Aerobic parameters of exercise as a function of body size during growth in children

Aerobic parameters of exercise as a function of body size during growth in children. J. Appl. Physiol.: Respirat. Environ. Exercise Physiol. 56(3): 628-634, 1984.-To examine the relationship between body weight in children and aerobic parameters of exercise, we determined the anaerobic threshold (AT), maximum O2 uptake (iro2,,,), work efficiency, and response time for O2 uptake (RT-~o~) in 109 healthy children (51 girls and 58 boys, range 6-17 yr old) using a cross-sectional study design. Gas exchange during ex- ercise was measured breath by breath. The protocol consisted of cycle ergometry and a linearly increasing work rate (ramp) to the limit of the subject’s tolerance. Both AT and Vogmax increased systematically with body weight, whereas work efficiency and RT-vos were virtually independent of body size. The ratio of AT to \j02max decreased slightly with age, and its mean value was 60%. AT scaled to body weight to the power of 0.92, not significantly different from the power of 1.01 for QO 2max. Thus both the AT and the TjO2max increase in a highly ordered manner with increasing size, and as judged by AT/ the onset of anaerobic metabolism during exercise occurred at a relatively constant proportion of the overall limit of the gas transport system. We conclude that in children cardiorespiratory responses to exercise are regulated at optimized values despite overall change in body size during growth.

-To examine the relationship between body weight in children and aerobic parameters of exercise, we determined the anaerobic threshold (AT), maximum O2 uptake (iro2,,,), work efficiency, and response time for O2 uptake (RT-~o~) in 109 healthy children (51 girls and 58 boys, range 6-17 yr old) using a cross-sectional study design. Gas exchange during exercise was measured breath by breath. The protocol consisted of cycle ergometry and a linearly increasing work rate (ramp) to the limit of the subject's tolerance. Both AT and Vogmax increased systematically with body weight, whereas work efficiency and RT-vos were virtually independent of body size. The ratio of AT to \j02max decreased slightly with age, and its mean value was 60%. AT scaled to body weight to the power of 0.92, not significantly different from the power of 1.01 for QO 2max. Thus both the AT and the TjO2max increase in a highly ordered manner with increasing size, and as judged by AT/ VO 2 max9 the onset of anaerobic metabolism during exercise occurred at a relatively constant proportion of the overall limit of the gas transport system. We conclude that in children cardiorespiratory responses to exercise are regulated at optimized values despite overall change in body size during growth.
anaerobic threshold; work efficiency; response time for oxygen uptake; maximum oxygen uptake; scaling; ramp exercise protocol THE PURPOSE of this study was to examine metabolic rates during exercise as a function of the changes in body weight accompanying growth in children. We measured four aerobic parameters of exercise in a cross-sectional study of normal children. One parameter, the maximum uptake (VOW max), often has been used to assess the aerobic response to exercise in children (1,6,7,10). However, \jozmax indicates only one aerobic factor: the limit of the organism's ability to utilize atmospheric 02 for cellular energetics.
Three other important aerobic factors are the anaerobic threshold (AT), the highest metabolic rate for which the energy requirements may be obtained solely from 02 uptake, and, hence, without concomitant anaerobiosis (28)(29)(30)(31); the work efficiency, the organism's proportional aerobic cost of mechanical work; and the response time of O2 uptake at the onset of exercise (RT-VO& the temporal coupling between the metabolic demand of exercise and the response dynamics of the gas transport system (i.e., the sum of the response time constant and any inherent system delays). These four aerobic parameters can be obtained noninvasively from a single brief exercise protocol developed in this laboratory (32). Neither the AT nor the RT-VOW have been measured in large numbers of normal children.

METHODS
Population. All subjects were volunteers obtained through local schools, community organizations, and children of the hospital staff. We excluded obese children, children with a history of chronic disease of any organ system, or children who were not allowed to participate in normal physical education programs at school. No attempt was made to select subjects who were particularly active; i.e., we did not recruit through physical education or sports programs.
One hundred nine children (58 boys and 51 girls, range 6-17 yr old) comprised the study population.
One hundred fourteen children were originally tested; however, five were excluded because the AT could not be determined by gas exchange techniques. Pubertal ratings were not done on the children, but for certain of our analyses, the subjects were divided into the four following age groups, which corresponded to the onset of puberty based on recent studies in boys and girls (19): 1) girls 6-11 yr old (younger girls); 2) girls 12-17 (older girls); 3) boys 6-13 (younger boys); and 4) boys 14-17 (older boys). Age, weight, and height profiles of our study population are provided in Table 1. All children were within the normal range for height and weight by reference tables of the National Center for Health Statistics. The children were predominantly of the middle socioeconomic class. Eighty-six percent of the subjects were Caucasian; the remainder consisted of Oriental, Hispanic, and black children. This project was approved by the Human Subjects Committee of Harbor-UCLA Medical Center. Informed consent was obtained from each child and guardian before participation.
Exercise protocol. Height and weight were measured before testing. The subjects were told that they would be doing one hard exercise test on a special bicycle, and that they would feel like they were riding up a steep hill. The protocol consisted of a ramp forcing function (32) utilizing an electronically braked cycle ergometer (Godart). Subjects began by cycling at 0 W (unloaded) work rate with a minimum of 3 min of a warm-up phase. For children 8 yr old and above, the warm-up period was 4 min. Work rate was then continuously incremented in a linear ramp pattern. An example of a ramp protocol is shown in Fig. 1. The slope of the ramp (the increase in work rate per minute) was determined so that the ramp would be greater than 5-and less than 12-min duration. For each child, a suitable slope was chosen based on our own initial studies. In general, for children from 6 to 9 yr of age, the ramp was 10 W/min; from 10-13, 15 W/min; and from M-17, 20 W/min. For certain adolescents who were deemed to be quite fit by history, ramp slopes as high as 40 W/min were chosen. The mean time for the ramp (not counting the warm up) was 9 min. The children were instructed to raise their hand when they could not continue, and on this signal, the work rate was reduced to 0 W. The children were actively encouraged periodically throughout the test.
The children were instructed to maintain as constant a pedaling rate as possible between 50 and 70 rpm. A pedaling rate meter was in full view of each subject, and VO, max Breath-by-breath response of gas exchange to ramp exercise protocol in Yyr-old boy. After a period of O-W pedaling, work rate increases (A WR) in linear manner. Note lag in response of O2 uptake after WR has started (RT-VO&, linear increase of 30, as WR increases, and plateau of vo2 despite an increasing WR (vozmar). Anaerobic threshold (AT) is determined by identifying vo2 where respiratory exchange ratio (R), ventilatory equivalent for O2 (7j,/irOz), and end-tidal 02 (PET~J abruptly increase while ventilatory equivalent for CO2 ( a metronome could be activated at the discretion of the investigator present. A servomechanism in the electronic braking system of the ergometer maintained the input work to an accuracy of 1% within a range of pedaling rates of 50-90 rpm. For the smaller children, the three following modifications were made in the ergometer itself: 1) an adj ustable seat designed to sit lower than the no Irma1 seat was used; 2) the handlebars were lowered; and 3) special attachments were used to effectively shorten the pedal cranks by 25%. We could, therefore, "fit" the ergometer to the size of the subject so that he or she was comfortable.
Measurement of gas exchange. The subjects breathed through low-resistance valves (Hans-Rudolph).
For children less than 12 yr old, a 40 -ml dead-space valv 'e was used; for those 12 and above, a go-ml dead-space valve was used. Inspiratory and expiratory airflow were measured by twd pneumotachographs (Fleisch no. 3) attached to the inspiratory and expiratory ports of the breathing valves and by two variable-reluctance manometers (Validyne, MP45) . The expiratory pneumotachograph was maintained at a constant temperature of 37°C by a thermal feedback device. This system was calibrated before each session by inputting known volumes of room air at various mean flows and flow profiles. Respired POP and PCO~ were determined by mass spectrometry (Perkins-Elmer MGA 1100) from a sample drawn continuously from the mouthpiece at 1 ml/s. Precision-analyzed gas mixtures were used for calibration of the mass spectrometer. The system was found to be stable throughout the period of the study (3,22).
The electrical signals from these devices underwent analog-to-digital computation (Hewlett-Packard model 1050) for the on-line breath-to-breath determination of O2 uptake (v+, STPD), Con output (ko2, STPD), expired ventilation (VE, BTPS), respiratory exchange ratio, ventilatory equivalentfor 02 and for COZ, respectively, (vE/ vo2, irE/TiCO,), and end-tidal partial pressures of 02 and C02, respectively, (PETE,, PET&. The data from each test were displayed on line (Beckman R711 dynagraph) and stored on digital tape for subsequent analysis. The breath-by-breath data of each subject were interpolated at l-s intervals, and moving averages could be obtained for smoothing of breath variation. studies with a great deal of breath-to-Analysis of Gas Exchange. Data Maximal 02 uptake. We took the VOzmax as the largest vo2 achieved by the subject. We distinguished between these values and the "true" \io, max, defined as an interval of at least 20 s in which Vo2 remained constant or decreased, despite an increasing work rate. C Anaerobic threshold. AT was determined from gas exhange data by finding the Vo2 at which VE increased out 0 If proportion to the increase in VOW. This is done most accurately by finding where VE/~O~ and PETE, increase (hyperventilation with respect to 02) without an increase in i7E/i7COz or a decrease in PERo,. Hyperventilation with respect to 02 without concomitant hyperventilation for CO2 only occurs during buffering of a metabolic acid by HCO:. Other forms of hyperventilation should cause PET~~ to increase and PET~~~ to decrease while vE/v02 and vE/irCOz increase together (23,32). In five subjects of our original sample, the AT could not be found because TjE/kOg increased at the same time as TjE/\jOz.
Response time for Oe uptake. In a ramp work rate protocol, RT-Vo2 is the lag in the response of \jo2 following the increase in work rate (32). This lag is determined by the method of least squares in the following way. A best-fit line is drawn through the steady-state Voz response during O-W pedaling, and another line through the linear phase of v02 response during the ramp. The RT-VOW is the time interval between the onset of work and the intersection of these two lines.
Work efficiency. In the ramp test, VOW increases linearly as work rate increases (32). Work efficiency is calculated from the ratio of the change in \jo2 to the change in work rate (33). The values of work rate and Voz are converted into equivalent energy units, and work efficiency is given by the following equation Since the calculated work efficiency is virtually unaffected by values of respiratory quotient in the physiological range, respiratory quotient is assumed to be 0.9, which gives the value 4.9 kcal/l 02.
To estimate differences in metabolic rates at the cellular level, comparative physiologists use allometric equations in which metabolic function is specifically scaled to body mass (weight) (15,17,21,23,25). This was used in part to analyze the data.
Allometric equations have the form Y 0~ Mb where Y is a metabolic function (e.g., \j02~~,J, M is an index of body size, usually measured as body weight, and b is the scaling factor. The scaling factor is found by using a log-log transform resulting in log Y 0~ b x log M where b is then solved by using least-squares technique to find the slope of the best-fit line. In the boys of our sample population, the height scaled to weight was to the power of 0.35, with a 95% confidence interval of 0.33-0.38, r = 0.95. In girls, this same value was 0.33, with a 95% confidence interval of 0.30-0.36, r = 0.95. For the population as a whole, the scaling factor was 0.34, with a 95% confidence interval of 0.32-0.37, r = 0.95. Statistical analysis. Standard techniques of linear regression were used. For significance testing of regression slopes and comparison of slopes of different regression equations, we used the two-tailed t test (12). Comparison of multiple means was done by analysis of variance and modified t test was done by the method of Bonferroni (27). Statistical significance was taken at the P < 0.05 level. Results are presented as means t SD.

RESULTS
Anaerobic threshold. The AT was highly correlated to body weight in children (Fig. 2). The scaling factor for the group as a whole was 0.92 ( Table 2). Slopes of the regression equations and the scaling factors were significantly higher in boys than in girls (Fig. 2, Table 2). Although the scaling factors differed in the children when grouped by age and gender (young boys 0.95; young girls 0.76; older boys 1.12; and older girls 1.02), these differences were not significant by analysis of variance. When the AT was normalized for body weight, there was no significant correlation with age (r = 0.13); however, the group of older girls had a significantly lower mean value (19 t 3 ml. min-' . kg-') than did the older boys (27 t 6 ml l min-l . kg-') or the younger children (boys 26 t 5 and girls 23 t 4 ml. min-' l kg-').
Maximum O2 uptake. VO 2 max also increased systematically with body weight (Fig. 3). The scaling factor for the group as a whole was 1.01 ( Table 2). Slopes of the regression equations and the scaling factors were significantly higher in boys than in girls (Fig. 4, Table 2). Although the scaling factors differed in the children when grouped by age and gender (young boys, 1.01; young girls, 0.79; older boys, 1.02; and older girls, 0.91), these differences were not significant by analysis of variance. When . VO2max was normalized for body weight, there was no significant correlation with age (r = 0.17). However, the group of older boys had a significantly greater mean value (50 t 8 ml. min-l *kg-') than did the three other groups (older girls, 34 t 4; younger boys, 42 t 6; and younger girls, 38 t 7 ml. min-' l kg-'). In addition, the mean of the older girls was significantly lower than the younger boys.  In 31 subjects (28%) (15 girls and 16 boys) a "true" . vo 2 max was observed.The regression equations and scaling factors of this group for both VO2max and AT were virtually indistinguishable from those obtained in the remaining 78 subjects.
In addition to the linear regression and log-log transforms done on the values for VO2max and AT, an analysis of quadratic regression equations (VOgmax and AT as a function of weight) was done. Goodness-of-fit testing demonstrated no significant improvement in model fitting using the second degree equations for either of these parameters in the boys or the girls. Relationship between anaerobic threshold and maximum O2 uptake. The regression slope of AT vs. body weight was significantly lower than that of VO2max (Figs.  2 and 3). The scaling factor of AT was lower than that of vO2maxy but the difference was not significant ( Table  2). The ratio of AT to VO2 max decreased slightly as weight increased (r = -0.28, P < 0.05, Fig. 4). The mean ratio for the study population was 60 t 9%. In the older boys, the ratio was 55 t 10%; younger boys, 64 t 6%; older girls, 58 t 8%; and younger girls, 61 t 7%. This ratio in the younger boys differed significantly from the older girls and older boys. The AT was highly correlated to the . vo 2 max. For the group as a whole, the correlation coefficient was 0.92; for boys 0.91; and for girls 0.89. Response time for O2 uptake. The RT-VOW was independent of body weight in children and adolescents. There were no differences in mean values for this parameter among the younger and older boys and girls (F ratio from analysis of variance was 0.66). For the study population as a whole, the RT-Vo2 was 43 t 15 s. Work efficiency. The work efficiency increased slightly with increasing body weight in children and adolescents (slope of regression equation = O.O7%/kg, r = 0.22). The slope of the regression was higher in girls (O.l5%/kg, r = 0.46) than in boys (O.O3%/kg, r = 0.13). The slope of the boys did not differ significantly from 0. Only the mean value of the older girls (33 t 3%) differed significantly (P < 0.05) from the other groups (younger girls, 29 + 5%; younger boys, 28 t 5%; and older boys, 29 t -47) 0 . WHIPP, AND WASSERMAN

DISCUSSION
We measured four aerobic parameters of exercise using a noninvasive brief exercise protocol in a large group of children. Both the VO 2max and the AT bear a systematic relationship to body weight, whereas the work efficiency and the RT-Vo2 are virtually independent of body mass. For the study population as a whole, the scaling factors of both the AT and the TjO2max are not statistically different from the power of 1.0. No previous studies of exercise in children have utilized the high density breath-by-breath measurement of gas exchange, which makes possible the measurement of aerobic function with the highest possible temporal resolution. To validate the use of our VO2max results as a true limit of the gas transport system, the subjects with true VO2max were compared with those who did not increased systematically with increasing body weight. Linear regression equations for VOzmax (ml/ min) as function of body weight (kg) were for boys, Y = 52.8.X -303.4, r = 0.94; for girls, Y = 28.5 l X + 288.2, r = 0.84; and for group as a whole, Y = 45.6. X -197.9, r = 0.86. Difference in slope between boys and girls was significant. achieve plateaus. The subjects with true TjOgmax were indistinguishable from the others, suggesting that normal children, if sufficiently encouraged, usually approach true limits of vo2, even without reaching a plateau in Vo2. These results were also similar to those of previous investigators using traditional methods of gas collection and exercise protocols (1,6,7). In addition, we reanalyzed the values of TO 2max obtained by Astrand et al. (1) in 63 boys and 63 girls (6-17 yr old) and found a scaling factor of 0.95 (95% confidence interval of 0.89-l.Ol), essentially the same as ours.
The relationship between AT and body weight (Fig. 2) implies that muscle mass is a major determinant of the anaerobic threshold during growth. Like the TjO2max, the AT indicates a physiological limit, the point at which 02 transport alone is apparently insufficient to meet the energy requirements of muscular work. There is evidence that the determinants of the AT include such factors as levels of hemoglobin (34), cardiac function (29), the density of capillaries and mitochondria in muscle tissue, and levels of oxidative enzymes (4,24).
The ratio of AT to the VO2max changed only slightly during growth (Fig. 4). The higher ratios found in the younger boys in particular may be related to such factors as the level of physical fitness, which is known in adults to increase the AT after periods of training (11). The mean ratio in adults tested in our laboratory was 56% (14), close to the mean of 60% that we found in children. It appears that regional 02 transport becomes limited during exercise at a work rate that is demarcated by the onset of anaerobic metabolism and which defines a range of cellular energy production. Our data show that this range comprises a relatively constant proportion of the maximal V?2 response throughout the growth period.
The RT-VOW was found to be highly variable, but the variability was the same in both large and small subjects. and tend to have higher levels of blood lactate during The RT-Vo2 is likely to be determined by intracellular exercise (18). There are several possible reasons for these enzymatic constraints and also by factors such as mus-observations. First, adolescent girls are prone to develop cular stores of "high-energy bond" phosphogens; intra-iron deficiency anemia (5), which could impede 02 transcellular stores of 02 bound to myoglobin, stores of O2 port. Second, the increased percentage of body fat that bound to hemoglobin, and anaerobic metabolism. Our occurs with adolescence in girls (19) may lead to lower finding that the RT-VOW is independent of body size values of VO 2max and AT relative to weight. In fact, suggests that the growth of each of the components of several investigators have shown that when estimates of the cardiorespiratory response to exercise (muscles, ves-lean body mass are used as an index of body size rather sels, heart, and lungs) is integrated so that the temporal than body weight, the differences in VO2max between boys coupling among them is preserved at an optimized value and girls is less marked (10). Additionally, social and despite the overall change in body size.
cultural pressures may inhibit teenage girls' participation The virtual independence of work efficiency on body in vigorous activity and result in comparatively poorer size implies that cellular mechanisms of energy utiliza-fitness levels. tion are mature early in childhood.
In fact, even in A. V. Hill (16) predicted that as animals increased in preparations of fetal mitochondria, there is very little size, body height would scale to body mass to the onedifference in pathways of oxidative phosphorylation third power and VO 2max would scale to body mass to the compared with mature systems (13). Our mean value for the work efficiency (29%) was not different from values two-thirds power. Although the relationship between height and weight in the children was consistent with obtained in adults previously tested in this laboratory Hill's prediction, the scaling factor for VO2max and the (33) but was higher than values in children and adults AT was significantly higher ( Table 2). This discrepancy obtained in other laboratories (1,26). Most likely, these has been observed by others as well (2,8,9). To resolve previous investigators overestimated the VOW for the work done. In their calculation of work efficiency, resting iro2 was subtracted from the vo2 at a particular work this discrepancy, it has been proposed that the peak force generated per unit cross-sectional area of muscle (considered independent of body size by Hill) could increase rate, and they neglected the Tj02 cost of moving the legs with increasing body mass and result in a larger scaling of unloaded cycling. Thus the denominator of the equa-factor for TjO2max than that predicted by Hill (20). This tion was falsely high and the work efficiency artifactually hypothesis is consistent with our findings. low. This problem is avoided when measuring the differ-Our results describe the relationship between growth ence in Vo2 between two levels of work or when usi the rate of change of VOW and work rate ,a .s is done w in the ramp protocol (32) (Fig. 1).
It is unlikely that the small but statistically significant and exercise as one in which certain aspects appear to be "optimized" relatively early in life. Both the work efficiency and the mean response time for VO2 are independent of age and size in our population despite their increases in work efficiency that we found in the older wide ranges. Both the AT and the VO2max increased in a girls represent true differences in substrate and O2 me-highly ordered manner with increasing size, and, as tabolism at the cellular level. Rather, these values may judged by the ratio of AT to VO2max (Fig. 4), the onset of have been caused by the low AT observed in these subjects. As a result, the Av02 may not have represented anaerobic metabolism during exercise occurred at a relatively constant proportion of the overall limit of the gas all of the ATP produced for the A work rate (viz., some ATP came from anaerobic sou rces); the denomi nator transport system. would be falsely low, and the work efficiency would be The authors thank the students, parents, and staff of the Palos artifactually high.
Verdes Unified School District for their cooperation. There were differences between boys and girls in the  Table 2). This resulted in large part from the low values in the older girls. Work D. M. Cooper is a Senior Investigator of the American Heart A ssociation, Greater Los Angeles Affiliate. from other laboratories has also shown that teenage girls have relatively lower VO 2max compared with boys (1, 10) R eceived 22 November 1982; accepted in final form 5 October 1983.