Abstract
present study was to determine if autonomic heart rate modulation, indicated by heart rate variability (HRV), differs during supine rest and head-up tilt (HUT) when sedentary and endur- ance-trained cyclists are compared. Eleven sedentary young men (S) and 10 trained cyclists (C) were studied. The volunteers were submit- ted to a dynamic ECG Holter to calculate HRV at rest and during a 70º HUT. The major aerobic capacity of athletes was expressed by higher values of V . O 2 at anaerobic threshold and peak conditions (P < 0.05). At rest the athletes had lower heart rates (P < 0.05) and higher values in the time domain of HRV compared with controls (SD of normal RR interval, SDNN, medians): 59.1 ms (S) vs 89.9 ms (C), P < 0.05. During tilt athletes also had higher values in the time domain of HRV compared with controls (SDNN, medians): 55.7 ms (S) vs 69.7 ms (C), P < 0.05. No differences in power spectral components of HRV at rest or during HUT were detected between groups. Based on the analysis of data by the frequency domain method, we conclude that in athletes the resting bradycardia seems to be much more related to changes in intrinsic mechanisms than to modifications
55.7 ms (S) vs 69.7 ms (C), P < 0.05. No differences in power spectral components of HRV at rest or during HUT were detected between groups. Based on the analysis of data by the frequency domain method, we conclude that in athletes the resting bradycardia seems to be much more related to changes in intrinsic mechanisms than to modifications in autonomic control. Also, HUT caused comparable changes in sympathetic and parasym- pathetic modulation of the sinus node in both groups. Correspondence F.S. Martinelli Laboratório de Fisiologia do Exercício Faculdade de Educação Física UNICAMP Av. Érico Veríssimo, 701 13083-970 Campinas, SP Brasil E-mail: chezfab@terra.com.br Research supported by FAPESP (No. 96/09945-0) and CNPq (No. 300528/85-0). Received July 14, 2003 Accepted January 24, 2005 Key words •Endurance •Power spectral analysis •Heart rate variability •Autonomic heart rate modulation •Training bradycardia •Orthostatic stress Introduction Several studies have reported that aero- bic physical training changes the sympatho- vagal balance of the sinus node and may contribute in part to the resting bradycardia observed in endurance athletes. However, it remains controversial whether the alterations in the autonomic function related to resting bradycardia are caused by attenuation of sym- pathetic tone and/or by enhanced vagal ac- tivity (1-6). Previous investigations have also reported that changes in intrinsic mechan- isms of the sinus node rather than autonomic
640 Braz J Med Biol Res 38(4) 2005 F.S. Martinelli et al. alterations are responsible for endurance- trained bradycardia (7-12). In addition to the possible role of vagal and sympathetic adaptations in resting brady- cardia induced by aerobic training, different autonomic cardiovascular responses may be activated during exercise itself and in pos- tural and respiratory tests when trained and untrained subjects are compared (11,13). Some studies have reported that during ortosthatic stress, trained and untrained sub- jects present different cardiovascular re- sponses, which have been associated with alterations in autonomic cardiovascular con- trol (5,14,15). However, other studies failed to show these differences when comparing trained and untrained subjects (16,17). In order to evaluate the cardiac auto- nomic activity non-invasively, many studies have used the measurement of the heart rate variability (HRV) that is obtained by calcu- lating the variations between the RR inter- vals of the ECG signal. These variations are calculated in the time and frequency do- mains; in this last condition they are ex- pressed as sine wave power values (power spectrum). High HRV is usually caused by an increase in vagal tone associated with a decrease in sympathetic tone (18,19). There is substantial controversy concern- ing the absolute and relative changes of power spectral components related to sympathetic and parasympathetic modulation, induced by aerobic training (1,4,6,14,20-23). Thus, the aim of this study was to evalu- ate HRV in the time and frequency domains in athletes and sedentary subjects in the su- pine position and after head-up tilt (HUT), when there is a change in sympathetic and parasympathetic balance in the sinus node. Material and Methods Subjects Two groups of young men, nonsmoking healthy volunteers, were studied: 11 seden- tary controls, who had not engaged in regu- lar physical activity for at least 6 months, and 10 trained cyclists of competitive level, who had been participating in an endurance training program for at least one year. They trained 90 min to 5 h/session, 6 days a week and all participated in official national and/ or international championships in this ath- letic modality. Their anthropometric charac- teristics are presented
physical activity for at least 6 months, and 10 trained cyclists of competitive level, who had been participating in an endurance training program for at least one year. They trained 90 min to 5 h/session, 6 days a week and all participated in official national and/ or international championships in this ath- letic modality. Their anthropometric charac- teristics are presented in Table 1. The study was approved by the Ethics Committee of the State University of Campi- nas, Campinas, SP, Brazil, and each subject was informed about the nature of the experi- ment and gave their informed written con- sent. Protocols The protocols were conducted in two sepa- rate sessions. In the first all subjects performed a dynamic physical exercise test on a cycloer- gometer (Quinton Instrument Company, Se- attle, WA, USA), using a progressive continu- ous protocol (ramp) up to physical exhaustion. The initial workload was 4 watts for 2 min for both groups. The workload increases were 15 W per min for the sedentary group and 30 W per min for the trained group. The subjects pedaled at a frequency of 60 rpm. The ventila- tory and metabolic variables were obtained by direct measurement using a gas analyzer sys- tem (metabolic measuring chart-horizon sys- tems; Sensormedics Corporation, Yorba Linda, CA, USA). The ECG was recorded continu- ously starting one minute before the beginning of the exercise and up to the end of the recov- ery period. The ventilatory anaerobic thresh- old was obtained by analyzing the loss of linearity of CO 2 production and pulmonary ventilation responses at submaximal power values (as averaged at 15-s intervals) (24). In the second session, which was held at least 24 h after the exercise test, all subjects initially rested for 40 min in the supine posi- tion with continuous recording of heart rate obtained by automatic counting of the R wave
641 Braz J Med Biol Res 38(4) 2005 HRV in athletes and nonathletes peaks of an ECG signal. After the supine resting period, a passive 70º HUT was per- formed without foot support using a seat at- tached to the table to support body weight. The subjects were tilted and remained in this posi- tion for up to 60 min. The ECG, the instantaneous heart rate, the pulse pressure, and the respiratory rhythm were recorded continuously with an eight- channel recording system (RS 3800; Gould Instruments Systems, Valley View, OH, USA). Tests were performed in a quiet air-con- ditioned (22-24ºC) room. The subjects were instructed to perform no exercise training 40 h before the day of the experiment and to avoid drugs and caffeine 12 h before the test. Heart rate variability The short-term HRV in the time and fre- quency domains was obtained using 24-h dynamic electrocardiography (Holter for Windows, v. 3.6-F, Rozinn Electronics, Glen- dale, NY, USA). Two or more 5-min win- dows that presented stationarity (visual in- spection) were selected from the RR interval series during rest and HUT. During HUT, HRV was analyzed after the initial transient response of this variable, always during periods of stationarity of the signal. Since autonomic heart rate control cannot be the same at the beginning and at the end of HUT, linear regression analysis was performed to determine if the time of tilting was correlated with the HRV re- sponses. This analysis showed that all 5-min stationary windows during this maneuver could be used. Then, each subject had a different number (average = 3) of 5-min intervals for calculating the HRV, with the first interval being analyzed at least 6 min after tilting. When these intervals were con- secutive for the subject, their arithmetic mean was used. The non-consecutive intervals were considered separately as repetitions. The following HRV variables were ana- lyzed: for the time domain - standard devia- tion of normal RR interval (SDNN, ms); for the frequency domain - low frequency power in normalized units (LFnu, 0.04-0.15 Hz), high frequency power in normalized units (HFnu, 0.15-0.4 Hz), and
the subject, their arithmetic mean was used. The non-consecutive intervals were considered separately as repetitions. The following HRV variables were ana- lyzed: for the time domain - standard devia- tion of normal RR interval (SDNN, ms); for the frequency domain - low frequency power in normalized units (LFnu, 0.04-0.15 Hz), high frequency power in normalized units (HFnu, 0.15-0.4 Hz), and LF/HF - ratio of absolute LF power to HF power. The spectral analysis was calculated by the MK5 Spectral Analysis Package, H4W - Rozinn Electronics, using the Fast Fourier Transform algorithm. The sampling rate was 128 samples per second and all data were analyzed in blocks of 100 s, with a frequency of resolution of 0.01 Hz. The spectra of six successive 100-s blocks were averaged to- gether. Visual inspection of ECG signal record- ings in the computer monitor showed ab- sence of artifacts or cardiac arrhythmias that could interfere with the HRV analysis per- formed by the system analyzer (Holter for Windows, v.3-6-F Rozinn Electronics). Statistical analysis Since most of the data did not present Gaussian distributions, nonparametric tests were used for statistical analysis. The statis- tical procedure used is the one proposed in McGill et al. (25) for exploratory data anal- ysis for various groups. Differences between groups were demonstrated by the confidence interval of the median presented in the boxplots, with the level of significance set at alpha = 0.05. If the intervals in the two boxes do not overlap, this indicates a difference in a location at roughly the 5% level. The same procedure is used for paired sample observa- tions when the differences of the pairs are used as a single sample of observations (26). The exact distribution for the statistical test used to compute the probability value when- ever possible, in case of no tie samples, is the Wilcoxon rank sum test for two sample data (equivalent to the Mann-Whitney test) or the Wilcoxon signed rank test for paired or one sample data.
the probability value when- ever possible, in case of no tie samples, is the Wilcoxon rank sum test for two sample data (equivalent to the Mann-Whitney test) or the Wilcoxon signed rank test for paired or one sample data.
642 Braz J Med Biol Res 38(4) 2005 F.S. Martinelli et al. Results Cardiorespiratory variables at rest and during exercise Table 1 shows the higher aerobic capac- ity of the athletes compared to control group, expressed by higher (P < 0.05) oxygen up- take and power values at anaerobic thresh- old (V . O 2AT) and at peak condition (V . O 2peak). The heart rate in the resting condition was significantly (P < 0.05) lower in athletes than in nonathletes (Table 1). Heart rate variability at rest Table 2 and Figure 1 show that the RR mean (medians) in the supine position was higher (P < 0.05) in athletes than in nonath- letes. Figures 2 to 5 show that, despite the significantly (P < 0.05) lower values of SDNN in the sedentary group, the spectral power components, LFnu, HFnu and LF/HF, were similar for the two groups. Heart rate variability during head-up tilt During HUT (Table 2 and Figures 1 to 5) both groups presented reduction in RR mean values (P < 0.05). Athletes also presented reductions in SDNN (P < 0.05). In both groups, spectral components showed the fol- lowing results: higher LFnu and LF/HF ratio (P < 0.05) and lower HFnu (P < 0.05). Higher values (P < 0.05) of mean RR and SDNN were found in athletes, but no differences in power spectral variables were found between groups. Discussion The higher aerobic capacity of the ath- letes was reflected in greater V . O 2 and power values (cycloergometer) at both anaerobic threshold and peak effort (Table 1). Also, the resting heart rate was significantly lower in cyclists than in nonathletes. Several inves- Table 1. Anthropometric and cardiorespiratory parameters of the sedentary and endur- ance-trained subjects at rest and during exercise. Sedentary group (N = 11) Cyclists (N = 10) Age (years) 21.82 (2.18) 20.80 (3.33) Height (cm) 175.36 (6.70) 175.55 (5.82) Weight (kg) 67.14 (8.56) 70.59 (9.61) Body surface area (m 2 ) 1.80 (0.14) 1.84 (0.15) HR rest 70.9 (10.75) 65.6 (10.49)* HR AT (bpm) 130.45 (15.41) 146.7 (15)* HR peak (bpm) 195.73 (12.04)
at rest and during exercise. Sedentary group (N = 11) Cyclists (N = 10) Age (years) 21.82 (2.18) 20.80 (3.33) Height (cm) 175.36 (6.70) 175.55 (5.82) Weight (kg) 67.14 (8.56) 70.59 (9.61) Body surface area (m 2 ) 1.80 (0.14) 1.84 (0.15) HR rest 70.9 (10.75) 65.6 (10.49)* HR AT (bpm) 130.45 (15.41) 146.7 (15)* HR peak (bpm) 195.73 (12.04) 189.7 (9.12) P AT (W) 108.48 (38.36) 209.25 (30.65)* P peak (W) 212.27 (29.12) 358.8 (19.68)* V . O 2AT (ml kg -1 min -1 ) 19.55 (6.53) 35.16 (6.26)* V . O 2peak (ml kg -1 min -1 ) 38.97 (6.79) 57.01 (4.82)* Data are reported as means (SD). HR rest = heart rate at supine rest; HRAT = heart rate at anaerobic threshold; HR peak = heart rate at peak effort; PAT = power at anaerobic threshold; P peak = power at peak effort; V . O 2AT = oxygen uptake at anaerobic thresh- old; V . O 2peak = oxygen uptake at peak effort. *P < 0.05 compared to sedentary group (Wilcoxon rank sum test). Table 2. Comparison of heart rate variability parameters in time and frequency do- mains in the supine and tilting position. Supine position Tilting position Sedentary group CyclistsSedentary group Cyclists (N = 11) (N = 10) (N = 11) (N = 10) RR mean (ms) 854.3 (105.7) 1076 (115.6)* 625.8 (188.5) + 794.2 (148)* + SDNN (ms) 59.1 (36.5) 89.9 (24.8)* 55.7 (8.8) 69.7 (18.9)* + LF (ms 2 /Hz) 3500 (2900) 8400 (3300) 4700 (1050) 8000 (7400) HF (ms 2 /Hz) 1500 (800) 2400 (900) 770 (550) 880 (975) LFnu (%) 71 (6) 68 (20) 85 (4.5) + 88 (8) + HFnu (%) 29 (6) 33 (18) 15 (4.8) + 12 (8.2) + LF/HF 2.47 (1.01) 2.08 (1.29) 5.56 (2.48) + 7.24 (6.83) + Data are reported as medians and interquartile range. SDNN = standard deviation of normal RR interval; LF, HF = low and high frequency power, respectively; LFnu = LF power in normalized units; HFnu = HF power in normalized units; LF/HF = ratio of absolute LF power to HF
+ LF/HF 2.47 (1.01) 2.08 (1.29) 5.56 (2.48) + 7.24 (6.83) + Data are reported as medians and interquartile range. SDNN = standard deviation of normal RR interval; LF, HF = low and high frequency power, respectively; LFnu = LF power in normalized units; HFnu = HF power in normalized units; LF/HF = ratio of absolute LF power to HF power values. Frequency ranges: LF: 0.04-0.15 Hz and HF: 0.15-0.4 Hz. *P < 0.05 compared to the sedentary group; + P < 0.05 between the supine and tilting position (Wilcoxon rank sum test and Wilcoxon signed rank test, respectively). Figure 1. Mean RR in the supine and head-up tilt (HUT) position for the sedentary group and cyclists. The boxplots show the minimum values, 1st quartile, medians (bold black line), 3rd quartile, maxi- mum, and confidence interval of the medians (hatched area). S = sedentary group (N = 11); C = cyclists (N = 10). *P < 0.05 for comparisons indicated by brack- ets (confidence interval of the median test). RR mean (ms) 1200 1234 1 23 4 1 23 4 1 23 4 1234 12345 1 234 5 1 234 5 1 234 5 1 234 5 1 234 5 1 234 5 12345 12345 1 234 5 1 234 5 1 234 5 1 234 5 1 234 5 12345 12345 1 234 5 1 234 5 1 234 5 1 234 5 1 234 5 12345 12345 1 234 5 1 234 5 1 234 5 1 234 5 1 234 5 1 234 5 1 234 5 1 234 5 12345 * * * * 1100 1000 900 800 700 600 500 S CSC Supine HUT
643 Braz J Med Biol Res 38(4) 2005 HRV in athletes and nonathletes Figure 2. Standard deviation of normal RR interval in the supine and head-up tilt position for the sedentary sub- jects and cyclists. The boxplots show the minimum values, 1st quartile, medians (bold black line), 3rd quartile, maximum, and confidence interval of the medi- ans (hatched area). S = sedentary group (N = 11); C = cyclists (N = 10); HUT = head-up tilt; SDNN = standard deviation of normal RR interval. *P < 0.05 for compari- sons indicated by brackets (confidence interval of the median test and Wilcoxon rank sum test). SDNN (ms) 120 123456 1 2345 6 1 2345 6 1 2345 6 1 2345 6 1 2345 6 1 2345 6 123456 12345 1 234 5 1 234 5 1 234 5 1 234 5 1234512345 1 234 5 12345 12345 1 234 5 1 234 5 1 234 5 1 234 5 1 234 5 1 234 5 1 234 5 1 234 5 1 234 5 1 234 5 1 234 5 1 234 5 1 234 5 12345 100 80 60 40 20 * * * SCSC Supine HUT Figure 5. Low frequency-high frequency ratio (absolute values) in the supine and head-up tilt position for the sedentary subjects and cyclists. The boxplots show the minimum values, 1st quartile, medians (bold black line), 3rd quartile, maximum, and confidence interval of the medians (hatched area). S = sedentary group (N = 11); C = cyclists (N = 10); HUT = head-up tilt; LH/HF = low frequency-high frequency ratio. *P < 0.05 for com- parisons indicated by brackets (confidence interval of the median test). 123456 1 2345 6 1 2345 6 1 2345 6 1 2345 6 1 2345 6 123456 12345 1 234 5 1 234 5 12345 12345 1 234 5 12345 12345 1 234 5 12345 S CSC LF/HF 25 20 15 10 5 0 Supine HUT * * 12345 1 234 5 1 234 5 1 234 5 12345 123456 1 2345 6 123456 12345 1 234 5 1 234 5
2345 6 1 2345 6 123456 12345 1 234 5 1 234 5 12345 12345 1 234 5 12345 12345 1 234 5 12345 S CSC LF/HF 25 20 15 10 5 0 Supine HUT * * 12345 1 234 5 1 234 5 1 234 5 12345 123456 1 2345 6 123456 12345 1 234 5 1 234 5 1 234 5 1 234 5 1 234 5 1 234 5 12345 123456 1 2345 6 1 2345 6 1 2345 6 1 2345 6 1 2345 6 123456 S CSC HFnu (%) 60 50 40 30 20 10 0 Supine HUT Figure 4. High frequency in normalized units at supine and tilting position in the sedentary subjects and cy- clists. The boxplots show the minimum values, 1st quartile, medians (bold black line), 3rd quartile, maxi- mum, and confidence interval of the medians (hatched area). S = sedentary group (N = 11); C = cyclists (N = 10); HUT = head-up tilt; HFnu = high frequency in normalized units. *P < 0.05 for comparions indicated by brackets (confidence interval of the median test). * * Figure 3. Low frequency in normalized units in the su- pine and head-up tilt position for the sedentary subjects and cyclists. The boxplots show the minimum values, 1st quartile, medians (bold black line), 3rd quartile, maxi- mum, and confidence interval of the medians (hatched area). S = sedentary group (N = 11); C = cyclists (N = 10); HUT = head-up tilt; LFnu = low frequency in normal- ized units. *P < 0.05 for comparisons indicated by brack- ets (confidence interval of the median test). 12345 1 234 5 12345 12345 1 234 5 1 234 5 1 234 5 12345 123456 1 2345 6 1 2345 6 1 2345 6 1 2345 6 1 2345 6 1 2345 6 123456 12345 1 234 5 1 234 5 1 234 5 1 234 5 1 234 5 1 234 5 12345 * * SC S C LFnu (%) 100 90 80 70 60 50 40 Supine HUT
Description
This research evaluates autonomic heart rate modulation in athletes and nonathletes.