Fitness's moderation of the facilitative effect of acute exercise on cognitive flexibility in older...

Gender Differences in Anaerobic Power ofthe Arms and Legs—A Scaling Issue

CLARE L. WEBER1, MICHAEL CHIA2, and OMRI INBAR3

1School of Physiotherapy and Exercise Science, Griffith University Gold Coast, Queensland, AUSTRALIA; 2PE andSports Science, National Institute of Education, Nanyang Technological University, SINGAPORE; and 3Department ofLife Sciences, Zinman College Wingate Institute, Netanya, ISRAEL

ABSTRACT

WEBER, C. L., M. CHIA, and O. INBAR. Gender Differences in Anaerobic Power of the Arms and Legs—A Scaling Issue. Med. Sci.

Sports Exerc., Vol. 38, No. 1, pp. 000–000, 2006. Purpose: Physiological variables must be scaled for body size differences to permit

meaningful comparisons between groups. Using allometric scaling, this study compared the anaerobic performance, using both arms

and legs, of men and women. Ten active male and 10 active female subjects performed the leg cycling and arm cranking in a 30-s all-out

Wingate test (WAnT). Regional measurements of the legs, gluteal area, arms, and torso taken using dual-energy x-ray absorptiometry

(DXA) served as indicators of lower body active musculature (LBAMM) and upper body (UBAMM) active musculature. Results:

Body mass (BM) was the best predictor (i.e., r � 0.93–0.96) for peak power (PP) and mean power (MP) generated from sprint cycling

and arm cranking. Sex differences for leg and arm power (i.e., PP and MP) were identified in absolute terms and then expressed in

ratio to BM1.0. When the same data were allometrically scaled to BM and expressed as power function ratios (Power;BMb), the sex

differences in PP and MP for sprint cycling were nullified (female:male ratio � 100: 100–103%), but remained for arm cranking

(female:male power ratio � 100: 69–84%). Conclusions: These results confirmed that anaerobic power of the upper body and lower

body were best normalized to BM and, when statistically appropriate methods were used to take into account differences in BM, PP,

and MP generated from sprint cycling were similar for both men and women. In contrast, after allometric scaling for BM, men remained

more powerful than women for the supramaximal the arm cranking task. Qualitative differences in the upper body musculature between

men and women are speculated to account for the more powerful performance of men, but confirmatory evidence using noninvasive

techniques is warranted. Key Words: ALLOMETRIC SCALING, SEX DIFFERENCES, ANAEROBIC PERFORMANCE, LOWER

AND UPPER LIMB POWER

In recent decades, there has been increased involvement

of female participants in high-intensity exercise and

sport. This has sparked important research questions

about male and female differences in exercise performance.

Although considerable data are found on the maximal short-

term lower limb power output in various adult populations

(3,8,17,43), considerable less information exists about the

sex differences in both lower and upper limb maximal

short-term power output of the same participants in a single

study (17). In apparently the only reported study on upper

and lower limb muscle peak power (PP) and mean power

(MP) in Israeli adults 30–40 yr of age, Ben-Ari and Inbar

(5) reported that male adults had significantly greater PP and

MP for legs and arms than female adults, even when the data

were ratio-scaled to body mass (BM). Data on other adults

will be useful as many team sports such as basketball,

volleyball, hockey and rugby involve exercise exertions of

the upper and lower limbs.

Ample data show that male participants generate signif-

icantly greater lower limb muscle power than female par-

ticipants (17), but the data are more ambivalent when the

power is expressed relative to body size or body composi-

tion (22,43). Limited data show that upper and lower limb

power output are significantly greater in men than in women

in absolute and mass-relative terms (5), but data from other

adult populations have not been adequately elucidated.

Moreover, in Ben-Ari and Inbar (5) study on upper and

lower limb muscle power, only ratio scaling had been used

to express PP and MP in the Wingate Anaerobic Test

(WAnT), relative to BM. Chia (8), among others (40),

argued that in the normalization of high-intensity exercise

data, considerations such as the (a) choice of body size

descriptor (i.e., independent variable such as BM, muscle

mass) to express the exercise performance (i.e., dependent

variable, such as power); (b) the scaling method used for

data normalization (i.e., ratio-standard, linear adjustment, or

log-linear adjustment); and (c) how best to express the

scaled data (i.e., power function ratios or adjusted means)

are important and affect the interpretation of the results,

especially in sex and intergroup comparisons.

The Wingate Anaerobic Test has been the test of choice

for the assessment of short-term sprint cycling and arm

cranking (17) and its use for the assessment of power has

widespread appeal among the sport science community be-

cause of ease of administration and the robustness-reliability

and validity of the test (7). WAnT data are often described

Address for correspondence: Omri Inbar, ZAN – CardioRespiratory,

Schlimpfhofer Strasse 14, D-97723 Oberthulba, GERMANY; E-mail:

[email protected].

Submitted for publication November 2004.

Accepted for publication July 2005.

0195-9131/06/3801-0001/0

MEDICINE & SCIENCE IN SPORTS & EXERCISE®

Copyright © 2006 by the American College of Sports Medicine

DOI: 10.1249/01.mss.0000179902.31527.2c

1

balt5/zms-mss/zms-mss/zms00106/zms6300-06a satterfa S�7 11/22/05 6:52 Art: MSS200129 Input-nlm

relative to BM. Chia (8), however, suggested that, in seated

exercise performance (e.g., sprint cycling or arm cranking),

where BM is supported, it may be more suitable to relate the

performance (i.e., power) to some indicator of muscle size

(e.g., lower or upper limb muscle mass) that is involved

during the WAnT (6,17,22,23,26,36). Using anaerobic in-

dices derived from all-out intensity cycle ergometer tests,

several authors reported that significant sex differences in

sprint cycling remained even after normalization for body

dimensions, using ratio scaling (26,29,36) or linear-regres-

sion methods (29,43). In contrast, other studies have dem-

onstrated that sex differences in anaerobic performance are

eliminated when data are ratio-scaled to remove the influ-

ence of body size and body composition differences (11,22).

Ratio scaling use, the simple division of the performance

variable (power) by the body size descriptor (body mass), is

supposed to allow sex or intergroup comparisons in exercise

performance that are free from the effects of body size. In

ratio scaling, it is assumed that a linear and proportional

relationship exists between exercise performance (power)

and the body size descriptor (body mass), with the line of

identity passing through the origin (41). The use of ratio

scaling is endemic without apparent attention to whether the

conditions for its use are appropriate. For example, Nevill et

al. (27) reported that when power was expressed in ratio to

BM (W�kg�1 BM) and correlated with BM (kg), a signifi-

cant negative relationship was the outcome. Therefore, ratio

scaling did not produce an index that was free from the

effects of BM. Instead, ratio scaling caused the absolute

power data to be overcorrected for BM. This accorded

lighter participants with higher power values and penalized

heavier participants with lower power values. A growing

conviction among researchers suggests that ratio scaling

may not always appropriately normalize exercise data for

body size differences or produce a size-independent variable

(40).

Log-linear or power function adjustment method (i.e.,

allometric scaling) presents a better model fit of the perfor-

mance-size data and, thus, allows for a more appropriate

accounting for differences in body size (8,40). The general

form of the allometric or power function relationship is

given by Y � A·Xb, where Y is the dependent variable

(power), A is the proportionality coefficient, X is the body

size descriptor (body mass), and b is the exponent (41). The

derived b exponent empirically and exactly describes the

relationship between power (dependent variable) and body

mass (independent variable). This allows for the computa-

tion of power function ratios in the form of Y/Xb (a size-free

index). Allometric modeling is usually more successful in

accounting for size differences, such that when the power

function ratio (Y/Xb) is correlated with the body size de-

scriptor (X), the result should not be significantly different

from zero (27,41). Still, Welsman and Armstrong (41) cau-

tion that the use of allometric scaling is conditional and that

its use is not always justified. What this means is that in

applying allometric scaling to data sets, it is essential to

verify that the residual errors following log transformation

of power and mass are constant or show homoscedasticity

(41). This can be done by checking for a lack of a significant

correlation between the residuals and log body mass, and

that the errors are normally distributed (28). Although allo-

metric scaling has been applied to lower limb muscle power

in adults (27) and in young people (8), it has apparently not

been applied to both lower and upper limb muscle power of

men and women in the same study.

It is hypothesized that in applying allometric scaling to

lower and upper limb muscle power generated by men and

women, a clearer picture about the relationship between

power and various body-size descriptors would emerge.

This study, therefore, was conducted to compare the lower

and upper limb muscle power of men and women, expressed

in absolute terms, expressed in ratio, and allometrically in

relation to the various body size descriptors (e.g., body

mass, lean body mass, upper and lower body active muscle

mass).

METHODS

Subjects. Ten men and women volunteered to partic-

ipate in the present study. Volunteers were considered rec-

reationally active, but none were involved in any specific

exercise-training program. Following familiarization with

all testing equipment and experimental procedures, written

informed consent was obtained from each subject. The Grif-

fith University ethics committee for human experimentation

approved the test procedures used in this study. All women

had regular menstrual cycles and performed both the WAnT

for legs (WAnT-leg) and for arms (WAnT-arm), during the

follicular phase of their menstrual cycles. Previously dem-

onstrated was that the peak power output obtained during

sprint cycling may be lower during the luteal phase com-

pared with the follicular phase (31).

Experimental protocol. Each adult attended three

sessions of testing on separate days within a 2-wk period.

Volunteers were instructed not to engage in vigorous phys-

ical activity for 48 h before visiting the exercise science

research laboratory. During the first session, participants

were required to perform either the WAnT-leg or the

WAnT-arm, the order of which was randomized by the

flipping of a coin. At least 48 h later, participants were asked

to return to the laboratory to perform the WAnT not done in

the first session. Venous blood samples were obtained be-

fore both of the WAnT tests and 3 min after the end of each

test. These samples were subsequently analyzed for blood

lactate concentration ([La�]). During the third session, body

composition and the active muscle mass (AMM) for leg

cycling (lower body active musculature, LBAMM) and arm

cranking (upper body active musculature, UBAMM) were

determined using dual-energy x-ray absorptiometry (DXA).

Wingate Anaerobic Test for leg cycling and arm

cranking. Both the WAnT-leg and the WAnT-arm were

performed on a mechanically braked cycle ergometer

(Monark Ergomedic 824E, Varberg, Sweden) adjustable for

both leg and arm exercises. Participants warmed up for the

WAnT-leg (men � 75 W; women � 50 W) and for the

WAnT-arm (men � 30 W; women � unloaded) by exer-

2 Official Journal of the American College of Sports Medicine http://www.acsm-msse.org


cising at 60 rpm on a Monark ergometer for 5 min. During

the warm-up period, participants were asked to perform

three all-out 5-s sprints on the command of the chief inves-

tigator. Following the warm-up, participants rested quietly

for 10 min. Immediately before the WAnT, the participant’s

hyperemic earlobe was sterilized with 70% ethanol and

punctured with a 1.5-mm lancet (Microlancet, Becton Dick-

son, Sandy, UT). The first drop of blood was wiped away

and 50 �L of free-flowing blood was collected in a capillary

tube and immediately dispensed into a prechilled Eppendorf

tube for subsequent analysis of blood lactate concentration

([La�]) (model 2700 SELECT, Yellow Spring Instruments,

Yellow Springs, OH). Participants were then directed by the

chief investigator to begin unloaded pedaling 10 s before

beginning the test. Again, on the command of the chief

investigator, each participant was instructed to accelerate

maximally against no load. The predetermined resistance

was then applied after 3 s of maximal acceleration. The

resistance applied for men was 0.931 and 0.601 N·kg�1

compared with 0.833 and 0.470 N·kg�1 for women during

the WAnT-leg and the WAnT-arm, respectively. Partici-

pants were instructed to remain seated throughout the test

and were given strong verbal encouragement to maintain an

all-out effort throughout the test. The pedal revolution count

began the instant the resistance was applied. A microswitch

triggered by an infrared beam on the pedal cogwheel and

recorded into a computer processor signaled pedal revolu-

tions. Revolutions were recorded for each 5-s period during

the test. Peak power (PP) was the highest power produced in

a 5-s segment of the test, whereas means power (MP) was

calculated as the average power during the test duration. PP

and MP were expressed in W and relative to body mass

(W·kg�1), lean body mass (W·kg�1 LBM), and active mus-

cle mass (W·kg�1 AMM). The fatigue index (FI) was cal-

culated as the absolute difference between the highest and

the lowest power output expressed as a percent of the

highest power output (%FI). The duration of the WAnT-leg

and the WAnT-arm was 30 s, but subjects continued to

pedal against no load for 5 min after the test to prevent

venous pooling. A single blood sample was obtained for

subsequent lactate analysis 3 min after the WAnT. Heart

rate and rhythm were monitored closely throughout the test

by using the standard electrocardiography (Elma 3-channel

recorder and scope (Seattle, WA) in the CM-5 configura-

tion).

Determination of body composition and active

muscle mass. Body composition was assessed using

DXA (model XR36, Norland, Fort Atkinson, WI). Whole-

body values were presented as total mass (g) and separately

for fat mass (g) and percent of total body mass and lean

body mass (LBM) (g). Regional measurements (legs, gluteal

area, arms, torso) were determined on the basis of bony

landmarks via manual analysis. The total lean mass for both

legs and the gluteal muscle group was measured and re-

ported as the LBAMM. The gluteal muscle mass, which has

been shown to be one of the major muscle groups involved

in cycling (33), has been largely ignored when traditional

methods of determining the LBAMM were used (43). The

total lean mass for the arms and the shoulders, as well as the

muscle groups of the back and chest (torso), was measured

and reported as the UBAMM. It is difficult to determine and

control the muscle groups recruited during arm cranking at

supramaximal work rates. Nevertheless, it is reasonable to

suggest that the muscles of the back and chest contribute to

work output during arm cranking (9,21). AMM is reported

independently of fat mass and bone mineral content (BMC).

Body composition values obtained by using DXA were used

to express PP and MP obtained during the WAnT relative to

the LBAMM and UBAMM in each subject.

Statistical analysis. Results are presented as mean �

standard error (SE) of the mean. Independent sample t-tests

were used to determine sex-dependent differences in phys-

ical characteristics in all variables obtained during exercise.

Paired sample t-tests were used to determine if significant

differences existed between power output values across

exercise mode within each sex. For allometric scaling, PP

and MP derived from sprint cycling and arm cranking were

logarithmically transformed (natural logarithms; i.e., Ln PP

and Ln MP) as were body size descriptors, BM, LBM,

UBAMM, and LBAMM (i.e., Ln body size descriptor). The

natural log transformation of the indices of power (i.e., Ln

PP and Ln MP) and the body size descriptors (Ln BM, Ln

LBM, Ln UBAMM, Ln LBAMM) was done using the SPSS

program (version 11.5), using the “compute” function and

selecting the appropriate variables for transformation. The

relationship between the power and body size was described

in the form log Y � log A � blogX (41), where Y was the

dependent variable (PP or MP), A was the constant (its

antilog was the constant multiplier of the allometric equa-

tion), which described the magnitude of the Y variable

between groups that shared a common b exponent (i.e., the

slopes of the regression lines were not significantly different

between groups). X was the body size descriptor. To derive

the common b exponents, analyses of covariance were con-

ducted, with Ln PP and Ln MP entered in turn, as the

dependent variable and Ln body size descriptor (i.e., BM,

LBM, UBAMM, and LBAMM) entered separately, as the

covariate. Sex was entered as a fixed factor in all cases of

analyses. Subsequently, the respective body function ratios

(pfr) were computed by dividing the Y variable by body size

raised to the value of b identified (i.e., Y/Xb) (41). Sex

differences in WAnT power were compared using ratio

scaling and power function ratios. An alpha level of 0.05

was used for all statistical tests.

RESULTS

Anthropometric characteristics. The physical char-

acteristics of the adults are presented in Table 1. The mean

age was not significantly different in the two groups. Men

were taller than women, whereas women had lower mean

values than men for BM, LBM, and muscle mass for

LBAMM and for UBAMM. The muscle mass measured in

men was 33.8 and 49.4% higher than in women for the

lower limbs and for the upper limbs, respectively. LBAMM

constituted 37.9 and 35.4% of BM in men and women,

ANAEROBIC POWER OF MEN AND WOMEN Medicine & Science in Sports & Exercise�

3

T1


respectively, whereas UBAMM constituted 29.3 and 24.5%

in men and women, respectively.

Relationships between Wingate Anaerobic Test

power and body size descriptors. Figures 1 and 2

depict the relationships between PP and MP for the sprint

cycling and arm cranking and body size descriptors BM,

LBM, UBAMM, and LBAMM, respectively in men and

women (pooled data). Results showed that BM had the

highest correlations with the power generated both from

sprint cycling and arm cranking.

Sprint cycling and arm cranking in the Wingate

Anaerobic Test. The peak heart rate (HR) of 186 � 1

beats�min�1 measured in the men was not significantly

different from the peak HR of 184 � 2 beats�min�1 mea-

sured in the women during sprint cycling. Table 2 presents

the PP and MP obtained for sprint cycling, expressed in

absolute terms, ratio-scaled, and allometrically scaled to

BM, LBM, and LBAMM.

For sprint cycling, the men generated significantly greater

PP and MP than the women when the WAnT data were

expressed in absolute terms, in ratio to BM (i.e., PP:BM1.0

and MP:BM1.0) and allometrically in relation to LBAMM

(i.e., PP:LBAMM0.29 and MP:LBAMM0.33). However, PP

and MP expressed in ratio to LBM (i.e., PP:LBM1.0 and

PP:LBM1.0) and allometrically in relation to LBM (i.e.,

PP:LBM1.00 and MP:LBM1.39) were not significantly dif-

ferent between the men and women. Men also accumulated

significantly higher peak lactate concentration following the

WAnT. Men and women, however, demonstrated similar

fatigue indices during sprint cycling in the WAnT.

Table 3 presents power data for arm cranking in the

WAnT. Heart rate readings attained during arm cranking in

the WAnT were similar for both the men and women (174

� 2 vs 174 � 2 beats�min�1, P � 0.05).

Results showed that the men generated significantly

greater PP and MP for arm cranking than did women when

FIGURE 1—Relationships between Wingate Anaerobic Test sprint

cycling power and body size descriptors using pooled data of men and

women.

TABLE 1. Physical characteristics of men and women.

VariableMen

(N � 10)Women

(N � 10)

Age (yr) 23.3 � 1.2 23.2 � 2.0Stature (cm) 177.9 � 1.3 169.1 � 1.3*Body mass, BM (kg) 79.4 � 1.8 63.5 � 1.2*Lean body mass, LBM (kg) 59.4 � 1.4 43.2 � 1.4*Upper body active muscle mass,

UBAMM (kg)23.3 � 0.5 15.6 � 0.9*

Lower body active muscle mass,LBAMM (kg)

30.1 � 0.8 22.5 � 0.6*

Data are mean � SE. BM, body mass; LBM, lean body mass; LBAMM, lower body activemuscle mass. UBAMM, upper body active muscle mass; BL, blood lactate concentra-tion. Dual energy x-ray absorptiometrically determined LBM and UBAMM did notinclude bone mineral content. LBAMM included the muscle mass of both legs andgluteal muscle groups. UBAMM included arm, shoulder and torso muscle groups.* Significantly different at P � 0.05.

FIGURE 2—Relationships between Wingate Anaerobic Test arm

cranking power and body size descriptors using pooled data of men

and women.


F1-2

T2

T3

IQ: 1


the data were expressed in absolute terms, ratio-scaled, or

allometrically scaled to BM, LBM, and UBAMM. The only

exception was for MP expressed in ratio to UBAMM where

the women generated similar power output as did the men.

When the WAnT power generated during arm cranking

and sprint cycling was compared (in absolute values), PP for

the arms was 70% that of the legs, and MP of the arms was

59% that of the legs, in the men. The respective comparative

values for the women were 46% for PP and 52% for MP

(Tables 2 and 3).

The men had significantly greater fatigue index during

arm cranking than during sprint cycling, whereas the women

had similar fatigue indices during arm cranking and sprint

cycling during the WAnT (Tables 2 and 3).

The pre-WAnT BL concentration obtained before sprint

cycling and arm cranking was not significantly different

between the men and women. For both arm cranking and leg

cycling, post-WAnT BL concentration was significantly

higher in the men compared with the women (Tables 2 and

3). Further, the post-WAnT peak BL was 20 and 19% higher

following sprint cycling compared with arm cranking for the

males and females, respectively.

Common sex b exponents.

Results showed that mean common b exponents (i.e., Ln

body size*sex interaction, P � 0.05) could be used to

describe the exact allometric relationships between WAnT

power of the arms and legs, and the respective body size

indicators for men and women. In other words, a single

regression equation could be used to explain the relationship

between logarithmic-transformed power and body size. Ex-

cept for PP during sprint cycling and arm cranking and when

allometrically expressed in relation to LBM (common b

exponents 1.00 and 1.01, respectively), all other identified

mean common b exponents differed from the assumed b

exponent of 1.0 that is used in ratio scaling (Tables 2 and 3).

Common sex b exponents for WAnT-sprint cycling were

BM1.80 and BM2.53, LBM1.00 and LBM1.39, LBAMM0.29

and LBAMM0.33, for PP and MP, respectively. For WAnT-

arm cranking, the common sex b exponents were BM1.74

and BM1.45, LBM1.01 and LBM0.76, UBAMM0.54 and

UBAMM0.47, for PP and MP, respectively.

Sex differences in performance in arm cranking

and sprint cycling. Table 2 and 3 present performance

data in the WAnT for arm cranking and sprint cycling for

men and women, whereas Figures 3 and 4 show the per-

centages of power achieved by women in comparison to

those achieved by men in the WAnT, when the data were

expressed relative to BM, LBM, UBAMM, and LBAMM.

With reference to Figure 3, in absolute terms, the power

percentages generated by the women for the arms and legs

(PP and MP) were 45 and 60% and 68 and 69%, respec-

tively, of the power attained by the men. When the same

data were expressed in ratio to the various body size de-

scriptors, the power generated by the women were 58 and

75% and 85 and 84% (BM), 65 and 83% (LBM), and 70 and

92% and 92 and 90% (AMM) of the power attained by the

TABLE 3. Arm cranking performance indices of men and women.

Wingate Anaerobic Test: Arm Cranking

Variable Men Women

Peak power, PP (W) 743 � 37 336 � 18*Mean power, MP (W) 453 � 23 273 � 12*PP (W�kg�1 BM) 9.3 � 0.3 5.4 � 0.2*MP (W�kg�1 BM) 5.7 � 0.2 4.3 � 1.2*PP (W�kg�1 BM1.74) 0.36 � 0.01 0.25 � 0.01*MP (W�kg�1 BM1.45) 0.79 � 0.03 0.66 � 0.02*PP (W�kg�1 LBM) 12.4 � 0.4 8.0 � 0.3*MP (W�kg�1 LBM) 7.6 � 0.3 6.3 � 0.3*PP (W�kg�1 LBM1.01) 11.9 � 0.4 7.7 � 0.3*MP (W�kg�1 LBM0.76) 20.2 � 0.7 15.6 � 0.6*PP (W�kg�1 UBAMM) 31.9 � 1.5 22.4 � 1.0*MP (W�kg�1 UBAMM) 19.4 � 0.9 17.8 � 0.9PP (W�kg�1 UBAMM0.54) 135.6 � 6.4 78.4 � 2.6*MP (W�kg�1 UBAMM0.47) 103.1 � 4.9 75.3 � 2.7*Fatigue index (%) 63.3 � 1.4 47.3 � 1.8*Pretest BL (mmol�L�1) 1.0 � 0.0 0.9 � 0.1Posttest BL (mmol�L�1) 12.7 � 0.3 9.0 � 0.5*

Data are mean � SE. PP, peak power; MP, mean power; BM, body mass; LBM, leanbody mass; UBAMM, upper body active muscle; BL, blood lactate concentration. Dualenergy x-ray absorptiometrically determined LBM and UBAMM did not include bonemineral content. UBAMM included arm, shoulder, and torso muscle groups.* Significantly different at P � 0.05.

TABLE 2. Sprint cycling performance indices of men and women.

Wingate Anaerobic Test: Sprint Cycling

Variable Men Women

Peak power, PP (W) 1055 � 42 725 � 28*Mean power, MP (W) 766 � 23 518 � 31*PP (W�kg�1 BM) 13.3 � 0.3 11.4 � 0.3*MP (W�kg�1 BM) 9.7 � 0.2 8.1 � 0.3*PP (W�kg�1 BM1.80) 0.40 � 0.01 0.41 � 0.07MP (W�kg�1 BM2.53) 0.01 � 0.02 0.01 � 0.01PP (W�kg�1 LBM) 17.8 � 0.5 16.8 � 0.4MP (W�kg�1 LBM) 12.9 � 0.2 12.0 � 0.5PP (W�kg�1 LBM1.00) 17.7 � 0.5 16.8 � 0.4MP (W�kg�1 LBM1.39) 2.6 � 0.1 2.8 � 0.1PP (W�kg�1 LBAMM) 35.2 � 1.2 32.5 � 1.4MP (W�kg�1 LBAMM) 25.5 � 0.4 23.2 � 1.4PP (W�kg�1 LBAMM0.29) 393.1 � 14.1 294.4 � 10.9*MP (W�kg�1 LBAMM0.33) 248.9 � 5.6 185.7 � 10.9*Fatigue index (%) 51.8 � 1.5 54.3 � 2.6Pretest BL (mmol�L�1) 1.0 � 0.0 0.9 � 0.1Posttest BL (mmol�L�1) 15.3 � 0.3 10.7 � 0.4*

Data are mean � SE. PP, peak power; MP, mean power; BM, body mass; LBM, leanbody mass; LBAMM, lower body active muscle mass; BL, blood lactate concentration.Dual energy x-ray absorptiometrically determined LBM did not include bone mineralcontent. LBAMM included the muscle mass of both legs and gluteal muscle groups.* Significantly different at P � 0.05.

FIGURE 3—Peak power and mean power for the legs and arms

attained by women in the Wingate Anaerobic Test, expressed as a

percentage of male values.


5

F3-4


men. With reference to Figure 4, power function ratios

computed for WAnT data for arm cranking and sprint cy-

cling for the women, expressed relative to BM, LBM, or

AMM (PP and MP), were 69 and 84% and 103 and 100%,

65 and 77% and 95 and 108%, and 58 and 73% and 75 and

75%, respectively, of the power attained by their male

counterparts.

DISCUSSION

Although many published studies describe the lower limb

muscle power of men and women, results from the present

study are unique in that only a few studies have examined

the power output derived from arm cranking and sprint

cycling in men and women in a single study. Also, appar-

ently no published studies show where power output gen-

erated by the upper and lower body is expressed in relative

terms to various body size descriptors (i.e., BM, LBM,

UBAMM, and LBAMM), using ratio and allometric scaling.

These results are insightful in that they allow researchers to

make informed discussions (e.g., which should be the body

size descriptor of choice, and how the data should be nor-

malized for differences in body size) about group and sex

comparisons in power performances that are based on sam-

ple-specific evidence rather than merely informed conjec-

ture.

Participant characteristics. The anthropometric

data of the men and women in terms of amount of muscle

mass and sex differences in stature, BM, and LBM were

typical of equivalent data reported for untrained and active

samples (10,22,26,30,32).

Relationships between body size descriptors

and WAnT power. Results from the present study show

that, among the body size descriptors, BM had the highest

correlations with PP and MP derived from arm cranking or

sprint cycling. This was followed by LBM and then

UBAMM and LBAMM. These results are somewhat sur-

prising in that previously it had been reported that the best

predictor for sprint cycling was LBAMM (8), albeit, the

latter data were on boys and girls. Apparently no equivalent

published data are available on adults for direct comparison.

In the cited study, which involved 48 girls and 38 boys, aged

between 13 and 15 yr, the results of stepwise regression

analyses, with stature, BM, and LBAMM entered in as

covariates, LBAMM emerged as the strongest predictor for

both PP and MP derived from a 30-s WAnT. Most related

studies in the literature do not present empirical evidence

why a particular body size descriptor is chosen over another,

but often merely assumed that in sprint cycling. For in-

stance, measures of fat-free mass and thigh cross-sectional

area were representative of active musculature (3), without

any confirmation from empirical evidence that was based on

the specific sample of subjects being tested. BM, on the

other hand, is often excluded as a body size descriptor in the

presence of more elaborate and often more expensive mea-

surements of muscle mass (42) because it has been demon-

strated that large group differences in body fat percentages

can confound power function analyses (38). In the light of

the results of the present study, caution is advised in select-

ing the body size descriptor of choice; in essence, it should

not only be based on informed conjecture of the researcher,

it would be prudent also to confirm the conjecture by sta-

tistical evidence generated by the sample.

An explanation why BM was a better predictor of upper

and lower limb WAnT performances rather than UBAMM

and LBAMM in the present study could be that the contri-

bution of muscle groups not apparently involved during arm

cranking and sprint cycling toward power generation cannot

be discounted. Indeed, two recent studies (1,2) reported that

PP generated during sprint cycling in a 20-s WAnT (i.e., 886

� 14 vs 815 � 151 W, P � 0.05), as well as postexercise

BL concentration (9.14 � 0.38 vs 7.62 � 0.65 mmol�L�1,

P � 0.05) were significantly higher in a protocol that

allowed for the gripping of the handle bars than in another

protocol without the gripping of the handle bars. Their

results demonstrated that the arms and the upper body were

involved in stabilizing the entire body so that the lower

limbs could exert forces downward onto the cycle pedals to

generate the mechanical power. They concluded that the

performance of traditional style leg-cycle ergometry, in

which subjects remain seated, requires a muscular contri-

bution from the whole body. Further, it is obvious that the

arm cranking protocol used, allowed for some muscular

contributions of the legs. Such contributions have been

previously demonstrated in studies that documented signif-

icantly greater power output when healthy subjects were

compared with persons with spinal cord injury or with lower

limb amputations (14). Thus all-out intensity sprint cycling

and arm cranking should be considered whole body exer-

cises rather than exercises that are limited to the lower limbs

or upper limbs, respectively.

Ratio versus allometric scaling of WAnT power

expressed in relation to body size descriptors. An-

other important result of the present study was that allomet-

ric scaling of WAnT power for arm cranking and sprint

cycling was statistically more appropriate in accounting for

differences in body size than the traditional ratio scaling.

Indeed, the common mean (i.e., applicable to male and

female data) b exponents, specifically derived for BM (for

arm cranking and sprint cycling), LBM (for arm cranking

FIGURE 4—Peak power and mean power, allometrically scaled to leg

and arm performances of women in the Wingate Anaerobic Test,

expressed as a percentage of male values.



and sprint cycling), UBAMM (for arm cranking), and

LBAMM (for sprint cycling), which described precisely the

allometric relationships between power and body size de-

scriptors, were, except for a single case (i.e., PP�kg�1 LBM

for sprint cycling), markedly different from 1.0. These re-

sults are also in agreement with a number of other studies

(3,8). Chia (8) reported common b exponents of 0.65 and

0.79 for PP and MP allometrically expressed in relation to

LBAMM in boys and girls. In another study that examined

the anaerobic performance of 12 men and 12 women using

the WAnT, Batterham and Birch (3) reported a common b

exponent of 0.1 for PP, allometrically expressed in relation

to fat-free mass. Together, the results of the present study

showed that the b exponents identified for PP and MP

differed markedly from 1.0, the b exponent used in the

traditional ratio scaling.

Sex difference in power generated in the Wing-

ate Anaerobic Test—lower body musculature. In

concordance with findings in the extant literature, men were

significantly more powerful, in terms of PP and MP, in

absolute terms for sprint cycling. Indeed, Figure 3 shows

that women generated 69 and 68% of the PP and MP

achieved by their male counterparts for sprint cycling. When

the same data were allometrically scaled to BM (the best

body descriptor), however, PP and MP achieved in sprint

cycling were similar for both the men and women. These

results were buttressed by the observation that men and

women demonstrated similar FI for the WAnT-legs (Table

2). Therefore, in line with several previous studies (11,22),

but contrary to others (26,29,36,43), the results of the

present study suggest that in the lower body musculature no

physiological qualitative differences exists between men

and women when performing an anaerobic-type task.

Hence, the previously reported sex differences in power

generation in the lower musculature could seemingly be

attributed to either scaling procedure, choice of body de-

scriptor, or training status of the compared muscle group.

Upper body musculature. In line with the few pub-

lished data (mostly on children and adolescents)

(6,15,16,29,34) men were also significantly more powerful

in terms of PP and MP in absolute terms for arm cranking.

Figure 3 shows that female participants were able to gen-

erate only 45 and 60% of the absolute PP and MP achieved

by the male participants for arm cranking. For arm cranking,

however, when using the appropriate b exponent to scale for

body size differences (instead of the assumed b � 1.0 used

in ratio scaling), a significant sex-related difference re-

mained in both PP and MP, with women generating only 69

and 84% of body size–independent PP and MP to those

achieved by the men (Fig. 4). These results were also

affirmed by the greater arm FI showed by men compared

with the women (Table 3).

Only a small number of studies on upper body anaerobic

performance in untrained women are available for compar-

ison, however. Nonetheless, those few studies (5,17,29)

showed results in agreement with those of the present study

(i.e., higher upper body PP and MP in men), both in absolute

and relative terms to various body size descriptors. In one of

the few studies comparing men and women for both leg and

arm power generation, Nindl et al. (29) reported that when

ratio normalization and ANCOVA are used to remove the

influence of body size (i.e., BM, FFM, and CSA), the

adjusted means for leg and arm PP and MP were signifi-

cantly higher in boys than in girls aged 16–17 yr. The

authors alluded that qualitative differences could exist be-

tween male and female muscle for both upper body and

lower body musculature. In the cited study, the authors

employed anthropometric methods and sex-specific body fat

equations to estimate fat-free mass. In the present study, the

use of DXA yielded more precise measurements of total and

segmented musculature (10), compared with those reported

by Nindl et al., (29).

Although no other measurements were secured, some

reasons for the size-independent sex differences in the upper

body PP and MP seem plausible. For instance, in the present

study, the upper body musculature comprised 39 and 36% of

the fat-free mass in men and women, respectively. These

findings concur with previous reports (19,25,30), albeit us-

ing magnetic resonance imaging (MRI) or computed tomog-

raphy (CT) technologies. They reported that, in women, the

cross-sectional areas of the biceps brachi and elbow flexors

were significantly smaller (40–60%) than that of men.

Because our power data were allometrically scaled to BM

(as best predictor of power), it is conceivable that the higher

percentage of muscle within the upper body in men than in

women was advantageous.

Further, although muscle biopsy data suggest similar fast-

to-slow muscle fiber type distribution in the upper (deltoid

and biceps) and lower (vastus lateralis, rectus femoris, and

gastrocnemius) muscle groups, in untrained adults (35),

these data are not very convincing because they are based

mainly on male participants, involve a very small sample

sizes, and have individuals who are trained and untrained

mixed together. It is plausible that differences in muscle

fiber type distribution could exist in the men and women in

the present study. The appreciable gender difference in FI,

exclusively during arm cranking, and the reported tight

association between power decay (FI) during the WAnT and

muscle fiber distribution (18) may point to such a possibil-

ity.

Other qualitative factors that could contribute to upper

body power differences between men and women include

hormonal, enzymatic, and neurological factors (15,39). For

instance, Bell and Jacobs (4) and others (34) reported sex

differences in electromechanical delay and in the rate of

force development during an all-out cycling task. In the

present study, the women might have been more adversely

affected than the men in the rapid generation of power

(during the WAnT-arm), and also in overcoming the inertia

of the flywheel, when given that the muscle mass of the

upper body in the men was 49.4% higher than in the women.

Sex differences are reported also in the activity of the

rate-limiting enzymes in the glycolytic pathway, with men

exhibiting 15–29% greater phosphofructokinase (PFK) ac-

tivity than women (12,20,37). The cited studies, however,

have been concerned mostly with the vastus lateralis muscle


7


group. In the deltoid muscle, no sex differences are reported

for the concentration of phosphagen stores or in glycolytic

enzymes activity (13). Future research should use state-of-

the-art noninvasive technologies (e.g., MR spectroscopy) to

tease out sex differences that are attributed to metabolic

dissimilarities.

In conclusion, our findings confirm that it is important to

use statistically and physiologically defensible methods to

normalize WAnT power of the upper and lower body for

differences in body mass. The present findings show that, if

properly accounting for anthropometrical differences, the

lower limb anaerobic power output of men and women is

qualitatively similar. Our results, however, suggest that men

have some advantage over women in the production of

power generated by the upper body musculature. The exact

underlying causative factors for the sex difference presently

remains to be elucidated.

REFERENCES

1. BAKER, J., E. BROWN, G. HILL, G. PHILLIPS, R. WILLIAMS, and B.

DAVIES. Handgrip contribution to lactate production and leg power

during high-intensity exercise. Med. Sci. Sports Exerc. 34:1037–

1340, 2002.

2. BAKER, J. S., J. GAL, B. DAVIES, D. M. BAILEY, and R. M. MORGAN.

Power output of legs during high intensity cycle ergometry: in-

fluence of handgrip. J. Sci. Med. Sport. 4:10–18, 2001.

3. BATTERHAM, A. M., and K. M. BIRCH. Allometry of anaerobic

performance: a gender comparison. Can. J. Appl. Physiol. 21:45–

62, 1996.

4. BELL, D. G., and I. JACOBS. Electromechanical response times and

rate of force development in males and females. Med. Sci. Sports

Exerc. 18:31–36, 1986.

5. BEN-ARI, E., and O. Inbar. Leg and arm anaerobic capacity of 30

to 40-year-old men and women. In: Biomechanics of Sports and

Kinanthropometry, Book VI, F. Landry and W. A. R. Organ

(Eds.). 1978.

6. BLIMKIE, C. J. R., P. ROACHE, J. T. HAY, and O. BAR-OR. Anaerobic

power of arms in teenage boys and girls: relationship to lean

tissue. Eur. J. Appl. Physiol. 57:677–683, 1988.

7. CHIA, Y. H. M. Anaerobic Fitness of Young People. Doctoral

dissertation. University of Exeter, UK, 1998.

8. CHIA, Y. H. M. Wingate anaerobic test power of boys and girls

expressed in relation to lower limb muscle mass as determined

using dual energy x-ray absorptiometry. Adv. Exerc. Sports

Physiol. 9:55–59, 2003.

9. FRANKLIN, B. A. Exercise testing, training and arm ergometry.

Sports Med. 2:100–119, 1985.

10. FULLER, N. J., M. A. LASKEY, and M. ELIA. Assessment of the

composition of major body regions by dual-energy x-ray absorp-

tiometry (DXA) with special reference to limb muscle mass. Clin.

Physiol. 12:253–266, 1992.

11. GLEIM, G. W., C. SMALL, M. J. LIEDERBACH, et al. Anaerobic power

of professional ballet dancers. Med. Sci. Sports Exerc. 16:193–

194, 1984.

12. GREEN, H. J., I. G. FRASER, and D. A. RANNEY. Male and female

differences in enzyme activities of energy metabolism in vastus

lateralis muscle. J. Neurol. Sci. 65:323–331, 1984.

13. HARALAMBIE, G. Skeletal muscle enzyme activities in female sub-

jects of various ages. Bull. Eur. Physiopathol. Respir. 15:259–267,

1979.

14. HUTZLER, Y., S. OCHANA, R. BOLOTIN, and E. KALINA. Aerobic and

anaerobic arm-cranking power outputs of males with lower limb

impairments: relationship with sport participation intensity, age,

impairment and functional classification. Spinal Cord 36:205–

212, 1998.

15. INBAR, O. Development of anaerobic power and local muscular

endurance. In: The Encyclopedia of Sports Medicine: The Child

and The Adolescent Athlete, 1996, pp. 42–53.

16. INBAR, O., and O. BAR-OR. Relationship between anaerobic and

aerobic arm and leg capacities in swimming performance of 8–12

year old children. In: Frontiers of Activity and Child Health, R. J.

Shephard and R. Lavellee (Eds.). Quebec: Pelican, 1977, pp.

283–292.

17. INBAR, O., O. BAR-OR, and J. SKINNER. The Wingate Anaerobic

Test. Champaign, IL: Human Kinetics Publication, 1996, 1–110.

18. INBAR, O., P. KAISER, and P. TESCH. Relationships between leg

muscle fiber type distribution and leg exercise performance. Int.

J. Sports Med. 3:154–159, 1981.

19. JANNSEN, I., S. B. HEYMSFIELD, Z. WANG, and R. ROSS. Skeletal

muscle mass and distribution in 468 men and women aged 18–88

yr. J. Appl. Physiol. 89:81–88, 2000.

20. KOMI, P. V., and J. KARLSSON. Skeletal muscle fiber types, enzyme

activities and physical performance in young males and females.

Acta Physiol. Scand. 103:212–218, 1978.

21. KOPPO, K., J. BOUCKAERT, and A. M. JONES. Oxygen uptake kinetics

during high-intensity arm and leg exercise. Respir. Physiol. Neu-

robiol. 133:241–250, 2002.

22. MAUD, P. J., and B. B. SCHULTZ. Gender comparisons in anaerobic

power and anaerobic capacity test. Br. J. Sports Med. 20:51–54,

1986.

23. MAUD, P. J., and B. B. SCHULTZ. Norms for the Wingate anaerobic

test with comparison to another similar test. Res. Q. Exerc. Sport

60:144–151, 1989.

24. MERCIER, B., J. MERCIER, P. GRANIER, D. LE GALLAIS, and C.

PREFAUT. Maximal anaerobic power: relationship to anthropomet-

ric characteristics during growth. Int. J. Sports Med. 13:21–26,

1992.

25. MILLER, A. E., J. D. MACDOUGAL, M. A. TARNOPOLSKI, and D. A.

SALE. Gender differences in strength and muscle characteristics.

Eur. J. Appl. Physiol. 66:254–262, 1993.

26. MURPHY, M. M., J. F. PATTON, and F. A. FREDRICK. Comparative

anaerobic power of men and women. Aviat. Space Environ. Med.

51:636–641, 1986.

27. NEVILL, A. M., R. RAMSBOTTOM, and C. WILLIAMS. Scaling phys-

iological measurements for individuals of different sizes. Eur.

J. Appl. Physiol. 65:110–117, 1992.

28. NEVILL, A. M., and R. L. HOLDER. Scaling, normalizing and per

ratio-standards: an allometric modeling approach. J. Appl. Physiol.

79:1027–1031, 1995.

29. NINDL, B. C., M. T. MAHAR, E. A. HARMAN, and J. F. PATTON.

Lower and upper body anaerobic performance in male and female

adolescent athletes. Med. Sci. Sports Exerc. 27:235–241, 1995.

30. NINDL, B. C., C. R. SCOVILLE, K. M. SHEEHAN, C. D. LEONE, and R.

P. MELLO. Gender differences in regional body composition and

somatotrophic influences of IGF-I and leptin. J. Appl. Physiol.

92:1611–1618, 2002.

31. PARISH, H. C., and P. M.JAKEMAN. The effects of menstruation

upon repeated maximal sprint performance. J. Sports Sci. 1:78,

1987.

32. PATTON, J. F., and A. DUGGAN. Upper and lower body anaerobic

power: comparison between biathletes and control subjects. Int.

J. Sports Med. 8:94–98, 1987.

33. RAASCH, C. C., F. E. ZAJAC, B. MA, and W. S. LEVINE. Muscle

coordination of maximum- speed pedaling. J. Biomech. 30:595–

602, 1997.

34. SAAVEDRA, C., C. P. LAGASSE, C. BOUCHARD, and J. A. SIMONEAU.

Maximal anaerobic performance of the knee extensor muscles

during growth. Med. Sci. Sports Exerc. 23:1083–1089, 1991.

35. SALTIN, B., J. HENRIKSSON, E. NYGAARD, P. ANDERSEN, and E.

JANSON. Fiber types and metabolic potentials of skeletal muscles in

sedentary men and endurance runners. Ann. N.Y. Acad. Sci. 301:

3–39, 1977.

36. SERRESSE, O., P. F. AMA, J. A. SIMONEAU, G. LORTIE, C. BOUCHARD,


AQ: 2

AQ: 3

AQ: 1


and M. R. BOULAY. Anaerobic performances of sedentary andtrained subjects. Can. J. Sport Sci. 14:46–52, 1989.

37. SIMONEAU, J. A., and C. BOUCHARD. Human variation in skeletalmuscle fiber type proportion and enzyme activities. Am.

J. Physiol. 257:E567–E572, 1989.38. VANDERBURG, P. M., M. T. MAHAR, and C. H. CHOU. Allometric

scaling of grip strength by: C. H. Welsman, Jr, and N. Armstrong.Scaling performance for differences in body size. In: Paediatric

Exercise and Sport Medicine, N. Armstrong and W. V. Mechlen(Eds.). Oxford: Oxford University Press, 2000, pp. 7–23.

39. VAN PRAAGH, E., and E. DORE. Short-term muscle power duringgrowth and maturation. Sports Med. 32:701–728, 2002.

40. WELSMAN, J. R., and N. ARMSTRONG. Longitudinal changes insubmaximal oxygen uptake in 11- to 13-year-olds. J. Sports Sci.

18:183–189, 2000.41. WELSMAN, J. R., AND N. ARMSTRONG. SCALING PERFORMANCE FOR

DIFFERENCES IN BODY SIZE. IN: PAEDIATRIC EXERCISE AND SPORT MED-

ICINE, N. ARMSTRONG AND W. V. MECHLEN (EDS.). OXFORD: OXFORD

UNIVERSITY PRESS, 2000, PP. 3–9.42. WINTER, E. M. Cross-sectional area of muscle. Med. Sci. Sports

Exerc. 28:934–935, 1996.43. WINTER, E. M., F. B. C. BROOKES, and E. J. HAMLEY. Maximal

exercise performance and lean leg volume in men and women. J.

Sports Sci. 9:3–13, 1991.


9

AQ: 4

balt5/zms-mss/zms-mss/zms00106/zms6300-06a satterfa S�7 11/22/05 6:52 Art: MSS200129 Input-sda(v)

JOBNAME: AUTHOR QUERIES PAGE: 1 SESS: 5 OUTPUT: Fri Nov 18 07:12:55 2005/balt5/zms�mss/zms�mss/zms00106/zms6300�06a

IQ1: INTERNAL— Ptr: Please key MSS20019 Tables 1–4; hard copy sent by courier (also in

Documentum).

AQ1: AUTHOR— Please cite reference 24 in text.

AQ2: AUTHOR— Please provide publisher, location, and page range for reference 5.

AQ3: AUTHOR— Please provide page range for reference 15.

AQ4: AUTHOR— Two chapter titles appear to have been provided in reference 38—please check

for accuracy.

AUTHOR QUERIES

AUTHOR PLEASE ANSWER ALL QUERIES 1

Fitness's moderation of the facilitative effect of acute exercise on cognitive flexibility in older...

Documents

Transcript of Fitness's moderation of the facilitative effect of acute exercise on cognitive flexibility in older...