INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

LOW HEART RATE VARIABILITY (HRV) and baroreflex sensitivity (BRS) are risk factors for cardiovascular mortality and vulnerability to life-threatening arrhythmias after acute myocardial infarction (1, 8, 11, 12, 14). Recent studies also suggest that low HRV is associated with occurrence of adverse cardiac events in a random population of middle-aged subjects (21). Endurance-trained athletes generally have resting bradycardia, and it has been suggested to be mediated by training-induced increase in vagal outflow (16, 19). Aerobic training was found to substantially increase both HRV and BRS in dogs, and it was suggested that endurance training increases vagal activity that may have an antifibrillatory effect during ischemia (9). Training has been shown to also improve HRV in patients with heart failure (3). However, it is not known how intensively, how often, and for how long one must exercise to obtain favorable changes in HRV in asymptomatic subjects. Furthermore, it is not known whether exercise training has any influence on HRV or BRS in middle-aged subjects with age-related impairment in autonomic regulation of heart rate. The aim of this study was to compare two exercise regimens of varying intensities on HRV and BRS in healthy, sedentary, middle-aged men.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Protocol

After a detailed medical examination and a life-style questionnaire were completed, the subjects were randomized as follows: 1) control group, no supervised exercise during the intervention and a maximum exercise of twice a week; 2) exercise 1, jogging or walking 4-6 times/wk, at a heart rate level corresponding to 55% of the O_2 max measured at baseline; or 3) exercise 2, jogging 4-6 times/wk at a heart rate level corresponding to 75% of the O_2 max measured at baseline. One training session per week was heart rate controlled (Polar Edge, Polar Electro Oy, Kempele, Finland) and supervised at the UKK Institute with adequate warm-up and cool-down. The remainder of the training sessions were performed at home, always pulse controlled, and recorded in a diary. The minimum duration of a session was 30 min at the target heart rate, and the intervention lasted for 5 mo. The study was approved by the Ethics Committee of the UKK Institute, and all subjects gave written, informed consent.

Measurements

Oxygen consumption. All subjects performed a maximal treadmill exercise test according to a standard protocol. A maximum test included at least two of the following criteria: subjective maximum stress by the Borg scale (19-20) was reached; age-predicted maximum heart rate (205  1/2 × age) was reached; oxygen consumption did not increase >150 ml during the last min; or respiratory quotient > 1.1 was observed. The length of the test was recorded as endurance time. Rest and stress ECGs were recorded by the modified Mason-Likar 12-lead system (Marquette Case 12, Marquette Electronics), and O_2 max was measured by a respiratory gas analyzer (Sensormedics 2900Z, Sensormedics BV, Bilthoven, The Netherlands).

BRS. Baroreflex sensitivity was assessed by the phenylephrine method. After 10 min of supine rest, a bolus of phenylephrine (100-200 µg) was administered through an intravenous line, a three-channel ECG was recorded, and blood pressure was measured continuously using a finger plethysmograph (Finapress 2300, Englewood, NJ) (20). An increase in systolic blood pressure (SBP) of at least 20 mmHg was required for a successful test. The test was then repeated three times. BRS was determined from the R-wave-to-R-wave (R-R) interval (RRI) change in relation to change in SBP [BRS = RRI (i + 1) vs. SBP (i) ms/mmHg], was calculated off line using commercial software (Medikro Cafts, Medikro Oy, Kuopio, Finland) and is expressed as the mean of three recordings. Regression lines with a correlation coefficient >0.80 were used for analysis.

HRV. The subjects underwent two-channel, 24-h Holter recording (Oxford Medilog MR-45, Oxford Instruments) on a normal working day, and the recordings were digitized for HRV analysis with a Medilog Excel (version 4.1 c, Oxford Medical) ECG software system. The ECG monitor was attached between 8:00 and 10:00 AM, and the recording was completed the following day at the same time. The ECG was recorded standardized, and the data on the tapes were digitized and transferred to a microcomputer for HRV analysis by a custom-made analysis program (Heart Signal, Kempele, Finland) as described previously (7). A linear detrend was applied to the R-R interval data segments of 512 samples in the spectral analysis of HRV. The R-R interval series was passed through a filter that eliminated premature beats and noise, and gaps were deleted. All time series were first edited automatically and then manually by careful visual inspection of the R-R intervals, as described previously (8). Time sequences corresponding to actual awake and sleeping periods were derived from the subject's diary to ensure correct assessment of HRV during active hours and sleep.

Statistical Methods

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All study subjects successfully completed the 20-wk intervention. However, two men (1 control, 1 exercise) were excluded from the analyses due to <8 h of ECG signal, and one (control) because of being at work during the night the Holter recording was made. Thus the results of 80 men were available for final analyses, and their baseline characteristics are presented in Table 1. Both exercise groups met the requirements for a minimum duration of training sessions at the predefined target heart rate level (with warm-up and cool-down periods). Recreational low-intensity exercise in the control group was allowed according to the study protocol as 2 sessions/wk but was somewhat exceeded. Table 2 summarizes the exercise data and ergospirometry results in the study groups. O_2 max improved in all groups [8.6% control, 11% exercise 1 (P = not significant vs. control), 14.6% exercise 2 (P = 0.05 vs. control)], but the endurance time improved significantly in both exercise groups compared with the control group (Table 2). There were no significant changes in body mass index in the exercise groups compared with the control group (ANOVA), and smoking habits remained unchanged (data not shown).

Resting heart rate, HRV for the 24-h period, and BRS data are summarized in Tables 3 and 4. Resting heart rate increased slightly in the control group, remained unchanged in the exercise 1 group (low intensity), and decreased significantly in the exercise 2 group (high intensity). Changes in BRS within and between the groups were not observed during the 5-mo intervention. Although a trend toward higher values in SDNN, HF, and LF powers were observed in the exercise 2 group compared with the control group, these changes did not reach statistical significance. Changes in HRV indexes during sleep were assessed separately to determine whether changes in a long-resting condition could be observed. Because the definition of sleeping hours was made according to the diary, an accurate time interval for computations was obtained for each subject. However, neither the SDNN, pNN50, nor the frequency domain measures of HRV changed significantly in the study groups during sleep (Tables 3 and 4).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The results of this randomized study suggest that even a 5-mo, well-controlled, successful, high-intensity endurance training program does not have a substantial effect on cardiac autonomic function assessed by measurement of HRV or BRS. Indexes reflecting tonic cardiac vagal outflow, SDNN, pNN50, and HF power (6) did not change significantly during the intervention. Furthermore, no consistent changes were observed in BRS, which is a measure of reflex cardiac vagal responsiveness. Even during sleep, when the sympathetic activity is withdrawn, the pNN50 and HF power were essentially unchanged after intervention.

It is well known that well-trained athletes have lower heart rates compared with their sedentary counterparts, but it has not been adequately addressed whether exercise training is able to improve the cardiovascular autonomic regulation (16). There are several cross-sectional studies that have compared resting heart rate and HRV between well-trained and sedentary subjects. These studies have shown that endurance-trained athletes have significantly higher HRV indexes than sedentary subjects (5, 10, 16, 19), although contradictory findings have also been reported (17). In all these studies, measures of HRV have been compared between well-trained young athletes and their sedentary, age-matched controls, but the results do not provide information as to whether cardiac autonomic function can be improved by exercise training in middle-aged sedentary subjects. Furthermore, heart rate itself significantly influences both the time and frequency domain measures of HRV, and it is not evident whether well-trained athletes really have better cardiac vagal function or merely have a reduced intrinsic heart rate compared with controls.

There are only few prospective, randomized, and controlled studies investigating the effects of exercise training on HRV (2, 15). Boutcher and Stein (2) studied sedentary middle-aged men (n = 19 intervention, n = 15 control) who engaged in moderate-intensity (60% heart rate range) exercise sessions 3 times/wk, for a total of 24 sessions in all. HRV was assessed from a 15-min ECG recording taken during resting conditions. The intervention group showed a marked increase in peak oxygen uptake, whereas the control group showed no change. Concurrent with the present findings, there was a significant reduction in the resting heart rate, but HRV remained unchanged. A comparable negative result in HRV was observed by Davy et al. (4), who studied eight postmenopausal women with elevated blood pressure. The training program included 12 wk of heart rate-controlled walking sessions (60-75% of individual maximal heart rate, 3-4 times/wk). O_2 max, supine resting heart rate, and time and frequency domain measures of HRV were unchanged. The results of these previous studies are limited by either the small sample size, the short duration of intervention, or the lack of randomization, and thus the results cannot be generalized.

Effects of exercise training on cardiovascular autonomic function have also been studied in patients with heart disease or in animal models. Leitch et al. (15) have studied patients with recent uncomplicated myocardial infarction. They randomized 49 patients either to a supervised, heart rate-controlled leg ergometry and circuit training group or to an unsupervised walking group. Exercise sessions were 3-4 times/wk, and the duration of the study was 6 wk. HRV was assessed from 24-h Holter recordings and BRS by the phenylephrine method (15). The endurance time improved significantly in the exercise, but not in the control, group, and no change was observed for O_2 max. In addition, significant improvements in BRS and HRV occurred within the study groups combined but not between the exercise and control groups. In fact, improvements in HRV were even higher in the control group, whereas the BRS improved slightly more in the intervention group, suggesting that the slight improvement in cardiac vagal responsiveness may be a spontaneous phenomenon after myocardial infarction and not necessarily a training effect. In our study, the net change in O_2 max in the high-intensity group was significant and was clearly above that observed by Leitch et al.; however, no change in BRS occurred. Hull et al. (9) studied changes in HRV and BRS in dogs and found that cardiac autonomic function improves after 6 wk of daily exercise and suggested that increased parasympathetic tone by regular exercise might be beneficial in the prevention of sudden cardiac death.

The discordant findings reported in previous studies are most obviously due to differences in patient populations, lack of randomization, and differences in measurement protocols and HRV indexes applied in these studies (2-4, 6, 9, 13, 15-20). The present study indicates that a 5-mo, heart rate-controlled endurance intervention, even when it effectively increases O_2 max and decreases the resting heart rate, is not effective in significantly increasing either the 24-h or sleep HRV indexes. Moreover, BRS was practically unchanged after the intervention, suggesting that significant improvement of vagal reflexes cannot be gained by regular exercise in healthy, sedentary, middle-aged men by a short-term training program. Higher plasma volume, together with increased venous return to the left ventricle, increases stroke volume, which is a well-known effect of endurance training, and results in lower resting heart rate. However, increased HRV and BRS may be a result of another mechanism, such as increased blood flow in the microcirculation of the autonomic nervous system, and may not necessarily be a result of changes in overall hemodynamics. There were no dropouts during the program, and, according to power calculations, the group size was sufficient. In addition, because training sessions were always heart rate controlled and the groups met the requirements (Table 2), the present results are not biased due to lack of sufficient training intensity or frequency or lack of power. Although statistical significance for some of the HRV variables was almost reached, it may be even harder to reach clinical significance in terms of prognosis after myocardial infarction (12). However, the present results cannot be generalized to patients with recent myocardial infarction.

A limitation of the present study is the relatively short duration of intervention, although it was longer than in the earlier randomized studies. These data and the previous studies suggest that, to obtain a clinically significant increase in HRV or BRS, endurance training should be practiced for a prolonged period, probably for many years, but further studies are needed to confirm this hypothesis. However, reduction in intrinsic heart rate occurs even during short-term exercise training that, by itself, without any significant changes in cardiac vagal outflow, may have protective effects on untoward cardiovascular events.