Impact of exercise training on arterial wall thickness in humans

Coronary and cerebrovascular diseases are leading causes of morbidity and mortality [1]. Exercise is associated with decreased cardiovascular risk [2,3] and higher fitness confers cardioprotection [4]. The beneficial effects of exercise on traditional risk factors may explain approximately half of the risk reduction associated with exercise [5,6], and it has been proposed that direct effects of exercise on the vessel wall may account for some of the remaining ‘risk factor gap’ [7,8]. One manifestation of exercise-mediated arterial adaptation is the change in carotid and peripheral wall thickness, which has not previously been reviewed. The purpose of the present review is therefore to describe the effect of exercise training on remodelling of the arterial wall in conduit arteries in healthy asymptomatic subjects and in those with cardiovascular risk factors or disease.

Atherosclerosis can begin in early life [9], with impair-ment of endothelial function a likely precipitating event [10], followed by gradual remodelling of the arterial wall [11]. Although the process of atherothrombosis is dynamic, plaque evolution is prolonged and occult. These considerations emphasize the potential utility of tools which assess pre-clinical atherosclerotic changes in vivo. High-resolution ultrasound is able to detect arterial wall thickness and contemporary edge-detection algorithms can accurately measure IMT (intima-media thickness), which is a validated surrogate marker for atherosclerosis [12].

Large follow-up trials, such as the ARIC (Atherosclerosis Risk in Communities) study [13,14] and the Rotterdam study [15–17], have established that carotid IMT is associated with increased risk for adverse cerebral events (e.g. stroke), which is independent of other risk factors [13,15,17]. Carotid IMT may possess superior predictive capacity for stroke than other measures of atherosclerotic risk, such as the presence of carotid plaques and the ankle–arm index [16].

Increased carotid IMT is also associated with increased risk for cardiac (e.g. angina pectoris and myocardial infarction) [13,15,18–20] and peripheral vascular events (e.g. peripheral artery disease and hypertension) [21,22]. The annual change in carotid artery IMT represents a surrogate marker for systemic atherosclerosis and provides prognostic information [11]. A meta-analysis found that a 0.1 mm increase in carotid artery IMT is associated with an increase in age- and sex-adjusted relative risk of 18% for stroke and 15% for myocardial infarction [23]. A recent study confirmed the independent predictive capacity of carotid IMT in asymptomatic subjects and indicated that carotid IMT provided similar risk stratification as the FRS (Framingham Risk Score) [24].

Arteries of the lower limbs are subject to the development of atherosclerosis, plaque formation and clinical complications (e.g. intermittent claudication). Indeed, thickening of the arterial wall in older subjects is even found in arteries of the upper limbs, such as the brachial artery [25,26], in which plaque formation/rupture is not typically observed. The presence of arterial wall thickening in atherosclerosis-prone and -resistant vessels supports the idea that wall thickening occurs systemically [27]. Whether an age-related increase in wall thickness in these vessels reflects systemic atherosclerotic remodelling or benign age-related thickening of the wall (in some or all vessels) is currently unknown.

Studies have also established the clinical and prognostic value of peripheral artery wall thickness measures. For example, increased femoral artery IMT strongly relates to traditional cardiovascular risk factors, such as blood pressure, waist circumference, cholesterol, insulin and smoking status [28–30]. Thickening of the femoral wall also correlates with measures of peripheral atherosclerotic disease, such as the ankle–brachial index [31], and a significant relationship is present between femoral IMT and the FRS [29,32]. This strong association with cardiovascular risk factors suggests that femoral IMT may possess prognostic relevance. Studies have found associations between femoral IMT and restenosis after percutaneous coronary intervention [33] and with the severity and extent of coronary artery disease [34].

Advanced age is associated with the thickening of atherosclerosis-resistant arteries, such as the brachial artery [25,26]. A correlation exists between brachial and carotid IMT [35], and the extent of atherosclerosis is also correlated between the brachial, carotid and coronary arteries [36]. These studies suggest that upper limb arterial wall thickening may be clinically and prognostically relevant. In a large population-based study, brachial artery IMT related strongly to the FRS in older subjects, with lipid levels [i.e. HDL (high-density lipoprotein); inversely] and oxidative stress [i.e. oxLDL (oxidized low-density lipoprotein); positively] related to brachial IMT [37]. Increased brachial IMT also correlates, independent of other cardiovascular risk factors, with the presence of coronary artery disease [38,39]. Previous studies also found that increased brachial IMT predicts future cardiovascular events in subjects undergoing coronary angiography [40] and that brachial IMT progression is linked with worsening disease severity in patients with heart failure [41]. Collectively, these findings suggest that brachial IMT is related to cardiovascular risk and may have a prognostic role in the prediction of cardiovascular events.

Studies which have examined the relationship between carotid IMT and daily PA levels have typically adopted a cross-sectional design and used questionnaires to assess PA (Table 1). When combined, studies including a total of >28000 subjects found an inverse relationship between self-reported PA and carotid IMT [42–47]. This association is supported by the finding that higher a priori PA levels are related to attenuated 3- or 6-year increases in carotid IMT [42,47]. It must be noted that self-reported PA represents a subjective measure and correlates only modestly with direct measures of physical fitness [48]. Owing to these limitations, questionnaires may only provide meaningful data in large cohorts and with specifically designed instruments to assess PA. Yamada et al. [49], in a small sample (n=149), found no relationship between IMT and PA level assessed using a non-specific questionnaire.

The limitations of PA measurement may explain inconsistent findings regarding the inverse correlation between PA levels and carotid IMT. For example, the association between higher leisure-time PA levels and lower carotid IMT in one analysis was present in never-smokers, but not smokers [45]. Data derived from the Tromso Study, a large population-based trial, identified sex and age as modulators of the inverse relationship between PA and carotid atherosclerosis [43,44]. More specifically, the inverse relationship between PA and carotid artery atherosclerosis was observed in men, but not women [43,44]. These researchers also found that the protective effects of a higher PA were more pronounced in older cohorts (i.e. 60–69 and >70 years) [43].

In summary, although measures of self-reported PA should be interpreted with caution, these studies indicate an inverse relationship between PA levels and carotid IMT. Moreover, limited cross-sectional evidence suggests that the effects of PA on carotid IMT may be more prominent in older subjects and in men.

The ‘gold-standard’ measure of cardiorespiratory fitness in humans involves direct assessment of peak or maximal oxygen consumption in response to an exercise test (Table 1). Cross-sectional studies, predominantly performed in middle-aged and older humans, have reported that low fitness correlates independently with increased carotid IMT [50–52] and the presence of carotid plaque [53]. Studies which have used surrogate measures of fitness in relatively large cohorts have observed that higher walking speeds correlate with a lower carotid IMT [54–56]. Although measurement of fitness and IMT at a single time point does not provide information about progression, Lakka et al. [57] examined the 4-year change in carotid IMT and found that lower fitness in middle-aged men was the strongest independent predictor of a 4-year increase in carotid IMT [57].

One of the first studies to examine the effect of exercise training on atherosclerosis examined carotid artery IMT in 137 endurance-trained and sedentary men and found no significant difference between the groups [58]. Subsequent studies performed in younger [59] or older [60] cohorts also observed no difference in carotid atherosclerosis between endurance-trained subjects and sedentary controls. A recent study [61], however, found a significantly lower carotid artery IMT in elite squash players compared with less active controls. The difference in training intensity and/or load may explain these disparate results, as elite squash players exercised >22 h/week at high intensity [61], whereas others classified endurance training when exercising >3 h/week [59] or >5 days/week [58,60].

Longitudinal studies involving exercise training have directly examined the effect of 8–12 weeks of aerobic exercise in middle-aged [58] and older [62] subjects, but found no evidence that exercise training altered carotid artery IMT. Similarly, 8-week resistance training in young men did not alter carotid artery wall thickness [63]. In another study, Rauramaa et al. [64] examined the 6-year change in carotid IMT in 140 middle-aged men who performed aerobic exercise training or no intervention. They found no effect of exercise training, but a ~40% lower 6-year progression of carotid IMT when participants on statins (n=15) were excluded, suggesting that the anti-atherosclerotic effects of statins may mask the impact of exercise. Another study examined the effect of a 4-year lifestyle intervention (PA+diet) in middle-aged women and found attenuated progression of carotid atherosclerosis in peri- and post-menopausal women, but not pre-menopausal women [65] (Table 2).

In summary, exercise training appears to have a modest effect on carotid artery atherosclerosis in young subjects (Figure 1) and changes may require intense exercise or interventions performed over prolonged time periods.

A limited number of cross-sectional studies have examined the effect of exercise training on the wall thickness of peripheral arteries. In contrast with findings in the carotid artery, lower femoral artery IMT was observed in endurance-trained men and women compared with their sedentary peers [66,67]. Moreau et al. [60] extended this observation by comparing femoral and carotid artery IMT between endurance-trained athletes and sedentary controls. Although they confirmed the lower femoral artery IMT in athletes, no significant differences were observed between groups for carotid IMT. These findings suggest that exercise training may have a larger effect on remodelling of the arterial wall in peripheral arteries (that supply the active regions) than in the carotid arteries. Recently, Rowley et al. [61] provided further support for this proposal (Figure 2) when they reported that elite squash players demonstrate a lower femoral, brachial and carotid artery IMT than sedentary controls. These findings support the presence of a generalized lower arterial wall thickness in highly trained athletes, than in controls.

Longitudinal training studies have also assessed changes in peripheral arterial wall thickness. Dinenno et al. [66] directly examined the impact of a 3-month aerobic exercise training programme on femoral artery IMT in middle-aged men and found a significant reduction in wall thickness. In addition, in older subjects, a decrease in peripheral arterial wall thickness (i.e. popliteal and brachial artery) was found after a 6-month aerobic exercise training [26], whereas no changes were observed after 8 weeks of training [62]. In a recent study, it was demonstrated that the increase in femoral IMT during 8 weeks of bed rest in young men can be (partly) prevented by (resistive) vibration exercise training [68]. Finally, a recent handgrip exercise training study found that localized exercise training can lead to a small, but significant, decreases in brachial artery wall thickness [69]. Taken together, exercise training studies performed in healthy subjects indicate that remodelling occurs in response to prolonged training interventions in peripheral arteries supplying the active skeletal muscle.

Longitudinal exercise training studies of arterial wall thickness have predominantly been performed in obese subjects. Studies have reported beneficial effects of 26 [70,71] and 58 [72] weeks of aerobic exercise training on carotid IMT in obese children. In contrast, no change in carotid IMT was found in adults with obesity after 16–38 weeks of lifestyle modification [73] or 52 weeks of resistance training [74].

Jae et al. [75] examined cardiorespiratory fitness and carotid artery atherosclerosis (defined as a wall thickness >1.2 mm) in 2532 hypertensive men. After adjusting for established risk factors, an inverse relationship was observed between fitness and carotid atherosclerosis. Another study in 87 hypertensive subjects demonstrated that higher self-reported PA was associated with a lower 6.5-year increase in carotid IMT [76]. However, these beneficial effects could not be confirmed by Anderssen et al. [77], who found no effect of lifestyle modification on the 4-year progression in carotid IMT in hypertensive subjects.

An inverse relationship between self-reported PA level and carotid atherosclerosis has been described in patients with T2D (Type 2 diabetes) [78]. Another study found that lifestyle modification prevented the 0.5-year increase in carotid IMT in T2D [79]. In contrast, children with T1D (Type 1 diabetes) demonstrate no relationship between physical fitness and carotid IMT [80] and no change in carotid IMT after exercise training [81] (Table 3).

Finally, a prospective study involving hypercholesterolaemic men found that lifestyle modification (including a recommendation of exercise training) resulted in a regression of the 2-year increase in carotid IMT [82]. Another study examined the effect of aerobic and resistance exercise in heart failure patients on brachial artery IMT. Although aerobic exercise was not associated with a change in brachial artery wall thickness, resistance training significantly reduced brachial artery IMT [83].

Taken together, the studies described above suggest that exercise modifies arterial wall thickness in subjects with existing cardiovascular disease and risk factors. Further research is required to determine the degree of impact that exercise of different forms can have on sub-clinical atherosclerotic progression in humans.

Remodelling of the arterial wall is believed to occur over the time frame of months and years. In keeping with this, longitudinal studies of exercise training typically indicate that changes in arterial wall thickness occur across longer time periods (Tables 2 and 4). These observations are also consistent with the hypothesis that exercise training may initially induce functional arterial adaptation, which is followed by structural adaptations in the arterial wall with continued training [84,85]. Studies which have reported changes in arterial wall thickness as a result of brief interventions (8–12 weeks) [68,69,83] have examined peripheral arteries that supply the exercising region, such as the brachial (i.e. handgrip) and popliteal (i.e. cycling and running) arteries. This suggests that peripheral arteries exposed to a large stimulus for remodelling are capable of adapting more rapidly or that peripheral arteries feeding skeletal muscle possess enhanced intrinsic plasticity than arteries such as the carotids. Another possible explanation relates to changes in resting vascular tone. In a recent study [86], we found that administration of an NO donor resulted in immediate change in carotid and femoral artery IMT, raising the possibility that changes in conduit artery wall thickness in response to exercise training, especially those observed after a relatively short duration training, may relate in part to changes in vascular tone, rather than true structural wall remodelling.

Nordstrom et al. [47] examined the 3-year change in carotid IMT and related this to different levels of leisure-time PA. They found a dose-dependent relationship between the self-reported PA level and the 3-year increase in carotid IMT. The presence of a dose-dependent relationship is supported by others, who have demonstrated with accelerometry that performance of vigorous leisure-time exercise [87], but not average daily PA levels [88], was related to an attenuated 3-year increase in carotid IMT (Table 1). Higher exercise intensity may therefore relate to enhanced remodelling of the arterial wall.

Exercise modality may impact on the change in wall thickness. Most training studies have examined the effect of aerobic exercise on arterial wall thickness (Tables 2 and 4), with some studies also examining alternative training forms, such as resistance training [63,74,83,89]. However, a direct comparison between exercise modality is rare. Maiorana et al. [83] recently compared the effect of aerobic compared with resistance exercise in heart failure patients on the brachial artery IMT. Although aerobic exercise was not associated with a change in brachial artery wall thickness, resistance training significantly reduced IMT. This finding may be specific to patients with congestive heart failure, in whom resistance training may be particularly beneficial [90,91]. Future studies are warranted to identify the optimal exercise training regime to improve arterial wall thickness.

In some studies, the relationship between PA and IMT is independent of cardiovascular risk factors [53,57]. This suggests that the effect of exercise training on arterial wall thickness cannot be entirely explained by exercise-mediated changes in traditional cardiovascular risk factors, such as lipid levels, adiposity and blood pressure. This is supported by studies which have demonstrated the impact of exercise training on IMT in the absence of changes in risk factors [66,69,79,83] and others that reported an inverse correlation between carotid IMT and fitness, independent of risk factors [53,57]. We summarize some alternative pathways to explain the change in IMT after exercise training.

Shear stress plays an important role in the regulation of large artery remodelling [92], and the development of carotid atheromatous plaques has been linked with the presence of low mean shear rate [93]. Development of carotid plaques has also been related to oscillatory shear stress [93,94]. Such oscillatory shear patterns are characterized by shear that is not unidirectional, but rather goes in both directions, i.e. forward and backward. This observation is in agreement with studies performed in vitro [95] and in vivo [96], which reported a pro-atherogenic endothelial cell phenotype or function when these cells were exposed to increased levels of shear into the backward direction (i.e. retrograde shear stress). Specifically, oscillatory shear has been associated with decreased endothelial NOS (nitric oxide synthase) mRNA and increased VCAM-1 (vascular cell-adhesion molecule-1), ICAM-1 (intercellular adhesion molecule-1) and ET-1 (endothelin-1) [97,98]. Although exercise leads to an increase in shear in exercising and non-exercising regions [99,100], few studies have examined whether shear stress contributes to arterial wall remodelling.

Recently, we compared brachial artery wall thickness and diameter between the preferred and non-preferred limbs of elite squash players, with the assumption that the preferred limb of squash players receives a larger shear stress stimulus than the non-preferred limb over a prolonged period [61]. This model allows the isolation of localized effects on arterial function and structure, given that central haemodynamics and sympathetic nervous system tone are controlled for using bilateral limb comparisons. Interestingly, wall thickness did not differ between the preferred and non-preferred limbs, despite a larger arterial diameter on the preferred side (Figure 2). In another study, we directly examined the role of shear stress by performing bilateral handgrip exercise training, while unilaterally manipulating shear stress using cuff inflation during repeated exercise bouts [69]. Although shear was kept near resting levels during handgrip exercise in one arm, exercise training induced a similar decrease in brachial artery wall thickness bilaterally. These studies suggest that systemic, rather than localized, shear stress plays an important role in adaptations of the arterial wall in response to exercise training.

During the cardiac cycle, blood pressure rapidly fluctuates and produces stretch on the arterial wall. Data collected in vitro demonstrate that chronic increases in blood pressure result in pro-atherogenic endothelial cell phenotypes, which are characterized by lower eNOS (endothelial NOS) mRNA expression and higher levels of VCAM-1, ICAM-1, ET-1 and ROS (reactive oxygen species) [98]. These findings support the clinical observation that a chronic increase in blood pressure relates to a thicker carotid arterial wall [101]. Moreover, a previous study in humans found that chronic elevations in local distending pressure in the carotid artery importantly contribute to arterial wall thickening [102]. Interestingly, exercise increases arterial pressure and, consequently, leads to a larger stretch on the arterial wall during exercise. Nonetheless, chronic exercise training is associated with a lower arterial wall thickness (Tables 2 and 4). It can therefore be speculated that up-regulation of pro- (e.g. VCAM-1, ICAM-1, ET-1 and ROS) and anti-atherogenic (e.g. eNOS) genes differ when the pressure stimulus involves chronic elevation compared with the transient, episodic and cyclical increases in blood pressure which occur in response to exercise [97,98]. Although findings are conflicting and primarily relate to responses of individual cells (excluding the cross-talk between cells), some evidence indicates that short-term cyclic elevation in pressure induces anti-atherogenic changes in the arterial wall (for a review, see [97]). However, this is speculative at present and research in this area is in its infancy. Future studies are needed to clarify the role of specific patterns of blood pressure as stimuli for arterial wall remodelling.

The carotid and femoral artery walls are directly influenced by changes in vascular tone, since any decreases in wall thickness are observed when vascular tone is acutely modified [86]. Previous studies have found that exercise training can alter peripheral vascular tone, possibly contributing to a decrease in arterial wall thickness.

Results from animal studies indicate that a sustained elevation of SNS tone stimulates smooth muscle cell hypertrophy [103]. Interestingly, SNS activity and femoral artery IMT in healthy young and older men are strongly correlated [101]. This suggests that SNS activity levels may contribute to changes in arterial wall thickness. Although exercise training is associated with decreased levels of resting SNS activity in clinical groups [104], conflicting results are apparent in healthy individuals [105], with some evidence supporting higher resting SNS activity in endurance-trained middle-aged and older subjects [106,107]. It is currently unknown whether changes in SNS activity and IMT as a result of exercise training are correlated, and whether this potential relation differs between groups.

Elevated levels of oxidative stress, which represent a balance between the production of ROS and the efficiency of antioxidant defences, induce endothelial dysfunction and are believed to contribute to the deve-lopment of atherosclerosis [108,109]. Exercise training, especially when performed at moderate-intensity levels, is associated with antioxidant effects [110]. Therefore exercise-training-associated improvement of antioxidant defences (e.g. superoxide dismutase) may lower oxidative stress in the arterial wall and consequently contribute to decreased arterial wall thickness. However, to date, no study has directly examined the potential link between oxidative stress and arterial remodelling as a result of exercise training and should be subject for future research.

The importance of inflammatory processes during the development of carotid atherosclerosis is commonly accepted [111]. Exercise training is associated with anti-inflammatory effects, which are believed to contribute to the cardioprotective effects of an active lifestyle [112]. On the basis of these observational findings, it is possible to speculate that exercise training may alter arterial wall thickness or plaque stability through an impact on inflammation. However, no direct evidence is currently available regarding this hypothesis. Future studies should examine the potential role of (anti-)inflammatory responses to contribute to the effect of exercise training on the arterial wall.

Exercise training can decrease arterial wall thickness in healthy asymptomatic subjects as well as in subjects with cardiovascular risk factors and/or disease that demonstrate increased arterial wall thickness a priori. However, differences exist between arteries, since exercise training has more pronounced effects on large peripheral arteries than the carotids. Modification of carotid wall thickness may require more intense or prolonged exercise exposure. Changes in arterial wall thickness may contribute to the cardioprotective effects of exercise, which are not fully explained by effects of traditional risk factors. However, many questions remain unanswered. Little is known regarding the duration, type and intensity of exercise training necessary to induce optimal benefits on the arterial wall and it is unclear whether these effects differ between healthy subjects and clinical groups. Moreover, the mechanisms that explain changes in the arterial wall as a result of exercise training are not fully understood. Identifying these stimuli will help in the design and recommendation of optimal exercise training protocols to attenuate atherosclerosis burden and risk. Technical advances and newer imaging technologies [e.g. MRI (magnetic resonance imaging) and strain echography] will assist in answering important questions regarding the impact of exercise training on arterial structure and remodelling in future.

D.J.G received research funding support from the National Heart Foundation of Australia and the Australian Research Council. D.H.J.T. is a recipient of an E. Dekker post-doctoral stipend from The Netherlands Heart Foundation [grant number 2009T064].

ET-1

endothelin-1

FRS

Framingham Risk Score

ICAM-1

intercellular adhesion molecule-1

IMT

intima-media thickness

NOS

nitric oxide synthase

eNOS

endothelial NOS

PA

physical activity

ROS

reactive oxygen species

SNS

sympathetic nervous system

T1D

Type 1 diabetes

T2D

Type 2 diabetes

VCAM-1

vascular cell-adhesion molecule-1