparent for stride length and was not statistically significant at either 90.3 km or 161.1 km, indicating that the stride length and stride rate were more influenced by other factors associated with having traveled such distances rather than by foot strike pattern.
Similar to Larson et al. (20), this study did not observe a significant difference in foot strike pattern between men and women. There was also no significant difference in foot strike pattern related to age, except at 16.5 km, where non-RFS runners averaged 4 years older than RFS runners. This may suggest that older runners, presumably also with more experience, may be more cognizant of the need to control the adverse cumulative effects from ground impact.
An observation not previously documented in prior studies was the presence of shufflers. This observation may be unique to in-race studies where the subjects are focused on race performance and not as likely to be influenced by what has been described as the “sampling effect” (27), which may be more common in simulated race conditions, as suggested by Morin et al. (28). The shufflers' movement was characterized by what seemed visually in real time to be a running pattern, but frame-by-frame analysis revealed the absence of a double float phase. The prevalence of shufflers peaked at the 90.3 km site (8.7%) and is likely an adaptation to muscular fatigue and pain. Although not previously examined, it is presumed that peak forces and loading rate would be decreased because of the brief double support phase, similar to the effect of the adaptations of running mechanics observed to by Morin et al. (26,28). As such, it may be an effective method of movement at slower speeds when this movement pattern was most evident.
Data from the downhill site demonstrated no significant difference in foot strike pattern compared with the level site at approximately the same distance. However, stride length was significantly greater at the downhill site than the level site, whereas stride rate was not affected by grade. These findings are consistent with previous observations during a 9-km mixed-gradient run in which the downhill sections were characterized by a longer stride length, although there was no change in stride rate compared with running on the level sections (36). Such gait adjustments are interesting given that they may increase impact forces (8,22). In contrast to our observed gait adjustments between level and downhill running, the presumed increase in speed during the final 250 m of the race was associated with an increase in stride length and stride rate. Therefore, changes in speed among fatigued runners seem to be regulated by different adjustments in stride length and stride rate depending on the situation.
Practical Applications
It is particularly important for runners to limit impact transients and to minimize muscular damage during an ultramarathon. Although impact transients can be reduced through the use of a non-RFS pattern compared with an RFS pattern, this study suggests that there is greater muscular damage from non-RFS patterns. As such, there was a relatively small percentage of runners in this mountain 161-km ultramarathon using a non-RFS pattern early in the race, and the percentage became even smaller by midway through the race. These findings, combined with the finding that there was minimal support for any relationship of performance level with foot strike pattern, suggest that ultramarathon runners should not be encouraged to alter their natural behavior for adoption of a non-RFS pattern.
Acknowledgments
This material is the result of work supported with resources and the use of facilities at the VA Northern California Health Care System. The work was also supported by the Western States Endurance Run Foundation. The authors thank the following individuals for their assistance with this work: Dr. Vanessa McGowan, Kevin Mullins, Mike Neal, Sarah Neal, Kyle Yang and Bob Read. The contents reported here do not represent the views of the Department of Veterans Affairs or the United States Government.
References
1. Cheung RT, Davis IS. Landing pattern modification to improve patellofemoral pain in runners: A case series. J Orthop Sports Phys Ther 41: 914–919, 2011.
2. Daoud AI, Geissler GJ, Wang F, Saretsky J, Daoud YA, Lieberman DE. Foot strike and injury rates in endurance runners: A retrospective study. Med Sci Sports Exerc 44: 1325–1334, 2012.
3. De Wit B, De Clercq D, Aerts P. Biomechanical analysis of the stance phase during barefoot and shod running. J Biomech 33: 269–278, 2000.
4. Divert C, Mornieux G, Baur H, Mayer F, Belli A. Mechanical comparison of barefoot and shod running. Int J Sports Med 26: 593–598, 2005.
5. Edwards WE, Taylor D, Rudolphi TJ, Gillette JC, Derrick TR. Effects of stride length and running mileage on a probabilistic stress fracture model. Med Sci Sports Exerc 41: 2177–2184, 2009.
6. Elliott BC, Blanksby BA. A cinematographic analysis of overground and treadmill running by males and females. Med Sci Sports Exerc 8: 84–87, 1976.
7. Giandolini M, Arnal PJ, Millet GY, Peyrot N, Samozino P, Dubois B, Morn JB. Impact reduction during running: efficiency of simple acute interventions in recreational runners. Eur J Appl Physiol 113: 599–609, 2013.
8. Gottschall JS, Kram R. Ground reaction forces during downhill and uphill running. J Biomech 38: 445–452, 2005.
9. Hamill J, Derrick TR, Holt KG. Shock attenuation and stride frequency during running. Hum Mov Sci 14: 45–60, 1995.
10. Hamill J, Miller R, Noehren B, Davis I. A prospective study of iliotibial band strain in runners. Clin Biomech (Bristol, Avon) 23: 1018–1025, 2008.
11. Hasegawa H, Yamauchi T, Kraemer WJ. Foot strike patterns of runners at the 15-km point during an elite-level half-marathon. J Strength Cond Res 21: 888–893, 2007.
12. Hausswirth C, Bigard AX, Guezennec CY. Relationships between running mechanics and energy cost of running at the end of a triathlon and a marathon. Int J Sports Med 18: 330–339, 1997.
13. Heiderscheit BC, Chumanov ES, Michalski MP, Wille CM, Ryan MB. Effects of step rate manipulation on joint mechanics during running. Med Sci Sports Exerc 43: 296–302, 2011.
14. Hobara H, Sato T, Sakaguchi M, Sato T, Nakazawa K. Step frequency and lower extremity loading during running. Int J Sports Med 33: 310–313, 2012.
15. Hoffman MD. Anthropomorphic characteristics of ultramarathoners. Int J Sports Med 29: 808–811, 2008.
16. Hoffman MD, Ingwerson JL, Rogers IR, Hew-Butler T, Stuempfle KJ. Increasing creatine kinase concentrations at the 161-km Western States Endurance Run. Wilderness Environ Med 23: 56–60, 2012.
17. Kasmer ME, Liu X, Roberts KG, Valadao JM. Foot-strike pattern and performance in a marathon. Int J Sports Physiol Perform 8: 286–292, 2013.
18. Kerr BA, Beauchamp L, Fisher V, Neil R. Footstrike patterns in distance running. In: Kerr B.A., ed. Biomechanical Aspects of Sports Shoes and Playing Surfaces: Proceedings of the International Symposium on Biomechanical Aspects of Sports Shoes and Playing Surfaces. Calgary, Alberta: University Press, 1983;135–142.
19. Krabak BJ, Hoffman MD, Millet GY, Chimes GP. Barefoot running. PM R 3: 1142–1149, 2011.
20. Larson P, Higgins E, Kaminski J, Decker T, Preble J, Lyons D, McIntyre K, Normille A. Foot strike patterns of recreation and sub-elite runners in a long-distance road race. J Sport Sci 29: 1665–1673, 2011.
21. Lieberman DE, Venkadesan M, Werbel WA, Daoud AI, D’Andrea S, Davis IS, Mang'Eni RO, Pitsiladis Y. Foot strike patterns and collision forces in habitually barefoot versus shod runners. Nature 463: 531–536, 2010.
22. Mercer JA, Devita P, Derrick TR, Bates BT. Individual effects of stride length and frequency on shock attenuation during running. Med Sci Sports Exerc 35: 307–313, 2003.
23. Millet GY, Hoffman MD, Morin JB. Sacrificing economy to improve running performance—A reality in the ultramarathon? J Appl Physiol (1985) 113: 507–509, 2012.
24. Millet GY, Tomazin K, Verges S, Vincent C, Bonnefoy R, Boisson RC, Gergele L, Feasson L, Martin V. Neuromuscular consequences of an extreme mountain ultra-marathon. PLoS One 6, 2011. Available at:
http://www.plosone.org/article/info:doi/10.1371/journal.pone.0017059.
25. Milner CE, Ferber R, Pollard CD, Hamill J, Davis IS. Biomechanical factors associated with tibial stress fracture in female runners. Med Sci Sports Exerc 38: 323–328, 2006.
26. Morin JB, Samozino P, Millet GY. Changes in running kinematics, kinetics, and spring-mass behavior over a 24-h run. Med Sci Sports Exerc 43: 829–836, 2011.
27. Morin JB, Samozino P, Peyrot N. Running pattern changes depending on the level of subjects' awareness of the measurements performed: a “sampling effect” in human locomotion experiments? Gait Posture 30: 507–510, 2009.
28. Morin JB, Tomazin K, Edouard P, Millet GY. Changes in running mechanics and spring-mass behavior induced by a mountain ultra-marathon race. J Biomech 44: 1104–1107, 2011.
29. Nett T. Foot plant in running. Track Tech 15: 462–463, 1964.
30. Novachek TF. The biomechanics of running. Gait Posture 7: 77–95, 1998.
31. Place N, Lepers R, Deley G, Millet GY. Time course of neuromuscular alterations during a prolonged running exercise. Med Sci Sports Exerc 36: 1347–1356, 2004.
32. Pohl MB, Hamill J, Davis IS. Biomechanical and anatomic factors associated with a history of plantar fasciitis in female runners. Clin J Sport Med 19: 372–376, 2009.
33. Pohl MB, Mullineaux DR, Milner CE, Hamill J, Davis IS. Biomechanical predictors of retrospective tibial stress fractures in runners. J Biomech 41: 1160–1165, 2008.
34. Robbins SE, Gouw GJ, Hanna AM. Running-related injury prevention through innate impact-moderating behavior. Med Sci Sports Exerc 21: 130–139, 1989.
35. Squadrone R, Gallozzi C. Biomechanical and physiological comparison of barefoot and two shod conditions in experience barefoot runners. J Sports Med Phys Fitness 49: 6–13, 2009.
36. Townshend AD, Worringham CJ, Stewart IB. Spontaneous pacing during overground hill running. Med Sci Sports Exerc 42: 160–169, 2010.
37. Williams DS, McClay IS, Manal KT. Lower extremity mechanics in runners with a converted forefoot strike pattern. J Appl Biomech 16: 210–218, 2000.
38. Willson JD, Kernozek TW. Plantar loading and cadence alterations with fatigue. Med Sci Sports Exerc 31: 1828–1833, 1999.
