Plantar loading and cadence alterations with fatigue : Medicine & Science in Sports & Exercise

Journal Logo

Applied Sciences: Biodynamics

Plantar loading and cadence alterations with fatigue


Author Information
Medicine & Science in Sports & Exercise 31(12):p 1828, December 1999.
  • Free


WILLSON, J. D., and THOMAS W. KERNOZEK. Plantar loading and cadence alterations with fatigue. Med. Sci. Sports Exerc., Vol. 31, No. 12, pp. 1828–1833, 1999.


The purpose of this investigation was to identify changes in loading characteristics of the foot associated with fatigue during running.


Nineteen healthy subjects ranging from 20 to 30 yr (mean = 22.3, SD = 2.4) were equipped with the Pedar in-shoe measurement system (Novel GmbH) for the assessment of plantar loading. After acclimation to the treadmill, subjects were progressed through the Ohio State protocol for exercise testing until fatigue was reported using Borg’s RPE scale. Six right footsteps were recorded at 150 Hz for each subject’s comfortable running pace under normal and fatigued conditions. A series of repeated measures multiple analysis of variance was performed for all dependent variables analyzed in this study including peak force (PF), force-time integral (FTI), peak pressure (PP), and pressure-time integral (PTI) for all regions of the plantar surface.


Decreased step time, significantly smaller values under the heel for PP, PF, FTI, CT, and PTI, and trends toward increased medial forefoot loading were identified while subjects were running under fatigued conditions (α < 0.05).


These results suggest that subjects change running technique and plantar surface loading characteristics in response to fatigued conditions through increased cadence, decreased loading of the heel, and increased medial forefoot loading.

There is little controversy regarding the value of a regular exercise routine. There is a plethora of convincing epidemiological evidence demonstrating that regular exercise protects against the development and progression of many chronic diseases and is an integral element of a healthy lifestyle (1). Two of the most common modes of exercise chosen in the United States are running and walking. Unfortunately, those individuals who engage in running protocols appear to be especially prone to musculoskeletal injury, including stress fractures, tendinitis, ankle sprains, and muscle strains (8,14,16). Annual overall incidence of injury among distance runners ranging from 25 to 65% for recreational and competitive athletes has been reported (8), with stress fractures comprising nearly 25% of these (7).

The etiology of these injuries has been the subject of numerous investigations. Research utilizing force platforms has shown that ground reaction forces during running are twice as high as those delivered to the foot during walking (13). In addition, as people run, forces are delivered to the foot in less than half of the time. Taken together, Perry (13) proposed that forces imposed on supporting tissues during running are four times those of walking. Jones et al. (8) further explained that this level of strain is likely to produce microtrauma to structures responsible for managing impact forces, leading to the creation of symptoms over time that are characteristic of running such as stress fractures, tendinitis, and muscle strains.

The ability of the body to absorb this impact may change as the runner becomes fatigued. Stokes et al. (18) used a mathematical model to analyze the forces acting on the metatarsals during the heel-lift phase of the gait cycle and hypothesized that fatigue of the plantar flexors may contribute to increased metatarsal bending moments causing strains and stress fractures. Sharkey et al. (16) followed this line of inquiry with a cadaveric study to examine the hypothesis that fatigue of the plantar flexors causes increased metatarsal loading. Although their results did not establish a definitive relationship between tension of the plantar flexors and metatarsal strain, they suggested a progressive increase in dorsal strain on the second metatarsal head as the number of plantar flexor muscles firing during heel lift decreased (16). Further, they reported that rapid and coordinated contractions of the plantar flexors was imperative in dynamic loading scenarios, such as running to offset the increased moments and loads delivered to the forefoot (16).

Nyland et al. (12) suggested that one of the consequences of running while fatigued was a diminished stabilizing capacity of the runner’s muscles. Therefore, they proposed that loads delivered to the feet while running must be absorbed by inert internal tissues such as ligaments, cartilage, and bones (12). Repetitive loading of these tissues has been shown to be a significant risk factor for many running related injuries including stress fractures, flattening of the longitudinal arch, and metatarsalgia (8,12,14,16).

Nigg et al. (11) stated that humans have the ability to change their running technique in response to the perception of harmful loads delivered to the feet. Other authors have suggested that changes associated with a fatiguing bout of exercise were made in an effort to increase running efficiency rather than to prevent injury. Williams et al. (19) reported an increase in step length, knee flexion angle during swing, and maximal thigh angle during hip flexion as competitive runners became fatigued during simulated 5000-m runs on a treadmill. Elliot and Ackland (4), on the other hand, reported a decrease in stride length during competitive 10,000-m runs and a more inconsistent leg position at foot-strike as the competitors became fatigued.

Nigg’s speculation is based largely on external measures of force due to the fact that the technology to adequately measure loads within the shoe has not been available until relatively recently. The development of in-shoe pressure analysis equipment allows investigators to objectively evaluate perhaps the most important link between the body and the environment, that of the interaction between the foot and the shoe (9). Thus, the consequences of adjusting running technique during fatigue to the human body may now be analyzed with less speculation. The purpose of this study is to identify changes in the loading characteristics of foot structures while running under fatigued versus rested conditions.

The hypothesis for this study stems from two related studies. First, physiological fatigue has been hypothesized to contribute to increased metatarsal strain and bending moments (18). Increased stress over a given surface area is believed to increase the risk of developing stress fractures (14). Second, as physical fitness and running have increased in popularity, there has been a proportional increase in metatarsal stress fractures (14). Therefore, it is hypothesized that loading characteristics of the forefoot will increase accompanied by a proportional decrease in heel loading values as runners become fatigued.


To analyze the effects of fatigue on loading variables of the feet while running, many characteristics of the materials and methods for this study are modeled after those of Kernozek et al. (9). These authors presented the intraclass correlation coefficients for the Pedar (Novel GmbH, Munich, Germany) in-shoe measurement system between 0 and 16 footsteps during treadmill walking. The following values were reported after six footsteps: peak force of the maximum pressure picture (PF) = 0.95, force time integral (FTI) = 0.91, peak pressure of the maximum pressure picture (PP) = 0.88, and pressure time integral (PTI) = 0.89 (9).


Twenty-two active volunteer participants were recruited for this study from a university campus with prior walking or running experience on a treadmill. All subjects were free from orthopedic injury or complaints and required to read and sign a written informed consent form approved by the institutional review board at the initiation of the testing procedures. In addition, each subject was fully informed of the procedures and risks of a graded exercise test (GXT) and given the opportunity to ask and have answered any questions they may have relative to this procedure. Finally, all subjects were asked to complete a health and lifestyle history for current illnesses, medical history, presence of cardiovascular risk factors, medications, and physical activity habits as recommended by the ACSM (1). Subjects who demonstrated contraindications to exercise testing were not invited to continue with this study. Nineteen subjects met the above criteria (11 female, 8 male; mean age = 22.4 (2.6)).


All subjects were weighed and requested to wear their own running shoes and hosiery for testing. Inside each shoe, a 2-mm thick, 99-sensor insole was inserted corresponding to the size of the subject’s foot. The leads for these insoles were secured to the subject’s lower legs with Velcro straps. The remainder of the Pedar in-shoe system consisted of an A/D conversion box and a cable connected to a laptop computer was attached to the participant’s waist. To enhance the sampling rate for running, 32 sensors in the arch region of the insole were inactivated. Therefore, 67 sensors in the subject’s right shoe were sampled at a rate of 150 Hz for a minimum of six footsteps during the participant’s running gait under both fatigued and rested conditions. Data was stored in the hard drive of the laptop computer for further analysis. Before fitting the participants with the equipment, each sensor of each insole was calibrated using calibration software and an air bladder that was inflated to load the insoles to various pressures throughout the measurement range (0–600 kPa).


The subjects were guided to the treadmill where resting heart rate (HR) and blood pressure (BP) were recorded. Throughout the testing procedure, HR was monitored every 2 min to gauge the subject’s reaction to the exercise protocol. Subjects were asked to straddle the treadmill belt and grasp the handrails as the treadmill started at a speed of 1.34 m·s−1. Subjects were then be asked to “paw” the belt to get the feel of the belt speed and then step onto the moving belt and begin walking. Subjects were given 5 min to acclimate themselves to the treadmill at this speed. At the conclusion of this 5 min, the subjects were asked to choose a comfortable running pace. The speed of this running pace was noted, and six right footsteps were recorded to represent the rested condition.

Once the baseline data had been recorded, the treadmill speed was adjusted to the first stage of the Ohio State Protocol for graded exercise testing (5). This protocol required participants to run at a constant speed (male subjects = 3.48 m·s−1, female subjects = 2.68 m·s−1; 7-min 41-s miles, and 10-min miles, respectively) while the treadmill platform was inclined 2° every 2 min. The subjects participated in this protocol until they demonstrated an inappropriate response to exercise as dictated by the ASCM guidelines for exercise testing or the subject indicated they were approaching maximum voluntary effort as measured by a score of greater or equal to 19 on Borg’s record of perceived exertion scale. At that time, the treadmill was returned to a 0% gradient and the speed the subject previously identified as their comfortable pace. Six right footsteps representing the fatigued condition were recorded and stored on the hard drive of the laptop computer for comparison to the rested condition.


Multimask software was utilized to outline the dimensions of seven anatomical regions on the plantar surface of the foot. Region 1 corresponded to the heel. Region 2 represented the arch region. Sensors in this plantar region were not sampled as to enhance the sampling rate of the remaining 67 sensors. Regions 3, 4, and 5 represented the first metatarsal, second and third metatarsals, and fourth and fifth metatarsals, respectively. Region 6 portrayed the hallux and region seven was over the lesser toes (Fig. 1). Cadence was determined from onset of one footstrike to onset of the next footstrike from the plantar loading data of three consecutive right footfalls for each subject during each condition.

Figure 1:
Seven anatomical regions or masks associated with the pressure distribution measurement: 1 = heel region, 2 = midfoot region, 3 = medial forefoot region, 4 = central forefoot region, 5 = lateral forefoot region, 6 = medial toe region, and 7 = lateral toe region.

A series of repeated measures multiple analysis of variance was performed for all dependent variables analyzed in this study including CT, PF, FTI, PP, and PTI for all plantar regions. Cadence was analyzed using a paired t-test. The independent variable in this experiment was fatigue. An alpha level of 0.05 was set for acceptance of the hypothesis stated above.


Plantar loading characteristics during the rested condition were consistent with previous studies (15). Peak force was highest under the heel followed by the second and third metatarsals (70.0%BW and 37.2%BW, respectively). Longer contact time on the forefoot led to higher FTI values under the second and third metatarsals than under the heel region.

The results of the RM MANOVA revealed significant multivariate effects under the heel and medial forefoot during the fatigued condition. Significantly reduced univariate values were isolated for the heel region for CT, PF, PP, FTI, and PTI during the fatigued run (Table 1).

Table 1:
Means and standard deviations for contact time (CT), peak force (PF), peak pressure (PP), force-time integral (FTI), and pressure-time integral (PTI) during rested and fatigued running conditions; significant multivariate effects were identified for regions 1 and 3 (heel and medial forefoot regions); follow-up univariate effects were identified for just the heel region.

During the fatigued condition, less force was delivered to the heel over a shorter period of time, leading to a significantly smaller impulse in this region with respect to the rested condition (Fig. 2). PF values changing from 70.0%BW in the rested condition to 62.5%BW in the fatigued condition in combination with reduced CT from 179.3 ms to 163.3 ms were responsible for the smaller FTI values recorded in this region. The same logic applies to the relationship between CT, PP, and PTI during fatigue.

Figure 2:
A typical footfall depicting the force-time integral (FTI) for the heel region during the rested (darker shading) and fatigued (lighter shading) conditions during running. Data from six right footsteps were time normalized and ensemble averaged. The shaded bands depict the standard deviation for each loading condition. The data displays the significantly smaller impulse delivered to the heel during the fatigued running condition.


This investigation into the consequences of fatigue differs from earlier studies in two key ways. First, earlier studies dealing with the effects of fatigue on running technique by previous authors chose to sample elite athletes (2,4,19). Our investigation focused on recreational runners only. This was an effort to make our results more generalizable to the public, and perhaps more applicable to health care professionals working with running related injuries of recreational athletes. Second, earlier investigations have mainly focused on over-ground running where speed may fluctuate. Our experiment was performed on a treadmill with speed controlled. Although most experts agree that runners tend to decrease their speed as fatigue sets in, plantar loading variables have been shown to change with velocity (2,4,9,19). Therefore, to ensure that changes identified during the fatigued state were not attributed to speed adjustments by the runners, it was important to control for this variable on the treadmill.

Subjects running while fatigued displayed three consistent differences from their rested technique. First, subjects demonstrated a significantly faster cadence (Table 2). Second, loading characteristics of the heel region including CT, PF, PP, FTI, and PTI were significantly less during the fatigued condition. Third, subjects tended to increase loading under the first metatarsal. Considering the treadmill speed was identical between conditions, the consequence of a faster cadence was a decreased step length.

Table 2:
Means and standard deviation for cadence during rested and fatigued running conditions; a significant difference was identified for the fatigued condition.

There are many possible reasons that a subject may decrease step length as fatigue sets in. First, the runners may shorten step length to decrease the deceleration forces at heel strike. This allows the runner to conserve forward momentum as their center of mass to passes over the stance leg. Indeed, competitive runners are often coached to make foot contact as close to their center of mass as possible to increase efficiency and performance (4). This hypothesis correlates well with other results of this study including significantly lower heel loading characteristics during the fatigued state.

A competing explanation proposed by Nigg et al. (11) is that runners change their technique in response to the perception of harmful loads. This hypothesis stems from the observation that runners change their technique to keep external impact forces constant when running with midsoles of different hardness. As mentioned above, the fatigue of dynamic supporting structures can mean that potentially harmful impact forces are absorbed by static structures. Perhaps runners choose to decrease step length to avoid harmful impact forces delivered through the shoe to the static foot structures during fatigue. Clarke et al. (3) has previously reported that stride rate and stride length are the two variables runners can most easily change and that reducing stride length decreases accelerometer values at the leg. Our results suggest that decreased leg shock values associated with a decreased step length may begin at the heel.

Further implications for a reduced stride length revolve around the ankle joint. Nigg (10) reported that runners who land on their heel produce a plantarflexion moment that is counteracted to a degree by the muscles on the anterior aspect of the lower leg. Runners who tend to land with more of a midfoot landing strategy produce a dorsiflexion moment that is counteracted by the much larger and stronger muscles on the posterior aspect of the lower leg. Therefore, overloading the anterior muscles of the lower leg can be avoided by changing the landing pattern from a heel-toe to midfoot landing strategy. In our experiment, there was a significant reduction in heel loading with trends toward increased medial metatarsal loading. This indicates that our subjects may have engaged a midfoot landing strategy in an attempt to unload the fatigued anterior muscles of the lower leg from their previous heel-toe running pattern. Future research may consider using electromyography to shed some light on the timing of these muscles under the two conditions.

Trends toward increased first metatarsal loading may be related to two factors. First, as noted above, subjects may have engaged a midfoot landing strategy to reduce the plantarflexion moment at the ankle and diminish the necessity for dynamic control by the anterior musculature. However, no significant loading changes were identified under the other four metatarsals, leaving this theory open to speculation. Second, increased first metatarsal loading may be a characteristic of greater pronation. Muscles such as the tibialis posterior that supinate the foot and support the medial longitudinal arch may have fatigued to the point where they can no longer maintain the rigid bony architecture necessary for propulsion. Therefore, the foot may fall into pronation to a greater degree or for a longer time during midstance during the fatigued condition. Prolonged pronation is thought to place the posterior tibialis in a compromised length-tension relationship during propulsion and has been shown to contribute to the etiology of “shin splints” (14).

There are limitations inherent in nearly every research design. Worthy of note in this experiment is the assumption that loading characteristics are similar between the right and left legs. Although we are aware of no evidence suggesting leg dominance affects loading characteristics during running, only right footsteps were collected. Second, we assumed that changes associated with the fatigued state were independent of gender. Finally, it was also our assumption that uphill running alone would have no lasting effect on running technique once the incline was eliminated and the “fatigued” condition was sampled. It is conceivable that the subjects changed their technique during the Ohio State Protocol and those changes, not fatigue alone, were at least partially responsible the results listed above. Despite these limitations, we hope these results inspire further investigation, especially in light of the overwhelming number of orthopedic injuries commonly associated with running.

We especially acknowledge Chris Dodge, director of the human performance laboratory, for organizing the use of the treadmill utilized in this investigation. We also thank Bob Strzelczyk for his past contributions to the clinical biomechanics laboratory.


1. American College of Sports Medicine. Guidelines for Exercise Testing and Prescription. Baltimore, MD: Williams and Wilkins, 1995, pp. 86–109.
2. Cavanagh, P. R., G. C. Andrew, R. Kram, M. M. Rodgers, D. J. Sanderson, and E. M. Hennig. An approach to biomechanical profiling of elite distance runners. Int. J. Sport Biomech. 1:36–62, 1985.
3. Clarke, T., L. Cooper, D. Clark, and C. Hamill. The effect of varied stride rate and length upon shank deceleration during ground contact in running (Abstract). Med. Sci. Sports Exerc. 15:170, 1983.
4. Elliot, B, and T. Ackland. Biomechanical Effects of Fatigue on 10,000 meter running technique. Res. Q. Exerc. Sport 52:160–166, 1981.
5. Fox, E. L., R. W. Bowers, and M. L. Foss. The Ohio State Protocol for exercise testing. In:The Physiological Basis of Physical Education and Athletics, 4th Ed. Dubuque, IA: William C. Brown, 1989, p. 667.
6. Frederick, E. C., and J. L. Hagy. Factors affecting peak vertical ground reaction forces in running. Int. J. Sport Biomech. 2:41–49, 1986.
    7. Fredericson, M. Common injuries in runners. Sports Med. 21:49–72, 1996.
    8. Jones, B. H., D. N. Cowan, and J. J. Knapik. Exercise, training and injuries. Sports Med. 18:202–14, 1994.
    9. Kernozek, T. W., E. E. LaMott, and M. J. Dancisak. Reliability of an in-shoe pressure measurement system during treadmill walking. Foot Ankle 17:204–09, 1996.
    10. Nigg, B. M. Biomechanics, load analysis, and sports injuries in the lower extremities. Sports Med. 2:367–378, 1985.
    11. Nigg, B. M., H. A. Bahlsen, S. M. Luethi, and S. Stokes. The influence of running velocity and midsole hardness on external impact forces in heel-toe running. J. Biomech. 20:310–315, 1987.
    12. Nyland, J. A., R. Shapiro, R. L. Stine, T. S. Horn, and M. L. Ireland. Relationship of fatigued run and rapid stop to ground reaction forces, lower extremity kinematics, and muscle activation. J. Orthop. Sports Phys. Ther. 20:132–37, 1994.
    13. Perry, J. Anatomy and biomechanics of the hindfoot. Clin. Orthop. 177:9–15, 1983.
    14. Reeder, M. T., B. H. Dick, J. K. Atkins, A. B. Pribis, and J. M. Martinez. Stress fractures: current concepts of diagnosis and treatment. Sports Med. 22:198–212, 1996.
    15. Rodgers, M. M. Dynamic biomechanics of the normal foot and ankle during walking and running. Phys. Ther. 68:1822–1830, 1988.
    16. Sharkey, N. A., L. Ferris, T. S. Smith, and D. K. Matthews. Strain and loading of the second metatarsal during heel-lift. J. Bone Joint Surg. Am. 77:1050–57, 1995.
    17. Sinacore, D. R., B. L. Bander, and A. Delitto. Recovery from a 1-minute bout of fatiguing exercise: characteristics, reliability, and responsiveness. Phys. Ther. 74:234–41, 1994.
    18. Stokes, A. F., W. C. Hutton, and J. R. R. Stott. Forces acting on the metatarsals during normal walking. J. Anat. 129:579–590, 1979.
    19. Williams, K. R., R. Snow, and C. Agruss. Changes in distance running kinematics with fatigue. Int. J. Sport Biomech. 7:138–162, 1991.


    © 1999 Lippincott Williams & Wilkins, Inc.