Standalone sauna vs exercise followed by sauna on cardiovascular function in non‐naïve sauna users: A comparison of acute effects

Abstract Background and aims Sauna bathing and aerobic exercise have each been shown to affect cardiovascular function. However, direct comparisons between standalone sauna bathing and a combination of exercise and sauna on vascular indices remain limited. Therefore, we conducted a cross‐over study using matched durations to explore the hemodynamic changes of sauna exposure when compared to a combination of aerobic exercise and sauna exposure. Methods Participants (N = 72) with at least one cardiovascular risk factor underwent, on two separate occasions: (a) a 30‐minute sauna at 75°C (SAUNA) and (b) the combination of a 15‐minute cycling exercise at 75% maximum heart rate followed by 15‐minute sauna exposure (EX+SAUNA). Relative changes to arterial stiffness (PWV), augmentation index (Alx), brachial systolic and diastolic blood pressure (SBP and DBP), central SBP (cSBP), mean arterial pressure (MAP), and heart rate (HR) were compared PRE‐POST and pre‐ to 30‐minutes post‐intervention (PRE‐POST30). Results Baseline SBP and DBP were 143 (SD 18) mmHg and 86 (SD 10) mmHg, respectively. From PRE‐POST, SAUNA had lower DBP (mean difference [95% CI] 2.5 [1.0, 4.1], P = .002) and MAP (2.5 [0.6, 4.3], P = .01). However, EX+SAUNA had lower SBP (−2.7 [−4.8, −0.5], P = .02), DBP (−1.8 [−3.3, −0.4], P = .01), and MAP (−2.0 [−3.5, −0.5], P = .009) PRE‐POST30. There were no statistically significant differences between SAUNA and EX+SAUNA for other measured parameters. Conclusion This study demonstrated that when matched for duration, EX+SAUNA and SAUNA elicit comparable acute hemodynamic alterations in middle‐aged participants with cardiovascular risk factors. The sauna is a suitable option for acute blood pressure reductions in those who are unable to perform aerobic exercise, and may be a viable lifestyle treatment option to improve blood pressure control.


| INTRODUCTION
Sauna bathing has been associated with a lower risk for cardiovascular disease (CVD) outcomes, 1 improved vascular endothelial and cardiac function, 2 lower blood pressure, 3 and positive alterations in several hemodynamic markers. 4 Sauna bathing exerts a strain on the cardiovascular system, to maintain blood pressure and sufficient blood flow to other organs and muscles. 5 A recent meta-analysis 6 found sauna use to play a positive role in improving cardiovascular function and functional capacity as well.
Likewise, aerobic exercise has been well documented to provide cardiovascular 7 and disease prevention 8 benefits, especially for older adults 9 and populations with adverse levels of risk. 10 Furthermore, aerobic exercise has also been shown to acutely reduce arterial stiffness, 11 although a recent study 12 suggests that improvements in arterial stiffness are only seen when an impairment is present. Arterial stiffness is a useful marker in the prediction of cardiovascular events and has been shown to be capable of predicting cardiovascular mortality independent of other traditional markers. 13 Nevertheless, several studies have noted the comparable physiological effects of sauna bathing and aerobic exercise. 14,15 Previous research investigated sauna bathing as a post-exercise intervention with promising results, 16,17 but the mechanisms of action for both modalities were not directly comparable due to methodological differences. Moreover, although some studies have investigated the conjunctive use of exercise and sauna in athletes, 18,19 this information remains somewhat limited in non-athlete populations. However, Rosenberg and colleagues 20 did speculate about the usefulness of adjunctive exercise and heat therapy, which was partially shown by our group, 21 where a combination of 15 minutes of aerobic exercise followed by 15 minutes of sauna bathing showed several notable improvements to hemodynamic function.
Long-term sauna therapy has been shown to effectively improve exercise tolerance, 22 while regular exercise and sauna exposure were able to improve cardiac function and autonomic nervous system activity, 23 albeit in populations with heart-related issues. Therefore, there is reason to believe that the combination of exercise and heat therapy in the form of sauna bathing may indeed complement each other. Both current and previous literature 24,25 are in support of this postulation, demonstrating heat augmented physiological responses.
As such, a comparison of the effects between these two interventions using an equal exposure time of 30 minutes will help to further elucidate any therapeutic potential that may exist, given the benefits seen from 30 minutes of acute 26 and regular aerobic exercise. 27 To the best of our knowledge, changes to arterial stiffness and hemodynamics between sauna bathing, and a combination of aerobic exercise followed by sauna bathing have yet to be investigated. Our group has previously explored the effects of acute sauna exposure, 4 and thus aim to extend our findings and compare the effects of sauna exposure alone against a shortened duration of exercise plus sauna exposure. This may be of considerable benefit for a broader population, including those who have lower exercise capacities. 28 Therefore, the purpose of the study was to explore the hemodynamic changes of a single session of sauna exposure compared to the combination of aerobic exercise and sauna exposure of matched duration, in a population with at least one cardiovascular risk factor.

| Participants
Participants (N = 72; females = 33, males = 39) were recruited from the city of Jyväskylä, Central Finland, through the local out-of-hospital health care center. To be eligible for inclusion, participants had to be free of a prior diagnosis of CVD and exhibited at least one of the following cardiovascular risk factors: a history of smoking, hyperlipidemia, hypertension, clinically diagnosed diabetes, obesity, or family history of coronary heart disease (CHD). Hypercholesterolemia was defined as the use of cholesterol drugs or serum low-density lipoprotein cholesterol over 3.5 mmol/L. Hypertension was defined as having a systolic reading of greater than 140 mm Hg, and/or a diastolic reading of greater than 80 mm Hg on two or more separate resting measurements. Obesity was defined as body mass index >30 kg/m 2 .
Family history of CHD was considered positive if father (<55 years) or mother (<65 years) had premature CHD. Menstrual status of female participants was not taken into account during recruitment; but none of them were menstruating at the time of the administra-  Participants were instructed to keep the cadence between 65 and 70 rpm for the entire 15 minutes. The cycling load in watts was monitored and adjusted throughout the duration of the exercise to ensure that the heart rate for each participant was kept at 75% of their individual maximum exercise heart rate, pre-calculated using data obtained from the exercise test.

| Experimental design
The temperature and humidity of the exercise room were 21 C and 25%, respectively. Participants were instructed to abstain from eating 2 hours, caffeine 12 hours, alcohol 24 hours, and exercise and/or sauna 48 hours prior to the measurements. Food intake was not standardized. Fluid was consumed ad libitum. A physician was in attendance at all times and participants were allowed to leave the sauna or stop the experiment at any time if they felt uncomfortable, but all participants underwent the two interventions successfully.

| Assessment of outcome measures
The measurements of arterial stiffness during this experiment adhered closely to published guidelines. 29 Supine brachial SBP and DBP were obtained using Microlife BP A200 (Microlife Corp., Taipei, Taiwan).
Two sequential readings were taken and the mean values were used.
All measurements were done on the right side of the body with the participant in the supine position. Transit distances were assessed by body surface measurements using a tape measure. There was a total of three transit distances, carotid artery site to suprasternal notch; carotid artery to femoral artery; and suprasternal notch to femoral artery. Carotid to femoral measurement was adjusted to 80% (common carotid artery À common femoral artery Â 0.8) for the calculation of PWV as recommended. 29 Brachial blood pressures and PWV as a measure of arterial stiffness were taken in their respective order at three different time points; before (PRE), immediately after (POST), and after a 30-minute recovery (POST30). All measurements were taken by a single trained operator of the tonometer to minimize ascertainment biases and to CIs were calculated using two repeated measurements taken 10 days apart (prior to the experiment) from 72 participants via the statistical software R, 30 based on a mean-rating (k = 2), absolute-agreement, two-way mixed-effects model. High-quality recordings (defined as an in-device quality index of more than 90% from an average of at least 10 cardiac cycles) were collected using the PulsePen device (DiaTecne s.r.l., Milan, Italy; www.pulsepen.com) using methods that have been documented in previous studies. 4,21 PWV, augmentation index (AIx), central systolic blood pressure (cSBP), and mean arterial pressure (MAP) were subsequently estimated via the software.
Participants were permitted to take a quick shower (<30 seconds) before POST measurements were taken. Water temperature of the shower was not controlled and participants could freely select their desired temperature. Thereafter, they were instructed to rest in a designated waiting lounge (21 C, humidity 25%) in a seated position for a duration of 30 minutes before the final measurement (POST30) was taken. Participants were kept in a supine position for 7 minutes prior to the measurements.

| Statistical analyses
The individual treatment effects, that is, the differences between preand post-intervention and follow-up measurements (PRE, POST, POST30) do not a priori follow a normal distribution. This was F I G U R E 1 Experimental design flow and order confirmed by 30% of the Shapiro-Wilk normality tests for PRE, POST, and POST30 effects giving P-values below .05. Graphical investigations revealed no remarkable differences in the distribution of responses related to age or gender, so controlling for these variables still leaves us with non-Gaussian distributions. As such, the assumptions of the basic parametric approaches, such as Student's t test and ANOVA were not satisfied. Therefore, a nonparametric approach was selected. Statistical inference in this work is based on the Neyman-Rubin causal model. [31][32][33] For the variable Y, the causal effect of treatment T with respect to control C for the individual u is defined as The effects of the two different interventions (S and ES) for the same individual cannot be simultaneously measured. This is known as the fundamen- The resulting distribution asymptotically approaches the true sampling distribution. As N = 72 the bootstrap samples are drawn from a representative approximation for the population sampling distribution. The analysis consisted of multiple tests for differences between interventions, so we use adjusted P-values for inference, where the adjusting has been done using the Benjamini-Hochberg method. 35 Unadjusted P-values are also reported for the purpose of completeness and meta-analyses.
The reliability of the statistical tests run depends on the validity of the assumption that the participants can be analyzed as a single group. However, it is possible that the responses might be dependent on for example, the age or biological sex of the participant, although this has largely been controlled for with the crossover study design and the inclusion/exclusion criteria. Nonetheless, we ran the analyses in several subgroups to ascertain whether the assumption of homogeneity holds. The analysis was for age, biological sex, initial SBP over 140 or DBP over 80, and menopausal status.
Similar responses were observed in each subgroup. Based on these assessments we concluded that controlling for these covariates would not make any noticeable difference in this data set.
The full-sample comparisons were pre-specified, with a priori determined significance level of P < .05 to be used for the adjusted Pvalues along with a two-tailed alternative hypothesis. The subgroup analyses and normality tests were exploratory in nature, and were aimed at determining whether parametric tests or models with covariates could be adopted in order to increase the statistical power.
Two-tailed alternative hypotheses were assumed here as well. The calculations were implemented with the statistical software R version 3.6.3 30 with the plots done using the ggplot2 package. 36 Continuous data are presented as means ± SD and categorical data as frequencies (percentage of the whole). To ensure that the statistical analyses were nonbiased, variables were coded and analyzed by an independent statistician who was completely blind to the experiment. In addition, the statistician carrying out the analyses was completely uninvolved in the participant recruitment and data collection processes.

| Characteristics of participants
The characteristics of the participants are presented in Table 1

| Changes in outcome measures from PRE to POST30
PRE-POST30 changes in outcome measures and their average causal effect are shown in Table 3. SBP, DBP, and MAP were lower in EX +SAUNA. The distribution of individual responses to the two interventions as measured between PRE and POST30 is displayed in  One of the first studies comparing between EX+SAUNA and SAUNA in a similar population group demonstrated that EX+SAUNA reduced SBP whereas no change was seen in DBP. 16  Passive heat exposure has been suggested to improve endothelial function and nitric oxide bioavailability through enhanced dilation of the arterial tree. 38 This is comparable to how exercise increases dilation of arteries supplying skeletal muscle, which has been documented to lead to reduced wave reflection. 39 The efficacy of heat stress from the sauna was inadvertently shown in a more recent study, 40

| Limitations
In spite of the evidence provided by the current study, several limitations must be noted. From a methodological standpoint, we did not include a standalone exercise intervention. This would have enabled a comparison between EX+SAUNA and an aerobic exercise session of the same duration. However, that was not the purpose of our investigation. The scope of our data interpretation is limited due to the lack of cardiac output measurements and thermal stress (temperature) data, although we did measure heart rate and arterial stiffness.
Menstrual phase was not controlled, and we are unable to determine the degree of influence that the differences in menstrual phases may have had. However, the majority of our female participants were postmenopausal. We did not control for the water temperature of the shower, which may have had an effect on some of the results seen, although we did try to minimize this by limiting the shower time to under 30 seconds. Finally, the majority of our participants were regular sauna users (85%) of at least once a week, which limits the generalizability of our results to a broader population who do not use sauna regularly. It is thus entirely possible that the current results are specific to the population studied, though the sauna-naïve would presumably demonstrate more pronounced responses as they would be less habituated. Nevertheless, it highlights the need for further investigation in this particular area.

| CONCLUSION
The current study shows for the first time the hemodynamic differences between sauna bathing, and a combination of a short bout of aerobic exercise followed by sauna bathing, in a representative F I G U R E 3 The distribution of individual responses to the two interventions as measured between PRE and POST30 population with cardiovascular risk factors. From an acute standpoint, sauna bathing is able to elicit responses that are comparable to a combination of aerobic exercise followed by the sauna, when matched for duration. For populations who are unable to perform aerobic exercise, sauna exposure may provide some similar benefits acutely.
The long-term adaptations of regular exercise in conjunction with passive heat such as the sauna is an area that needs more attention, and experimental trials are needed to better understand the sauna intricately, as it has shown compatibility with aerobic exercise. In the management of hypertension, emphasis is often given to improving ones' diet, performing regular exercise, or weight control. Sauna use might also be a worthwhile lifestyle treatment option to improve BP control.