Oxygen cost and oxygen uptake dynamics and recovery with 1 min of exercise in children and adults

Oxygen cost and oxygen uptake dynamics and recovery with 1 min of exercise in children and adults. 3. Appl. Phys-iol. 71(.3): 993-998, 1991.-To test the hypothesis that 0, up- take (VO,) dynamics are different in adults and children, we examined the response to and recovery from short bursts of exercise in 10 children (7-11 yr) and 13 adults (26-42 yr). Each subject performed 1 min of cycle ergometer exercise at 50% of the anaerobic threshold (AT), 80% AT, and 50% of the difference between the AT and the maximal 0, uptake (%70,~~) and 100 and 125% vo2max. Gas exchange was measured breath by breath. The cumulative 0, cost [the integral of vo2 (over base- line) through exercise and 10 min of recovery (ml 0,/J)] was independent of work intensity in both children and adults. In above-AT exercise, 0, cost was significantly higher in children [0.25 t 0.05 (SD) ml/J] than in adults (0.18 ? 0.02 ml/J, P < 0.01). Recovery dynamics of 00, in above-AT exercise [mea- sured as the time constant (&02) of the best-fit single expo- nential] were independent of work intensity in children and adults. Recovery &o,.was the same in both groups except at 125% Tjo 2 max, where do, was significantly smaller in children (35.5 2 5.9 s) than

ence between the AT and the maximal 0, uptake (%70,~~) and 100 and 125% vo2max. Gas exchange was measured breath by breath. The cumulative 0, cost [the integral of vo2 (over baseline) through exercise and 10 min of recovery (ml 0,/J)] was independent of work intensity in both children and adults. In above-AT exercise, 0, cost was significantly higher in children [0.25 t 0.05 (SD) ml/J] than in adults (0.18 ? 0.02 ml/J, P < 0.01). Recovery dynamics of 00, in above-AT exercise [measured as the time constant (&02) of the best-fit single exponential] were independent of work intensity in children and adults. Recovery &o,.was the same in both groups except at 125% Tjo 2 max, where do, was significantly smaller in children (35.5 2 5.9 s) than in adults (46.3 _t 4 s, P < 0.001). VO, responses (i.e., time course, kinetics) to short bursts of exercise are, surprisingly, largely independent of work rate (power output) in both adults and children. In children, certain features of the VO, response to high-intensity exercise are, to a small but significant degree, different from those in adults, indicating an underlying process of physiological maturation. anaerobic threshold; maturation; oxygen uptake kinetics THE NATURAL GROWTH and development of the integrated cellular and cardiorespiratory adjustments to exercise are not yet fully understood. Some data suggest that the time course of 0, uptake (VO,) at the onset of low-intensity exercise [work performed below the anaerobic threshold (AT) or lactate threshold] is the same in children and adults (6), suggesting a fully mature ability of children 27 yr to increase VO, in response to physical activity. Other investigations demonstrated that children compared with adults have faster on-transient 60, dynamics to maximal exercise (14), reach lower levels of blood lactate for similar degrees of heavy exercise (16, Zl), and are less able to sustain exercise in the high-intensity range (2). These observations imply that there do exist growth-related differences in cellular 0, metabolism or transport in response to exercise. The goal of this study was to test the hypothesis that the dynamics of Vo2 in response to exercise in children are different from those in adults. We did this by devising protocols that would, to some extent, mimic patterns of exercise found in the daily life of children and that would also allow a precise analysis of the dynamic VO, response. Thus we used a cross-sectional study design and compared the VO, response with l-min bursts of exercise at a variety of work intensities in a group of adults and children.

Population
Ten healthy children [6 boys and 4 girls, aged 7-11 yr, mean 9.0 t 1.3 (SD)] and 13 healthy adults (10 males and 3 females, aged 26-42 yr, mean 32.6 t 4.8) comprised the study population. All were volunteers, had no chronic diseases, and did not smoke or use medication. The study was approved by the Human Subjects Committee of Harbor-UCLA Medical Center. Informed consent was obtained from each subject and guardian when appropriate. In addition, we reanalyzed maximal VO, (Vo2 m,) and work rate (WR, also power output) data from 65 randomly selected healthy children (age range 6-18 yr, 36 boys) who participated in a study of cardiorespiratory responses to exercise in this laboratory published previously (7).

Protocol
Progressive exercise test. Each subject performed a ramp-type progressive exercise test on a electromagnetically braked cycle ergometer to determine the VO, max and the AT (also known as the lactate or ventilatory threshold) (18, 19).
Burst-exercise tests. The subjects performed 1-min bursts of varying-intensity exercise: 50% of the AT, 80% of the AT, 50% of the difference between the VO, m8x and the AT (we refer to this difference as A), 100% of the . vo Zmaxp and 125% of the Vo2 max. Each subject exercised at unloading pedaling (~7-12 W) for 3-5 min before exercise and for 10 min after exercise. The tests were performed in randomized order and usually required two or three separate sessions. When studies were performed on the same day, a sufficient interval (30-90 min) between studies was allowed so that all gas exchange variables and heart rate had returned to preexercise values.  (VE) and gas exchange were measured breath by breath. The subjects breathed through a mouthpiece connected to a turbine flowmeter and a lowresistance inspiratory-expiratory valve for continuous measurement of inspired and expired volume. The apparatus dead space was 90 ml. CO, and 0, concentrations were measured by a mass spectrometer that sampled continuously from the mouthpiece at 1 ml/s. VE (BTPS), VO, (STPD), CO, output (STPD), end-tidal PO,, and end-tidal PCO, were computed on-line breath by breath as previously described (3). Heart rate was measured beat by beat by a standard lead I electrocardiogram with the use of three electrodes placed on the chest. The data from each test were stored on digital tape for further analysis. Data Analysis .
vo 2max and AT. VO, m8X was taken as the peak vo2 achieved by each subject before cessation of exercise. The AT was measured noninvasively from the gas exchange data obtained during the progressive exercise (7). AT was defined as the vo2 at which the ventilatory equivalent for 0, (vE/v02) and end-tidal PO, increased without an increase in the ventilatory equivalent for CO, (VE/CO~ output) and end-tidal Pco,.

Normalization
To compare the VO, responses from the adults and children, we used several strategies. First, the data were normalized to body weight by using the increase in VO, above baseline (i.e., AVo,/kg). Second, the data were normalized to the actual work performed in two ways as described below.
Cumulative 0, cost per joule. We defined the cumulative 0, cost of exercise bursts as the integral of VO, over baseline values (i.e., unloaded cycling) from the onset of exercise, through the exercise period, and for 10 min of recovery ( Fig. 1). To facilitate comparison of results from adults and children, we divided the cumulative VO, values by the external work done (ml 0,/J). In this manner, differences observed indicated true difference in the relationship between Vo2 and work performed. To determine whether the lo-min period allowed for complete recovery of the VO,, we compared the mean VO, during the last 30 s of the preexercise period with the mean value of the last 30 s of the recovery phase. This analysis was done only for the highest work rates (i.e., lOO%max and lE%max).
To gain additional insight into the physiological mechanism of the 00, response, we measured the cumulative 0, volume that occurred during recovery (Fig. 1). For each subject, we calculated the ratio of these cumulative costs (i.e., recovery to total) at each WR.
End-exercise O2 cost per watt. In addition to the cumulative 0, cost, we calculated the end-exercise 0, cost by dividing the mean Vo2 (above baseline) over the last 6 s of exercise by the WR (ml 0, l min-' l W-l).

Kinetics of 1702
On-transient. As noted, 60-s periods of data collection are not sufficient to allow the achievement of a steady state for VO,. Thus modeling these data to exponential functions is confounded by the lack of knowledge of the asymptotic steady state. To empirically quantify the ontransient response, we calculated the half time (tl,2) as the time required to reach one-half the value of the preexercise to peak vo2 (the largest VO, achieved by the subject).
Recovery time. VO, recovery time was determined by fitting the recovery data (i.e., after the I-min WR) from each subject with a single-exponential equation. Nonlinear techniques were used to calculate the parameters of the characteristic equation (12) h2(t) = Aeekt + C (1) where Vo2(t) is the value of 00, at time t after exercise, A is a parameter, k is the rate constant, and C is the asymptotic baseline value. The time constant (7 = l/k) was used to quantify the recovery time and indicates the time required to achieve 63.2% of the difference between peak and baseline values. On the basis of previous analysis of Vo2 responses to high-intensity exercise (l), where the on-transient data were better fit by models more complex than a single exponential, the recovery data for the highest WR (125%max) were fitted both with a single-exponential (as described above) and a two-exponential model VO 2 ( t) = A 1 eklt + A 2 ekat + C Statistical Analysis Analysis of variance (repeated measures) with subsequent modified t tests (Duncan and Bonferroni method) Values are means t SD. Cumulative 0, cost was not affected by increasing work intensity in children and adults. However, cost was significantly higher in children than in adults at 50%A (*P < O.OOl), lOO%max, and 125%max (**P < 0.01).
was used for the statistical analysis of work-intensity related differences within and between the groups of adults and children. In children, the noise-to-signal ratio for the lowest WR (50% AT) was too high to permit accurate data analysis. A paired t test was used to compare preexercise and recovery values of VO,. In the analysis of data obtained in our previous study (7), linear regression was used to assess the correlation between the ratio Vo2max/ WR,,, and both height and age. Statistical significance was taken at P < 0.05. Data are expressed as means t SD . There were no significant differences in the VO, max per kilogram between adults (41.5 t 8.5 ml 0, l min-l l kg-') and children (41.7 t 5.8 ml 0, l mine1 l kg-'). Similarly, no differences were found between the AT per kilogram in children (24.8 t 5.6 ml OZ. min-' l kg-') and adults (22.6 t 3.5 ml 0, l mine1 l kg-'). Despite this, the WR normalized to body weight was significantly higher in adults for all the tests above AT (Table 1).

Cumulative 0, Cost per Joule
There was no difference between the preexercise and the end-recovery VO, in either group, indicating that recovery for this variable was complete within the 10 min of observation. The cumulative 0, cost of exercise per joule was independent of the work intensity in children and adults (Fig. 2). For WRs above AT (i.e., 5O%A, lOO%max, Values are means t SD in ml l min-' l W-'. * Greater than 80% AT, 5O%A, lOO%max, and 125%max, P < 0.001. t Greater than lOO%max and 125%max, P < 0.05. $ Greater than adults at comparable work intensity, P < 0.001.  125%max), 0, cost per joule was significantly higher in children (mean 0, cost 0.25 t 0.05 ml/J) than in adults (0.18 t 0.02 ml/J, P < 0.01). In adults the recovery 0, cost (as percent cumulative 0, cost) at the highest WR (125%max) was significantly higher (62.3 t 4.2%) than at the two WRs below AT (55.2 t 13.4% at 50% AT and 54.3 t 4.1% at 80% AT, P < 0.05). At 50%A and lOO%max the ratio was 57.8 t 2.7 and 59.3 t 3.5%, respectively. There was no significant effect of WR on the ratio in children (51 t 12% at 80% AT, 57.5 t 5.8% at 5O%A, 56.1 t 5.6% at 1007 omax, 59.7 t 4.1% at 125%max). No significant difference was found between the values in children and adults at any of the WRs studied.

End-Exercise 0, Cost per Watt
In adults, the end-exercise 0, cost decreased as WR increased (Table 2). By contrast, the values in children did not change as WR increased. Moreover, the children's results were higher than those of adults for all the WRs above AT.

Reanalysis of Previous Data
The higher 0, cost in children was an unexpected finding. Thus we thought it would be useful to reanalyze previously collected data to corroborate the results. There was a small but significant decrease in the VO, max/WR,, as both height (r = -0.48, P < 0.001) (Fig. 3) and age (r = -0.53, P < 0.001) increased in the 65 children previously studied. The correlation between VO, maX/WR,,, and height in the current subjects was significantly from 50 to 80% AT (P < O.Ol), but ye found no effect of work intensity on 7v02 in children.
At 125%max 7V0, was significantly higher in adults than in children (*P < 0.001). study are consistent with thi s observati .on: the mean cumulative 0, cost (converted to ml 0, l min-l l W-l) was 11.5 t 0.65 in adults compared with 0, costs of 10.1 ml 0, l min-' l W-l measured from steady-state tests (10).
Similarly in children, the 0, cost was 15.3 t 0.65 ml l min-' l W-l compared with previously reported values of ~10.2 (7).
The reanalysis of previously collected data demonstrates a higher VO, per watt at maximal WR in children than in young adults even though the VO, m,/kg was the same (7) (Fig. 3). These results are consistent with the findings of the present study in which the cumulative 0, cost per joule was consistently greater in the children than in the adults (Fig. 2). Our calculations of the cumulative 0, cost represented the metabolic energy expenditure of the work performed for two reasons: first, because the Vo2 had returned to baseline values during the observation period (see RESULTS), meaning that we had measured all the increase in Vo2 resulting from the perfor-Recovery Time Constant mance of exercise; and second, because we subtracted the Vo2 due to unloaded pedaling and, thereby, eliminated In adults the Vo2 recovery time constant increased be-energy expenditure related to inertial and frictional costs tween the two lowest WRs (50 and 80% AT) (P < O.Ol), incurred as the subject moved his or her legs in the perforbut the time constants were independent of WR for the mance Of 'ycle ergometryo remaining protocols. In the children, the VO, recovery There are several possible mechanisms for the higher time constants were independent of WR. The recovery total 0, costs per joule in children than in adults. Subtime constants were lower in children than in adults, but strate utilization itself can modify the relationship beonly at 125% of the VO 2 max was there a statistically signif-tween VO, and WR because the high-energy phosphateicant difference (P < 0.001) (Fig. 4). We observed no to-O, ratio is 2.82 for fats and 3.00 for glucose (17). If differences in the recovery time constants of females children metabolized solely fat and adults only glucose compared with males in either the adults or children. We (which is highly unlikely), then the difference in 0, cost also attempted to fit a two-exponential equation for the could be no more than 6%. The difference we observed 125%max exercise intensity recovery data. In four chil-was, on average, almost 40% greater in children and, dren and six adults, the two-exponential model resulted therefore, inconsistent with a growth-related difference in slightly better fits than the single-exponential.
in substrate utilization The larger 0, cost in children could reflect a greater DISCUSSION influence of O,-requiring processes such as thermoregulation that accompany muscular work (13). Heat loss dur-The response to and recovery from 1-min bursts of exercise reveal much about the integrated cardiorespiratory adjustment to changes in metabolic rate. It is noteworthy that in both adults and children >50% of the total 0, cost occurred during the recovery period. The significantly greater cumulative 0, cost per joule in children at virtually all WRs was the major difference between the groups. In addition, the VO, recovery time was prolonged in adults at the highest WR. Surprisingly, we found that .
ing exercise in children is likely to be facilitated by their larger ratio of surface area to body mass, and this would tend to reduce rather than increase any additional 0, requirement for the maintenance of temperature homeostasis. Other O,-dependent processes like cardiac or respiratory work are small in magnitude and unlikely to account for the differences in cumulative 0, cost that we observed between adults and children (4).
Margaria et al. (15) were among the first to postulate the Yo, recovery times and cumulative 0, cost per joule the mechanisms responsible for 0, kinetics at the onset were minimally affected by WR in both adults and chilof exercise. One of the determinants of VO, at the ondren. transient of exercise was the amount of 0, stored in the Previous investigators have examined the dynamics of VO, at the beginning and end of exercise by the use of the tissues available for aerobic energy metabolism. At the end of exercise, these stores would be replenished. The stored 0, (largely venous blood) accounted for a small but measurable amount of the 0, deficit. It is noteworthy that children have slightly smaller hemoglobin concentrations than do adults, and smaller stores might result in more rapid VO, kinetics. However, we found no apparent growth-related differences in VO, on-transient kinetics in either the present study or in our previous investigation of VO, kinetics during low-intensity exercise (6). Moreover, the cumulative 0, cost is not likely to be affected by the size of the 0, stores if they return to their preexercise values in a relatively short period of time. It is important to note that the brief exercise period (i.e., 1 min) virtually precluded the possibility of evaluating all but the simplest VO, kinetics at the onset of exercise. For example, we would not be able to find age-related differences in slower exponential components of the on-transient, even if they existed, using the current protocol and data analysis.
Biomechanical factors (e.g., use of additional muscle groups for limb acceleration) have been used to explain greater energy expenditure during walking and running in children than in adults (5, 8). However, these studies were limited because one cannot measure the precise work done by the use of treadmill exercise. In addition, cycle ergometry is likely to involve comparable muscle groups in adults and children because pedaling on the ergometer constrains limb movement to a greater degree than does walking or running.
The end-exercise 0, cost per watt did not entirely parallel the values obtained for the cumulative 0, cost per joule. The end-exercise 0, cost was consistently higher in children than in adults (Table 2). Although the end-exercise 0, cost fell in adults as WR increased, it was unaffected by the work intensity in children. These observations may be related to the maturation process associated with anaerobic metabolism.
Lower levels of blood lactate during exercise have been found in children than in adults (16,21). The metabolism of pyruvate to lactate can, for a limited period of time, act as an O,-sparing mechanism during the exercise response by providing a source of ATP rephosphorylation that does not require 0,. As judged by the fall in the 0, cost at higher work intensity in adult subjects, a greater proportion of the energy expenditure at the end of the l-min exercise bout was derived from anaerobic sources. Note, however, that the total 0, cost per joule did not change (Fig. 2), implying greater Vo2 during recovery. In fact, at 125%max, the 0, cost during recovery (as percent cumulative 0, cost) in adults was significantly greater than for below-AT exercise. Moreover, the finding of generally higher end-exercise 0, costs per watt in children during high-intensity exercise suggests a greater dependence on O,-dependent ATP generation. Ultimately, this issue will be resolved either by measurements of blood lactate during exercise in children (few healthy children or their parents agree to blood sampling during exercise tests) or, alternatively, by using noninvasive estimates of cellular energy metabolism using magnetic resonance spectroscopy.
There are, in fact, limited data suggesting that the level of phosphofructokinase, a key muscle glycolytic enzyme, is lower in children than in adults (9). This observation supports the concept that children have less "anaerobic capacity" than do adults because glycolysis provides the pyruvate necessary for anaerobic ATP production. Whether the greater 0, dependence in children represents 1) less ability for anaerobic metabolism, 2) less efficient coupling of oxidative metabolism to force generation in the muscle, or 3) a more efficient ability to increase uptake of atmospheric 0, has yet to be determined.
It was surprising to find that the kinetics of the recovery phase were only minimally WR dependent in adults and virtually independent of work intensity in children (Fig. 4). Previous studies in this laboratory demonstrated that constant WR performed above the subject's AT results in on-transient kinetics that are not described well by single-exponential functions. Accurate modeling of the high -intensi .tY VO, response requires th e addition of another slower (i. e., longer time constant) exponential term to adequately fit the data.
The increasing complexity may result from O,-dependent processes of lactate metabolism [e.g., lactate oxidation or lactate conversion to glucose (Cori cycle)] once excess lactate has been produced during high-intensity exercise (17). Thus it is reasonable to expect that the metabolic processes associated with high-intensity exercise, once initiated, would prolong the recovery dynamics of VO,. However, models more complex than the single exponential did little to improve the fit of the data. Why did short-term burst exercise lead to relatively simple Vo2 kinetics during the recovery phase? One possible explanation is that the 1-min period of exercise was too brief to induce the slow metabolic adaptations characteristic of longer bouts of high-intensity exercise. The response dynamics of the slower process (related to the second exponential term in the on-transient) are so slow that there is only a very small effect on the recovery after 1 min of exercise.
Although the time constant for 0, recovery (80~) tended to be smaller (i.e., faster recovery) in children at all work intensities, 7V02 was significantly faster than in adults only at 125%max. As noted, a greater anaerobic metabolism may explain the slower VO, responses, thus the slower TO, kinetics at the highest WR in the adults.
Finally, there may be a practical implication of our data for clinical investigations of abnormalities in cardiorespiratory responses to exercise in children. Above-AT exercise protocols are advanta .geous because the noise-to-signal ratio is often lower th an in below-AT exercise, but the metabolic acidosi .S and stress tha t accompany high-intensity exercise are n ot advisabl .e in patients with impaired function. Using short bursts of exercise in individuals with heart or lung disease could mitigate this problem. Moreover, choosing the appropriate WR to ensure a valid comparison with normal children is easier because the on-transient t1,2, the cumulative 0, cost, and the recovery do2 are virtually independent bf WR in healthy children.