Main

HR recovery depends on the cardiorespiratory fitness level of the individual and the intensity of exercise (1–3) and is greatly affected by autonomic nervous activity during the recovery period (4). According to previous studies, HR recovery is faster in children than in adults (3); however, the precise mechanism of the rapid HR decline in children is unknown. Recently, measurement of HRV has enabled us to estimate cardiac autonomic nervous activity (5) to some extent. If cardiac autonomic nervous activity is related to HR recovery, analysis of HR recovery should provide important information in estimating cardiac autonomic function for investigators working with pediatric cardiac patients. Therefore, we investigated the relationship between HR recovery after exercise and cardiac autonomic nervous activity in children, including arterial baroreflex function, which is important in maintaining blood pressure by regulating HR. We compared the results with those obtained from young adults.

METHODS

Subjects

Seven boys and two girls aged 9 to 12 y (mean, 10.4 y; group A) and six young adult men and two women aged 17 to 21 y (mean, 19.1 y; group B) participated in this study. According to secondary sex characteristics (genitalia and pubic hair), the patients in group A were all Tanner stage 1 (6), i.e. they were all prepubertal. The Tanner stage of all subjects in group B was stage 5, i.e. postpubertal. All 17 subjects were being followed at our institute because of a history of Kawasaki disease. Mean follow-up period was 12 ± 5 y (5 to 20 y). However, no stenotic coronary arterial lesions or abnormal hemodynamic data were detected at follow-up cardiac catheterization and selective coronary angiography. Echocardiography and cardiac catheterization studies were performed during the hospital stay, during which all the subjects performed the exercise tests of the present study. No significant mitral valve regurgitation or abnormal wall motion of the left ventricle was demonstrated in any of the subjects. Some subjects showed slight dilation of one coronary artery, and others showed regression of previously dilated coronary arteries. The subjects had no history of chronic pulmonary disease and had normal cardiopulmonary function (Table 1). No abnormal ECG findings were observed in any subject during an exercise test, and they all lived a normal life without any limitations. All of the subjects had some experience with treadmill exercise testing in outpatient clinics and were familiar with this kind of exercise test.

Table 1 Study subjects Abbreviations: LVEDVI, left ventricular end-diastolic volume index (LVEDV/body surface area); LVEF, ejection fraction of the left ventricle. Values are mean ± SE. †p < 0.001 compared with young adults.
Full size table

Consent

Informed consent was obtained from all subjects or their parents. The present study was approved by the Ethics Committee of the National Cardiovascular Center.

Pulmonary Function Tests

All 17 patients underwent pulmonary function tests. We measured the vital capacity (L), the percent forced expired volume in 1 s (%) (Spirosift, SP-600, Fukuda Denshi, Tokyo, Japan), functional residual capacity (mL), residual volume (mL), and total lung capacity (mL) (Ellopse-1000 System, Fukuda Denshi). The vital capacity values are presented as a percent of normal, and the residual volume to total lung capacity ratio (%) was calculated.

Exercise Protocol

Peak exercise.

The subjects performed a ramp-like progressive exercise test on a treadmill (Q-5000 system, Quinton, Seattle, WA). After a 4-min rest, the patients completed a 3-min warm-up walk at a speed of 1.5 km/h (grade, 0%) and then exercised with progressive intensity until exhausted. After a 30-s walk-down period, the patients sat on a chair and cardiorespiratory variables during the next 4 min were obtained. In our original protocol, 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 determining the slope of V˙O2 as a function of work rate according to the equation MATH where V˙ and G are the velocity (km/h) of the belt and its grade (%), respectively (7), we used a value of 3.5 mL·kg−1·min−1 (= 1.0 metabolic unit) as reported in adults. The value was derived from the resting V˙O2 for a 70-kg 40-y-old male because of difficulty in determining an actual metabolic unit in children. Using this treadmill protocol, we demonstrated a high correlation between the values of V˙O2 at AT and the lactate threshold, and we established its clinical usefulness for determining the AT and evaluating cardiorespiratory tolerance in patients with congenital heart disease (8). Endurance time (min) was defined as the duration of ramp exercise of the protocol.

Twelve standard ECG leads were placed to monitor HR during exercise testing. SBP was measured every 2 min during exercise and 1, 2, and 4 min after peak exercise. Because it is difficult to measure blood pressure with a mercury sphygmomanometer during dynamic exercise, especially in younger patients, we measured SBP by the palpation method. In a preliminary study, SBP obtained by this method during treadmill exercise testing in 12 patients with cardiac disease correlated with measurements taken using a mercury sphygmomanometer (r = 0.98, p < 0.0001, our unpublished data, 1996).

Constant-load exercise.

Each subject performed a 4-min constant-load exercise test on a separate day. After a 3-min standing period, the patients performed a 4-min constant-load walking test at the AT-level exercise intensity and then immediately sat on a chair, and cardiorespiratory variables during the next 4 min were obtained. Because of difficulty in determining AT-level exercise intensity on a treadmill protocol, the stage after which the AT had occurred during our 30-s incremental protocol was adopted as the intensity of constant-load exercise.

Gas exchange measurements.

Ventilation and gas exchange were measured by the breath-by-breath method. Each 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) was used for continuous measurements of oxygen and carbon dioxide partial 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 subjects taller than 150 cm, which had a dead space of 100 mL. In the breath-by-breath protocol, derived respiratory parameters including the respiratory rate, tidal volume, minute ventilation, ventilatory equivalents for oxygen and carbon dioxide, and the respiratory gas exchange ratio were computed in real time and displayed with HR and V˙O2 on a monitor. A personal computer (PC-9801, NEC, Tokyo) 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 ventilatory equivalents for oxygen and end-tidal PO2 increased without increases in the ventilatory equivalents for carbon dioxide and end-tidal PCO2, alternatively determined by the V-slope method (9, 10).

Assessment of Cardiac Autonomic Nervous Activity

Measurement of HRV.

After each subject had rested lying down for 10 min in a quiet room, ECG signals were recorded in the supine position for an additional 5 min at a rate of 1000 samples per second. The data were stored in a personal computer (PC-9821 La10, NEC). The recordings were checked visually to ensure that all QRS complexes on the ECG were correctly labeled. The beat-to-beat fluctuations were transformed into frequency domains by using fast Fourier transformation. The spectral HRV was expressed as a LF component at 0.04 to 0.15 Hz and a HF component at 0.15 to 0.40 Hz. Because the values were skewed toward larger values, the logarithmic values of the frequency components, log LF, log HF, and log L/H, were used (11). When premature contractions were detected, removal of ectopy was performed (12).

Central parasympathetic efferent activity to the heart is inhibited during inspiration and the inhibitory magnitude diminishes during expiration (13), so the value of log HF represents the central parasympathetic efferent nervous activity in normal subjects. On the other hand, log LF can be generated by both the α-adrenergic effector mechanism controlling total peripheral arterial resistance and the consequently relatively slow fluctuation of HR regulated by the arterial baroreflex for the fluctuation of arterial blood pressure. Thus, log LF is an index of cardiac sympathetic nervous activity as well as baroreflex-mediated cardiac parasympathetic nervous activity (14), and log L/H is considered to mirror sympathovagal balance or to reflect sympathetic modulation (5).

BRS measurement.

Arterial blood pressure was monitored using a finger photoplethysmograph (Finapres Ohmeda 2300) that was capable of providing accurate and reproducible beat-to-beat systolic and diastolic values (15). The respiratory patterns (i.e. the phases of inspiration and expiration) were monitored using an impedance device positioned at the midabdomen. To measure the BRS, phenylephrine was administered in increasing doses (1–2 μg/kg) to determine the dose that would increase the SBP by 20 mm Hg or more. After each bolus injection, the ECG, arterial blood pressure, and respiration were recorded continuously on a multichannel recorder at a speed of 100 mm/s until the peak blood pressure was reached. The R-R intervals obtained during expiration were plotted against the SBP of the preceding beats on a beat-to-beat basis. A computerized linear fit was performed to establish the linear portion of the line of best fit; data points beyond the linear portion of the curve were discarded. Only regression lines that had a correlation coefficient of more than 0.80 or that were statistically significant (p < 0.05) were accepted for analysis. The final slope was obtained by calculating the mean value of at least two measurements; this value (ms/mm Hg) was considered to represent the BRS (16). The BRS values of the invasive and noninvasive methods have proven to be highly correlated (17).

Analysis of HR and SBP recovery.

Although the HR recovery pattern after light or severe exercise is thought to be characterized by one or at least two exponential declines (18), it is clinically convenient and useful to measure the absolute HR or to calculate the change in HR as a function of recovery time when specific conditions, such as patients with cardiac disease, are under consideration. Therefore, values for HR 1, 2, 3, and 4 min after peak exercise and those at 30 s, 1, 2, and 4 min after constant-load exercise were measured, and the changes in HR after peak HR during each exercise protocol were also calculated (during peak exercise protocol: ΔHR1, 2, 3, and 4, respectively; and during constant-load exercise protocol: ΔHR30, 60, 120, and 240, respectively).

The changes in SBP after peak exercise were also calculated during the peak exercise protocol (ΔSBP1, 2, and 4, respectively).

Statistical analysis.

Differences between the two groups in mean cardiorespiratory values during exercise were assessed by 2-way repeated-measures ANOVA. Differences between the two groups at each exercise intensity were evaluated by unpaired t test, and changes in cardiorespiratory variables during exercise were evaluated by paired t test. Data are expressed as mean ± SE, and p values < 0.05 were considered significant.

RESULTS

Cardiopulmonary function at rest.

Although lung volume was significantly lower in group A than in group B, no significant difference in percent predicted values for these variables was observed between the groups. No significant difference between groups A and B was observed either in the left ventricular end-diastolic volume index (mL/m2) or in the ejection fraction of the ventricle.

BRS and HRV.

Values for BRS and HRV in the study groups are summarized in Table 2. No significant difference in BRS was observed between groups A and B. The value of log HF was significantly higher in group A than in group B (p < 0.01), whereas the value of log L/H was significantly lower in group A than in group B (p < 0.05).

Table 2 HRV and BRS for the study groups at rest Values are mean ± SE. *p < 0.05, **p < 0.01 compared with young adults.
Full size table

Cardiovascular response during progressive exercise.

Although there was no difference in change of HR during progressive exercise, SBP was significantly higher in group B than in group A at rest and at peak exercise (Table 3). Because there was no difference in V˙O2 at the AT and at peak exercise, physical fitness was equal in both groups.

Table 3 Cardiovascular variables during progressive exercise Values are mean ± SE. **p < 0.01 compared with young adults.
Full size table

HR and SBP recovery after peak exercise.

Although there was no difference in peak HR between groups A and B, HR at 1 and 2 min after peak exercise was significantly lower in group A than in group B (Table 4, Fig. 1a, p < 0.05), and the changes in HR from peak exercise to each recovery time were significantly greater in group A than in group B (p < 0.01).

Table 4 HR and SBP after peak and constant-load exercise Abbreviations: ΔHR, decrease in HR from peak exercise to the point where the HR was measured after peak exercise; ΔSBP, decrease in SBP from peak exercise to the point where the SBP was measured after peak exercise. Values are mean ± SE. *p < 0.05; **p < 0.01 compared with young adults.
Full size table
Figure 1
figure 1

a, Change in HR after peak exercise testing. In this and subsequent Figures, open and solid circles represent values for children and younger subjects, respectively. ★p < 0.05 vs young adults. b, Change in HR after constant-load exercise testing. ★p < 0.05 vs young adults.

Full size image

Although the values for SBP 1 and 4 min after peak exercise were significantly lower in group A than in group B, the change in SBP during the recovery period was not different between groups A and B (Table 4).

HR after constant-load exercise.

Gas exchange ratio, HR, and V˙O2 were not different between groups A and B at the end of constant-load exercise (Table 4). The HR values in groups A and B were 125 ± 3 and 126 ± 3, respectively, and these values were not different from those obtained at the AT exercise intensity during the peak exercise protocol. The V˙O2 values at the end of constant-load exercise in groups A and B were 19.0 ± 0.6 and 20.5 ± 1.3, respectively, and these two values were not significantly higher than those obtained at AT exercise level during the peak exercise protocol. In addition, because the gas exchange ratios in groups A and B at the end of the constant-load exercise test were 0.92 ± 0.01 and 0.94 ± 0.02, respectively, the exercise intensity used in the present study did not exceed the intensity of the AT exercise level.

Lower HR values were observed throughout the recovery period in group A compared with those in group B, as demonstrated during peak exercise (Fig. 1b). Although only the HR 60 s after constant-load exercise was significantly lower in group A than in group B, the changes in HR from peak values were significantly greater in group A than in group B throughout the recovery period (Table 4). The HR values throughout the recovery period were all significantly higher in group B, whereas those values in group A were not significantly different from HR at rest except at 30 s after the end of constant-load exercise. Conversely, HR 60 s after constant exercise tended to be lower than those at rest and at 180 and 240 s after the end of constant-load exercise (p < 0.08).

Correlations between HRV, BRS, and decrease in HR afterexercise.

Correlations between HRV, BRS, and the change in HR after peak and constant-load exercise testing are summarized in Table 5. BRS did not correlate with the change in HR at any time after either exercise protocol. The log HF value correlated negatively with all the changes in HR after peak exercise, especially with that 1 min after peak exercise (Fig. 2a). On the other hand, the change in HR 30 and 240 s after constant exercise correlated with log HF (Fig. 2b). The value for log L/H also correlated with all the changes in HR after peak exercise, and, the later the period, the stronger the relationship became. However, there were no significant correlations between log L/H and the changes in HR immediately after constant-load exercise except 240 s after exercise.

Table 5 Correlations between HRV, BRS, and decrease in HR after peak and constant exercise tests
Full size table
Figure 2
figure 2

a, Relationship between log HF and decrease in HR 1 min after peak exercise (ΔHR1). b, Relationship between log HF and decrease in HR 30 min after constant-load exercise (ΔHR30).

Full size image

DISCUSSION

HR response during exercise and its recovery after exercise is regulated by several factors: cardiac autonomic nervous system, increased circulating hormones, elevated body temperature, and mechanical stretch of the heart. Therefore, HR recovery after exercise depends on the intensity of exercise and the cardiorespiratory fitness of the subject (2, 3). In the present study, because there was no difference in V˙O2 at the AT and at peak exercise intensity between the two study groups, factors other than physical fitness must affect HR recovery. As mentioned above, because of the influence of exercise intensity on HR recovery, we used two exercise protocols: AT level and peak exercise tests on a treadmill.

Imai et al. (19) demonstrated that the time constant for the first 30 s after exercise was almost independent of exercise intensity and that HR recovery 30 s after exercise, especially at AT exercise intensity, could be a specific index for parasympathetic nervous-mediated HR recovery. In addition, because the elevation of circulating catecholamines (20, 21) and body temperature is probably small at less than AT exercise intensity, a major contributing factor to HR recovery is autonomic nervous activity. Actual parasympathetic nervous activity during the early period was not measured in the present study. However, because ΔHR30 correlated well with log HF, there was probably a tight correlation between parasympathetic nervous activity (tone) at rest and the magnitude of HR recovery early (30–60 s) after light to moderate exercise. Therefore, it is likely that greater central parasympathetic modulation of HR in children is responsible for the faster HR decrease immediately after AT-level exercise compared with that in young adults.

A persistently elevated level of catecholamines after peak exercise (22) and submaximal exercise has been demonstrated, and it has been thought that this elevation affects HR recovery after moderate to severe exercise (23). Although the plasma concentration of norepinephrine does not always reflect sympathetic nervous activity, the relationship between plasma norepinephrine concentration and HR recovery was unclear in the present study because the level of catecholamines was not measured. However, the higher value of log L/H and the lower magnitude of HR recovery suggest that the relatively high sympathetic nervous activity in young adults creates a significant delay in HR recovery compared with that in children.

It is possible that the change in blood pressure after exercise brings about reflex HR change by the arterial baroreceptors, which are important in attenuating the effects of rapid change in blood pressure during daily perturbations (24). Therefore, this reflex may have an important role in HR recovery. However, because no significant differences in BRS at rest or relative change in SBP after peak exercise were observed between children and young adults, it is unclear whether the arterial baroreflex affected HR recovery in the present study. Walgenbach et al. (25, 26) demonstrated that although the arterial baroreflex functioned at the end of exercise to return pressure to normal levels, cardiac output and HR recovery were almost independent of baroreflex function in conscious dogs. According to these reports, BRS may not be related to HR recovery after exercise. The reason HR recovery is faster in children than in adults is unclear from the present study. However, fast HR recovery may be necessary for children to prevent excessive cardiac work after exercise. Thus, enhanced central parasympathetic modulation may play an important role in protecting cardiac overload in children with frequent short intervals of physical activity.

BRS is reset to a higher operating point to permit the arterial blood pressure to increase (27), but it is unknown whether the re-reset of the BRS to a resting level occurs immediately after exercise. However, because HR recovery after AT exercise level, in which little metabolic changes occur, did not correlate with BRS, it seems that the HR recovery after mild to moderate exercise is not related to the peripheral arterial baroreflex function.

The inhibitory mechanism from the pulmonary stretch receptors to the cardiac parasympathetic nerve, which modifies the HF power of the HRV, must be taken into consideration when evaluating the parasympathetic activity from the CNS to the heart (13, 28). However, tidal volume to vital capacity ratio at 30 and 60 s after constant-load exercise in children was not significantly different from that in young adults (30 s: children = 29 ± 2%, young adults = 29 ± 2%; 60 s: children = 21 ± 1%, young adults = 23 ± 2%, NS). Therefore, the magnitude of inhibition from the pulmonary stretch receptor to cardiac efferent parasympathetic nervous activity in children is approximately the same as in young adults, and thus the central efferent parasympathetic control of HR is greater in children than that in young adults.

Study limitations.

First, our subjects cannot be considered entirely normal. However, no significant differences in cardiorespiratory response to exercise between patients with a history of Kawasaki disease and normal subjects have been demonstrated, and even patients with significant coronary arterial lesions showed no significant abnormal findings (29). In our subjects, no stenotic coronary arterial lesions or abnormal hemodynamic parameters were detected by follow-up cardiac catheterization and selective coronary angiography. Consequently, we believe that our present data are comparable to those from normal healthy subjects.

Second, we estimated the cardiac autonomic nervous activity on the basis of analysis of HRV. Pharmacologic methods, as demonstrated in adult subjects by several investigators, may be preferable, but this approach is not always feasible in children. High doses of atropine sometimes cause dizziness and/or thirst, which may influence the results to some extent. Parasympathetic nervous activity estimated by HRV was greater in children than in adults, and this finding is similar to the previously reported results (30).

Our study was also limited by the small number of subjects; however, the data reached statistical significance despite the small sample population.

In conclusion, central parasympathetic efferent nervous activity to the heart is greater in children than in young adults. The greater central cholinergic modulation of HR in children may be responsible in part for children's faster HR recovery, whereas the relatively high contribution of cardiac sympathetic nervous activity in young adults may account for a significant delay in HR recovery after exercise compared with that in children.