The ability to run fast is a major factor determining level of performance in various sporting and recreational physical activities. However, with aging, the ability to perform short sprints declines and may eventually inhibit participation in physical activities. Given the growing number of people remaining highly physically active into older age, information on the mechanisms responsible for the decline in maximal locomotor performance may be of increasing overall importance and have implications for the planning of exercises to preserve and potentially improve speed performance in older people. Accordingly, examination of the specific biomechanical and physiological changes that impair maximum running velocity (Vmax) in older athletes is needed to understand the degree to which these changes are preventable by regular training. So far, studies of age-related changes in Vmax have focused mainly on stride cycle parameters (12,20). However, stride parameters do not provide information on the factors underlying changes in performance; hence, it would be important to determine the interrelationship between stride cycle parameters, ground reaction forces (GRFs), and neuromuscular factors.
The two basic mechanisms that determine Vmax are the forces generated to the ground and the speed at which the legs can be swung forward and backward (36). On the basis of studies in young adults, Vmax is primarily related to the magnitude of contact forces that the runners are capable of producing during the short contact phase. For example, in the study by Weyand et al. (36), a 1.8-fold intersubject difference in Vmax (from 6.2 to 11.1 m·s−1) and 1.7-fold difference in stride length were attributable to a 1.3-fold change in vertical GRF. The faster runners also had higher stride frequency. However, this was achieved by reduced contact time, whereas the time taken to swing the limb into the position for the next stride did not vary between runners. These findings are in line with the studies by Mero (25) and Mero and Komi (26) that used resultant GRF as a specific force indicator in young male and female runners with different sprinting abilities. The higher Vmax and longer strides of the fastest runners were explained by their superior ability to produce greater ground contact forces.
Skeletal muscle characteristics have been commonly identified as critical determinants of sprint performance. Investigations in young athletes have suggested that sprinting ability is limited by muscle mass (3,23,35), fiber type composition (5,25,27), and fascicle length (3,23). However, it is not uncommon to observe wide variation in muscle properties among athletes of similar performance level (2,27). This suggests that the ability to run fast depends on a combination of factors. In addition to muscle structural characteristics, several studies have shown that whole-muscle mechanical performance is associated with sprinting ability (11,39). On the other hand, the association between muscle strength and sprint performance may be influenced, among other factors, by the type of contraction. For example, there is some evidence to suggest that rapid muscle force capacity is more closely related to sprint performance than maximal strength (9,22). However, the extent to which these findings in young sprint athletes apply to older runners remains unknown.
Accordingly, the main purpose of this study was to examine specific biomechanical characteristics that may determine the ability to run at maximum speed with increasing age. To address this aim, we measured both GRF and stride cycle parameters in male sprinters aged 17-82 yr. The secondary aim was to identify age-related differences in the morphological and mechanical properties of the lower limb muscles and to determine the relationship between the muscle characteristics and sprint performance.
This study was part of a larger investigation on aging, speed performance, and skeletal muscle characteristics (19). Eighteen young adult (17-33 yr) and 59 elite master (40-82 yr) male sprinters were chosen for the present study. The subjects arrived from all over Finland. All 77 athletes were included in sprint performance tests, but because of the subjects' decisions not to complete certain assessment tasks (fear of injury, timing of testing not suitable for the current training/competition preparation schedule), the muscle strength and structural characteristics data were obtained from 58 to 77 subjects. The missing data were mainly from the youngest age group. The subjects had achieved good national- or international-level results in 100- to 400-m sprint events and were continuing to train and compete systematically. The runners in different age groups had similar relative performance level: the 60-m running times in this study ranged from 109 ± 3% (17-33yr) to 106 ± 4% (70-82 yr) of the age-group world record times. Furthermore, there were no significant age-group differences in the reported all-time personal best 100-m times (ranged from 10.96 ± 0.28 s in the 17-33 yr age group to 11.46 ± 0.30 s in the 70-82 yr age group, n = 37). A questionnaire indicated a decrease in total training hours (from 10.6 to 6.4 h·wk−1, P < 0.001) and sessions (from 5.7 to 3.9 times per wk, P < 0.001) from the youngest to the oldest group, whereas training years increased with age (from 12.8 to 27.3 yr, P < 0.05). With increasing age, the number of hours spent on speed and strength training decreased (from 5.3 to 3.1 h·wk−1, P < 0.001, and from 4.8 to 1.0 h·wk−1, P < 0.001, respectively), whereas the amount of other exercises (aerobic running, cross-country skiing, ball games) tended to increase with age (from 0.5 to 2.3h·wk−1, P = 0.08). A written consent was obtained from all subjects. The Ethics Committee of the University of Jyväskylä, in conformity with the Declaration of Helsinki, approved the study.
Body height was measured with a height gauge, and body mass was measured with a balance beam scale. Total body fat percentage was assessed using bioelectrical impedance (Spectrum II; RJL Systems, Detroit, MI). Upper leg length was determined as the distance from the lateral condyle of the femur to the greater trochanter, and lower leg length was determined as the distance between the lateral malleolus of the fibula and the lateral condyle of the femur. Leg length was calculated as the sum of the upper and lower leg lengths and was used for normalization of the biomechanical variables.
Knee extensor (KE) and ankle plantar flexor (PF) morphological characteristics were measured using B-mode ultrasonography (SSD-1400; Aloka, Tokyo, Japan), as described earlier (21,23), with slight modifications. Briefly, a 5-cm linear-array probe (7.5 MHz) was positioned perpendicular to the surface of the muscle, and in the ultrasonic images, the subcutaneous adipose tissue layer, superior and inferior aponeurosis, and fascicles between the aponeuroses were identified. The thickness of the KE muscles (vastus lateralis (VL), vastus medialis (VM), vastus intermedius (VIM), rectus femoris (RF)) and PF muscles (gastrocnemius medialis (GM), gastrocnemius lateralis (GL)) was determined as the distance from the adipose tissue-muscle interface to the intermuscular interface. The measurements for the VM were taken at 30% and those for the VL, VIM, and RF at 50% of the distance between the lateral condyle of the femur and the greater trochanter (for VIM two locations: under VL and under RF), whereas the anatomic site for the GM and GL at 30% proximal between the lateral malleolus of the fibula and the lateral condyle of the tibia. The sum of the muscle thicknesses of all the knee extensor and plantar flexor muscles (KE + PF muscle thickness) was used as an indicator of muscle mass. The pennation angle was measured as the angle between the fascicle and the deep aponeuroses, and fascicle length was estimated from the muscle thickness and pennation angle as follows: Fascicle length = Muscle thickness × sin (pennation angle)−1. Fascicle pennation angle and length were determined from the VL, GM, and GL. The ratios of VL fascicle length to upper leg length and of GM and GL to lower leg length were also examined. The interday repeatability (coefficient of variation, CV%) of the muscle architectural measurements using this method has previously been shown to be 2.5% for muscle thickness, 3.8% for pennation angle, and 5.0% for fascicle length (21). In addition, previous studies have provided evidence that muscle thickness, as determined by ultrasonography, is a good estimate for muscle volume (28) and muscle mass (32).
Muscle biopsy and histochemical analyses
Details of the muscle sampling procedure and myofibrillar ATPase histochemical analyses have been described earlier (19). Briefly, muscle biopsy samples were taken from the middle portion of the vastus lateralis of the dominant leg (take-off foot in jumping events). Serial cryosections (10 μm) were stained for myofibrillar ATPase after acid (pH 4.37, 4.60) and alkaline (pH 10.30) preincubations. The number and cross-sectional area of the various fiber types were analyzed using a microscope combined with a computer-assisted image analysis system (Tema; Scanbeam, Hadsund, Denmark). For the purpose of this study, the fibers were divided into Types I and II fibers. Fiber-type characteristics were calculated from an average of 492 fibers (range, 89-1100) in each biopsy sample.
Dynamic and isometric strength of the lower limb muscles was determined as described previously (8,19). Briefly, one repetition maximum (1-RM) dynamic strength of the leg extensors was measured using a concentric half-squat exercise (from 90° knee angle) in the Smith machine (Frapp fitness, Joensuu, Finland). Dynamic speed-strength ability was assessed by a countermovement jump (CMJ) on a contact mat. Maximum vertical jump height was assessed by determining the rise of the body's center of gravity (calculated from flight time). During the test, the hands were kept on the hips to minimize differences in technique.
Bilateral isometric force of the leg extensors was measured by an electromechanical dynamometer (University of Jyväskylä) with the subject seated with knees and hip at 107° and 110° flexion, respectively. On a verbal command, the subjects performed an isometric leg extension as hard and as fast as possible during 2.5-4 s. Maximal force was defined as the highest force value recorded, and maximal rate of force development (RFD), a measure of isometric speed-strength quality, was defined as the greatest increase in force in a given 5-ms period using 1-ms moving intervals (Δforce/Δtime). Furthermore, specific force (maximal force/KE thickness) and specific RFD (maximal RFD/KE thickness) were determined to provide an insight into the muscles' intrinsic force/speed generation capacities. The subjects performed approximately three to four trials for each strength test (up to five trials in squat 1-RM) until their performance no longer improved.
The subjects ran a maximal 60-m sprint twice on an indoor synthetic track with spiked running shoes (rest between runs of 7 min). The athletes started from a static forward-lean standing position with the front leg 70 cm behind the starting line (first photocell gates). Further, to minimize the aging effect on starting actions (reaction time, technique), own start without commands were used. Vertical and horizontal anteroposterior GRFs and timing parameters were measured during the maximal speed phase (from 30 m onward) using a special 9.4-m-long force platform system. It consisted of nine tartan-surface force plates (five two-dimensional and three three-dimensional force plates, 0.9/1.0 m; natural frequency, ≥170 Hz; nonlinearity, ≤1%; cross talk, ≤2%, TR test, Finland; and one Kistler 3-dimensional force plate, 0.9/0.9 m; natural frequency, 400 Hz; Honeycomb, Kistler, Switzerland) connected in series. The force signals were sampled at 1000 Hz and stored on a microcomputer via an AT Codas A/D converter card (Dataq Instruments, Inc, Akron, OH).
The maximum 10-m running velocity (30-40 m) and the 60-m trial time (t60m) were obtained by four pairs of double-beam photocell gates. In addition, a laser radar (Laveg Sport; Jenoptik, Jena, Germany), positioned 6 m behind the start line, was used to analyze the instantaneous running speeds ofthe athletes. The results indicated that 91% of the athletes had reached their Vmax by the 40-m mark (range, 19-48 m; mean 30 ± 6 m), and those athletes who continued to accelerate up to 40-48 m achieved 99% of their Vmax by the 40-m mark (data not shown). Thus, the 30- to 40-m distance used for force platform measurements represented the phase of maximum velocity sprinting for all the athletes.
The force platform data were analyzed using custom-written software (University of Jyväskylä, Finland). The vertical and horizontal GRFs were integrated with respect to time phases and were then combined to obtain the average resultant GRFs. A horizontal GRF curve was used to divide the force components into the braking and push-off phases (26). Time of total contact (tc), braking phase (tbrake), push-off phase (tpush), swing (tsw), and aerial phase (taer) were obtained from the force platform records (Fig. 1). Stride cycle time (tstr) was defined as the time between consecutive footfalls of the same foot, and stride frequency (Freqstr) is the inverted value of the stride cycle time [1/tstr, (= step frequency / 2)] (36). Duty factor was defined as time of contact relative to entire stride cycle duration (tc/tstr). Stride length (Lstr), the distance between successive contacts of the same foot (step length × 2), was measured from spike marks on a thin paper sheet firmly attached over the force platform. In previous measurements in young sprinters, the accuracy of this method has been ±2 cm compared with film analysis (A. M., unpublished observations).
Each trial involved four to six contacts on the force platform. For consistency, the first four contacts of the two trials were taken for the analyses. Because no significant bilateral differences were found in the biomechanical variables, the results for the right and left sides were averaged for each subject. The intrasubject variability of the biomechanical variables was assessed for 18 young (24 ± 4 yr) and 25 older (70 ± 4 yr) runners. All variables had a coefficient ofvariation (calculated from a total of 8 contacts/4 strides) of0.9%-6.0%. No age-related differences were observed intheCVs.
The amplitudes of the resultant GRFs (Fbrake and Fpush) are given as average net forces relative to body weight (F=(N − Nbw) / kgbw), and the direction of GRF are given as the mean angle of the average net resultant GRF in the braking and push-off phases (25,26). Further, to examine the RFD applied to the ground, the Fbrake and Fpush were divided by the respective tbrake and tpush (expressed as N·s−1·kg−1). Leg length showed significant correlations with Freqstr (r = −0.39, P < 0.001), Lstr (r = 0.39, P < 0.001), and tsw (r = 0.28, P < 0.05) when controlled for speed (partial correlation). Because there was an age-related difference in leg length (Table 1), both absolute and relative leg length-adjusted Lstr, Freqstr, and tsw (i.e., divided by 1/leg length) were analyzed. Figure 1 illustrates the biomechanical parameters examined in this study.
Linear and quadratic regression analyses were performed to determine the association between sprint performance and muscle morphological and functional characteristics with age (Figs. 2-6). The mean percent changes in selected variables across the age range (% per decade) were calculated from the difference in the mean values of the youngest, 17-33 yr (mean, 23.8 yr), and the oldest, 70-82 yr (mean, 74.9 yr), groups. ANOVA was used to determine differences between 10-yr age groups. In the event of significant age effect, Tukey's post hoc test was used to identify the significant differences between each pair of age groups. The mean statistical power for detecting significant (P < 0.05) age effect in various comparisons was 0.95. Three parameters showed power values less than 0.80 (Fpush-angle° = 0.78; pennation angle of VL = 0.74; Leg extension specific RFD = 0.75). Pearson's correlation coefficient was used to examine the relationships between continuous variables, and where appropriate, partial correlation was used to control the effect of age in the relationship. Stepwise multiple regression analyses were conducted to find out the combination of muscle characteristics and age that explained the most variance in Vmax and GRFs and to examine the association of Fpush, Fbrake, and age with Vmax, Lstr, and tc. Statistical significance was set at P < 0.05 for all analyses.
There was an age-related decline in body mass (r = −0.40, P < 0.001), body height (r = −0.63, P < 0.001), and leg length (r = −0.55, P<0.001), whereas percent body fat increased with age (r= 0.30, P < 0.01). Leg length relative to height did not correlate with age (Table 1).
The 60-m running times increased from 6.98 ± 0.17 s in 17- to 33-yr-old runners to 9.23 ± 0.41 s among 70- to 82-yr olds (P < 0.001; +6% per decade; Appendix 1 for age group comparisons). There was a progressive age-related decline in Vmax (−5% per decade) and Lstr (−4% per decade; Fig. 2), which both became significant by the 40- to 49-yr age group in comparison with the youngest group (Appendix 1). Freqstr also showed a reduction with age (−1% per decade; Fig. 2). However, after the significant decline from the youngest (2.25 Hz) to the 40- to 49-yr group (2.13 Hz), Freqstr remained the same until the oldest group (2.12 Hz; Appendix 1). When expressed relative to leg length, the age-related decline remained significant for Lstr (r = −0.77, P < 0.001; −3% per decade) but not for Freqstr (r = 0.10; Appendix 1).
Of the temporal variables (Fig. 3), there was an age-related increase in tc (+5% per decade), tbrake (+9% per decade), and tpush (+2% per decade) and a decrease in taer (−2% per decade), whereas tsw remained unchanged. However, tsw normalized to leg length increased with age (r = 0.44, P < 0.001, +1% per decade). The tpush/tbrake ratio showed nonlinear reduction with age (r2 = 0.17, P < 0.001), reaching significance for the oldest group (Appendix 1). Duty factor (tc/tstr) increased with age (r = 0.73, P < 0.001).
The GRFs of the braking (Fbrake) and push-off (Fpush) phases declined gradually with age (−4% per decade and −6% per decade, respectively; Fig. 4A). The Fpush/Fbrake ratio showed lower values in the 70- to 82-yr-old group (Fig. 4B). The mean angle of Fpush became more vertically oriented with age, whereas no age effect existed in the braking GRF angle (Fig. 4C). The age-associated increase in the mean angle of Fpush reached significance only in the 60- to 69-yr-old group (Appendix 1). Further, the decline in GRFs along with increase in contact times with age led to clear reductions in the rate of GRF development in both braking (Fbrake/tbrake: −9% per decade) and push-off phases (Fpush/tpush: −8% per decade).
There was an age-related decline in KE (−6% per decade), PF (−2.5% per decade), and KE+ PF muscle thickness (−5.5% per decade; Fig. 5; Appendix 2). When compared with the youngest groups, KE and KE + PF thickness showed a decline by age 50-59 yr and thereafter, whereas the age group differences in PF thickness did not reach significance (Appendix 2).
The mean cross-sectional area of Type I fibers showed no significant association with age, but a reduction in Type II fiber area (−7% per decade) and Type II-to-I fiber area ratio (−4.5% per decade) was observed (Fig. 5B). The relative number of Types I and II fibers did not correlate with age.
Pennation angle showed a progressive age-related decline for VL (r = −0.43, P < 0.001) but was not associated with age for GM or GL. For VL, a significant decline in pennation angle was observed in the oldest age group in comparison with the youngest group (Appendix 2).
There were no significant age-related differences in fascicle length for VL, GM, or GL. Furthermore, fascicle length relative to leg length did not correlate with age in the VL, GM, or GL.
There was an age-related decline in maximal dynamic strength (concentric half-squat 1-RM, −9% per decade), maximal isometric force (bilateral leg extension, −8% per decade), vertical jump height (CMJ, −11% per decade), and maximal rate of isometric force development (leg extension RFD, −10% per decade; Fig. 6; Appendix 3). Isometric force normalized to KE muscle thickness was not associated with age (r = −0.17), whereas the isometric RFD/KE thickness ratio declined with age (r=−0.37, P < 0.01). ANOVA of RFD/KE thickness ratio showed an overall decline in with age (P < 0.05), but in the post hoc analysis, the differences between the age groups did not reach significance (Appendix 3).
Relationships among biomechanical parameters
Table 2 shows the simple and age-adjusted correlations between the biomechanical parameters. In the overall sample, Vmax was significantly associated with all parameters, except tsw. Of the parameters, Lstr, tc, tbrake, and Fpush were the best correlates of Vmax. When controlled for age, most of the significant correlations between the parameters remained, with tc showing the strongest relationship with Vmax.
Stepwise multiple regression analysis, with the inclusion of Fpush, Fbrake, and chronological age in the models, was used to determine the association of GRFs with Vmax, Lstr, and tc. Age and Fpush jointly explained 91% and 81%, respectively, of the variance in Vmax and Lstr. For the tc, age, Fpush, and Fbrake were all significant predictors and together explained 70% of the total variance in tc. When age was intentionally excluded from these regression models, both Fpush and Fbrake entered in the equations and together accounted for 53%, 46%, and 61%, respectively, of the variance in Vmax, Lstr, and tc.
Relationships among muscle morphology, strength, and biomechanical parameters
Table 3 shows the age-adjusted correlations among the muscle structure, strength, and running parameters. KE + PF thickness was associated with Vmax and t60m. No other morphological characteristics were related to running parameters (data not shown). Of the muscle strength measures, squat 1-RM, CMJ, and maximal isometric force correlated with the sprint parameters, but no associations were found between isometric RFD and sprint performance (not shown).
Table 4 shows the stepwise regression models of the muscle predictors for Vmax, Fbrake, and Fpush. The variables entered were KE + PF thickness, Type II fiber percentage, Type II-to-I fiber area ratio, maximal isometric force, CMJ, and chronological age (n = 54). Age was the strongest predictor of Vmax and explained 89% of the total variance, with maximal isometric force being the only other factor (1%) to appear in the model. When age was intentionally excluded from the model, CMJand KE + PF thickness appeared in the model and together explained 81% of the variance in Vmax (not shown).
In the GRFs, KE + PF thickness was the only variable to enter the model for Fbrake (26%), whereas maximum CMJheight was the only significant predictor of Fpush (34%). Age did not enter the models for Fbrake and Fpush.
The present study showed that in competitive male sprinters the slowing of maximum running speed with age was characterized by a decline in stride length and an increase in ground contact time along with a lower magnitude of GRFs. The athletes demonstrated age-related changes in muscle structure and force production capacity of the lower limb muscles that contributed to the deterioration in sprint running performance. Figure 7 shows a simplified diagram summarizing the observed biomechanical and skeletal muscle changes with age that may be connected to the decline in sprinting ability.
Kinematic stride cycle parameters
On the first level of the mechanical analysis, we found that a decrease in Lstr contributed more to the decline in Vmax than Freqstr (Fig. 2). Further, the reduction in Freqstr was associated entirely with increased tc because tsw did not vary between runners of different ages. These results are in line with age-related changes in stride cycle parameters in previous studies (12,20). However, because differences in body size may influence stride variables (13), we also adjusted Lstr and Freqstr to leg length. The decline with age was still evident for Lstr but did not persist for Freqstr. This suggests that it was primarily the Lstr aspect of velocity that was affected by age.
Some studies have proposed that the ability to reposition the limbs in the air is an important determinant of sprinting speed (34). However, other studies have challenged this view and concluded that, despite having a more powerful muscle profile, a faster sprint specialist has a similar minimum swing duration to that of slower runners (36). In this connection, it is believed that the mechanical energy required for the repositioning of the swinging leg can occur passively through elastic recoil and energy transfer between body segments rather than by power generated within muscles (36). Our results showed a lack of difference in tsw between the young and older runners (Fig. 3C). However, when tsw was corrected for dimensional changes (i.e., divided by leg length), a significant increase in swing duration was observed in the three oldest age groups in comparison to the youngest group. Therefore, it cannot be ruled out that tsw or the ability to rotate the legs backward and forward is a contributing factor in determining sprinting speed in older age groups. An evaluation of the swing phase from other points of view, e.g., segment and joint angular velocities/accelerations during forward and backward swing phases, might provide further information on this topic. For example, studies involving young elite sprinters have indicated that differences in swing-back velocity immediately before ground contact can explain differences in maximum velocity between runners (4).
Ground reaction forces
The present study demonstrated that the magnitudes of GRFs in sprinting are reduced substantially with age (Fig. 4A) and reflected in changes inLstr, tc, and, consequently, in Vmax. Our findings are in general agreement with those of previous studies that have examined GRF and stride characteristics in young runners with different sprinting abilities (18,25,26,36). For instance, the treadmill running study by Weyand et al. (36) found thatVmax (1.8-fold difference: 6.2-11.1 m·s−1) was highly sensitive to a small variation (1.26-fold difference) in vertical forces. Higher vertical forces had a positive effect on both the maximal Lstr and minimal tc that runners were able to achieve.
In this investigation, resultant GRF was used as a specific force indicator, thereby providing an insight into the interaction between vertical and horizontal force components. Although vertical force dominates the resultant GRF (Fig. 1B) and is likely to have a strong effect on the minimum time needed to be spent on the ground to produce sufficient vertical impulse to support body weight and to create taer long enough for repositioning the swing leg, horizontal GRFs may also play a major role in attaining higher sprinting speed (15). Our results indicated a small but significant change in the mean angle of Fpush (i.e., decrease in the horizontal push-off/vertical GRF ratio) in older runners (Fig. 4C). However, to what extent these age differences in GRF direction, even minor, may impair the acceleration of the body in the optimal horizontal direction and thus affect sprint velocity is unknown at present.
The present study also indicated an age-related decline in Fpush/Fbrake as well as tpush/tbrake ratios, but these decrements were evident only in the oldest group (Fig. 4B). It could be hypothesized that high eccentric impact loads are less tolerated at older ages, resulting in a longer braking phase and a decreased elastic energy/force potentiation during the concentric phase. The consideration of suboptimal stretch-shortening cycle action during ground contact of running in elderly men seems to be supported by the recent study of Cavagna et al. (7). Their results indicated a reduced elastic recovery (−20%) in old (74 ± 6 yr) compared with young (21 ± 2 yr) subjects running at moderate speeds (4.2-4.7 m·s−1). More detailed analysis of the effects of age on support leg movements (14) and spring-like function of muscle-tendon complex (31) during maximum speed running awaits further studies.
Muscle morphology and sprint performance
Studies in young athletes have shown that although sprinters vary in body height, they are typically heavier and have greater muscle mass along with a faster fiber-type profile in their leg muscles than middle- and long-distance runners (5,16). Furthermore, the fascicle length and pennation angle may vary according to running event specialization (3) and could predict overall sprint performance in sprinters (2,23). The present results imply that the decline in the contractile force and velocity potential of muscle with age were mainly attributable to reductions in muscle thickness and Type II fiber area (Fig. 5) because the muscle fiber-type distribution and the muscle architectural characteristics remained largely unchanged. Furthermore, on the basis of our earlier study (19), regular sprint training does seem to reduce the typical age-related decrease in the single-fiber mechanical properties (specific force, shortening velocity) and thus might not be a factor inthe deterioration in the strength and sprint performance in theseathletes.
The combined KE + PF muscle thickness was the strongest predictor of Fbrake in the regression analysis, explaining 26% of the total variance (Table 4). It is worth noting that the lower body weight in older sprinters is partially caused by age-related loss of muscle mass, and the present approach of controlling the average net GRFs for body mass (i.e., (N − Nbw) / kgbw) may underestimate "true" prediction of muscle thickness to GRFs. When the average net GRFs were used in the stepwise regression analysis, muscle thickness showed an increased prediction for Fbrake (50%) and was also a primary predictor for Fpush (54%) with CMJplaying a secondary role (5%; data not shown). Hence, in aging sprinters, muscle volume does seem to play a significant role in the ability to tolerate the great contact forces needed to achieve higher sprinting speeds. This view is consistent with the results of a recent study by Weyand and Davis (35) that estimated the vertical force requirements for different velocities in young male and female athletes. It was concluded that, for runners of same stature and body composition, the larger muscle masses of the faster sprint specialists are directly related to the greater ground contact forces necessary for attaining higher velocities (35).
Muscle strength and sprint performance
Maximal dynamic and isometric leg strength as well as CMJ performance and rate of isometric force development showed a gradual age-related decline (Fig. 6). When isometric force production was normalized to leg muscle thickness, the age-associated decline disappeared for maximal force but remained true for RFD. This slowing of isometric RFD is consistent with our observation of age-related preferential atrophy of fast fibers, although other factors such as rapid neural activation of motor units may also play a role in well-trained older athletes (29). On the basis of the training data, possibly the deterioration in muscle function and associated neuromuscular properties is affected by the lack of proper strength training needed for the effective stimulation of fast motor units. This assumption is also consistent with our recent results (8), indicating selective hypertrophy of Type II fibers along with improved rapid neural activation of muscle, when supplementing sprint training with maximal and explosive strength exercises in elite older sprinters.
Among the muscle functional parameters, CMJwas the only predictor of GRFs, explaining 34% of the variance in Fpush (Table 4). Several studies in young athletes have also indicated that the CMJperformance is related to overall sprint times, Vmax, and stride cycle parameters (9,22,27,39). Apparently, despite many performance dissimilarities (e.g., acyclic vertical vs cyclic horizontal), CMJ, to some extent, simulates fast stretch-shorten cycle contraction of sprint stride contact. On the other hand, the level of prediction of CMJfor Fpush (34%) as well as the age-adjusted associations of the maximal and rapid strength measures with sprint performance were weak in this study (r = 0.23-0.36; Table 3), which is in contrast to many (27,38,39) but not all (37) studies in young athletes. One possible explanation for these low correlations between strength qualities and sprint performance could be that, with age, sprint performance becomes relatively more dependent on technical factors. For example, the elderly sprinters may show greater limb movement constraints (decreased range of motion) because of reduced joint function, which would directly affect stride length (12).
Specificity of training adaptations
Previous findings in untrained (24) and strength-trained (6,30) men have suggested that the age-related decline in rapid muscle force/power-generating capacity occurs at a higher rate than that in maximal isometric and dynamic strength. For example, the study by Pearson et al. (30) among elite master Olympic weightlifters (40-87 yr) and age-matched controls indicated that the rate of loss in maximal isometric knee extension force (−5% to −6% per decade) was approximately half of that observed for leg extension muscle power (−12% to −13% per decade). To address this issue, we calculated changes in strength measures and sprinting velocity on the basis of the values observed in the young athlete group (Fig. 8). Our data suggest that in continuously sprint-trained athletes, the decline in rapid isometric force and vertical jumping capacity (∼−10% to −11% per decade) proceeds only a slightly faster than that in maximal strength (∼−8% to −9% per decade). Furthermore, it was somewhat unexpected that sprint performance (Vmax and 60-m time), which is likely to impose higher requirements on the integration of muscle force production and neuromuscular coordination than the present simple strength tasks, was the least affected by age in these sprinters (∼−5% to −6% per decade). This result seems to confirm the concept of high training specificity and the adaptability of the neuromuscular performance characteristics during aging, and that in large part, the decline in performance in older people is due to a reduction in specific exercise stimulus rather than aging per se. This knowledge is encouraging for aging people who aim to exercise regularly with specific training modes with a view to continuing their higher-intensity sporting and recreational physical activities.
Because of the cross-sectional nature of the study, the findings may have been affected by genetic and constitutional factors. Longitudinal studies are needed to provide more definitive insights into the age-related change in maximum running speed. A potential confounding factor is the assessment of muscle thickness by ultrasonography because this method is unable to distinguish between muscle and intramuscular fat. However, this may not significantly influence our results because the amount of intramuscular fat in the calf muscles of these master sprinters was minimal when examined by computer tomography (H. S., unpublished observations). Nevertheless, a more reliable assessment of muscle mass loss warrants the use of other techniques, such as magnetic resonance imaging, and should also focus on other sprint-specific muscle groups (knee flexors, hip extensors, hip flexors). Although we recruited highly competitive sprinters to minimize the effect of physical activity on our measures, the fact that there were age-related reductions in training volume, especially in strength training, may have contributed to the rate of decline in variables. A similar trend of overall reduction in training stimulus with age has been noted in many studies on endurance athletes and may be explained by age-related reductions in motivation and "intrinsic drive" to train intensively (33) and/or decreased trainability due to physiological changes, e.g., impaired recovery and muscle repair processes (10). Finally, a limitation of the study is that there were incomplete data sets for muscle structure and strength tests. Although no bias is expected because of homogeneity ofthe athletes in different age groups, the sprint performance predictions on the basis of muscle characteristics could be improved with a larger sample size.
The findings of the present study suggest that the age-related decline in maximum velocity sprinting is primarily related to reduction in stride length and increase contact time secondary to decreased ability to generate ground contact forces. Selective muscular atrophy and reduced maximal and rapid muscle force capacity all could contribute to the age-related deterioration in maximal running performance in sprint-trained athletes.
Valuable information on the effects of aging and long-term training on skeletal muscle and physical performance characteristics can be obtained by studying master athletes from different sports. A smaller decrease in sprinting speed compared with strength performance in these athletes supports the concept of training specificity and the favorable effect of regular sprint training on complex locomotor skills during aging (Fig. 8). The present data indicate, however, limited benefit of sprint training alone in prevention of Type II muscle fiber atrophy. This result suggests that the degree of recruitment of fast motor units by short-duration, high-velocity contractions are insufficient for maintaining fast muscle mass and force production. Given the present aswell as other findings (1,17) on the effect of long-term aerobic versus resistance training on muscle characteristics, optimal training should also include intensive strength exercises.
This study was supported by grants from the Finnish Ministry of Education, National Graduate School of Musculoskeletal Disorders and Biomaterials, Finnish Cultural Foundation, Peurunka Medical Rehabilitation Foundation, and Ellen and Artturi Nyyssönen Foundation. The results of the present study do not constitute endorsement by ACSM.
The authors thank Timo Annala, Milan Sedliak, Erkki Helkala, Tuovi Nykänen, and the Master's and Ph.D. students for valuable assistance with the data collection and analysis and all the subjects participating in this study. The authors also to thank the anonymous reviewers for their insightful comments on an earlier draft of the manuscript.
1. Aagaard P, Magnusson PS, Larsson B, Kjaer M, Krustrup P. Mechanical muscle function, morphology, and fiber type in lifelong trained elderly. Med Sci Sports Exerc
2. Abe T, Fukashiro S, Harada Y, Kawamoto K. Relationship between sprint performance and muscle fascicle length in female sprinters. JPhysiol Anthropol Appl Human Sci
3. Abe T, Kumagai K, Brechue WF. Fascicle length of leg muscles is greater in sprinters than distance runners. Med Sci Sports Exerc
4. Ae M, Ito A, Suzuki M. The men's 100 meters. New Stud Athl
5. Andersen JL. Muscle fiber type characteristics of the runner. In: Bangsbo J, Larsen HB, editors. Running & Science
. Copenhagen (Denmark): Munksgaard; 2001. p. 49-65.
6. Anton MM, Spirduso WW, Tanaka H. Age-related declines in anaerobic muscular performance: weightlifting and powerlifting. Med Sci Sports Exerc
7. Cavagna GA, Legramandi MA, Peyre-Tartaruga LA. Old men running: mechanical work and elastic bounce. Proc Biol Sci
8. Cristea A, Korhonen MT, Häkkinen K, et al. Effects of combined strength
and sprint training on regulation of muscle contraction at the whole-muscle and single fibre levels in elite master sprinters. Acta Physiol (Oxf)
9. Cronin JB, Hansen KT. Strength
and power predictors of sports speed. JStrength Cond Res
10. Fell J, Williams D. The effect of aging on skeletal-muscle recovery from exercise: possible implications for aging athletes. JAging Phys Act
11. Häkkinen K, Keskinen KL. Muscle cross-sectional area and voluntary force production characteristics in elite strength
- and endurance-trained athletes and sprinters. Eur JAppl Physiol Occup Physiol
12. Hamilton N. Changes in sprint stride kinematics with age in master's athletes. JAppl Biomech
13. Hoffmann K. Stature, leg length, and stride frequency. Track Technique
14. Ito A, Fukuda K, Kijima K. Mid-phase movements of Tyson Gayand Asafa Powell in the 100 metres at the 2007 World Championships in Athletics. New Stud Athl
15. Keränen T, Nummela A. Ground reaction forces at submaximal and maximal running speeds. In: Strength Training for Sport, Health, Aging and Rehabilitation. 5th International Conference on Strength Training; Oct 18-21
. Denmark: University of Southern Denmark; 2007. p. 344-5.
16. Khosla T. Standards on age, height and weight in Olympic running events for men. Br JSports Med
17. Klitgaard H, Mantoni M, Schiaffino S, et al. Function, morphology and protein expression of ageing skeletal muscle: a cross-sectional study of elderly men with different training backgrounds. Acta Physiol Scand
18. Kobayashi K, Tsuchie H, Matsuo A, Fukunaga T, Kawakami Y. Changes in sprint performance and kinetics during acceleration phase of running of a world record holder. In: The 13th Annual Congress of the European College of Sport Science; July 9-12
. Estoril (Portugal); 2008. p. 591.
19. Korhonen MT, Cristea A, Alen M, et al. Aging, muscle fiber type, and contractile function in sprint-trained athletes. JAppl Physiol
20. Korhonen MT, Mero A, Suominen H. Age-related differences in 100-m sprint performance in male and female master runners. Med Sci Sports Exerc
21. Kubo K, Kanehisa H, Azuma K, et al. Muscle architectural characteristics in young and elderly men and women. Int JSports Med
22. Kukolj M, Ropret R, Ugarkovic D, Jaric S. Anthropometric, strength
, and power predictors of sprinting performance. JSports Med Phys Fitness
23. Kumagai K, Abe T, Brechue WF, Ryushi T, Takano S, Mizuno M. Sprint performance is related to muscle fascicle length in male 100-m sprinters. JAppl Physiol
24. Lauretani F, Russo CR, Bandinelli S, et al. Age-associated changes in skeletal muscles and their effect on mobility: an operational diagnosis of sarcopenia. JAppl Physiol
25. Mero A. Electromyographic activity, force and anaerobic energy production in sprint running with special reference to different constant speeds ranging from submaximal to supramaximal [PhD dissertation]. Jyväskylä (Finland): University of Jyväskylä; 1987.
26. Mero A, Komi PV. Force-, EMG-, and elasticity-velocity relationships at submaximal, maximal and supramaximal running speeds in sprinters. Eur JAppl Physiol
27. Mero A, Luhtanen P, Viitasalo JT, Komi PV. Relationships between the maximal running velocity, muscle fiber characteristics, force production and force relaxation in sprinters. Scand JSports Sci
28. Miyatani M, Kanehisa H, Ito M, Kawakami Y, Fukunaga T. The accuracy of volume estimates using ultrasound muscle thickness measurements in different muscle groups. Eur JAppl Physiol
29. Ojanen T, Rauhala T, Häkkinen K. Strength
and power profiles of the lower and upper extremities in master throwers at different ages. JStrength Cond Res
30. Pearson SJ, Young A, Macaluso A, et al. Muscle function in elite master weightlifters. Med Sci Sports Exerc
31. Roberts TJ. The integrated function of muscles and tendons during locomotion. Comp Biochem Physiol A Mol Integr Physiol
32. Sanada K, Kearns CF, Midorikawa T, Abe T. Prediction and validation of total and regional skeletal muscle mass by ultrasound in Japanese adults. Eur JAppl Physiol
33. Tanaka H, Seals DR. Endurance exercise performance in masters athletes: age-associated changes and underlying physiological mechanisms. JPhysiol
34. van Ingen Schenau GJ, de Koning JJ, de Groot G. Optimisation of sprinting performance in running, cycling and speed skating. Sports Med
35. Weyand PG, Davis JA. Running performance has a structural basis. JExp Biol
. 2005;208(Pt 14):2625-31.
36. Weyand PG, Sternlight DB, Bellizzi MJ, Wright S. Faster top running speeds are achieved with greater ground forces not more rapid leg movements. JAppl Physiol
37. Wilson GJ, Murphy AJ. The use of isometric tests of muscular function in athletic assessment. Sports Med
38. Wisloff U, Castagna C, Helgerud J, Jones R, Hoff J. Strong correlation of maximal squat strength
with sprint performance and vertical jump height in elite soccer players. Br JSports Med
39. Young W, McLean B, Ardagna J. Relationship between strength
qualities and sprinting performance. JSports Med Phys Fitness
Biomechanical Characteristics of Running by Age Groups
Muscle Structural Characteristics by Age Groups
Muscle Strength Characteristics by Age Groups