Endurance exercise performance: The physiology of champions

Abstract

Efforts to understand human physiology through the study of champion athletes and record performances have been ongoing for about a century. For endurance sports three main factors--maximal oxygen consumption (.VO(2,max)), the so-called 'lactate threshold' and efficiency (i.e. the oxygen cost to generate a given running speed or cycling power output)--appear to play key roles in endurance performance. and lactate threshold interact to determine the 'performance .VO(2)' which is the oxygen consumption that can be sustained for a given period of time. Efficiency interacts with the performance .VO(2) to establish the speed or power that can be generated at this oxygen consumption. This review focuses on what is currently known about how these factors interact, their utility as predictors of elite performance, and areas where there is relatively less information to guide current thinking. In this context, definitive ideas about the physiological determinants of running and cycling efficiency is relatively lacking in comparison with .VO(2,max) and the lactate threshold, and there is surprisingly limited and clear information about the genetic factors that might pre-dispose for elite performance. It should also be cautioned that complex motivational and sociological factors also play important roles in who does or does not become a champion and these factors go far beyond simple physiological explanations. Therefore, the performance of elite athletes is likely to defy the types of easy explanations sought by scientific reductionism and remain an important puzzle for those interested in physiological integration well into the future.

Full text

J Physiol 586.1 (2008) pp 35–44 35
TOPICAL REVIEW
Endurance exercise performance: the physiology
of champions
Michael J. Joyner1and Edward F. Coyle2
1Departments of Anaesthesiology and Physiology, Mayo Clinic College of Medicine Rochester, MN 55905, USA
2Department of Kinesiology and Health Education, University of Texas Austin, Austin, TX 78712, USA
Efforts to understand human physiology through the study of champion athletes and record
performances have been ongoing for about a century. For endurance sports three main factors –
maximal oxygen consumption ( ˙
VO2,max), the so-called ‘lactate threshold’ and efficiency (i.e. the
oxygen cost to generate a give running speed or cycling power output) – appear to play key roles
in endurance performance. ˙
VO2,max and lactate threshold interact to determine the ‘performance
˙
VO2‘ which is the oxygen consumption that can be sustained for a given period of time. Efficiency
interacts with the performance ˙
VO2to establish the speed or power that can be generated at this
oxygen consumption. This review focuses on what is currently known about how these factors
interact, their utility as predictors of elite performance, and areas where there is relatively less
information to guide current thinking. In this context, definitive ideas about the physiological
determinants of running and cycling efficiency is relatively lacking in comparison with ˙
VO2,max
and the lactate threshold, and there is surprisingly limited and clear information about the
genetic factors that might pre-dispose for elite performance. It should also be cautioned that
complex motivational and sociological factors also play important roles in who does or does
not become a champion and these factors go far beyond simple physiological explanations.
Therefore, the performance of elite athletes is likely to defy the types of easy explanations sought
by scientific reductionism and remain an important puzzle for those interested in physiological
integration well into the future.
(Received 24 August 2007; accepted after revision 26 September 2007; first published online 27 September 2007)
Corresponding author M. J. Joyner: Departments of Anaesthesiology and Physiology, Mayo Clinic College of Medicine,
200 First Street SW, Rochester, MN 55905, USA. Email: joyner.michael@mayo.edu
Introduction
Faster, Higher, Stronger: these simple descriptions have
been of interest to humans since the beginning of recorded
history. In this context, integrative physiology has long
been served by so-called ‘experiments in nature.’ These
include asking fundamental questions about the ability of
various animal species to function in harsh environments
and studies on unique human patients with clinical
conditions that offer the opportunity to ask important
questions about physiological regulation (for examples see
Hagberg et al. 1982; Faraci et al. 1984; Schrage et al. 2005).
Along similar lines, studies on both human and animal
performance in athletic events can provide important
insights and raise critical questions about oxygen
transport, muscle performance and metabolism, cardio-
vascular control, and the operation of various components
of the nervous system (Joyner, 1991).
Historical note
One of the first analyses of world records from A. V. Hill
in 1925 (Hill, 1925) related the decline in running speed
as race distance increased to the topic of muscle fatigue
(Fig. 1). Even before then, the Italian physiologist Mosso,
who was interested in fatigue associated with manual
labour, noted ‘It is not will, not the nerves, but it is the
muscle that finds itself worn out after the intense work of
the brain.’ But Mosso hedged his bets and also commented
that ‘fatigue of brain reduces the strength of the muscles’
(DiGiulio et al. 2006).
In Hill’s analysis he speculated that the factors limiting
performance in events of less than a minute and more
than an hour are probably not dependent solely on the
energy supply to the contracting muscles and discussed the
physiological determinants of performance in the context
of ideas about energy stores, oxygen demand and oxygen
C
2008 The Authors. Journal compilation C
2008 The Physiological Society DOI: 10.1113/jphysiol.2007.143834

36 M. J. Joyner and E. F. Coyle J Physiol 586.1
debt. He also speculated that there were ‘three types of
fatigue’ (a concept he found to be inexact) including:
(1) one associated with short violent efforts; (2) the
exhaustion ‘which overcomes the body when an effort
of moderate intensity is continued for a long time’; and
(3) fatigue associated with a more general ‘wear-and-tear’.
The first two types of fatigue were thought to be primarily
‘muscular’.
Additionally, for the second and third types of fatigue,
which occur during endurance exercise, Hill speculated
thatas distancesincrease beyondabout 10miles (∼16 km),
‘The continued fallin the curve,as theeffort isprolonged, is
probably due to the second and third types of fatigue which
we discussed above, either to the exhaustion of the material
of the muscle, or to the incidental disturbances which may
make a man stop before his muscular system has reached
its limit. A man of average size running in a race must
expend about 300 g of glycogen per hour; perhaps a half
of this may be replaced by its equivalent of fat. After a very
few hours therefore the whole glycogen supply of his body
will be exhausted. The body, however, does not readily use
fat alone as a source of energy; disturbances may arise in
the metabolism; it will be necessary to feed a man with
carbohydrate as the effort continues. Such feeding will be
followed by digestion; disturbances of digestion may occur
– other reactions may ensue. For very long distances the
case is far more complex than for the shorter ones, and
although, no doubt, the physiological principles can be
Figure 1. A. V. Hill’s (Hill, 1925) original plot of world record performance time on the X-axis versus
performance speed on the Y-axis
The top tracing is for speed-skating, the middle tracing is for running by males, and the bottom tracing is for
running by women. The shape of the curve led to Hill’s original ideas about differing causes of muscle fatigue for
exercise bouts of different durations.
ascertained, we do not know enough about them yet to be
able further to analyse the curves.’ These comments and
the work of Scandinavian physiologists in the 1930s set the
stagefor theconceptof carbohydrateloading anda number
of dietary and feeding strategies that have been shown to
delay fatigue (Christensen, 1939; Sherman & Costill, 1984;
Murray, 1998).
Focus of this review
In this review for The Journal of Physiology’s 2008
Olympic Issue, we will focus on current models of human
performance, review the physiological ‘ideas’ that led to
these models, and ask what these models explain and more
importantly do not explain. Figure 2 presents our concepts
using a model like Hill’s that is focused on ‘performance
velocity’ and how it is determined by maximum rates
of aerobic energy production, anaerobic capacity and
how efficiently the energy being used is converted to
movement.
In general, we focus on endurance exercise performance
because it is our area of expertise, and there are relatively
more data on the physiological adaptations that contribute
to endurance performance, and (especially for running)
there are accurate records extending for more than
100 years. There is also at least some physiological data
on champion athletes over almost the same period of
time.
C
2008 The Authors. Journal compilation C
2008 The Physiological Society

J Physiol 586.1 Factors that make champions 37
Overview of current ideas about human performance
As noted above, most models of athletic performance focus
on distance running and endurance cycling. First, there
are excellent records and standard events. Second, there
is comprehensive physiological data on a large number
of elite athletes. Third, it is possible, using treadmills and
cycle ergometers, to reasonably simulate in a laboratory
what is happening during actual competition. We should
also note that for the purposes of this review we assume
that environmental conditions are ideal and do not add
anyadditional challengesto physiological regulation(most
notably the challenges associated with high altitude and/or
high environmental temperatures).
˙
VO2,max.Several well-accepted concepts (Joyner, 1991,
1993; Coyle, 1995; Bassett & Howley, 2000) have emerged
related to endurance exercise performance velocity and the
first component issue is the level of aerobic metabolism
that can be maintained during a race (i.e. performance
˙
VO2; Fig. 2). The upper limit for this is ‘maximal’ oxygen
uptake. This is usually achieved during relatively large
muscle mass exercise and represents the integrative ability
of the heart to generate a high cardiac output, total body
haemoglobin, high muscle blood flow and muscle oxygen
extraction, and in some cases the ability of the lungs to
oxygenate the blood (Mitchell et al. 1958; Kanstrup &
Ekblom, 1984; Rowell, 1986; Dempsey, 1986; Saltin &
Strange, 1992; Bassett & Howley, 2000). By the 1930s
very high values for ˙
VO2,max in athletes were observed and
identified as a marker of elite performance (Robinson
et al. 1937). Champion endurance athletes have ˙
VO2,max
values of between 70 and 85 ml kg−1min−1, with values in
women typically averaging about 10% lower due to lower
haemoglobin concentrations and higher levels of body fat
Figure 2. Overall schematic of the multiple
physiological factors that interact as
determinants of performance velocity or
power output
This figure serves as the conceptual framework
for the ideas discussed in this review.
(Saltin & Astrand, 1967; Pollock, 1977; Durstine etal. 1987;
Pate et al. 1987).
In summary, ˙
VO2,max values 50–100% greater than those
seen in normally active healthy young subjects are seen
in champion endurance athletes and the most striking
adaptations to training that contribute to these high
˙
VO2,max values include increased cardiac stroke volume,
increased blood volume, increased capillary density and
mitochondrial density in the trained muscles (Costill et al.
1976). Of these, the most dominant factor is a high stroke
volume (Ekblom & Hermansen, 1968; Coyle et al. 1984;
Martin et al. 1986).
Once it became reasonably clear that elite runners
had high values for ˙
VO2,max it also became clear that for
events lasting beyond 10 or 15 min, most or all of the
competition was performed at an average pace that did not
evoke ˙
VO2,max, with much of the 42 km marathon run at
approximately 75–85% ˙
VO2,max while 10 km is performed
at 90–100% ˙
VO2,max and 5 km at close to ˙
VO2,max (Costill
et al. 1973; Bassett & Howley, 2000). Along these lines, it
has recently been shown that maximal aerobic metabolism
can decline acutely during the course of a 5–8 min
laboratory performance bout. This decline is caused by
a fall in stroke volume and accelerated muscle fatigue
due to reduced blood and oxygen delivery and increased
anaerobic metabolism (Gonzalez-Alonso & Calbet, 2003;
Mortensenet al.2005). This doesnot invalidatethe concept
of ˙
VO2,max, but rather indicates that the maximal rate of
aerobic ATP resynthesis during a race is dynamic and that
truly accurate models of energy turnover during actual
competition would require instantaneous measurements
and calculation of fluxes through multiple metabolic
pathways (e.g. total ATP turnover with contributions from
both aerobic and anaerobic components as well as energy
stores).
C
2008 The Authors. Journal compilation C
2008 The Physiological Society

38 M. J. Joyner and E. F. Coyle J Physiol 586.1
Figure 3. Plot or blood lactic acid concentration
versus race distance (Costill, 1970)
This figure is an example of the diminishing contribution
of so-called ‘anaerobic’ energy sources as race distance
increases. This paper also set the stage for a number of
later investigations related to the fraction of ˙
VO2,max
(e.g. performance ˙
VO2) that could be sustained in
competition.
Lactate threshold. Based on the concepts above the
question then became what fraction of ˙
VO2,max might be
sustained for periods of time extending to several hours
(i.e. the marathon) and what is the rate of glycolysis in
the active muscles at this rate of mitochondrial oxidation.
This question led to observations showing a curvilinear
relationship between blood lactate values during exercise
and the distance of the effort (Fig. 3; Costill, 1970) and
Figure 4. Individual record of treadmill velocity and ˙
VO2versus
blood lactate concentration in subject capable of breaking
2:30 h for the marathon (Farrell et al. 1979)
In untrained subjects the upturn in lactic acid concentrations is seen at
about 60% of ˙
VO2,max. Trained subjects can usually exercise at
75–85% of ˙
VO2,max before there is a marked increase in blood lactate
concentration. This figure also illustrates the concept of performance
˙
VO2and performance velocity.
led to the concept that the rate of aerobic metabolism
maintained during a performance bout (i.e. performance
˙
VO2; Fig. 2) can be better described by the degree of muscle
glycolytic stress reflected in lactate production in addition
to ˙
VO2,max (Farrell et al. 1979; La Fontaine et al. 1981).
In this context, as running speed or power output on
a cycle ergometer increases in untrained subjects there is
typically no sustained rise in blood lactate concentration
until about 60% of ˙
VO2,max is reached. In trained
subjects this value can be 75–90% of ˙
VO2,max (Fig. 4). There
is a long history of investigation about what causes this
rise in blood lactate levels and also how lactate (and/or
hydrogen ion) does or does not contribute to fatigue. For
this review the important summary points include: (1) the
initial appearance of blood lactate is not synonymous with
hypoxia in the skeletal muscle, and (2) the lactate molecule
per se does not ‘cause’ muscle fatigue (Holloszy et al. 1977;
Holloszy & Coyle, 1984; Robergs et al. 2004).
What appears to be occurring is that the maximum
rate of fat oxidation is inadequate to meet the ATP
demands of muscles contracting at moderate and high
intensities. This causes intracellular signalling events to
occur which stimulate glycogenolysis and glycolysis and
ultimatelythe rateofpyruvate deliverytothe mitochondria
progressively exceeds the ability of the mitochondria to
oxidize pyruvate and this leads to accelerated generation
of lactic acid (Holloszy et al. 1977; Holloszy & Coyle, 1984;
Robergs et al. 2004). The associated hydrogen ion is then a
likely culprit in muscle fatigue and also activates group III
and IV skeletal muscle afferents that evoke important
cardiovascular and autonomic reflexes (Pryor et al. 1990).
While the physiological determinants of the lactate
threshold are extremely complex, they are determined
mainly by the oxidative capacity of the skeletal muscle
(Holloszy et al. 1977; Davies et al. 1982; Holloszy & Coyle,
1984; Gregg et al. 1989a,b). This capacity is highly plastic
C
2008 The Authors. Journal compilation C
2008 The Physiological Society

J Physiol 586.1 Factors that make champions 39
and can essentially increase more than twofold in the
trained skeletal muscle of humans or animals who engage
in 20–120 min of training at a requisite intensity (Holloszy
et al. 1977; Dudley et al. 1982; Holloszy & Coyle, 1984).
This more than doubling of oxidative capacity is one of
the factors that is linked to the high ‘lactate threshold’
values seen in elite endurance athletes (Fig. 2). As noted
above, these elite athletes have ˙
VO2,max values that are
50–100% above those seen in normally active sedentary
young people and their lactate threshold occurs at a higher
percentage of their ˙
VO2,max. This means that in elite athletes
the absolute oxygen consumptions (power output and/or
speed) that can be generated for long periods of time
before reaching the lactate threshold is essentially doubled
allowing sustained running speeds of 20 km h−1or cycling
power outputs of 400 W.
Other key factors that reduce muscle fatigability and
lactate production during exercise at 85–90% ˙
VO2,max,
when only a fraction of the total limb muscle mass
is simultaneously recruited, is the quantity of muscle
mass that the athlete can recruit to share in sustaining
power production (Fig. 2). Elite cyclists appear capable
of rotating power production through 20–25% more
muscle mass throughout a 1 h bout of cycling, thus
reducing the relative power production and stress on a
givenfibre(Coyleet al. 1988; Coyle, 1995). Additionally,
this ‘power sharing’ among fibres would also reduce the
glycolytic stress and lactate production per fibre due to
more total mitochondrial sharing for a given rate of
aerobic metabolism. These factors should operate in
a complementary way that reduces the stress per
mitochondria and muscle fibre.
As exercise extends beyond about 2 h the problem
becomes one of fuel availability as (Hill predicted) the
glycogen content in skeletal muscle becomes depleted and
the modest ability of active muscle to take up glucose
from blood (via either the liver or from feeding) can
limit the rate of oxidative ATP generation and thus
the pace that can be sustained. In some (but not all)
subjects the associated reductions in blood glucose evoke
frank symptoms of hypoglycaemia that limit the ability
of the individual to continue exercising (Christensen,
1939; Coyle et al. 1983, 1986). Other highly trained
subjects show remarkable resistance to hypoglycaemia and
for these athletes muscle glycogen depletion is probably
more important. In response to these events, a number
of pre-competition dietary strategies and during-exercise
energy replacement regimens and products have been
developed (Murray, 1998). When these are used in an
optimal manner muscle glycogen stores can be augmented
by 40% before exercise, and hypoglycaemia can be avoided
with the net effect being that the duration of exercise
at about the lactate threshold can be extended by about
one-third (from 2 to 3 h to 4 h) (Coyle et al. 1983, 1986;
Sherman & Costill, 1984).
Performance ˙
VO2and anaerobic metabolism. Without
practical direct calorimetric methods to measure
instantaneous rates of heat and work production during
endurance exercise (Webb et al. 1988; Scott, 2000),
the best practical estimation of the rates of actual
metabolic energy production and ATP turnover is
obtained from measures of oxygen consumption (i.e.
indirect calorimetry) during an endurance performance
bout. During marathon running the relative amount
of anaerobic metabolism is small yet in events lasting
13–30 min (i.e. 5 and 10 km running), it will be significant,
contributing perhaps 10–20% of total ATP turnover. This
anaerobic contribution to ATP turnover during endurance
performance bouts is noted in Fig. 2 and has classically
been estimated from measures of post-exercise oxygen
consumption and may equal the energy provided by
50–80 ml kg−1of oxygen uptake (Fig. 2) (Bangsbo et al.
1993). However, the rate at which this energy might
be generated and consumed is difficult to estimate in a
definitive way.
Figure 2 also makes the point that the rate of total
ATP turnover during endurance performance reflects
the interplay of aerobic and anaerobic metabolism with
lactate generation serving to maintain the NAD+needed
for continued glycolysis and generation of pyruvate. An
example of this interplay appears to be the influence of
high skeletal muscle capillary density, serving to remove or
recycle within muscle fatiguing metabolites (e.g. hydrogen
ions). As shown in Fig. 5, exercise time to fatigue at 88%
˙
VO2,max in a population of cyclists (n=14, individually
numbered) possessing the same ˙
VO2,max (i.e. 4.9 l min−1),
as expected, was related to the percentage of ˙
VO2,max at
the blood lactate threshold. However, some subjects (see
upper line in Fig. 5) were able to exercise longer than
normal (see lower line in Fig. 5) even when accounting
for their lactate threshold (i.e. subjects 1, 2, 7 and 8 in
Fig. 4). For the most part, these individuals (i.e. 1, 2, 7 and
8) possessed an unusually high muscle capillary density
which may have allowed their exercising muscles to better
tolerate anaerobic metabolism and lactic acid production.
For this reason, Fig. 2 indicates that ‘Performance ˙
VO2,
might be directly influenced by muscle capillary density,
independent of its important role in delivering oxygen
and reducing diffusion gradients, but also by removing
waste products and limiting acidosis in the contracting
muscles.
An additional point from Fig. 5 is that much remains
to be learned about subtle factors that delay or accelerate
fatigue during events performed at intensities above
80–90% of ˙
VO2,max. Small increases in total energy
expenditure or reductions in oxygen delivery will have
disproportionate effects and accelerate fatigue (Mortensen
et al. 2005) during very intense exercise. At this time it
remains unclear if laboratory tests can detect the subtle
adaptationsin thevery bestperformers who seemto beable
C
2008 The Authors. Journal compilation C
2008 The Physiological Society

40 M. J. Joyner and E. F. Coyle J Physiol 586.1
manage their metabolism in a way that permits maximum
efficient energy use.
Efficiency. The next factor that makes an important
contribution in endurance exercise performance velocity
has been termed ‘economy’ or ‘efficiency.’ In the above
sections we outlined how ˙
VO2,max and the lactate threshold
operate to determine ‘Performance ˙
VO2, (Fig. 2). The
next question then is how much speed or power can
be generated for that level of oxygen consumption? The
oxygen cost of endurance running (ml kg min−1)atagiven
speed can vary about 30–40% among individuals (Farrell
et al. 1979; Conley & Krahenbuhl, 1980; Joyner, 1991), as
shown in Fig. 6. When cycling at a given power output,
the oxygen cost and thus gross mechanical efficiency also
varies from oneperson toanother,but by asomewhat lesser
amount compared with running (i.e. 20–30%) (Coyle,
1995).
Gross mechanical efficiency when endurance-trained
cyclists generate 300 W can vary from 18.5 to 23.5% and
it appears that more than one-half of this variability is
related to the percentage of type I (slow twitch) muscle
fibres of the vastus lateralis muscle (Coyle et al. 1992).
The efficiency with which the chemical energy of ATP
hydrolysis is converted to physical work depends greatly
on the velocity of sarcomere and muscle fibre shortening.
Type I (slow twitch) fibres display greater mechanical
efficiency when cycling at cadences of 60–120 r.p.m.
Therefore, it is not surprising that elite endurance cyclists
Time to
Fatigue at
88% VO2max
(min) 9
9
8
87
7
2
2
1
1
3
3
4
4
5
5
6
6
11
11
10
10
12
12
13
13
14
14
%VO2max at LT
%VO2max at LT
Time to
Fatigue at
88% VO2max
(min)
75
75
60
60
45
45
30
30
15
15
Subjects with High Capillary Density
Subjects with High Capillary Density
60
60 70
70 80
80 90
90
Figure 5. Time to fatigue during exercise at 88% of ˙
VO2,max
plotted against lactate threshold (LT) in 14 highly trained cyclists
and triathletes (data plotted from Coyle et al. 1988; Coyle, 1995)
These athletes all had similar ˙
VO2,max values and uniformly high muscle
oxidative enzymes. A subgroup of 4 athletes (subjects 1, 2, 7 and 8)
with exceptionally high capillary density seemed to ‘overachieve’ in
comparison with their lactate threshold values compared with other
members of the group.
typically possessa higherpercentage of type Imuscle fibres,
given that they are more efficient. Although type I muscle
fibres in untrained humans possess higher mitochondrial
density compared with type II fibres (fast twitch), it is
important to note that with intense interval training,
mitochondrial activity can be increased to equally high
levels in both fibre types (Chi et al. 1983). Thus, with
intense endurance training over years, the main functional
advantage of type I fibres appears to be efficiency when
cycling rather than total oxidative ability, although type I
fibre seem to retain a greater ability to oxidize fat.
It is also of note that many champion cyclists chose
pedal cadences of around 90 r.p.m. This is a cadence that
may actually increase whole body oxygen consumption
slightly for a given total body power output from the
˙
VO2minimum which usually occurs at 50–60 r.p.m.
In a comprehensive engineering/physiology analysis of
this problem Hansen et al. (2002) noted that subjects
with higher levels of myosin heavy chain I (MHC I, the
predominant myosin in type I fibres) self-selected higher
pedal rates and these rates closely matched the rate of peak
mechanical efficiency. In this context, they speculated that
motor control patterns in these subjects might favour a
faster cadence so that relatively low total muscle forces
(probably from fatigue-resistant motor units) per pedal
stroke could generate the needed power so that the higher
force (and more fatigable) motor units could be conserved.
On a speculative note, with lower force per contraction
there might be less compression of the microcirculation
Figure 6. Regression lines for high, average and low running
economy (efficiency) in elite endurance athletes based on
values gleaned from a number of sources (Joyner, 1991)
Since there has been little systematic data collected above
∼18 km h−1the filled triangles in the figure are individual data from a
limited number of champions with exceptional running economy. This
figure emphasizes the importance of efficiency among groups of elite
performers with relatively uniform ˙
VO2,max and lactate threshold
values. It is also of note that the physiological determinants of
efficiency (especially for running) are poorly understood.
C
2008 The Authors. Journal compilation C
2008 The Physiological Society

J Physiol 586.1 Factors that make champions 41
in the active muscle and better distribution of blood flow
in a way that is consistent with the concepts presented in
Fig. 5.
Running is a more complicated movement than cycling
in that it elicits more stretch on the muscle prior
to contraction and there is more potential to capture
mechanical energy in the elastic elements of tissue.
However, although there has been long-standing interest
in identifying the biomechanical and anatomical factors
that allow one person compared with another to expend
30–40% less energy per kilogram of body to move at
a given velocity, the aetiology of differences in running
economy generally remain a mystery, and biomechanical
descriptions of running are not good predictors of running
economy (Kyr¨
ol¨
ainen et al. 2001; Williams, 2007).
Elite endurance runners typically possess a
predominance of type I muscle fibres and it would
seem logical that they are more mechanically efficient
at the velocities of distance running (Costill et al. 1976;
Fink et al. 1977; Bosco et al. 1987). However, running and
walking economy has not often been highly correlated
with a person’s percentage of type I muscle fibres (Morgan
& Craib, 1992; Hunter et al. 2005). This agrees with the
idea that running economy reflects the interaction of
numerous factors including muscle morphology, elastic
elements and joint mechanics in the efficient transfer of
ATP to running speed.
The extent to which cycling efficiency or running
economy can be improved with training has also been
of long-standing interest. Until recently, it was generally
believed that cycling efficiency and running economy did
not improve much with training (Moseley et al. 2004).
At best, running economy might sometimes increase
slightly over the course of 2 months when explosive-type
weight training is added to an endurance training program
(Paavolainen et al. 1999; Millet et al. 2002). However, the
conclusion that efficiency does not change with training
was based on cross-sectional comparisons of relatively
small numbers of endurance athletes (Moseley et al. 2004).
In this context, there are no comprehensive longitudinal
data on groups of endurance athletes followed over several
years to determine the trainability of cycling efficiency or
running economy. However, there are at least two cases
reporting that running economy can be improved over
years of training in elite athletes (Conley et al. 1984;
Jones, 2006). In fact, the current world record holder
for the women’s marathon displayed a remarkable 14%
improvement in running economy over the course of
5 years of training (Jones, 2006). Furthermore, cycling
efficiency was observed to increase 8% over the course
of 7 years in an elite endurance cyclist (Coyle, 2005). In
general, these case reports suggest that muscular efficiency
and running economy might indeed improve with
continued endurance training at a rate of approximately
1–3% per year. One possible contributing factor is that
at least some of the fast myosin in endurance-trained
muscle shifts to a different and perhaps more efficient iso-
form (Green et al. 1984). Additionally, in some models of
extreme muscle use there can be a complete conversion of
fast twitch to slow twitch muscle fibres, whether this occurs
in elite athletes who train for two or more hours per day
for many years is not known and it is further not known if
such a shift would explain any improvements in efficiency
that might occur with years of training (Pette, 2001).
Integrating current ideas about physiological
limiting factors
The concepts above and in Fig. 2 suggest that ˙
VO2,max and
lactate threshold interact to determine how long a given
rate of aerobic and anaerobic metabolism can be sustained
(i.e. performance ˙
VO2) and efficiency then determines how
much speed or power (i.e. performance velocity) can be
achieved at a give amount of energy consumption. These
relationships were hinted at by Hill in his 1925 paper (Hill,
1925) and were clearly defined in the period between 1970
and the early 1990s (Costill, 1970; Costill etal. 1973; Joyner,
1991; Joyner, 1993; Coyle, 1995; Bassett & Howley, 2000).
The Marathon. In 1991, these concepts were used (Joyner,
1991) to predict that a much faster world record for the
marathon was ‘physiologically’ possible based on the idea
that marathon running speed was essentially predicted by
the equation:
˙
VO2,max ×lactate threshold percentage×running economy
When reasonable estimates of the ‘best’ values ever
recorded for these three parameters were used in this
equation a predicted optimal marathon time of around
1:45 h emerged. Even when assumptions about wind
resistance were added, times well under 2 h seemed
possible. In retrospect, one overlooked possibility was
that the ˙
VO2,max values used in the estimates came
from laboratory studies typically conducted while the
subjects ran up a grade of 5–10% and these values may
be ∼10% higher than those seen during level running
(Morgan et al. 1989). However, even if the highest ˙
VO2,max
values seen during graded running protocols are not
attainable during level running in many people, there
are high enough ˙
VO2,max and lactate threshold values that
might result in sustainable oxygen uptakes, which in
combination with outstanding running economy, would
generate a marked improvement in current world record
time. These comments reinforce the conclusions of this
earlier modelling effort that either there are unknown
factors that operate at high speeds that make such time
‘not’achievable orthat forsome reason‘best inclass’ values
for every factor are unlikely to occur in the same person
(Lucia et al. 2002).
C
2008 The Authors. Journal compilation C
2008 The Physiological Society

42 M. J. Joyner and E. F. Coyle J Physiol 586.1
Some unanswered questions
In the context of the ideas above there are a number
of fundamental unanswered questions. We have already
highlighted questions about the determinants of efficiency
especially for running, and for both running and cycling a
key question is how ‘trainable’ this factor is. Additionally,
we have discussed several factors beyond mitochondrial
content and oxidative enzymes that may permit some
athletes to operate for prolonged periods at especially
high fractions of their ˙
VO2,max. These factors may also
be important in events like long distance cycling and
cross country skiing that occur over varied terrain and
are not conducted at an even physiological pace. In these
competitions there are frequent bursts of more intense
near-maximal activity lasting from a few seconds to a few
minutes that are followed by periods of relative recovery.
A fundamental question is the role genetics plays in the
attainment of world class status and truly elite athletic
performance. There are a number of studies showing that
key elements of the response to training in sedentary
persons is widely variable and has a genetic component
(Rankinen et al. 2006). There have also been reports
suggestingthat commonsingle nucleotidepolymorphisms
might be over represented in either groups of elite
endurance athletes or in sedentary subjects that respond
most to training. The most notable example is the idea
that I (for insertion) variant of the angiotensin converting
enzyme (ACE) gene is over represented among elite end-
urance athletes. However, in the largest cohort of elite
athletes who have been both rigorously phenotyped and
genotyped this association has not been confirmed and
to date there are no genetic markers identified in humans
that have been clearly shown to be more frequent in elite
endurance athletes (Rankinen et al. 2000).
Another interesting example relates to the gene
encoding for the skeletal muscle isoform of AMP
deaminase. There is a common mutation of this gene
that may be associated with lower exercise capacity and
‘trainability’ in untrained subjects (Rico-Sanz et al. 2003).
While the frequency of the gene may be lower in elite
endurance athletes there are still a number of elite
performers who carry it so it does not appear to pre-
clude the attainment of elite status and there is at least
one example of an elite performer with essentially no AMP
deaminase activity (Lucia et al. 2007; Rubio et al. 2005).
In this context, finding genetic markers that are
strongly predictive of either success in endurance athletic
performance or somehow preclude it is likely to be
a daunting task because of the many cultural and
environmental factors that contribute to success in
sport, the many physiological factors that interact as
determinants of performance, and the heroic nature of the
training required. Ideas about culture are highlighted by
the observation that while East African runners currently
dominate international competition previously athletes
from Australia and New Zealand, preceded by Eastern
Europeans and even earlier the Finns showed remarkable
levels of success. This geographical diversity argues against
a simple genetically based set of answers to the problem of
elite performance in endurance competition. In a parallel
way, there is a low signal to noise ratio for many proposed
genetic factors that might contribute to multifactorial
medical conditions like heart disease, diabetes and
hypertension and clear causal associations between
genotype and phenotype are slow to emerge (Morgan et al.
2007).
Concluding remarks
Our concepts about factors that regulate and potentially
limit endurance performance are not a radical departure
from the intuitive logic introduced by Mosso and Hill. Elite
athletic performance involves integration of muscular,
cardiovascular and neurological factors that function
cooperatively to efficiently transfer the energy from
aerobic and anaerobic ATP turnover into velocity and
power. The past four decades of research have described
in great detail the cardiovascular and muscular factors
that govern oxygen delivery to active muscles, oxidative
ATP rephosphorylation and markers of metabolic stress.
However, little advancement seems to have been made in
identifying neurological factors that might alter motor
unit recruitment during prolonged exercise in ways
that limit fatigue. Although it has become increasingly
apparent that muscular efficiency and economy are hugely
important, the physiological determinants of running
economy remain a mystery while myosin type appears
important to cycling efficiency at cadences chosen in
competition.
The outcome of all Olympic endurance events is
decided at intensities above 85% ˙
VO2,max and most require
athletes to be relatively fatigue resistant at intensities that
stimulate significant anaerobic metabolism. At this time,
the literature contains insufficient data that specifically
describe the actual total energy demands of competition,
the amount of muscle that is active during competition,
and the complex neural patterns by which power and
velocity are maintained as fatigue and failure develop in the
nervous, cardiovascular and muscular systems. Such data
are needed both in absolute and temporal terms. In this
context, more work is needed on highly trained athletes
performing very intense exercise in real or simulated
competitions.
References
Bangsbo J, Michalsik L & Petersen A (1993). Accumulated O2
deficit during intense exercise and muscle characteristics of
elite athletes. IntJSportsMed14, 207–213.
C
2008 The Authors. Journal compilation C
2008 The Physiological Society

J Physiol 586.1 Factors that make champions 43
Bassett DR Jr & Howley ET (2000). Limiting factors for
maximum oxygen uptake and determinants of endurance
performance. Med Sci Sports Exerc 32, 70–84.
Bosco C, Montanari G, Ribacchi R, Giovenali P, Latteri F,
Iachelli G, Faina M, Colli R, Dal Monte A & La Rosa M
(1987). Relationship between the efficiency of muscular
work during jumping and the energetics of running.
Eur J Appl Physiol Occup Physiol 56, 138–143.
Chi M, Hintz CS, Coyle EF, Martin WH 3rd, Ivy JL, Nemeth
PM, Holloszy JO & Lowry OH (1983). Effects of detraining
on enzymes of energy metabolism in individual human
muscle fibers. Am J Physiol Cell Physiol 244, C276–C287.
Christensen EH (1939). Untersuchungen uber die
Verbrennungsvorgange bei langdauernder, scwherer
Muskelarbeit. Skand Arch Physiol 81, 152–161.
Conley DK, Burkett LN & Millar AL (1984). Following Steve
Scott: physiological changes accompanying training.
Physician Sports Med 12, 103–106.
Conley DL & Krahenbuhl GS (1980). Running economy and
distance running performance of highly trained athletes.
Med Sci Sports Exercise 12, 357–360.
Costill DL (1970). Metabolic responses during distance
running. J Appl Physiol 28, 251–255.
Costill D, Fink WJ & Pollock ML (1976). Muscle fiber
composition and enzyme activities of elite distance runners.
Med Sci Sports 8, 96–100.
Costill DL, Thomason H & Roberts E (1973). Fractional
utilization of the aerobic capacity during distance running.
Med Sci Sports 5, 248–252.
Coyle EF (1995). Integration of the physiological factors
determining endurance performance ability. Exerc Sport Sci
Rev 23, 25–63.
Coyle EF (2005). Improved muscular efficiency displayed as
Tour de France champion matures. J Appl Physiol 98,
2191–2196.
Coyle EF, Coggan AR, Hemmert MK & Ivy JL (1986). Muscle
glycogen utilization during prolonged strenuous exercise
when fed carbohydrate. J Appl Physiol 61, 165–172.
Coyle EF, Coggan AR, Hopper MK & Walters TJ (1988).
Determinants of endurance in well-trained cyclists. JAppl
Physiol 64, 2622–2630.
Coyle EF, Hagberg JM, Hurley BF, Martin WH, Ehsani AA &
Holloszy JO (1983). Carbohydrate feeding during prolonged
strenuous exercise can delay fatigue. J Appl Physiol 55,
230–235.
Coyle EF, Martin WH 3rd, Sinacore DR, Joyner MJ & Holloszy
JO (1984). Time course of loss of adaptations after stopping
prolonged intense endurance training. J Appl Physiol 57,
1857–1864.
Coyle EF, Sidossis LS, Horowitz JF & Beltz JD (1992). Cycling
efficiency is related to the percentage of type I muscle fibers.
Med Sci Sports Exerc 24, 782–788.
Davies KJA, Maguire JJ, Brooks GA, Dallman PR & Packer L
(1982). Muscle mitochondrial bioenergetics, oxygen supply,
and work capacity during dietary iron deficiency and
repletion. Am J Physiol Endocrinol Metab 242,
E418–E427.
Dempsey JA (1986). Is the lung built for exercise? Med Sci
Sports Exerc 18, 143–155.
DiGiulio C, Daniele F & Tipton CM (2006). Angelo Mosso and
muscular fatigue: 116 years after the first Congress of
Physiologists: IUPS commemoration. AdvPhysiolEduc30,
51–57.
Dudley GA, Abraham WM & Terjung RL (1982). Influence of
exercise intensity and duration on biochemical adaptations
in skeletal muscle. J Appl Physiol 53, 844–850.
Durstine JL, Pate RR, Sparling PB, Wilson GE, Senn MD &
Bartoli WP (1987). Lipid, lipoprotein, and iron status of elite
women distance runners. Int J Sports Med 8, 119–123.
EkblomB&Hermansen L (1968). Cardiac output in athletes.
J Appl Physiol 25, 619–625.
Faraci FM, Kilgore DL Jr & Fedde MR (1984). Oxygen delivery
to the heart and brain during hypoxia: Pekin duck vs
bar-headed goose. Am J Physiol Regul Integr Comp Physiol
247, R69–R75.
Farrell P, Wilmore JH, Coyle EF, Billing JE & Costill DL (1979).
Plasma lactate accumulation and distance running
performance. Med Sci Sports 25, 1091–1097.
Fink WJ, Costill DL & Pollack ML (1977). Submaximal and
maximal working capacity of elite distance runners. Part II.
Muscle fiber composition and enzyme activities. Ann NY
Acad Sci 30, 323–327.
Gonzalez-AlonsoJ&CalbetJA(2003). Reductions in systemic
and skeletal muscle blood flow and oxygen delivery limit
maximal aerobic capacity in humans. Circulation 107,
824–830.
Green HJ, Klug GA, Reichmann H, Wiehrer W & Pette D
(1984). Exercise-induced fibre type transitions with regard to
myosin, parfalbumin, and sarcoplasmic reticulum in muscles
of the rat. Pflugers Arch 400, 432–438.
Gregg SG, Mazzeo RS, Budinger TF & Brooks GA (1989a).
Acute anemia increases lactate production and decreases
clearance during exercise. J Appl Physiol 67, 756–764.
Gregg SG, Willis WT & Brooks GA (1989b). Interactive effects
of anemia and muscle oxidative capacity on exercise
endurance. J Appl Physiol 67, 765–770.
Hagberg JM, Coyle EF, Carroll JE, Miller JM, Martin WH &
Brooke MH (1982). Exercise hyperventilation in patients
with McArdle’s disease. J Appl Physiol 52, 991–994.
Hansen EA, Andersen JL, Nielsen JS & Sjogaard G (2002).
Muscle fibre type, efficiency, and mechanical optima affect
freely chosen pedal rate during cycling. Acta Physiol Scand
176, 185–194.
Hill AV (1925). Athletic records. Lancet 5, 481–486.
Holloszy JO & Coyle EF (1984). Adaptations of skeletal muscle
to endurance exercise and their metabolic consequences.
J Appl Physiol 56, 831–838.
Holloszy JO, Rennie MJ, Hickson RC, Conlee RK & Hagberg
JM (1977). Physiological consequences of the biochemical
adaptations to endurance exercise. Ann NY Acad Sci 301,
441–450.
Hunter GR, Bamman MM, Larson-Meyer DE, Joanisse DR,
McCarthy JP, Blaudeau TE & Newcomer BR (2005). Inverse
relationship between exercise economy and oxidative
capacity in muscle. Eur J Appl Physiol 94, 558–568.
Jones AM (2006). The physiology of the world record holder
for the women’s marathon. Int J Sports Sci Coaching 1,
101–116.
C
2008 The Authors. Journal compilation C
2008 The Physiological Society

44 M. J. Joyner and E. F. Coyle J Physiol 586.1
Joyner MJ (1991). Modeling: optimal marathon performance
on the basis of physiological factors. J Appl Physiol 70,
683–687.
Joyner MJ (1993). Physiologic limiting factors and distance
running: Influence of gender and age on record
performances. ExercSportSciRev21, 103–133.
Kanstrup I-L & Ekblom B (1984). Blood volume and
hemoglobin concentration as determinants of maximal
aerobic power. Med Sci Sports Exerc 16, 256–262.
Kyr ¨
ol¨
ainen H, Belli A & Komi PV (2001). Biomechanical
factors affecting running economy. Med Sci Sports Exerc 33,
1330–1337.
La Fontaine TP, Londeree BR & Spath WK (1981). The
maximal steady state versus selected running events. Med Sci
Sports Exercise 13, 190–193.
Lucia A, Hoyos J, P´
erez M, Santalla A & Chicharro JL (2002).
Inverse relationship between VO2max and
economy/efficiency in world-class cyclists. Med Sci Sports
Exerc 34, 2079–2084.
Lucia A, Martin MA, Esteve-Lanao J, San Juan AF, Rubio JC &
Oliv´
an & Arenas J (2007). C34T mutation of the AMPD1
gene in an elite white runner. Br J Sports Med 40, e7.
Martin W 3rd, Coyle EF, Bloomfield SA & Ehsani AA (1986).
Effects of physical deconditioning after intense endurance
training on left ventricular dimensions and stroke volume. J
Am Coll Cardiol 7, 982–989.
Millet G, Jaouen B, Borrani F & Candau R (2002). Effects of
concurrent endurance and strength training on running
economy and. VO2kinetics. Med Sci Sports Exerc 34,
1351–1359.
Mitchell JH, Sproule BJ & Chapman CB (1958). The
physiological meaning of the maximal oxygen intake test.
J Clin Invest 37, 538–547.
Morgan DW, Baldini FD, Martin PE & Kohrt WM (1989). Ten
kilometer performance and predicted velocity at VO2max
among well-trained male runners. Med Sci Sports Exerc 21,
78–83.
Morgan DW & Craib M (1992). Physiological aspects of
running economy. Med Sci Sports Exerc 24, 456–461.
Morgan TM, Krumholz HM, Lifton RP & Spertus JA (2007).
Nonvalidation of reported genetic risk factors for acute
coronary syndrome in a large-scale replication study. JAMA
297, 1551–1561.
Mortensen SP, Dawson EA, Yoshiga CC, Dalsgaard MK,
Damsgaard R, Secher NH & Gon´
alez-Alonso J (2005).
Limitations to systemic and locomotor limb muscle oxygen
delivery and uptake during maximal exercise in humans.
J Physiol 566, 273–285.
Moseley L, Achten J, Martin JC & Jeukendrup AE (2004). No
differences in cycling efficiency between world-class and
recreational cyclists. IntJSportsMed25, 374–379.
Murray R (1998). Rehydration strategies – balancing substrate,
fluid, and electrolyte provision. Int J Sports Med 19,
S133–S135.
Paavolainen L, H¨
akkinen K, H¨
am¨
al¨
ainen I, Nummela A &
Rusko H (1999). Explosive-strength training improves 5-km
running time by improving running economy and muscle
power. J Appl Physiol 86, 1527–1533.
Pate RR, Sparling PB, Wilson GE, Cureton KJ & Miller BJ
(1987). Cardiorespiratory and metabolic responses to
submaximal and maximal exercise in elite women distance
runners. Int J Sports Med 8, 91–95.
Pette D (2001). Historical perspectives: plasticity of
mammalian skeletal muscle. J Appl Physiol 90, 1119–1124.
Pollock ML (1977). Submaximal and maximal working
capacity of elite distance runners. Part 1. Cardiorespiratory
aspects. Ann NY Acad Sci 301, 310–322.
Pryor SL, Lewis SF, Haller RG, Bertocci LA & Victor RG (1990).
Impairment of sympathetic activation during static exercise
in patients with muscle phosphorylase deficiency (McArdle’s
disease). J Clin Invest 85, 1444–1449.
Rankinen T, Bray MS, Hagberg JM, P´
erusse L, Roth SM,
WolfarthB&BouchardC(2006). The human gene map for
performance and health-related fitness phenotypes: the 2005
update. Med Sci Sports Exerc 38, 1863–1888.
Rankinen T, Wolfarth B, Simoneau JA, Maier-Lenz D,
Rauramaa R, Rivera MA, Boulay MR, Chagnon YC, P´
erusse
L,KeulJ&BouchardC(2000). No association between the
angiotensin-converting enzyme ID polymorphism and elite
endurance athlete status. J Appl Physiol 88, 1571–1575.
Rico-Sanz J, Rankinen T, Joanisse DR, Leon AS, Skinner JS,
Wilmore JH, Rao DC & Bouchard C; HERITGE Family Study
(2003). Associations between cardiorespiratory responses to
exercise and the C34T AMPD1 gene polymorphism in the
HERITAGE Family Study. Physiol Genomics 14, 161–166.
Robergs RA, Ghiasvand F & Parker D (2004). Biochemistry of
exercise-induced metabolic acidosis. Am J Physiol Regul
Integr Comp Physiol 287, R502–R516.
Robinson S, Edward HT & Dill DB (1937). New records in
human power. Science 85, 409–410.
Rowell LB (1986). Human Circulation Regulation During
Physical Stress, pp. 1–416. Oxford University Press, New York.
Rubio JC, Martin MA, Rabad´
an M, G´
omez-Gallego F, San Juan
AF, Chicharro JL, P´
erez M, Arenas J & Lucia A (2005).
Frequency of the C34T mutation of the AMPD1 gene in
world-class endurance athletes: does this mutation impair
performance? J Appl Physiol 8, 2108–2112.
SaltinB&AstrandP-O(1967). Maximal oxygen uptake in
athletes. J Appl Physiol 23, 353–358.
SaltinB&StrangeS(1992). Maximal oxygen uptake: ‘old’ and
‘new’ arguments for a cardiovascular limitation. Med Sci
Sports Exerc 24, 30–37.
Schrage WG, Wilkins BW, Dean VL, Scott JP, Henry NK,
Wylam ME & Joyner MJ (2005). Exercise hyperemia and
vasoconstrictor responses in humans with cystic fibrosis.
J Appl Physiol 99, 1866–1871.
Scott CB (2000). Energy expenditure of heavy to severe exercise
and recovery. J Theor Biol 207, 293–297.
Sherman WM & Costill DL (1984). The marathon: dietary
manipulation to optimize performance. AmJSportsMed12,
44–51.
Webb PW, Saris WH, Schoffelen PF, Van Ingen Schenau GJ &
Ten Hoor F (1988). The work of walking: a calorimetric
study. Med Sci Sports Exerc 20, 331–337.
Williams KR (2007). Biomechanical factors contributing to
marathon race success. Sports Med 37, 420–423.
C
2008 The Authors. Journal compilation C
2008 The Physiological Society