Presented at NACOB 98:
North American Congress on Biomechanics
Canadian Society for Biomechanics - American Society of Biomechanics

University of Waterloo
Waterloo, Ontario, Canada
August 14-18, 1998


I. McClay 1,2 and D. Williams 1
1 Motion Analysis Laboratory, University of Delaware, Newark, DE 19716
2 Joyner Sportsmedicine Institute, Harrisburg, PA 17111


Runners have been known to alter their footstrike patterns for a variety of reasons. This study was undertaken to compare lower extremity mechanics produced by a converted forefoot strike pattern (CFFS) to those of a practiced forefoot striker (FFS). The peak vertical ground reaction force and peak ankle plantarflexion moment were found to be significantly lower in the CFFS group with values being closer to that of rearfoot striker (RFS). However, all other kinematic and kinetic peak variables and all patterns of movements were found to be similar between these groups. This suggests that when one changes from RFS to a FFS pattern, they immediately adopt a pattern that is similar to a practiced forefoot striker.


Approximately 20% of distance runners are mid/forefoot strikers (Kerr et al., 1983). Some of these runners have reported that they adopted this type of running in response to coaches suggestions in attempt to improve their speed. Others have altered their strike pattern as a result of an injury.

Significant differences have been noted both in joint kinematics, kinetics and ground reaction forces in forefoot strikers (McClay and Manal, 1995, Cavanagh et al, 1980, Oakley et al, 1988, Harrison et al, 1988). These differences include greater rearfoot plantarflexion and inversion and knee flexion at footstrike and greater peak dorsiflexion and eversion velocities. In addition, larger peak vertical and anteroposterior GRF have been noted along with larger peak powers at the rearfoot and lower peak powers at the knee in the FFS group. Vertical loading rates are also lower in FFS. RFS that have injuries thought to be associated with the force transient present with heel contact may do well to adopt a forefoot strike pattern resulting in the elimination of the impact peak and reduction of the loading rate. However, it is not known whether simply instructing someone to "run on their toes" results in a biomechanical profile characteristic of a practiced or natural FFS.

Since only 20% of runners are FFS, recruitment of these subjects for study is sometimes difficult. If a RFS is able to simply alter their strike pattern and immediately adopt a FFS pattern that is similar in mechanical composition, the same subjects can be used to study both patterns of running. Therefore, the purpose of this research is to compare the lower extremity kinematics of runners with natural forefoot strike pattern to those of rearfoot strikers who are asked to alter their pattern to a FFS.


Eighteen recreational runners (9 FFS and 9 RFS) volunteered for this study. The dominant leg was tested, unless this foot exhibited excessive eversion (> 15 deg) as determined during an initial treadmill screening.

Four retroreflective markers were affixed to two velcro-backed polyform shells and attached to the thigh and shank via a neoprene wrap. In addition, three markers were attached to the heel counter of the shoe. Additional markers were placed over various body landmarks to establish the anatomical coordinate systems in which the motion would be described. After a standing calibration trial was collected, the anatomical markers were removed.

Subjects then ran along a 75 ft. runway at a speed of 3.35 m/s (10%). The RFS first ran with their normal strike pattern. They were then instructed to run on their toes. They were given a few practice trials until they were comfortable running within the speed range striking the forceplate with a forefoot strike pattern. The FFS ran with their normal strike pattern. Five trials were collected for each footstrike pattern. Data were sampled at 120 hz with a 5 camera VICON (Oxford Metrics, UK) motion analysis system. GRF were collected with a Bertec (BERTEC Corp, OH) forceplate being sampled at 480 hz. Three-dimensional joint kinematics and kinetics were calculated using MOVE3D software (NIH Biomechanics Laboratory). The variables of interest were those that have been noted to be different between RFS and FFS. Independent t-tests (adjusted p<0.025) were run for the chosen variables comparing the CFFS and FFS data. RFS data are presented for comparison.


Table 1: Rearfoot Kinematics (mean, sd)

PF at FS (°) -6.2 (7.2) ns -4.3 (7.4) +7.6 (5.3)
Inv at FS (°) -8.5 (4.1) ns -8.9 (3.2) -2.5 (2.8)
DF excursion (°) 29.9 (7.1) ns 31.7 (6.1) 22.0 (2.9)
EV excursion (°) 17.5(5.7) ns 15.8 (5.0) 11.5 (1.6)
DF Vel (°/s) 420.5 (93.3) ns 394.4 (74.6) 316.2 (32.9)
EV Vel (°/s) 361.4 (82.8) ns 286.9 (96.3) 243.2 (56.6)

Table 2: Knee Kinematics (mean, sd)

FL at FS (°) -20.1 (8.6) ns -26.1 (6.6) -14.0 (5.7)
ER at FS (°) -15.2 (8.2) ns -12.4 (7.0) -10.9 (5.6)
FL Excursion (°) -25.2 (7.6) ns -25.4 (4.9) 28.9 (2.3)
IR Vel (°/s) 239.8 (79.5) ns 194.7 (54.6) 177.4 (36.7)

Table 3: Kinetics Results (mean, sd)

V GRF (bw) 2.82 (0.18) sig 2.43 (0.22) 2.58 (0.16)
Ant GRF (bw) 0.38 (0.06) ns 0.34 (0.06) 0.31 (0.06)
V Load Rate 35.57 (13.43) ns 28.85 (10.02) 64.18 (21.08)
Ankle PF Mom. -1.83 (0.16) sig -1.41 (0.23) -1.51 (0.21)
Ankle Pow. Abs. -7.25 (1.39) ns -5.35 (1.67) -4.98 (1.07)
Ankle Neg. Work -51.99 (26.37) ns -54.82(12.35) -32.72 (11.57)
Ext Mom. 1.45 (0.30) ns 1.68 (0.36) 1.63 (0.28)
Knee Power Abs. -4.97 (2.16) ns -6.20 (0.78) -8.65 (1.61)
Knee Neg. Work -26.97 (14.32) ns -40.22(13.86) -49.89(15.72 )

Mom. in Nm/kg·ht, Power in Watts/kg·ht,

Figure 1: Ground Reaction Forces.

Figure 2: Moments and Powers


The variables studied were those that have been previously noted as different between RFS and FFS. It was surprising to find that all but one comparison were statistically similar between the groups. Along with discrete variables, kinematic and kinetic patterns of motion were visually similar between groups.

One might expect that it would require some practice to adapt to the new strike pattern. However, it appears that the simple alteration of striking the ground with the forefoot results in predictable changes up the kinematic chain. This is due, in part, to the closed chain nature of gait. For example, landing with the foot plantarflexed functionally lengthens the lower extremity. Increasing knee flexion at footstrike is one way in which to functionally shorten the limb to compensate for the plantarflexed ankle. In addition, inversion is mechanically linked with plantarflexion of the ankle and is therefore greater in a forefoot strike pattern.

Anteroposterior ground reaction forces were also found to be similar with greater peak propulsive forces in both CFFS and FFS compared with RFS. Both FFS and CFFS exhibited the characteristic biphasic pattern of the braking component. It appears slightly attenuated in fig. 1. This is due to the smoothing of the curve that occurs with averaging.

The lower peak VGRF seen in the CFFS compared with the FFS was associated with a lower speed of running (although both were within the 10% range) The VGRF pattern of the CFFS exhibited the missing impact peak, which is characteristic of the practiced FFS. In addition, loading rates were similar between the two groups and both values were lower than that of the RFS. This implies that changing to a FFS pattern immediately reduces the vertical loading rate which has been suggested by some to be related to injury.

However, an anterior strike pattern may increase the load on the posterior calf musculature. Although peak plantarflexion moments were signiificantly lower in the CFFS, dorsiflexion velocities, peak power absorption and negative work at the ankle were greater in both the FFS and CFFS compared with the RFS. This may overwork the gastrocsoleus muscle group and increase the risk for injury such as achilles tendinitis. Conversely, both FFS and CFFS demonstrated lower power absorption at the knee which may result in lower demands of the quadriceps muscle group.

In summary, it appears that runners are able to quickly adapt their gait pattern from a RFS to a FFS that is mechanically similar to that of a practiced FFS. Further study is needed to determine whether this strike pattern alters the risk for specific running-related injuries.


Cavanagh et al, J. Biom, vol 13, pp397-406, 1980

Harrison et al, Biomechanics in Sport, pp81-88, 1988

Kerr et al,Biomechanical Aspects of Sport Shoes and Playing Surfaces, 1983

McClay & Manal, Proc. of ASB, 211-214, 1995

Oakley et al, Clin. Biom, vol 3, pp159-165, 1988


The authors would like to acknowledge the Whitaker Foundation for their support of this work.