Physical inactivity is an important risk factor for the development of various cardiometabolic diseases. Importantly, physical activity (PA) has been associated positively with cardiac vagal activity, as measured by heart rate variability (HRV) and BRS (5,15,18,21,24). Cardiovascular autonomic function is known to be related to cardiometabolic risk factors (11,30), and its improvement with increased PA has been postulated as one potential mechanism that leads to reduced cardiovascular risk, independently of traditional risk markers (9). Impaired cardiovascular autonomic function, manifested as depressed vagal and augmented sympathetic activity, has been shown to increase the susceptibility to cardiac events and death in various human populations (13,14).
The effects of lifelong PA (PALife) on HRV in retrospective studies are inconclusive (10,32), and as far as we are aware, there are no studies that have investigated the relationship between PALife and BRS. Notably, some cross-sectional studies have not detected any effect of PA on cardiovascular autonomic function that is independent of demographic, metabolic, and genetic factors uniformly in both sexes (3,6,11,25). The strong age and sex dependency of cardiovascular autonomic function may attenuate the contribution of PA to HRV and BRS (6,11,30).
The purpose of the present study was to determine whether PALife contributes independently to cardiovascular autonomic function in middle-age men and women in large and prospective Northern Finland Birth Cohort 1966 (NFBC1966). We hypothesized that the high level of PALife would be associated with better cardiovascular autonomic function in these now middle-age subjects of NFBC1966.
The subjects whose expected date of birth fell between January 1 and December 31, 1966, in Northern Finland (96.3% of all 1966 births, n = 12,058 live births) were included in the prospective NFBC1966 study. Because of their mothers’ recruitment during their first visit to the maternity health centers, data have been collected on the subjects’ health, lifestyle, and socioeconomic status. The study was conducted according the Declaration of Helsinki and approved by the Ethical Committee of the Northern Ostrobothnia Hospital District in Oulu, Finland. The subjects and their parents provided their written informed consent for the study.
Postal questionnaires enquiring about the participant’s health status and lifestyle was conducted in 1980, 1997–1998, and 2012–2014, i.e., when the participants were age 14, 31, and 46 yr. The response rate was 97% at age 14 yr (n = 11,399), 75% at age 31 yr (n = 8767), and 66% at age 46 yr (n = 6825). Subjects who were living at known addresses in Finland (n = 10,282) were invited to attend the clinical examinations in one of three laboratory units (Oulu, Southern, and Northern Finland). A total of 5861 subjects (57%) participated in clinical examinations between April 2012 and March 2014. The subjects entered the laboratory between 7:00 and 11:00 a.m. after overnight fasting (12 h) and abstained from smoking and drinking coffee during the examination day. Venous blood samples were drawn for the analysis of glycemic and lipid status. Seated systolic (SBP) and diastolic blood pressures (DBP) were measured three times (the two lowest values averaged, Omron M10; Omron Healthcare, Kyoto, Japan) after 15 min of rest. After anthropometric measurements and other examinations, the participants had a light meal 60–90 min before the assessments of cardiovascular autonomic function. On a separate day, an oral glucose tolerance test was conducted according to the recommendations in those participants without medication for diabetes (1).
Measurement of cardiovascular autonomic function
Each participant sat down on a chair to allow instrumentation and a review of the protocol. An HR monitor (RS800CX; Polar Electro Oy, Kempele, Finland) was used to record R-R intervals (RRi). In half of the participants (Oulu laboratory unit), spontaneous BRS was also assessed. Standard lead-II ECG (Cardiolife; Nihon Kohden, Tokyo, Japan), breathing frequency (MLT415/D, Nasal Temperature Probe; ADInstruments, Bella Vista, New South Wales, Australia), and blood pressure (BP) by finger photoplethysmography (Nexfin; BMEYE Medical Systems, Amsterdam, the Netherlands) were recorded with a sampling frequency of 1000 Hz (PowerLab 8/35, ADInstruments). The finger cuff was adjusted so that SBP and DBP assessed by finger photoplethysmography (left arm, supported by an arm sling) did not differ by more than 10 mm Hg from the values measured by automated sphygmomanometer (right arm, Omron M10). The physiological calibration of finger BP was then turned off. After these procedures (5–10 min), there was at least a 1-min period allowing the stabilization of HR before the recording that included 3 min at the seated and 3 min at standing position while breathing spontaneously. Standing position was included because the measures of BRS reflect baroreflex physiology better during upright position and are more reproducible (8,28). The first 150 s of recording in the seated position and the last 150 s in the standing position were analyzed.
Analysis of HRV
Artifacts and ectopic beats were removed and replaced by the local average (Hearts 1.2; University of Oulu, Oulu, Finland). However, sequences with ≥10 consecutive beats of noise or ectopic beats were totally deleted. The RRi series with ≥80% accepted data were included in the analyses. A total of 5679 subjects took part in the RRi recordings, and of these, 5473 (96%) had eligible HRV data for both phases of the protocol (seated and standing). Mean HR and root mean square of successive differences in RRi (rMSSD, ms), a robust measure of cardiac vagal activity, were analyzed along with short-term fractal-like scaling exponent (α1), which was used as an estimate of sympathovagal balance (20,27,31).
Analysis of BRS
Continuous ECG, BP, and respiration signals were imported to a custom-made standalone Matlab-based software (Biosignal Processing Team, University of Oulu) where RRi and SBP values were extracted (12). Artifacts and ectopic beats were replaced using linear interpolation (<5% for accepted recording) and, thereafter, resampled at 2 Hz and detrended (<0.04 Hz, Savitzky–Golay method). A fast Fourier transform (Welch method, segments of 128 samples with 50% overlap) was performed to analyze the low-frequency (LF, 0.04–0.15 Hz) power of RRi and SBP oscillations for subsequent analysis of BRS by the alpha method, if sufficient coherence (≥0.5) between LF oscillations in RRi and SBP was verified. Out of 2726 recordings, BRS was successfully calculated for 2599 subjects (95%) in both the seated and standing positions.
Assessment of PA
PA was self-reported at 14, 31, and 46 yr of age (PA14y, PA31y, and PA46y). At the age of 14 yr, the subjects were asked how often they participated in sports after school hours with the following alternatives: 1) daily, 2) every other day, 3) twice a week, 4) once a week, 5) every other week, 6) once a month, and 7) generally not at all. Options 6 and 7 were combined. At the ages of 31 and 46 yr, the subjects were asked how often they participated in brisk PA/exercise during their leisure time. The term “brisk” was defined as PA causing at least some sweating and getting out of breath, corresponding to moderate- to vigorous-intensity PA. The six response alternatives were 1) daily, 2) four to six times a week, 3) two to three times a week, 4) once a week, 5) two to three times a month, and 6) once a month or less often. For the age-specific PA variables, the two consecutive answers were combined.
Latent class analysis was used to obtain clusters in which the individuals had a similar profile of PA from adolescence to middle age (Auvinen et al. manuscript). Three PALife trajectory groups (active, semiactive, and inactive) were formed according to the self-reported frequency of PA at 14, 31 and 46 yr of age (PALife). In latent class analysis, the number of clusters is increased until the most appropriate model is found (19). We assessed models with a cluster number of one to seven and determined the best-fitting cluster solution of these candidates by calculating the Bayesian Information Criterion. We considered the interpretability of the classification, the conceptual meaningfulness of the models, and the sizes of the subgroups while choosing the best solution (17). The clustering was conducted for all the applicable subjects and both sexes together, not only for those subjects with measured HRV and BRS.
Lifestyle factors at age 46 yr
Based on the questionnaire, subjects were defined as current smokers if they smoked regularly ≥1 cigarette per day on ≥2 d·wk−1. The amount of alcohol consumed per day was estimated from the questions measuring the frequency and the usual amount beverages consumed on one occasion. The subjects were then categorized into two groups based on the highest sex-specific deciles rounded to the closest 10 g·d−1 (40 g·d−1 for men, 20 g·d−1 for women). Total sitting time during waking hours was established by asking the subjects how many hours on average they sat on weekdays (at work, at home, in vehicle and elsewhere) and the sum of sitting times was dichotomized based on previously reported cutoff value of 11 h·d−1 (33). To evaluate the sufficiency of sleep, the subjects were asked how tired they typically felt in the morning during the first half hour after awakening (very/somewhat tired = insufficient, somewhat/well rested = sufficient).
Among subjects with eligible HRV data (n = 5473), PALife could be determined in 4330 subjects. Among them, 255 patients with previous or newly diagnosed diabetes (1,34), cardiac (n = 88) or respiratory disease (n = 246, including asthma), antihypertensive medication (n = 570), or missing data for any covariate (n = 125) were excluded. The final population included 1283 men and 1779 women for HRV and 662 men and 807 women for BRS [Table 1; see Table, Supplemental Digital Content 1, Age-specific frequency of brisk PA and its changes according to the lifelong physical activity (PALife), http://links.lww.com/MSS/A679].
Gaussian distributions of the dependent variables were verified by visual inspection of the histograms. In case of skewed distribution, the variable was transformed into a natural logarithm before statistical analyses (rMSSD and BRS). One-way ANOVA was used to compare the PA groups and sexes, which was followed by Bonferroni’s post hoc test when applicable. Sex differences in categorical variables were assessed by chi-square test. The linearity and collinearity of associations were assessed by the linear and quadratic regression models or contrasts estimated by ANOVA. Thereafter, the multivariate linear regression analysis (enter method) was used to determine the total contribution of the most potential confounders (Block I: smoking, alcohol consumption, sitting time, sufficiency of sleep, and body mass index) and mediators (Block II: waist-to-hip ratio, brachial SBP and DBP, glycated hemoglobin, fasting plasma glucose, serum total and high-density lipoprotein cholesterol, and triglycerides) of the relationship between PA and cardiac autonomic function. Low-density lipoprotein cholesterol was excluded from Block II because of its significant collinearity with total cholesterol. Subsequently, the independent association of PA to autonomic measures was established (Block I + PA and Block I + II + PA). Complementary step- and blockwise multivariate regression (Block I [incl. PALife] + II) was applied to establish the most significant associations of lifestyle and cardiometabolic factors to autonomic measures (data supplements). The data were analyzed using SPSS software (IBM SPSS Statistics 21, IBM Corp., New York). A P-value <0.05 was considered significant.
In men, univariate analysis detected that PALife was associated with all measures of cardiovascular autonomic function, except for α1 in the seated position (Table 2, Fig. 1). A similar pattern was observed with PA46y, whereas the effect of PA31y was only observed in seated HR, seated rMMSD, and standing HR. In the multivariate analysis (Block I + PALife), PALife was significantly associated with HR in the seated and standing position but not with the other measures of autonomic function (Table 2). For example, current smoking was more strongly associated with outcomes in autonomic function than PALife in men (see Table, Supplemental Digital Content 2, Most significant determinants of cardiovascular autonomic function in the multivariate linear regression in men, http://links.lww.com/MSS/A680). However, PA46y was independently related to the measures of cardiovascular autonomic function (Table 2).
In women, PALife was related to all measures of cardiovascular autonomic function in both the seated and standing positions, except α1 (Table 3, Fig. 2). PA46y was associated with seated BRS approaching borderline statistical significance (P = 0.075). In the multivariate analysis (Block I + PALife), all significant univariate associations between PALife and cardiovascular autonomic function remained significant, with exception of the seated BRS (Table 3). Body mass index and BP were also potent factors underlying the variance in autonomic function (see Table, Supplemental Digital Content 3, Most significant determinants of cardiovascular autonomic function in the multivariate linear regression in women, http://links.lww.com/MSS/A681).
This is the first prospective study to reveal that PALife associates with cardiovascular autonomic function, as measured by HRV and BRS, in a middle-age population. In women, PALife was independently associated with vagally mediated HRV and BRS. In men, however, this association was not independent of the other lifestyle and cardiometabolic factors that may mediate the autonomic adaptations to PALife. These findings underscore the importance of long-term PA in prevention of autonomic dysfunction and related cardiovascular risk.
The present study revealed an independent association between current level of PA at middle age and cardiovascular autonomic function in both men and women. In contrast to some previous findings (3,6,11,25), the present study supports the proposal that PA exerts a significant effect of cardiovascular autonomic function that is not completely explained by the other cardiometabolic and lifestyle factors (5,9,15,18,21,24). The intensity of the PA may explain some of the previous conflicting findings. A less significant effect of PA on cardiovascular autonomic function has been observed in those studies that did not estimate the amount of PA at higher intensities (3,11,25), whereas investigations detecting positive associations between cardiovascular autonomic function and PA have found that the relationship exists particularly with moderate- to vigorous-intensity PA (5,21).
The present association between PALife and cardiovascular autonomic function indicated that there are important sex-related factors involved. Uusitalo et al. (32) have reported that PALife was not independently related to cardiovascular autonomic function in middle-age men. The current documentation is in good agreement with the findings by Kaikkonen et al. where PALife was independently associated with HRV in obese, mostly female individuals (10). One possible reason for these sex-specific differences may be that autonomic function exhibits greater adaptability to prevailing level of PA in middle-age men than women (2). This theory is supported by our data showing that the other lifestyle factors at middle age were also more closely related to cardiovascular autonomic function in men than in women, particularly smoking and sufficiency of sleep (see Table, Supplemental Digital Content 2, Most significant determinants of cardiovascular autonomic function in the multivariate linear regression in men, http://links.lww.com/MSS/A680; see Table, Supplemental Digital Content 3, Most significant determinants of cardiovascular autonomic function in the multivariate linear regression in women, http://links.lww.com/MSS/A681). We also observed that PA at age of 31 yr was a more potent determinant of cardiovascular autonomic function in women compared with men, which also supports this hypothesis. Although the earlier PA affects the PA at later life (26), it is important to note that PA14y was not associated with cardiovascular autonomic function at midlife. Therefore, PA during adulthood seems to be the key for the present observations. It is noteworthy that we cannot establish the causality between PA and cardiac autonomic function. Therefore, there may be factors during the growth and aging affecting both long-term PA and the outcomes in autonomic nervous system that remain to be established.
Age and sex are the two most important modifiers of cardiovascular autonomic function (6,11,30). Along with large sample size, the strength of the present study is that age was strictly controlled in the sex-specific analyses. This, however, also explains that in total, independent factors accounted for only ~20% of variance in HRV and BRS in men, and even less in women. However, this is at least as large as the age- and sex-adjusted contribution of the other demographic and lifestyle factors on HRV and BRS in previous studies (6,11,25,32).
The present study provides novel information about the effects of prospectively assessed PALife on BRS, which is known to be a strong determinant of cardiovascular risk (13,14). PALife was associated with BRS in both sexes but associated independently only in women when BRS was assessed in the standing position. Notably, spontaneous BRS method reflects baroreflex physiology with better reproducibility, particularly when evaluated in the standing position (8,28). Like HRV, BRS has been typically measured at supine rest, which is why the results may not be fully comparable with the previous reports. In the present study, seated BRS and rMMSD values were lower than those reported previously from the recordings at supine rest in comparable populations (11,23). Importantly, BRS reflects autonomic outflow to sinoatrial node in relation to the sensation of arterial transmural pressure by baroreceptors that is related to mechanical properties of the vessel (16) that may improve with exercise training (22).
Fractal-like scaling exponent α1, an estimate of sympathovagal balance, was almost totally unrelated to any measured variable. Although Fagard et al. (6) have reported modest effects of PA on the low- to high-frequency ratio, a controversial measure of sympathovagal balance (7), several studies failed to confirm this association (10,25,30,32). However, Kaikkonen et al. (10) observed that physical fitness was related to the low- to high-frequency ratio in the obese individuals but, as with the present fractal α1, the contribution of the other determinants remained low. Although the assessments of sympathetic activity and sympathovagal balance by HRV methods are susceptible (7), the effect of PALife on these features of autonomic regulation remains to be determined.
Body mass index and BP were important factors underlying variance in HRV and BRS in both sexes although the former’s independent value decreased when including cardiometabolic markers in the model (see Table, Supplemental Digital Content 2, Most significant determinants of cardiovascular autonomic function in the multivariate linear regression in men, http://links.lww.com/MSS/A680; see Table, Supplemental Digital Content 3, Most significant determinants of cardiovascular autonomic function in the multivariate linear regression in women, http://links.lww.com/MSS/A681). This suggests that lifestyle changes targeting on these risk factors may also benefit in terms cardiac autonomic function. For example, a unit decrease in rMSSD (ln) has corresponded ~70% increase in adjusted risk for cardiac event during the mean follow-up of 3.5 yr in middle-age general population (29). When assuming causality, the present results suggest that the cardiac risk-related decreased rMSSD might be decreased to one half by improving all unhealthy habits in men, with the potential effects of lifestyle changes being lesser in women. However, it must be acknowledged that the effects of lifestyle factors on cardiac morbidities are not exclusively mediated via autonomic nervous system.
The HRV analyses from the short-term recordings are less reproducible compared with longer-term ambulatory recordings (4). For example, cardiovascular autonomic function is affected by the time from the previous meal, which although controlled was relatively short. This was a tradeoff related to other competing priorities resulting in some compromises in the extensive protocol of NFBC1966 study. Spontaneous breathing may confound the spectral analysis of HRV and BRS (20). In subjects with a breathing signal (n = 1441), ~10% had a breathing frequency at LF, but this was not associated with PA. Moreover, rMSSD is less vulnerable to slow breathing than the spectral indices. Instead of objective measurements, the PALife and all the other lifestyle factors, including sufficiency sleep and sitting time, were based on self-report. Nevertheless, the present PA assessment was simple and based on the one question at each age, although quite infrequently, being easily reproducible. Finally, the population may not fully represent the whole NFBC1966 because of incomplete participation to follow-up measurements and exclusions by cardiorespiratory and metabolic diseases and medications affecting autonomic function.
High level of PALife was independently related to cardiovascular autonomic function in middle-age women. In men, this association was not independent of other lifestyle and health related factors that, in turn, may mediate the effects of PALife on cardiovascular autonomic function. Notably, current PA was independently related to cardiac autonomic function also in men. The present findings highlight the importance of PALife in preventing cardiac events mediated by impaired cardiovascular autonomic function, particularly among women, in whom PA seems to be the most important target for lifestyle intervention in this respect.
This work was supported by the University of Oulu (grant no. 24000692); the Oulu University Hospital (grant no. 24301140); the European Regional Development Fund (grant no. 539/2010 A31592); the Academy of Finland (grant nos. 267435 and 285547); the Ministry of Education and Culture (grant no. 86/686/2014); the Sigrid Juselius Foundation; the Finnish Foundation for Cardiovascular Research, Helsinki, Finland; and the European Union’s Horizon 2020 research and innovation program (grant no. 633595). We thank the late professor Paula Rantakallio (launch of NFBC1966), the participants in the 46-yr study, and the NFBC project center. Meri-Maija Ollila, BM, is gratefully acknowledged for her help with processing of HRV data. The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Alberti KG, Zimmet PZ. Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1: diagnosis and classification of diabetes mellitus provisional report of a WHO consultation. Diabet Med
2. Carter JB, Banister EW, Blaber AP. The effect of age and gender on heart rate variability
after endurance training. Med Sci Sports Exerc
3. De Geus EJ, Boomsma DI, Snieder H. Genetic correlation of exercise
with heart rate and respiratory sinus arrhythmia. Med Sci Sports Exerc
4. Dekker JM, de Vries EL, Lengton RR, Schouten EG, Swenne CA, Maan A. Reproducibility and comparability of short- and long-term heart rate variability
measures in healthy young men. Ann Noninvasive Electrocardiol
5. Dietrich DF, Schindler C, Schwartz J, et al. Heart rate variability
in an ageing population and its association with lifestyle and cardiovascular risk factors: results of the SAPALDIA study. Europace
6. Fagard RH, Pardaens K, Staessen JA. Influence of demographic, anthropometric and lifestyle characteristics on heart rate and its variability in the population. J Hypertens
7. Heathers JA. Sympathovagal balance from heart rate variability
: an obituary. Exp Physiol
8. Herpin D, Ragot S. Mid- and long-term reproducibility of noninvasive measurements of spontaneous arterial baroreflex sensitivity in healthy volunteers. Am J Hypertens
. 1997;10(7 Pt 1):790–7.
9. Joyner MJ, Green DJ. Exercise
protects the cardiovascular system: effects beyond traditional risk factors. J Physiol
. 2009;587(Pt 23):5551–8.
10. Kaikkonen KM, Korpelainen RI, Tulppo MP, et al. Physical activity and aerobic fitness are positively associated with heart rate variability
in obese adults. J Phys Act Health
11. Kardos A, Watterich G, de Menezes R, Csanády M, Casadei B, Rudas L. Determinants of spontaneous baroreflex sensitivity in a healthy working population. Hypertension
12. Kiviniemi AM, Tiinanen S, Hautala AJ, et al. Frequency of slow oscillations in arterial pressure and R-R intervals during muscle metaboreflex activation. Auton Neurosci
13. Kiviniemi AM, Tulppo MP, Hautala AJ, et al. Prognostic significance of impaired baroreflex sensitivity assessed from Phase IV of the Valsalva maneuver in a population-based sample of middle-aged subjects. Am J Cardiol
14. La Rovere MT, Bigger JT Jr, Marcus FI, Mortara A, Schwartz PJ. Baroreflex sensitivity and heart-rate variability in prediction of total cardiac mortality after myocardial infarction. ATRAMI (Autonomic Tone and Reflexes After Myocardial Infarction) Investigators. Lancet
15. La Rovere MT, Pinna GD. Beneficial effects of physical activity on baroreflex control in the elderly. Ann Noninvasive Electrocardiol
16. Lucini D, Palombo C, Malacarne M, Pagani M. Relationship between carotid artery mechanics and the spontaneous baroreflex: a noninvasive investigation in normal humans. J Hypertens
17. Marsh HW, Lüdtke O, Trautwein U, Morin AJ. Classical latent profile analysis of academic self-concept dimensions: synergy of person- and variable-centered approaches to theoretical models of self-concept. Struct Eq Model
18. Molgaard H, Hermansen K, Bjerregaard P. Spectral components of short-term RR interval variability in healthy subjects and effects of risk factors. Eur Heart J
19. Nylund KL, Asparouhov T, Muthen BO. Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Struct Eq Model
20. Penttilä J, Helminen A, Jartti T, et al. Time domain, geometrical and frequency domain analysis of cardiac vagal outflow: effects of various respiratory patterns. Clin Physiol
21. Rennie KL, Hemingway H, Kumari M, Brunner E, Malik M, Marmot M. Effects of moderate and vigorous physical activity on heart rate variability
in a British study of civil servants. Am J Epidemiol
22. Seals DR, Desouza CA, Donato AJ, Tanaka H. Habitual exercise
and arterial aging. J Appl Physiol (1985)
23. Sinnreich R, Friedlander Y, Luria MH, Sapoznikov D, Kark JD. Inheritance of heart rate variability
: the kibbutzim family study. Hum Genet
24. Soares-Miranda L, Sattelmair J, Chaves P, et al. Physical activity and heart rate variability
in older adults: the Cardiovascular Health Study. Circulation
25. Stolarz K, Staessen JA, Kuznetsova T, et al. Host and environmental determinants of heart rate and heart rate variability
in four European populations. J Hypertens
26. Tammelin T, Näyhä S, Hills AP, Järvelin MR. Adolescent participation in sports and adult physical activity. Am J Prev Med
27. 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
28. Taylor JA, Eckberg DL. Fundamental relations between short-term RR interval and arterial pressure oscillations in humans. Circulation
29. Tsuji H, Larson MG, Venditti FJ Jr, et al. Impact of reduced heart rate variability
on risk for cardiac events. The Framingham Heart Study. Circulation
30. Tsuji H, Venditti FJ Jr, Manders ES, et al. Determinants of heart rate variability
. J Am Coll Cardiol
31. Tulppo MP, Kiviniemi AM, Hautala AJ, et al. Physiological background of the loss of fractal heart rate dynamics. Circulation
32. Uusitalo AL, Vanninen E, Levalahti E, Battie MC, Videman T, Kaprio J. Role of genetic and environmental influences on heart rate variability
in middle-aged men. Am J Physiol Heart Circ Physiol
33. van der Ploeg HP, Chey T, Korda RJ, Banks E, Bauman A. Sitting time and all-cause mortality risk in 222 497 Australian adults. Arch Intern Med
EXERCISE; AUTONOMIC NERVOUS SYSTEM; HEART RATE VARIABILITY; BARORECEPTORS
Supplemental Digital Content
© 2016 American College of Sports Medicine