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

---

THE RELATIONSHIP BETWEEN RUNNING SMOOTHNESS AND INJURIES

Alan Hreljac1, Robert N. Marshall2
1Department of Exercise and Movement Science, University of Oregon
2Sport and Exercise Science Department, University of Auckland

INTRODUCTION

Chronic running injuries are thought to be related to excessive and repeated loads placed on joints and other anatomical structures (Nigg, 1985). It has been suggested (Bergmann et al., 1995) that the only means to reduce joint loading during jogging is to use a smoother gait pattern. If this statement is true, then it follows that running with a smooth gait pattern would reduce injury potential, and gait patterns which lack a smooth pattern would increase joint loading, thereby contributing to chronic injuries. Although the degree of smoothness is a characteristic that is often cited by coaches and other observers when describing running strides, it is not clear what constitutes a smooth gait pattern. Recently, stride smoothness has been quantified during gait (Hreljac & Martin, 1993; 1994) by evaluating the endpoint jerk-cost (JC) function, defined as the time integral of the mean squared jerk (third derivative of position) function. During gait, these researchers have concluded that the heel best represents the endpoint of gait.

In a previous study (Hreljac & Martin, 1994), it was reported that the running stride of competitive runners was smoother than the stride of recreational runners, assessed by the endpoint JC criterion. However, most of the difference between these groups was found during the swing phase. This may explain why elite running appear to be smoother than their recreational counterparts, but if smoother running strides reduce injury potential, it could only be during the stance phase (when joints are loaded) where this difference would be important. In this project, smoothness during the stance phase of running was evaluated and compared between a group of runners who had suffered chronic running injuries, and a group of runners who had remained injury-free throughout their running careers. It was hypothesized that the JC of the injury-free group would be less than that of the injured group.

METHODS

Two groups of 20 runners were recruited from various running, orienteering, and triathlon clubs to participate in this project. An injury-free (NI) group was comprised of runners who had not sustained any chronic injuries throughout their running careers (minimum of three years), while an injured (I) group consisted of runners who had experienced injuries which could be attributed to running. At the time of the study, none of the I group subjects were experiencing pain while running. Three-dimensional kinematic data of a reflective marker, placed on the posterior aspect of the landing heel, were recorded during the stance phase of a single gait cycle by a system of four Motion Analysis Falcon cameras (120 Hz) as subjects ran over a force platform, after running down a 20 m runway at 3, 4, and 5 m·s ^-1 . Speed was verified by three sets of photocell timers. Data were collected for both legs as subjects ran at each experimental speed. Landing leg and speed were randomly ordered.

Coordinate data were smoothed using a 4th order, zero-lag Butterworth filter before calculating the second derivative (acceleration) using finite difference equations. Acceleration data were then smoothed, and the first derivative of acceleration (jerk) was calculated. This "double" smoothing has been shown to be the best of several methods tested (unpublished work from this laboratory) in reproducing jerk curves from known functions to which random noise was introduced. Components of JC were calculated for the stance phase of each trial. In addition to the JC in the horizontal (JCx), medio-lateral (JCy) and vertical (JCz) directions, JC was also partitioned into magnitudinal (JCm) and directional (JCd) components (Schneider & Zernicke, 1989). The total JC (JCtot) was defined as the sum of JCm and JCd. A multivariate analysis of variance (MANOVA) was utilized to test for differences between groups for the six dependent variables during the stance phase of each speed and landing leg condition (p<0.05).

RESULTS

No significant differences were found in any variable (including stance time as measured from force platform records) between landing foot conditions at any running speed within either group. During left foot landing, the I group generally exhibited greater JC values than the NI group at 3 and 4 m·s ^-1 , although no differences were observed at 5 m·s ^-1 (Table 1). No significant differences were found between groups for any component of JC at any speed during the right foot landing condition (Table 2).

Table 1. JC (x 105 m ² ·s ^-5 ± 1 SD) for Left Foot Landing Condition.

Table 2. JC (x 105 m ² ·s ^-5 ± 1 SD) for Right Foot Landing Condition.

When left and right foot data were combined by averaging values for each subject, only JCy was significantly greater for I than NI at 3 m·s ^-1 , while JCx, JCz, JCmag, and JCtot were all significantly greater for I than NI at 4 m(s-1. There were still no significant differences between groups at 5 m·s ^-1 . There were no significant correlations coefficients found between the left and right foot values of specific variables for either group at 3 and 4 m·s ^-1 , although five of the six dependent variables showed significant positive correlations (r=0.50 to 0.74) within the NI group at 5 m·s ^-1 .

DISCUSSION

At testing speeds of 3 and 4 m·s ^-1 , the NI group was generally smoother than the I group during the stance phase of running. The differences in smoothness were most evident at the 4 m·s ^-1 pace, which was reported by a majority of subjects to be the most comfortable of the testing speeds. Since the variables of interest during this study were speed dependent, it was necessary to control the testing speeds, rather than allow subjects to run at their typical training pace. It was felt that the kinematics at the middle testing speed (4 m·s ^-1 ) were most representative of the typical kinematic patterns of subjects during training, although the typical training pace of all subjects in this study was 3.2 m·s ^-1 . The average training run of subjects from both groups was 14.8 km (no differences between groups), so it is likely that subjects were more comfortable running at a slightly faster speed during the short laboratory run.

The results of this study appear to indicate that the smoothness of the stance phase of running may be related to chronic running injuries, as indicated by the fact that the I group was generally not as smooth as the NI group while running at speeds close to their typical training pace. However, at these speeds (3 and 4 m·s ^-1 ), individuals in both groups did not demonstrate consistent smoothness patterns between left and right foot landings, as indicated by the lack of significant bilateral correlations, even though there were no group differences. It may be concluded that although the uninjured runners generally exhibited a smoother stance phase than the injured runners while running at speeds near typical training speeds, the relationship is not strong enough, nor consistent enough to allow smoothness to be used as an indicator of injury potential on an individual basis.

REFERENCES

Bergmann, G. et al. J. Biomech., 28, 817-827, 1995.

Hreljac, A. et al. Biol. Cybern., 69, 213-218, 1993.

Hreljac, A. et al. Proc. VIIIth CSB, 92-93, 1994.

Nigg, B.M. Sports Med., 2, 367-379, 1985.

Schneider, K. et al. Biol. Cybern., 60, 221-230, 1989.