Inverse relationship between VO2max and economy/efficiency in world-class cyclists ALEJANDRO LUCiA, JESOS HOYOS, MARGARITA PEREZ, ALFREDO SANTALLA, and JOSE L. CHICHARRO Facuitad de Ciencias de la Activated Fisica y el Depone, Universidad Europea de Madrid, Madrid, SPAIN; Asociacion Depo-tiva Banesto, Madrid, SPAIN; Universidad Alfonso XEl Sabio, Madrid, SPAIN; and Departamento de Enfermeria, Universidad Complutense, Madrid, SPAIN
ABSTRACT LUCtA, A., J. HOYOS, M. PEREZ, A. SANTALLA, and J. L. CHICHARRO. Inverse relationship between VO^ and economy/ efficiency hi world-class cyclists. Med. Sci Sports Exerc, Vol. 34, No. 12, pp. 2079-2084, 2002. Pnrpctc: To determine the relationship that exists between VO^^ and cycling economy/efncieDcy during intense, submaxunal exercise m world-class road professional cyclists. Methida: Each of 11 male cyclists (26 ± 1 yr (mean ± SEM); VOjM>: 72.0 ± 1.8 mL-kg"' -mm"') performed: 1) a ramp test for VOtou determination and 2) a constant-load test of 20-min duration at the power output eliciting 80% of subjects' VOj^, during the previous ramp test (mean power output of 385 ± 7 W). Cycling economy (CE) and gross mechanical efficiency (GE) were calculated during the constant-load tests. Results: CE and GE averaged 85.2 ±' 2.3 W-ir'-min"' and 24.S ± 0.7%, respectively. An inverse, significant, correlation v«s found between 1) V02mm (mL-kg"°K-min~') and both CE (r = - 0.71; P - 0.01) and GE (-0.72; P = 0.01), and 2) VO^ (mL-kg-'imn-') and bo* CE (r = - 0.65; P = 0.03) and GE (-0.64; P = 0.03). Conclusions: A high CE/GE seems to compensate for a relatively low TOj^, in professional cyclists. Key Werdt: PERFORMANCE, GROSS EFFICIENCY, POWER OUTPUT, PROFESSIONAL CYCLING, CYCLE ERGOMETRY
revious studies have analyzed the main physiological determinants of performance in endurance sports. These include, among other variables, maximal oxygen uptake fv*O2mKi), Jactate/ventilatory thresholds, and economy/efficiency (11). Concerning the latter, a better economy or efficiency will decrease the percentage of ^Qinux required to sustain a given mechanical work and thus might be advantageous to endurance performance. Previous research has indeed shown the importance of economy in endurance running performance (5,17,21,26,30). For instarce, the superior performance of Kenyan runners during the last decades compared with their European counterparts is attributable, at least partly, to their greater running economy (26). Other variables such as VO^,,,,,, do not appear to differ between Europeans and Africans. The relatively low VO2roiut values (~ 70 mL-kg~ '-min"1) that sometimes are found in world-class male endurance runners can be compensated for by a great running economy (12,17,18). Furthermore, an inverse relationship has been reported in highly trained runners between VO2max and running economy (18,23).
To the best of our knowledge, no previous study has analyzed whether VO2nHX and economy/efficiency are inversely related in top-level cyclists (i.e., professional riders), as it occurs in elite runners. Although high VO2mlx values (5.0 to 5.5 L-min~' or 70-75 mL-kg^'-min"1) are usually found in world-class cyclists, VO^^ is not the main performance determinant in this sport (14,15). For instance, amateur, well-trained cyclists show similar VO2mlM values to those of professional riders (14,15). Provided a minimum level of VO2lnax is reached (e.g., > 65 mL-kg^'-min"1), cycling economy (CE) and gross mechanical efficiency (GE) might be especially important in top-level endurance cycling (9,14). Indeed, professional riders are considerably more economical and efficient than amateur riders despite similar VO2n)M values in both groups (14,15). The purpose of die study was to determine if there exists a relationship between VO^,^ and CE/GE during intense, submaximal exercise in a group of world-class cyclists.
METHODS Subjects. Eleven professional male road cyclists (age (mean ± SEM): 26 ± 1 yr) were selected for this investigation. Written informed consent was obtained from each participant, and the institutional Ethics Committee (Complutense University of Madrid) approved the study. A previous physical examination (including ECG and echocardiographic evaluation within the previous month) ensured that each participant was in good health. Several of the present subjects are among the best cyclists in the world, according to the ranking of the International Cycling Union. To ensure that all of them could be really considered as
Address for correspondence: Alejandro Lucia, MJD, PhD, Departamento de Ciencias Mofctogicas y Fisiologia, Universidad Europea de Madrid, E-28670 VDlavicusa de Odon, Madrid, Spam; E-mail: [email protected] Submitted for publication March 2002. Accepted ftr publication August 2002. 0195-9131/D2/3412-2079/$3.0«D MEDICINE & SCIENCE IN SPORTS A EXERCISED Copyright O 2002 by the American College of Sports Medicine DO1:10.1249/01.MSS.0000039306.92778.DF
"world-class" riders, they were required to meet the following requirements: 1) bave participated in the mean competition's of the professional category (e.g., 3-wk tour races) and 2) have woo at least one major professional race (e.g., one or more individual stages and/or final classification of a major 1-wk or 3-wk race (Giro d'ltalia, Tour de France, or Vuelta a Espana), or Top 3 in World Championships). Hemoglobin and hematocrit levels were measured in each subject before participating in the experiments. Mean values of hemoglobin and hematocrit averaged 14.7 ± 0.3 g-dL~' (range, 12.8-16.1)and43.5 ±0.7% (range, 39.9-46.5)and thus were within normal, physiological limits for endurance athletes (27). Study protocol Each subject reported to our laboratory on two consecutive days during the months of January or February, before the start of the competition season. During the days before testing and the test days, the subjects followed a similar type of high-carbohydrate (CHO) diet (~ SCO g CHO-d~'). On the first day, they performed a maximal exercise test (ramp protocol) for VO2mM determination, and on the second they performed a subraaximal, constant-load test to measure CE and QE. Both tests were performed on the same electromagnetically braked cycle ergometer (Ergometrics 900, Ergo-line; Barcelona, Spain). The torque measuring unit was calibrated before each testing session (4-5 tests per session) with a known weight of 4.0 kg. All the components of the ergometer were checked by an experienced technician before the start of the study. Before this investigation, the ergometer was equipped with a new chain. This cycle ergometer has been used in numerous studies conducted in our laboratory with professional cyclists (13-15). During the tests, the subjects adopted the conventional (upright sitting) cycling posture. This posture was characterized by a trunk inclination of - 75° and by the subject placing his hands on the handlebars with elbows slightly bent (~ 10° of flexion). All the tests were performed under similar environmental conditions (21-24°C, 45-55% relative humidity). Subjects were allowed to choose their preferred cadence within the range 70-90 rpm during both type of tests. This is known to better simulate actual cycling conditions compared with tests performed at a fixed cadence. During actual racing, indeed, the preferred pedalling cadence of professional riders ranges from 70 rpm (hill climbs) to 90 rpm (flat terrains or individual time trials) (13). A pedal-frequency meter was used by each subject to maintain his pedalling cadence within the aforementioned range. The subjects were cooled with a fan throughout the bouts of exercise. ; Maximal exercise test For the maximal test, a ramp protocol was followed until exhaustion. This type of protocol has been used for the VO2nMX determination of professional cyclists in several previous studies (13-15). Starting at 20 W, the workload was increased by 25 W'tnin"1. The tests were terminated when pedal cadence could not be maintained at 70 rpm (at least). Verbal encouragement was given to the subjects to continue the test until they were exhausted. All the participants had previous experience with 2080
Official Journal of the American College of Sports Medirina
this type of protocol. Heart rates (HR, in bpm) were monitored during the tests from modified 12-lead ECG tracings (EK56; Hellige; Freiburg, Germany), and gas exchange data were collected continuously using an automated breath-bybreath system (CPX; Medical Graphics; St. Paul, MM). With this system, O2 and CO2 are measured with rapid analyzers, while a disposable flowmeter, which is based on the principle of differential pressure measurement by two sensitive differential pressure transducers, analyzes ventilatory flow. This type of flowmeter has been shown to be accurate (to within 2% of the target value obtained from Douglas bag collections) and reproducible for the measurement of minute ventilation during exercise (25). The mean percentage difference and the correlation coefficient between the VO2 measurements provided by the breatb-by-breath system used in the present study and the Douglas bag method is 2.2% and 0.995 (P = 0.0001), respectively (unpublished data provided by the manufacturer from maximal tests performed in 15 subjects of varying fitness levels). The O2 and CO2 analyzers and the flowmeter were calibrated before each single test with reference gases (Praxair, Madrid, Spain) at a concentration of 15.99% for O2 and 4.00% for CO2, and a 3-L syringe (25), respectively. For each test, VO2mwc was recorded as the highest V02 value obtained for each 1-min interval, and the maximal power ou was computed as follows (22):
where Wf is the value of the last completed workload (in W), t is the time the last uncompleted workload was maintained (in s), 60 is the duration of each completed workload (in s), and 25 is the power output difference between (be last two workloads. Constant-toad test at 80% VO^,,,,. The submaximal, constant-load tests were performed over a 20-inin period at a fixed power output. For each subject, the latter was identified on the VO2 (average for each l-min interval): power output curve of the previous ramp test by straight linear interpolation, as shown in Figure 1 . Each 20-min test was preceded by a 15-min warm-up period, consisting of 5 min at 70 W, 3 inin at 60% of the maximal power output reached during the previous ramp test, and 2 min of gradual workload increases until the target power output was attained. Gas exchange data and HR were monitored as in the maximal tests. In addition, blood variables were determined as detailed below. Before the start of the experimental protocol, a 21 -gauge butterfly needle was inserted into the antecubital vein of each subject. The catheter was kept patent by periodic flushing with a heparinized saline solution. Blood samples were collected every 5 min during the tests. During each sampling period (~ 15 s), a 1-mL aliquot was initially withdrawn to clear the catheter, and a 1.5-mL blood sample was subsequently collected vsing a heparinized syringe for the immediate estimation of PCO2 and pH using an automated blood gas analyzer (ABL5; Radiometer; Copenhagen, Denmark). Bicarbonate concentration [HCO3~] was calculated using the pH and PCO2 values. Capillary blood http://www.acsm-msse.org
11 ' I ' ' " 1'' " I " '' I ' ' " I ' ' " I ' " ' I " '.' I " •' 300 325 350 375 400 425 450 475 500 525 550
gg C S S S SS
SS 3 S R 3 SS
Power Output (W) FIGURE 1—Example ef determination »f the power output for the constant-load test! at 89% VO^^ in one study subject The target ontpal was Identified by straight linear Interpolation on this curve, which thaws the rdati.nshlp obtained fat die ramp tests between VO2 (average value for each 1-mta workload) expressed »» % ef VOmn, »d power outpet. In this particular case, 199% VO3mM was 5257 mL and the closest power oatput eBcJUng 80% Vo]BU was 395 W. The smallest change ia power output that can be applied to the erguneter fa ± 5 Wj therefore, the target power output for each subject was rounded off at mnlUpla of 5 (e.g^ 350 W, 355 W, 360 W, etc.).
samples were taken from fingertips (25 fjiL) every 5 min during the tests and immediately after exercise for the determination of blood lactate concentration (BLa) using an electro-enzymatic analyzer (YSI 1500; Yellow Springs, OH). Average values of CE and GE during the constant-load test were calculated. CE was expressed in (W-L~'-min~') (7), and GE was calculated as the ratio of work acconplished-min" ' (i.e., W converted to kcal-min"1) to energy expended-min"1 (i.e., in kcal-min"'), as described elsewhere (7). Energy expended was calculated from VO2 and respiratory exchange ratio (RER) using the tables of Lusk(16). Statistical analysis. Pearson product-moment correlation coefficients were calculated to determine whether there was a significant relationship between VO2roax and both CE and GE. VOj,^ was expressed in absolute units (L-rnin" ') and in relative units (mL-kg^-min"1 and mL-kg~a32min~'). The later was performed following the recommendation by Padilla et at. (22) to express physiological values relative to mass exponents of 032 and 1 in order to adequately evaluate level and uphill cycling ability, respectively, in elite cyclists. The level of significance was set at 0.05. To discard any possible influence of individual variations in pedalling cadence on CE/GE, correlation coefficients were also calculated between these variables. Results are expressed as means ± SEM.
S 85 S S R SR
SS S R
RESULTS Individual characteristics of the subjects (demographic and physical characteristics, history of cycling performance in the professional category) and the results of both ramp *OJKW< AND ECONOMY/EFFICIENCY
Medicine & Science in Sports & Exercise.
i = -0.65 (P = 0.03)
T = -0.71 (P = 0.01}
d1TO > ft 66
CE (W-l/'-min'1) FIGURE 2— Relationship between cydtag economy (CE)t
and constant-load tests are shown in Table 1. The mean values of pH and [HCO3~] obtained during the constantload tests averaged 7.3 8 ±0.01 (range, 7.30-7.45) and 19.1 ± 0.9 mM (17.2-21.4), respectively. The following significant, inverse correlations were found: VO2tm (mL-min"1) versus both CE (r = -0.61; P = 0.047) and GE (-0.63; P = 0.04); VO^ (mL-kg~OJ2-mur') versus both CE (r = -0.71; P = 0.01) and GE (-0.72; P = 0.01) (Figs. 2 and 3); and VOj,^ (mL-kgr'-mijT1) versus both CE (r = -0.65; P = 0.03) and GE (-0.64; P = 0.03) (Figs. 4 and 5). No significant correlation (P > 0.05) was found between pedalling cadence and either CE (r = 0.02) or GE (r = 0.002).
DISCUSSION The main finding of our study was that, in professional world-class cyclists, both CE and GE are inversely correlated to VO2max (either expressed in absolute or relative units). It follows that a high CE/GE could compensate for a relatively low J^O2ma in these athletes. Although compa-
FIGURE 4—RelrttensMp between VO,M, (mL-kg"'-mlB"1) and cycling economy (CE),
rable findings have been obtained in highly trained distance runners (12,17,18,23), to the best of our knowledge, no previous study has assessed the possible relationship between VO2max and CE/GE in cyclists of this high fitness level. In addition, no data are available about the CE or GE of humans able to tolerate such high power outputs during prolonged endurance cycling (i.e., average of ~ 400 W in our subjects and s 400 W in four of them) before significant lactic acidosis occurs (average values of BLa were relatively low and pH was maintained within normal limits). The values of GE obtained in the present study (~ 24%) are similar to those recently reported in professional riders at the power outputs eliciting the lactate threshold (LT) and the respiratory compensation point (RCP) during a ramp test (14), and higher than those previously measured in not highly trained cyclists (average of-20%) (19,20). Although GE is not an accurate measure of muscle efficiency (7), it is a good indicator of whole body efficiency and thus might be relevant from a practical point of view (3). In addition, GE measurements performed during laboratory testing have been proved to be reliable (19). Although the physiological M ^-x
r = -0.72 (P = 0.01)
BO ™ 78
GE (%) FIGURE 3—Relationship between VOj_» grosi mccbinfcal efficiency (GE).
GE (%) -') »nd
Official Journal of the American College of Sports Medicina
FIGURE 5—Relationship between VO^., (mL-kg~1-min~>) and gross mechanical efllciency (CE).
and metabolic determinants of this variable remain to be fully understood (19), several factors can have an influence on GE, such as pedalling cadence (4), diet (24), overtraining (1), genetics (2), or fiber type distribution (7). It is unlikely thai the three first factors could have significantly influenced the present results given that 1) we found no significant correlation between pedalling cadence and GE (r = 0.002; P > 0.05) and individual values of preferred pedalling cadence ranged within relatively narrowed limits (71-88 rpm; 2) the diet of the subjects was standardized as specified in the Methods section; and 3) all the subjects were tested before the competition period, and none of them showed symptoms or signs of overtraining. At the present moment, it is not possible to determine the influence of the remaining two factors, genetics and fiber type distribution, on the GE of professional road cyclists. For instance, scarce data are available in the literature showing the results of musc'e biopsies in professional riders of the highest competition level. Nevertheless, the GE of humans is positively related to the percentage distribution of Type I fibers in exercising muscles. Previous research with endurance trained cyclists has shown that a higher percentage of Type I fibers in one of the main muscles involved in cycling (vasais lateralis) is associated with a greater GE during prolonged (1 h) exercise of either high (>LT) or moderate intensity (400 W) during prolonged periods at the lowest possible metabolic cost Moreover, once a certain fitness level is reached (e.g., the amateur category), submaxhnal variables such as GE at the LT (~ 70%vX>2lKJ or at the RCP (~ 90%VO2mK) are more important determinants of cycling performance than VO2nax (14). CE averaged ~ 85 W-L '-min"' in our subjects, although a considerable variability existed among subjects (range, 72-98). This mean value is clearly above those values (mean of ~ 75 WlT'-min"1) previously reported by Coyle and coworkers (6) in amateur highly trained riders of a lower competition level, during a simulated time trial of 1-h duration at power outputs ranging between 325 and 376 W. < AND ECONOMY/EFFICIENCY
In line with our findings, CE also showed important variations among subjects. Biomechanical/anatomical factors can have a significant influence on running economy (26). In contrast, the variations of CE and GE that occur among elite riders of a lower fitness level than the present subjects are largely attributable to variations in the percentage distribution of Type I fibers in knee extensor muscles. The best rider in the present study (e.g., two-time world champion) showed a relatively low VO2max value (slightly below 70 mL-kg~1-min""1) but very high values of both CE and GE (clearly above 90 W-L^-min"1 and 25%, respectively). The 20-min constant-load bouts were performed at the power output eliciting 80% of the subjects' VO2jMX during the previous ramp tests. Average exercise intensity increased up to ~ 86%VO2mM throughout the constant-load bouts because of the so-called "VO2 slow component"— that is, the gradual increase in VO2 thai inevitably occurs in all humans during intense, submaximal exercise, and that largely reflects an increased recruitment of inefficient Type II fibers (8). Most fibers (including both Types I and D) of the main muscles involved in pedalling are indeed recruited at the relative intensity at which the constant-load tests were performed, as shown in previous research (10,28). The most important phases of endurance cycling races (mountain ascents, time trials) are also held at high, submaximal intensities, i.e. around the RCP (14). On the other hand, both the time duration of the constant-load tests and the selected work rate (slightly below the subjects' RCP, within the so-called "isocapnic buffering phase" (29)) were well tolerated by the cyclists. For this reason, we propose that the type of constant-load exercise protocol used here could be included in the "routine" evaluation of competitive cyclists. Although thorough research has been conducted on those predictors of cycling performance that can be evaluated during more conventional gradual tests (e.g., VO2mn, LT, or RCP), to date, less is known about the possible influence of GE and CE on top-level performance in this sport. Similarly, little data are available concerning the potential trainability of GE/GE hi elite cyclists and the influence of genetic endowment on both variables. In summary, both CE and GE are inversely correlated to VOjnro in world-class endurance cyclists. As it occurs in elite runners, a high CE/GE could compensate for a relatively low VO2nuot. Further research is needed in this field, particularly to determine which aspects of training (e.g., technique modification, high-intensity intervals versus lowintensity training, etc.) have the greatest impact on CE/GE. We propose that constant-load exercise protocols as the one used in this study could be included in the "routine" tests that most competitive cyclists perform several times over the season. The information provided by constant-load bouts (particularly CE and GE) is of practical applicability and complementary to that obtained from the more conventional gradual tests to exhaustion (e.g., VO2m>x, LT, or RCP). This study was financed by Asociaddn Deportiva Banesto. Medicine & Science in Sports & Exercise*
REFERENCES 1. BAKU, R., P. OFSIAD, I. MEDBO, and 0. SEIERSTED. Strenuous prolonged exercise elevates resting metabolic rate and causes reduced mechanical efficiency. Acta PtiysioL Scand 141:555-563, 1991. 2. BUEMANN, B., B. SCHTCRNING, S. TOUBRO, et aL The association between die val/ala-55 polymorphism of the uncoupling protein 2 gene and exercise efficiency. Int. J. Oba. Relat. Metdb. Disord. 25:467-471,2001. 3. COAST, J. R. Optimal pedalling cadence. In: High-Tech Cycling, E. R. Burke (Ed.). Champaign, JL: Human Kinetics, 1996, pp. 1C1-117. 4. COAST, J. R,, R. H. Cox, and H. G. WELCH. Optimal pedalling rate in prolonged bouts of cycle ergometry. Med Sci. Sports Exerc. 18225-230,1986. 5. CONLEY, D. L., and G. S. KRAHENBUHL. Running economy and distance nmning performance of highly trained athletes. Med. ScL Sports Exerc. 12:357-360,1980. 6. COYLE, E. F., M. E. FELTNER, S. A. KAUTZ, et a). Physiological and biomechankal factors associated with elite endurance cycling performance. Med. ScL Sports Exerc. 23:93-107,1991. 7. COYLE, E. F., L. S. SIDOSSB, J. F. HOROWITZ, and 1. D. BBLTZ. Cycling efficiency is related to the percentage of type I muscle fibers. Med Sci. Sports Exerc. 24:782-788,1992. 8. G IESSER, G. A., and D. C. POOLS. The slow component of oxygen uptake kinetics in humans. Exerc. Sport Sci Rev. 2435-71,1996. 9. HOROWITZ, J. F., L. S. Smossis, and E. F. COYLE. High efficiency of type I muscle fibres improves performance, hi. J. Sports Med. 15:152-157, 1994. 10. IVY, J. I~, M-Y. Cm, C. S. HINTA W. M. SHERMAN, R. P. HELLBNDALL, and O. H. LOWRY. Progressive metabolic changes in individual muscle fibers with increasing work rates. Am. J. Physio!. 252(Ce/i Phyaioi 21):C630-C639, 1987. 11. JONES, A. M., and H. CARTER. The effect of endurance training on parameters of aerobic fitness. Sports Med. 29:373-386,2000. 12. LONDEREE, B. R. The use of laboratory test results with long distance runners. Sports Med. 3:201-213, 1986. 13. LI-CIA, A., I. HOYOS, and J. L. CHICHAHRO. Preferred pedalling cadence in professional cycling. Med Sci. Sports Exerc. 33:13611366,2001. 14. LLTIA, A., J. HOYOS, A. SANTALLA, M. PEREZ and J. L. CHICHARRO. Kinetics of VO2 in professional cyclists. Med. Sc. Sports Exerc. 34 326-331, 2002. 15. LL-OA, A, J. PARDO, A. DURANTEZ, J. HOYOS, and J. L. GDCHARRO. Physiological differences between professional and elite road cyclists. Int. J. Sports Med. 19:342-348, 1998.
Official Journal of the American Cotega of Sports Medidrra
16. LUSK, G. The Elements of the Science of Nutrition, 4th Ed Philadelphia: W. B. Sawders, 1928, pp. 400-446. 17. MORGAN, D. W., D. R. BKANSFORD, and D. L. COSTTLL. Variation in the aerobic demand of running among trained and untrained subjects. Med. ScL Sports Exerc. 27:404-409,1995. 18. MORGAN, D. W., and 3. T. DANIELS. Relationship between VOjmax and the aerobic demand of running in elite distance runners. Int. J. Sports Med. 15:426-429,1994. 19. MOSELEY, L., and A. E. JEOKENDRUP. The reliability of cycling efficiency. Med. ScL Sports Exerc. 33:621-627, 2001. 20. NICKLEBERRY, B. L. Jr., and G. A. BROOKS. No effect of cycling experience on leg cycle ergomeler efficiency. Med Sci. Sports Exerc. 28:1396-1401. 1996. 21. NOAKES, T. D. Implications of exercise testing for prediction of athletic performance: a contemporary perspective. Med Sci. Sports Exerc. 20:319-330, 1988. 22. PAMLLA, S., I. MVJKA, G. CUESTA, and J. I. GOOUEVA. Level ground and uphill cycling ability in professional road cycling. Med Sci Sports Exerc. 31:878-8853, 1999. 23. PATTE, R. R., C. A. MACERA, S. P. BAILEY, W. P. BARTOU, and K. B. POWELL. Physiological, antfaropometrk, and training correlates of nmning economy. Med ScL Sports Exerc. 24:1128-1133, 1995. 24. POOLE, D. C., and L. C HENSON. Effect of acute caloric restriction on work efficiency. Am. J. din. Nutr. 47:15-18, 1988. 25. PORSZASZ, J., T. J. BARSTOW, and K. W. WASSERMAN. Evaluation of a symmetrically disposed Pilot tube flowmeler for measuring gas flow during exercise, J. Appl Physiol. 77:2659-2665,1994. 26. SALTDJ, B., C. K. KM, N. TERRADOS, H. LARSEN, J. SVEDENHAG, and C. ROLF. Morphology, enzyme activities and buffer capacity in leg muscles of Kenyan and Scandinavian runners. Scand. J. Med. ScL Sports 5:222-230, 1995. 27. SARIS, W. H. M., J. M. G. SENDSN, and F. BROUNS. What is a normal red-blood cell mass for professional cyclists? Lancet 352: 1758,1998. 28. SMMOHARA, M, and T. MORITANI. Increase in neuromuscular activity and oxygen uptake during heavy exercise. Ann. Physiol AnthropoL 11:257-262, 1992. 29. SKINNER, J. S., and T. H, MCLELLAN. The transition from aerobic to anaerobic metabolism. Res. Q. Exerc. Sport 51:234-248, 1980. 30. WESTON, A. R., Z. MBAMBO, and K. H. MYBURGH. Running economy of African and Caucasian distance runners. Med. Sci. Sports Exerc. 32:1130-1134,2000.