Aortic haemodynamics: the effects of habitual endurance exercise, age and muscle sympathetic vasomotor outflow in healthy men

The individuals studied here have been reported on previously to address separate a priori study aims (Wakeham et al. 2019; Lord et al. 2020; Talbot et al. 2020). Forty-six healthy (free of chronic disease), non-smoking, normotensive and non-obese males were recruited into groups of endurance-trained runners and recreationally active nonrunners. Young runners had been training for 8 ± 5 years and were currently completing 65 ± 14 miles per week. The middle-aged runners had been training for 29 ± 15 years and were currently completing 35 ± 10 miles of per week. All endurance-trained runners were ranked within the top 30% of their age-category in the UK for 5 km in the year in which they were studied, with season best times of 15:22 ± 00:37 min:s and 19:31 ± 00:57 min:s for young and middle-aged runners, respectively. Nonrunners were recruited based upon self-report of completing at least 150 min of moderate to vigorous physical activity per week in accordance with the UK Physical Activity Guidelines for adults aged 19–64 years of age from the United Kingdom Chief Medical Officers (2019), but were not undertaking structured exercise training. All middle-aged men presented with a normal electrocardiogram at rest and during a maximal exercise test (via an incremental ramp protocol) on a cycle ergometer (Lode Corival, Groningen, Netherlands). Maximal exercise tests were used to determine cardiorespiratory fitness (\(\dot{V}{\text{O}}_{2}\) Peak), via automated indirect calorimetry (Oxycon Pro, Jaeger, Hoechberg, Germany); the Oxycon Pro has been validated across a range of exercise intensities (Rietjens et al. 2001). This study conformed to the Declaration of Helsinki, other than registration in a database, and all procedures were approved by the Cardiff School of Sport and Health Sciences Research Ethics Committee (16/7/02R). Prior to testing, all participants provided both written and verbal informed consent.

All visits were completed in a temperature-controlled laboratory (~ 22 °C). Participants attended the laboratory on two occasions having abstained from caffeine, alcohol, and strenuous exercise for twenty-four hours and fasted for six hours. The first visit involved the assessment of aortic haemodynamics, followed by the measurement of \(\dot{V}{\text{O}}_{2}\) Peak. On a separate day, with a minimum of 1 week between testing days, participants underwent recordings of MSNA at rest, as outlined previously (Wakeham et al. 2019).

To address the first aim of this study, after confirming compliance with basic parametric assumptions, analysis of variance (ANOVA) was used to compare participant characteristics. Subsequently, a general linear model was used to compare aortic haemodynamics between groups; that is to determine the main effects of runner and age, and the runner*age interaction. In the presence of a significant runner*age interaction, SIDAK post-hoc multiple comparisons were assessed. In the presence of significant main effects of runner or age for Converted aPWV or AP, as well as for the potential covariates (heart rate and aMAP [or height, for AP only]), then ANCOVA analyses were performed. The statistical model used for the ANCOVA was dependent on the presence of a significant runner*age interaction. If there was no significant runner*age interaction, then a general linear model was used with the inclusion of the significant covariates; however, if there was a significant interaction, separate one-way “posthoc” ANCOVA analyses were conducted for each of the four primary between-group comparisons (i.e. effects of runner in young [1] and middle-aged men [2], as well as the effects of age in nonrunners [3] and runners [4]). Adjusted group standard deviations (SD) were calculated from the sum of the square root of the sample size and the standard error of the estimated marginal means from the adjusted statistical models (Atkinson and Batterham 2013). Mean difference [95% confidence intervals] and Hedges’ g, an estimate of effect size ([−]0.2 small effect, [−]0.5 medium effect, [−]0.8 large effect) (Cohen 1992), are also reported for primary outcomes, namely adjusted aPWV, adjusted AP, and aSBP. To address the second aim of this study, two-tailed Pearson product-moment correlation coefficients were generated to assess the relationship between MSNA and AP for the merged samples of young (n = 23) and middle-aged (n = 23) men. Data are presented as mean ± standard deviation or as mean difference [95% confidence intervals] and α was set a priori at < 0.05. All statistical analyses were conducted using Statistics Package for Social Sciences (SPSS) for Windows (version 26, Chicago, IL).

In both young and middle-aged men, body mass and BMI were significantly lower in runners compared to nonrunners, whereas \(\dot{V}{\text{O}}_{2}\) Peak was higher in runners (Table 1). \(\dot{V}{\text{O}}_{2}\) Peak was lower in middle-aged compared to young runners, with no difference between middle-aged and young nonrunners.

When the runners and nonrunners were merged within each age group to address the second study aim, middle-aged men were older (55 ± 4 vs 23 ± 3 years, P < 0.001) and shorter (175.1 ± 6.5 vs 179.1 ± 5.4 cm, P = 0.025) than young men. Body mass (72.6 ± 11.4 vs 72.8 ± 12.9 kg, P = 0.946), BMI (24 ± 3 vs 23 ± 4 kg m², P = 0.433) and \(\dot{V}{\text{O}}_{2}\) Peak (43.8 ± 10.9 vs 50.1 ± 14.6 mL kg⁻¹ min⁻¹, P = 0.067) were not different between middle-aged and young men. MSNA burst frequency (30 ± 10 vs 17 ± 8 bursts min⁻¹, P < 0.001) and incidence (62 ± 21 vs 32 ± 18 bursts 100hb⁻¹, P < 0.001) were higher in middle-aged men.

As 4 young men had AP ≥ 0 and 1 middle-aged man had an AP of 0, separate analyses for positive and negative AP values was not possible in each age group. Correlation analyses were, therefore, conducted in 19 young men and 22 middle-aged men, with only negative and positive AP values, respectively. We observed no significant relationship between MSNA burst frequency and AP in either young or middle-aged men (Fig. 3). In addition, there was no significant correlation between MSNA and AIx in either young (r = − 0.03, P = 0.893) or middle-aged (r = 0.12, P = 0.574) men.

In the present study, HR and aMAP adjusted aPWV was not influenced by habitual exercise in either young or middle-aged men. In the studies conducted to date aPWV has not been adjusted for HR and aMAP (Vaitkevicius et al. 1993; Tanaka et al. 1998; McDonnell et al. 2013; Pierce et al. 2013, 2016; Tarumi et al. 2013, 2021). Notably, the adjustment of aPWV markedly changed our study findings. Converted aPWV (i.e. unadjusted) was lower with habitual exercise in both young and middle-aged men, as reported in most (Vaitkevicius et al. 1993; Tanaka et al. 1998; McDonnell et al. 2013; Pierce et al. 2013, 2016; Tarumi et al. 2013, 2021; Otsuki et al. 2007a, b), but not all (Bjarnegard et al. 2018; McDonnell et al. 2013; Shibata et al. 2018) previous studies. Our findings suggest that habitual exercise may not directly influence intrinsic aortic stiffness (i.e. aPWV independent of confounders) per se, rather that habitual exercise influences operating aortic stiffness (i.e. measured aPWV) due to training-related effects on heart rate and arterial pressure. Thus, future studies should also determine adjusted aPWV when appropriate, dependent on differences in heart rate and/or arterial pressure, to avoid reporting erroneous conclusions (Van Bortel et al. 2020) regarding the effects of habitual exercise on intrinsic aortic stiffness. These data suggest that habitual exercise does not change intrinsic aortic stiffness in healthy normotensive men.

Alongside aortic stiffness, aortic systolic pressure augmentation is an important determinant of aortic blood pressure, which independently increases cardiovascular risk (Wang et al. 2010). When adjusted for heart rate and aMAP, there were no significant effects of habitual exercise on AP (or AIx) in young or middle-aged men. Again, these findings differ when these variables are unadjusted, as AP was not different in young men, but AP was higher in middle-aged runners compared to nonrunners. AP@75 was, however, lower in young runners compared to nonrunners, whereas the normalisation to a heart rate of 75 bpm removed the difference between middle-aged groups. Thus, the adjustment of AP for heart rate and aMAP highlights the importance of including aMAP as a covariate when comparing AP between the trained and untrained men here, as adjusted AP was not effected by habitual exercise in either age group. The use of ANCOVA for the adjustment of AP or AIx for heart rate, as opposed to the automated adjusted to 75 bpm within the SphygmoCor, is beneficial as this statistical method adjusts the dependent variable according to the group means and not based upon an arbitrary calculation from previous data, as highlighted previously (Stoner et al. 2014). Furthermore, the adjustment to 75 bpm is limited to heart rates between 40 and 110; thus, it is not possible to generate these data in endurance athletes who present with heart rates below 40 bpm, as was the case for four runners included here. Previous studies have reported that systolic pressure augmentation (AP, AIx or AIx@75) is either lower (Edwards and Lang 2005; Denham et al. 2016; McDonnell et al. 2013; Bjarnegard et al. 2018) or not different (Knez et al. 2008; McDonnell et al. 2013; Shibata et al. 2018) with habitual exercise. The mechanism(s) mediating the higher raw AP in middle-aged runners compared to nonrunners are unclear; future studies are required to determine the influence of habitual exercise across the lifespan on forward, backward and reservoir pressures. The disparity in findings is likely due to the differences in the measure (AP vs AIx), or equipment used, whether, or not, heart rate and aMAP were adjusted for, as well as the magnitude of differences in cardiorespiratory fitness and/or participant age. Together, the data presented here suggest that the habitual endurance exercise related difference in aortic systolic pressure augmentation is mediated by heart rate and aMAP.

The effects of habitual exercise on aSBP were age-dependent; whereby, aSBP was only lower in young runners compared to nonrunners, despite no significant differences in adjusted aPWV or AP in either age group. This age-dependent effect is likely due to the greater magnitude of difference in non-augmented aSBP (i.e. P1) in young runners compared to nonrunners (-6 mmHg); the mean difference was smaller (− 2 mmHg) between middle-aged groups. As P1 is determined by the interaction between the left ventricular ejection wave and ascending aortic stiffness, the greater ascending aortic compliance in young endurance-trained men (Tarumi et al. 2021) likely contributes to this difference in non-augmented aSBP. The effects of habitual exercise on ascending aortic stiffness in middle/older age are currently unclear. It is important to note that our data do contradict a previous study which reported the inverse, with effects of habitual exercise on aSBP in middle-aged but not young individuals (McDonnell et al. 2013). However, this discrepancy is likely related to the study of only normotensive men in both age groups here, unlike the previous study. Therefore, in terms of cardiovascular risk as determined by aSBP (Stamatelopoulos et al. 2006; Vlachopoulos et al. 2010; Lamarche et al. 2020), the benefit of habitual endurance exercise is apparent in young but not middle-aged normotensive men.

With advancing age aortic stiffness, systolic pressure augmentation, and aortic blood pressure increase (McEniery et al. 2005). Many studies have suggested that lifelong habitual exercise can slow age-related aortic stiffening (Vaitkevicius et al. 1993; Tanaka et al. 1998; Gates et al. 2003; Pierce et al. 2013, 2016; McDonnell et al. 2013), which could also influence AP. However, some of these previous studies compared middle/older aged endurance-trained individuals with young sedentary, but not young endurance-trained, individuals (Vaitkevicius et al. 1993; Pierce et al. 2013, 2016), which could underestimate the age-related difference with the middle/older aged endurance-trained groups. Indeed, we report that the age-related differences in aPWV are similar for chronically endurance-trained runners and recreationally-active nonrunners. Importantly, the middle-aged runners here had been training for ~ 30 years, which is similar to the age difference between our young and middle-aged groups (~ 32 years). This finding opposes the view that habitual endurance exercise has “anti-ageing” effects on aortic stiffness, as reviewed recently (Nowak et al. 2018; Rossman et al. 2018; Seals et al. 2018; Tanaka 2019; Seals et al. 2019; Parry-Williams and Sharma 2020). The age-related difference in AP was greater (larger effect size) in runners compared to nonrunners. This also contradicts a previous report of no effect of habitual exercise on the age-related difference in AP has been reported previously (McDonnell et al. 2013); however, as mentioned previously, the older groups studied by McDonnell and colleagues were prehypertensive and hypertensive, unlike the present study, and the active group were likely not as homogenous as the sample studied here. Thus, more studies are warranted to provide further insight into whether habitual endurance exercise influences “aortic ageing”, utilising a similar four-group study design as utilised here, with appropriate adjustments of aPWV and AP.

In the present study, the age-related difference in aSBP was greater in runners compared to nonrunners. Despite this, the middle-aged runners did not present with higher aSBP or adjusted AP when compared to their recreationally active age-matched counterparts (Table 2). This is of importance for their cardiovascular risk profile, as absolute aSBP is predictive of cardiovascular morbidity and mortality (Stamatelopoulos et al. 2006; Vlachopoulos et al. 2010; Lamarche et al. 2020).

In the current study, we found no relationship between MSNA burst frequency and AP, in young or middle-aged normotensive men. Specifically, we observed this in a sample of young men where AP did not contribute to aSBP in 91% of individuals, but did contribute to aSBP in 96% of middle-aged men, due to the return of the backward pressure wave occurring in diastole and systole in young and middle-aged men, respectively. MSNA and AP (or AIx) have been shown to positively correlate in young men (Casey et al. 2011; Smith et al. 2015); however, 44% (Casey et al. 2011) and 53% (Smith et al. 2015) of the samples studied previously had positive AP values. Whether these previously reported correlations between MSNA and AP in young men would remain when assessed for positive and negative AP values separately, is unclear. The previous studies in middle/older age report disparate findings. Millar et al. (2019) reported no correlation between MSNA and AP in healthy normotensive individuals (75% male), where AP contributed to aSBP in 88% of participants studied. The discordance in findings between this study and those conducted previously likely reflects differences in the distribution of AP values and/or levels of aortic pressure. Indeed, higher arterial pressure is associated with higher AP (Mitchell 2014). Positive relationships between MSNA and AP may only exist in individuals with elevated arterial pressure and AP, as shown previously in prehypertensive and hypertensive postmenopausal women (Hart et al. 2013) and heart failure patients with reduced ejection fraction (Millar et al. 2019). Together, in normotensive young or middle-aged men, without and with pressure raising AP values respectively, resting peripheral sympathetic vasomotor outflow does not correlate with AP (or AIx).

Casey et al. (2011) suggested that MSNA likely contributes to AP in young men via alterations in peripheral vascular tone, and subsequent increases in backward travelling pressure waves, as well as increases in aortic stiffness (Casey et al. 2011). However, the authors made this conclusion despite reporting no significant correlation between total peripheral resistance and AP in young men (Casey et al. 2011). Regarding the influence of MSNA on aortic stiffness, and subsequently AP, it has been shown that MSNA significantly correlates with aPWV at rest (Swierblewska et al. 2010; Holwerda et al. 2019) and there are pressure-independent increases in aPWV during acute sympathoexcitation (Nardone et al. 2018; Holwerda et al. 2019). Notably, we observe no significant correlation between MSNA and aPWV in either age group studied here (young: r = 0.162, P = 0.461; middle-aged: r = − 0.367, P = 0.085), which may contribute to the lack of relationships between MSNA and AP in this study. Thus, a positive relationship between MSNA and aortic stiffness may underlie the previously reported positive correlations between MSNA and AP (Casey et al. 2011; Smith et al. 2015; Hart et al. 2013; Millar et al. 2019); an effect likely due to the pressure-raising effect higher aortic stiffness has upon aortic reservoir pressure, and subsequently AP (Davies et al. 2010). This may especially be the case in the presence of high blood pressure and/or cardiovascular disease where AP raises aSBP (Mitchell 2014).

There are some limitations which warrant discussion. First, as we only studied Caucasian men here it is not possible to determine if there are similar or divergent effects of habitual exercise and/or ageing on aortic haemodynamics between the sexes or different racial groups. The effects of habitual exercise on central haemodynamics have been studied in women previous, with reports of no differences between trained and untrained women in either young (Bjarnegard et al. 2018) or middle/older age (Tanaka et al. 1998). Further, there are no effects of sex on aortic pressure in either African American or Caucasian individuals, despite lower aPWV in women (Yan et al. 2014). Future studies are required to understand the interaction between the effects of habitual exercise, age, sex, and other socio-cultural factors on aortic haemodynamics. Second, as a cross-sectional study design was utilised, it is not possible to truly determine the effects of habitual exercise on the age-related changes in aortic haemodynamics. Future studies should complete longitudinal follow-up studies in endurance-trained and untrained men and women to discern the effects of age, sex and habitual exercise on aortic haemodynamics. Furthermore, as we did not record objective measures of physical activity here, we cannot determine the effects of different physical activity and/or sedentary time on aortic haemodynamics. This was by design, as we studied healthy but untrained men who were meeting the current physical activity guidelines. Third, we assessed the relationships between resting MSNA and aortic AP which were recorded on separate days. However, the measurement of MSNA is repeatable within individuals over the short, medium and long-term (Fagius and Wallin 1993; Kimmerly et al. 2004; Notay et al. 2016; Grassi et al. 1997; Fonkoue and Carter 2015). Finally, it is not possible to directly record sympathetic outflow to the aorta in humans. Nevertheless, the main aim of the correlational analyses performed here was to determine whether peripheral sympathetic vasomotor outflow (i.e. MSNA) is correlated to AP with consideration of the effect of negative and positive AP values.

We report an age-dependent effect of habitual endurance exercise on aortic haemodynamics in normotensive men; whereby, aSBP is lower in young, but not middle-aged, runners compared to nonrunners, with no difference in adjusted aPWV or AP in either age-group. The difference in findings between the raw and adjusted aPWV and AP data highlight that the appropriate adjustments of aortic stiffness and systolic pressure augmentation are important when studying the effects of habitual exercise between groups of young and middle-aged men to avoid reporting erroneous conclusions (Van Bortel et al. 2020). However, the differences in unadjusted aPWV and AP with habitual exercise reported here are the functionally relevant operating aortic stiffness and augmentation pressure values, which are of clear physiological and potential clinical relevance. Nevertheless, our study highlights that intrinsic aortic stiffness and adjusted systolic augmentation pressure are not influenced by habitual exercise. Thus, the habitual exercise related differences in operating aortic stiffness and systolic augmentation pressure appear to be mediated by differences in other hemodynamic indices, namely heart rate and blood pressure. In addition, we report no significant correlation between peripheral sympathetic vasomotor outflow and systolic pressure augmentation in either young or middle-aged normotensive men. Future studies should consider performing separate correlational analyses with positive and negative AP values, as only positive AP values represent clinically relevant physiological effects on aSBP.