Journal Logo


Ground Reaction Forces and Kinematics in Distance Running in Older-Aged Men


Author Information
Medicine & Science in Sports & Exercise: July 2003 - Volume 35 - Issue 7 - p 1167-1175
doi: 10.1249/01.MSS.0000074441.55707.D1
  • Free


Although the stance phase biomechanics of distance running in younger-aged runners has been studied extensively over the past few decades, investigations on older-aged runners (defined here as 55 yr and older) are, to the best knowledge of the author, not present in the literature. However, in the process of aging, several degenerative musculoskeletal changes occur that may significantly affect the mechanics of running. For example, loss of lower-extremity joint flexibility (29,31) may reduce the joint ranges of motion (ROM) during running. In addition, progressive weakness in muscle and bone (18,27,32), together with atrophy and loss of elasticity of plantar fat tissue (13,14), reduce the shock-absorbing capacity of the body and may lead to higher impact forces. Finally, loss of strength and contraction velocity of major lower-extremity muscle groups (18,32) may reduce the ground reaction forces (GRF) during the propulsive phase of running. This raises the question whether older- and younger-aged runners are biomechanically similar in the way they run. If significant differences in ground reaction forces and stance phase kinematics exist between older- and younger-aged runners, then this may have implications for their relative susceptibility to lower-extremity overuse injuries and for the prescription of exercise in older-aged runners. Moreover, this may have consequences for the design and prescription of running shoes, which, to date, has been based on existent biomechanical data from younger-aged runners.

Both younger- and older-aged runners are frequently exposed to injuries, and a yearly incidence rate between 37% and 56% has been reported in each of these groups (19). In older-aged runners this problem may be underestimated due to the “healthy runner effect” in which only older runners who remain free of injury continue to run (19,20). The majority of sports medicine literature has suggested that overuse due to repetitive high-impact loading of the lower extremities is a large contributor to injuries (12,16,19,30), also in older-aged runners (20,26). As a consequence, the prescription of exercise for older-aged persons is mainly directed toward activities with lower impact such as (brisk) walking, cycling, and swimming (26). In the design of running shoes, the suggested association between both high impact forces and excessive pronation and lower-extremity injury (7,12,16) has lead to a major emphasis on the cushioning of impact forces and rearfoot stability. Defining the conditions experienced by the musculoskeletal system during impact as well as defining the degree of rearfoot control in older-aged runners should give further insight into the assumptions made on the magnitude and causes of lower-extremity injuries as well as the need for modification of the prescription of exercise or a specific shoe design and prescription for older-aged runners.

The present study aims at answering the question whether older- and younger-aged runners are biomechanically similar in the way they run by comparing the ground reaction forces and kinematics of running in a group of older- and younger-aged male runners. Based on the above-mentioned degenerative effects of older age on musculoskeletal function, it is hypothesized that older-aged runners will show smaller lower-extremity ROM, higher impact GRF, and lower propulsive GRF than younger runners.



Sixteen older-aged (55–65 yr) and 13 younger-aged (20–35 yr) male runners participated in this study. All subjects were well-trained distance runners and heel strikers at the speeds tested (confirmed through video analysis) in order to eliminate the type of foot strike as a possible confounding variable (3). None of the participants had any known neurological or musculoskeletal disorder affecting their gait, nor did any subject have a lower-extremity injury in the 2 months preceding the study. Each subject gave written informed consent to undergo the experimental procedures, which were approved by the research ethics committee of TNO Industrial Technology. The personal and training characteristics of the subjects are summarized in Table 1.

Personal and training characteristics of the younger- and older-aged subjects expressed in mean (SD) values; running distance per week was estimated by the subjects.

Data collection and analysis.

A Precision Motion Analysis System (PRIMAS, TU Delft, The Netherlands), consisting of four infrared cameras measuring at a sample frequency of 100 Hz, was used to collect the three-dimensional (3D) trajectories of clusters of three markers that were attached to the thigh, shank, and foot/shoe segments of the right lower extremity. Markers were retroreflective spheres of 2-cm diameter that were rigidly fastened to a stiff metal plate. The metal plate was attached directly to the shoe to represent the foot/shoe segment, and to shin guards, which were fastened to the subject using sports tape, to represent the shank and thigh segments. Right-to-left symmetry in running was assumed (22).

Custom software was used to process the kinematic data. The data were low-pass filtered (bi-directional) at a cut-off frequency of 9 Hz using a fourth-order Butterworth filter. Direct linear transformation was applied to reconstruct the 2D image coordinates of the markers into 3D space coordinates. All trajectories of the marker clusters were related to an upright standing reference position of the subject to determine the segmental linear and joint angular displacements in the three planes of motion.

Knee flexion at heel strike and vertical impact speed were selected as variables for statistical analysis because of their reported influence on impact force (8,9,21,23). Vertical impact speed was calculated over six motion frames just before heel strike by numerical differentiation of the vertical position data of the shank segment. Maximum subtalar eversion angle was included in the analysis because of its suggested importance in the etiology of running injuries (7,16). Knee and ankle joint range of motion (ROM) in the sagittal plane and subtalar joint ROM in the frontal plane were also included in the analysis to describe overall kinematics. Angle conventions are presented in Figure 1.

Convention for angular displacements in the sagittal plane (left) of the knee joint (θK) and the ankle joint (θA) and in the frontal plane (right, posterior view) of the subtalar joint (θS) of the right lower extremity.

The method of using an upright standing reference position of the subject for calculating joint angles implies that, despite careful instructions, minor deviations in this position, such as standing with slightly bent knees, may have been present. This leads to an overestimation of the absolute value for knee joint angle. Therefore, the values for knee joint angle at heel strike should be regarded with caution. This error does not, however, affect the range of motion values in the study because range of motion is an interval measure of two absolute values at different instances, which cancels out this error.

The vertical, anteroposterior, and mediolateral orthogonal components of the ground reaction force (GRF) were measured at a sample frequency of 1000 Hz using a Kistler force plate (type 9281, Kistler Instrument Corp., Winterthur, Switzerland). The force plate was located in the middle of a 17-m-long runway and was concealed with a thin black cloth to avoid targeting. A fourth-order bi-directional Butterworth filter was used to low-pass filter the GRF data at a cut-off frequency of 100 Hz. Maximal initial loading rate, impact and active peak forces, and impulse in the vertical force direction together with braking and propulsive peak forces and impulses in the anteroposterior force direction were selected as representative GRF descriptors (3,22,23). Maximal initial loading rate was defined as the maximal slope in the GRF curve between heel strike and the instant of impact peak force. Mediolateral GRF were not analyzed because of their relatively low values and large variability across individuals (3,22).

Running speed was assessed using photocells positioned at shoulder height along the runway, just in front of the force plate. Stride length was determined using tape marks placed 10 cm apart across the floor in the middle of the runway and a video camera (25 Hz) that captured a full stride of the subject. Stride length was defined as the distance between two successive foot strikes of the left foot. Stride frequency was calculated by dividing running speed with stride length.

All subjects wore the same type of running shoe (Asics A6-gel, biomechanically neutral) in order to minimize the effect of shoes on the data (7,23). The shoes were worn without socks to provide a snug fit of the foot inside the shoe.


Subjects were tested at both a self-selected running speed (SRS) and a controlled running speed (CRS) of 3.3 m·s−1. Running at a self-selected speeds was chosen to allow a biomechanical comparison during the conditions in which both groups normally run, whereas running at a controlled speed provides the opportunity to assess the effect of age on the biomechanics of running, independent of speed effects (22,23). The chosen CRS was comparable to speeds reported for distance running (2,6,22). The SRS condition was executed first to avoid an influence of the CRS on the SRS. After their usual individual warm-up, subjects were given as many practice runs as needed to habituate themselves with the test-procedures and the shoes worn. This familiarization period was also used to determine the SRS. Subjects were instructed to run along the runway in a natural manner. Eight successful trials per condition had to be completed. A successful trial was defined as one in which the right foot fully hit the force plate, the subjects’ running speed remained within a margin of 5% of the predetermined level, and the subject displayed his normal running pattern.

Statistical analysis.

Due to the absence in the literature of GRF and kinematics data from older-aged runners at speeds typical for distance running, a statistical power analysis was not performed before data collection. However, a post hoc power analysis of the results obtained from the 13 younger- and 16 older-aged subjects in the study demonstrated a sufficient level of power of 76% or greater (range 76–99%) in all variables hypothesized to be different between the age groups.

For each dependent variable, the mean of the eight trials per condition per subject was used for statistical analysis. Independent t-tests using SPSS statistical software (SPSS Inc., Chicago, IL) assessed statistical significance between age groups. In view of the explorative nature of this study, no Bonferroni adjustments for multiple comparisons were made (28). A significance level of 0.01 was chosen for the purpose of increasing the biological relevance of any significant difference found in the study. For the CRS condition, Pearson correlation values were obtained between selected kinematic and kinetic variables (P < 0.01) in order to give indications for possible explanations of the differences found between the younger- and older-aged subjects.


The SRS of the older runners was significantly lower than the SRS of the younger participants (P < 0.001, Table 2). In both speed conditions, stride length was significantly shorter and stride frequency significantly higher in the older subject group (P < 0.001).

Mean values (SD) for all dependent variables in the younger- and older-aged runners for the self-selected (SRS) and controlled (CRS) running speed conditions.

Because the results on most kinematic variables were similar in the SRS and CRS conditions (Table 2), only the results for the CRS condition are reported here. However, where a significant difference between the groups was present at the SRS only, this is mentioned. Visual inspection of Figures 2–4 shows that larger differences between the groups were present in the knee joint when compared with the ankle and subtalar joints. At heel strike, knee flexion angle was significantly larger in the older runners compared with the younger subjects (P < 0.01). Knee flexion ROM in the first half and knee extension ROM in the second half of stance were substantially smaller in the older subjects (P < 0.001). In the first half of stance, only minor differences in ankle joint motion were present between the groups (Fig. 3). Ankle dorsiflexion ROM was significantly smaller in the older runners at the SRS (P < 0.01). Ankle plantar flexion ROM during the second half of stance was significantly smaller for the older subjects at the CRS (P < 0.01). Range of motion in the subtalar joint (Fig. 4) was more variable among the subjects than ROM in the ankle and knee joints, as evidenced by a coefficient of variation (CV) ranging from 27 to 31% for subtalar eversion ROM as opposed to 10–16% for ankle and knee joint ROM (calculated from the values shown in Table 2). Neither eversion ROM nor maximum eversion angle was significantly different between the groups. Vertical impact speed was significantly higher in the older runners at the CRS (P < 0.01); no significant age effect was found at the SRS.

Mean knee joint angle expressed as a function of contact time for the younger- and older-aged subjects in the controlled running speed condition.Error bars represent one standard deviation. HS denotes heel strike, TO toe-off.
Mean ankle joint angle expressed as a function of contact time for the younger- and older-aged subjects in the controlled running speed condition.Error bars represent one standard deviation. HS denotes heel strike, TO toe-off.
Mean subtalar joint angle expressed as a function of contact time for the younger- and older-aged subjects in the controlled running speed condition.Error bars represent one standard deviation. HS denotes heel strike, TO toe-off.

In the SRS condition, no differences were found between the groups in maximal initial loading rate and impact peak force (Table 2 and Fig. 5). The active peak vertical GRF (P < 0.001) and vertical force impulse (P < 0.01) were significantly reduced in the older runners. All anteroposterior GRF variables (braking and propulsive peak forces and impulses) also showed significantly lower values in the older runners. In the CRS condition, maximal initial loading rate and impact peak force were significantly higher for the older runners than for the younger subjects (P < 0.01, Table 2 and Fig. 6). From the other selected GRF variables, only the vertical force impulse was significantly different between the age groups, with lower mean values in the older subjects (P < 0.001). The vertical force impulse between 30% and 80% of stance in the CRS condition was 0.28 BW·s−1 (SD 0.01) for the younger and 0.25 BW·s−1 (SD 0.01) for the older subjects, which was a statistically significant difference (P < 0.001).

Mean vertical (top) and anteroposterior (bottom) ground reaction forces in units of body weight (BW) expressed as a function of contact time for the younger- and older-aged subjects in the self-selected running speed condition.Error bars represent one standard deviation. Data are slightly attenuated as a result of point-to-point averaging of asynchronous occurring events across individuals.
Mean vertical (top) and anteroposterior (bottom) ground reaction forces in units of body weight (BW) expressed as a function of contact time for the younger- and older-aged subjects in the controlled running speed condition.Error bars represent one standard deviation. Data are slightly attenuated as a result of point-to-point averaging of asynchronous occurring events across individuals.

Correlation analysis.

Stride length and frequency were both significantly correlated with vertical and anteroposterior GRF in the propulsive phase of stance (Table 3). Impact peak force and maximal loading rate were significantly correlated with vertical impact speed in the combined group of younger and older subjects (P < 0.01), highly correlated in the subgroup of younger runners (P < 0.001), but not correlated in the older age group (Fig. 7). No significant correlations were present between impact peak force or maximal loading rate and other parameters that have been reported to contribute to the level of impact in running, such as stride length and frequency and knee flexion angle at heel strike (6,8,21). Active peak vertical GRF, vertical force impulse, and anteroposterior propulsive impulse were significantly correlated with ankle plantar flexion ROM, knee flexion ROM, and knee extension ROM.

Correlation coefficients between selected variables for the combined group of younger- and older-aged subjects (N = 29), the younger-aged subject group (N = 13) and the older-aged group of runners (N = 16) at the controlled running speed.
Scatter plot for the correlation between vertical impact speed at heel strike and impact peak force in the older- and younger-aged subject groups at the controlled running speed condition. Linear regression lines with R-squared values are also included.


Ground reaction forces.

The results under controlled speed conditions showed that impact peak force and maximal initial loading rate were significantly higher in the older subjects when compared with the younger subjects. Several authors suggest that impact force and loading rate are positively related to stride length and vertical impact speed, and inversely related to knee flexion angle at heel strike (6,8,21,23,24). Gerritsen et al. (9) used forward dynamics simulations to model the impact phase of running and found knee flexion angle at heel strike and vertical impact speed to be strong independent determinants of impact peak force. Despite that the older runners had a significantly shorter stride length and significantly larger knee flexion angle at heel strike than the younger runners in the present study, impact peak force and loading rate were significantly higher in the older subjects. Moreover, these basic stride and kinematic variables were not significantly correlated with the impact-loading parameters. In contrast, vertical impact speed was significantly higher in the older runners and significantly positively related to impact peak force and loading rate in the total group of 29 subjects. Therefore, this variable may play a role in explaining the differences found during impact between the older and younger runners. However, the variance of impact peak force and loading rate explained by vertical impact speed was largely discrepant between the younger (74% and 69%, respectively) and the older runners (0% and 0%, respectively). The reason for this discrepancy as well as for the difference in vertical impact speed between the age groups is not clear from the results of this study; the correlation analysis failed to show any significant association between vertical impact speed and other kinematic or kinetic variables. Fat pad status in the heel was not assessed in this study but is an intrinsic factor that influences impact force and loading rate in running (24). The heel pad has good shock-absorbing properties (1,13) but is known to atrophy and loose elasticity in aging, which compromises shock attenuation (13–15). Such a loss of heel pad function may have been present in the older runners and therefore could have explained the higher impact peak forces and loading rate when compared with the younger subjects.

Irrespective of their cause, the higher impact loads in the older runners at the controlled running speed condition suggest a decrease in the shock-absorbing capacity of the musculoskeletal system in the older runners and, subsequently, an increase of the load on bone, joints, and soft tissue in the lower extremities. However, while running under normal conditions, i.e., at a self-selected speed, the levels of impact peak force and loading rate were similar between the younger- and older-aged runners. This may suggest that older-aged runners, conscious or not of their inherent decrease in capacity for shock absorption, “adjust” their gait pattern in such a way (i.e., by lowering speed) that impact peak force and loading rate are reduced. Nevertheless, the cumulative effect of these similar impact levels is, despite the potential presence of such a protective mechanism, still expected to be more adverse in older- than in younger-aged runners for two reasons. One, involuntary musculoskeletal degeneration at older age presumably makes the locomotor system more susceptible to damage by a given repetitive high level of impact force as occurs in running (26). If this damage and resulting loss of function would occur in musculoskeletal structures used for shock absorption during impact, then higher impact forces may be expected, which would then again be more threatening to the biological system. Two, older runners seem to undergo the same levels of impact more frequently than younger runners in their usual running schedule due to a significantly smaller stride length together with a similar self-reported weekly running distance (Table 1); older-aged runners simply take more steps than younger runners do in which the repetitive pounding action of high impact may have a detrimental effect. A simple calculation of the amount of steps taken per week by the older- and younger-aged subjects using information on self-reported weekly running distance (22.3 km for both groups in the present study) and stride length (2.91 m in the young and 2.41 m in the older subjects) illustrates that the older runners take 21% more steps per week of running than the younger runners (18,506 vs 15,326 steps).

Even though studies have shown that high-impact exercise, such as running or jump training can be beneficial for bone strength, muscle performance, and aerobic capacity (11,26), repetitive high impact loading is generally assumed to play a causative role in the etiology of overuse injuries in running (12,16,26). Recently, Hreljac et al. (12) found significantly higher impact forces and loading rates in runners who sustained an overuse injury when compared with injury-free runners and suggested that runners with higher impact loading are at an increased risk of incurring overuse running injuries. The incidence of lower-extremity injuries is high in both younger- and older-aged runners (17,19,20,26,30). A few studies report a higher rate of overuse injuries in older-aged runners or athletes (20,26), whereas others do not find an age effect in lower-extremity injury incidence (17,30). Additionally, the injury rate in older runners may be higher than reported due to a “healthy runner effect” where older-aged injured runners, because of their advanced age, discontinue their running activities and are therefore missed in surveys on injury incidence among active runners (19,20). If older runners do sustain more overuse injuries than younger runners, then the repetitive action of high impact peak forces and loading rates may play a role in explaining this difference. For the design and prescription of running shoes, these increased impact loads show that greater emphasis should be put on providing optimal cushioning for older-aged runners in order to decrease shock at heel strike. The absence of data on injury incidence in this study does not allow drawing definitive conclusions for the prescription of exercise in this age group. However, the reduction of the shock-absorbing capacity of the older-aged body suggests that older-aged runners should be cautious with running under high impact conditions and may be better off with running at lower speeds as a means of reducing impact on the body (22,23) to a, presently unknown, level that is still beneficial for improving general health and bone strength but may reduce the potential for injury.

It was hypothesized that under controlled speed conditions the older runners would show significantly smaller peak GRF and impulses in the propulsive phase of stance due to an expected loss of lower-extremity muscle strength and contraction velocity. Such a loss of motor function is known to occur above the age of 50, independent of physical activity throughout life (18,32). Lower GRF and impulses are then seen as a necessary consequence in order to reduce the percentage of available strength needed to propel the body toward toe-off, which seems logical from the perspective of economics in distance running where the goal is to minimize metabolic energy expenditure (2).

However, despite a trend toward lower GRF values for the older-aged runners, with borderline significant P-values for the active peak vertical GRF (P = 0.03), peak propulsive anteroposterior GRF (P = 0.02), and propulsive anteroposterior impulse (P = 0.02), only the vertical GRF impulse was significantly lower (P < 0.001) in the older runners when compared with the younger subjects. This absence of significant differences may suggest that lower-extremity muscle strength was not reduced in the older-aged runners examined or that this factor does not influence peak forces during running. However, as can be seen clearly in Figure 6, a lower vertical GRF was consistently present in the older runners during propulsion, which was demonstrated quantitatively by a significant decrease in GRF impulse between 30% and 80% of stance in the older subject group (P < 0.001). This wide range of lower forces may well be the result of reduced lower-extremity muscle strength in the older runners, even though definitive conclusions cannot be drawn because strength was not measured. The vertical GRF impulse was strongly inversely associated with stride frequency (r = −0.86). This could be expected because the cumulative impulse produced over a certain distance (impulse times number of steps) should be constant across subjects when speed is controlled. Cause and effect in this relationship is, however, unclear. Vertical force impulse could be the dependent variable if a certain combination of stride length and frequency is chosen for safety or economic reasons or, for example, because joint stiffness is increased. This implies that other factors than reductions in muscle strength may have also been responsible for the lower vertical forces and impulses in the second half of stance in the older runners.

The large number of significant differences in vertical and anteroposterior propulsive GRF in the self-selected running speed condition seems to be the primary result of differences in running speed between the older and younger subject groups.

Kinematics and basic stride variables.

The joint kinematics of the older runners was characterized in both conditions by more knee flexion at heel strike, and substantially smaller knee flexion and extension ROM during the first and second half of stance, respectively. These results are in agreement with findings in the literature on the kinematics of competitive running and sprinting in older people (10,25). Okada et al. (25) studied the kinematics of 46 male competitors in a 5000-m race, ranging in age between 40 and 81 yr. Knee joint ROM was negatively correlated with age. However, because all competitors ran at their self-selected speed, which was substantially lower in the older age groups, the effects of age and speed on these findings could not be distinguished. Hamilton (10) studied the kinematics of sprinting in 162 men and women between the ages of 30 and 90, and also found that knee joint ROM was inversely correlated to age, independent of running speed.

In the present study, knee and ankle joint ROM showed multiple highly significant positive correlations with propulsive GRF (Table 3). Therefore, differences in GRF may explain a large percentage of the variance in the kinematics in both groups even though this assumption should be treated with caution because cause-and-effect relationships could not be established. Chances might be even that differences in joint ROM explain the findings on GRF. A larger ROM, for example, increases the arc over which force can be produced, resulting in higher peak values, as suggested by Hamilton (10).

Lower-extremity joint stiffness was not measured in this study, but is known to increase with age (29,31). Possible larger degrees of joint stiffness in the older-aged runners may have diminished their joint ROM in running. The larger degrees of knee flexion at heel strike in the older-aged runners may also be a direct result of increased joint stiffness, which limits these older subjects in actively stretching the legs before touchdown. Alternatively, more strategic reasons might have played a role too, in that older runners obtain larger flexion angles at heel strike in order to attempt to reduce the load on the musculoskeletal system directly after impact. Together with the findings of reduced ROM in the joints of the lower extremity, these results reflect a “cautious” gait pattern that older runners may obtain because they strive for more stability and a lower risk of falling.

In the ankle joint, smaller and fewer significant differences were found between the older- and younger-aged runners than in the knee. In support of this, Okada et al. (25) found no significant correlation between age and ankle ROM in competitive running. As an explanation for these results, increases in joint stiffness with age may simply be larger in the knee than in the ankle, although there is no evidence for this in the literature. Alternatively, adaptive strategies in the knee may be considered by older-aged runners to have a greater effect on impact loading than adaptive strategies in the ankle. Several studies have demonstrated the impact reducing effect of larger knee flexion angles at heel strike (8,9,21). Similar effects from changes in ankle joint angle are not known to occur. Therefore, older runners may alter their running pattern in the knee and not in the ankle for this purpose.

Subtalar joint eversion ROM and maximum eversion angle are representative parameters for rearfoot control in running. Rearfoot control has, in addition to cushioning, high priority in the design of running shoes because excessive pronation has been suggested to cause lower-extremity injuries (7,16). Improved cushioning, recommended for older-aged runners due to their high levels of impact, can occur at the cost of rearfoot stability and vice versa. The present results indicate that this may not be a more important issue for the older runner because no significant differences in maximal eversion angle and eversion ROM were found between the age groups.

As expected, the self-selected running speed of the older runners was lower than that of the younger runners. The large age group difference in stride length (0.50 m) was directly responsible for this finding. The difference in stride length between the age groups at the CRS was much smaller (0.18 m). In contrast, the age differences in stride frequency were not much different across speed conditions (0.08 and 0.10 Hz for at the SRS and CRS, respectively). These results confirm findings from Hamilton (10) in sprinting, Okada et al. (25) in competitive 5000-m running, and Cavanagh and Kram (2) in younger-aged distance running and draw to the conclusion that decreases in running speed depend largely on reductions in stride length, not on changes in stride frequency.

Stride length and frequency were not scaled to body height even though the older runners were significantly shorter than the younger participants and despite the fact that scaling is common practice in gait and running studies. However, the correlation value between height and stride length at the CRS was low for the younger age group (r = 0.29) and even inverse for the older-aged runners (−0.16); both were not statistically significant (P > 0.3). Additional evidence in the literature confirms that taller individuals do not necessarily take longer strides at a given running speed. For example, Cavanagh and Williams (5) and Cavanagh and Kram (2) have reported correlation values between body height and stride length in running of 0.09 and 0.26, respectively. Cavanagh et al. (4) reported an inverse correlation of −0.10 between leg length and stride length. This suggests that anthropometric measures such as height are not the primary determinants of basic stride variables in running and questions the need for normalization. More likely, other factors, such as ground reaction force production, determine the magnitude of stride length and frequency as indicated by the many significant correlations between most vertical and anteroposterior propulsive GRF parameters and these stride variables (Table 3). Alternatively, as mentioned before, joint stiffness, different demands for safety (i.e., injury avoidance) and/or minimization of metabolic cost (running economics) (5) may be significant contributors to the stride length and frequency obtained by an individual distance runner, and may partly explain the significant differences found between the older and younger runners.

There is no reason to suggest that the subjects in this study represent an unusual subset of younger- and older-aged male runners from the general running population. The younger subjects were significantly taller than the older subjects, and they were also slightly lighter. Body type was therefore quite different between the groups, as indicated by a significantly higher body mass index (BMI) in the older than in the younger subjects (Table 1). To determine whether the difference in BMI may have affected running style and with that the results in the study, we calculated the correlation values between BMI and all kinematic and kinetic variables in the study for each age group within a condition. No significant correlations were found with a highest correlation of 0.51 between BMI and knee angle at heel strike in the younger-aged subjects at the CRS. These results suggest that differences in body type did not affect the results of the study.

The subjects did not wear their own running shoes during the experiments but wore a shoe that was designated as neutral in its biomechanical properties, meaning that no anti-pronation or anti-supination elements were incorporated in the outsole of the shoe. The unfamiliarity with the experimental shoes may have influenced the individual data to an unknown degree. However, only 4 of 13 shoes of the younger subjects and only 2 of 16 shoes of the older subjects had antipronation or antisupination elements incorporated or had a special inlay. Thus, the large majority of subjects wore shoes that were biomechanically similar to the shoes used in the experiment. Additionally, all subjects were given as much time as needed to familiarize themselves with the experimental procedures before data collection, which included getting used to the shoes worn. We therefore assume that all runners felt comfortable with the shoes worn during data collection and do not expect that this issue has affected the comparison between the older and younger subjects to a significant degree.


This study has shown that older- and younger-aged subjects are biomechanically different in the way they run. Older-aged runners use shorter steps at a higher frequency and display smaller knee ranges of motion, higher vertical impact speeds, higher impact peak forces, and higher initial loading rates than younger-aged runners. The causes for these differences are presently unknown, but they may be the result of the delicate interaction between kinematic and kinetic factors influenced by such factors as musculoskeletal degeneration, demands for safety, and/or minimization of metabolic cost in distance running.

The findings of increased impact peak force and loading rate in the older subjects, indicate that the shock absorbing capacity of the musculoskeletal system is compromised in older-aged runners. This may increase their susceptibility to lower-extremity overuse injuries and may explain a possible higher incidence of overuse injuries in older than in younger-aged runners. Moreover, these findings emphasize the focus on cushioning properties in running shoe design and prescription for older runners. Finally, these results suggest that older-aged runners should be cautious with running under conditions of high impact and may be better off running at lower speeds as a way of reducing shock at heel strike. Prospective data on injury incidence would be necessary to make definitive recommendations on exercise prescription for older-aged runners. It is clear from the results that biomechanical data from older-aged runners and not existent data on younger-aged runners should be used in the future in these applied fields of research. The combination of testing subjects at self-selected as well as controlled speeds proved useful for determining the contribution of age, independent of speed effects, to the biomechanics of running, as well as for discussing the implications of running at older age for locomotor function and the prescription of shoes and exercise.

Future research should expand on the presented information by determining the direct relationships between specific musculoskeletal changes as a consequence of aging (e.g., loss of fat pad function, decrease in muscle strength, increase in joint stiffness) and biomechanical characteristics of running in larger groups of runners stratified to age, preferably in a prospective study design. This way, a better understanding of the mediating intrinsic factors in altered biomechanics in older-aged runners can be expected. The calculation of joint moments and power was beyond the scope of the present study, but this may also provide more insight into the mechanisms behind changes in kinematics and GRF in older-aged runners. Case-control or prospective cohort studies on the incidence of overuse injuries in younger- and older-aged runners of whom the levels of impact force and loading rate during running are known should give more insight into the risk of higher impact loading in older-aged runners for developing overuse injuries.

The author is grateful to Cor Verwaaijen for his technical assistance, Huub Maas for his help in data collection and analysis, and Pauline Kok, René Wijlens, and Dr. Jos de Koning for their advisory roles in this research project. Appreciation is extended to Dr. Ruud Selles and Dr. Maarten Bobbert for their useful comments on the manuscript.


1. Bennett, M. B., and R. F. Ker. The mechanical properties of the human subcalcaneal fat pad in compression. J. Anat. 171: 131–138, 1990.
2. Cavanagh, P. R., and R. Kram. Stride length in distance running: velocity, body dimensions, and added mass effects. Med. Sci. Sports Exerc. 21: 467–479, 1989.
3. Cavanagh, P. R., and M. A. Lafortune. Ground reaction forces in distance running. J. Biomech. 13: 397–406, 1980.
4. Cavanagh, P. R., M. L. Pollock, and J. Landa. A biomechanical comparison of elite and good distance runners. Ann. N. Y. Acad. Sci. 301: 328–345, 1977.
5. Cavanagh, P. R., and K. R. Williams. The effect of stride length variation on oxygen uptake during distance running. Med. Sci. Sports Exerc. 14: 30–35, 1982.
6. Clarke, T. E., L. B. Cooper, C. L. Hamill, and D. E. Clark. The effect of varied stride rate upon shank deceleration in running. J. Sports Sci. 3: 41–49, 1985.
7. Clarke, T. E., E. C. Frederick, and C. L. Hamill. The effects of shoe design parameters on rearfoot control in running. Med. Sci. Sports Exerc. 15: 376–381, 1983.
8. Denoth, J. Load on the locomotor system and modeling. In: Biomechanics of Running Shoes, B. M. Nigg (Ed.). Champaign, IL: Human Kinetics, 1986, pp. 63–116.
9. Gerritsen, K. G., A. J. van den Bogert, and B. M. Nigg. Direct dynamics simulation of the impact phase in heel-toe running. J. Biomech. 28: 661–668, 1995.
10. Hamilton, N. Changes in sprint stride kinematics with age in master’s athletes. J. Appl. Biomech. 9: 15–26, 1993.
11. Heinonen, A., P. Kannus, H. Sievanen, et al. Randomised controlled trial of effect of high-impact exercise on selected risk factors for osteoporotic fractures. Lancet 348: 1343–1347, 1996.
12. Hreljac, A., R. N. Marshall, and P. A. Hume. Evaluation of lower extremity overuse injury potential in runners. Med. Sci. Sports Exerc. 32: 1635–1641, 2000.
13. Hsu, T. C., C. L. Wang, W. C. Tsai, J. K. Kuo, and F. T. Tang. Comparison of the mechanical properties of the heel pad between young and elderly adults. Arch. Phys. Med. Rehabil. 79: 1101–1104, 1998.
14. Jahss, M. H. Geriatric aspects of the foot and ankle. In: Clinical Geriatrics, 3rd Ed, I. Rosmann (Ed.). Philadelphia: Lippincott, 1986, pp. 567–578.
15. Jahss, M. H., F. Kummer, and J. D. Michelson. Investigations into the fat pads of the sole of the foot: heel pressure studies. Foot Ankle 13: 227–232, 1992.
16. James, S. L., B. T. Bates, and L. R. Osternig. Injuries to runners. Am. J. Sports Med. 6: 40–50, 1978.
17. Lane, N. E., D. A. Bloch, P. D. Wood, and J. F. Fries. Aging, long-distance running, and the development of musculoskeletal disability: a controlled study. Am. J. Med. 82: 772–780, 1987.
18. Larsson, L., G. Grimby, and J. Karlsson. Muscle strength and speed of movement in relation to age and muscle morphology. J. Appl. Physiol. 46: 451–456, 1979.
19. Marti, B., J. P. Vader, C. E. Minder, and T. Abelin. On the epidemiology of running injuries: the 1984 Bern Grand-Prix study. Am. J. Sports Med. 16: 285–294, 1988.
20. Matheson, G. O., J. G. Macintyre, J. E. Taunton, D. B. Clement, and R. Lloyd-Smith. Musculoskeletal injuries associated with physical activity in older adults. Med. Sci. Sports Exerc. 21: 379–385, 1989.
21. Mcmahon, T. A., G. Valiant, and E. C. Frederick. Groucho running. J. Appl. Physiol. 62: 2326–2337, 1987.
22. Munro, C. F., D. I. Miller, and A. J. Fuglevand. Ground reaction forces in running: a reexamination. J. Biomech. 20: 147–155, 1987.
23. 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: 951–959, 1987.
24. Nigg, B. M., G. K. Cole, and G. P. Bruggeman. Impact forces during heel-toe running. J. Appl. Biomech. 11: 407–432, 1995.
25. Okada, H., M. Ae, T. Kotoh, and K. Asano. Biomechanical study on changes in running kinematics with aging in master’s distance runners. In: Abstracts of the XIVth Congress of the International Society of Biomechanics. Paris: Pergamon, 1993, pp. 982–983.
26. Pollock, M. L., J. F. Carroll, J. E. Graves, et al. Injuries and adherence to walk/jog and resistance training programs in the elderly. Med. Sci. Sports Exerc. 23: 1194–1200, 1991.
27. Riggs, B. L., and L. J. Melton, III. Involutional osteoporosis. N. Engl. J. Med. 314: 1676–1686, 1986.
28. Rothman, K. J. No adjustments are needed for multiple comparisons. Epidemiology 1: 43–46, 1990.
29. Such, C. H., A. Unsworth, V. Wright, and D. Dowson. Quantitative study of stiffness in the knee joint. Ann. Rheum. Dis. 34: 286–291, 1975.
30. van Mechelen, W. Running injuries: a review of the epidemiological literature. Sports Med. 14: 320–335, 1992.
31. Vandervoort, A. A., B. M. Chesworth, D. A. Cunningham, D. H. Paterson, P. A. Rechnitzer, and J. J. Koval. Age and sex effects on mobility of the human ankle. J. Gerontol. 47: M17–M21, 1992.
32. Vandervoort, A. A., and A. J. Mccomas. Contractile changes in opposing muscles of the human ankle joint with aging. J. Appl. Physiol. 61: 361–367, 1986.


©2003The American College of Sports Medicine