European Journal of Applied Physiology

, Volume 102, Issue 6, pp 685–694

Role of muscle mass on sprint performance: gender differences?

Authors

  • Jorge Perez-Gomez
    • Department of Physical EducationUniversity of Las Palmas de Gran Canaria
  • German Vicente Rodriguez
    • Department of Physical EducationUniversity of Las Palmas de Gran Canaria
  • Ignacio Ara
    • Department of Physical EducationUniversity of Las Palmas de Gran Canaria
  • Hugo Olmedillas
    • Department of Physical EducationUniversity of Las Palmas de Gran Canaria
  • Javier Chavarren
    • Department of Physical EducationUniversity of Las Palmas de Gran Canaria
  • Juan Jose González-Henriquez
    • Department of MathematicsUniversity of Las Palmas de Gran Canaria
  • Cecilia Dorado
    • Department of Physical EducationUniversity of Las Palmas de Gran Canaria
    • Department of Physical EducationUniversity of Las Palmas de Gran Canaria
    • Departamento de Educación FísicaCampus Universitario de Tafira
Original Article

DOI: 10.1007/s00421-007-0648-8

Cite this article as:
Perez-Gomez, J., Rodriguez, G.V., Ara, I. et al. Eur J Appl Physiol (2008) 102: 685. doi:10.1007/s00421-007-0648-8
  • 1.8k Views

Abstract

The aim of this study was to determine if gender differences in muscle mass explain the gender differences in running and cycling sprint performance. Body composition (dual-energy X-ray absorptiometry), and running (30 and 300 m test) and cycling (Wingate test) sprint performance were assessed in 123 men and 32 women. Peak power (PP) output in the Wingate test expressed per kg of lower extremities lean mass (LM) was similar in males and females (50.4 ± 5.6 and 50.5 ± 6.2 W kg−1, P = 0.88). No gender differences were observed in the slope of the linear relation between LM and PP or mean power output (MP). However, when MP was expressed per kg of LM, the males attained a 22% higher value (26.6 ± 3.4 and 21.9 ± 3.2 W kg−1, P < 0.001). The 30 and 300-m running time divided by the relative lean mass of the lower extremities (RLM = LM × 100/body mass) was significantly lower in males than in females. Although, the slope of the linear relationship between RLM and 300-m running time was not significantly different between genders, the males achieved better performance in the 300-m test than the females. The main factor accounting for gender differences in peak and mean power output during cycling is the muscle mass of the lower extremities. Although, the peak power generating capability of the muscle is similar in males and females, muscle mass only partially explains the gender difference in running sprints, even when expressed as a percentage of the whole body mass.

Keywords

Anaerobic capacityCycle ergometryShort sprintGenderExercise

Introduction

Sprint performance depends on the capacity to generate power and to achieve a high ratio between body mass and power (Mero et al. 1981; van Ingen Schenau et al.1994; Weyand et al. 2000; Chelly and Denis 2001; Cronin and Sleivert 2005). Weyand et al. (2000) showed that the main biomechanical variable determining differences in sprint running performance is the mean force applied during the ground contact phase of each step divided by the weight of the whole body. Chelly et al. (2001) reported in 11 handball players that acceleration significantly correlated with forward power only when expressed per unit of body mass (W kg−1: specific power), whereas the maximal running velocity only correlated with the total forward power of the body (W: absolute power). These authors reported that the leg muscle volume (estimated by anthropometry) was correlated with maximal running velocity in 11 teenage handball players (Chelly and Denis 2001). However, they did not perform a direct assessment of muscle mass. Since muscle mass is a major determinant of maximal force (Madsen et al. 1997; Ford et al. 2000; van Langendonck et al. 2004), and power results from the product of force and velocity, we hypothesised that sprint running performance depends, among other factors, on the muscle mass of the lower extremities in male and females adjusted for the whole mass (body weight) of the sprinter.

There are two major differences between sprint running and sprint tests on a static cycle ergometer. First, during sprint running, work has to be done to transport the own body mass; this fact may be taken into account by adjusting the power developed during static cycling by the body mass. Second, during running, the stance phase may be decomposed into an eccentric and a concentric phase (Luhtanen and Komi 1980). Cavagna et al. (1971) suggested that a sprint runner uses the work absorbed in his leg muscles (negative work, eccentric phase) at high speed to release further positive work (concentric phase) and thus increase power output. Thus, muscle mass may play different roles during sprint cycling and sprint running, and this effect may show gender dymorphism, due to the greater capacity of females to store and utilise elastic energy during the stretch-shortening cycle (Komi and Bosco 1978).

Therefore, the main aims of this study were: (1) to determine if the muscle mass of the lower extremities influences sprint performance in events with high and low recruitment of the stretch-shortening cycle, such as running and cycling; (2) to determine if muscle mass impacts sprint performance in males and females differently; and (3) to assess if sprint cycling performance has a similar predictive value for sprint running performance in males and females.

Methods

Subjects

One hundred and twenty-three men physical education students (age 23.6 ± 2.8 years, height 176.1 ± 6.3 cm, body mass 74.1 ± 8.6 kg; mean ± SD), and 32 women physical education students (age 23.3 ± 2.6 years, height 165.1 ± 6.4 cm, body mass 60.3 ± 5.7 kg; mean ± SD) participated in the study (Table 1). The study was performed in accordance with the Helsinki Declaration of 1975 as regards the conduct of clinical research, being approved by the Ethical Committee of the University of Las Palmas de Gran Canaria. Subjects provided their written consent before participating in the study.
Table 1

Subject’s physical characteristics and test results

Variables (mean ± SD)

Men

Women

Age (years)

23.7 ± 2.8

23.3 ± 2.6

Body mass (kg)

74.1 ± 8.6

60.3 ± 5.7*

Height (cm)

176.1 ± 6.3

165.1 ± 6.4*

Percentage of body fat (%)

15.4 ± 5.3

26.6 ± 6.2*

Lower limb muscle mass (kg)

19.5 ± 2.2

12.9 ± 1.2*

Lean total mass (kg)

58.5 ± 5.7

41.0 ± 3.3*

Peak power (W)

981.2 ± 145.0

652.8 ± 109.7*

Mean power (W)

701.0 ± 85.4

465.0 ± 58.2*

30-m time (s)

4.4 ± 0.2

5.0 ± 0.2*

300-m time (s)

46.5 ± 3.0

57.8 ± 3.8*

VO2max (ml kg−1 min−1)

45.6 ± 6.4

35.1 ± 7.0*

VO2max (ml (kg lower limb muscle mass)−1 min−1)

173 ± 25

163 ± 31 (P = 0.10)

P < 0.05 men versus women

General overview

All subjects were explored with a dual X-ray absorptiometer to determine their body composition and performed an incremental exercise to exhaustion (shuttle running test), a 30 and a 300-m running test, and a Wingate test. The running tests and the Wingate test were carried out on different days.

Lower limbs (LM) and total lean mass (TM)

Total lean mass and lean mass of the lower limbs [lower limb mass − (lower limb fat mass + lower limb bone mass)] was assessed by dual-energy X-ray absorptiometry (DXA) (QDR-1500, Hologic Corp., Software version 7.10, Waltham, MA) as reported in Calbet et al. (2001) and Ara et al. (2006). DXA equipment was calibrated using a lumbar spine phantom and following the Hologic guidelines. Subjects were scanned in supine position and the scans were performed in high resolution. Lower limb lean mass (kg) was calculated from the regional analysis of the whole body scan and it has been considered equivalent to the lower limb muscle mass. In addition, the relative lean mass of the lower extremities was calculated as [(lower extremities lean mass) × 100]/(whole body mass).

Running sprint tests

Subjects performed three maximal indoor short sprint trials, each separated by at least 5 min. The time required to cover 30 m was recorded indoors with photoelectric cells (General ASDE, Valencia). The timer is automatically activated when the subject crossed the first cell, and every 5-m thereafter. The subjects were encouraged to run as fast as they could. A standing start was used and the best of the three trials was selected as the representative value of this test (Vicente-Rodriguez et al. 2004). On a separate day, an all-out 300-m running test was carried out on a 400 m track outdoors; the time was recorded manually with a digital stopwatch.

All-out 30-s sprint test

Stop start Wingate tests were performed on a modified mechanically braked ergometer (Monark 818E, Monark AB, Vargerg, Sweden) equipped with a SRM power meter (Schoberer, Germany) with a braking load equivalent to 10 and 8% of body mass for men and women, respectively (Calbet et al. 2003). This test was preceded by at least two familiarisation Wingate tests in the preceding days. During the Wingate tests double-toe stirrups and straps were used to tightly fix the feet to the pedals. The subjects carried out a standardised warm-up consisting of 10 min of continuous cycling at an intensity close to 80 W followed by five maximal accelerations lasting 6 s every minute since minute 6. Then the participants rested for 5 min and performed an all-out 30-s effort with verbal encouragement. Peak power output (PP) was calculated as the highest work output performed during 1 s interval, and mean power output (MP) as the average work performed during the 30 s.

Aerobic maximal power

The maximal oxygen uptake (VO2max) was estimated using the maximal multistage 20-m shuttle run (Leger et al. 1988). This test predicts VO2max depending on the speed attained at exhaustion, which also depends on anaerobic capacity. Thus, estimated VO2max data were not used in the analysis of the factors determining sprint performance.

Statistical analysis

Means and standard deviations (SD) are given as descriptive statistics. The relationship between variables was tested using linear regression. To test the similarity of slopes and intercepts of these relationships, the corresponding t test was applied for the model: Yij = αi + βiXij + εij for i = 1, 2 (1 = men, 2 = women) and j = 1,..., n1, being εij independent and identically distributed random variables following a distribution N(0, σ1). Variables that correlated better with performance were included in a linear stepwise multiple regression analysis to determine which variables are more relevant to predict performance. Some of the comparisons between genders were carried out using ANOVA with gender as a factor with two levels. SPSS package (SPSS Inc., Chicago, IL) for personal computer was used for the statistical analysis. Statistical significance was set at P < 0.05.

Results

In men and women, the absolute lean mass of the lower extremities was linearly related to the peak and mean Wingate test power output (r = 0.66 – 0.77, P < 0.01) (Fig. 1). In women, the absolute lean mass of the lower extremities was linearly related to the 300-m running time (r = −0.53, P < 0.01) (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs00421-007-0648-8/MediaObjects/421_2007_648_Fig1_HTML.gif
Fig. 1

Relationship between lower extremities limbs mass and 30-m, 300-m, and Wingate performance. No significant differences between genders were observed in the slope of the linear relationship between the lean mass of the lower extremities and the peak or mean power output achieved in the Wingate test. However, the intercept of the linear relationship between the lean mass and mean power output was significantly lower in females than in males

The relative lower extremities lean mass (RLM) (RLM = Lean mass × 100/body mass) was linearly related to the 30-m and 300-m running time, and to the peak and MP output in men (r = −0.42, −0.38, 0.21, 0.23, respectively, all P < 0.01) and to the 300-m running time in women (r = −0.51, P < 0.01) (Fig. 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs00421-007-0648-8/MediaObjects/421_2007_648_Fig2_HTML.gif
Fig. 2

Relationship between the relative lean mass of the lower extremities (RLM = lean mass × 100/body mass) and 30-m, 300-m, and Wingate performance. The slope of the linear relation between RLM and 300-m running time was not significantly different between genders. However, the intercept was higher for the females

No significant differences between genders were observed in the slope of the linear relationship between the lean mass of the lower extremities and the peak or MP output achieved in the Wingate test. Lean mass-normalised PP output was also almost the same [50.4 ± 5.6 and 50.5 ± 6.2 W (kg of muscle mass)−1, in males and females, respectively, P = 0.88]. Lean mass-normalised MP output was 22% higher in males than females [26.6 ± 3.4 and 21.9 ± 3.2 W (kg of muscle mass)−1, respectively, P < 0.001].

The 30-m running time divided by RLM was significantly lower in males than in females [168.8 ± 16.9 and 233.7 ± 25.1 ms (% of muscle mass)−1, males and females, respectively, P < 0.001]. For a given percentage of RLM, the males achieved better performance in the 300-m test than the females [1,776.9 ± 198.8 and 2,719.5 ± 358.7 ms (% of muscle mass)−1, males and females, respectively, P < 0.001].

Bivariate correlations

A correlation matrix between running sprint performance and measured laboratory variables is shown in Table 2. When males and females were pooled together into a single group, the strongest relationship between the results in the Wingate test and body composition was found between MP output and whole body and lower extremities lean masses (r = 0.90, P < 0.01; Table 2).
Table 2

Bivariate relationships between Wingate test performance, body composition, and running sprint performance

 

300-m

30-m

%BF

LM

TM

Men

PP

−0.25*

−0.36*

−0.02

0.66*

0.67*

MP

−0.27*

−0.34*

−0.01

0.73*

0.74*

Women

  PP

−0.53*

−0.66*

0.03

0.66*

0.57*

  MP

−0.44*

−0.34

0.10

0.77*

0.68*

Men and women

  PP

−0.70*

−0.72*

−0.45*

0.84*

0.84*

  MP

−0.74*

−0.72*

−0.49*

0.90*

0.90*

PP peak power, MP mean power, BM body mass, %BF percentage of body fat, LM lower limb lean mass, TM total lean mass, 30-m time to cover 30-m running test, 300-m time to cover 300-m running test

P < 0.05

The 30-m time in men was significantly correlated with PP (r = −0.36, P < 0.05) and MP (r = −0.34, P < 0.05), while in women the correlation was observed only with PP (r = −0.66, P < 0.05) (Table 2). However, normalising power output in the Wingate test to whole body mass (W kg−1) improved these correlations to r = −0.55 in males, and r = −0.73 in females (both P < 0.05). Normalising PP by lower extremities lean mass produced slightly lower correlation coefficients (r = −0.39 and r = −0.70, in males and females, respectively, both P < 0.05). Likewise, normalising the MP output for whole body mass gave better correlation between MP and 30-m running time (r = −0.58 and r = −0.40, in males and females, respectively, both P < 0.05).

The 300-m running time was significantly correlated with PP (r = −0.25 and r = −0.53, in males and females, respectively, both P < 0.05). Normalising PP to whole body mass resulted in better correlations (r = −0.41 and r = −0.64, in males and females, respectively, both P < 0.05). However, the correlation between 300-m and PP were slightly lower when PP was normalised to the lean mass of the lower extremities (P = −0.24 and r = −0.30, in males and females, respectively, both P < 0.05).

The correlation between 300-m running time and MP was r = −0.27 and r = −0.44, in males and females, respectively (both P < 0.05). Normalising by body mass MP resulted in better correlations (r = −0.46 and r = −0.56, in males and females, respectively, both P < 0.05).

Multiple regression analysis: 300-m as the dependent variable

Linear stepwise multiple regression analysis to predict 300-m running time performance in men, women, and men and women combined into a single group, are depicted in Table 3. In men, the percentage of body fat (%BF), MP output in the Wingate test, and age are significant predictors of 300-m running time, explaining each 19, 8 and 4% of the variability in running time. When 30-m sprint performance was added to the regression analysis, it had the highest predictive value for 300-m time in men, explaining 23% of variance in 300-m time, while the %BF, MP and age contributed each to explain 7, 4 and 3% of the variance in 300-m time, respectively.
Table 3

Linear stepwise multiple regression analysis to predict 300-m running time performance in men, women, and men and women combined into a single group

 

300-m performance

R

R2

Men

  %BF

0.45

0.19

  MP

0.53

0.27

  Age

0.58

0.32

300-m time (ms) = 308.94 (age) − 10.60 (MP) + 233.95 (%BF) + 43,088.46

  30-m

0.48

0.23

  %BF

0.56

0.30

  Age

0.60

0.34

  MP

0.63

0.37

300-m time (ms) = −7.08 (MP) + 308.50 (age) + 160.21 (%BF) + 4,583.71 (30-m) + 21,475.87

 Women

  PP

0.58

0.32

  %BF

0.68

0.42

300-m time (ms) = 195.55 (%BF) − 18.47 (PP) + 64,595.71

  30

0.60

0.34

  TM

0.72

0.48

300-m time (ms) = −0.42 (TM) + 9,311.70 (30-m) + 28,394.59

 Men and women

  MP

0.73

0.53

  %BF

0.84

0.70

  Age

0.85

0.71

300-m time (ms) = 244.74 (age) + 345.46 (%BF) − 23.28 (MP) + 52,152.06

  30

0.81

0.66

  TM

0.86

0.74

  %BF

0.88

0.77

  Age

0.89

0.78

300-m time (ms) = 223.31 (age) + 186.21 (%BF) − 0.20 (TM) + 8,379.86 (30) + 12,967.24

Two different equations were calculated for each distance. For the first equation, the variables included in the multiple regression analysis were: peak power (PP), mean power (MP), lower extremities lean mass (LM) and total lean mass (TM), percentage of body fat (%BF), and age. In the second equation, the variables included in the multiple regression analysis were: peak power (PP), mean power (MP), lower extremities lean mass (LM) and total lean mass (TM), percentage of body fat (%BF), age, and the time to cover 30-m (ms) (when the predicted variable was 300-m time)

In women, the main contributor to the variance in 300-m time was the PP output achieved during the Wingate test which explained 32% of variance in running time, while the %BF explained 10% of the variance in running time. When the 30-m running time was considered in regression analysis, it was able to explain 34% of the variance in the 300-m time, while the whole body lean mass accounted for 14%.

When men and women were combined into a single population, the MP output in the Wingate test was a better predictor of 300-m running time, accounting for 53% of variance in 300-m time. The %BF and age explained 17 and 1%, respectively, of the variance in 300-m running time. When the 30-m running time was included into the regression analysis, this variable alone was able to explain 66% of the variance in 300-m time, while the whole body lean mass, %BF, and age accounted for 8, 3 and 1% of the variance, respectively, in 300-m running time. The PP output achieved during the Wingate test, the whole body lean mass, and the lean mass of the lower extremities explained 11, 6, and 3% of the variance in 300-m running time.

Multiple regression analysis: 30-m as the dependent variable

Linear stepwise multiple regression analysis to predict 30-m running time performance in men, women, and men and women combined into a single group, are depicted in Table 4. In men, the percentage of body fat was the best predictor of 30-m running time which accounted by itself for 13% of the variance in 30-m time. The PP achieved during the Wingate test, the whole body lean mass, and the lower extremities lean mass accounted for 11, 6 and 3% of the variance in 30-m running time.
Table 4

Linear stepwise multiple regression analysis to predict 30-m running time performance in men, women, and men and women combined into a single group

 

30-m performance 

R

R2

Men

  %BF

0.37

0.13

  PP

0.50

0.24

  TM

0.56

0.30

  LM

0.60

0.33

30-m time (ms) = − 1.08 (LM) + 0.02 (TM) − 0.36 (PP) + 13.54 (%BF) + 4,387.53

 Women

  PP

0.66

0.42

  TM

0.75

0.53

30-m time (ms) = 0.03 (TM) − 1.73 (PP) + 4,977.34

 Men and women

  MP

0.72

0.52

  %BF

0.79

0.61

  TM

0.80

0.63

  PP

0.81

0.64

30-m time (ms) = −0.57 (PP) + 0.01 (TM) + 16.97 (%BF) − 1. 22 (MP) + 4,921.11

Two different equations were calculated for each distance. For the first equation, the variables included in the multiple regression analysis were: peak power (PP), mean power (MP), lower extremities lean mass (LM) and total lean mass (TM), percentage of body fat (%BF), and age. In the second equation, the variables included in the multiple regression analysis were: peak power (PP), mean power (MP), lower extremities lean mass (LM) and total lean mass (TM), percentage of body fat (%BF), and age

In women, the PP achieved during the Wingate test was the best predictor of the 30-m running time, explaining 42% of the variance in 30-m running time, while whole body lean mass accounted for 9% of the variance in 30-m running time.

When men and women were pooled together into a single population, the MP output in the Wingate test was the best predictor of the 30-m running time, explaining 52% of the variance in 30-m running time. Additional variables with predictive value for the 30-m running time were %BF, whole body lean mass, and PP output achieved in the Wingate test, which accounted for 9, 2, and 1% of the variance in 30-m running time.

Discussion

The main aim of this study was to determine if the muscle mass of the lower extremities influences sprint performance in activities with high and low recruitment of the stretch-shortening cycle, such as running and cycling, and to assess if there are gender differences in the role played by muscle mass in sprint performance.

Gender differences in sprint performance have received limited attention (Cheuvront et al. 2005). It is known that males have a higher absolute power than females and also have a higher relative power than females when power is expressed normalised to whole body mass (Bar-Or 1987; Vandewalle et al.1987; Green 1995). Our study clearly shows a linear association between absolute lean mass of the lower extremities and peak power output during the Wingate test. Moreover, this relationship is similar in men and women to the extent that in both groups, the mean peak power output achieved during the Wingate was close to 50 W kg−1 of lean mass in the lower extremities. In addition, the regression analysis shows that peak power output increases linearly in both genders with the amount of lean mass present in the lower extremities (slopes and intercepts of the linear relationships were similar). The fact that power output normalised to lower extremities muscle mass is the same in males and females strongly suggest that the main reason for the gender difference in peak power output during the Wingate test is that females have a lower muscle mass.

Skeletal muscle single fibre studies have shown two main factors that determine fibre power: maximal shortening velocity and maximal specific tension (maximal isometric force divided by cross-sectional area) (Malisoux et al. 2006). Maximal shortening velocity depends on several factors among which the most relevant are the predominant isoform of myosin heavy chain, with a smaller influence of light chain isoforms (Bottinelli 2001). In turn, myosin heavy chain composition depends principally on the pattern of muscle recruitment (type of training) and genetic factors (Simoneau and Bouchard 1995). A genetic polymorphism of the gene codifying the alpha-actinin-3, a Z-disc structural protein found only in type II muscle fibres, has been associated with muscle-power performance in elite athletes (MacArthur and North 2004). About 18% of Caucasians are deficient in alpha-actinin-3 owing to homozygosity for a premature stop codon polymorphism, R577X, in the ACTN3 gene (MacArthur and North 2004). Homozygosity for this mutation is exceptional in elite sprinters (Niemi and Majamaa 2005; Lucia et al. 2007). Experiments with transgenic mice have shown that loss of alpha-actinin-3 results in lower LDH activity (indicative of reduced anaerobic capacity), without affecting muscle fibre type distribution (Macarthur et al. 2007).

Maximal isometric force is mainly dependent on cross-sectional area (Tesch and Karlsson 1978). In turn, the fibre’s cross-sectional area depends on many factors including genetic factors, hormones, growth factors, nutritional, and mechanical factors (Booth and Thomason 1991; Russell et al. 2000). No gender differences have been reported in fibre types in humans with different athletic status (Costill et al. 1976; Prince et al. 1977; Schantz et al. 1983; Sale et al. 1987; Alway et al. 1989). Several structural factors such as, muscle cross-sectional area, specific tension (force per cross-sectional area), tendon stiffness, penneation angle, fibre length, and fascicle length could explain gender differences in force generation capabilities. However, marked gender differences have been reported only for muscle cross-sectional area (Komi and Karlsson 1978; Schantz et al. 1983; Alway et al. 1989; Miller et al. 1993; Abe et al. 1998; Kumagai et al. 2000; Abe et al. 2001) and tendon stiffness (Blackburn et al. 2006), both being greater in males than in females.

Sex differences in passive stiffness have been reported for the knee flexors (Gajdosik et al. 1990; Blackburn et al. 2004), knee joint complex (Oatis 1993), ankle joint complex (Riemann et al. 2001), tendon structures (tendon and aponeurosis) in the medial gastrocnemius muscle (Kubo et al. 2003), and elbow flexors (Chleboun et al. 1997). Similarly, sex differences in active tendon stiffness have been described for the knee flexors (Granata et al. 2002b; Blackburn et al. 2004) and the lower extremity (Granata et al. 2002a).

However, the gender difference in tendon stiffness is not greater than the gender difference in muscle strength, and in each gender, tendon stiffness is proportional to the strength of the muscles acting on it (Bamman et al. 2000). Although, some gender differences in fibre length, and angle of pennation have been reported (Kubo et al. 2003), the muscle with greater force producing capabilities as a function of architecture would be expected to have a greater capacity to resist changes in musculotendinous length (i.e. display greater stiffness) when comparing two muscles of equal absolute size. Consequently, gender differences in tendon stiffness could hardly account for the gender differences in power performance, particularly during sprint cycling.

In theory, the more complying tendons of females should allow for a greater storage and release of energy during stretch-shortening activities such as sprint running, as reported during jumping (Komi and Bosco 1978). However, our results did not confirm or disprove this hypothesis. This study shows that while differences in absolute muscle mass could explain the totality of the gender differences in peak power output during sprint cycling, this is not the case during sprint running. The multiple regression analysis applied in the present investigation indicates that other factors could play a role, such as the percentage of body fat.

Weyand et al. (2000) showed that faster top running speeds are achieved with greater ground forces and not more rapid leg movements. It is likely that this also applies to the male–female comparison (Korhonen et al. 2003). Top male sprinters are about 7.3% faster than their female counterparts (Cheuvront et al. 2005), likely because male sprinters are able to generate higher ground reaction forces and, hence, longer strides (Korhonen et al. 2003). How this is brought about remains a mystery.

This study also showed that differences in muscle mass account for part of the sex differences in mean power output during the Wingate test and 300-m running time. No significant gender differences were observed in the slope of the linear relationship between the lean mass of the lower extremities and mean power output in the Wingate test. The latter indicates that mean power output improves with the increase in muscle mass almost in the same magnitude in males and females. However, the fact that the intercept of this relationship was lower in females and that mean power output in the Wingate test was 22% higher in males, suggests that other factors, in addition to muscle mass, must contribute to the gender differences in performance during prolonged sprints. This finding is slightly different than that reported by Weber et al. (2006). These authors, based on an allometric study on ten men and ten women, have reported that after accounting for anthropometrical differences, the lower limb anaerobic power output of men and women is qualitatively similar. The discrepancy between our study and that of Weber et al. may depend on the fact that Weber et al. lacked sufficient statistical power to appropriately compare, between genders, the slopes and intercepts of the relationship between muscle mass and power output. Weber et al. did not include this type of analysis in their study.

The capacity to synthesise the ATP required to sustain muscle contraction during a prolonged sprint is higher in males than females, likely due to the greater aerobic and anaerobic power of men. Males have a greater anaerobic capacity than females, particularly due to their higher glycolytic capacity (Komi and Karlsson 1978; Green et al. 1984; Simoneau and Bouchard 1989; Jaworowski et al. 2002). Type II fibres have higher glycolytic capacity than type I fibres (Essen-Gustavsson and Henriksson 1984), and although there are no sex differences in fibre type distribution, the area occupied by type II fibres is higher in men than women (Jaworowski et al. 2002). However, when the activities of LDH and PFK were adjusted by gender differences in cross-sectional area, the gender differences in LDH and PFK activities disappeared (Jaworowski et al. 2002). Thus, part of the gender difference in anaerobic capacity is due to the fact that men have a greater mass of type II fibres than females. This agrees with the findings obtained in the present investigation showing that muscle mass accounts for a great part of the gender difference in mean power output during the Wingate test. Nevertheless, our results also indicate that factors other than differences in absolute and relative muscle mass should also play a role. In fact, it has been reported that men also have greater anaerobic capacity than women even when expressed per kg of active muscle mass (Weyand et al. 1993; Weber and Schneider 2000; Mayhew et al. 2001).

During a sprint lasting for 30 s, 20–30% of the ATP consumed is produced through aerobic metabolism (Medbo and Tabata 1993; Calbet et al. 1997; Parolin et al. 1999; Calbet et al. 2003), a figure that reaches a value close to 50% of overall energy yield when the duration of the sprint is close to 60 s (Medbo and Tabata 1993). VO2max and the oxygen consumed during the Wingate test are highly correlated, meaning that subjects with higher VO2max are able to provide a greater amount of ATP during the Wingate test through oxidative metabolism (Calbet et al. 2003). Males have a higher VO2max (Lewis et al. 1986) and peak muscle oxidative capacity than females (Green et al. 1984; Borges and Essen-Gustavsson 1989). Since our males had higher estimated VO2max (per kg of body mass) than the females, it is probable that during the sprints, males were able to synthesise aerobically a greater amount of ATP (per kg of body mass) than females (Hill and Smith 1993). However, it remains unknown if males are able to provide a greater amount of ATP per kg of active muscle mass during sprint exercise.

Predictive value of Wingate power output: absolute versus relative power

Our results indicate that power output needs to be normalised to body mass for power output to improve its predictive value on sprint performance in either short (4–5 s) or long sprints (50–70 s). This finding is in agreement with previous studies (Meckel et al. 1995; Driss et al. 1998; Baker and Nance 1999). The studies having more heterogeneity in subjects performance such as that by Meckel et al. (1995), who studied a group of 30 subjects with varied sprint ability [mean 100 m time ranged from 11.1 s in the fastest group (n = 10) to 14.2 s in the slowest group (n = 10)] report the highest predictive value for sprint performance of power output normalised to body mass. In general, peak power output normalised to body mass has greater predictive value for running performance in short sprints than the absolute peak power output. Using the lean mass of the lower extremities as normalising variable does not improve the predictive value of peak power output beyond that obtained when the whole body mass is used as the normalising variable.

In agreement with our results, mean power output (W kg−1) in the Wingate test has been reported to be correlated (r = −0.88) with the running time in 100 m sprint (Meckel et al. 1995) in females. Likewise, in males, the mean power output developed during a 10-s sprint on the cycle ergometer has also been reported to correlate with running time in a 40-m sprint (r = −0.46) (Nesser et al. 1996). However, our study clearly shows that the peak power output is a more accurate predictor for short sprint performance than mean power output attained during the Wingate test.

In summary, this study shows that the main factor accounting for gender differences in peak power output during cycling is the muscle mass of the lower extremities. However, gender differences in prolonged sprint performance can be explained only partially by the sex differences in muscle mass and body fat, meaning that other factors, likely related with the capacity to sustain a high rate of ATP resynthesis, should account for the sexual dymorphism in prolonged sprint performance. Although, the lower gender difference in short sprint performance during cycling compared to running may be caused by the higher body fat mass of women, more studies are required to identify the mechanisms implicated. The predictive value of the power output achieved during a Wingate test for running sprint performance is higher when the power is expressed relative to body mass rather than in absolute values.

Acknowledgments

The authors wish to thank José Navarro de Tuero for his excellent technical assistance and to Fiona Wong for her wonderful editorial skills. This study was supported by grants from the Ministerio de Educación y Ciencia (BFI2003-09638, DEP2006-56076-C06-04/ACTI and FEDER) and the Gobierno de Canarias (PI2005/177). Special thanks are given to all subjects who volunteered for these experiments.

Copyright information

© Springer-Verlag 2007