Familial aggregation of submaximal aerobic performance in... : Medicine & Science in Sports & Exercise (original) (raw)
Cardiorespiratory fitness and its response to regular exercise are characterized by marked interindividual differences (2,12,20). Results from twin and family studies suggest that genetic factors are important in determining this variability (5). Most studies on the heritability of cardiorespiratory fitness were based on twin data and used maximal oxygen uptake (V̇O2max) as a phenotype. One of the studies with the largest sample size and comprising 436 pairs of monozygotic (MZ) and 622 pairs of dizygotic (DZ) twins reported an MZ intraclass correlation of 0.62 for predicted V̇O2max compared with 0.29 for DZ twins (21), suggesting an heritability of more than 60% for V̇O2max. Another twin study based on smaller number of pairs (29 MZ pairs and 19 DZ pairs) but with a direct measure of V̇O2max and after adjustment for body weight, body fat, and sports participation reported an heritability of 66%(8). The most recent twin study was based on a sample of 105 10-yr-old twins, and their parents reported a heritability above 65% for V̇O2max(13), but this estimate is difficult to interpret as V̇O2max was not adjusted for body mass. The first genetic studies of V̇O2max based on family data (11,14) suggested that about 40% of the variance in V̇O2max could be accounted for by genetic factors. More recently, HERITAGE Family study reports have shown that V̇O2max in the sedentary state (4) and in response to a standardized endurance training program (3) were characterized by a significant familial resemblance with estimates of heritability reaching about 50% of the phenotypic variance.
Compared with V̇O2max, relatively few studies have investigated the role of genetic factors in submaximal aerobic performance and its response to exercise training. Some family data from the Quebec Family Study (16,17) and from a nationally representative sample of the Canadian population (10,15) suggested that physical working capacity measured at a heart rate of 150 beats·min−1 was characterized by a significant familial resemblance that was mainly accounted for by shared familial environmental factors. Studies with monozygotic (MZ) twins trained under standardized endurance cycle exercise programs for periods of 15 or 20 wk revealed that O2 consumption changes measured at a given submaximal power output was characterized by a significant within MZ twin pair resemblance (5), suggesting that genetic factors are involved in the trainability of these phenotypes. To our knowledge, the role of genetic factors in determining the response of submaximal aerobic performance to exercise training was not investigated in any population-based family study. Thus, the purpose of the present study was to determine whether or not submaximal exercise capacities and their responses to 20 wk of endurance training were characterized by a significant familial resemblance and assess the heritability of the corresponding phenotypes using data from the HERITAGE Family study.
METHODS
Sample.
Subjects of the HERITAGE Family study were used for the purpose of this study. The HERITAGE study is a multicenter study designed to investigate the effects of regular exercise on several cardiovascular disease and diabetes risk factors and to determine the role of genetic factors in the cardiovascular, metabolic, and hormonal adaptations to exercise training. The specific aims, design, and measurements of the study have been described in detail elsewhere (6).
For the present study, a total of 483 whites from two-generation families (184 parents and 299 biological offspring) and ranging in age from 17 to 65 yr were available. Subjects were required to be sedentary at baseline, defined as engaging in no regular physical activities over the previous 6 months, and to be free of any condition or disease that could be aggravated by exercise training. Obese individuals (BMI > 40 kg·m−2) were excluded because of potential metabolic abnormalities and exercise difficulties, unless they were able to meet the demands of the training program as judged by a physician. Individuals with resting blood pressure greater than 159 mm Hg for systolic and/or greater than 99 mm Hg for diastolic or those on antihypertensive medications were also excluded. More details about exclusion criteria can be found in Bouchard et al. (6).
Training protocol.
Subjects trained on cycle ergometers three times a week for 20 wk using a standardized protocol. Subjects worked at a heart rate (HR) corresponding to 55% of their baseline maximal V̇O2 (V̇O2max) for 30 min per session at the beginning. The intensity or duration of the training program were adjusted every 2 wk until the 14th week, at which time subjects trained at a HR associated with 75% of their baseline V̇O2max for 50 min during the remaining of the training protocol. Training intensities were adjusted individually by a computer system recording training data and automatically adjusting the power output (PO) of the cycle ergometer to keep each subject’s heart rate within 5 beats of the programmed heart rate during the training sessions. Details about the training program can be found elsewhere (20).
Exercise tests.
Three exercise tests were conducted on separate days, before and after training, on SensorMedics ErgoMetrics 800S cycle ergometers (Yorba Linda, CA). These tests, described in detail elsewhere (20), were done at about the same time of the day and with at least 48 h between two tests. First, subjects completed a maximal exercise test. The subjects started pedaling at a PO of 50 W for 3 min with increases of 25 W every 2 min thereafter until volitional exhaustion. Second, subjects performed a submaximal exercise test during which they exercised 8–12 min at an absolute PO of 50 W and for 8–12 min at a relative PO equivalent to 60% of their initial V̇O2max. Finally, a submaximal/maximal exercise test was performed, starting with the same protocol as in the submaximal test and then followed by 3 min of exercise at 80% of the subject’s initial V̇O2max. The resistance was then increased to the highest PO attained in the first test and by 25 W every 2 min thereafter until exhaustion.
Gas exchange variables (V̇O2, V̇CO2, V̇E, and RER) were recorded using a SensorMedics 2900 metabolic measurement cart throughout each exercise test and reported as the rolling average of the last three 20-s intervals of each exercise stage. The criteria for V̇O2max were: RER > 1.1, plateau in V̇O2 (changes of < 100 mL·min in the last three 20-s intervals) and an HR within 10 beats·min−1 of the maximal HR predicted by age. HR was monitored with an electrocardiogram and values were recorded during the last 15 s of each exercise stage of the maximal test and once steady state was achieved at each of the submaximal work rates during the submaximal and submaximal-maximal tests. All subjects achieved a V̇O2max by one of these criteria in at least one of the two maximal exercise tests, both pre- and post-training. The submaximal exercise phenotypes that were used in the present study were V̇O2 at 50 W (V̇O250W), 60% (V̇O260%), and 80% (V̇O280%) of V̇O2max and power outputs (W) at 60% (PO60%) and 80% (PO80%) of V̇O2max. Because duplicate measures were available at 50 W and at 60% of V̇O2max, the average of the two measures were taken as values for V̇O250W, V̇O260%, and PO60%. These submaximal power output levels were selected because we wanted both a measure of O2 consumption in a 10- to 12-min steady state at an absolute power output level (50 W) that could be sustained by all participants and also higher intensity power output levels which, in this case, were defined in terms of percentages of V̇O2max. The response to training was computed as the difference (Δ) between the posttraining and baseline measurements of the same measures. A paired _t_-test on the response scores was used to test for the effects of endurance training.
Data adjustments.
Baseline phenotypes were adjusted for the effects of age, sex, and body mass, whereas response phenotypes were adjusted for the effects of age, sex, and baseline value. These adjustments were performed within each of the four sex by generation groups by using a stepwise multiple regression procedure retaining only those terms that were significant at the 5% level. The phenotypes used in the genetic analysis were the residuals from the regression standardized to a 0 mean and a standard deviation of 1. In general, the effects of age and body mass were significant in all groups for submaximal oxygen consumption (V̇O250W, V̇O260%, and V̇O280%) measured at baseline, with about 14–54% of the variance accounted for, whereas for PO60% and PO80% only the age effects were significant (9–21%). For the response phenotypes, age terms were generally not significant, whereas baseline values accounted for 16–23% of the variance in V̇O250W and 3–12% for the other response phenotypes.
Familial correlation model.
Familial aggregation in the baseline and response phenotypes was investigated by computing familial correlations. An ANOVA comparing the between- and the within-family variances was first performed to test whether or not the phenotypes aggregate in families. The familial correlation model was based on four groups of individuals {fathers (f), mothers (m), sons (s), and daughters (d)}, giving rise to eight interindividual correlations in three familial classes {1 spouse (fm), 4 parent-offspring (fs, fd, ms, md), and 3 sibling (ss, dd, sd)}. The maximum likelihood computer program SEGPATH (18) fitted the model directly to the family data under the assumption that the phenotypes within a family jointly follow a multivariate normal distribution. A general model with all eight familial correlations and several reduced models (see Appendix A) testing specific null hypotheses were fitted to the data. Three broad classes of reduced models were considered. First, null hypotheses on sex and/or generation differences in the familial correlations were tested, including no sex differences in the offspring, no sex differences in parents or offspring, and no sex nor generation differences. Second, null hypotheses testing the nature and strength of the familial resemblance were also conducted by familial class, including no sibling resemblance (model 5: ss = dd = sd = 0), no parent-offspring resemblance (model 6: fs = fd = ms = md = 0), no spouse resemblance (model 7: fm = 0), and a environmental model where all eight correlations were equated. Third, five models of maternal inheritance (where mother-offspring and sibling correlations are expected to be equal, i.e., ms = md = sd = ss = dd), with or without restrictions about the father’s contribution, were tested. In model 9, the hypothesis of maternal inheritance without any assumption about the father’s contribution was tested. Maternal inheritance was also tested under the constraint that the father-offspring correlations were independent of sex (fs = fd, ms = md = sd = ss = dd), that the father’s contribution is environmental rather than genetic (fm = fs = fd, ms = md = sd = ss = dd), that the father’s contribution is not significant (fm = fs = fd = 0, ms = md = sd = ss = dd), and that the father-mother correlation is not significant (fm = 0, fs = fd, ms = md = sd = ss = dd). These reduced models were tested against the general model using the likelihood ratio test, which is approximately distributed as a χ2, with the degrees of freedom being the difference in the number of parameters estimated in the two competing hypotheses. The most parsimonious model was derived by combining all nonrejected null hypotheses. In addition to the likelihood ratio test, Akaike’s (1) Information Criterion (AIC), which is −2 ln L plus twice the number of estimated parameters, was used to judge the fit of nonnested models. The most parsimonious model by AIC is the one with the smallest value. Maximal heritabilities were computed from the maximum-likelihood estimates of the familial correlations obtained under the most parsimonious model as follows (19) :
This heritability estimate includes both genetic and nongenetic sources of variance and is adjusted for the degree of spouse resemblance. For the maternal inheritance models, the maximal maternal heritability was based on the common mother-offspring and sibling correlations.
RESULTS
Table 1 presents the sample sizes, means, and standard deviations for age, body mass, and submaximal exercise capacity before and after training, separately in each of the sex and generation groups (fathers, mothers, sons, and daughters). Endurance training resulted in significant (P < 0.0001) improvements of all indicators of submaximal exercise capacity in parents and offspring of both sexes.
Descriptive statistics of the sample at baseline and after training in each of the sex and generation groups.a
The results of the ANOVA performed to test the significance of the familial aggregation indicated significant familial resemblance for all indicators of submaximal exercise capacity and their responses to endurance training (results not shown). There were about 2–5 times more variance between families than within families for the baseline phenotypes, with 38–56% of the variance accounted for by family lines. The response to exercise training was also characterized by a significant familial resemblance which accounted for 30–43% of the variance in the response phenotypes, independent of the baseline values.
The model fitting results for the baseline and the response phenotypes are presented in appendix B. The results for indicators of submaximal V̇O2 and POs at baseline indicate significant parent-offspring and sibling correlations for all indicators of submaximal working capacity, whereas the spouse correlation was significant only for V̇O250W, PO60% and PO80%. Except for ΔV̇O250W and ΔV̇O280%, the model testing maternal inheritance (model 8) could not be rejected, suggesting that familial resemblance could be partly attributable to specific maternal genetic and/or environmental factors.
The familial correlations under the general, the most parsimonious and the best maternal inheritance models (when significant and different from the most parsimonious model) are presented in Tables 2 and 3 for baseline and response phenotypes, respectively. The maximal heritabilities derived from the most parsimonious model and the best maternal inheritance model (where ms = md = sd = ss =dd) are presented in Figure 1. For all baseline phenotypes except V̇O250W, the best maternal inheritance model was the most parsimonious model. Heritability estimates were 70% and 48% for V̇O250W; 52% and 30% for V̇O260%; 48% and 29% for V̇O280%; 68% and 38% for PO60%; and 74% and 44% for PO80%, under the most parsimonious and the best maternal inheritance models, respectively. The response phenotypes were characterized by lower heritabilities (bottom panel, Fig. 1) than the baseline values. Heritability estimates were 57% and 0% for V̇O250W; 23% and 15% for V̇O260%; 44% and 0% for V̇O280%; 33% and 19% for PO60%; and 45% and 26% for PO80%, under the most parsimonious and the best maternal inheritance models, respectively.
Maximum likelihood estimates of familial correlations (± standard error) under the general and most parsimonious models for baseline phenotypes.a
Maximum likelihood estimates of familial correlations (± standard error) under the general and most parsimonious models for the response phenotypes.a
Maximal and maternal heritabilities of submaximal oxygen consumption and power outputs in baseline (top) and in response to endurance training (bottom) derived from the most parsimonious and the best maternal inheritance models, respectively.
DISCUSSION
The results of the present study suggest that submaximal exercise V̇O2 and PO in sedentary subjects strongly aggregate in families. The changes in the same phenotypes brought about by 20 wk of endurance training, after adjustment for pretraining levels, were also characterized by significant familial resemblance. The heritability estimates derived from the familial correlations computed from spouses, parent-offspring, and siblings ranged from 48% to 74% for the baseline phenotypes and from 23% to 57% for the response phenotypes. These estimates reflect the contribution of both genetic and shared familial environmental factors as, in most cases, the spouse correlations were significant.
The familial correlation model used in the present study cannot distinguish between the contributions of genetic and familial environmental factors in the familial resemblance. For this reason, the heritability estimates presented in Figure 1 are considered as “maximal heritabilities” because they could result from the transmission of both genetic and familial environmental factors. However, the pattern of correlations observed among spouses, parent-offspring, and siblings can be used to make inferences about the relative importance of genetic versus nongenetic factors in the heritabilities. Thus, a pattern of significant parent-offspring and sibling correlations with no spouse correlation would suggest that the familial resemblance is primarily attributable to genetic factors, whereas the presence of significant spouse correlations in addition to the parent-offspring and sibling correlations would suggest the contribution of shared family environment in addition to genetic factors. The latter pattern of familial correlations was observed for all baseline phenotypes. Indeed, we found that the spouse correlations were significant and equal to the parent-offspring correlations, ranging from 0.12 for V̇O280% to 0.45 for V̇O250W. The sibling correlations were higher than the spouse and parent-offspring correlations (except for V̇O250W), ranging from 0.29 to 0.44.
This pattern of familial correlations is similar to the one reported in two other family studies. In the Quebec Family Study (QFS), the PO measured at a heart rate of 150 beats·min−1 and expressed per kg of body weight (PWC150/kg) exhibited significant familial resemblance with significant spouse (0.21), parent-offspring (0.14), and sibling (0.25) correlations (16). In another study performed with the same population, maximal heritability, which included the transmission of both genetic and nongenetic factors, reached 22%(17). Familial aggregation of submaximal PO derived from a step test was also investigated in a large sample of the Canadian population involving 13,804 subjects who participated in the 1981 Canada Fitness Survey (15). Familial correlations computed for PWC150·kg−1 reached 0.17, 0.17, and 0.26 for spouses, parent-offspring, and sibling pairs, respectively. Analysis of these correlations with a path model, assuming the transmission of both genetic and environmental factors from parent to offspring (without the possibility to distinguish between them), translated into a maximal heritability of 28%(15).
The endurance training program resulted in significant improvements in submaximal performance. The V̇O2 at 50 W was reduced by 10–15% on average, whereas it was significantly increased at the relative POs of 60% (12–16%) and 80% (14–18%). Changes induced by endurance training and adjusted for pretraining values were also characterized by a significant familial resemblance, with maximal heritabilities ranging from 23% to 57% (Fig. 1, bottom panel). These findings suggest that trainability of submaximal working capacities is strongly influenced by familial and/or genetic factors. For V̇O280% and PO80%, the spouse correlation was not significant, suggesting that genetic factors may be more important than familial environmental factors in determining the trainability of submaximal exercise performed at a high intensity. Other evidence that genetic factors could be involved in determining the trainability of submaximal aerobic performance comes from twin studies. In a study of six pairs of monozygotic (MZ) twins, total PO during a 90-min ergocycle exercise test was measured before and after 15 wk of endurance training. A highly significant within-pair resemblance was observed for the training gains, with the intraclass coefficient for twin resemblance reaching 0.83 (9). In another study, seven pairs of male MZ twins exercised twice daily while being kept on a constant daily energy intake for 3 months (7). The exercise training protocol resulted in significant improvements in submaximal V̇O2 and in the heart rate measured at fixed POs. Again, changes were characterized by a significant within-pair resemblance. For the changes in V̇O2 measured at 50 W in the latter study, there were 15 times more variance between pairs than within pairs and the intraclass correlation for the resemblance in the response reached 0.87 (7). The results of these twin studies, along with those of the present family study, strongly suggest that the trainability of submaximal exercise capacities is influenced by genetic factors.
Despite similar trends in the familial correlations, the maximal heritabilities reported in this study are higher than those reported in the few other family studies of submaximal working capacities. Besides differences in the analytical strategies used in these studies that could account for some of the differences in the heritability estimates, it is important to keep in mind that the subjects of the present study had to be sedentary. A stringent control over initial physical activity, an important environmental determinant of interindividual differences in submaximal aerobic performances, probably contributed to a reduction of the phenotypic variance and thus to an increase in the heritabilities.
The analytical approach used to investigate the familial resemblance of submaximal aerobic performance in the present study allowed us to specifically test different models of maternal inheritance in which the mother-offspring and sibling correlations were forced to be equal. Our results showed that all submaximal aerobic performance phenotypes measured at baseline could be characterized by a significant maternal component, with the best maternal inheritance model assuming that the father’s contribution is environmental rather than genetic (model 10 in Appendix B). To the best of our knowledge, this is the first study to show a significant contribution of maternal inheritance for submaximal aerobic performance phenotypes, as such evidence was not found in the other two family studies for PWC150·kg−1(15,16). However, a study undertaken in the same cohort of families found a significant maternal heritability for V̇O2max(4). Although the molecular basis of this maternal inheritance is not known, possible mechanisms include 1) contribution of genes from mitochondrial DNA (which are transmitted from the mother only because all mitochondria in a fertilized egg come from the egg only); 2) in utero maternal effects not driven primarily by mitochondrial DNA; 3) a genetic effect expressed only when the gene is inherited from the mother; or 4) a maternal cultural transmission.
Heritabilities reported in the present study for submaximal working capacities and their responses to training are slightly higher than those reported recently for V̇O2max in the same population (4). Using the same familial correlation model as the one used here, the maximal heritability of V̇O2max adjusted for age and body mass was 52%. As shown in Figure 1 (top panel), the heritability of V̇O2 measured at a relative PO of 60% and 80% of V̇O2max is very similar to the one reported for V̇O2max in the study of Bouchard et al. (4) and accounts for about 50% of the phenotypic variance after adjustment for age and body mass. However, V̇O2 measured at low intensity exercise appears to be characterized by a stronger genetic effect with a maximal heritability reaching 70% for V̇O250W. The smaller phenotypic variance of V̇O250W, which was found to be more influenced by age, sex, and body mass differences than the other baseline phenotypes, could explain the higher heritability of this phenotype.
In summary, results of the present study reveal that submaximal exercise capacities measured in sedentary subjects before and in response to endurance training aggregate in families and that genetic and other familial factors contribute to this familial resemblance. We estimated the maximal heritabilities of the phenotypes to range from about 50–70%. Moreover, we showed the contribution of specific maternal effects in the heritability of these traits, with maternal heritabilities ranging from about 15–48%. Finally, we showed that the trainability of submaximal aerobic performance, independent of the pretraining value, is influenced by genetic factors, confirming the results reported for V̇O2max in the same cohort of families.
The HERITAGE study is supported by the NHLBI through the following grants: HL45670 (C. Bouchard, PI), HL47323 (A. S. Leon, PI), HL47317 (D. C. Rao, PI), HL47327 (J. S. Skinner, PI), and HL47321 (J. H. Wilmore, PI). Arthur Leon is partially supported by the Henry L. Taylor endowed Professorship in Exercise Science and Health Enhancement. Claude Bouchard is partially supported by the George A. Bray Chair in Nutrition. Thanks are expressed to all the co-principal investigators, investigators, co-investigators, local project coordinators, research assistants, laboratory technicians, and secretaries who have contributed to the study.
Address for correspondence: Louis Pérusse, Ph.D., Physical Activity Sciences Laboratory, Division of Kinesiology, PEPS, Laval University, Québec G1K 7P4, Canada; E-mail: [email protected].
APPENDIX A
APPENDIX B
FIGUREFIGUREFIGUREFIGUREFIGURE
Continued.
Continued.
Continued.
Continued.
REFERENCES
1. Akaike, H. A new look at the statistical model identification. IEEE Trans. Automat. Cont. 19: 716–723, 1974.
2. Bouchard, C. Heredity and trainability. In:Sports Medicine for the Mature Athlete, J. R. Sutton, and M. R. Brock, editors. Indianapolis, IN: Benchmark Press Inc., 1986, pp. 81–89.
3. Bouchard, C., P. An, T. Rice, et al. Familial aggregation of VO2 max response to exercise training: results from the HERITAGE Family Study. J. Appl. Physiol. 87:1003–1008, 1999.
4. Bouchard, C., E. W. Daw, T. Rice, et al. Familial resemblance for VO2max in the sedentary state: the HERITAGE family study. Med. Sci. Sports Exerc. 30: 252–258, 1998.
5. Bouchard, C., F. T. Dionne, J. A. Simoneau, and M. R. Boulay. Genetics of aerobic and anaerobic performances. Exerc. Sport Sci. Rev. 20: 27–58, 1992.
6. Bouchard, C., A. S. Leon, D. C. Rao, J. S. Skinner, J. H. Wilmore, and J. Gagnon. The HERITAGE Family Study: aims, design, and measurement protocol. Med. Sci. Sports Exerc. 27: 721–729, 1995.
7. Bouchard, C., A. Tremblay, J-P. Després, et al. The response to exercise with constant energy intake in identical twins. Obes. Res. 2: 400–410, 1994.
8. Fagard, R., E. Bielen, and A. Amery. Heritability of aerobic power and anaerobic energy generation during exercise. J. Appl. Physiol. 70: 357–362, 1991.
9. Hamel, P., J-A. Simoneau, G. Lortie, M. R. Boulay, and C. Bouchard. Heredity and muscle adaptation to endurance training. Med. Sci. Sports Exerc. 18: 690–696, 1986.
10. Katzmarzyk, P. T., L. Pérusse, D. C. Rao, and C. Bouchard. Familial risk ratios for high and low physical fitness levels in the Canadian population. Med. Sci. Sports Exerc. 32: 614–619, 2000.
11. Lortie, G., C. Bouchard, C. Leblanc, et al. Familial similarity in aerobic power. Hum. Biol. 54: 801–812, 1982.
12. Lortie, G., J. A. Simoneau, P. Hamel, M. R. Boulay, F. Landry, and C. Bouchard. Responses of maximal aerobic power and capacity to aerobic training. Int. J. Sports Med. 5: 232–236, 1984.
13. Maes, H. H., G. P. Beunen, R. F. Vlietinck, et al. Inheritance of physical fitness in 10-yr-old twins and their parents. Med. Sci. Sports Exerc. 28: 1479–1491, 1996.
14. Montoye, H. J., and R. Gayle. Familial relationships in maximal oxygen uptake. Hum. Biol. 50: 241–249, 1978.
15. Pérusse, L., C. Leblanc, and C. Bouchard. Inter-generation transmission of physical fitness in the Canadian population. Can. J. Sport Sci. 13: 8–14, 1988.
16. Pérusse, L., C. Leblanc, A. Tremblay, et al. Familial aggregation in physical fitness, coronary heart disease risk factors, and pulmonary function measurements. Prev. Med. 16: 607–615, 1987.
17. Pérusse, L., G. Lortie, C. Leblanc, A. Tremblay, G. Thériault,and C. Bouchard. Genetic and environmental sources of variation in physical fitness. Ann. Hum. Biol. 14: 425–434, 1987.
18. Province, M. A., and D. C. Rao. General purpose model and a computer program for combined segregation and path analysis (SEGPATH): automatically creating computer programs from symbolic language model specifications. Genet. Epidemiol. 12: 203–219, 1995.
19. Rice, T., J. P. Després, E. W. Daw, et al. Familial resemblance for abdominal visceral fat: the HERITAGE family study. Int. J. Obes. 21: 1024–1031, 1997.
20. Skinner, J. S., K. M. Wilmore, J. B. Krasnoff, et al. Adaptation to a standardized training program and changes in fitness in a large, heterogeneous population: the HERITAGE Family Study. Med. Sci. Sports Exerc. 32: 157–161, 2000.
21. Sundet, J. M., P. Magnus, and K. Tambs. The heritability of maximal aerobic power: a study of Norwegian twins. Scand. J. Med. Sci. Sports 4: 181–185, 1994.
Keywords:
OXYGEN UPTAKE,; POWER OUTPUT,; EXERCISE TRAINING,; GENETICS,; HERITABILITY
© 2001 Lippincott Williams & Wilkins, Inc.