Main

There have been several reports of difference in ventilatory response to exercise between children and adults(1–4), and it is thought that the smaller body size, smaller CO2 storage capacity, smaller production of CO2 at cellular level, and immaturity of the respiratory center of children account for the different ventilatory responses of children and adults. These prior studies were, however, based on noninvasive analyses of ventilatory response during exercise. In light of the growth of body size and maturation of muscle metabolism and the ventilatory center in children, the degree to which these noninvasive variables are correlated with data obtained by invasive techniques should be evaluated. The accurate measurement of PaCO2 is important in studies of ventilatory control during exercise. However, indwelling arterial catheters for sampling arterial blood are not always feasible or desirable, especially in children. In addition, because repeated studies of the same subjects may be unacceptable, a noninvasive estimation of PaCO2 is preferable. There have been few studies of children in which a detailed evaluation of the relationship between PaCO2 and ETCO2 during exercise was made with pulmonary function. The measurement of ETCO2 has been used to estimate PaCO2 during rest, but it has been shown that ETCO2 significantly overestimates PaCO2 during exercise in adults, and it is unclear whether ETCO2 overestimates PaCO2 during exercise in children. If the relationship obtained in children is different from that in adult subjects, this difference is very important for evaluating ventilatory mechanisms under a variety of conditions, such as during exercise. This would also be important and useful information for investigators dealing with pediatric patients with cardiac and/or pulmonary dysfunction for estimating ventilatory responses under stress conditions. We, therefore, investigated the ventilatory response and arterial blood gases during exercise in children and compared their data with those obtained from young adults.

METHODS

Subjects. Seven boys, aged 8 to 11 y (mean, 9.7 y [group A]), and young adult men, aged 14 to 21 y (mean, 18.0 y [group B]) participated in this study. According to secondary sex characteristics (genitalia and pubic hair), the classifications of sex maturity stage (Tanner stage) in group A were all in stage 1(5), i.e. they were all in the prepubertal stage. Five of the six subjects in group B were in Tanner stage 5, i.e. the postpubertal stage. One subject (age 14 y), categorized in group B, was considered to be in stage 4, i.e. in the late pubertal stage. One female, aged 14 y, was excluded from the present study because we thought that it was possible to eliminate the influence of the gender difference in cardiorespiratory response during exercise. The 13 subjects were all being followed at our institute because of a history of Kawasaki disease; however, no stenotic coronary arterial lesions or abnormal hemodynamic data were detected by follow-up cardiac catheterization and selective coronary angiography. Some subjects showed a slight dilation of the coronary artery and others showed the regression of a dilated coronary artery. The subjects also had no history of chronic pulmonary disease and showed normal pulmonary function (Table 1). No abnormal ECG findings during an exercise test were observed in any subjects, and they all lived a normal life without any limitations. We concluded that the subjects had normal cardiorespiratory function because patients with Kawasaki disease who had significant coronary aneurysms had cardiorespiratory responses similar to those of healthy control subjects during exercise testing(6). We considered the present subjects to be good substitutes for healthy control subjects.

Table 1 Study subjects
Full size table

All of the subjects had some experience with treadmill exercise testing in an outpatient clinic and were familiar with this kind of exercise testing. In addition, before any invasive study was undertaken, each subject first performed the treadmill exercise test at least once to become accustomed to the protocol and to set a baseline for exercise tolerance.

Consent. Informed consent was obtained from each subject or his parents. This protocol was approved by the Ethics Committee of the National Cardiovascular Center.

Pulmonary function tests. All 13 patients underwent pulmonary function tests. We measured the VC (L), the percent forced expired volume in 1 s (%) (Spirosift, SP-600, Fukuda Denshi, Tokyo, Japan), functional residual capacity (mL), RV (mL), and TLC (mL) (Ellopse-1000 System, Fukuda Denshi). The VC values were presented as the percent of normal (% VC), and the RV to TLC ratio (%) was calculated.

Exercise protocol. The subjects performed a ramp-like progressive exercise test on a treadmill (Q-5000, Quinton, Seattle, WA). We previously demonstrated a high correlation between AT and the lactate threshold using this treadmill test, and we established its clinical usefulness for determining AT and evaluating cardiorespiratory tolerance in patients with congenital heart disease(7). The exercise intensity was increased by 0.7 metabolic units every 30 s with completion of the incremental part of the exercise test in approximately 10 min. When selecting the slope of the V˙O2 as a function of the work rate, a value of 3.5 mL/kg/min (= 1.0 metabolic unit) was used because of the difficulty in determining this metabolic unit in children. After a 4-min rest, the patients performed a 3-min WU walk at a speed of 1.5 km/h and then exercised with progressive intensity until exhausted. Twelve standard ECG leads were placed to monitor the heart rate during exercise testing.

Gas exchange measurements. Ventilation and gas exchange were measured by the breath-by-breath method. The subject breathed through a mask connected to a hot-wire anemometer (Riko AS500, Minato Medical Science, Osaka, Japan) to measure inspired and expired volume continuously. A mass spectrometer (MG-300, Perkin-Elmer, St. Louis, MO) was used for continuous measurements of partial oxygen and carbon dioxide pressures. Two sizes of full face masks were used: one for children from 120- to 150-cm tall, which had a dead space of 80 mL, and another for the subjects taller than 150 cm, which had a dead space of 100 mL. In the breath-by-breath protocol, derived respiratory parameters, including the RR, VT, V˙E, V˙E/V˙O2, V˙E/V˙CO2, and the respiratory gas exchange ratio (R), were computed in real time and displayed with heart rate and V˙O2 on a monitor. A personal computer (PC-9801, NEC, Tokyo, Japan) was used for data acquisition and storage. Breath-by-breath data were averaged to provide one data point for each 30-s period. The delay times and response characteristics of the oxygen and carbon dioxide analyzers were carefully checked before each exercise test.

The metabolic rate above which anaerobic metabolism supplements the production of aerobic energy production and leads to lactic acidosis corresponds to the AT. This threshold was defined as the V˙O2 at which the V˙E/V˙O2 and the end-tidal PO2 increased without increases in the V˙E/V˙CO2 and ETCO2, alternatively determined by the V-slope method(8,9).

Arterial blood gas analysis. Arterial blood samples were taken from an indwelling 22-guage angiocath placed in a radial or brachial artery. The samples were taken at rest, 3 min after WU walking began, at AT, and at peak exercise. Samples were stored on ice (≤ 20 min) until the time of analysis for pH, PaO2, and PaCO2 (ABL3 blood gas analyzer, Radiometer, Copenhagen, Denmark).

CALCULATIONS

In this study, because a steady state did not exist at AT or at peak exercise, end-tidal PO2 was used instead of the partial alveolar O2 tension, and ETCO2 was used instead of the partial alveolar CO2 tension (which is difficult to measure clinically). The P(ET-a)DO2 represents mainly the low ventilation-perfusion space in the lung, and the P(a-ET)DCO2 represents mainly the high ventilation-perfusion space in the lung. During progressive exercise, these relationships may not hold, but it was expected that changes in these parameters may prove useful(10).

V˙A and VD/VT. V˙A was quantified from the measurement of V˙CO2 and ideal alveolar CO2 concentration estimated from the PaCO2 as shown in the following equation: Equation 1 VD/VT was calculated using Enghoff's modification of the Bohr equation with PaCO2 (aVD/VT) and with ETCO2 (etVD/VT) as shown in the following equations: aVD/VT = (PaCO2 - PECO2)/PaCO2 etVD/VT = (ETCO2 - PECO2)/ETCO2 where PECO2 is mixed with expired PCO2.

Calculation of total V˙CO2 and V˙A. Considering delay (time constant) of V˙CO2 response to exercise, total V˙CO2, during the periods from 30 s after the beginning of the ramp exercise to the AT point (aerobic exercise) and from the AT point to peak exercise (anaerobic exercise), was calculated separately. By assuming that the increase in V˙CO2 was linear during ramp exercise, the total amount of V˙CO2 during aerobic exercise could be calculated as an area of a triangle (S1) and total amount of V˙CO2 during anaerobic exercise could be calculated as an area of a trapezoid (S2), as shown in Figure 1. Total V˙A was calculated in the same manner.

Figure 1
figure 1

Change in V˙CO2 during exercise testing (sample from 9-y-old boy from group A). Assuming that the increase in V˙CO2 is linear during ramp exercise, total amount of V˙CO2 during aerobic exercise can be calculated as an area of a triangle (S1), and total amount of V˙CO2 during anaerobic exercise can be calculated as an area of a trapezoid (S2).

Full size image

Statistical analysis. Differences between two groups in mean respiratory values during exercise were assessed by a two-way repeated measure ANOVA. Differences between two groups at each exercise intensity were evaluated by unpaired t test, and changes in respiratory variables during exercise were evaluated by paired t test. Data are expressed as mean ± SEM. p values < 0.05 were considered significant.

RESULTS

Pulmonary function at rest. TLC, VC, and forced expired volume in 1 s were normal in all subjects. Although the lung volume was significantly lower in group A than in group B, no significant difference in percent predicted values for these variables were observed between groups A and B. The RV/TLC ratio was lower in group A than in group B (Table 1).

Respiratory response to exercise. The respiratory variables at rest, during WU, at AT, and at peak exercise are shown in Table 2. Although RR at rest in group A was not significantly different from that in group B, the increase in RR during exercise was greater in group A than in group B. There was no significant difference in Tv per body weight (Tv/kg) during exercise between groups A and B except at rest.

Table 2 Ventilatory variables at each measured exercise intensity for the study groups during progressive exercise
Full size table

The V˙CO2 at rest and during WU was greater in group A than in group B; however, there was no significant difference in V˙CO2 between the two groups at AT and at peak exercise. Although V˙CO2 and R at peak exercise in group A showed higher values compared with those in group B, this difference did not reach statistical significance.

The V˙a throughout the exercise test except at AT was significantly greater in group A than in group B, and the V˙A/V˙CO2 during exercise was greater in group A than group B. Thus, the children showed excessive ventilation during exercise compared with the young adults.

Relationship between total V˙CO2 and total V˙A during exercise. The total V˙CO2 during both aerobic and anaerobic exercise was significantly lower in group A than in group B (Table 3). However, no significant difference in total V˙A between groups A and B was observed during both aerobic and anaerobic exercise. As a result, although no significant difference in the total V˙A to total V˙CO2 ratio (dV˙A/dV˙CO2) during the period from WU to AT was observed, dV˙A/dV˙CO2 during the periods from AT to peak exercise and from WU to peak exercise was significantly greater in group A than in group B.

Table 3 Carbon dioxide output and alveolar ventilation during exercise
Full size table

Arterial blood gas tensions during exercise. Although the pH values during aerobic exercise were not significantly different between groups A and B, pH at peak exercise was significantly higher in group A than in group B (group A: 7.33 ± 0.01 versus group B: 7.28 ± 0.01; p < 0.05).

There was no significant difference in the change in PaO2 during exercise between groups A and B. PaO2 at AT was lower than that at rest in group A, and PaO2 in group A was significantly decreased at AT and at peak exercise compared with that at rest in group B. Although group A demonstrated higher values for PaO2 during exercise, this difference was not significant (Fig. 2a).

Figure 2
figure 2

(a, b) Changes in PaO2 and PaCO2 during exercise. Solid and open circle represent values for children and younger subjects, respectively. * p < 0.05 vs rest, ** p < 0.01 vs rest, § p < 0.05 between two groups, §§ p < 0.01 between two groups.

Full size image

Significantly different changes in PaCO2 were demonstrated during exercise between groups A and B. PaCO2 decreased significantly from the AT point to peak exercise in group A (p < 0.01); whereas the PaCO2 in group B increased from rest to AT (p < 0.05) and thereafter decreased slightly, and the PaCO2 at peak exercise was not significantly different from that at rest. Group A showed lower values for PaCO2 throughout the study period, and PaCO2 at AT and at peak exercise were significantly lower in group A than in group B (Fig. 2b). dV˙A/dV˙CO2 was well correlated inversely with the change in PaCO2 during the period from AT to peak exercise (Fig. 3) and peak PaCO2 (r = -0.90; p < 0.001).

Figure 3
figure 3

Correlation between total V˙A to total V˙CO2 ratio (dV˙A/dV˙CO2) and change in PaCO2 during the period from the AT to peak exercise (anaerobic exercise).

Full size image

Alveolar-arterial gas tension differences during exercise. There was no significantly different change in P(ET-a)DO2 during exercise between groups A and B, and P(ET-a)DO2 was significantly higher at peak exercise than that at rest in both groups (Fig. 4a). A significantly different change in P(a-ET)DCO2 was demonstrated during exercise between the two groups (p < 0.001). Although there was no significant difference in P(a-ET)DCO2 at rest between the two groups, a significant difference from zero was observed in group B, whereas no significant difference from zero was observed in group A. Although P(a-ET)DCO2 was significantly different from zero at the AT in group A, no significant difference in P(a-ET)DCO2 from zero was observed at peak exercise. In group B, P(a-ET)DCO2 decreased progressively as the exercise intensity became greater, and the values at the AT and at peak exercise were significantly lower than zero. P(a-ET)DCO2 was significantly lower in group B than group A at peak exercise (Fig. 4b).

Figure 4
figure 4

(a, b) Changes in P(ET-a)DO2 and P(a-ET)DCO2 during exercise. Solid and open circle represent values for children and younger subjects, respectively. * p < 0.05 vs rest, ** p < 0.01 vs rest, *** p < 0.001, § p < 0.05 between two groups, # p < 0.05 vs zero, ## p 0.01 vs zero.

Full size image

P(a-ET)DCO2 at peak exercise was inversely correlated with Tv (L) and V˙CO2 (mL) (Tv: r = -0.76, p < 0.01; V˙CO2: r = -0.58, p < 0.05; Fig. 5, a and b).

Figure 5
figure 5

(a, b) Correlation between tidal volume and P(a-ET)DCO2 at peak exercise, and correlation between V˙CO2 and P(a-ET)DCO2 at peak exercise. Solid and open circle represent values for children and younger subjects, respectively.

Full size image

VD/VT during exercise. A significantly different change in aVD/VT was demonstrated during exercise between the two groups (p < 0.001, Fig. 6). Although aVD/VT at rest was lower in group A than in group B and its value at the AT was significantly lower compared with that at rest, the improvement in aVD/VT was small throughout the study period in group A. The, aVD/VT at rest was significantly higher in group B than in group A; however, the improvement in aVD/VT during exercise was greater in group B, and as a result, the aVD/VT value at peak exercise was significantly lower in group B than in group A (p < 0.05).

Figure 6
figure 6

Change in aVD/VT calculated with PaCO2. Solid and open circle represent values for children and younger subjects, respectively. ** p < 0.01 vs rest, § p < 0.05 between two groups.

Full size image

There was no significant difference between aVD/VT and etVD/VT during exercise in group A, and the etVD/VT at the AT point was not significantly different from that at rest as demonstrated with aVD/VT. The etVD/VT in group B was, however, higher than aVD/VT during exercise and the difference was significant at peak exercise (Fig. 7). Thus, dead space ventilation was overestimated in group B compared with group A when it was assessed by VD/VT calculated using ETCO2.

Figure 7
figure 7

(a, b) Changes in VD/VT in groups children (left) and younger subjects (right). Solid circle and triangle represent values for VD/VT calculated with PaCO2 (aVD/VT) and with end-tidal PCO2 (etVD/VT) in children, respectively, and open circle and triangle represent values for aVD/VT and etVD/VT, in younger subjects, respectively. *p < 0.05 vs rest, ** p < 0.01 vs rest, *** p < 0.001, §§§ p < 0.001 between two groups.

Full size image

DISCUSSION

Our study demonstrated that the magnitude and change in P(a-ET)DCO2 during exercise were smaller in children than in young adults(11), and PaCO2 during exercise was estimated somewhat more accurately by ETCO2 in children compared with young adults. Williams et al.(12) demonstrated that PaCO2 during exercise is better estimated by ETCO2 in old subjects, contrary to what is observed in young adults. Those authors attributed this finding in part to an increase in alveolar dead space caused by normal decline in lung function which occurs with aging and small CO2 production during exercise. Because of the anatomic and functional development of the respiratory system, including low compliance of the lungs(13) and the relatively larger size of the upper airway in the respiratory system of children compared with that in adults, it is likely that the extrathorax anatomic dead space is larger in children than in adults. In terms of dead space ventilation, this could account for the similarity of the response of P(a-ET)DCO2 in children to that demonstrated in old subjects by Williams et al.(12). The fluctuation of alveolar PCO2 during the breathing cycle has been considered as one factor in the magnitude of P(a-ET)DCO2(14), and increasing V˙CO2 and decreasing VT as expiration result in greater P(a-ET)DCO2 in young adults(12,15). Therefore, in addition to a less efficient ventilatory response to exercise caused by the incomplete process of maturation of the respiratory system during childhood(16), the smaller VT and smaller total V˙CO2 during exercise observed in the present study (probably caused by a lower production of lactate) could be responsible for the smaller P(a-ET)DCO2 during progressive exercise. In addition, it is likely that smaller V˙CO2 in working muscles results in a smaller elevation of mixed venous Pco2, and this also could be responsible for smaller P(a-ET)DCO2 during exercise(14,17,18). The positive mean P(a-ET)DCO2 in both children and young adults at rest is probably the result of the underperfused lung apexes caused by the gravitational effect(11), and this effect might be smaller in children because of their smaller lung volume.

Even if VT is the most important single determinant of P(a-ET)DCO2, as demonstrated by Jones et al.(15), our present results are in agreement with their findings and, in fact, we observed a relatively good correlation between P(a-ET)DCO2 and VT and V˙CO2 at peak exercise.

As shown in alveolar equation, V˙a = 863·V˙CO2/PaCO2, then PaCO2 = 863·V˙CO2/V˙A. Therefore, PaCO2 is determined by the balance of V˙CO2 and V˙A, and the difference of PaCO2 between children and young adults during exercise is determined by the difference in V˙CO2 and V˙A between these groups.

V˙CO2. Because of the high basal metabolic rate per body weight (i.e. high V˙O2/kg) in children, V˙CO2/kg is higher in children than in young adults. However, additional total production of CO2 during the periods from WU to AT and from AT to peak exercise is significantly smaller in children than in young adults. Cooper et al.(2) demonstrated faster responses of V˙E and V˙CO2 to constant work rate exercise in children compared with young adults, and speculated that this finding was caused by a smaller capacity of CO2 storage in children compared with young adults and not caused by a lesser production of CO2. However, it seems that because the CO2 storage capacity was saturated to some extent during the WU period, the difference in CO2 storage capacity might have had no influence on the results obtained in the present study. Although it is unclear why the V˙CO2 during exercise is smaller in children than in young adults, it may be caused by the immaturity of cellular metabolism in working muscle in children. According to a report by Eriksson et al.(19) the production of lactate in working muscle in a 13-14-y-old boy is smaller even during relatively mild exercise compared with that in adults. This might mean that the CO2 production by the HCO3 buffering system is somewhat smaller in children than in adults, even at exercise intensity below AT. At exercise intensity above AT, of course, the CO2 production is smaller in children than in adults. R (=V˙CO2/V˙2) at peak exercise is lower in children than in adults, and this finding has been thought to be the result of a lower production of lactate in children at peak exercise(1). In the present study, R in children was lower than that in young adults, although not significant.

V˙A. Although the V˙A/kg was higher in the children than in the young adults, the increases in the total amount of V˙A/kg from WU to AT and from AT to peak exercise in the children were not significantly different from those during the same periods in the young adults. Although the respiratory system of children is undergoing the process of maturation, during which the higher respiratory rate and smaller VT result in a less efficient ventilatory response to exercise compared with adults(1), children are able to increase their respiratory rate enough to increase as much V˙a volume as that in young adults. Because a strong linear relation is observed between VC and body height, children have enough lung volume corresponding to their body size, regardless of their metabolic development, and it is therefore possible for them to obtain sufficient minute ventilation during exercise. Thus, the increase in V˙A during exercise is determined by a subject's body size and is not always dependent on his or her cellular metabolic development. In addition, the incomplete maturation of the peripheral chemoreceptors for hypoxia(3) and/or the central mechanism of ventilatory acceleration during childhood might contribute to the well-maintained V˙a during progressive exercise(20).

Therefore, it appears that Paco2, which is determined by the V˙A/V˙2 value, is lower in children than in young adults during moderate-to-severe exercise, and children may be undergoing relatively alveolar hyperventilation during exercise(21). The present data suggest that a "mismatch" between the immaturity of the cellular metabolic condition in the working muscle and the well-maintained capacity for increasing V˙a in children, despite their smaller lung volume, is responsible for the different responses of Paco2 and P(a-ET)DCO2 from those of young adults during exercise.

The changes in P(ET-a)DO2 during exercise observed in the children in the present study are similar to those reported in adults(11). In addition, we found that there is no significantly different change in P(ET-a)DO2 between children and young adults during progressive exercise. Because P(ET-a)DO2 represents mainly the magnitude of right-to-left shunting and/or a low ventilation-perfusion area in the lung, it seems that these physiologic parameters in children were not significantly different from those in young adults during exercise. The reason why there was no significant difference in P(ET-a)DO2 between these two groups in the present study is unclear. According to the report by Wagner et al.(22), for up-to-moderate exercise intensity, the small increases in the P(ET-a)DO2 caused by alveolar-capillary diffusion limitation and those caused by ventilation-perfusion mismatch are approximately the same. At a severe exercise intensity, however, most of the increase in P(ET-a)DO2 is attributable to an alveolar-end-capillary diffusion limitation(22). In athletes, a short pulmonary transit time results in a decrease in Pao2 because of high pulmonary flow(22). Because of the smaller increases in cardiac output (pulmonary flow) during the period from rest to peak exercise in children compared with that in adults(23), the pulmonary transit time to ensure gas equilibration in the small lung might be enough in children to maintain high Pao2 values compared with low Pao2 in adults, although a statistical difference in Pao2 at peak exercise between groups A and B was not observed in the present study. However, because the magnitude of dead space ventilation (i.e. smaller decrease in aVD/VT as demonstrated in the present study) is greater in children than in adults, there may be no significant difference between these groups in P(ET-a)DO2 during heavy exercise.

Study limitations. In the present study, our subjects were not considered entirely healthy. However, to our knowledge, there has been no report about ventilatory dysfunction in patients with a history of Kawasaki disease with no significant coronary arterial lesions, and our subjects showed no abnormal pulmonary findings. As mentioned, because patients with Kawasaki disease who have had significant coronary aneurysms have cardiorespiratory responses similar to those of healthy control subjects during exercise testing(6), we believe that our present data are comparable to those from normal healthy subjects.