Share this article on:

Exercise Responses and Adaptations in Rowers and Spinal Cord Injury Individuals


Medicine & Science in Sports & Exercise: May 2006 - Volume 38 - Issue 5 - p 958-962
doi: 10.1249/01.mss.0000218131.32162.ce
APPLIED SCIENCES: Physical Fitness and Performance

Elite rowers (ROWERS) and those who have had a spinal cord injury (SCI) are different physically in many realms. Both have physical activity histories that affect their lower-extremity extensor muscles in a dramatically different fashion. ROWERS can sustain a 500-W power output during their 5- to 6-min race. After a complete SCI, a 75-W power output might be achieved during a VO2peak test. Elite SCI wheelchair racers can achieve a higher value that is similar to that of a sedentary able-bodied person. ROWERS can attain a O2 max of more than 7.5 L·min−1 and can tolerate a blood lactate of 30 mmol·L−1. After a complete SCI in which muscles become markedly atrophied, a peak O2 of 2 L·min−1 and a blood lactate of 10 mmol·L−1 might be achieved. ROWERS rely on the 75% slow-twitch fiber composition of their trained thigh muscles to train and race. Such activity modestly increases fiber size and markedly increases mitochondrial content. After a complete SCI, affected muscle fibers markedly atrophy, maintain most of their mitochondrial content, and become fast-twitch. These data suggest remarkable plasticity of physical function to the extreme that a marked increase in energy demanding, rather continuous physical activity can make a muscle more "slow-twitch"; so it will demand less energy when contracted. In contrast, SCI eventually causes muscle to be composed of more fast-twitch fibers. Molecular biologists may explain why fast-twitch fibers, which appear ideal for some athletes because of their high power output, are abundant in muscles that are seldom recruited. Until then, our results indicate that the fiber type composition of muscle in humans is stable unless extreme alterations in physical activity are endured.

1Department of Biomedical Sciences, College of Osteopathic Medicine, Ohio University, Athens, OH; 2The Miami Project to Cure Paralysis, University of Miami, Coral Gables, FL and Center of Excellence in Functional Recovery in Chronic Spinal Cord Injury, Miami VA Medical Center, Miami, FL; 3Department of Kinesiology, The University of Georgia, Athens, GA; and 4Crawford Research Institute, Shepherd Center, Atlanta, GA

Address for correspondence: Gary A Dudley, Ph D, Department of Kinesiology, 115 M Ramsey Center, University of Georgia, 330 River Rd, Athens, GA 30602; E-mail:

Submitted for publication March 2005.

Accepted for publication November 2005.

The interest of this brief review is to contrast the physiological adaptations to exercise of international-caliber rowers (ROWER) with those of the non-physically active, paraplegic spinal cord injury (SCI) individuals to highlight the plasticity of human exercise ability. Most of this review focuses on the thigh extensor muscle because it is trained extensively in ROWERS as they push the boat forward in the water as they row versus the same muscle that is essentially unused after a complete SCI except for an occasional spasm (9,22). A complete SCI that leaves someone paraplegic and being a crew member in an elite, international-caliber rowing crew have some vivid contrasts that affect the lower extremities and influence exercise ability. Each markedly affects the muscles of the lower extremities by demanding extreme alterations in loading history and physical activity. A number of studies have emphasized the plasticity of the skeletal muscle in its response to extreme exercise (12,15,16,23), whereas others have shown very opposite effects from extreme detraining or injury, which leads to a significant reduction in voluntary contraction of muscle and subsequent dysfunction of joint action (3,4,8,9). As might be expected, extreme physiologic-metabolic adaptations are accompanied by histochemical and ultrastructural adaptations in muscle. Plasticity of the human body can be assessed, as will be done in this brief review, by comparing responses and adaptations to physical activity between these two very different physically demanding situations. We also evaluate the muscles that meet the energy demands and briefly recommend a few topics for study using molecular biology techniques. We have studied ROWERS, both in the cross-section model and longitudinally over the past four decades, and are internationally known over the past decade for our study of individuals with SCI, especially their muscles and exercise ability. Both groups have physical activity abilities that are markedly influenced by the loading history of their lower extremities, thus we felt it would be of value to contrast their adaptations to these extremes in loading and share our decades of experiences.

Back to Top | Article Outline


The sport of competitive rowing may appear, on initial observation, to be primarily an upper-body exercise. A closer examination, however, clearly indicates that boat movement depends, for the most part, on repeated and powerful contractions of the hip and thigh extensor muscles (i.e., the gluteus maximus and quadriceps femoris muscle groups). This is because competitive racing boats are equipped with sliding seats that ergonomically permit the rower to perform rapid flexion of the hips and knees and, thus, coiling or compacting the body in preparation for vigorous hip, lower-extremity, and back extension that serve as the "drive"; or power phase of the rowing stroke (Fig. 1). Muscular power is transferred to the oars that impart acceleration to the boat. The power phase is immediately followed by the compacting phase of the stroke that results in brief deceleration of the boat. Elite oarsmen can average in excess of 500 W for a 5- to 6-min competitive effort. However, their intermittent power application, even for only a brief period during each rowing stroke, is responsible for relative low mechanical efficiency; only 18-22% compared with those of elite male cyclists and runners who average between 25 and 30%. The continuous, alternating limb movements of cyclists and runners versus the simultaneous limb movement with interruption by the rowers during the coiling recovery phase can explain the mechanical efficiency differences.



Elite rowing athletes can be physically characterized as mature, tall, and lean. Male rowers average 26 yr, 194 cm in height, 92 kg in weight, and have between 8 and 9% body fat. Their female counterparts average 26 yr and are shorter and lighter at 177 cm and 77 kg. They have between 15 and 16% body fat (12). Successful rowers exhibit a special combination of excellent muscular power and endurance accompanied by their extraordinary aerobic capacity. The application of these physiological attributes to a competitive event is unique among aerobic sports because, with the exception of the single sculling event, all international competition, including Olympic racing, is conducted with multiple-oared crews. Thus, the sport of rowing is the only predominantly aerobic physical activity in which athletes compete as teams, and this means that each multiple-oared boat can be only as fast as its weakest human link.

As opposed to the remarkable physical performance abilities of international-caliber rowers, nontrained individuals after a complete SCI that leaves a person a paraplegic represent humans with very low exercise ability. In contrast to the 500-W power output that can be maintained during simulated competitive rowing, nontrained persons with complete, paraplegic SCI might achieve 75 W during a peak O2 test when the atrophied muscle groups of the affected lower extremity receive neuromuscular electrical stimulation (NMES) of the extensor muscle groups (Fig. 2 as has been reported, for example (6,13,33). Such individuals with paraplegia with atrophied lower limbs can perform ambulation in the upright posture using a Parastep ergometer; however, similar to rowing, this type of exercise is very inefficient (19) Fig. 3). Those with paraplegia are inefficient in "walking" because of excessive use of their upper bodies. Even if the unaffected arms are used simultaneously with the lower limbs receiving NMES to exercise on a Nu-Step (Fig. 2), the power output is at best about one fourth that of international-caliber rowers.





Back to Top | Article Outline

Aerobic Power

We have spent years studying our volunteers and have often studied the same subject consistently over years, probably accumulating the most definitive data on our subjects of anyone studying such extremes. The rowing data are important because they add high performance information to sport science, and the SCI data have aided in the clinical treatment of such individuals. Briefly, it is well known that absolute maximal oxygen consumptions of elite rowers are among the highest recorded (11,13,14,18,24,28-30) and have been increasing steadily among our National and Olympic Team candidates over a 40-yr period of study. Recently, we have observed absolute O2max values exceeding 7 L·min−1 in several elite U.S. oarsmen, the highest being 7.5 L·min−1, and more than 5 L·min−1 for a number of female rowers. These absolute responses translate into relative values that are higher than 80 mL·kg−1·min−1 for men and more than 65 mL·kg−1·min−1 for women. These impressive aerobic capacities and mechanical power outputs translate into very high energy expenditures, which were first produced by Astrand and Rodahl in 1977 (2) and reproduced with coauthors in 2003. These textbook power and energy equivalents are estimates, but are comparable to many other similar power and energy tables and norms reproduced in current exercise science texts (26).

Anaerobic power contributes only about 20-25% of the total energy to a 2-km competitive rowing effort compared with the 75-80% that is the relative contribution of the aerobic energy systems to this effort (12,18). Anaerobic energy systems, however, are maximized at critical stages of a 2-km competitive effort: at the start of the race (initial 100-200 m) and again during the final sprint of about 200 m. If one accounts for lactic acid production and clearance by skeletal muscles, postcompetitive capillary blood lactates greater than 18 mmol·L−1 for women and 30 mmol·L−1 for men represent significant anaerobic responses (12,18).

This brings us to a discussion of the energy sources during exercise in individuals with complete, paraplegic SCI as compared with international-caliber rowers. O2peak of complete SCI elite wheelchair racers can be one fourth that of international-caliber rowers, more than 2 L·min−1 versus almost 8 L·min−1 (33). Anaerobic capacity of those with SCI has usually not been measured, but their low O2 and slow nature of physical activity make it unlikely that anaerobic energy sources, blood lactate as high as 10 mmol·L−1 (6), are relied on by individuals just trying to ambulate. We finalize this portion of this brief review noting that all is not lost after a complete SCI, in large part, because of the physical effort of a few individuals and the societal view that all individuals, irrespective of the physical traumas they have survived, should and are given the opportunity to compete. A peak aerobic power in the 30s (mL·kg−1·min−1) that is similar to that of sedentary able-bodied individuals can be achieved by elite complete SCI wheelchair racers (33).

Back to Top | Article Outline

Muscle Fiber Type and Mitochondrial Content

Although oxygen delivery and use by rowers are extraordinary, another physical characteristic that defines these athletes is the presence of an unusually large percentage of type I (slow-twitch) muscle fibers in the primary power muscles of rowing, the quadriceps femoris muscle group, represented by m. vastus lateralis. Several studies have revealed that highly trained, elite rowers, both men and women, have a preponderance of type I fibers (15,16,21). Although it is not surprising that elite endurance athletes exhibit a dominance of type I fibers (7), results of our analyses of vastus lateralis biopsy samples of rowers are very interesting. Competitive rowing has been identified as primarily dependent on aerobic energy sources for successful efforts over 2000 m (22). Unlike most other endurance athletes whose competitive events are of relative longer duration, a rower's competition lasts only 5-7 min. Rowers possess extraordinary leg power in addition to their excellent aerobic capacities (16).

Training for rowing encompasses many hours of long, steady-state sessions and shorter power-rowing efforts. Elite rowers tend to be unique athletes utilizing a rare combination of muscular power and endurance to ensure success in their sport. Mean percentages of 72-77% for type I fibers have been consistently observed in the vastus lateralis of elite male rowers (17). Type IIA fibers accounted for 25-32% of the total fibers analyzed, whereas fast-twitch type IIX fibers were almost nonexistent in both male and female rowers. Larsson and Forsberg (21) have observed similar proportions of fast- versus slow-twitch fibers in male rowers' limb muscles.

Only limited muscle biopsy sampling has been obtained from female rowers, and currently no biopsy data have been reported for international-caliber female rowers, other than our data favoring type I preponderance as compared with healthy, sedentary controls (31). Clarkson et al. (5) found the ratio of type I to types IIA and IIX in the m. vastus lateralis of elite oarswomen to be about 60:40, whereas a 55:45 ratio favoring type I fibers was observed in biopsy samples removed from the biceps brachii of these women.

We have also observed significantly greater mitochondrial size and density and increased activity of oxidative enzymes in all fiber types examined in rowers when compared with controls and elite power lifters, especially in type I fibers (10,16). These findings might help explain the extraordinary aerobic capacities of elite rowers. Increased oxidative capacity in the muscle of oarsmen was predicted by Larsson and Forsberg (21), who observed capillary densities in trained rowers' muscles that nearly doubled those reported for the same muscles in untrained subjects (1,31). Measurements of cross-sectional areas of all fiber types in rowers revealed unusually large diameters when compared with the same fiber types in controls and other endurance athletes (21,29). This finding was especially exaggerated in the slow-twitch fiber population. Although IIA fiber types remained as the largest of the three basic fiber types in oarsmen, several observations of type I fibers were analyzed from male rowers that exceeded 6000 μm2, and some were slightly larger than 7000 μm2. Type IIA fibers in male oarsmen were similar in cross-sectional areas (8000 μm2) to those reported for the fast fiber population in weight lifters (15). The proportion and the diameter of the slow-twitch, type I fibers of rowers are very special among predominately aerobic athletes. Rowers must have an abundance of oxidative fibers to meet the extraordinary aerobic demands of their sport. At the same time, rowers must develop unusually large slow-twitch fibers resulting from the high power requirements of their sport; 200 hard strokes are required to successfully negotiate the 2-km competitive distance, which means 200 near-maximal trunk, hip, and knee extensions in order to finish.

Skeletal muscle fibers in complete SCI individuals also have several unique characteristics. Interestingly, using histochemical techniques, we have shown that fiber type composition during the first several months after a complete SCI is not altered in the affected m. vastus lateralis (4). These results may seem surprising when it is considered that the extreme loading imposed on elite rowers is associated with an abundance of type I fiber. Thus, the extreme unloading of SCI should decrease the abundance of type I fibers. We have confirmed our results using electrophoretic analyses of myosin heavy chains (32). Nonetheless, it is well accepted that eventually slow-twitch fibers transform into fast-twitch fibers after complete SCI (4). This response takes several years, however. When these results are considered with those obtained from studies of elite rowers, several conclusions can be drawn. Most importantly, voluntary loading or the lack of it can alter the fiber type composition of human muscle by up- or downregulation expression of type I myosin heavy chain. This requires extreme alterations in loading history and apparently several years to be accomplished.

The aerobic enzyme capacity of affected muscle fibers is not altered meaningfully by unloading (25) or by the first several months after complete SCI (4). Others using similar methods of estimating a fiber's aerobic capacity found comparable results (27). When aerobic capacity is expressed per muscle weight, as is often done in humans and has been done in lower mammals, total enzyme content during unloading decreases because of preferential loss of contractile material (20). This loss of contractile material is associated with extreme muscle fiber atrophy. (4). Affected muscle fiber cross-sectional area (CSA) about 6 months after a complete SCI is approximately 2500 μm2, about one half the diameter of fibers unaffected by complete SCI in a sedentary, able-bodied individual. O2max also plummets when fiber size of the major muscles engaged in whole-body exercise decreases to about one half. Thus, fiber CSA must be sufficient in able-bodied individuals to allow performance of aerobic-type physical activities. In contrast, reductions in fiber CSA with the disuse caused by complete SCI apparently compromise whole-body O2peak.

In retrospect, several aspects of our study of extreme human adaptations to alterations in physical activity history warrant further evaluation using molecular biology. The tendency for muscle to eventually contain mainly fast-twitch fibers when it is extremely atrophied should be examined. The atrophy occurs from disuse, but this inactivity does not cause cell death. Why? No one knows. This extreme disuse does, however, make muscle more susceptible to injury, even action-induced cell death, from the very thing this tissue is exquisitely designed to do: contract (5). Quite confusing is the observation that fast-twitch, type IIA fibers are the largest fibers of the three major fiber types in able-bodied males, but muscle becomes fast-twitch when it shows extreme atrophy from inactivity and unloading of a complete SCI. The atrophy can be countered with activities that cause hypertrophy, and this has apparent health benefits (22). In contrast, why muscle transforms its fast-twitch fibers to slow-twitch fibers with extreme increases in physical activity while increasing its size also begs examination. Muscle fiber size is probably not markedly increased to limit diffusion distance of oxygen from the capillaries to the abundant mitochondrial make-up of the fibers. Although this has yet to be demonstrated, the "low" energy demand during contraction of slow-twitch fibers is probably met by their mitochondria. This allows very high power output, high energy input, and continuous performance to be maintained, whereas individuals with SCI present the other extreme of human adaptive models; very low power output with considerably less or nonexistent anatomical or physiological adaptive responses to improve exercise ability.

Back to Top | Article Outline

Unifying Thoughts

First, male ROWERS are unique in their physical characteristics and stature for athletes well equipped to excel in events that require a substantial aerobic energy supply to compete. These individuals are often 195 cm tall, weigh 100 kg, and often run marathons in 2 h 30 min or less. Clearly, these "aerobic" individuals with a O2max of more than 80 mL·kg−1·min−1 are not typical of most marathon runners (i.e., small, light, and able to run). Even "large" individuals can be exceptional aerobic athletes. Incomplete paraplegics, on the other hand, find it extremely difficult to walk (Fig. 3); they gain adipose tissue after injury (9), have a typical body stature, and can achieve an aerobic power of 20 mL O2·kg−1·min−1. For both, the cardiorespiratory system must adapt to their unique loading histories and supply the energy substrate. Second, male ROWERS can exercise for several minutes and supply the energy, both aerobically and anaerobically, to burn a 500-W light bulb during the whole effort. This remarkable exercise ability is evidenced by individuals who typically achieve a blood lactate above 30 mmol·L−1. Surely, this ability comes from their unique "boat races" used during training. As opposed, "walking" is sufficiently difficult for incomplete paraplegics that anaerobic power is not a concern. Third, thigh extensor muscles of ROWERS are almost all slow-twitch, whereas the same muscles of individuals with complete SCI become all fast-twitch fibers within a few years of injury. The typical individual has a 50-50% makeup. These data strongly suggest that the loading history of muscle can eventually allow it to respond to its demands. The more continuous, extreme demand results in less energy-demanding slow-twitch fibers, whereas no demand, complete SCI, results in the most energy-demanding, but unusually small, fast-twitch fibers. Loading history and physical activity can increase the mitochondrial content to meet the increased energy demand by the less demanding fiber type, slow-twitch fibers. In contrast, minimal physical activity as occurs after a complete SCI does not reduce mitochondrial content to supply energy per fiber volume; rather, there are now just smaller, more energy-demanding fast-twitch fibers. Fourth, contrast of these unique individuals might allow muscle molecular biologists to explain the molecular control of expression on myosin heavy chain, contractile machinery content, and mitochondrial content.

The research work presented by the authors was supported in part by U.S. Rowing and the Ohio University Foundation (to F.H.), The Paralyzed Veterans of America: Spinal Cord Research Foundation (to G.A.D.) and the NIH (HD 39676 and HD 39676S2 to G.A.D.).

Back to Top | Article Outline


1. Anderson, P. Capillary density in skeletal muscle of man. Acta. Physiol. Scand. 95:203-205, 1975.
2. Astrand, P. O., and K. Rodahl. Textbook of Work Physiology, 2nd ed. New York: McGraw-Hill, 1977, p. 454.
3. Bickel, C. S., J. M. Slade, and G. A. Dudley. Long-term spinal cord injury increases susceptibility to isometric contraction-induced muscle injury. Eur. J. Appl. Physiol. 91(2-3):308-313, 2004.
4. Castro, M. J., D. F. Apple Jr., R. S. Staron, G. E. R. Campos, and G. A. Dudley. Influence of complete spinal cord injury on skeletal muscle within six months of injury. J. Appl. Physiol. 86:350-358, 1999.
5. Clarkson, P. M., J. Graves, A. M. Me1chionda, and J. Johnson. Isokinetic strength and endurance and muscle fiber type of elite oarswomen. Can. J. Appl. Sports Sci. 9:127-132, 1984.
6. Cooper, R. A., F. D. Baldini, M. L. Boninger, and R. Cooper. Physiological responses to two wheelchair-racing exercise protocols. Neurorehabil Neural Repair 15(3):191-195, 2001.
7. Coyle, E. F., L. S. Sidossis, J. F. Horowitz, and J. D. Beltz. Cycling efficiency is related to the percentage of type I muscle fibers. Med. Sci. Sports Exerc. 24:782-788, 1992.
8. Dudley, G. A., M. Castro, S. Rogers, and D. F. Apple Jr. A simple means of increasing muscle size after SCI: a pilot study. Eur. J. Appl. Physiol. 80:394-396, 1999.
9. Elder, C. P., D. F. Apple, C. S. Bickel, R. A. Meyer, and G. A. Dudley. Intramuscular fat and glucose tolerance after spinal cord injury-a cross-sectional study. Spinal Cord 42(12):711-716.
10. Gregory, C. M., K. Vandenborne, and G. A. Dudley. Metabolic enzymes and phenotypic expression among human locomotor muscles. Muscle Nerve 24:387-393, 2001.
11. Hagerman, F. C. Physiology and nutrition for rowing. In: Perspectives in Exercise Science and Sports Medicine, P. R. Lamb, H. G. Knuttgen, and R. Murray (Eds.). Carmel, IN: Cooper, pp. 221-302.
12. Hagerman, F. C. Applied physiology of rowing. Sports Med. 1:303-326, 1984.
13. Hagerman, F. C., W. W. Addington, and E. A. Gaensler. A comparison of selected physiological variables among outstanding competitive oarsmen. J. Sports Med. Phys. Fitness. 12:12-22, 1972.
14. Hagerman, F. C., W. W. Addington, and E. A. Gaensler. Severe steady state exercise at sea level and altitude in Olympic oarsmen. Med. Sci. Sports. 7:275-279, 1975.
15. Hagerman, F. C., S. J. Walsh, R. S. Staron, et al. Effects of high-intensity resistance training on untrained older men. Strength cardiovascular, and metabolic responses. J. Gerontol. Biol. Sci. 55A:B336-B346, 2000.
16. Hagerman, F. C., and R. S. Staron. Seasonal variations among physiological variables in elite rowers. Can. J. Appl. Sport Sci. 83:143-148, 1983.
17. Hagerman, F. C., G. R. Hagerman, and T. C. Mickelson. Physiological profiles of elite rowers. Phys. Sports Med. 7:74-81, 1979.
18. Hagerman, F. C., M. C. Connors, J. A. Gault, G. R. Hagerman, and W. J. Polinski. Energy expenditure during simulated rowing. J. Appl. Physiol. 45:87-93, 1978.
19. Jacobs, P. L., B. Johnson, and E. T. Mahoney. Physiologic responses to electrically assisted and frame-supported standing in persons with paraplegia. J. Spinal Cord Med. 26(4):384-389, 2003.
20. Kandarian, S. C., R. C. Boushel, and L. M. Schulte. Elevated interstitial fluid volume in rat soleus muscles by hindlimb unweighting. J. Appl. Physiol. 71:910-914, 1991.
21. Larsson, L.,and A. Forsberg. Morphological muscle characteristics in rowers. Can.J. Appl. Sports Sci. 5:239-244, 1980.
22. Mahoney, E., C. Elder, C. Black, C. S. Bickel, and G. A. Dudley. The effect of NMES induced resistance exercise on skeletal muscle morphology and glucose tolerance after chronic complete SCI. Arch. Phys. Med. Rehabil. 86:1502-1504, 2005.
23. Mickelson, T. C., and F. C. Hagerman. Anaerobic threshold measurements of elite oarsmen. Med. Sci. Sports Exerc. 14:440-444, 1982.
24. Nowacki, P. E., R. Krause, and K. Adam. Maximal oxygen uptake by the rowing crew winning the Olympic gold medal. Pflugers Arch. 312:R66-R67, 1969.
25. Ohira, Y., B. Jiang, R.R. Roy, et al. Rat soleus muscle fiber responses to 14 days of spaceflight and hindlimb suspension. J. Appl. Physiol. 73(2 Suppl):51S-57S, 1992.
26. Powers, S. K., and E. T. Howley. Exercise Physiology, Theory and Application to Fitness and Performance, 4th ed. Boston: McGraw-Hill, 2001, pp. 235-241.
27. Rochester, R., M. J. Barron, C. S. Chandler, R. A. Sutton, S. Miller, and M. A. Johnson. Influence of electrical stimulation of the tibilias anterior muscle in paraplegic subjects. 2. Morphological and histochemical properties. Paraplegia 33:514-522, 1995.
28. Saltin, B.,and P. O. Astrand. Maximal oxygen uptake in athletes. J. Appl. Physiol 23:353-358, 1967.
29. Secher, N. H. Physiological and biomechanical aspects of rowing. Sports Med. 15:24-42, 1993.
30. Secher, N. H., O. Vaage, and R. C. Jackson. Rowing performance and maximal aerobic power in oarsmen. Scand. J. Sports Sci. 4:9-11, 1982.
31. Staron, R. S., F. C. Hagerman, R. S. Hikida, et al.Fiber type composition of the vastus lateralis muscle of young men and women. J. Histochem. Cytochem. 48:623-629, 2000.
32. Talmadge, R. J., M. J. Castro, D. F. Apple Jr., and G. A. Dudley. Phenotypic adaptations in human muscle fibers 6 and 24 wk after spinal cord injury. J. Appl. Physiol. 92:147-154, 2002.
33. van der Woude, L. H., C. Bouten, H. E. Veeger, and T. Gwinn. Aerobic work capacity in elite wheelchair athletes: a cross-sectional analysis. Am. J. Phys. Med. Rehabil. 81(4):261-271, 2002.


©2006The American College of Sports Medicine