Endurance training has been suggested to result in an adaptation of the autonomic nervous system (ANS) reflected by lower exercise-induced norepinephrine levels, reduced resting and exercising heart rates, and most often increased heart rate variability measures (6,7,10). The sympathetic activity after training is first of all reduced during submaximal exercise but possibly also at rest. Findings from research on adaptations within the parasympathetic part of the ANS after training have, however, not been consistent (14,18).
The measure of fitness level referred to in studies investigating effects of endurance training has often been maximal oxygen uptake (V̇O2max). However, in addition to changes in V̇O2max, a training effect could also be reflected by changes in peripheral indices like muscle strength and lactate production. Relationships between fitness and other variables may depend on whether the actual training mainly affects peripheral or central (cardiorespiratory) systems because they may not be intimately related to each other.
Under certain conditions, sympathetic stimulation improves skeletal muscle performance (8), whereas parasympathetic activity has no known efferent effect on muscles. On the other hand, afferent signals from chemoreceptors in working muscles seem to regulate both sympathetic and parasympathetic outflow to the heart via the so-called metaboreflex (15,16). Thus, the metaboreflex may be one source of reported relationships between training indices and autonomic activity.
Calculating the spectral densities of selected components of the heart rate variability (HRV) gives some insight into the cardiac autonomic control. Parasympathetic activity is reflected in the high-frequency oscillations (HF: 0.15–0.45 Hz) of the HRV spectrum (2,20). Sympathetic modulations of peripheral vascular resistance are believed to be responsible for the low-frequency oscillations in heart rate (LF: 0.04–0.15 Hz) although these oscillations also depend on parasympathetic activity (4,9). The purpose of this study was to investigate whether changes in muscle performance and aerobic power are associated with quantitative or qualitative changes in HRV measures after a 7-month training period.
MATERIAL AND METHODS
We studied 24 subjects of whom 17 were cross-country (X-C) skiers (9 women, 8 men) and 7 were canoeists (2 women, 5 men). Age was 18.5 ± 1.8 yr (mean ± SD). The X-C skiers and canoeists were on regional and national elite standard, and both groups contained world championship participants. They all signed informed consent with parents’ consent as well for subjects under 18. Subjects were informed about the aims and procedure and they all volunteered for the study. The study was approved by the ethical committee of Umeå University.
Muscle strength was measured in an isokinetic dynamometer (Biodex Co., Shirley, NY) to give an estimate of peripheral work capacity. The canoeists performed a shoulder extension-flexion test consisting of 25 repetitions at 120°·s−1. The right arm was used, and the lever allowed subjects to have their elbow flexed in a comfortable position; the lever also allowed the subjects to rotate/invert the wrist. Range of motion was 140° (from −20°, i.e. arm lifted slightly above horizontal, to a fully extended angle of 120°). The shoulder extension in the dynamometer was made to resemble the propelling motion in canoeing. The X-C skiers performed a leg extension test in the same dynamometer. Leg strength was measured in a five-repetition knee flexion-extension test at 90°·s−1, and only right leg data are presented. Range of motion was 90°, i.e. from 90° (flexed) to 0° (extended). Subjects were exhorted to perform maximal contractions from the first repetition. From the strength tests, peak torque (Tq; Nm) and time to peak torque (TiTq; ms) from the best repetitions as well as total work (Wrk; Nm) were obtained. Only extension data are presented. Changes between tests were calculated as relative changes because X-C skiers and canoeists tested different muscles. The relative changes are presented as ΔTq, ΔTiTq, etc., and were calculated as percentages, i.e., the difference between tests divided with the initial value.
Maximal oxygen uptake (V̇O2max) was assessed from an incremental treadmill test after an 11-min warm-up (5-min walking and 6-min running). Slope was increased every 3 min until exhaustion. Every 30 s, oxygen uptake was analyzed in an EOS Sprint gas exchange analyzer (Erich Jaeger GmbH & Co., Würzburg, Germany). When subjects could no longer keep up with the speed of the treadmill, a maximal test was considered to be obtained. At exhaustion, a blood sample for lactate analysis (YSI 1500 Sport, YSI Inc., Yellow Springs, OH) was drawn from an antecubital vein to yield “Lamax.” For each individual, the treadmill tests were identically performed at the two test occasions. Changes in V̇O2max and Lamax during training were converted to relative changes when included in statistical analyses.
Heart rate variability recordings.
After 5 min of supine rest, ECG was recorded during spontaneous breathing for 5 min followed by a 1-min period of controlled breathing at 0.2 Hz (12 breaths·min−1). Subjects were then tilted to a 70° head-up position and recordings proceeded for 5 min.
The two-lead ECG was recorded using a single channel monitor (Hellige Servomed, Freiburg, Germany). Power spectral analysis (frequency domain) of cubic spline interpolated heart rate data was performed by autoregressive modelling as previously described (3). Spectral power of HF (0.15–0.45 Hz) and LF (0.04–0.15 Hz) heart rate variability components were calculated and log transformed. Resting supine HRV data represent recordings during controlled breathing (12 breaths·min−1). Power is presented both as absolute (e.g., LF) and normalized values (LFn). In normalized units, the power of the actual HRV-component was divided by the sum of the LF and HF components and then multiplied with 100.
HRV recordings, strength tests, and treadmill tests were performed twice; immediately after a low intensity period (i.e., September for the X-C skiers and February for the canoeists), and after the competitive season (April and end of August for X-C skiers and canoeists respectively). HRV recordings were performed in the morning, always before the physical performance tests.
For comparisons of data before and after the observation period, a paired t-test was used. For bivariate correlations, the Pearson correlation coefficient was calculated. A P-value less than 0.05 led to rejection of the null hypothesis. When investigating whether changes in V̇O2max was associated with HRV variables, a repeated measures ANOVA was performed. For this analysis, the group was divided into two equally big subgroups based on the subjects’ relative changes in V̇O2max. The SPSS® statistical software was used.
Of the HRV variables, LF-power in standing was lower after training than before (in absolute but not in normalized values) (P < 0.05). Total and HF-powers were not significantly changed after training. Resting heart rate decreased from 60 ± 5 to 57 ± 6 beats·min−1 (P = 0.01).
Of the strength measures, Tq was increased and TiTq was reduced after the observation period (P < 0.05). Total work also increased although not significantly (0.05 <P < 0.10). The change in peak torque (ΔTq) was similar in the canoeists and X-C skiers. Regarding TiTq (contraction velocity), the reduction was almost exclusively seen in the skiing group (only 1 of 7 canoeists). However, they were tested in different muscles and they also performed different training.
Bivariate correlations and analysis of variance.
The change between tests in absolute LF-power of HRV in the tilted position (ΔLFT) was inversely related to ΔTq (r = −0.46, P < 0.05;Fig. 1 A) and to the change in maximal lactate, ΔLamax (r = −0.42, P < 0.05). The change between tests in normalized LF-power on tilt (ΔLFnT) was significantly correlated with the change in time to peak torque (ΔTiTq; r = 0.60, P < 0.01;Fig. 1 B). ΔLFnT was further inversely correlated with the change in V̇O2max (r = −0.56, P < 0.01) and with the change in Lamax (r = −0.60, P < 0.01;Fig. 1 C and D). Together the results show that a shift of HRV toward the LF-component (in the tilted position) was associated with a reduced peripheral (Tq, TiTq, and Lamax) and central (V̇O2max) performance in our subjects. The change in total work (ΔWrk) did not show any linear relationship with HRV variables.
Repeated-measures ANOVA showed that subjects who increased their V̇O2max during the training period presented with clearly higher HF (both absolute and normalized units) and total power throughout the study compared with those who showed reduced V̇O2max (P < 0.05;Fig. 2). The mean change in V̇O2max in these two subgroups (both N = 12) were −0.16 L·min−1 (range: −0.42 to 0.01) and 0.15 L·min−1 (range: 0.04–0.37), respectively. With all subjects in the calculation, the changes in V̇O2max between tests were however not followed by corresponding changes in HF or total heart rate variability. LF variability after tilt (nu) was affected by gender with female subjects showing lower values than the male subjects.
The purpose of this study was to investigate whether long-term changes in muscular performance and changes in V̇O2max, representing peripheral and central measures of fitness, are associated with quantitative or qualitative changes in heart rate variability. Between tests we found decreased LF-power in the tilted position but no change in HF or total HRV components. Regarding muscle performance, maximal Tq was uniformly improved in both canoeists and X-C skiers, but time to Tq was improved, i.e., shortened, only in the skiers. Different training and testing modalities in the two groups might explain this discrepancy. It should further be stressed that V̇O2max and most HRV parameters did not show any significant differences on group level after versus before training. This may partly depend on the high level of fitness already at the start of the training period. Besides, the correlations presented on this relatively small sample are not strong enough to be conclusive but may bring new viewpoints to future research on relationships between fitness and cardiac autonomic control.
Parasympathetic activity has often been estimated from the heart rate response to atropine, respiratory sinus arrhythmia (RSA), or other measures of respiratory related HRV (11,13,14). Although several reports, using the different methods above, have shown higher parasympathetic activity in endurance-trained athletes compared with controls, this response to training has been questioned by others (14). When comparing subjects with varying aerobic capacities, Goldsmith et al. (7) showed a relationship between the power in the HF component of HRV and aerobic power. According to previous research, HF variability and also total HRV have been shown to reflect parasympathetic activity (1,20). In our study, neither the changes in V̇O2max, nor the changes in peripheral measures (i.e., Tq, TiTq, and Lamax) were followed by corresponding changes in HF or total HRV. However, we found consistently higher HF and total power in the subjects who improved V̇O2max during the season. Based on the findings above, one would suggest that higher levels of HF and total power constituted a favorable condition to increase V̇O2max in our subjects, rather than being an effect of training. If that is true, maybe those athletes who achieve the greatest V̇O2 increases, and thus the highest V̇O2 values, have higher HRV (HF and total) already from the beginning. A resulting speculation is that the level of parasympathetic activity (estimated from HF and total HRV) reflects an inherited quality that affects the cardiovascular responsiveness to training. The lack of a linear relationship between the changes in fitness measures and parasympathetic activity indicate that training per se did not seem to have a major influence on the parasympathetic adaptation. Such a conclusion should, however, be interpreted with caution because the relatively small sample size, together with small changes in V̇O2max and widely distributed HRV data, make it difficult to show significant relationships. The small changes in V̇O2max are probably due to the high aerobic level at start in these subjects. Although there were no significant changes in V̇O2max on group level, we found it interesting that some factors (HF and total HRV levels) might have influenced why some of the subjects in fact did improve V̇O2max after training whereas others did not.
Explanations of the different findings regarding relationships between fitness (V̇O2max) and parasympathetic activity in previous reports could be differences in methods used (to assess parasympathetic activity and fitness) or more complex relationships between the two variables. Besides, many of the studies investigating relationships between parasympathetic activity and fitness are cross-sectional and cannot determine the specific effect of training (5,7,14,17). Another problem is the definition of fitness, which can be referred to as levels of measured or estimated V̇O2max, training volume, or other, less specific measures. Of course, training itself is a prerequisite for fitness in general, but if fitness is estimated from V̇O2max, fitness should also depend on the cardiovascular response to the experienced training. Our data support the theory of higher parasympathetic activity in the more well-trained state but suggest that the higher parasympathetic activity, at least in these fit subjects, rather was a cause than an effect of a further increase in aerobic fitness. There may however be factors not controlled for here, such as changes of HRV during the preceding years of training, that affect our results.
When planning the study, we assumed that an increase in peak torque (Tq) and a reduction in time to peak torque (TiTq) during arm and leg extension would reflect an improved muscle performance after training. Summarizing the findings of the bivariate correlations, there was a relationship between changes in LF-power and changes in muscle performance as well as changes in V̇O2max, i.e., when LF-power (tilt) decreased, muscular performance and V̇O2max increased. LF-power, particularly after a tilt, has in several reports been used to represent sympathetic activity although, in more recent reports, the source of LF has been debated. In general, LF-power is thought to reflect both sympathetic and parasympathetic activities (2). The association with sympathetic activity originates from the blood-pressure changes caused by sympathetic vasomotion, which via the baroreflex affects heart rate (4). Thus, sympathetic modulation of the vascular response after a tilt should appear in the LF component of HRV. If ΔLFnT reflects a change in sympathetic activity, the relationships reported above suggest that an improvement in both muscle and aerobic performance is associated with reduced sympathetic activity at rest.
The change in LFn in the tilted position (ΔLFnT) showed a highly significant linear relationship with changes in both strength measures and aerobic capacity. This suggests that the observed changes in LFnT reflect a general adaptation to training and also, in contrast to HF-power, responded to training over time in our subjects. Even though the changes in performance and LF-power were intimately related to each other, they may be two independent results of the training program.
One possible source of the relationships between HRV and muscle performance (and aerobic capacity) may be the metaboreflex, which is induced by muscle ischemia. When activated, e.g., during exercise, this reflex has been shown to first increase sympathetic activity and afterward also increase parasympathetic activity and heart rate variability (15,16). Activation of skeletal muscle may thus lead to adaptations in HRV. Further, sympathetic activation has a direct effect on skeletal muscle by improving performance in fatigued fast-contracting muscle (refs in (8)). On the other hand, muscle blood flow, and thus oxygen consumption, seems to be reduced during sympathetic stimulation both at rest and during exercise (12,19). These findings could correspond with our results of an inverse relationship between LF-power and peripheral and central (aerobic) capacities.
In our subjects, high levels of HF and total heart rate variability seemed to constitute a favorable condition to increase V̇O2max rather than being a result of increased aerobic fitness. The factors responsible for the LF variations in heart rate, on the other hand, seemed to respond over time in inverse proportion to changes in both peripheral and central measures of fitness.
The study was financially supported by grants from the Swedish National Center for Research in Sports. The authors thank Tom Pietilä for assistance with strength measurements, and Erkki Jakobsson, Lennart Burlin, and Rolf Hörnsten for assistance during treadmill tests and HRV recordings.
Address for correspondence: Rikard Hedelin, Sports Medicine Unit, Umeå University, S-901 87 Umeå, Sweden; E-mail: [email protected]
1. Akselrod, S., D. Gordon, J. B. Madwed, N. C. Snidman, D. C. Shannon, and R. J. Cohen. Hemodynamic regulation: investigation by spectral analysis. Am. J. Physiol. 249: H867–H875, 1985.
2. Akselrod, S., D. Gordon, F. A. Ubel, D. C. Shannon, A. C. Barger, and R. J. Cohen. Power spectrum analysis of heart rate fluctuation: a quantitative probe of beat-to-beat cardiovascular control. Science 213: 220–222, 1981.
3. Blom, H., C. Andersson, B.-O. Olofsson, P. Bjerle, U. Wiklund, and F. Lithner. Assessment of autonomic nerve function in acute intermittent porphyria: a study based on spectral analysis of heart rate variability. J. Intern. Med. 240: 73–79, 1996.
4. Deboer, R. W., J. M. Karemaker, and J. Strackee. Hemodynamic fluctuations and baroreflex sensitivity in humans: a beat-to-beat model. Am. J. Physiol. 253: 680–689, 1987.
5. Dixon, E. M., M. V. Kamath, N. Mccartney, and E. L. Fallen. Neural regulation of heart rate variability in endurance ath-letes and sedentary controls. Cardiovasc. Res. 26: 713–719, 1992.
6. Ekblom, B., Å. Kilbom, and J. Soltysiak. Physical training, bradycardia, and autonomic nervous system. Scand J. Clin. Lab Invest. 32: 251–256, 1973.
7. Goldsmith, R. L., J. T. Bigger, Jr., D. M. Bloomfield, and R. C. Steinman. Physical fitness as a determinant of vagal modulation. Med. Sci. Sports Exerc. 29: 812–817, 1997.
8. Grassi, C., and M. Passatore. Action of the sympathetic system on skeletal muscle. Ital. J. Neurol. Sci. 9: 23–28, 1988.
9. Grasso, R., F. Schena, G. Gulli, and A. Cevese. Does low-frequency variability of heart period reflect a specific parasympathetic mechanism? J. Autonom. Nerv. Syst. 63: 30–38, 1997.
10. Hartley, L. H., J. W. Mason, R. P. Hogan, et al. Multiple hormonal responses to graded exercise in relation to physical training. J. Appl. Physiol. 33: 602–606, 1972.
11. Janssen, M. J., J. de Bie, C. A. Swenne, and J. Oudhof. Supine and standing sympathovagal balance in athletes and controls. Eur. J. Appl. Physiol. 67: 164–167, 1993.
12. Joyner, M. J., L. A. Nauss, M. A. Warner, and D. O. Warner. Sympathetic modulation of blood flow and O2
uptake in rhythmically contracting human forearm muscles. Am. J. Physiol. 263: H1078–H1083, 1992.
13. Katona, P. G., and F. Jih. Respiratory sinus arrhythmia: noninvasive measure of parasympathetic cardiac control. J. Appl. Physiol. 39: 801–805, 1975.
14. Katona, P. G., M. Mclean, D. H. Dighton, and A. Guz. Sympathetic and parasympathetic cardiac control in athletes and nonathletes at rest. J. Appl. Physiol. 52: 1652–1657, 1982.
15. Nishiyasu, T., N. Tan, K. Morimoto, M. Nishiyasu, Y. Yamaguchi, and N. Murakami. Enhancement of parasympathetic cardiac activity during activation of muscle metaboreflex in humans. J. Appl. Physiol. 77: 2778–2783, 1994.
16. O’Leary, D. S. Autonomic mechanisms of muscle metaboreflex control of heart rate. J. Appl. Physiol. 74: 1748–1754, 1993.
17. Rossy, L. A., and J. F. Thayer. Fitness and gender-related differences in heart period variability. Psychosom. Med. 60: 773–781, 1998.
18. Smith, M. L., D. L. Hudson, H. M. Graitzer, and P. B. Raven. Exercise training bradycardia: the role of autonomic balance. Med. Sci. Sports Exerc. 21: 40–44, 1989.
19. Thompson, L. P., and D. E. Mohrman. Blood flow and oxygen consumption in skeletal muscle during sympathetic stimulation. Am. J. Physiol. 245: H66–H71, 1983.
20. Uusitalo, A. L., K. U. Tahvanainen, A. J. Uusitalo, and H. K. Rusko. Non-invasive evaluation of sympathovagal balance in athletes by time and frequency domain analyses of heart rate and blood pressure variability. Clin. Physiol. 16: 575–588, 1996.