P1: OTA/XYZ P2: ABC

JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Bioenergetics of Exercising Humans

George A. Brooks*1

ABSTRACT:
Human muscles, limbs and supporting ventilatory, cardiovascular, and metabolic systems are
well adapted for walking, and there is reasonable transfer of efficiency of movement to bicycling.
Our efficiency and economy of movement of bipedal walking (≈30%) are far superior to those
of apes. This overall body efficiency during walking and bicycling represents the multiplicative
interaction of a phosphorylative coupling efficiency of ≈60%, and a mechanical coupling effi-
ciency of ≈50%. These coupling efficiencies compare well with those of other species adapted
for locomotion. We are capable runners, but our speed and power are inferior to carnivorous
and omnivorous terrestrial mammalian quadrupeds because of biomechanical and physiological
constraints. But, because of our metabolic plasticity (i.e., the ability to switch among carbohydrate
(CHO)- and lipid-derived energy sources) our endurance capacity is very good by comparison to
most mammals, but inferior to highly adapted species such as wolves and migratory birds. Our
ancestral ability for hunting and gathering depends on strategy and capabilities in the areas of
thermoregulation, and metabolic plasticity. Clearly, our competitive advantage of survival in the
biosphere depends in intelligence and behavior. Today, those abilities that served early hunter-
gatherers make for interesting athletic competitions due to wide variations in human phenotypes.
In contemporary society, the stresses of regular physical exercise serve to minimize morbidities
and mortality associated with physical inactivity, overnutrition, and aging.  C 2012 American

Physiological Society. Compr Physiol 2:537-562, 2012.

Introduction and possess greater endurance (106). For example, since the
Englishman Sir Roger Bannister achieved the first sub-4 min
Whether stated or not, the matter of human bioenergetics is mile (3:59.4 in 1954), that performance of running 4 min at a
implicitly fundamental to studies of human performance and speed of 15 mph has been repeated and surpassed countless
metabolism in health and disease. How the body acquires, times. However, at the time of this writing the World Record
stores, and utilizes energy has allowed our ancestors to hunt for the 2-mile run is 7:58.61 by the Kenyan Daniel Komen in
and gather, to escape predation, to survive famines and mi- 1997. That is an extraordinary feat, by a highly trained human
grations, to deal with heat, cold, the hypoxia of high altitude, athlete who is atypical for our species. By comparison, wild
and to carry children to term. Accordingly, the history of stud- wolf packs can cover 30 miles of rough terrain in an hour (98),
ies of human bioenergetics is broad and deep, ranging from and migratory birds can cover 12,000 kilometers in 12 days,
archeology and anthropology, to studies of tissue metabolism over open water without stopping for water or food (106).
and muscle mitochondrial energetics. Related are studies from Comparative physiologists (74) and anthropologists (17) hold
comparative physiology on the metabolic capabilities of ter- that traits of altitude tolerance and endurance running capacity
restrial, aquatic, and aerial animals. Human history, as re- are ancestral in humans, traceable to human origins in the high
flected in our human phenotypes confers success to some of plains of east Africa. Today, however, the capacity for humans
us in daily life as well as very different types of sports and to survive at high altitude (136) and run down antelope in
games; but, ironically, in contemporary developed societies the bush may have more to do with strategy and the human
some of the same capabilities that led to biological success capacity for temperature regulation than physical prowess.
of past generations can lead to the development of chronic In terms of the efficiency of walking and bicycling, in-
diseases related to physical inactivity. vestigators (36, 50, 51, 80, 107, 110, 132) commonly com-
In this article emphasis is on the bioenergetics of humans. pute whole-body muscular efficiency to approximate 30%.
So far as is known, we share similar cardiovascular and muscle That value for total body efficiency during steady-rate sub-
designs as with other animal species, but we are adapted for maximal exercise conditions agrees closely with the results
upright, bipedal locomotion freeing the hands and arms for
important tasks. But, how do we compare to other animals * Correspondence to [email protected]
in terms of physical prowess? Though celebrated in legend 1 Department of Integrative Biology, University of California, Berkeley,
(e.g., Phidippides of Marathon fame) and in the annals of the California
modern athletics (e.g., Abebe Bikila in the 1960 and 1964 Published online, January 2012 (comprehensivephysiology.com)
Olympiads) our capacity for locomotion is paltry compared DOI: 10.1002/cphy.c110007
to those of other species that are more powerful, faster (17) Copyright 
C American Physiological Society

Volume 2, January 2012 537

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Bioenergetics of Exercising Humans Comprehensive Physiology

Figure 1 Antoine-Laurent de Lavoisier measures oxygen consumption on co-investigator Armand

Seguin during foot treadle exercise, circa 1780. Drawings ascribed to his wife Marie Anne Paulze
Lavoisier, who depicted herself at the table on the far right. The drawing is entitled “Expérience
sur la respiration humaine” (Experiments into Respiration). Courtesy of the Division of Rare and
Manuscript Collections, Cornell University Library.

from ex vivo-measured efficiencies of “phosphorylative-” Antoine-Laurent de Lavoisier, co-discoverer of oxygen, was

(oxidative, 60%) and “mechanical-coupling” efficiencies the first to make determinations on resting and exercising
(50%) [e.g., (0.60) (0.50) = 0.30]. The terms phosphorylative- humans (Fig. 1). Subsequently, it came to be known that
and mechanical-coupling efficiencies are, respectively, used as with unicellular organisms, isolated muscles could work
to denote the percentage of energy released in the catabolism in the absence of oxygen [for reviews see (22, 54)]. Such
of energy substrates captured as adenosine triphosphate knowledge gave rise to the concepts of aerobic (oxidative)
(ATP), and the percentage of energy in ATP converted to and anaerobic (nonoxidative) sources of energy. But, whereas
mechanical work in the hydrolysis of ATP by interactions of in the early 20th century technology rapidly progressed in the
the contractile proteins actin-myosin. The value of human lo- ability to measure energy supply from oxidative metabolism
comotor efficiency during walking is superior to that of great in resting and exercising humans, it remained until the later
apes (124), but is less than that of other bipeds such as kanga- part of that century and the advent of muscle biopsy and
roos that are capable of harnessing kinetic energy from bound isotope tracer technologies to be able to estimate energy
to bound (30). supply from nonoxidative energy systems. Still, to their credit
In an elemental sense, the ability to undertake and sustain knowing that muscles possessed oxidative and nonoxidative
human muscular activity can be described in terms of the capacities for energy delivery, investigators in the early 20th
balance of energy demand and supply. Different forms of century were resourceful in determining oxygen equivalents
activity require (demand) different levels of energy which is for the energy supplied by nonoxidative energy sources.
supplied in the form of ATP. However, cellular levels of ATP One strategy was to compute the oxygen missed, that is,
are miniscule requiring real-time resupply of ATP used to that which would have been needed to meet energy demand
do cellular work, always, and particularly in exercise when solely by oxidative metabolism. Typically computed for the
muscle power output and work of supporting systems rise rest to exercise transition, the oxygen missed was commonly
more than an order of magnitude. With this understanding, termed the “oxygen deficit.” Similarly, recognizing that
power output exceeding the capacity of supply it cannot be exercise tasks of sufficient intensity to result in blood lactate
sustained. As a first step in broaching this immense topic accumulation were accompanied by a prolonged postexercise
of human bioenergetics the power and capacity of energy metabolic response, Hill and associates (67-70) developed
systems in muscle are described. the “oxygen debt” concept and used it in an attempt to com-
pute an oxygen equivalent of the nonoxidative energy flux
during exercise. Subsequently, with the discovery of the im-
Power and Capacity of Metabolic portant role of phosphagen [ATP and phosphocreatine (phos-
phorylcreatine or PCr)], using the biexponential recovery O2
Energy Systems in Human Muscle debt curve Margaria and associates (95) segmented the recov-
Knowing the sources of muscle energy has been an issue ery oxygen volume into two, “lactacid” and “alactacid” com-
almost since the beginnings of modern biology. Indeed, ponents. Although the terms are no longer in general use, it is

538 Volume 2, January 2012

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Comprehensive Physiology Bioenergetics of Exercising Humans

important to know these terms, their derivations, intents and Adenylate kinase is expressed in human muscle and
limitations (51). catalyzes the reaction 2 ADP → ATP + AMP. Adenosine
monophosphate rises significantly in working human skeletal
muscle (119), and through kinase activation AMP is thought
Immediate energy sources in human muscle to be an important signaling molecule stimulating both short-
ATP is utilized as the chemical energy source for actin-myosin and long term metabolic adjustments. However, the energy
interactions and most other endergonic, energy-requiring yield from myokinase is small. Estimates vary (21, 38),
reactions in muscle such as Ca++ resequestration into the sar- but capacity is likely less than 1.0 kcal in 30 kg of human
coplasmic reticulum (SR) and plasma membrane Na+ -K+ ex- muscle.
change. There are three “immediate” or “phosphagen” energy
sources available in skeletal muscle: ATP, PCr, and adenylate
kinase (myokinase). Measuring the use of these energy stor- Glycogenolysis and glycolysis
age forms is difficult during exercise for several reasons. The In human muscle, the phosphagen energy stores possess char-
stores are finite, turnover is very rapid, and assessment by acteristics of high turnover, but low capacity requiring support
muscle biopsy and 31 P-MRS (magnetic resonance spectrom- from glycolytic and oxidative energy sources. During steady-
etry) is challenging. Fortunately, however, in some respects state submaximal exercise, glycolytic and oxidative energy
Nature’s design makes it possible to estimate energy flux rates fluxes can be estimated from the combination of isotope tracer
and capacities. Even though ATP is the immediate energy and indirect calorimetry techniques (vide infra). Although sig-
source for actin-myosin interactions and related processes, nificant lactate production occurs during steady-state exercise
the design of the muscle energy system is to maintain [ATP] when muscle and blood lactate concentrations are constant,
homeostasis. Therefore, despite very rapid turnover, there is most is disposed of via oxidation, so the rate of oxygen
little net change in muscle [ATP] during physical exercise consumption accurately includes and represents the energy
(16, 77, 119). Hence, whether muscle [ATP] is estimated by contribution of “aerobic glycolysis.” This means that during
enzymatic analysis following muscle biopsy or NMR spec- short-term, high intensity exercise the net rise in the body
troscopy, the estimate of ATP storage capacity (5-6 mMol/kg lactate pool represents the energy contribution of “nonoxida-
wet weight) is small (<1.0 kcal/kg muscle) and the absence tive,” “anaerobic” glycolysis. Reasonable assumptions are
of net [ATP] change during exercise leads to discounting of that 1.0 ATP is produced per lactate anion accumulated when
ATP as a major muscle energy storage form. glucose is the precursor, and that 1.5 ATP is produced per
In contrast to ATP content, [PCr] is five to six times lactate anion accumulated when glycogen is the precursor. As
greater in human muscle (16, 38), and [PCr] does decline described below, from muscle biopsies and arterial-venous
during exercise in proportion to relative exercise intensity (a-v) difference measurements of net lactate release and
(REL) (16). Assuming a G of –11 kcal/mol PCr, muscle lactate net accumulation can be estimated during nonsteady-
[PCr] = 30 mMol/kg, and an active muscle mass of 30 kg state exercise (5). Alternatively, from the rise in blood
in a 70 kg human, then the human muscle storage capac- lactate accumulation following maximal exercise efforts, and
ity of PCr approximates 10 kcal. Considering the capacity assuming that lactate mixes rapidly with total body water
for human muscle to hydrolyze ATP during maximal efforts [an assumption now justified with the finding of ubiquitous
(45 kcal/min) (94), without replenishment, PCr would be ex- expression of lactate (monocarboxylate) transporter (MCT)
hausted within a fraction of a minute during hard muscular isoforms], nonoxidative energy production can be estimated.
exercise (Table 1). In the past measurements of “lactate tolerance” by Rodolfo
Margaria, Paolo Cerretelli, Pietro diPrampero, Guido Ferretti
Table 1 Estimates of the power and capacity of human muscle en- and colleagues (the Milan School), have yielded estimated of
ergy systems 15 to 20 kcal energy capacity in a 70 kg human with 30 kg
of human muscle (94). In agreement, others have produced
Power Capacity Capacity oxygen Duration similar estimates (21, 84). As with phosphagen energy stores,
(kcal/min) (kcal) equivalent (L) (s)
the capacities for glycogenolysis and glycolysis in human
muscle are characterized by high turnover (> 20 kcal/min),
Immediate 45a 11b 1.4 15
(phosphagen)
but low capacity which means during maximal efforts the
(alactic) energy system is exhausted in 30 to 40 s such that continued
Glycolytic 22a 15 3 40 efforts and recovery require support from oxidative energy
(lactic)
sources.
Oxidative 25c 1,200 240 ∝d

Assumptions: 70 kg body weight, 30 kg active muscle.

a Power estimates from Margaria et al. (94). Aerobic energy power and capacity
b Phosphagen energy capacities from Edwards (38) and Karlsson et al.
Compared to immediate and glycolytic energy production
(84).
c VO
2max = 5 L/min.
pathways, oxidative metabolism is denoted by slow response
d 80% VO
2max sustainable for 60 min. [t/2 ≈ 30 s (29)], low to moderate power, but very large

Volume 2, January 2012 539

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Bioenergetics of Exercising Humans Comprehensive Physiology

capacity. In overnight fasted humans habituated to a mixed able evidence that the price of speed in fast muscle is a loss
diet, either resting or engaged in prolonged, submaximal of economy (123).
exercise eliciting a pulmonary respiratory exchange ratio
(RER = VCO2 /VO2 ) of 0.82, approximately 4.82 kcal are
derived per liter O2 consumed. From a resting baseline of
approximately 2.6 to 3.3 ml/kg body mass per minute (27), Human Muscle Power Output and
resting metabolic rate approximates 1.0 kcal/min with no ap- Oxygen Consumption In Vivo
preciable contributions from phosphagen pools or anaero-
In the process of oxidative phosphorylation, working mus-
bic glycolysis. From a resting baseline, peak aerobic energy
cle consumes oxygen and produces carbon dioxide. However,
production will depend on maximum oxygen consumption
measuring the metabolic response of human muscles to grada-
(VO2max ), such that a person capable of raising their rest-
tions in power output is not trivial. In the past, most measure-
ing VO2 20-fold will have twice the capacity for aerobic
ments were whole-body (pulmonary) oxygen consumption,
metabolism (i.e., 20 multiples of resting metabolism, METS)
vide infra (Figs. 16-18). In using pulmonary gas exchange to
compared to someone capable of only 10 METS. Further, a
reach conclusions about working muscle investigators assume
person with a large MET capacity will typically accomplish a
that working muscle dominates the whole-body metabolic re-
given exercise task with a lower pulmonary RER, and presum-
sponse. In retrospect, that assumption proves to be correct
ably lower muscle respiratory quotient (RQ), than a person
because of the extraordinary and innovative efforts of a few
with a lesser aerobic capacity while exercising at the same
investigators who have employed combinations of pulmonary
power output. Consequently, an individual with a greater aer-
and muscle measurements (12,103,134). More recent studies
obic capacity will derive a lower percentage of energy from
of muscle metabolic measurements have involved determina-
CHO-derived fuel sources (glycogen, glucose, and lactate),
tions of indirect calorimetry and NMR spectroscopy (89).
and a greater percentage of energy from lipid energy sources
[plasma FFA and intramuscular triglycerides (IMTGs)] com-
pared to an individual with a lesser aerobic capacity.
Assessments of muscle respiration by
simultaneous pulmonary and working
muscle measurements
Energetics of isolated muscles and muscle fibers Working leg muscle oxygen-consumption rates are rare due
As described by Mommaerts (100), early muscle physiolo- to the invasiveness and difficulties in measuring blood flow
gists such as Hill (65) used thermopiles to measure the en- and femoral arterial and venous oxygen and carbon dioxide
ergies (heats) of muscle activation and shortening in nonper- contents. Still more rare in the literature are measurements in
fused frog muscles stimulated to contract at room temperature. which leg respiratory quotient (RQ = VCO2 /VO2 ) were deter-
Over decades the device was improved (66), and subsequently mined over a range of power outputs so that working muscle
initial (contraction) and latent (recovery) heats could be ob- efficiency could be determined. An example of a study in
served not only in amphibian muscles, but also in mammalian which pulmonary and working leg muscle rates of oxygen
muscle bundles. Ultimately, initial heats could be described consumption were determined simultaneously over a range
as representing separate heats of activation and shortening of exercise intensities (leg ergometer cycling) is that of David
(100). Activation heat is measured during isometric or iso- Poole, Peter Wagner and colleagues (107) (Fig. 2). Results
tonic contractions and is now accepted to be associated with of that study were subsequently reproduced (47). In Poole
ATP hydrolysis attributable to ion pumping, specifically the et al., from the inverse of the regression of the caloric equiv-
cell membrane Na+ -K+ -ATPase and, quantitatively more im- alent of pulmonary VO2 on the caloric equivalent of external
portantly, the Ca++ -ATPase associated with calcium ion re- power output during leg cycling muscular efficiency deter-
sequestration into the SR (43, 100). mined from pulmonary gas exchange averaged 29.1 ± 0.6%
In his initial experiments, Hill (65) observed that if mus- (95% CI 27.9-30.3%). Simultaneously, muscle efficiency cal-
cles were allowed to shorten additional heat was released; culated from the leg VO2 averaged 33.7 ± 2.4% (95% CI 29.0-
this was termed “shortening heat” and subsequently became 38.4%) (Fig. 2). If for these experiments the phosphorylative
associated with the work done (42), and now ascribed to coupling efficiency is taken to be 60%, then the mechanical
myosin-actin cross-bridge interactions in addition to those coupling efficiency for human leg muscle is 56%.
used for isometric contraction (100). Fast muscle fibers are In the above example, the confidence intervals for muscle
distinguished by unique myosin isoforms that have high AT- are larger than those for pulmonary VO2 . Among the technical
Pase activity and rapid cross-bridge turnover (6-8). As well, problems associated with the muscle are catheter placement
fast fibers are arranged in larger motor units, innervated by (ante- or retrograde), the admixture of blood from nonmuscu-
larger, more heavily myelinated alpha motor neurons with lar tissues or auxiliary muscles, and determination of arterial
more extensive motor end plates (8). Given its design fea- and venous oxygen contents. In the example cited, blood CO2
tures, there is every reason why fast fibers contract and relax contents were not measured, and so a RQ of 0.95 as seen
more rapidly than do slow fibers. As well, there is consider- in previous studies and subsequently reconfirmed by us (12),

540 Volume 2, January 2012

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Comprehensive Physiology Bioenergetics of Exercising Humans

4 1.05
Pretraining
1 Posttraining
Pulmonary 45% Pre
3 r = 0.999 0.95 65% Pre
Y = 0.68 + 0.0099X 65% Old (ABT)
VO2 (L.min–1)

0.9 65% New (RLT)

RQ
2 Two Leg 0.85
r = 0.989
Y = 0.29 + 0.0092X 0.8
1
0.75

0.7
0 Rest Exercise
0 50 100 150 200 250 300 350
* Significantly different from pretraining (45%) at P < 0.05.
Work Rate (Watts) Δ Significantly different from rest P < 0.05.

Figure 2 Group mean values (± SE) of studies on 17 men showing Figure 3 Working muscle respiratory quotient (RQ) as determined
linear and parallel increments in pulmonary and working muscle (leg) by femoral arterial and venous CO2 and O2 concentration difference
oxygen uptake. Pulmonary and leg VO2 measurements yield delta ef- measurement. Values are means ± SEM for eight subjects. Exercise
ficiencies of measurements (29.1 ±0.6%) and (33.7 ± 2.4%) for whole data are means of last 30 min of exercise. Subjects studied 2× before
body and working muscles, respectively. The difference between y in- training (i.e., @ 45% and 65% VO2peak ), and twice after training (i.e.,
tercepts shows that the body provides significant metabolic support for ABT = same absolute intensity as 65% pretraining, and RLT = same
processes outside the exercising legs, but that these metabolic “costs,” relative intensity as pretraining, i.e., 65% VO2peak ). *Significantly dif-
such as the work of breathing and gluconeogenesis, change in pro- ferent from pretraining (45%) (P < 0.05);  significantly different from
portion to muscle power output. From Poole et al. (107) and used with rest, P < 0.05). Results show working muscle to be carbohydrate de-
permission. pendent, both before and after training. From Bergman et al. (12) and
used with permission.

was assumed to compute muscle RQ and the caloric equiva-

lent of VO2 . Considering literature values for working muscle working muscle RQ to be slightly higher than pulmonary
RQ (12, 47, 103), the assumption of an RQ of 0.95 was ap- RQ. The differences are small (0.03), but consistent suggest-
propriate, especially for higher power output exercises. As a ing relatively greater CHO, and less lipid oxidation in working
result of parallel slopes in pulmonary and working muscle res- muscle as compared to the whole body. In our studies, glu-
piratory in responses to muscle power output requirements it cose (13,47,48) and lipid (12,47,48,79,129) whole-body and
can be concluded that human muscle contracts with an overall working muscle fluxes were measured by the combination of
efficiency on the order of 30%. Again, that value for an over- tracers and limb net balance techniques. Like the results of
all efficiency agrees well with ex vivo estimates of 50% for others (112) our measurements showed that most (≥90%) of
mechanical coupling efficiency and 60% for phosphorylative glucose disposal (Rd) was accounted for by net uptake by
coupling efficiency, respectively. The efficiency and economy working muscle (13). As well, we could not observe a net
of human muscle exercise are discussed below. change in muscle IMTG content by muscle biopsy analysis
(12,49,129), and net releases of glycerol (a surrogate for intra-
muscular lipolysis) was small. Overall, these results support
A pulmonary respiratory exchange the notion of a shunt of available carbohydrate energy stores
ratio—working muscle respiratory to working muscle with a relatively greater reliance on lipid
quotient paradox? energy sources by the remainder of the body during muscle
By classic as well as contemporary standards measuring pul- exercise (12, 13).
monary carbon dioxide production, hence RER (or R =
VCO2 /VO2 ) is a standard technique well within the technical
capabilities of most clinical and research laboratories. How-
ever, measuring arterial and venous O2 and CO2 contents as
Classic Compared to
well as tissue (usually working limb) blood flow is technically Theoretical-Thermodynamic
demanding. Of the parameters to be measured, the CO2 con- Approaches to Estimating Energy
tent of arterial and venous blood is perhaps most complex, Yield from Substrate Oxidation
and not always attempted, as in Poole et al. (107) (Fig. 2).
Still, knowing the “combustion coefficient” of working mus- Working muscle oxygen consumption, RQ, P/O,
cle provides invaluable information on the energy-substrate and energy yield
partitioning in working muscle. From the first law of ther- There are two approaches to solving the problem of determin-
modynamics (principle of conservation of mass) one would ing the energy yield per unit of oxygen consumed. The classic
anticipate that over time RER and RQ would be equivalent. approach is to utilize indirect calorimetry to determine pul-
However, in two investigations, we (12, 47) (Fig. 3) found monary VO2 and VCO2 , and then to use standardized tables to

Volume 2, January 2012 541

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Bioenergetics of Exercising Humans Comprehensive Physiology

determine the energy expenditure in kcal/min (or kJ). A more 7,826 kcal/mol. In contrast, for glucose (180 g/mol), the en-
recent approach is to use indirect calorimetry to determine ergy yield is (4.2 kcal/g × 180 g/mol =) 756 kcal/mol, or
VO2 , and from there calculate the ADP to ATP phosphoryla- one-tenth that of lipid.
tion rate using assumptions from contemporary biochemistry;
the latter is referred to as the “theoretical-thermodynamic
approach,” a term first used by Whipp and Wasserman Theoretical-thermodynamic approach
(132) and adopted by others (50, 80). The two methods, to determining metabolism
classic and theoretical-thermodynamic, yield very similar In contrast to the use of standard tables derived at the end of
results, with advantages being in favor of the classic method the 19th century, and modified early in the 20th century used
in terms of ease of use, particularly when employing the by metabolic biochemists and physiologists, a structural bio-
RER for evaluating effects of variables (nutrition, exercise, chemist’s accounting involves determining the ATP yield per
training, gender, aging, and environment) on energy substrate unit of substrate oxidized. For glucose oxidation (C6 H12 O6 +
partitioning. 6 O2 → 6 CO2 + 6 H2 O), and depending on the cytoplasmic
to mitochondrial redox shuttle used (i.e., glycerol-phosphate,
malate-aspartate, or lactate shuttle), the ATP yield from
The classic approach glucose oxidation will be 36 to 38 mol ATP/mol glucose (91),
to determining metabolism making the P/O (inorganic phosphate, or ADP phosphory-
In classic metabolic biochemistry based on the cumula- lated, per atom of oxygen consumed) for glucose oxidation
tive efforts of Antoine-Laurent de Lavoisier, John Scott ≈ 36/12 to 38/12, or 3.0 to 3.17. This is because the malate-
Haldane (57), Francis Benedict (9, 10), C. Gordon Dou- aspartate and lactate shuttles are NADH-linked, and the
glas (37), A. Monmouth Smith (118), Wilbur Olin Atwa- P/O for NADH is 3.0 (90), whereas the glycerol-phosphate
ter (1), Edward Bennett Rosa (1, 2), and Nathan Zuntz shuttle uses FADH2 as the reducing equivalent (hydride ion)
(137) and their colleagues, if the pulmonary RER equals carrier, with the P/O for FADH2 is 2.0. If NADH-linked
1.0 (RER = VCO2 /VO2 = 1.0) carbohydrate is the cytoplasmic to mitochondrial lactate shuttles are assumed
fuel (e.g., for glucose: C6 H12 O6 + 6 O2 → 6 CO2 + (P/O=3), then for glucose an ATP yield of 38 ATP/mol
6 H2 O, RQ = VCO2 /VO2 = 6/6 = 1.0) and the energy yield glucose and G of −11 kcal/mol ATP yields −418 kcal/mol
from bomb calorimetery (81) is 5.05 kcal/L O2 . If on the glucose, or 418 kcal/180 g, or 2.32 kcal/g effectively trapped
other hand, the pulmonary RER approximates 0.71 (e.g., for in the form of ATP from aerobic glycolysis with a P/O of
the most common triglyceride Trioleate: C57 H104 O6 + 80 O2 3 to 3.2.
→ 57 CO2 + 52 H2 0, RQ = 57/80 = 0.71), lipid is the fuel Similarly, for the oxidation of a typical fatty acid such
and the energy yield is 4.69 kcal/L O2 . As originally described as palmitate (16 C), a structural biochemist would calculate
by Zuntz and Schumburg (137), and subsequently modified that the ATP yield from palmitate oxidation to be 129 mol
by Lusk (93), the relative contributions of lipids and carbo- ATP per mol palmitate. In palmitate oxidation (C16 H31 O2 +
hydrates to the fuel energy source and their contributions to 23 O2 → 16 CO2 + 145 H2 O, RQ 16/23 = 0.70), with
energy flux could be interpolated for RER values between one ATP (yielding AMP) used for activation, seven cycles of
1.0 and 0.71. Consequently, using CHO combustion as the the β-oxidation pathway will yield 7 NADH and 7 FADH2,
standard, 7.7% more energy is liberated per unit oxygen con- or (3+2=) 5 ATP/β-oxidation cycle, or 35 ATP from β-
sumed if CHO is the fuel compared to lipid. Hence, to a oxidation cycling associated with the catabolism of a single
metabolic biochemist the advantage of CHO combustion is palmitate molecule. As well, each β-oxidation cycle yields an
obvious when oxygen supply is limited. Acetyl-CoA; each TCA cycle yields 3 NADH, 1 FADH2, and
To a metabolic biochemist (73) or physiologist (20,21,49), 1 GTP (≈ 1 ATP), or 12 ATP/cycle; so TCA cycle activity
the advantage of lipid oxidation is revealed when the en- from palmitate yields 96 ATP associated with the catabolism
thalpy/unit weight of lipid is considered with a yield of of a single palmitate molecule. Hence, for palmitate the yield
9.1 kcal/g (38 kJ/g, 1 kcal = 4.19 kJ) as opposed to a value is 96 ATP from β-oxidation + 35 ATP from tricarboxylic
of 4.2 kcal/g (18 kJ/g) for CHO oxidation. Hence, given the acid cycle (TCA) cycle activity – 2 ATP for activation =)
2.17 ratio of energy per gram of lipid compared to energy 129 ATP with an overall P/O (129 ATP/46 O) = 2.8. In terms
per gram of carbohydrate, lipid yields 117% greater yield per of energy from the oxidation of palmitate, and ATP yield of
unit weight making lipid the preferred energy storage form in 129 mol/mol palmitate and a G of −11 kcal/mol ATP, the
biology. The superior energy yield of lipids over carbohydrate energy yield from palmitate is −1,419 kcal/mol palmitate,
is further amplified because glycogen is hydrated in vivo. Es- or 1,419 kcal/256 g, or 5.54 kcal/g effectively trapped in the
timates vary, but a value ≥ 2.7 g H2 O/g glycogen is widely form of ATP from the oxidation of 1 g of palmitate, again
used (116). Superiority of the energy density (compactness) with a P/O of 2.8. As implied from the greater energy yield of
of lipid over glycogen energy storage is further emphasized CHO compared to lipid oxidation (5.05 vs. 4.69 kcal/L O2 , a
when the energy/mol equivalent is computed. For an aver- 7.8% difference), the P/O from the oxidation of carbohydrate-
age triglyceride, 860 g/mol is typically assumed (12, 454, derived oxidative catabolism (3.0 vs 2.8 mol ATP/mol atom
46, 129); so the energy yield is (9.1 kcal/g × 860 g/mol =) oxygen) represents a 7.1 % difference.

542 Volume 2, January 2012

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Comprehensive Physiology Bioenergetics of Exercising Humans

Before leaving this section, it is appropriate to note the nitrogen balance associated with physical exercise, for studies
conclusion of a greater energy yield per atom of oxygen for of energy substrate partitioning during exercise, investigators
CHO dervied, as opposed to lipid-derived substrates from (e.g., 11, 128) have not bothered to correct pulmonary RER
classic indirect calorimery and the stoichiometry of biochem- measurements for amino acid and protein use during exercise.
istry is also supported by studies on isolated mitochondria By default then, the general view became that carbohydrate-
(18). However, because of shattering of the mitochondrial and lipid-derived fuels are preferred energy substrates for
reticulum during isolation of mitochondrial vessicles, proton exercise.
leaks across inner mitochondrial membranse are significant With the advent of isotope tracer methodology several
to the extent the P/Os are affected, making the difference be- sets of investigators evaluated the effects of various stresses
tween energy yields from lipid- and CHO-derived fuels 15% on amino acid metabolism in muscle preparations (102), intact
per atom of oxygen. Nonetheless, it is clear that regardless rats (61, 133), and intact functioning humans (99, 135). Such
of method, CHO oxidation provides more energy when oxy- studies forced recognition that the oxidation of some classes
gen supply is limited, such as during hard exercise. Energy of amino acids, such as branched-chain amino acids, increase
yield per gram of lipid is, however, far greater than for CHO during exercise. However, it is also apparent that while the
oxidation in a unit mass basis (vide supra). oxidation of ketogenic amino acids such as leucine scales
Whether by classic or theoretical-thermodynamic ap- to metabolic rate during exercise, overall total flux changes
proaches there is no evidence that exercise training improves little, with increased disposal via oxidation compensated for
the mitochondrial or glycolytic P/O, which represents the by decreased non-oxidative disposal (99).
efficiency by which the chemical potential energy in dietary- So far as the effect of exercise training on oxidative dis-
derived energy substrates is converted to ATP. However, train- posal of key amino acids is concerned, from earlier studies
ing does raise the maximum rate of oxygen consumption and on trained and untrained rats (61) it appeared that training
decreases the RER at which a given submaximal power output increased oxidative disposal of leucine during exercise. The
can be accomplished. As well, the learning of complex neuro- result made sense as training increased muscle mitochondrial
muscular tasks may improve the economy of movement, and mass (35,75). However, in studies on men and women studied
in the aggregate these contribute toward the improvements both before and after endurance training McKenzie et al. (97)
in performance associated with training. However, as is cur- showed that training suppressed leucine oxidation whether
rently known, the P/Os for CHO- and lipid-derived substrates measured at given absolute or relative intensities, especially
approximate 3 and 2.8, respectfully. If training shifts energy- in young women who were also capable of greater lipid oxi-
substrate partitioning to greater lipid use, then the measured dation than their male cohorts.
VO2 to accomplish a task may rise slightly. While exercise has minimal effects on amino acid flux
rates and nitrogen balance, it remains that amino acids and
proteins play important roles in sustaining physical exercise;
Protein and amino acids as fuels but, of those roles, that of an energy substrate during exercise
Dietary proteins provide essential building blocks to body is perhaps the least important. Still, the assumption that zero
structures, whereas dietary carbohydrates and lipids serve to amino acid oxidation occurs during exercise is unjustified and
provide the energy for diverse body functions including pro- a matter of concern. Estimates of the roles played by amino
tein synthesis. All three classes of dietary energy sources acids as fuel sources during exercise vary from 4% to 10%, but
require digestion, assimilation, distribution, and cellular up- there is little security in those estimates as individual amino
take prior to entry into final common metabolic pathways of acids behave differently and tracing one gives little confidence
catabolism. In the process of becoming energy sources, amino about what is happening to the others assessed by leg RQ
acids require one additional step, nitrogen removal, that is ac- or pulmonary RER determinations. This situation is better
complished by trans- or deamination. Hence, in terms of the in estimating plasma FFA oxidation, as one fatty acid (e.g.,
time from assimilation, and the metabolic power of the three palmitate or oleate) can be traced, and knowing its absolute
classes of dietary energy sources the hierarchy of substrate and relative concentrations as well as its flux and oxidation
use is CHO > lipid > amino acids. rates, the assumption can be made that what happens to the
Early on in the study of human nutrition, it became traced plasma FFA is representative of the total.
obvious that physical exercise had little effect on urinary Another approach to estimating the relative role of pro-
nitrogen excretion (28, 86, 87). Subsequently, in detailed teins and amino acids to the fuel energy source can be made
studies of total body nitrogen balance involving strict from dietary records (79). From urinary nitrogen excretion
dietary controls and the collection of urinary, fecal, and measurements or food records obtained on well-nourished
sweat nitrogenous products Todd et al. (126) established and weight stable individuals the contribution of amino acids
that so long as dietary energy was adequate to cover need, and proteins to daily energy expenditure can be made. For in-
physical activity did not result in negative nitrogen balance. stance, if dietary proteins and amino acids supply 15% of daily
Subsequently, with knowledge that little additional urinary energy intake, then the contribution of proteins and amino
nitrogen is excreted because of exercise, and because of acids to pre-exercise energy expenditure can be assumed
the inconveniences and difficulties in measuring changes in to be 15%. Next, assuming that total amino acid oxidation

Volume 2, January 2012 543

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Bioenergetics of Exercising Humans Comprehensive Physiology

60 60

Glycogen utilization (mmol glucosyl units/kg/min)

Glycogen util Glycogen
50 Glucose Ra 50

FFA and glucose Ra (umol/kg/min)

FFA Ra
Coggan et al. 1995
Glucose
40 40

30 30

20 20
FFA

10 10

0 0
0 20 40 60 80 100
Percentage of VO2 max

Figure 4 Results of an extensive literature search showing blood glucose and free
fatty acid flux rates (Ra) and net muscle glycogenolysis (77) as functions of relative
exercise intensity (REL) as given by % VO2max . This form of analysis indicates exponential
increments in muscle glycogenolysis and glucose Ra as functions of relative exercise
intensity. In contrast, the analysis shows multicomponent polynomial response of plasma
FFA flux, with easy to moderate intensity exercise (i.e., 25%-40% VO2max ) eliciting a large
rise in flux, but crossover and decreasing flux at approximately 50% VO2max . Note that
plasma free fatty acid (FFA) flux is predicted to reach minimal values as VO2max is
approached. For glycogen utilization, y = 2.11 e(0.04×), r2 = 0.87; for glucose Ra, y
= 9.8 e(0.02×), r2 = 0.84, and for FFA Ra, y = −0.833 + 1.14× − 0.013×2, r2 =
0.87. From Brooks and Trimmer (24) and used with permission.

remains constant during exercise in which total energy ex- which rates of O2 consumption and CO2 production are sub-
penditure rises, perhaps 15-fold, then amino acids contribute maximal and constant (steady), both before and after 10 to
1% of total energy expenditure. If, on the other hand, amino 12 weeks of endurance training. This design permits pre- and
acid oxidation rises 50 % (97) or 400% during exercise (99), posttraining comparisons to be made at given absolute power
estimates are that amino acids contribute 2% to 4% of total outputs and relative exercise intensities. On the basis of those
energy substrate, which has the effect of lessening the calcu- investigations as well as data from a variety of sources, it is
lated lipid oxidation by a corresponding amount. Hence, as clear that results follow predictions of the Crossover Concept
originally suspected from the lack of effect on urinary nitro- (23) (Fig. 4) such that CHO oxidation predominates in work-
gen excretion, the fuel energy role of amino acids is relatively ing muscle and at the whole-body level and that lipid is used
small in well-nourished individuals, with the result being that sparingly in working skeletal muscle, especially if exercise
the total body RER is a reasonable estimate of the nonprotein intensity is greater than 65% VO2max , or if exercise duration
RQ during exercise, and a correction for amino acid oxidation is short (12, 49, 63, 64, 82, 103, 129).
is typically not used even when available (79). While developed on humans, (Fig. 4) the Crossover Con-
cept appears applicable to other mammalian species with data
obtained on dogs goats and rats all seeming to follow the
human pattern (Fig. 5) (113). In the figure shown, given a
Exercise and Energy Substrate composite of lipid and CHO oxidation rates across ranges of
Partitioning in Humans relative exercise intensities, the comparisons between dogs
and goats are interesting. The animals were of similar body
Crossover Effects masses, but whereas running goats have aerobic capacities
Over the last several decades both cross-sectional and longitu- like those of humans (40-60 ml/kg/min VO2max ), running dogs
dinal training studies to determine effects of gender, exercise have much higher aerobic capacities, similar to those of rats
and exercise training on lipid metabolism in young men and (80-100 ml/kg/min VO2max ). Still, despite significant differ-
women have been conducted. For example, in our studies ences in aerobic capacities and body sizes, when expressed
we have studied subjects engaged in moderate (45%-50% relative to VO2max the balance of lipid and CHO use is well
VO2peak ) and hard (65% VO2peak ) intensity exercises during preserved across mammalian species.

544 Volume 2, January 2012

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Comprehensive Physiology Bioenergetics of Exercising Humans

100 1.05

0.95
80

RER
0.9

0.85
Fuel oxidation (% MO2)

60
0.8

0.75
10 20 30 40 50 60 70 80 90
% VO2 peak
40

Figure 6 Relationship between respiratory gas exchange ratio

RER = VCO2 /VO2 in trained (T) and untrained (UT) men during sus-
tained exercise. Subjects were studied in fed and over night fasted
(postabsorptive) conditions. Trained and fasted individuals have the
20 lowest RERs, but only during easy to mild-intensity exercise. Overall,
results show predominance of carbohydrate (CHO) over Lox during ex-
ercise regardless of training state or dietary condition. From Bergman
and Brooks (11) and used with permission.

0
0 40 60 80 100 seven category two cyclists during graded exercise from 20%
Exercise intensity (% MO2 max) to 80% VO2peak after a day of rest to normalize muscle glyco-
gen levels. To assure steady pulmonary gas exchange and
Figure 5 Effect of exercise intensity on the balance of lipid (◦) and blood lactate values, measurements were made over periods
carbohydrate () oxidation in four mammalian species (goats, dogs, of 45 min (80% VO2peak ), 90 min (60% VO2peak ), and 120 min
rats, and humans). Mean (± SEM) data on dogs and goats are shown
in the center and quadrants; other data on rats and humans have been
(20% and 40% VO2peak ). Subjects were studied after either a
included from the literature. Regardless of body size and configuration 12-hour overnight fast, or 3 h after a 500 kcal (53% CHO,
or aerobic capacity, the same patterns of energy substrate partitioning 2% fat, and 16% protein), low glycemic index breakfast. Ef-
during physical activity are apparent. In this respect the data are to be
compared to those in Figure 1. Redrawn, with permission, from two
fects of exercise, prior endurance training, and recent nutrition
separate figures in Roberts et al. (113) by GAB. were evident. Training decreased RER (increased lipid oxi-
dation) most notably at low relative power outputs, especially
when trained subjects were studied overnight fasted. However,
Exercise, exercise training, and preexercise when relative power output increased from mild (20%-40%
nutrition effects on whole body energy VO2peak ) to moderate (60% VO2peak ) and hard (80% VO2peak )
substrate partitioning as assessed from RER exercises, RER increased to values close to unity indicating
determinations predominance of CHO oxidation in all subjects regardless of
By two actions endurance training increases the ability to uti- training state or dietary condition (Fig. 6).
lize lipid energy sources. First, training raises VO2max , thereby
allowing men and women to perform a given task at a lower
relative intensity (as given by % VO2max ). Second, by increas- Exercise and exercise training effects on muscle
ing mitochondrial mass, endurance training increases the sen- energy substrate partitioning as assessed leg
sitivity of respiratory control by raising cytosolic adenylate respiratory quotient determinations
energy charge (ATP/ADP) and redox status (NADH/NAD+ ), Endurance training raises fat metabolism during exercise by
thus downregulating glycolysis and relieving its inhibitory allowing men and women to perform a given task at a lower %
effects of mitochondrial FFA uptake and oxidation (19, 119). VO2max . For example, in a longitudinal training study on nine
At the whole body level, measurements of pulmonary O2 and men, by means of femoral arterial and venous catheterization
CO2 exchange allow computation of the whole body RER and limb blood flow measurements we determined glycerol,
(=VCO2 /VO2 ), which in turn allows assessment of the rel- FFA, glucose, and lactate exchanges across the working limb.
ative contributions of CHO and lipid to the total fuel mix Femoral and arterial blood samples were taken for determi-
under steady-state conditions. To assess effects of exercise nations of CO2 and O2 contents so that limb RQ, and the
intensity, training, and recent nutrition on energy substrate intramuscular balance of CHO and lipid oxidation could be
partitioning, we (11) (Fig. 6) studied seven untrained men and determined. As well, muscle biopsies were taken before and

Volume 2, January 2012 545

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Bioenergetics of Exercising Humans Comprehensive Physiology

after 1-h exercise bouts to determine net changes in mus- (A) 8 Glucose Rd
cle glycogen and IMTG contents could be compared with
CHO and lipid oxidation rates determined from pulmonary 7
Rest: Pretraining
gas exchange and leg RQ determinations. Men were stud- Rest: Posttraining
6 45%: Pretraining (45UT)
ied during 1 h of cycle ergometry at two intensities before 65%: Pretraining (65UT)

(mg/kg • min)
training (45% and 65% VO2peak ) and after training [65% pre- 5 65% old: Posttraining (ABT)
65% new: Posttraining (RLT)
training VO2peak , same absolute workload (ABT), and 65%
4
posttraining VO2peak , same relative intensity (RLT)]. Super-
vised training involved nine weeks of leg cycle ergometry 3
(1 h/day, 6 days/week @75% VO2peak ). Over the course of
2
a given exercise bout, arterial pH and [lactate] values were
constant (12, 49) suggesting that a-v blood gas values would 1
reflect tissue substrate utilization. Prior to training, during
exercise at 45% VO2peak leg RQ was 0.89±0.05 (Fig. 2). 0
Pre Post Pretraining Posttraining
Otherwise, during the 65% VO2peak trials leg RQ was in the
range of 0.95 to 1.0. These data (Fig. 3) indicate little lipid Rest Exercise
oxidation in the working leg. Very similar results have been (B)
obtained by Odland et al. in Heigenhauser’s group (103). As 7.0
Women y = 2.506 × 100.006x
with results of the cross-sectional study (Fig. 6), results of our 6.5
Men y = 2.449 × 100.006x
longitudinal study indicate that working muscle oxidizes pre- 6.0
dominantly CHO, even after months of endurance training.
Data in Figure 3 are consistent with other results obtained (Rd mg/kg • min) 5.5

in the same study such as the absence of glycerol release 5.0

(Fig. 11), very small net FFA uptake, and no net decrease in 4.5
IMTG content (12, 129). In the aggregate, these results indi- 4.0
cate limited lipid oxidation or IMTG mobilization in working
3.5
skeletal regardless of training state.
3.0
2.5
0 10 20 30 40 50 60 70 80
Effects of exercise and exercise training on (Rest)
% of VO2 max
blood glucose and glycogen use
Plasma glucose use (rate of disposal or disappearance, Rd) is Figure 7 (A) Effect of exercise intensity and training on the plasma
glucose rate of disappearance (Rd). Values are means ±SE of last
significant in resting postprandial individuals, but glycemia 15 and 30 min for rest and exercise, respectively, for 17 women.
is maintained because hepatic glucose production (rate of (B) Relationship between glucose rate of disappearance (Rd) exercise
appearance, Ra) is equivalent. Depending on time since last intensity as given by %VO2max in 19 men and 17 women, before and
after 10 to 12 weeks of endurance training. Note the exponential rise in
eating, glucose is a major source of the carbohydrate oxidized glucose use as a function of exercise power output. Values are means
in resting individuals. As seen Figure 7A (48), endurance ± SE;  significantly different from rest, P < 0.05. *Significantly different
training has no significant effect of glucose Rd (or Ra) in from 45UT (untrained), P < 0.05. From Friedlander et al. (48) and used
with permission.
resting individuals. However, both before and after training,
glucose disposal rises as an exponential function of REL as
given by %VO2max (Fig. 7B) (48).
Not only does glucose flux rise during exercise, but also In normal sized adults, with a blood volume approximat-
endurance training has significant effects on glucose use, as ing 5 L and a blood [glucose] of 100 mg/dL, the circulat-
measured in Rd and reflected also in Ra. Seen in the com- ing blood glucose pool amounts to a diminutive 5 grams, or
parison between second and third exercise histogram bars roughly 21 kcal. Consequently, while exercise significantly
(Fig. 7A), 10 to 12 weeks of endurance training decreases glu- raises blood glucose use (Fig. 7A and B), whether at rest or
cose flux for a given absolute exercise power output (ABT), during exercise, glucose provides only a small portion of the
in this case that which elicited 65% of VO2max before train- carbohydrate energy used, and even a smaller part of the total
ing (65UT) and 52% of VO2max after training. However, as energy used when contributions of muscle glycogen and lipid
seen by comparing the fourth with other exercise histogram energy sources are considered (Fig. 8) (48). Hence, while glu-
bars, training increases glucose flux for a given REL. After cose use scales exponentially to REL (Figs. 4 and 7B), the
training, the REL task at 65% of posttraining VO2max elicited gain is small compared to general rise in metabolic rate. This
a 43% greater than that during exercise at 65% of VO2max mechanism of a restraint on blood glucose use during exercise
pretraining (65UT) reflecting a significantly greater capacity points to the need to protect glycemia for cerebral metabolism
for hepatic glucose production to maintain glycemia. under all conditions.

546 Volume 2, January 2012

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Comprehensive Physiology Bioenergetics of Exercising Humans

0.20 (A) FFA Rd

# 14
0.18 Glucose + Rest: Pretraining
13
Other CHO Rest: Posttraining
0.16 12 +
45%: Pretraining (45UT)
0.14 Lipid 11 65%: Pretraining (65UT) +
10 65% old: Posttraining (ABT)
(kcal/kg • min)

0.12 65% new: Posttraining (RLT)

(μmol/kg • min)
9
0.10 8
7
0.08
6
0.06 5
4
0.04
3
0.02 2
1
0.00
Pre Post 45UT 65UT ABT RLT 0
Rest Exercise Pre Post Pretraining Posttraining
Rest Exercise
Figure 8 Contributions of energy from different substrate sources (B)
14
during rest and exercise before and after training. SE and statisti-
cal symbols are for total energy expenditure only. Values are for
nine subjects. CHO, carbohydrate.  Significantly different from rest; 12
+
+
*significantly different from 45UT; significantly different from 65UT; +

FFA Rd (μmol/kg • min)

and # significantly different from ABT, P < 0.05. From Friedlander et al. 10
(48) and used with permission.
8 ∗

Effects of exercise and exercise training on the 6

use of plasma free fatty acids and other lipid 4
energy sources
On arising after 8 h of sleep and having eaten 10 to 12 h pre- 2
viously, lipids comprise most of the energy utilized (Fig. 6).
0
Of the lipid energy sources used, free fatty acids derived from Pre Post Pretraining Posttraining
adipose triglycerides, as well as other lipid energy sources Rest Exercise
are used. As seen in Figure 9, in individuals adequately nour-
ished, endurance training has no significant effect on resting Figure 9 (A) Effect of exercise intensity and training on plasma FFA
FFA Rd (or Ra). A similar statement can be made about glyc- Rd. Values are means 6 SE of the last 15 and 30 min for rest and ex-
ercise, respectively; n = 8 young women.  Significantly different from
erol flux (a marker of adipose lipolysis) in resting exercise rest; *significantly different from 45UT; + significantly different from
trained and untrained individuals. However, both before and 65UT; and #significantly different from ABT (P < 0.05). From Fried-
after training, FFA use (Rd) rises during easy to moderate lander et al. (46). (B) Effect of exercise intensity and training on plasma
FFA rate of disappearance (Rd) in 10 young men before and after 10
REL as given by %VO2max (Fig. 9). However, as shown in weeks of supervised endurance training. Values are means ± SE of the
Figures 4 and 9, FFA Rd does not rise exponentially as does last 15 and 30 min for rest and exercise, respectively; n = 9 subjects.

glucose flux, but rather FFA flux and oxidation as well as total *Significantly different from pretraining (45UT); significantly different
+
from rest; significantly different from 65UT; and # significantly differ-
lipid oxidation rates are described as inverted hyperbolas. As ent between resting conditions, P < 0.05. From Friedlander et al. (45)
shown in Figure 9A and B for women (46) and men (45), and used with permission.
respectively, prior to 12 weeks of endurance training FFA
Rd increases significantly during easy, 45% VO2max exercise
(45UT). However, as seen by comparing the first and second by comparing the fourth with other exercise histogram bars,
exercise histogram bars, in untrained individuals FFA Rd is training increases FFA flux and rate of oxidation (Rox, not
less during hard 65% VO2max exercise (65UT) compared to shown) for a given REL. After training, for young women
that during easy to moderate intensity exercise at 45% VO2max the REL task at 65% of posttraining VO2max elicited a 43%
(45UT). greater than that during exercise at 65% of VO2max pretraining
Not only does FFA flux rise during exercise, but also (65UT) reflecting a significantly greater capacity for fatty acid
endurance training has significant effects of FFA mobilization mobilization, circulation, and use.
and use in young women, as measured in Rd and reflected Comparison of the posttraining results for women
also in Ra. Seen in the comparison between second and third (Fig. 9A) and men (Fig. 9B) shows gender dimorphism with
exercise histogram bars (Fig. 9A), in young women 10 to 12 respect to the effect of training on FFA use during hard (65%
weeks of endurance training increases FFA flux for a given VO2max ) exercise. Whereas in women FFA Rd is greatest in the
absolute exercise power output (ABT). In women, as seen REL trial, after training men show suppression of fatty acid

Volume 2, January 2012 547

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Bioenergetics of Exercising Humans Comprehensive Physiology

disposal during the hard (REL) posttraining trial. This result (A) Total lipid metabolism
is significant in several ways. Recalling first that both men 35
and women demonstrate increased FFA Rd and Rox when FFA reesterification
30 Other FFA oxidation ∗
exercising at a given power output (i.e., the ABT vs. 65UT Δ
comparisons) it is to be appreciated that endurance training Plasma FFA oxidation ∗
25 Δ ∗
both increases lipid use (Fig. 9) and decreases dependence on Δ

(μmol/kg • min)
glucose and other CHO-derived fuels (Figs. 7 and 8). Recall 20
also that after training, the power output needed to achieve Δ
65% VO2max is 40% to 50% greater than before training. Seen 15
in this context the rise of FFA Rd and Rox by women during 10
the REL trial is particularly significant as they are better at
resisting downregulation of FFA mobilization and use during 5
hard exercise than are men. This gender difference may be
related to lesser sympathetic drive that subsequently lessens 0
Pre Post 45UT 65UT ABT RLT
glycogen degradation rate (48) in women, and also to other Rest Exercise
factors such as a greater percentage of Type I muscle fibers.
Given complexity of substrate-substrate interactions dur- (B) Oxidative energy sources
ing exercise, in Figure 10 an attempt is made to reconcile data 120
obtained using D2-glucose, [1-13 C]glucose, D5-glycerol, [1- Total CHO Other FFA Plasma FFA
13
C]palmitate tracers to describe the fates of FFAs mobilized 100
during exercise and their relative roles as energy substrates
during exercise. Figure 10A shows that a significant portion Percent contributions
80
of FFAs mobilized during exercise are returned to storage
via reesterification. While significant, the values (approxi- 60
mately 20% during rest and exercise) in young women pale
in comparison to the values in postmenopausal women in 40
whom reesterification accounts for more than 50% of fatty
acids mobilized (74). As well, Figure 10A is instructive in 20
showing that plasma FFA oxidation does not account for all
of total body lipid oxidation. Use of these “other” lipid en- 0
ergy sources including IMTGs, are discussed below. Again, Pre Post 45UT 65UT ABT RLT
Figure 10B is instructive in showing that plasma FFA and Rest Exercise
other lipids provide most of the energy used in postprandial
Figure 10 (A) Contributions of different lipid sources to total lipid
humans. However, when exercise starts, CHO-derived energy metabolism during rest and exercise before and after training. SE and
sources predominate regardles of gender or age (82). statistical symbols are for total lipid metabolism only; n = 8, mean ± SE.
 Significantly different from rest and *significantly different from 45UT
(P < 0.05). (B) Contributions of energy from different substrate sources
during rest and exercise normalized to percent energy expenditure; n
Exercise and exercise training effects on muscle = 8. From Friedlander et al. (46).
IMTG net use
In the same study providing results of working limb RQ
(Fig. 3), IMTG use during exercise was assessed in two VO2peak ) exercise, either before or after training, glycerol net
ways. First, biopsies taken before and after exercise showed release was insignificant (Fig. 11). Moreover, an integrated
no significant net change (12). Admittedly, however, sam- analysis of Lox during exercise, based on pulmonary RER,
pling errors in biopsy studies make interpretation difficult working limb RQ, glycerol release (approximately zero dur-
(130). Fortunately, measurements of blood [glycerol] is more ing exercise), net FFA uptake (very small during exercise),
sensitive, reliable, and consistent than are biopsy measure- and insignificant net IMTG content change, inexorably leads
ments. Most importantly, results of a-v differences in glyc- to the conclusion that CHO, not lipid energy sources sustain
erol concentration and muscle IMTG measurements yielded working human skeletal muscle, regardless of training state.
consistent results. Significance of those results is to be appre-
ciated as it is realized that muscle and adipose tissues lack
glycerol kinase. Hence, glycerol released from the working Effects of exercise and exercise training on
limb would be consistent with IMTG mobilization. Regret- circulating lipoproteins and lipoprotein
tably, however, in no condition studied was significant net exchange across working muscle
glycerol release observed. Rather, net limb glycerol release In a subsequent study to corroborate and extend previ-
was a feature of postabsorptive rest, whereas in all exercise ous results (12), determine the availability of lipoproteins
conditions, whether it is easy (45% VO2peak ) or hard (65% and their triglyceride and cholesterol contents, as fuels for

548 Volume 2, January 2012

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Comprehensive Physiology Bioenergetics of Exercising Humans

[Glycerol)v-a 15 A Net leg TG balance

0.07
10
0.06 45% Pre
65% Pre 5
0.05 65% Old (ABT)
65% New (RLT) 0
0.04

mg • min–1
mmol

0.03 –5

0.02 –10

0.01 –15
Release
0 –20
Uptake
–0.01
–25
UT-rest T-rest 45% Pre 65% Pre ABT RLT
–15 0 5 15 30 45 60
Exercise time (min)
# Exercise mean for 65% pre different from 45% pre and 65% old at P < 0.05. 15 Net leg LDL-C balance

Figure 11 Effect of exercise intensity and training on glycerol 10

B
venous-arterial difference. Values are mean ± SEM for eight subjects. 5
Little net glycerol release occurs from working limb muscle indicative
of insignificant intramuscular triglyceride (IMTG) mobilization during 0

mg • min–1
contractions. Symbols represent moderate intensity exercise before en-
durance training (45% Pre), hard exercise before training (65% Pre), af- –5
ter training the same absolute (ABT) intensity that elicited 65% VO2peak –10
before training (65% Old), and exercise that elicited 65% of the post-
training VO2peak [new, relative hard exercise (RLT)]. From Bergman –15
et al. (12) and used with permission.
–20
–25
UT-rest T-rest 45% Pre 65% Pre ABT RLT
working human muscle. Accordingly, we studied eight ad-
ditional men, during rest and exercise, before and after nine
weeks of endurance training. Previously, based on working 15 Net leg HDL-C balance
limb RQ (Fig. 3) and net glycerol release (Fig. 11), we as- 10
C
sumed insignificant contributions of lipoproteins. Based on
5
recently obtained results (79), the assumption of insignifi-
cant muscle lipoprotein uptake and oxidation during exercise 0
mg • min–1

in earlier studies (vide supra) is justified. Importantly, nine –5

weeks of endurance training favorably affected the postab- –10
sorptive circulating lipoprotein patterns. These “long-term”
–15
results are in contrast to those in Figure 6 which show no
acute effect of exercise, whether the intensity be moderate –20
(45% VO2peak ) or hard (65% VO2peak ), either before (UT) or –25
UT-rest T-rest 45% Pre 65% Pre ABT RLT
after training (T).
Figure 12 Effects of graded exercise and exercise training on rest-
ing and working limb triglyceride (A) and cholesterol LDL-C (B) and
HDL-C (C). Exercise and exercise training effects are physiologically
Studies of energy substrate use during exercise insignificant. From Jacobs et al. (79) and used with permission.
and recovery from exercise using indirect
calorimetry
Investigators have long debated whether the exercise-induced ing total energy expenditure associated with physical activity,
elevation in metabolism should be considered as part of the especially if a goal is to know the nutritive requirements of
energy cost of exercise. Classically, the oxygen consumption physically active individuals (20). To evaluate the hypothesis
during the postexercise (O2 debt) period was determined to that physical exercise, even rigorous and prolonged exercise,
assess the extent of energy derived from nonoxidative energy results in large energy expenditure, but little lipid oxidation,
sources during exercise (70, 95). On the contrary, for reasons we examined healthy men and women during and after two
reviewed elsewhere (51), the period of excess postexercise O2 exercise tasks [89 min @ 45% and 60 min @ 65% of peak rate
consumption (EPOC) has multiple causes, reflecting a general of oxygen consumption (VO2peak )] as well as a time of day-
exercise-induced effect on metabolism, and not quantitatively matched resting control trial. Exercise bouts were matched
equivalent to ATP produced by lactate formation during exer- for energy expenditure, the harder exercise task being shorter
cise. However, the EPOC period possesses utility for assess- in duration (60 min), than the easier, but longer task (89 min).

Volume 2, January 2012 549

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Bioenergetics of Exercising Humans Comprehensive Physiology

12.0

Men-65% VO2peak
10.0

Substrate oxidation (kcal/min)

Lipid oxidation
8.0
CHO oxidation

6.0

4.0

2.0

0.0
–30 –15 0 15 30 45 60 75 90 105 120 135 150 165 180 195 210 225 240
Rest exercise Recovery time (min)

Figure 13 Absolute substrate oxidation rates (mean ± SEM) in men before during
and after 60 min of exercise at 65% VO2peak . Lipid and CHO oxidation rates
shown by solid and crosshatched bars, respectively. Data show dominance of CHO
oxidation during physical activity, and crossover to lipid oxidation during recovery.
Data on women are not shown. From Kuo et al. (88) and used with permission.

RERs and VO2 data were used to calculate lipid and CHO Fat as a fuel in the absence of others
oxidation during recovery. The results showed that men and The preponderance of evidence, on humans (Fig. 4) and other
women switched from predominant CHO oxidation during mammals (Fig. 5) is that when energy demand is high, gly-
exercise to lipid oxidation during recovery (88). The switch colytic flux is high and lipid oxidation is downregulated. Be-
to predominant lipid oxidation was true even during recovery lief in the CHO dependence of athletes is empirically based,
from the harder (65% VO2peak ) trial in which CHO was the and universal with athletes across the endurance spectrum
major fuel during activity, particularly in men (Fig. 13). The seeking CHO nutrition for energy. For example, human sprint
results, subsequently confirmed (63,64), showed the presence athletes depend heavily on glycogenolysis and glycolysis
of a gender paradox wherein women derive a greater percent-
age of energy from lipid oxidation during exercise and men
depend more on lipid, and less on glucose oxidation during FA Ra: Men
recovery. Start End End
exercise 65% 45%

25 Control
45% VO 2peak
65% VO 2peak
Ra (μmol/kg FFM/min)

Studies of energy substrate use during exercise 20

and recovery from exercise using tracers 15
Given results of Kuo et al. (88) indicating a shift to reliance
on lipid oxidation during postexercise recovery, we studied 10
10 young lean men and women before and after 90 min @
45% and 60 min @ 65% VO2peak . To determine exercise and 5
postexercise effects on metabolite fluxes, combinations of D2-
0
glucose, [1-13 C]palmitate, and D5-glycerol to determine glu- 0 100 200 300 400
cose, FFA and glycerol fluxes, respectively. As well, separate Time (min)
time of day resting trials, and trials using [13 C]bicarbonate to
evaluate the extent of tracer retention in CO2 and carbonic acid Figure 14 Plasma FFA rate of appearance (Ra) as determined from
pools control trials were conducted (62). Results of metabolite continuous infusion of [1−13 C]palmitate in 10 men studied at rest and
during and after two exercise intensities (45% VO2peak for 90 min and
tracer studies corrected for [13 C]bicarbonate retention (63,64) 65% VO2peak for 60 min). The same subjects were also studied on a
(Fig. 14) confirm and extend those of Kuo et al. (88) show- nonexercise day (control) to account for diurnal variations. Plasma FFA
ing that exercise causes increased fat mobilization, flux, and Ra rose significantly during exercise, compared to preexercise rest, and
remained elevated in 3 h of recovery whether compared to preexercise
oxidation, and that these exercise-induced effects persist well rest or time of day matched resting control. From Henderson et al. (64)
into recovery. and used with permission.

550 Volume 2, January 2012

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Comprehensive Physiology Bioenergetics of Exercising Humans

(Table 1), and as well, endurance athletes, such as ultra- molecule p38 mitogen-activated protein kinase (MAPK) that
distance runners, bicyclists, and triathletes crave carbohy- phosphorylates and activates PGC-1α.
drate foods both during and after exercise. The bottles carried Realizing that training on low-CHO diets might stimulate
by amateur and professional cyclists contain dilute aqueous muscle adaptations, but that CHO is necessary for good per-
blends of carbohydrates (lactate polymers, glucose, fructose, formance, some investigators, for example, Havemann et al.
sucrose, and glucose polymers) (4,81). Few have tried to make (59) trained cyclists on a high fat diet (68% of energy intake)
a go of endurance cycling with vegetable oil in their bottles. for six days, followed by one day of a high-CHO, glycogen
Despite the habits of successful human athletes, are there loading diet. This treatment had the effect of lowering RER
examples in nature, athletics or human experience demon- (increasing lipid oxidation) during a 100 km bicycle time trial,
strating that fat feeding promotes lipid oxidation that in turn, but performance was unaffected. In contrast, performance in
engenders exercise endurance? In the animal kingdom, ex- a 1 km sprint was negatively affected by the high-fat diet fol-
traordinary access and use of body lipid stores is demonstrated lowed by one day of CHO loading. Other than measuring VO2
by migratory birds (106). Among mammals, extraordinary and RER, there is a dearth of efforts using isotope tracers to
running endurance is demonstrated by sled dogs fed high-fat know the effects of fat feeding followed by a period of rest,
diets (58). In human experience, individuals and populations with and without CHO loading on energy substrate partition-
have adapted to very different diets, ranging from running suc- ing and criterion exercise performance. In this context, what
cesses on grain-based diets used by Kenyan Olympians (104), do the results show?
to dependence on marine mammals and fish by humans native Once again, results of the fat-feeding maneuvers demon-
to northern climates and arctic explorers of European descent strate the dedication of athletes to their sport and the equally
(117). Tales of arctic explorers who live on fatty meats alone remarkable ability of muscles and related organ systems, such
are legendary; the ability of early 20th century English ex- as liver to provide energy substrates for exercise. The human
plorers to toil by sledge hauling and exist on fatty meats when body can adapt to a ketotic diet, and this may be appropri-
necessary borders on the unimaginable (32). ate for an arctic explorer, but short- or long-term benefits of
Knowing metabolic plasticity as well as having read the fat feeding on exercise performance are not apparent. To the
accounts of Lieutenant Frederick Schwatka, Aspley Cherry- contrary, those who have studied the acute effects recommend
Garrard and others, Phinney and colleagues set out to evaluate against high-fat feeding as a means to enhance exercise perfor-
whether members of a professional bicycling team could adapt mance (25, 26, 60). As well, from the standpoint of long-term
to a low CHO, ketotic diet (105). Realizing that adaptation health, high-fat feeding is counterproductive (20).
to a ketotic diet takes several weeks, Phinney studied athletes
before and after four weeks of a high fat, high protein, and
low CHO diet. After the ketotic diet, athletes were able to
maintain capacities for VO2max and submaximal endurance at Steady Rate Ergometry and Problems
62% to 64% VO2max . However, the performance of athletes in Computing the Efficiency of
during hill climbing was diminished (S.D. Phinney, personal Human Locomotion
communication). The latter effect was likely attributable to
decreased muscle glycogen content. Realizations about the integrated functions of muscle energy
More recently Hawley and Burke (60) reviewed results transduction systems and the use of energy substrates to op-
of their efforts and those of others who have explored the erate the apparatus for excitation-contraction coupling lead
possibility of increasing the stress on muscle during training to concepts of parsing the totality of energy conversions into
by high-fat and low-CHO feeding. As they point out, prepa- two components: the phosphorylative coupling and mechan-
ration for competition involves some or all of the following: ical coupling efficiencies. These discrete efficiencies can be
training in the morning after an overnight fast, training in- measured on isolated systems ex vivo (vide supra). Alter-
volving weight loss, twice a day training, and prolonged train- natively, phosphorylative coupling and mechanical coupling
ing sessions without taking nutrition. All of these maneuvers efficiencies can be estimated on intact working individuals
cause athletes to work while becoming glycogen depleted with when one or the other can be determined. Realizations about
glycemia being supported by hepatic gluconeogenesis. Also limitations in our ability to measure components of coupling
as they point out, CHO-deprivation significantly decreases efficiencies during nonsteady states has placed emphasis on
power output in training, but the relatively greater training making determinations during steady-rate conditions and then
stress may increase factors that stimulate muscle GLUT4 employ mechanical maneuvers and computational modalities
protein expression and mitochondrial biogenesis. The pur- to devolve the separate coupling efficiencies and their interac-
ported mechanisms are postulated to involve activation of tions. Measurements of the efficiency and economy of human
AMP-activated protein kinase (AMPK), the expression of locomotion suggest maximal efficiencies of 25% to 35% in
transcription factors, specifically nuclear respiratory factors- leg cycling (50) and similar, if not slightly higher efficien-
1 and -2 (NRF-1 and NRF-2), and the expression of per- cies for normal gait walking (36). Differences in efficiencies
oxisome proliferator-activated receptor gamma coactivator-1 measured during cycling versus walking may relate to the
alpha (PGC-1α), as well as the nutrient-sensitive signaling ability to transfer the kinetic energy of one step to the next in

Volume 2, January 2012 551

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Bioenergetics of Exercising Humans Comprehensive Physiology

walking whereas there is little kinetic energy transfer within 12

or between limb muscles during leg cycling. Other modes
of transport, such as slow walking (30, 109) and swimming 11
(40, 76) yield lesser values.
10

9
The baseline hurdle
Pulmonary and muscle metabolism rises in response to the 8
challenge of doing muscle work (Fig. 1), but how much of the

Kcal/min
7
rise is attributable to muscle work is a longstanding question.
The assumption that basal and other housekeeping processes 6
continued during exercise led to several approaches to ad-
just (correct by subtraction) for the background energy costs 5
unrelated to muscle work; this is the matter of the so called 40 RPM
“baseline correction” (50, 51, 120). 4
60 RPM
80 RPM
3
100 RPM
Net efficiency
2
One approach to correcting for basal and housekeeping func- 0
tions during exercise is to subtract a constant quantity, such 0 200 400 600 800
as the preexercise resting metabolic rate from the metabolic Kg-m/min
rate observed during exercise. “Net” efficiency is then:
Figure 15 Effect of work rat and speed of movement on a leg cycle
ergometer (mean ± SE) of 12 young males during steady-rate exercise.
Ef (Net) (%) = (100) external power (kcal)/ Caloric values determined from VO2 and RER. The essentially linear
relationship with perhaps a slight exponential rise in caloric output at
exercise VO2 -restingVO2 (kcal). higher work rates dictates either constant or decreasing efficiency. Be-
cause the caloric cost of each exercise power output increases with
increments in pedaling speed, decreasing efficiency with increasing
Using resting metabolic rate as a baseline correction has speed is indicated. Note that the y-intercepts measured during un-
the effect of depressing the calculated efficiency at low power loaded cycling deviate from rest (≈ 1 kcal/min) and significantly from
outputs, but as power output rises subtracting the resting value linearity for all but the slowest cycling cadence making “gross,” “net,”
and “work” efficiency calculations invalid, and indicating use of the
from exercise energy expenditure has less of an effect, and “delta” method of efficiency calculation. From Gaesser and Brooks
the calculated net efficiency rises. However, with reference to (50) and used with permission.
Figure 15, the no (0) work y-intercept differs from the resting
metabolic rate (≈ 1 kcal/min) as well as the “zero power,”
unloaded cycling rate that is shown at the y-intercept.
Delta (), instantaneous and work efficiencies
As opposed to fixed baseline corrections, there are three other
Gross efficiency approaches to calculating muscular efficiency. In common,
Another approach to correcting for basal and housekeep- these approaches attempt to use an exercise baseline to correct
ing functions during exercise is to make no correction for for basal, housekeeping and other metabolic costs extraneous
the metabolic power associated with unloaded exercise, but to delivering muscle power. The first of these is “work effi-
zero is a constant and has similar effects as does the rest- ciency”, in which unloaded movement, such as free wheel cy-
ing metabolic rate subtraction in determining net efficiency. cling is conducted and the values used to correct the metabolic
“Gross” efficiency is then: cost when a load is applied. Although insightful and simple
in concept, even for bicycling in practice it is difficult to mea-
sure unloaded cycling. For example, as seen in the y-intercepts
Ef (Gross) (%) = (100) external power (kcal)/
in Figure 15, frictional loads and eddy currents in cycle er-
exercise VO2 (kcal). gometers cause the y-intercept determined during unloaded
cycling to be off the regression lines for most conditions stud-
Yet again, with reference to Figure 15, the no (0) work y- ied. Hence, for leg cycling because of problems in controlling
intercept differs from zero, and the computed gross efficiency the ergometer at zero power, and also likely because of varia-
rises as exercise power output rises (Fig. 22B, vide infra). This tions in human movement patterns during unloaded cycling, it
apparent rise in exercise efficiency occurs while the slope of has proven difficult to obtain a “zero” work exercise baseline.
rise of energy expenditure is constant or increasing (Fig. 15). And, to this writer’s knowledge, while partial unweighting
Hence, both net and gross efficiency calculations suffer from by suspension has been attempted for walking subjects (56),
the same two artifacts. and while other forms of pneumatic and aquatic suspension

552 Volume 2, January 2012

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Comprehensive Physiology Bioenergetics of Exercising Humans

are used in rehabilitation medicine, physical suspension with 35

counterweights, springs and other means of unweighting have
not been successfully applied to the computation of work ef- 40 RPM
ficiency during walking.
30
Delta () efficiency

Efficiency (%)
The () approach to calculating muscular efficiency utilizes
a floating baseline that varies as the metabolic response to
25
graded exercises changes. In this way efficiencies can be com-
puted all along a power output curve such as seen in Figure 15.
“Delta” efficiency is then: Theoretical delta
Delta
20
Ef () (%) = (100) change in external power (kcal)/
change exercise VO2 (kcal).

200 400 600 800

Or, as shown in Figures 2 and 15, inverse of the regression
of the caloric equivalent of pulmonary VO2 on the caloric Kg-m/min
equivalent of external power output during leg cycling for
Figure 16 The effect of work rate on delta () efficiency (mean
power outputs > 200 kg-m/min. ± SE) for 12 young men pedaling at a slow cadence (40 RPM) on a leg
A variation on  efficiency is “instantaneous efficiency” cycle ergometer. In contrast to other (gross, net, and work) modes of
which is computed as the reciprocal of the first derivative calculation, the data (From Fig. 15) demonstrates decreasing efficiency
with increments in power output. From Gaesser and Brooks (50) and
of the equation describing the relationship between energy used with permission.
expenditure and work rate (36). This approach has the effect
of smoothing the curve, but some information may be lost at
40
ends of the power output curve. Similarities and differences
between delta and instantaneous efficiencies are shown below
(Fig. 22). 100 RPM

35
Human muscular efficiency In cycling
and walking
In a metabolic state when power output is submaximal and
Efficiency (%)

steady such that the rate of oxygen uptake (VO2 ), pulmonary 30

respiratory gas exchange ratio (RER) and blood [lactate] are
constant (steady rate determinations for constant VCO2 /VO2
and [blood] are typically made over periods of 3-10 min),
the “energy cost” of exercise can be calculated from standard 25
tables developed by Zuntz and Schumburg (137) developed
at the end of the 19th century and modified by Lusk early Theoretical delta
in the 20th century (62), or by a theoretical-thermodynamic Delta
approach (50,132). These computations assume that energy is 20
derived from the oxidation of carbohydrate and lipid energy
sources. From the data in Figure 15 an example of the use
of energy expenditure data to compute the efficiency of leg
ergometer cycling at low pedal cadence (40 rpm) is given 200 400 600 800
in Figure 16, where the effect of increasing power output is Kg-m/min
shown to decrease efficiency. The similar computation for
high cadence (100 rpm) cycling (Fig. 17), again shows a Figure 17 The effect of work rate on delta () efficiency (mean ±
SE) for 12 young men pedaling at a rapid cadence (100 RPM) on a leg
decline in computed efficiency as power output increases. cycle ergometer, calculations based on data in Figure 15. As with slow
Efficiency values are slightly higher for cycling at 100 vs cadence pedaling (Fig. 16), the results demonstrate decreasing effi-
ciency with increments in power output when cycling at a fast cadence.
40 rpm, but are in the same range. Note that the range of computed efficiencies (25%-35%) is similar for
Some of the unresolved problems in evaluating the low and high cadence pedaling. From Gaesser and Brooks (50) and
effects of changes in speed and power on the exercise used with permission.
efficiencies shown in Figures 16 and 17 can be seen by close

Volume 2, January 2012 553

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Bioenergetics of Exercising Humans Comprehensive Physiology

inspection of Figure 15 which gives the data upon which those

efficiencies were calculated. The literature contains extensive
commentary on those issues (36,50,51,80,120,132). Beyond
the assumption of no change in basal energy expenditure is 44
that of the changing y-intercept. Specifically, in comparing 42
cycling cadences of 40 and 100 rpm the caloric expenditure V3, vertical at 6.0 Km/hr
40 V2, vertical at 4.5 Km/hr
for high cadence cycling is clearly greater than when the
same external power output is accomplished during cycling 38 V1, vertical at 3.0 Km/hr
at lower cranking rates (Fig. 15). Recognition that energy 36
cost represented by the y-intercepts in Figure 15 likely
includes energy associated with internal work accelerating 34

Delta efficiency (%)

and decelerating the legs and increases in the metabolic costs 32
of ion pumping associated with more frequent, although less
30
forceful contractions. But, who can claim that internal work
that gives rise to increased energy expenditure is not truly 28
“work” as is measured in a Newtonian sense? Therefore, 26
in the example shown (Fig. 17), the mode of computation
24
attempts to account for both the effects of increased power
and greater internal energy cost of cycling at high, 100 rpm 22
by using as the baseline delta in metabolism the energy cost 20
measured during 200 kg-m/min less power and cycling at
20 lower rpm. 18
In aggregate, comparing Figures 15 and 17 illustrates how
investigators need to exercise judgment when interpreting and 0 250 500 750
Kg-m/min
applying steady-rate efficiency calculations. While the com-
puted efficiencies for cycling at 40 and 100 rpm (Figs. 16 Figure 18 Effects of work rate on delta () efficiency (mean ± SE) for
and 17), and the actual metabolic energy cost for cycling at nine young men during treadmill gradient walking at 3.0, 4.5, and 6.0
100 rpm is greater than for cycling at 40 rpm (Fig. 15), no km/h. Results gradient (vertical) work as well as horizontal work against
an impeding force demonstrate decreasing efficiency with increments
competitive or recreational bicyclist would choose pedaling in muscle power output. As with results in [Figs. (16, (17)], computations
at 40 rpm over pedaling at 80 to 100 pm to maintain a given based on caloric values determined from VO2 and RER during steady-
ground speed. This is because the greater force required to rate, submaximal exercise. From Donovan and Brooks (36) and used
with permission.
push the pedals at slow cranking rate requires greater force
and recruitment of fast-fatiguable fibers (55), a decision that
would result in poor, fatigue-limited performance. with increasing exercise power outputs, as it does in leg cycle
Values given in Figure 15 and corresponding efficiency ergometry (Figs. 16 and 17).
calculations (Figs 16 and 17) are for leg cycling, a mode of
laboratory ergometry that is ideal for investigators and which
is amenable to human locomotary patterns. In contrast, walk-
ing is a movement pattern for which humans are evolved,
Classic Compared to
but which presents problems for investigators interested in Theoretical-Thermodynamic
relating the energy cost of movement to external physical Approaches to Estimating Human
work done. For humans walking or running on a horizontal Exercise Efficiency In Vivo
treadmill, no external work is accomplished. Accordingly, to
know external work accomplished, investigators have used The presence of parallel and superimposable
inclined treadmill exercise where work of lifting the body plots lends confidence in the underlying
can be determined (118). As well, to know the work accom- assumptions
plished in forward locomotion, investigators have used wind Figures 16 and 17 show two lines each, one line represent-
tunnel exercise (108), or a horizontal impeding force (92), ing delta efficiency computed by the “classic” method, and
as originally applied to measuring horsepower (86). As well, the other computed by “theoretical-thermodynamic” method.
engineers have calculated the work done in accelerating and As explained above (50, 80, 120, 132), delta efficiency was
decelerating limbs during walking (31, 109, 110). One exam- based on the regression of metabolic energy expenditure (de-
ple of the “vertical” work of treadmill climbing is depicted termined by indirect calorimetry) on external power output.
in Figure 18 (36). In that experiment, treadmill speed and In the calculation of delta efficiency, ergometer power output
gradient were manipulated, and knowing subject mass the ef- is changed and the corresponding changes in oxygen con-
ficiency of vertical work could be estimated. In the example sumption and carbon dioxide production are observed and
shown, the efficiency of gradient climbing (Fig. 18) decreases used to calculate the change in metabolic energy expenditure

554 Volume 2, January 2012

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Comprehensive Physiology Bioenergetics of Exercising Humans

using the tables of Zuntz and Schumburg (137), as modified The calculation of theoretical delta efficiencies as in
by Lusk (93). The change in metabolic energy expenditure is Figures 16 and 17 provide invaluable insight into how to
then compared to the change in external power output, and compare whole body respiratory data with those obtained on
the computation becomes the inverse of the slope of metabolic isolated fragments of the mitochondrial reticulum. Seeing that
power output. As illustrated in Figures 16 and 17, a leg cycle efficiencies computed by classical and theoretical means, and
ergometer can be used to determine the metabolic responses that both methods respond to increases in speed and power
to ordered changes in mechanical power output over a wide output is, in fact, remarkable. In the view of this writer, the
range. Delta efficiency calculations such as those in Figures 16 close correspondence of results is the best evidence that the
and 17 have the advantage of being sensitive to changes in use in theoretical calculations (i.e., P/O for NADH-linked
RER, and, therefore, energy expenditure as external power substrates, P/O = 2 for FADH2-linked substrates, ATP yield
output rises the relative use of CHO oxidation increases. from glucose = 36-38, and GATP = −10.5 kcal/mol) are
The “theoretical delta” efficiencies were based on re- either correct, or very close to being correct.
sults of measurements made using assumptions made ex vivo. The values cited above and used in the calculations pre-
Specifically, the assumptions were that the P/O was 3, that the sented here are taken from Lehninger (91) and other classical
free energy of ATP (G’) for ATP was –11 kcal/mol, and that sources. In fairness, it needs to be said that in the past (71,94),
the ATP yield from glucose oxidation was 38 mol ATP/mol as well as more recently (18) investigators have reported lesser
glucose. Accordingly using data from Lehninger (91), the values, for example, 30 to 32 ATP/glucose and P/Os of 2.5
theoretical-thermodynamic calculations for glucose oxidation for NADH-linked and 1.5 for FADH2-linked substrates, re-
were obtained: efficiency = [(38 ATP/glucose) (−11 kcal/mol spectively. However, substrate-level phosphorylation of ADP
ATP) = −418 kcal captured as ATP]. For, glucose with an en- occurs in glycolysis, and in vivo most glycolytic carbon flux
thalpy (heat of combustion, G’ = −686 kcal/mol glucose), is from glycogen increasing the P/O from glycolysis 1.0/G6P.
the theoretical coupling efficiency of oxidative phosphory- As well, variability in measuring the P/O of various substrates
lation related to glucose oxidation is then (418 kcal/mol ÷ in isolated mitochondria can be understood from the perspec-
686 kcal/mol = 0.61, or) 61%. As shown clearly in the figures, tive of new concepts of the mitochondrial reticulum (85) and
the two efficiencies assume mechanical coupling efficiencies mitochondrial dynamics (39).
of 50% and produce overlapping results. These results ob- In muscle, and every other tissue examined, the respiratory
tained on healthy exercising adult humans are the strongest apparatus is arranged as a network, a mitochondrial reticulum
available to support validity of the P/O and G’ results ob- (85). Moreover, the reticulum is not static, but turns over
tained on systems ex vivo. continuously at rates depending on several factors, mostly
For glucose degradation, there are two segments in related to energy state (39). Accordingly, what we obtain
catabolism, glycolysis and subsequent oxidation of the prod- when preparing mitochondria for examination ex vivo are, in
ucts of glycolysis (lactate anions). Because the individual fact, fragments of the mitochondrial reticulum. With that view,
enthalpies of the two segments are known, the separate effi- it is unrealistic to expect the fragments to work as well as they
ciencies of the two segments can be estimated. For the con- do in vivo. Accordingly, the data from leg cycle ergometry
version of glucose to lactate, G’ = −38 kcal/mol. From the are best assessing mitochondrial efficiency in vivo. Simply,
energy released in glycolysis, the net ATP yield is 2 mol lac- and energy yields of 20 to 32 ATP/glucose and P/Os of 1.5 to
tate/mol glucose, or [(2 ATP) (−11 kcal/mol) = 22 kcal/mol]. 2.5 would not allow classical and theoretical-thermodynamic
The efficiency then = [(−22 kcal/mol) / (−38 kcal/mol) = calculation of human muscle efficiency (Figs. 16 and 17) to
0.58 or] 58%. be as close as they are. As well, because delta efficiency
In the second segment, for the oxidation of lactate G’ = is determined from slopes of changes in metabolic versus
−326 kcal/mol and the ATP yield is 17 mol ATP/mol glu- mechanical power outputs, small changes in P/O and ATP
cose. Accordingly, energy yield captured as ATP = [(17 yield per unit substrate are compensated for in the mode of
ATP/glucose) (−11 kcal/mol glucose = 187 kcal]. Hence, calculation.
for lactate oxidation the coupling efficiency = 187 kcal/
mol ÷ 326 kcal/mol = 0.57, or] 57%. In reality then, phos-
phorylative coupling efficiency for glucose, as well as for the
glycolytic and mitochondrial segments are similarly efficient The Steady VO2 Rate and the
regardless if calculated from traditional indirect calorimetry Assumption of Minimal ATP
or using assumptions derived from data obtained ex vivo. Production via Anaerobic Glycolysis
The small differences obtained are likely attributable to the
assumptions used. For example, the P/O ratio of 3.0 is a the- Steady-state data obtained in the last several minutes of 6 to
oretical maximum for NADH-linked substrates, whereas the 8 min of constant rate submaximal exercise the assumption is
P/O for FADH2-linked substrates is 2 such that a P/O ra- that VO2 and VCO2 measurements accurately represent ATP
tio of 2.83 might be appropriate in vivo, as might a G’ of turnover during the measurement period. This means constant
–10.5 kcal/mol ATP and an ATP yield of 36 mol ATP/mol levels of muscle adenosine triphosphate ATP, PCr, and lac-
glucose. tate. These assumptions about the phosphagens ATP and PCr

Volume 2, January 2012 555

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Bioenergetics of Exercising Humans Comprehensive Physiology

are likely justified for steady-rate, submaximal exercise (77).

This assumption acknowledges that while the ATP and PCr 10
pools constantly turn over; the rates of hydrolysis to support
cell work are matched by the energy captured in oxidative 8
phosphorylation. However, the question arises, do constant

Work (J)
muscle and blood lactate levels mean the absence of signifi- 6
cant “anaerobic” ATP production in glycolysis? The answer
is almost certainly “yes.” 4
Recent knowledge that lactate is formed continuously,
even in resting individuals and that perturbations such as car- 2
bohydrate eating (44) and physical exercise (15, 96, 121, 122)
increase lactate production under fully aerobic conditions 0
0 5 10 15 20
(111). At issue then is whether the elevated lactate production
Cost (J)
during muscle exercise results in significant ATP production
not accounted for in the measured oxygen consumption. Be- Figure 19 Relationship between external work and metabolic cost
cause most (75%-80%) lactate is disposed of within working of human interosseous (hand) muscle in which metabolic cost was
skeletal muscle (15,120), heart (52), and brain (127), and most determined by 31 P-MRS. The least squares linear regression of data
yields a very high, 68% mechanical coupling efficiency (i.e., external
of the remainder converted to glucose (14, 47, 48) is subse- work from ATP hydrolysis) [Work output (J) = 0.68 ± 0.09 ATP cost
quently oxidized, measures of tissue and pulmonary oxygen (J) − 2.2 ± 0.9 (J)]. Assuming a phosphorylative coupling efficiency
uptake account for glycolytically produced ATP. of 50%, overall  efficiency approximates 34%. Used with permission.
From Jubrias et al. (83) and used with permission.

Real-time Measurements of the tosol to mitochondrial redox balance, are accepted by most
Energetics of Working Human Muscle biochemists. However, for physiologists it is certain that, rel-
ative to whole-body exercise, during small muscle exercise
The assessment of working muscle energetics by means of there are surfeits in capacities for cardiac output and mus-
a-v difference and blood flow measurements such as those of cle blood flow during the postischemic, hyperemic period.
Poole et al. (107) and others (49) provide extra muscular mea- More importantly for them is the assumption that the rate
surements of the energetics of contraction. Recognizing those of PCr restoration postischemia equals the ATP flux during
limitations, at the University of Washington Martin Kush- free flow exercise. Important also is that the biomechanics of
merick and colleagues (including Kevin Conley and Sharon small muscle movements in the magnet bore differ from large
Jubrias) and others have developed highly sophisticated tech- muscle exercise. And, perhaps most difficult for the method
niques of phosphorous MRS and muscle ergometry to parse to overcome is the assumption that glycolysis makes lactic
the components of contraction energetics within working hu- acid, the investigators assuming a 1:1 stoichiometry between
man muscle (83, 89). 31 P-MRS provides measurements of net lactate production and H+ accumulation, an assumption
intramuscular phosphate (Pi), PCr, and ATP. From the chem- that has been hotly disputed as the terminal step in glycolysis
ical equilibrium of creatine kinase, the levels of ADP can be catalyzed by lactate dehydrogenase (LDH) consumes protons
determined; and, assuming a value for the buffering capacity and produces lactate− (114). Nonetheless, phosphorus NRS
of muscle and chemical shift (splitting) in the Pi peak, muscle offers a unique means to noninvasively and continuously de-
hydrogen ion (H+ ) concentration is calculated. To go from termine the ATP flux during human exercise.
static measurements of ATP, PCr, and [H+ ] a period of is- One such example of the types of data that can be obtained
chemia is imposed followed by return to free flow conditions is shown in Jubrias et al. (83) (Fig. 19). Knowing the ATP flux
all the while MRS spectra are acquired. Confinement to the and with a reasonable estimate of the free energy of hydrolysis
bore of NMR magnets and the necessity to use small, nonfer- for a working human first interosseous dorsal (FID) hand
rous ergometers obviates measurements on whole-body exer- muscle, the investigators could calculate a value for efficiency
cises such as cycling, walking, running, or swimming. Still, as the ratio of physical work accomplished (from ergometry)
accurate and discrete measurements can be made on small to the metabolic cost (from the ATP) flux. Hence, for the
muscle groups made to contract in a confined space. As is interosseous muscle, the mechanical coupling efficiency was
always the case, the method requires numerous assumptions 68% (Fig. 19). This is a probably the highest value reported for
that the investigators have systematically evaluated. Of those, any mammalian muscle. Then, assuming an efficiency of 60%
assumption about the creatine kinase equilibrium, free energy for phosphorylative coupling, overall efficiency approximates
of ATP hydrolysis, P/O ratio and magnitude and constancy 41%. This value is significantly (∼5%) higher than for human
of muscle buffering capacity are certain. Other assumptions leg cycling (Figs. 16 and 17), but similar to that for slow
such as that the chemical shift in the Pi peak is due to anaer- walking (Fig. 18). Seemingly then, for slow movements, the
obic glycolysis, and glycerol phosphate shuttle explains cy- efficiency of human muscle work can be as high as 40%.

556 Volume 2, January 2012

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Comprehensive Physiology Bioenergetics of Exercising Humans

Human Muscle Energetics in the 7

Nonsteady State 6
Considering the difficulties in estimation human muscle effi-

(mmol ATP kg–1 d.w. s–1)

ciency in the metabolic steady state, the difficulties encoun- 5
tered by those wishing to know the efficiency of short-term,

ATP turnover
high intensity exercise are enormous. Knowing the efficiency 4
Aerobic
of human muscle during intense exertion has been of interest Others
since the early 1920’s when Krogh and Lindhard (87) and Hill, 3
ATP
Lactate release
Long, and Lupton (67-70), and Smith (118) could reliably Lactate accumulation
measure oxygen consumption in exercising humans. In par- 2
CP
ticular, Smith was focused on steady-state human energetics,
but others recognized the need to estimate energy costs from
1
nonsteady-state exercise. For those purposes, the seminal in-
vestigators devised “O2 deficit” and “O2 debt” methodologies.
0
Regrettably, for numerous reasons reviewed elsewhere (51), Ex1 Ex2
those indirect measures cannot reliably be used to estimate
the nonoxidative energy turnover during exercise. How then Figure 20 Rate of muscle ATP turnover (mmol ATP/kg dry wt/s) dur-
to estimate the energetics of short-term, high power output ing 0-5, 5-15, and 15-180 s of two bouts of intense knee extensor
exercise (EX1) and (EX2) separated by 3 min of rest. ATP turnover esti-
exercise? mated as the sum of muscle anaerobic energy production determined
Using an impressive array of technologies including direct as energy release related to utilization of CP (hatched part of bar),
and indirect calorimetry, muscle biopsies, and measurements net lactate production determined as the sum of accumulation in mus-
cle (open bar) and release to the blood (horizontally lined bar), net
of tissue metabolite exchange, Bangsbo and colleagues (5) ATP utilization (vertically lined bar), others sources, and aerobic energy
reported ATP flux rates on men during 3-min intense bouts of production (filled bar), determined from muscle oxygen uptake and es-
leg kicking exercise. As explained by them, leg kicking was timated utilization of oxygen from myoglobin. Values are means ± SE.
Modified from Figure 4 in Bangsbo et al. (5) and used with permission.
chosen to emphasize quadriceps exercise, a muscle whose
mass could be estimated, that could be biopsied and whose
metabolite content (product of concentration and blood flow) in Figures 12-16, but the investigators undertook a unique
could be measured. The rate of muscle oxygen consumption computation that was to estimate efficiency from the ratio
plus an estimation of 50% of myoglobin desaturation gave a of physical work accomplished to the heat liberated during
value for “aerobic metabolism.” Anaerobic metabolism was exercise. Symbols in the figure indicate different efficiencies
estimated from the decrements of ATP and PCr, muscle lac- at the beginning and end of the 3-min trial, but no differences
tate accumulation and net release. Oxidative disposal of lac-
tate was covered in the limb VO2 measurement. Moreover, 70
they investigated the potential effects of energy provided by
“other” energy sources such as pyruvate and alanine release 60
and accumulation, lactate uptake by inactive tissues in the
Mechanical efficiency (%)

limb. Quantitative estimates of myokinase activity were as- 50

sumed to be minor and not attempted. Their summed estimates #
of ATP use in two 2-min exercise bouts separated by 6 min 40
(Fig. 20) show the predominant role of oxidative metabolism #
beyond 15 s of the 3-min effort; estimates for periods <15 30
s are uncertain and not shown. PCr pool depletion provides
approximately 6% of the energy derived. Estimates for the 20
role of lactate in energy production from net release and ac-
cumulation show a 25% contribution. The nonsteady state and 10
absence of a lactate tracer precludes an estimate of the role of
lactate disposal via oxidation, that quantity being included in 0
the estimate of aerobic metabolism. 15 – 180
Given that they could estimate the ATP flux and because
they knew the external power output the investigators could Figure 21 Muscle mechanical efficiency, determined as work per
total energy production during the interval between 15-180 s of an
calculate a “mechanical efficiency” as the ratio of external intense knee extensor exercise, in which total energy production was
power output to energy (ATP) expenditure; that efficiency determined from metabolic measurements (open bars) and as the sum
computation (Fig. 21) corresponded to “gross efficiency” as of total heat production and work performed. Values are means ± SE.
#significantly (P < 0.05) different from values determined during the
defined by us (50) and others (80). The computed values first 15 s of exercise. Values are means ± SE. Modified from Figure 6
are not comparable to the delta efficiency values as reported in Bangsbo et al. (5) and used with permission.

Volume 2, January 2012 557

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Bioenergetics of Exercising Humans Comprehensive Physiology

between efficiencies determined by direct calorimetry and at the higher (140 W) exercise power output studied. Assum-
ATP turnover. ing that the power output studied (50 kJ/min, 12 kcal/min)
represented steady-state submaximal exercise for both sets of
athletes, the results support the notion that fast-twitch fibers
Muscle Fiber Type and Energetics may be as efficient, or perhaps more efficient than slow-twitch
fibers during high power output exercises (33).
Thermopile studies on isolated rat soleus, comprised mostly
of slow-red fibers (53), and extensor digitorum longus, com-
prised of mostly of fast-white fibers (131) provided support
for the idea that slow and fast muscles are evolved for different Economy of Movement and O2 Cost
purposes, with slow fibers adapted for maintaining tension, as of Transport
in postural control, whereas fast muscles are adapted for rapid
limb movement (33, 50). In this context, it is notable that some Parameters such as VO2max , %VO2max , % efficiency, RER,
investigators (3) argue that fast fibers are adapted to function and glucose and FFA flux rates are very important labora-
at higher speeds and greater power outputs, and are perhaps tory measures related to the energetics of human locomotion.
more efficient working under those circumstances than are However, our present capacities evolved without appreciation
slow-twitch fibers. These data on ex vivo systems appear to of those concepts or the ability to measure them. Further,
be relevant to the human condition. while related to performance, they are nothing more than de-
Regardless of the data from isolated frog and mammalian pendent variables measured during exercise. For this reason,
muscles, studied in thermopiles and other devices at low tem- some comparative physiologists (30, 125) and human physi-
peratures ex vivo, and the disagreements around methodology ologists (34) are particularly interested in understanding de-
and interpretation of such data (3), it is of interest to know terminants of the “O2 cost of transport” and the “economy of
the applications to understanding the bioenergetics of freely movement”; as measures of the metabolic costs of movement
moving humans. It is for certain that there exists heterogene- these parameters are defined as VO2 /distance and VO2 /speed.
ity of muscle fiber type in human populations (55), but little For human runners, those with high levels of VO2max may
attention has been paid to the effects of muscle fiber type on not be the most successful performers, especially if the VO2 to
muscle energetics. One of the few attempts was that of Stuart run at a given speed is high, or in other words, economy is low
and colleagues (123) who conducted leg cycle ergometry on (41). Seen in this context, a high VO2 for an athlete running
track and field sprinter and long-distance runner teammates at a given speed represents a physiological strain, shifting en-
at a major US university. These sprinters and endurance ath- ergy substrate partitioning to greater reliance on CHO-derived
letes were selected to represent individuals possessing fast- fuels (Fig. 4). Because as of yet the existing evidence is that
and slow-twitch fiber types, respectively, and leg cycling was elite runners are neither extraordinarily gifted with regard to
used for experimental purposes as well as to provide a neutral either cardiovascular capacity or muscle biochemistry (115),
exercise modality that neither group had trained extensively and because there is evidence of greater running economy in
on. An assumption inherent in using runners to study leg cy- elite runners (76), at present it is reasonable to suspect that
cling economy and efficiency is that both groups were equally superior running economy has its basis in biomechanics.
naı̈ve to cycling. That assumption seems justified as in a pre- In terms of the energetics of performance, it is clear that
vious study on highly experienced compared to recreationally a greater economy of locomotion translates to a lesser cost
experienced college males produced superimposable curves of transport, with biomechanics as determinate. For instance,
of the regression of metabolic rate on external power output professional bicyclists can maintain speeds almost three times
(101). In that context the results of Stuart et al. (Fig. 19), sup- greater and for three times longer than Olympic marathon run-
port studies on isolated fibers indicating greater efficiency of ners. As well, Olympic 100 m sprint runners are almost five
muscle contraction for individuals expressing predominantly times faster than are Olympic freestyle sprinters, but 500 m
slow-twitch muscle fiber types. In Figure 22A (their Fig. 3) speed skaters are almost twice as fast as 400 m sprint run-
the energy cost of performing identical exercise power out- ners. With assumptions of similar cardiovascular, metabolic,
puts is higher in sprinters than in distance runners. Greater and muscle characteristics, in terms of performance, matters
energy cost, and hence lesser efficiency of the fast-twitch related to biomechanics become supremely important.
predominant sprinters is suggested. Accordingly, whether by
gross efficiency (Fig. 19B, their Fig. 2), or delta (Fig. 19C,
their Fig. 1) efficiency computation, at low to moderate leg Summary and Conclusion
cycling power outputs distance runners were more efficient.
As well, Stuart et al. (115) also calculated instantaneous effi- Human muscles, limbs and supporting ventilatory, cardiovas-
ciency over the range of observed energy expenditures (34). cular, and metabolic systems are well adapted for walking,
That computation (Fig. 22D, their Fig. 4) had the effect of and there is reasonable transfer of efficiency of movement
smoothing the delta work efficiency curves (Fig. 19C). The to bicycling. Our efficiency and economy of movement by
results are similar to the extent of showing similar efficiencies walking (≈ 30%) are far superior to those of apes. This

558 Volume 2, January 2012

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Comprehensive Physiology Bioenergetics of Exercising Humans

(A) (B)

22.5
50
Distance runners
Energy expenditusre (kJ • min –1)

Sprinters 20.0
40

Gross efficiency (%)

17.5
30

15.0
20
12.5
10
10.0

0 0
0 20 40 60 80 100 120 140 160 0 40 60 80 100 120 140 160
Work rate (watts) Work rate (watts)

(C) (D)
40
40

35
Instantaneous efficiency (%)

35
Delta efficiency (%)

30

30
25

25
20

15 20
0 0
0 40 60 80 100 120 140 0 40 60 80 100 120 140
Work rate (watts) Work rate (watts)

Figure 22 A compilation of results of Stuart et al. (123) who studied leg ergometer cycling efficiency of sprinters (fast twitch) and
distance runners (slow twitch) athletes. (A) Higher rates of energy expenditure in sprinters exercising at given exercise power outputs
indicate lesser efficiency. (B) Higher metabolic costs of exercise in sprinters makes for a lesser computed “gross” efficiency. (C, D) For both
groups, delta () and instantaneous calculations show decreasing exercise efficiencies as exercise power outputs increase. Interestingly,
while greater slopes of caloric expenditure regressed on power output make computed  and instantaneous efficiencies less at lower
exercise power outputs, the computed efficiencies converge, or cross over at higher power outputs. Used with permission.

overall body efficiency during walking and bicycling rep- mammals, but inferior to highly adapted species such as
resents the multiplicative interaction of a phosphorylative wolves and migratory birds. Our ancestral ability for hunt-
coupling efficiency of ≈ 60%, and a mechanical coupling ing and gathering depends on strategy and capabilities in the
efficiency of ≈ 50%. These coupling efficiencies compare areas of thermoregulation, and metabolic plasticity. Clearly,
well with those of other species adapted for locomotion. our competitive advantage of survival in the biosphere de-
We are capable runners, but our speed and power are in- pends in intelligence and behavior. Today, those abilities that
ferior to carnivorous and omnivorous terrestrial mammalian served early hunter-gatherers make for interesting athletic
quadrupeds because of biomechanical and physiological con- competitions due to wide variations in human phenotypes. In
straints. But, because of our metabolic plasticity (i.e., the abil- contemporary society, the stresses of regular physical exercise
ity to switch among CHO- and lipid-derived energy sources), serve to minimize morbidities and mortality associated with
our endurance capacity is very good by comparison to most physical inactivity, overnutrition, and aging.

Volume 2, January 2012 559

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Bioenergetics of Exercising Humans Comprehensive Physiology

Acknowledgements 23. Brooks GA, Mercier J. The balance of carbohydrate and lipid utilization
during exercise: The ‘crossover’ Concept, Brief Review. J Appl Physiol
76: 2253-2261, 1994.
The author thanks the following for reading and commenting 24. Brooks GA, Trimmer JK. Literature supports the Crossover Concept.
on the article: Chi-An W. Emhoff, Sean Gross, Rajaa Hussien, (Letter to the Editor). J Appl Physiol 80: 1073-1074, 1995.
25. Burke LM, Kiens B. “Fat adaptation” for athletic performance: The nail
and Matthew Johnson. in the coffin? J Appl Physiol 100: 7-8, 2006.
26. Burke LM, Kiens B, Ivy JL. Carbohydrates and fat for training and
recovery. J Sports Sci 22: 15-30, 2004.
27. Byrne NM, Hills AP, Hunter GR, Weinsier RL, Schutz Y. Metabolic
References equivalent: One size does not fit all. J Appl Physiol 99: 1112-1119.
28. Cathcart EP, Burnett WA. Influence of muscle work on metabolism of
varying conditions of diet. Proc Roy Soc (Biol) 99: 405-426, 1926.
1. Atwater WO, Benedict FG. A Respiration Calorimeter with Appliances 29. Casaburi R, Whipp BJ, Wasserman K, Koyal SN. Ventilatory and gas
for the Direct Determination of Oxygen. Washington, D.C.: Carnegie exchange responses to cycling with sinusoidally varying pedal rate. J
Institute of Washington, 1905. Appl Physiol 44: 97-103, 1978.
2. Atwater WO, Rosa EB. Description of a New Respiration Calorimeter 30. Cavagna GA, Heglund NC, Taylor CR. Mechanical work in terrestrial
and Experiments on the Conservation of Energy in the Human Body, locomotion: Two basic mechanisms for minimizing energy expenditure.
Bulletin 63 Washington, D.C.: U.S. Department of Agriculture, Office Am J Physiol 233: R243-R261, 1977.
of Experiment Stations, Government Printing Office, 1899. 31. Cavagna GA, Kaneko M. Mechanical work and efficiency in level
3. Awan MZ, Frearson N, Goldspink G, Waterson SE. Biochemical ef- walking and running. J Physiol 268: 467-481, 1977.
ficiency of smooth muscle and different types of striated muscle. J 32. Cherry-Girard, A. The Worst Journey in the World: Antarctic 1910-
Mechanochem Cell Motil 1: 225-320, 1972. 1913. Geo. H. Doran Publ., 1922.
4. Azevedo JL, Tietz E, Two-Feathers T, Paull J, Chapman K. Lactate, 33. Coetzer P, Noakes TD, Sanders B, Lambert MI, Bosch AN, Wiggins
fructose and glucose oxidation profiles in sports drinks and the effect T, Dennis SC. Superior fatigue resistance of elite black South African
on exercise performance. PLoS One 2(9): e927, 2007. distance runners. J Appl Physiol 75: 1822-1827, 1993.
5. Bangsbo J, Krustrup P, González-Alonso J, Saltin B. ATP production 34. Daniels J, Daniels N. Running economy of elite male and elite female
and efficiency of human skeletal muscle during intense exercise: Effect runners. Med Sci Sports Exerc 24: 483-489, 1992.
of previous exercise. Am J Physiol Endocrinol Metab 280: E956-E964, 35. Davies KJA, Packer L, Brooks GA. Biochemical adaptation of mito-
2001. chondria, muscle, and whole animal respiration to endurance training.
6. Bárány M. ATPase activity of myosin correlated with speed of muscle Arch Biochem Biophys 209: 539-559, 1981.
shortening. J Gen Physiol 50: 197-218, 1967. 36. Donovan CM, Brooks GA. Muscular efficiency during steady-rate ex-
7. Bárány M, Close RI. The transformation of myosin in cross-innervated ercise II: Effects of walking speed and work rate. J Appl Physiol 43:
rat muscles. J Physiol 213: 455-474, 1971. 431-439, 1977.
8. Barnard RJ, Edgerton VR, Furukawa T, Peter JB. Histochemical, bio- 37. Douglas CG. A method for determining the total respiratory exchange
chemical, and contractile properties of red, white, and intermediate in man. J Physiol 42: 17-18, 1911.
fibers. Am J Physiol 220: 410-414, 1971. 38. Edwards RH, Dawson MJ, Wilkie DR, Gordon RE, Shaw D. Clinical
9. Benedict FG, Carpenter TM. Respiration Calorimeters for Studying use of nuclear magnetic resonance in the investigation of myopathy.
the Respiratory Exchange and Energy Transformations of Man, Bul- Lancet 1(8274): 725-731, 1982.
letin 123 Washington, D.C.: U.S. Department of Agriculture, Office of 39. Fan X, Hussien R, Brooks GA. H2 O2 -induced mitochondrial fission
Experiment Stations, Government Printing Office, 1910. in C2 C12 myocytes. Free Radicals in Biology and Medicine 49: 1646-
10. Benedict FG, Smith HM. The Influence of athletic training upon basal 1654, 2010.
metabolism. Proc Natl Acad Sci U S A 1: 102-103, 1915. 40. Faulkner JA. Physiology of swimming. Res Q 137: 41-54, 1966.
11. Bergman BC, Brooks GA. Respiratory gas exchange ratios during 41. Faulkner JA, Daniels JT, Balke B. Effects of training at moderate al-
graded exercise in fed and fasted trained and untrained men. J Appl titude on physical performance capacity. J Appl Physiol 23: 85-89,
Physiol 86: 479-487, 1999. 1967.
12. Bergman BC, Butterfield GE, Wolfel EE, Casazza, Brooks GA. An 42. Fenn WO. A quantitative comparison between the energy liberated
evaluation of exercise and training on muscle lipid metabolism. Am J and the work performed by the isolated sartorius muscle of the frog. J
Physiol 276: E106-E117, 1999. Physiol (London) 58: 175-203, 1923.
13. Bergman BC, Butterfield GE, Wolfel EE, Lopaschuk GD, Casazza 43. Fiehn W, Peter JB. Properties of the fragmented sarcoplasmic reticulum
GA, Horning MA, Brooks GA. Muscle net glucose uptake and glucose from fast twitch and slow twitch muscles. J Clin Invest 50: 570-573,
kinetics after endurance training in men. Am J Physiol 277: E81-E92, 1971.
1999. 44. Foster D. Banting lecture 1984. From glycogen to ketones-and back.
14. Bergman BC, Horning MA, Casazza GA, Wolfel EE, Butterfield GE, Diabetes 33: 1188-1199, 1984.
Brooks GA. Endurance training increases gluconeogenesis during rest 45. Friedlander AL, Casazza GA, Horning MA, Brooks GA. Endurance
and exercise in men. Am J Physiol Endocrinol Metab 278: E244-251, training increases fatty acid turnover, but not fat oxidation, in young
2000. men. J Appl Physiol 86: 2097-2105, 1999.
15. Bergman BC, Wolfel EE, Butterfield GE, Lopaschuk GD, Casazza 46. Friedlander AL, Casazza GA, Horning MA, Budinger TF, Brooks GA.
GA, Horning MA, Brooks GA. Active muscle and whole body lactate Effects of exercise intensity and training on lipid metabolism in young
kinetics after endurance training in men. J Appl Physiol 87: 1684-1696, women. Am J Physiol 275: E853-E863, 1998.
1999. 47. Friedlander AL, Casazza GA, Horning MA, Huie MJ, Brooks GA.
16. Bergström J. Local changes of ATP and phosphorylcreatine in human Training-induced alterations of glucose flux in men. J Appl Physiol 82:
muscle tissue in connection with exercise. In: Chapman CB, editor. 1360-1369, 1997.
Physiology of Muscular Exercise. New York: American Heart Associ- 48. Friedlander AL, Casazza GA, Horning MA, Huie MJ, Piacentini MF,
ation, 1967, pp. 191-196. (Monograph No. 15). Trimmer JK, Brooks GA. Training-induced alterations of carbohydrate
17. Bramble DM, Lieberman DE. Endurance running and the evolution of metabolism in young women: Women respond differently from men. J
Homo. Nature 432: 345-352, 2004. Appl Physiol 85: 1175-1186, 1998.
18. Brand, M. The efficiency and plasticity of mitochondrial energy trans- 49. Friedlander AL, Jacobs KA, Fattor JA, Horning MA, Hagobian TA,
duction. Biochem Soc Trans 33: 897-899, 2005. Bauer TA, Wolfel EE, Brooks GA. Contributions of working muscle to
19. Brooks GA. Master regulator or readout: The wisdom of distributed whole body lipid metabolism vary with exercise intensity and training.
control. Focus on “Pyruvate suppresses PGC1alpha expression and sub- Am J Physiol Endocrinol Metab 292: E107-E116, 2007.
strate utilization despite increased respiratory chain content in C2C12 50. Gaesser GA, Brooks GA. Muscular efficiency during steady-rate ex-
myotubes”. Am J Physiol Cell Physiol 299: C216-C217, 2010. ercise: Effects of speed and work rate. J Appl Physiol 38: 1132-1139,
20. Brooks GA, Butte NF, Rand WM, Caballero BH. Chronicle of the IOM 1975.
physical activity recommendation: How a physical activity recommen- 51. Gaesser GA, Brooks GA. Metabolic bases of excess post-exercise oxy-
dation came to be among dietary recommendations. Am J Clin Nutr 79: gen consumption: A review. Med Sci Sports Exerc 16: 29-43, 1984.
921S-930S, 2004. 52. Gertz EW, Wisneski JA, Stanley WC, Neese RA. Myocardial substrate
21. Brooks GA, Fahey TD. Exercise Physiology: Human Bioenergetics and utilization during exercise in humans. Dual carbon-labeled carbohy-
its Applications. New York: , John Wiley and Sons, 1984. drate isotope experiments. J Clin Invest 82: 2017-2025, 1988.
22. Brooks GA, Gladden LB. Metabolic systems: Non-oxidative (Gly- 53. Gibbs CL, Gibson WR. Energy production of rat soleus muscle. Am J
colytic and phosphagen). In: Tipton CM, editor. Exercise Physiology: Physiol 223: 864-871, 1972.
People and Ideas. American Physiological Society, 2003, pp. 322- 54. Gladden LB. Lactate metabolism: A new paradigm for the third mil-
360. lennium. J Physiol 558: 5-30, 2004.

560 Volume 2, January 2012

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Comprehensive Physiology Bioenergetics of Exercising Humans

55. Gollnick PD, Armstrong RB, Saubert CW, Piehl K, Saltin B. En- 83. Jubrias SA, Vollestad NK, Gronka RK, Kushmerick MJ. Contraction
zyme activity and fiber composition in skeletal muscle of untrained coupling efficiency of human first dorsal interosseous muscle. J Physiol
and trained men. J Appl Physiol 33: 312-319, 1972. 586: 1993-2002, 2008.
56. Griffin TM, Tolani NA, Kram R. Walking in simulated reduced gravity: 84. Karlsson J. Lactate and phosphagen concentrations in working muscle
Mechanical energy fluctuations and exchange. J Appl Physiol 86: 383- in man. Acta Physiol Scand 358: 313-329, 1971.
390, 1999 85. Kirkwood SP, Munn EA, Packer L, Brooks GA. Mitochondrial reticu-
57. Haldane JS. A convenient form of gas analysis apparatus. J Hyg (Lond) lum in limb skeletal muscle. Am J Physiol 251: C395-C402, 1986.
6: 74-76, 1906. 86. Kleiber M. The fire of life: An introduction to animal energetics. New
58. Hammel EP, Kronfeld DS, Ganjam VK, Dunlap HL Jr. Metabolic York: Wiley, 1961, pp. 116-128.
responses to exhaustive exercise in racing sled dogs fed diets contain- 87. Krogh A, Lindhard J. The changes in respiration at the transition from
ing medium, low, or zero carbohydrate. Am J Clin Nutr 30: 409-418, work to rest. J Physiol (Lond) 53: 431-437, 1920.
1977. 88. Kuo CC, Fattor JA, Henderson GC, Brooks GA. Effect of exercise
59. Havemann L, West SJ, Goedecke JH, Macdonald IA, St Clair Gibson intensity on lipid oxidation in fit young adults during exercise recovery.
A, Noakes TD, Lambert EV. Fat adaptation followed by carbohydrate J Appl Physiol 99: 349-356, 2005.
loading compromises high-intensity sprint performance. J Appl Physiol 89. Kushmerick MJ, Conley KE. Energetics of muscle contraction: The
100: 194-202, 2006. whole is less than the sum of its parts. Biochem Soc Trans 30: 227-231,
60. Hawley JA, Burke LM. Carbohydrate availability and training adap- 2002.
tation: Effects on cell metabolism. Exerc Sport Sci Rev 38: 152-160, 90. Lee CP, Gu Q, Xiong Y, Mitchell RA, Ernster L. P/O ratios reassessed:
2010. Mitochondrial P/O ratios consistently exceed 1.5 with succinate and
61. Henderson SA, Black AL, Brooks GA. Effects of training on leucine 2.5 with NAD-linked substrates. FASEB J 10: 345-350, 1996.
turnover and oxidation during exercise. Am J Physiol 249: E137-E144, 91. Lehninger AL. Bioenergetics. Menlo Park, CA: W.A. Benjamin, Inc.,
1985. 1971.
62. Henderson GC, Fattor JA, Horning MA, Faghihnia N, Johnson ML, 92. Lloyd BB, Zacks RM. The mechanical efficiency of treadmill running
Luke-Zeitoun M, Brooks GA. Retention of intravenously infused against a horizontal impeding force. J. Physiol (London) 223: 355-363,
[13 C]bicarbonate is transiently increased during recovery from exer- 1972.
cise of hard but not moderate intensity. J Appl Physiol 103: 1604-1612, 93. Lusk G. Animal calorimetry: Analysis of the oxidation of mixtures of
2006. carbohydrate and fat. J Biol Chem 59: 41-42, 1924.
63. Henderson GC, Fattor JA, Horning MA, Faghihnia N, Johnson ML, 94. Margaria R, Cerretelli P, Mangili F. Balance and kinetics of anaerobic
Luke-Zeitoun M, Brooks GA. More precise postexercise glucoregula- energy release during strenuous exercise in man. J Appl Physiol 19:
tion in women than men. Am J Clin Nutr 87: 1686-1694, 2008. 623-628, 1964.
64. Henderson GC, Fattor JA, Horning MA, Faghihnia N, Johnson ML, 95. Margaria R, Edwards RHT, Dill DB. The possible mechanisms of
Mau TL, Luke-Zeitoun M, Brooks GA. Lipolysis and free fatty acid contracting and paying the oxygen debt and the role of lactic acid in
metabolism during the post-exercise recovery period. J Physiol 584: muscular contraction. Am J Physiol 106: 689-715, 1933.
963-981, 2007. 96. Mazzeo RS, Brooks GA, Schoeller DA, Budinger TF. Disposal of [1-
65. Hill AV. The energy degraded in recovery processes of stimulated 13C]-lactate during rest and exercise. J Appl Physiol 60: 232-241, 1986.
muscles. J Physiol (Lond) 46: 28-80, 1913. 97. McKenzie S, Phillips SM, Carter SL, Lowther S, Gibala MJ, Tarnopol-
66. Hill AV. Myothermic apparatus. Proc R Soc Lon Ser B 103: 117-137, sky MA. Endurance exercise training attenuates leucine oxidation and
1928. BCOAD activation during exercise in humans. Am J Physiol Endocrinol
67. Hill AV, Long CNH, Lupton H. Muscular exercise, lactic acid and the Metab 278: E580-E587, 2000.
supply and utilization of oxygen. Pt I-III. Proc Royal Soc B 96: 438-475, 98. Mech LD, Boitani L. Wolves: Behavior, ecology and conservation.
1924a. University of Chicago Press, 2003.
68. Hill AV, Long CNH, Lupton H. Muscular exercise, lactic acid and 99. Millward DJ, Davies CT, Halliday D, Wolman SL, Matthews D, Rennie
the supply and utilization of oxygen. Pt IV-VI. Proc Royal Soc B 97: M. Effect of exercise on protein metabolism in humans as explored with
84-138, 1924b. stable isotopes. Fed Proc 41: 2686-2691, 1982.
69. Hill AV, Long CNH, Lupton H. Muscular exercise, lactic acid and the 100. Mommaerts WF. Energetics of muscular contraction. Physiol Rev 49:
supply and utilization of oxygen. Pt VII-VIII. Proc Royal Soc B 97: 427-508, 1969.
155-176, 1924c. 101. Nickleberry B, Brooks GA. No effect of cycling experience on leg cycle
70. Hill AV, Lupton H. Muscular exercise, lactic acid and the supply and ergometer efficiency. Med Sci Sports Exerc 28: 1396-1401, 1996
utilization of oxygen. Quar J Med 16: 135-171, 1923. 102. Odessey R, Goldberg AL. Oxidation of leucine by rat skeletal muscle.
71. Hinkle PC, Kumar MA, Resetar A, Harris DL. Mechanistic stoichiome- Am J Physiol 223: 1376-1383, 1972.
try of mitochondrial oxidative phosphorylation. Biochemistry 30: 3576- 103. Odland LM, Heigenhauser GJ, Wong D, Hollidge-Horvat MG, Spriet
3582, 1991 LL. Effects of increased fat availability on fat-carbohydrate interaction
72. Hochachka PW. Metabolic biochemistry and the making of a during prolonged exercise in men. Am J Physiol 274: R894-R902, 1998.
mesopelagic mammal. Experientia 48: 570-575, 1992. 104. Onywera VO, Kiplamai FK, Boit MK, Pitsiladis YP. Food and
73. Hochachka PW. Oxygen, homeostasis, and metabolic regulation. Adv macronutrient intake of elite Kenyan distance runners. Int J Sport Nutr
Exp Med Biol 475: 311-335, 2000. Exerc Metab 14: 709-719, 2004.
74. Hochachka PW, Rupert JL, Monge C. Adaptation and conservation of 105. Phinney SD, Bistrian BR, Evans WJ, Gervino E, Blackburn GL. The
physiological systems in the evolution of human hypoxia tolerance. human metabolic response to chronic ketosis without caloric restriction:
Comp Biochem Physiol A Mol Integr Physiol 124: 1-17, 1999. Preservation of submaximal exercise capability with reduced carbohy-
75. Holloszy JO. Biochemical adaptations in muscle. Effects of exercise drate oxidation. Metabolism 32: 769-776, 1983.
on mitochondrial oxygen uptake and respiratory enzyme activity in 106. Piersma T, Reneerkens J, Ramenofsky M. Baseline corticosterone
skeletal muscle. J Biol Chem 242: 2278-2282, 1967. peaks in shorebirds with maximal energy stores for migration: A general
76. Holmér I. Physiology of swimming man. Exerc Sport Sci Rev 7: 87-123, preparatory mechanism for rapid behavioral and metabolic transitions?
1979. Gen Comp Endocrinol 120: 118-126, 2000.
77. Hultman E, Bergström J, Anderson NM. Breakdown and resynthesis 107. Poole DC, Gaesser GA, Hogan MC, Knight DR, Wagner PD. Pul-
of phosphorylcreatine and adenosine triphosphate in connection with monary and leg VO2 during submaximal exercise: Implications for
muscular work in man. Scand J Clin Lab Invest 19: 56-66, 1967. muscular efficiency. J Appl Physiol 72: 805-810, 1992.
78. Hultman E, Spriet LL. Dietary intake prior to and during exercise. 108. Pugh LG. The influence of wind resistance in running and walking and
In: Horton ES, Terjung RL, editors. Exercise, Nutrition and Energy the mechanical efficiency of work against horizontal or vertical forces.
Metabolism. New York: Macmillan, 1988, p. 132-149. J Physiol (London) 213: 255-276, 1971.
79. Jacobs KA, Krauss RM, Fattor JA, Horning MA, Friedlander AL, Bauer 109. Ralston HJ. Energy-speed relation and optimal speed during level walk-
TA, Hagobian TA, Wolfel EE, Brooks GA. Endurance training has ing. Intern. Z. Angew. Physiol 17: 277-283, 1958.
little effect on active muscle fatty acid, lipoprotein, or triglyceride net 110. Ralston HJ, Lukin L. Energy levels of human body segments during
balances. Am J Physiol Endocrinol Metab 29: E656-E665, 2006. level walking. Ergonomics 12: 39-46, 1969.
80. Jéquier E, Flatt J-P. Recent advances in human energetics. NIPS 1: 111. Richardson RS, Noyszewski EA, Leigh JS, Wagner PD. Lactate efflux
112-114, 1986. from exercising human skeletal muscle: Role of intracellular PO2 . J
81. Jeukendrup AE. Carbohydrate and exercise performance: The role of Appl Physiol 85: 627-634, 1998.
multiple transportable carbohydrates. Curr Opin Clin Nutr Metab Care 112. Richter EA, Kiens B, Saltin B, Christensen NJ, Savard G. Skeletal
13: 452-457, 2010. muscle glucose uptake during dynamic exercise in humans: Role of
82. Johnson ML, Zarins Z, Fattor JA, Horning MA, Messonnier L, Lehman muscle mass. Am. J. Physiol 254E555-E561, 1988.
SL, Brooks GA. Twelve weeks of endurance training increases FFA 113. Roberts TJ, Weber JM, Hoppeler H, Weibel ER, Taylor CR. Design
mobilization and reesterification in postmenopausal women. J Appl of the oxygen and substrate pathways. II. Defining the upper limits of
Physiol 109: 1573-1581, 2010. carbohydrate and fat oxidation. J Exp Biol 199: 1651-1658, 1996.

Volume 2, January 2012 561

P1: OTA/XYZ P2: ABC
JWBT335-c110007 JWBT335/Comprehensive Physiology November 1, 2011 7:53 Printer Name: Yet to Come

Bioenergetics of Exercising Humans Comprehensive Physiology

114. Robergs RA, Ghiasvand F, Parker D. Biochemistry of exercise-induced 126. Todd KS, Butterfield GE, Calloway DH. Nitrogen balance in men with
metabolic acidosis. Am J Physiol Regul Integr Comp Physiol 287: R502- adequate and deficient energy intake at three levels of work. J Nutr 114:
R516, 2004. 2107-2118, 1984.
115. Saltin B, Kim CK, Terrados N, Larsen H, Svedenhag J, Rolf CJ. 127. van Hall G, Strømstad M, Rasmussen P, Jans O, Zaar M, Gam C,
Morphology, enzyme activities and buffer capacity in leg muscles of Quistorff B, Secher NH, Nielsen HB. Blood lactate is an important
Kenyan and Scandinavian runners. Scand J Med Sci Sports 5: 222-230, energy source for the human brain. J Cereb Blood Flow Metab 29:
1995. 1121-1129, 2008.
116. Schmidt-Nielsen K. Animal physiology: Adaptation and environment. 128. Venables MC, Achten J, Jeukendrup AE. Determinants of fat oxidation
Cambridge, UK:. Cambridge University Press, 1997. during exercise in healthy men and women: A cross-sectional study. J
117. Schwatka F. The long arctic search. Mystic, CT: , Maine Historical Appl Physiol 98: 160-167, 2005.
Association, Inc., 1965. 129. Wallis GA, Friedlander AL, Jacobs KA, Horning MA, Fattor JA, Wolfel
118. Smith HM. Gaseous exchange and physiological requirements for EE, Lopaschuk GD, Brooks GA. Substantial working muscle glycerol
level walking. Washington, D.C.: Carnegie Institution of Washington, turnover during two-legged cycle ergometry. Am J Physiol Endocrinol
1922. Metab 293: E950-E957, 2007.
119. Spriet LL, Howlett RA, Heigenhauser GJF. An enzymatic approach to 130. Watt MJ, Heigenhauser GJ, Spriet LL. Intramuscular triacylglycerol
lactate production in human skeletal muscle during exercise. Med Sci utilization in human skeletal muscle during exercise: Is there a contro-
Sports Exerc 32: 756-763, 2000. versy? J Appl Physiol 93: 1185-1195, 2002.
120. Stainsby WN, Barclay JK. Exercise metabolism: O2 deficit, steady level 131. Wendt IR, Gibbs CL. Energy production of rat extensor digitorum
O2 uptake and O2 uptake for recovery. Med Sci Sports 2: 177-181, longus muscle. Am J Physiol 224: 1081-1086, 1973.
1970. 132. Whipp BJ, Wasserman K. Efficiency of muscular work. J Appl Physiol
121. Stanley WC, Gertz EW, Wisneski JA, Morris DL, Neese RA, Brooks 26: 644-648, 1969.
GA. Systemic lactate turnover during graded exercise in man. Am J 133. White TP, Brooks GA. [U-14C]-glucose, -alanine, and -leucine oxida-
Physiol 249: E595-E602, 1985. tion in rats at rest and two intensities of running. Am J Physiol 240:
122. Stanley WC, Gertz EW, Wisneski JA, Neese RA, Morris DL, Brooks E155-E165, 1981.
GA. Lactate metabolism in exercising human skeletal muscle: Evidence 134. Willis WT, Jackman MR. Mitochondrial function during heavy exer-
for lactate extraction during net lactate release. J Appl Physiol 60: 1116- cise. Med Sci Sports Exerc 26: 1347-1353, 1994.
1120, 1986. 135. Wolfe RR, Wolfe MH, Nadel ER, Shaw JH. Isotopic determination of
123. Stuart MK, Howley ET, Gladden LB, Cox RH. Efficiency of trained amino acid-urea interactions in exercise in humans. J Appl Physiol 56:
subjects differing in maximal oxygen uptake and type of training. J 221-229, 1984.
Appl Physiol 50: 444-449, 1981. 136. Wolfel EE, Groves BM, Brooks GA, Butterfield GE, Mazzeo RS, Moore
124. Taylor CR, Rountree VJ. Running on two or on four legs: Which LG, Sutton JR, Bender PR, Dahms TE, McCullough RE, Reeves JT et al.
consumes more energy? Science 179: 186-187, 1973. Oxygen transport during steady-state submaximal exercise in chronic
125. Taylor CR, Schmidt-Nielsen K, Raab JL. Scaling of energetic cost hypoxia. J Appl Physiol 70: 1129-1136, 1991.
of running to body size in mammals. Am J Physiol 219, 1104-1107, 137. Zuntz N, Schumburg D. Studien Zu Einer Physiologie Des Marches.
1970. Berlin: Verlag von August Hirchwald, 1901.

562 Volume 2, January 2012