Heart rate variability (HRV), the spontaneous fluctuations around the mean heart rate, is a simple noninvasive measure that reflects the autonomic balance. A reduced HRV is associated with increased incidence of total mortality and cardiac events in both post infarction patients (3,12,16) and apparently healthy middle aged and elderly subjects (8,26,27). It is assumed that a reduced HRV is a reflection of elevated sympathetic activity, a condition that may decrease the fibrillation threshold and thus predispose to ventricular fibrillation (23). With increasing age the respiratory variation in heart rate decreases (20) and baroreceptor reflexes attenuate (11,25) and HRV is reported to be lower in elderly people (22).
It is widely presumed that regular physical activity induces adaptations in the autonomic nervous system. One of the possible adaptations is an increase in parasympathetic activity and HRV (9,10,14,24). Until now most intervention studies have been performed in chronic heart failure or myocardial infarction patients (1,2,7,15,18,19). The majority of these studies report a considerable and significant increase in HRV after a period of physical training (2,7,16,19). Only two studies have been performed in healthy middle aged and elderly subjects, using relatively short recordings. These show inconsistent results. Seals et al. (24) observed a 15% significant increase in HRV (SD of RR-intervals of a 5-min recording) at rest in healthy middle aged and elderly sedentary men, who followed a strenuous and prolonged endurance training program. In contrast, Boutcher et al. (5) recently failed to show an increase in HRV in healthy middle aged men after a training period. However, the training protocol of Boutcher was shorter and of moderate intensity.
Since the elderly have both increased incidence of cardiac events and a reduced HRV, it is important to learn more about the effects of regular physical activity on HRV in this age group. Therefore, we have studied change in HRV, both in the time and frequency domain in older men and women who were randomly allocated to a 6 months' training program and in a control group, using 24-h ECG recordings.
The study population comprised Caucasian people, 60- to 80-yr old, living in a middle sized city in the Netherlands. Participants took part in an intervention study which investigated the effect of a six months' training program on various cardiovascular risk factors. These subjects were recruited via their general practitioner. Of the invited subjects, 38% were willing to participate in the original study. Participants were younger (men: 68.5 ± 5.5 vs 71.5 ± 5.6 yr; women: 67.8 ± 5.4 vs 70.5 ± 5.4 yr) and less often considered themselves inactive as compared with their peers (men: 49% vs 77%; women: 48% vs 77%), in comparison with those who refused. In total, 74 older men and women, randomly recruited from the intervention and control groups, participated in the present study. These subjects complied with the following inclusion criteria: normal resting and exercise electrocardiogram, no myocardial infarction or stroke in the past 2 yr, no insulin-dependent diabetes mellitus, no heart failure, no unstable angina pectoris, and no use of beta blockers. Fourteen preintervention ambulatory electrocardiogram (ECG) recordings could not be used, because of bad quality tapes (N = 12), pacemaker (N = 1), and atrial fibrillation (N = 1). From the remaining 60 subjects, three provided unsuitable postintervention HRV measurements. One subject dropped out during the intervention and five subjects were not able to do the postintervention ECG measurement. Therefore, we were left with a group of 51 older subjects with an adequate pre- and postintervention ECG measurement. All participants provided written informed consent, and the study was approved by the Medical Ethical Committee of the Agricultural University Wageningen.
Ambulatory ECG Recording
At the beginning and end of the intervention period, subjects were visited at home by the research assistant who attached the ECG electrodes, started the 3-lead Marquette 8500 Holter recorder (Marquette Electronics Inc. Milwaukee, WI) and instructed the participants how to disconnect the Holter recorder after 24 h. All tapes (pre- and postintervention measurements) were analyzed after the intervention period, in a manner blinded to group and order, to avoid bias.
Assessment of HRV
Signal processing. The electrocardiogram was analyzed with a Marquette Series 8000 Holter Analyzer. Accurate determination of the onsets of the QRS-complexes was accomplished by an extensive review and edit procedure and a computer program from the Marquette 8000 Holter research software modules. The resulting interbeat interval series were analyzed further on a personal computer (4). In case of ectopic beats, the surrounding intervals were replaced by interpolated ones. Five-minute episodes containing more than 10% of isolated ectopics were excluded from the analysis. Episodes containing ectopic runs were also excluded from the analysis.
Power spectral analysis. Frequency domain analysis was performed by calculating the power density spectrum using the fast Fourier algorithm. We computed high frequency (HF, 0.15-0.40 Hz), low frequency (LF, 0.05-0.15 Hz), and very low frequency power (VLF, 0.01-0.05 Hz) for every 5-min interval (4).
Time domain analysis. The SD of all normal RR intervals (SDNN) and the percentage of differences between adjacent normal RR intervals exceeding 50 ms (pNN50) were calculated for every 5-min interval.
Maximal oxygen consumption (O2max) and maximal exercise capacity (Wmax) were determined in a maximal bicycle ergometry test before and after the intervention period. The protocol started with a load of 30 W for women and 60 W for men, and every 3 min the workload was increased with 30 W. We used this protocol to approach a steady state at each level to be able to evaluate intercurrent ischemia. Body mass index was calculated as body weight (kilograms) divided by height squared (m2) before and after the intervention. Smoking status was assessed during an interview. Physical activity was assessed with the Zutphen Study Questionnaire, which has been described (6) and validated (28). "(In)active in sport" was defined by the respective answers "no" and "yes" to the question "Have you performed any sport activity lately?"
The 74 subjects who participated in this study were randomly allocated to a training (N = 36) and a control group (N = 38). During half a year, subjects of the training group gathered three times per week in a sports facility for a 45-min supervised training session. The training started with a 10-min warm-up, consisting of walking or slow jogging, followed by some flexibility and stretching exercises. Then the training was continued with a 30-min main program. The main program consisted of aerobic exercise, such as outdoor jogging, exercise to music (once a week), a ball game involving running such as basket ball or indoor baseball (once a week), or calisthenics (once a week). During the main program the instructor saw to it that the intensity of the training was adapted to the individual's capacity. The aim was to exercise at a level that was between 60-80% of an individual's maximum capacity. The attendance of subjects finally included in the analysis was 72% for men and 81% for women. Subjects in the control group were asked to maintain their habitual activities during the 6 months but were not restrained from undertaking (more) physical activity.
Variables with a skewed distribution (HF, LF, and VLF) were log transformed. Means of SDNN and pNN50 and arithmetic and geometric mean of HF, LF, and VLF (all from 5-min averages) were calculated for the complete 24 h recording, during the active period of the day (between 6 a.m. and 6 p.m.), and during sleep (between 2 and 5 a.m.). General characteristics at baseline were calculated for all subjects with an acceptable Holter preintervention measurement.
Mean change in physical fitness (O2max, Wmax) and the various HRV measures in the intervention and control groups were compared using Student's t-test. To investigate the effect of the training program in more sedentary subjects, we repeated the analysis including only subjects who were not engaged in sport activities (such as swimming, gymnastics, etc.) at baseline. To evaluate whether changes in HRV parameters were affected by changes in physical fitness and mean heart rate, a regression analysis was performed in which HRV parameters were dependent variates and assignment (intervention, control) and change in physical fitness and mean heart rate were independent variates.
The association between physical fitness (O2max and Wmax) and HRV, both baseline levels as well as change during the intervention period, were evaluated using regression analysis. In the baseline analysis, adjustment was made for sex, smoking, and age.
In Table 1 baseline characteristics of the study population are depicted separately for all subjects in the intervention and control groups who had a usable preintervention Holter tape. Subjects in the intervention group were younger (P < 0.05) and had a higher body mass index. About half the subjects were already regularly active in sports at baseline. Subjects with a missing postintervention measurement (loss to follow up) did not significantly differ from subjects who completed the protocols with respect to age, body mass index, blood pressure, or HRV characteristics at baseline.
The mean baseline SDNN, LF, and VLF, but not HF, were associated with the baseline fitness level (Table 2); however, after adjustment for sex, smoking, body mass index, and age the association attenuated and became insignificant.
Table 3 shows the mean values of physical fitness and HRV at baseline and after the 6 months' training period for subjects who completed the protocols. Despite randomization, mean values of baseline HF, SDNN, and pNN50 were lower in the intervention group as compared with the control group, although not statistically significant. After the intervention period, subjects in the training group significantly increased both O2max as well as Wmax as compared with the control group. The mean SDNN and pNN50 of the total 24 h recording were modestly increased in the intervention group (5% and 13%, respectively), whereas the control group showed a small decrease (P-value of difference: 0.11 and 0.32). The average changes in 24-h HF, LF, and VLF were positive in the intervention group and negative in the control group; however, they were rather small and not significantly different between the intervention and control groups. During the day (between 6 a.m. and 6 p.m.), mean increases in SDNN, pNN50, LF, and VLF in subjects in the intervention group were more pronounced (6, 16, 15, and 10%) and, except for pNN50, significantly different from those in the control group (P = 0.02, P = 0.32, P = 0.05, and P = 0.02, respectively). Mean change in HRV characteristics between 7 a.m. to 7 p.m. or between 8 a.m. to 8 p.m. showed similar results.
When we restricted analysis to subjects who were inactive in sports at baseline (N = 25), we observed a larger increase in HRV in the intervention group. Mean increases in 24-h SDNN, LF, and VLF were 11, 42, and 20%, with P-values for the difference with the control group of 0.08, 0.03, and 0.04, respectively. Again, among the subjects of the intervention group, increase in HRV was most pronounced during the day (SDNN: + 17%, P < 0.01, pNN50: + 104%, P = 0.08, HF: +40%, P = 0.09, LF: +54%, P < 0.01, VLF: +31%, P < 0.01).
Change in SDNN and HF were significantly associated with a change in heart rate (crude: −0.98 and −0.015 (10−3 s(2)) per 1 beat·min−1, respectively), but not with change in O2max, or Wmax. However, the difference in change of the HRV parameters between the intervention and control groups was not attenuated after adjustment for changes in heart rate.
In a randomized intervention study, we studied the effect of a 6 months' training program on heart rate variability, a characteristic of autonomic nervous function, from 24 h ambulatory ECG recordings of an older population. Subjects in the intervention group showed a small increase in total 24-h HRV (SDNN) and in the very low and low frequency component (VLF, LF), but a moderate and significant increase in daytime-HRV, compared with the control group in which a small decrease was observed. The differences were most pronounced and significant among subjects who were sedentary at baseline and independent of change in mean heart rate and weight.
Increasing HRV may be important since low HRV, particularly common in the elderly, was recently observed to be associated with increased incidence of cardiac events and total mortality in a healthy population (8,26,27). In these studies all HRV measures in the time and frequency domain were associated with increased risk of cardiac events (an-'gina pectoris, myocardial infarction, coronary heart disease (CHD) death, or congestive heart failure). One SD decrement in log SDNN was associated with a relative risk of 1.5 (95% confidence interval 1.2 to 1.9). One SD decrement in log LF was associated with a 1.7 times greater hazard of total mortality.
The observed small reduction in HRV among subjects in the control group may be caused by a reduction in physical activity during the intervention period. The study was conducted during the winter months to have a lower level of background physical activity and to minimize absence at the training sessions because of holidays. Also our control subjects had spent less time on walking (− 65 min·wk−1), and gardening (− 3.5 h·wk−1), but not on sports or bicycling, than during the summer.
Since baseline levels of some HRV measures (SDNN, HF, and especially pNN50) were somewhat different between the intervention and control groups, regression toward the mean in these parameters cannot be ruled out. It remains unclear to what extent this phenomenon may have affected the results of the study.
The ambulatory ECG measurement may have been subject to possible errors. As previously mentioned, a large proportion of the tapes were of insufficient quality (11%). We believe that these instances of poor quality were random since they were caused by bad signal conduction and (except for two) not by abnormalities in the ECG.
Unlike the short-term standardized recording performed in a laboratory, the 24-h Holter recording enables studying HRV under natural conditions, and during wake and sleep. Differences in physical activity pattern during Holter recording might hamper comparability between the intervention and control groups. However, participants were instructed not to exercise, and since the Holter recorder discouraged undertaking other strenuous activities, it is not likely that the activity pattern was markedly different between the intervention and control groups nor within the groups in pre and postintervention measurements. Still, differences cannot be completely ruled out.
Subjects who were using beta-blockers were excluded from the study, and other medications that can influence the heart or the central nervous system were well balanced in the intervention and the control groups. This medication was not altered during the intervention period. Only two subjects had suffered from CHD in the past, and one subject had non-insulin-dependent diabetes mellitus. Exclusion of these subjects did not alter the outcome of the study.
Other Intervention Studies
The majority of intervention studies investigating the effect of physical training on HRV has been performed in CHD patients (2,7,16,18,19) and most of these studies have observed significant and considerable increases in HRV. However, to our knowledge, only five studies have reported the effect of regular exercise during longer periods on HRV in healthy subjects. Three of these were carried out in young and two in middle aged and older subjects. The studies performed in young subjects show an increase in SDNN (21) and LF (13,21) of about 26%. The change reported in HF was inconsistent: two studies reported an unexpected decrease in HF in the range of 12 to 60% (13,21), whereas one study reported a significant increase in HF in 9 of the 11 individuals who followed a 6-wk intensive training program (2).
The studies performed in middle aged and elderly subjects show inconsistent results. Seals and Chase (24) observed a moderate increase of 15% in SDNN after 30 wk of training, while Boutcher did not find any effect after a moderate training program of only 3 months. The training effect we observed in our study is similar to the results of Seals et al. (24) The results so far may suggest that training modulates HRV beneficially only when it is performed vigorously enough.
Until now most training intervention studies in healthy subjects and patients have used short-term ECG recordings. Those who have reported an effect obviously found it during the day. Only one study has performed a 20-h recording in congestive heart patients (16), but also in this study increase in HRV (in the high frequency domain) was observed only during the day. Our results are in line with these studies. We do not, however, have an explanation for this other than a greater challenge to circulatory homeostasis during the day.
HRV represents vagal and sympathetic cardiac control. HF reflects uniquely vagal control, while the lower frequencies reflect unknown mixes of vagal and sympathetic control. In the intervention group, during the day SDNN, LF, and VLF increased significantly. HF and pNN50 (also considered an index for vagal control), did not change significantly, although the average value of pNN50 after intervention was 16% larger than at baseline. This study demonstrated that exercise training of older people modifies their HRV. Interpretation of these changes is difficult. The increase in pNN50 denotes increased vagal control, but this difference did not reach statistical significance. LF heart rate variability, also known as Mayer waves, is often interpreted as baroreflex resonance. The increase in LF might therefore be the result of an increased baroreflex sensitivity, a measure that is known to increase with exercise training. Nevertheless, if the association between physical activity and HRV, on the one hand, and the association between HRV and incidence of cardiac events, on the other, is confirmed, physical activity may have preventive implications.
In conclusion, a 6 months' intensive training program resulted in a moderate increase in heart rate variability, primarily during the day, among a group of older subjects in both time domain (SDNN) as well as frequency domain (VLF, LF). Hence, in older subjects physical training may be an effective means to modify positively a factor that is associated with increased risk of cardiac events.
1. Adamopoulos, S., P. Ponikowski, E. Cerquetani, et al. Circadian pattern of heart rate variability
in chronic heart failure patients: effects of physical training. Eur. Heart J.
2. Alani, M., S. M. Munir, M. White, J. Townend, and J. H. Coote. Changes in RR variability before and after endurance training measured by power spectral analysis and by the effect of isometric muscle contraction. Eur. J. Appl. Physiol. Occup. Physiol.
3. Bigger, J. T., Jr., J. L. Fleiss, L. M. Rolnitzky, and R. C. Steinman. Frequency domain measures of heart period variability to assess risk late after myocardial infarction. Circulation
4. Bootsma, M., C. A. Swenne, H. H. Van Bolhuis, P. C. Chang, V. M. Cats, and A. V. G. Bruschke. Heart rate and heart rate variability
as indexes of the sympathovagal balance. Am. J. Physiol. (Heart Circ. Physiol)
5. Boutcher, S. H. and P. Stein. Association between heart rate variability
and training response in sedentary middle aged men. Eur. J. Appl. Physiol.
6. Caspersen, C. J., B. P. M. Bloemberg, W. H. M. Saris, R. K. Merritt, and D. Kromhout. The prevalence of selected physical activities and their relation with CHD risk factors in elderly
men: the Zutphen Study. Am. J. Epidemiol.
7. Coats, A. J., S. Adamopoulos, A. Radaelli, et al. Controlled trial of physical training in chronic heart failure: exercise performance, hemodynamics, ventilation, and autonomic function. Circulation
8. Dekker, J. M., E. G. Schouten, P. Klootwijk, J. Pool, C. A. Swenne, and D. Kromhout. Heart rate variability
from short ECG recordings predicts mortality from all cause in middleaged and elderly
men: The Zutphen Study. Am. J. Epidemiol
9. De Meersman, R. E. Heart rate variability
and aerobic fitness. Am. Heart J.
10. Dixon, E. M., M. V. Kamath, N. McCartney, and E. L. Fallen. Neural regulation of heart rate variability
in endurance athletes and sedentary controls. Cardiovasc Res.
11. Duke, P. C., J. G. Wade, R. F. Hickey, and C. P. Larson. The effect of of age on baroreceptor reflex function in man. Can. Anesthesiol. J.
12. Farrell, T. G., Y. Bashir, T. Cripps, et al. Risk stratification for arrhythmic events in postinfarction patients based on heart rate variability
, ambulatory electrocardiographic variables, and the signal-averaged electrocardiogram. J. Am. Coll. Cardiol.
13. Furlan, R., S. Piazza, S. Dell'orto, et al. Early and late effects of exercise and athletic training on neural mechanisms controlling heart rate. Cardiovasc. Res.
14. Goldsmith, R., J. T. Bigger, R. Steinman, and J. Fleiss. Comparison of 24 hour parasympathetic activity in endurance-trained and untrained young men. J. Am. Coll. Cardiol.
15. Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. Heart rate variability
: standards of measurement, physiological interpretation and clinical use. Circulation
16. Kiilavuori, K., L. Toivonen, H. Näveri, and H. Leinonen. Reversal of autonomic derangements by physical training in chronic heart failure assessed by heart rate variability
. Eur. Heart J.
17. Kleiger, R. E., J. P. Miller, J. T. Bigger, Jr., and A. J. Moss. The Multicenter Post-infarction Research Group: decreased heart rate variability
and its association with increased mortality after acute myocardial infarction. Am. J. Cardiol.
18. La Rovere, M. T., A. Mortara, G. Sandrone, and F. Lombardi. Autonomic nervous system adaptations to short-term exercise training. Chest
19. Malfatto, G., M. Facchini, R. Bragato, G. Branzi, L. Sala, and G. Leonetti. Short and long term effects of exercise training on the tonic autonomic modulation of heart rate variability
after myocardial infarction. Eur. Heart J.
20. Pfeiffer, M. A., C. R Weinberg, D. Cook, J. D. Best, A. Reenan, and J. B. Halter. Differential changes of autonomic nervous system function with age in man. Am. J. Med.
21. Sacknoff, D. M., G. W. Gleim, N. Stachenfeld, and N. L. Coplan. Effect of athletic training on heart rate variability
. Am. Heart J.
22. Schwartz, J. B., W. J. Gibb, and T. Tran. Aging effects on heart rate variation. J. Gerontol.
23. Schwartz, P. J. and S. G. Priori. Sympathetic nervous system and cardiac arrhythmias. In: Cardiac Electrophysiology: From Cell to Bedside.
D. P. Zipes and Y. Jalife (Eds.). Philadelphia: W. B. Saunders, 1990, pp. 330-334.
24. Seals, D. R. and P. B. Chase. Influence of physical training on heart rate variability
and baroreflex circulatory control. J. Appl. Physiol.
25. Shimada, K., K. Tadao, H. Ogura, N. Sadakane, and T. Ozawa. Differences in age-independent effects on blood pressure on baroreflex sensitivity between normal and hypertensive subjects. Clin. Sci.
26. Tsuji, H., F. J. Venditti, E. S. Manders, et al. Reduced heart rate variability
and mortality risk in an elderly
cohort: The Framingham Heart Study. Circulation
27. Tsuji, H., M. G. Larson, F. J. Venditti, et al. Impact of reduced heart rate variability
on risk for cardiac events: The Framingham Heart Study. Circulation
28. Westerterp, K. R., W H. M. Saris, B. P. M. Bloemberg, K. Kempen, C. J. Caspersen, and D. Kromhout. Validation of the Zutphen Physical Activity
Questionnaire for the Elderly
with doubly labeled water. Med. Sci. Sports Exerc.