See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/22014772

Muscular efficiency during steady-rate exercise: effects of speed and work

rate

Article  in  Journal of applied physiology · July 1975

DOI: 10.1152/jappl.1975.38.6.1132 · Source: PubMed

CITATIONS READS

497 3,779

2 authors:

Glenn A Gaesser George A Brooks

Arizona State University University of California, Berkeley
257 PUBLICATIONS   9,588 CITATIONS    357 PUBLICATIONS   23,190 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Thomas D. Fahey View project

Smart Walk View project

All content following this page was uploaded by George A Brooks on 10 March 2014.

The user has requested enhancement of the downloaded file.

JOURNALOFAPPLIEDPHYSIOLOGY
Vol. 38, No. 6, June 1975. Printed in U.S.A.

Muscular efficiency during steady-rate exercise:

effects of speed and work rate

GLENN A. GAESSER AND GEORGE A. BROOKS

Exercise Physiology Laboratory, Department of Physical Education, University of CaliJornia, Berkeley, CaliJornia 94720

GAESSER, GLENN A., AND GEORGE A. BROOKS. Muscular efi- Traditionally, caloric equivalents of substrate utilized
ciency during steady-rate exercise: effects of speed and work rate. J. Appl. have been determined from the 02 consumption (vo,)
Physiol. 38(6) : 1132-l 139. 1975.-In a comparison of traditional and respiratory exchange ratio (R). Recently it was pro-
and theoretical exercise efficiency calculations male subjects were posed that the work efficiency calculation best describes
studied during steady-rate cycle ergometer exercises of “0,” 200,
muscular efficiency (27). The rationale was based on a
400, 600, and 800 kgm/min while pedaling at 40, 60, 80, and 100
novel approach to efficiency calculations involving theo-
rpm. Gross (no base-line correction), net (resting metabolism as
base-line correction), work (unloaded cycling as base-line correc- retical-thermodynamic considerations (i.e., P/O = 3,
tion), and delta (measurable work rate as base-line correction) AG for ATP = CP = - 11 .O kcal/mol, and phosphoryla-
efficiencies were computed. The result that gross (range 7.5- tive-coupling efficiency = 60 70). However, this theoretical-
20.4%) and net (9.8-24.1 s) e ffi ciencies increased with increments thermodynamic approach cannot be used to validate any
in work rate was considered to be an artifact of calculation. ,4 one particular method of efhciency calculation because,
linear or slightly exponential relationship between caloric output irrespective of the approach, the estimation of energy ex-
and work rate dictates either constant or decreasing efficiency pended by the subject is still obtained from J?oz.
with increments in work. The delta efficiency (24.4-34.0%) Although it has been demonstrated that at constant-load
definition produced this result. Due to the difficulty in obtaining work muscular efficiency decreases as the speed of move-
0 work equivalents, the work efficiency definition proved difficult
ment increases (4, 10, 1 Z), data on the effect of work rate
to apply. All definitions yielded the result of decreasing efficiency
with increments in speed. Since the theoretical-thermodynamic on efficiency are inconsistent (2, 3, 4, 7, 10, 15, 30). Pur-
computation (assuming mitochondrial P/O = 3.0 and AG = poses of the present study were to 1) further investigate
- 11 .O kcal/mol for ATP) holds only for CHO, the traditional similarities and differences between traditional and theo-
mode of computation (based upon ~OZ and R) was judged to be retical-thermodynamic approaches to calculating em-
superior since R < 1 .O. Assuming a constant phosphorylative- ciency, -3) elucidate effects of speed and work rate on
coupling efficiency of 60 yO, the mechanical contraction-coupling muscular ef-ficiency, and 3) evaluate adequacy of the various
efficiency appears to vary between 41 and 57 yO. base-line corrections in describing relationships between
caloric expenditure, work rate and speed of movement.
exertion; respiration; indirect calorimetry; energetics; oxidative
phosphorylation METHODS

Subjects were 12 well-conditioned males between the

ages 19 and 24 yr. Each subject exercised on a Quinton
(QI-844) electrically braked cycle ergometer at 40, 60, 80,
MUSCULAR EFFICIENCY of an individual during steady-rate and 100 rpm. Exercise at each pedaling speed was per-
exercise is expressed as the ratio of work accomplished to forrned on separate days, 1 wk apart. Subjects were tested
energy expended. Calculations of efficiency during the the same time of day each week at least 2 h after their
steady rate are based on open-circuit indirect calorim- last meal. At each pedaling speed work rates of “0,” 200,
etry and the assumption that energy requirements are 400, 600, and 800 kgrn/min were performed. “0” work
met by respiration. It is obvious, however, that selection
of the base-line correction factor will change estirnates of work accomplished W
= -_____ x 100
energy expenditure and efficiency. This problem of base- energy expended above El - E,
work efficiency =
line selection resulted in the development of gross, net, that in cycling without
work, and delta definitions of eficiency? a load
~~~-~ delta work accomplished = AW
delta efficiency = z x 100
l Definitions of muscular efficiency delta energy expended
work accomplished W where W = caloric equivalent of external work performed; E =
gross efficiency = =- E x100
energy expended gross caloric output, including resting metabolism; e = resting
caloric output; El = caloric output, loaded cycling; E, = caloric
work accomplished W output, unloaded cycling, AW = caloric equivalent of increment in
=- x 100
net efficiency = energy expended above E - e work performed above previous work rate; AE = increment in caloric
that at rest output above that at previous work rate.
1132
CYCLE ERGOMETER STEADY-RATE WORK EFFICIENCY 1133

was defined as cycling with the ergometer switched off. TABLE 1. &fect of pedaling speed on
To achieve the steady rate (23, 26), the duration of exercise metabolic rate at ” 0” work
.at 0, 200, and 400 kgm/min was 6 min, while at 600 and --
,800 kgm/min each subject exercised for 8 min. Rest in- Metabolic Rate
wm 902 l/min* Caloric output, kcal/mint
tervals between exercise bou ts were of the sarne duration
as the exercise period prior to the rest interval. Resting 10 0.435 zt 0.012 2.03 zt 0.06
Vo2 was observed prior to exercise, as the subject sat quietly 60 0.448 & 0.019 2.13 zt 0.11
80 0.584 z!z 0.032 2.81 ziz 0.15
*on the cycle ergometer. 100 0.933 * 0.040 4.52 zfi: 0.21
Subjects inspired room air through a Daniel’s low-re-
sistance valve. During the last 2 min of each exercise period Values are expressed as means =t SE (N = 12). * y inter-
*expired air was collected in a 120-liter Tissot spirometer. cept of regression of Vo2 on work rate. t y intercept of re-
gression of caloric output on work rate.
Gas analysis was performed by pumping expired air from
the spirometer over color indicator CaS04 and through
Beckman E2 02 and Beckman LB-l CO 2 analyzers. These requisite to the concept of steady-rate work. Therefore,
analyzers were calibrated from tanks of known gases (es- the rate of ATP utilization is directly related to the Vo2.
tablished by Haldane analysis) before, during, and after By using the standard molar volume of 22.4 l/mol, liters
*each experiment. of 02 utilized can be converted to moles of 0 2 utilized.
Steady-rate To 2 and CO 2 production (%02) were Second, a theoretical average P/O ratio of 3.0 for the
calculated (8; 11, p. 300-309), and the respiratory exchange mitochondrial electron transport chain is assumed (19).
ratio (R) was used to estimate caloric output (11, p. 628). Third, a free energy change for the hydrolysis of ATP of
The data obtained on each subject were used to calculate - 11 .O kcal/mol is considered for conditions approximating
efficiency using all definitions (gross, net, work, and delta), physiological (28, 29). Th e above considerations allow for
and both traditional and theoretical-thermodynarmic ap- a calculation of the mechanical contraction-coupling em-
proaches. ciency, which represents the ability of muscle to utilize the
To calculate the efliciency using the theoretical-thermo- available chemical bond energy of ATP for muscle con-
<dynamic approach four assumptions must be made. First, traction. And fourth, assuming the other assumptions to
in steady-rate work the levels of ATP and CP are main- define a phosphorylative-coupling efhciency of 60 7% (18,
tained by the catabolism of foodstuffs. This assumption 19), then the total efficiency of the body in muscular work
*appears valid based on muscle biopsy studies (16) and is can be obtained.

RESULTS
12
r
2 The relationship between caloric output and work rate
II /- at each pedaling speed is presented in Fig. 1. Zero work
/
/ data obtained on our cycle ergometer clearly yielded an
IO / z unacceptable base line and could not be used to represent
,/”
the energy cost of moving the legs without a load. There-
9 A’ z ,”
/ fore, for work efhciency calculations, the y intercept of the
./ / /’ regression of V0 2 and caloric output on work rate served
8 ./ as the base-line correction factor (Table 1). Between the
J? /y’
c
a- *A’ work rates of 200 and 800 kgm/min the increase in caloric
.A ,A
\7E ./ output was essentially linear (Fig. 1). However, there was
-e / 4” 0 /
.-.-.-. a slight exponential rise in caloric output at the higher
z -4’ H 0/
6
,/ , / work rates, particularly between 600 and 800 kgm/min.
In addition, Fig. 1 demonstrates that at any given work
,/’ ,///
5 ,
00 rate the caloric output increases exponentially with pedaling
__----- -d 00 speed.
- 40 RPM
4 ----- Data for gross, net, and work efhciency calculations are
-c/ AAH 60 RPM
HH-
/-
--- - presented in Tables 2 and 3. Data for the effects of speed
80 RPM
3 / and work rate on delta eficiency calculations are presented
---e-e- 100 RPM
in Table 3 and Fig. 2, respectively. For gross and net ef-
2 ~~ ficiencies Table 2 contains values that can be used to de-
Kg-m/min
0 scribe trends for changes in eficiency as a function of both
0 200 400 630 800
speed and work rate. For work efficiency computations
FIG. 1. Effect of work rate and speed on caloric output (X -4
Table 2 contains data only on the effect of work rate on ef-
SE) of human subjects during steady-rate exercise. Caloric outputs ficiency. Since in the work and delta eficiency calculations
determined from VO, and R. Essentially linear relationship with the base-line correction is influenced by both speed and
perhaps a slightly exponential rise in caloric output at higher work work rate, a separate Table 3 is needed to demonstrate
rates dictates either a constant or decreasing efficiency. Since the
the effect of speed on work and delta efficiency calcula-
caloric cost of each work rate increases with increments in pedaling
speed, decreasing efficiency with increasing speed is indicated. Note tions. For work efficiency, data obtained for 0 work at 40
that y intercepts deviate significantly from linearity. rpm served as base line for all calculations. For delta efhci-
1134 G. A. GAESSER AND G. A. BROOKS

TABLE 2. effect of speed, work rate, and method TABLE 3. effect of speed on work and delta
of calculation on cycle ergometer ejiciency cyde ergometer ejiciency ca hula tions

40 rpm 200 kgm/min

Method of 200 kgm/min 400 kgm/min 800 kgm/min Method of calculation 40 rpm 60 rpm 80 rpm 100 rpm
calculation 600 kgm/min
I
I I I I
Theoretical
Work work I 24.8
25.4 &
rt 726.5
.727.5
I zfi
& 2.5
2.7 118.1
18.9 ~fr
zt .6
.5*111.311.6 zt
rt .6*
.5
Gross 19.3 =t .2 20.2 rt .4
Theoretical 19.2 z!z .2 20.1 rt .4
gross
400 kgm/min
Net 18.8 zk .7 22.9 zt .6 23.9 zt .4 23.8 zt .6
Theoretical 18.3 + .7* 22.6 -+I .5:’ 23.8 zk .3 23.8 rfI: .5 Work 26.9 rt .426.1 zt .8 21.7 -+: .3 16.4 zt .5
net Theoretical work 27.3 zt .426.7 zt .7 22.3 rt .4* 16.7 rt .4

Work 24.8 =t .7 26.9 zt .4 26.8 zfi: .4 25.7 zt .6 Delta 29.0 rt 2.1 19.7 * .9 15.6 rf 1.1
Theoretical 25.4 -+: .7 27.3 rk .4 27.1 sf .2 26.2 zt .5 Theoretical delta 29.5 rt 2.2 20.2 & 1.0 15.7 zt 1.0
work
600 kgm/min
60 rpm
Work 26.8 =t .426.7 rt .8 23.6 rt .3 19.3 zt .4
Theoretical work 27.1 I+ .227.2 zt .8*24.2 & .2* 19.7 I!Z .4*
Gross 12.1 It .3 16.6 zfi. .3 19.2 & .4 20.4 rt .4
Theoretical 11.8 & .3* 16.5 zt .2 19.1 * .4 20.3 sfi: .3
Delta 30.5 rt 5.1 20.4 rt .9 16.1 rt .8
gross
Theoretical delta 31.7 =t 5.3* 21.1 It 1.0* 16.4 =t .9
Net 18.9 zt .8 22.3 31 .5 23.8 zt .5 24.1 =t .5
Theoretical 18.6 rt .8*22.0 rt .4 23.7 zfi .5 24.1 zk .4 800 kgm/min
net
Work 25.7 -+: .626.2 s+ .6 23.6 I+ .4 20.3 31 .5
Work 26.4 zt .8 26.8 dz .6 127.1 zk .6 26.6 zt .7 Theoretical work 26.2 & .526.4 rt .5 24.1 rt .3* 20.7 =t .3*
Theoretical 26.7 zt .7 27.1 r+ .5 27.4 It .5 26.8 =t .6
work Delta 26.3 rt 2.3 18.2 rt .9 14.6 rf .7
Theoretical delta 26.7 zk 2.3 18.7 xi= 1.0 14.8 =t .7
80 rpm
Values are expressed as means & SE (N = 12). * Significant
Gross 110.2 zk 2 14.8 & .2 17.6 zt .2 18.8 If .3 at P < 0.05 for paired t-test comparisons of traditional and theo-
Theoretical 10.1 It 2 14.7 zk .2 17.6 zt .2 18.9 =t .2 retical calculations of efficiency.
gross
The results of paired t-test analyses (Tables 2, 3, and
Net 14.7 zt .4 19.0 =t .2 21.4 =t .4 21.9 * .4
Theoretical 14.6 zt .4 19.0 rt 2
Fig. 2) demonstrated that in 55 of 81 comparisons there
net 21.5 It -3 !22*o * l3 were no statistically significant differences between tra-
ditional and theoretical-thermodynamic approaches to
Work 26.2 zt 1.1 26.8 =t .8 27.4 zt .9 26.4 It .9 calculating efficiency. In all instances where significant
Theoretical 126.6 xt 1.0” 27.1 =t .8 27.9 =t .8* 26.9 zt .9
differences occurred, the differences were small (0.2-l .2 %).
work
However, the theoretical-thermodynamic approach con-
100 rpm
sistently gave lower values than the traditional approach
in the gross and net efficiencies, whereas the opposite was
Gross 7.6 zt .3 12.1 =t .3 15.1 zt .2 16.6 * .3 true for work and delta efficiencies (in which the theoretical-
Theoretical 7.5 & .2 12.1 rt .2 15.1 zt .2 16.7 zt .2 thermodynamic approach gave greater values than the
gross traditional approach to calculating efficiency).
9.8 zk .5 14.9 =t .4
In Fig. 3 gross, net, and work efficiencies are plotted as
Net 17.8 & .4 119.0 zt .4
Theoretical 9.8 zk .5 15.1 =t .5 17.9 r+ .3 19.2 =t .3 a function of work rate at 60 rpm. Since the essentially
net linear with constant slope relationship between caloric
output and work rate holds at each pedaling speed (Fig. l),
Work 26.7 r+ .7 29.5 =t .9 27.8 zt .7 with differences existing only in magnitude of y intercepts,
Theoretical 27.5 zt .7” 30.3 xt 1.1*28.5 =t .8*
work just one curve for each method of calculation (using both
approaches) is presented to illustrate trend. With gross and
Values are expressed as means zfi SE (lv = 12). * Significant
at P < 0.05 for paired t test comparisons of traditional and theo- work efficiency at 600 kgm/min, 80 rpm =
retical calculations of efficiency. caloric equivalent of 600 kgm/min
. x 100 = 23.6y0
caloric output at 600 kgm/min, 80 rpm -
caloric output at 0 kgm/min, 40 rpm
ency, changes in speed with each change in work rate
were used to obtain base-line corrections.2 delta efficiency at 600 kgm/min, 80 rpm =
caloric equivalent of 200 kgm/min
x 100 = 20.4y0
2 Effect of speed on work and delta. efficiency calculations (sample caloric output at 600 kgm/min, 80 rpm -
calculations) caloric output at 400 kgm/min, 60 rpm
CYCLE ERGOMETER STEADY-RATE WORK EFFICIENCY 1135

35

40 RPM 60 RPM

^o
o\ 30
-

z
z
W
3

E 25
W

A-------A Theoretical Delta

C+---0 Delta o-----o De.ta

Kg-m/min Kg; m,/min

200 400 600 800 200 400 600 800

40 40

80 RPM - 100 RPM

-
35

*.,

’ ‘. ‘1
-
-.
.
.
.
-
-

-- --
& A Theoretical Delta
Theoretical Delta
(y----c) Delta

20 20 ~
/j
. Kg- m/min , Kg- m /min
I I I
200 400 600 800 200 400 600 800
FIG. 2. Effect of work rate on delta efficiency calculations (x rf= increments in work rate. Result at 60 rpm is due to the fact that the
SE) for human subjects during steady-rate exercise at 40, 60, 80, and most linear relationship between caloric output and work rate was
100 rpm. In contrast to the other definitions (Fig. 3), delta effi- observed at this speed (see Fig. 1).
ciency demonstrates a definite trend for decreasing efficiency with

net efficiency curves it appears that efficiency is increasing In considering changes in efhciency as a function of
with increments in work rate (from 12.1 and 18.9 %, respec- increments in pedaling speed, all methods of calculation
tively, to 20.4 and 24.1 % respectively). The work efficiency, demonstrate a decreasing efliciency as speed of movement
however, remains relatively constant (between 26.4 and increases (Fig. 4).
27.1 %). At the other pedaling speeds, however, work
efficiency tends to increase with increments in work rate, DISCUSSION
particularly between 200 and 600 kgm/min, with a slight
decrease in efficiency between 600 and 800 kgm/min Data on R (Table 4) indicate the metabolic rates of these
(Table 2). In contrast to these methods, delta efficiency subjects to have been well below their anaerobic threshold
calculations demonstrate decreased efficiency as metabolic (23). Therefore, a requirement for the validity of indirect
load increases (Fig. 2). This is especially apparent at 40, calorimetry as being representative of the metabolic rate
80, and 100 rpm. At 60 rpm the delta efficiency demon- was met. The pedaling speeds used (40-100 rpm) were
strates relatively little change in efficiency. This is in agree- those dictated by technical limitations of the Quinton
ment with the observation that the most linear relation- ergometer and encompassed those utilized by bicycle
ship between caloric output and work rate was at this racers and previous investigators experimenting with the
speed (Fig. 1). cycle ergometer.
1136 G. A. GAESSER AND G. A. BROOKS

In discussing the efliciency of the human body during

cycle ergometer exercise it is obvious that any efficiency
calculations, and resulting plots, must be able to accurately
describe the relationships between pedaling speed, work
rate, and the caloric output. Since the relationship between
caloric output and work rate is essentially linear, or slightly
exponential (Fig. 1), this logically implies, or rather dictates,
25
either a constant or decreasing efliciency with increments
in metabolic load. The gross, net, and work efficiency
calculations (Table 2, Fig. 3) do not yield results in agree-
ment with those dictated by the original steady-rate work
data (Fig. 1). These efficiency calculations are considered
to be artifacts of the computation. In gross and net efli-
ciency definitions the base-line vo2 used in obtaining the
denominator of the efficiency calculation does not approxi-
mate the y intercept of the regression of vo2 on work rate.
Therefore, these definitions distort what is implied by the
near-linear relationship between whole-body Vo 2 and
work rate by making it appear (in the efficiency denomina- t----a Work
tor) that vo2 responses at higher work rates are pro- + + Theoretical Work
portionately less than at lower work rates. In contrast to Net o-----o Gross

these results, the delta efficiency calculation (Fig. 2) does Theoretical Net A-- Theoretical Gross

accurately describe the relationship between caloric output

and work rate and is, therefore, concluded to be the most
appropriate method of calculating efficiency. This method I RPM t I

also yielded the highest calculated efficiencies, ranging 40 60 80 100

from 24.4 to 34.0 % with respect to increments in work FIG. 4. Effect of pedaling speed on caloric output of human sub-
rate, and 14.6 to 3 1.7 % with respect to increments in jects during steady-rate exercise at 400 Kgm/min calculated by
pedaling speed. These values are considerably higher than gross, net, work, and delta definitions of efficiency (x =Is: SE). All
gross eficiency (range 7.5-20.4 %) and net efficiency definitions produce the result of decreasing efficiencies with respect
to increments in speed.

30
TABLE 4. effect of speed and work rate on

- =0 -_
_----p-,
re.$iratory exchange ratio (R = l&02/ VOW)
-
I __--------c”__--- - - - -4’- Work Rate, kgm/min
wm
200 400 600 800

40 0.82 rt 0.02 0.84 & 0.02 0.88 r+ 0.03 0.90 + 0.03

60 0.85 zt 0.04 0.86 zt 0.03 0.89 zt 0.03 0.90 zt 0.02
80 0.88 III 0.02 0.89 I+ 0.02 0.91 IIZ 0.02 0.92 zt 0.02
100 0.88 zt 0.02 0.89 zt 0.02 0.91 zt 0.02 0.93 Ifi 0.02

Values are expressed as means zt SE (N = 12). R at rest was

0.89 -+: 0.04; VOZ at rest was 0.286 rfi 0.010 l/min; caloric output
qt rest was 1.40 =t 0.05 kcal/min.

(range 9.8-24-l %), and slightly higher than work efficiency

(range 24.8-30.3 Y’o with respect to increments in work
Net rate, and 11.3-27.5 % with respect to increments in pedal-
A A Theoretical Net ing speed).
W-----O Gross The range of values for work and delta efliciency calcu-
b A Theoretical Gross lations with respect to changes in speed are lower than
values obtained in calculations with respect to changes in
work rate. This result is due to the method of computation
1 which considers both a work rate change and a pedaling
Kg-m /min I I frequency change. Since the relationship between caloric
200 400 600 800 output and work rate was slightly exponential, the effect of
FIG. 3. Gross, net, and work efficiencies (J’? ZII SE), using both work rate on efficiency is compounded with (additive to)
traditional and theoretical-thermodynamic methods of calculation, the effect of speed alone. Although this procedure of calcu-
are plotted as a function of work rate at 60 rpm. Apparent result of
increasing efficiency by gross and net definitions is considered to be
lation results in lower efficiencies, it is the only reasonable
an artifact of the methods of calculation. Note that work efficiency way to determine the effect of speed in work and delta
remains relatively constant. efficiency calculations (see footnote 2). As with gross and
CYCLE ERGOMETER STEADY-RATE WORK EFFICIENCY 1137

net efficiency calculations, in which there is no possibility term “delta efficiency” has not appeared in the literature
of speed or work rate influencing the base-line correction, before, the concept is not new (4, 7, 10). Benedict and
the work and delta efhciencies do describe the trend of Cathcart (4), in using a light but measurable load (0.5 A)
decreasing efficiency with increasing speed (Fig. 4). This as a base line, observed delta efficiencies to average about
method of calculating the effects of speed on efficiency has 30 %. This value agrees very favorably to the grand mean
not, to our knowledge, appeared in the literature before. of 28 % in the present investigation (based upon delta
For purposes of comparison we will restrict the discussion efficiency calculations with respec t to i ncrements in work
to the effect of work rate on work and delta efficiency calcu- rate at a constant pedaling speed).
lations. The literature on the effect of work rate on muscular
It is proposed that the gross, net, and work efhciency efficiency is very inconsistent (224, 7, 10, 15, 30). Benedict
definitions not be used in calculations of muscular efli- and Cathcart .observed gross (9-20 %), net (13-23 %), and
ciency. The gross efficiency calculation appears to be of work (22-41 70) efficiencies to increase with increases in
little value in determining muscular efficiency. It “indi- work rate. I. Astrand (3), using steady-rate exercise data
cates little of the potentialities of the human body for observed between the work rates of 50 and 600 kgm/min,
severe muscular work, and gives no conception of the reported net efhciency to increase from 10.5 to 23.5 %,
possible efficiency of the human body as a machine (4). ” although the relationship between vos and work rate was
The net efficiency, which has been the most frequently linear. These combined results are very similar to those in
used method in calculating what has also been termed the the present investigation and support the conclusion that
mechanical efficiency (3, 20, ZZ), still does not appear to these methods of calculation lead to computational arti-
be a satisfactory method of calculating muscular efliciency. facts in efficiency calculations. In contrast to the above
In this calculation the energy cost of moving the legs is observations, efficiency has been reported to remain con-
not considered in the estimation of work done by the exer- stant (lo), and decrease (2, 7, 15, 30) with increments in
cising subject (27). Tl lis undoubtedly results in an under- work rate. These inconsistencies may be a result of the
estimation of the “true” muscular efficiency. Since the varying methodologies employed, including the selection of
net efhciencv calculation involves the use of a constant appropriate base-line corrections to the efhciency calcula-
which falls considerably below the line describing the re- tion, use of postexercise Tjog in estimates of caloric cost of
lationship between caloric output and work rate (Fig. l), exercise, selection of work rates, and the time at which
use of resting metabolism as a base line leads to computa- vo2 is measured.
tional artifacts as illustrated in Fig. 3, which could be The present study demonstrates how the use of different
interpreted to mean that efficiency increases with work rate. definitions of efliciency yields varied results and probably
The work efficiency definition, which has been considered spurious conclusions about muscular efficiency. Of further
in several investigations (4, 10, 27), is good in theory but, consideration is the Vo2 measured to use in the estimation
depending upon the ergometer available, in practice is of energy expended by the exercising subject. In the present
very difficult to obtain (4, 10). We were unable to experi- experiments only steady-rate vo2 was used in efficiency
mentally obtain a Vo 2 that satisfactorily represented the calculations. IJnder these conditions the vo2 is assumed to
energy cost of moving the legs without a load (Fig. 1). be representative of the energy expended by the individual
Alternatively, this value rnay be obtained with regression to meet the metabolic requirements of the body. Therefore,
lines. If the relationship between caloric output and work postexercise vo2 should not be used in any calculations of
rate is linea r, which dictates a constant efficiency within muscular efficiency. Although the entirety of the post-
the given range of work rates, then any point on the line exercise 902 may legitimately be considered a part of the
(including the 0 work value obtained with the regression total metabolic cost of exercise, it certainly cannot be
equation) may be used as a base line, and the resulting equated with the anaerobic metabolism occurring during
efficiency calculation will accurately describe the metabolic exercise (1, 5, 6, 14, 2 1). Its inclusion into the caloric cost of
response to increases in work intensity. However, if the exercise would distort estimates of the energy expenditure
relationship between caloric output and work rate is not of working muscle during exercise.
perfectly linear (as in the present experiments), the use of The selection of work rates represents another problem
the y intercept for work efficiency calculations will lead in determining muscular eficiency from steady-rate er-
to erroneous results as do gross and net efficiencies. gometer work. It has been demonstrated that the time
The delta efficiency calculation utilizes a base line based taken to reach steady .-rate conditions increases with in-
uoon the first derivative of the increases in caloric cost of creases in work intensity (26). The work intensity and dura-
exercise with respect to ordered increases in work rate. In tion of exercise not only influence the kinetics of Vo2
using this “ floating base-line” changes at any point along during the work task but also influence the kinetics of
the regression of caloric output on work rate will be repre- the postexercise Vo2 (17, 26). Since both the duration and
sented in the efficiency estimate and, therefore, will always intensity of exercise have the effect of increasing the mag-
yield results consistent with those logicallv , implied from nitude of Vo2 after exercise (17), inclusion of postexercise
the data. If each increment in caloric exsenditure can be vo2 into the metabolic cost of exercise invariably results
assumed to represent the additional 902 of’only the working in the conclusion that intense exercise is accomplished at
muscles, to meet the additional metabolic requirements a decreased efficiency (2, 15, 17). Since many investigators
imposed by the increases in work rate, then the delta have utilized both light and heavy work loads, and did
efficiency calculation represents the most accurate estimate not calculate efficiency from steady-rate conditions, but
of muscular efficiency. Although to our knowledge the rather included postexercise Vo2 in their estimations of
1138 G. A. GAESSER AND G. A. BROOKS

energy expenditure, the inconsistencies with respect to the cannot be used to justify any one particular method of
effect of work rate on muscular eficiency may be explained. efficiency calculation. The theoretical-thermodynamic ap-
From the present data it can be concluded that regardless proach does provide the grounds for calculating exercise
of the approach used, whether traditional, utilizing TjTo2 efficiency, involving expressions for both the phosphoryla-
and R, or theoretical, involving thermodynamic considera- tive-coupling efhciency and the mechanical contraction-
tions, the efhciencies obtained are the same (in approxi- coupling efficiency (2 7).
mately 70 70 of the comparisons). The differences that were If we assume the phosphorylative-coupling efhciency to
statistically significant were small, 0.2-l 2 %. The con- remain constant at 60 Yo, the mechanical contraction-
sistent nature of these differences may be a result of one coupling efficiency appears to vary between 41 and 57 %
or rnore possibilities. One or a combination of the assump- (for delta efficiency calculation, effect of work rate). This
tions utilized in the theoretical-thermodynamic calculations range appears to be consistent with the value of 4970 ob-
(P/O = 3.0, AG for ATP = - 11 .O kcal/mol, phosphoryla- tained by Whipp and Wasserman (27), and with values
tive-coupling efficiency = 60 %) may be slightly in error. ranging from 35 to 50 % for isolated muscles (9). The vari-
In gross and net efficiency calculations the theoretical- ation in efficiencies found in isolated muscles was a result
thermodynamic approach yields values lower than the of the contractile properties of the muscle, with slow-twitch
traditional approach. This may be a result of an overesti- muscles demonstrating a greater efficiency than fast-twitch
mation of the P/O ratio, which would tend to inflate the muscles. These results, along with the recent data of Gibbs
denominator in the efhciency ratio and thus decrease the and Gibson (13) and Wendt and Gibbs (24), also demon-
estimated efhciency. The P/O ratio of 3.0 used in all calcu- strating energetic differences between the slow-twitch and
lations is a theoretical maximum, and is based upon the fast-twitch fibers in mammalian skeletal muscle, may ex-
oxidation of glucose (19). The P/O ratio for fats is less plain the present observation of a decreased efficiency
(for example, the P/O ratio for palmitate is 2.83). Since with both increased speed and work rate. In the latter two
R was always less than 1 .O in our experiments (Table 4), studies the efficiency of, the slow-twitch soleus was nearly
indicating utilization of some fat as substrate, a P/O ratio twice that of the fast-twitch extensor digitorum longus.
of 3.0 is undoubtedly an overestimate. Consequently, in Those authors proposed that the energetic differences may
the theoretical-thermodynamic calculations the energy be related to both the mechanically linked heat (represent-
expended by the subject was overestimated, and hence ing actomyosin activity) and the heat of activation (as-
lower gross and net efficiencies were calculated (Table 2). sociated with the movement of calcium out from and back
In contrast to the above results, the theoretical-thermo- into the sarcoplasmic reticulum). If during exercise more
dynamic approach yielded emciencies greater than the energetically inefficient fast-twitch fibers are selectively
traditional approach in the work and delta efficiency calcu- recruited as the speed of movement and/or intensity of
lations. This suggests that the base-line corrections in- work increases, then the results of the present experiments
fluence the nature of the differences. In the work efficiency may be explained. Data on R (Table 4) tends to support
calculations regressions for both voz and caloric output this hypothesis that fast-twitch fibers are recruited as speed
on work rate were used to obtain the y intercepts. Since R of movement and/or work rate increase. The observation
increases with work rate (Table 4), it has the effect of that R increases with both increased speed and work rate
increasing the caloric equivalent of the J?o2 with increments indicates a greater contribution of carbohydrates as sub-
in work, and thus creating a regression line with greater strate and hence an augmentation in glycolysis. Since
slope. Therefore, compared to the regression of caloric fast-twitch fibers are high in glycolytic capacity, more
output on work rate, the regression of vo2 on work rate may be preferentially recruited at higher speeds and/or
results in a greater relative base-line correction in the work rates, resulting in a reduction in muscular efficiency.
denominator and hence a greater efficiency using the The observation that theoretical-thermodynamic calcu-
theoretical-thermodynamic approach. The same principle lations of efficiency produce values closely corresponding
is applicable to the delta efficiency calculations where to those achieved by traditional means is perhaps the most
the change in caloric output is relatively greater than the powerful basis upon which to assess validity of the experi-
change in Vo2 for each increment in work rate. This again ments performed in vitro, from which the assumptions for
results in a greater estimate of energy expenditure in the the theoretical-thermodynamic calculations are derived.
traditional approach and, therefore, a lower efficiency.
Since the traditional approach considers the substrate
The authors thank Drs. G. L. Rarick and L. Packer for advice.
utilized in the estimation of energy expenditure, this ap- This research was supported by University of California Biomedical
proach is perhaps of more value in terms of the accuracy Research Development and Faculty Research Grants.
of determining muscular efficiency.
It is clear that the theoretical-thermodynamic approach Received for publication 17 October 1974.

REFERENCES

ALPERT, N. R. Lactate production and removal and the regula- BENEDICT, F., AND E. P. CATHCART. Muscular Work. Washington,
tion of metabolism. Ann. N. Y. Acad. Sci. 199: 995-1012, 1965. D.C. : Carnegie Institution of Washington, Publ. 187, 19 13.
ASMUSSEN, E. Aerobic recovery after anaerobiosis in rest and work.
BROOKS, G. A., K. E. BRAUNER, AND R. G. CASSENS. Glycogen
$cta Physiol. Stand. 11: 197-Z 10, 1946.
synthesis and metabolism of lactic acid after exercise. Am. J.
ASTRAND, I. Aerobic work capacity in men and women with
special reference to age. Acta Physiol. Stand. Suppl. 169 : 7-92, Physiol. 224 : 1162- 1166, 1973.
1960. BROOKS, G. A., K. J. HITTELMAN, J. A. FAIJLKNER, AND R. E.
CYCLE ERGOMETER STEADY-RATE WORK EFFICIENCY 1139

BEYER. Temperature, skeletal muscle mitochondrial. functions, and 19. LEHNINGER, A. L. Bioenergetics. Menlo Park, Calif. : Benjamin,
oxygen debt. Am. J. Physiol. 220: 1053-1059, 1971. 1971, p. 86-89.
7. CAMPBELL, J. M. H., C. DOUGLAS, AND F. HOBSON. The respira- 20. MOREHOUSE, L. E. Laboratory Manual for PhysioZogy of Exercise.
tory exchange of man during and after muscular exercise: Trans. St. Louis, MO. : Mosby, 1972, p. 153-154.
Roy. Sot., London, Ser. B 210 : l-47, 1920. 21. STAINSBY, W. N., AND J. K. BARCLAY. Exercise metabolism: 02
8. CONSOLAZIO, C. F., R. E. JOHNSON, AND L. T. PECORA. Physi- deficit, steady level 0 2 uptake and 02 uptake in recovery. Med.
ological Measurements of Metabolic Functions in Man. New York: Sci. Sports 2: 177-186, 1970.
McGraw, 1963, p. 5-9. 22. WASSERMAN, K., A. L. VAN’KESSEL, AND G. G. BURTON. Inter-
9. DAVIES, R. E. Bioenergetics of muscular contraction. In: Control action of physiological mechanisms during exercise. J. ApPZ.
of Energ_y Metabolism, edited by B. Chance, R. W. Estabrook, and Physiol. 22 : 7 l-85, 1967.
J. R. Williamson. New York: Academic 1965, p. 383-392. 23. WASSERMAN, K., B. J. WHIPP, S. N. KOYAL, AND W. L. BEAVER.
10. DICKENSEN, S. The efficiency of bicycle pedaling as affected by Anaerobic threshold and respiratory gas exchange during exer-
speed and load. J. Physiol., London 67 : 242-255, 1929. cise. J. A@Z. Physiol. 35 : 236-243, 1973.
11. Documenta Geigy-ScientiJic Tables, edited by K. Diem. New 24. WENDT, I. R., AND C. L. GIBBS. Energy production of rat ex-
York : Geigy Pharmaceuticals, 1962. tensor digitorum longus muscle. Am. J. Physiol. 224: 1081-1086,
12. DUFFIELD, F. A., AND J. MACDONALD. Relationship between speed 1973.
and efficiency. J. Physiol., London 58, Proc. xiii, 1923. 25. WHIPP, B. J., C. SEARD, AND K. WASSERMAN. Oxygen deficit-
13. GIBBS, C. L., AND W. R. GIBSON. Energy production of rat oxygen debt relationships and efficiency of anaerobic work. J.
soleus muscle. Am. J. Physiol. 223 : 864-87 1, 1972. AppZ. Physiol. 28: 452-456, 1970.
14. HARRIS, P. Lactic acid and the phlogiston debt. Cardiovascular 26. WHIPP, B. J., AND K. WASSERMAN. Oxygen uptake kinetics for
Res. 3 : 381-390, 1969. various intensities of constant load work. J. A@Z. Physiol. 33:
15. HENRY, F. M., AND J. DEMOOR. Metabolic efficiency of exercise 351-356, 1972.
in relation to workload at constant speed. J. Ap~l. Physiol. 8: 27. WHIPP, B. J., AND K. WASSERMAN. Efficiency of muscular work.
608-14, 1956. J. AppZ. Physiol. 26: 644-48, 1969.
16. HULTMAN, E. Studies on muscle metabolism of glycogen and ac- 28. WILKIE, D. R. Heat work and phosphorylcreatine break-down
tive phosphate in man with special reference to exercise and diet. in muscle. J. Physiol., London 195 : 157-183, 1968.
&and. J. CZin. Lab. Invest. SuppZ. 94 : 1-63, 1967. 29. WILKIE, D. R., AND R. C. WOLEDGE. The application of irre-
versible thermodynamics to muscular contraction. Comments
17. KNUTTGEN, H. G. Oxygen debt after submaximal physical exer-
on a recent theory by S. R. Caplan. Proc. Roy. Sot., London, Ser.
cise. J. APPZ. Physiol. 29 : 65 l-657, 1970.
B 169: 17-29, 1967.
18. KREBS, H. A., AND H. L. KORNBERG. Energy Transformations in 3. . WINSLOW, C. A., AND L. HERRINGTON. Temperature and Human
Living Matter. Berlin : Springer, 1957, p. 213. Life. Princeton, N. J. : University Press, 1949.

View publication stats