Arterial adaptations to training among first time marathoners

For the ultrasound assessment, participants entered a quiet room with limited lighting and reclined in a supine position for 10 min. Following the rest period, right side carotid (1–3 cm inferior to the carotid bifurcation), brachial (1–3 cm superior to the anticubital fossa), and popliteal (1–3 cm superior to the popliteal fossa) images were obtained with high resolution, B–mode ultrasonography (Mindray, DC-6, 21 Mahwah, NJ) using a probe that was set at 10 MHz.

Analyses of the arterial images were performed with methods described by Veller et al. [34] and others [12, 34]. The only exception to the methodology described by Veller et al. [34] was that ImageJ software system (National Institute of Health; http://​rsb.​info.​nih.​gov/​ij/​) was used in place of the CalComp digitizing system. All ultrasound measurements preceded exercise testing and were recorded in mm. Arterial IMT was defined as the distance from the edge of the lumen-intima interface to the edge of the media-adventitia interface; and, TWT was defined as the distance between the deep lining of the intima (endothelium) and the superficial wall of the adventitia [34]. Vessel LD was defined as the distance from one edge of the lumen-intima interface to the edge of the opposing lumen-intima interface; and, W: L was calculated by dividing TWT by LD [34].

A graded exercise test was conducted to determine maximal oxygen uptake (VO_2max), submaximal blood lactate concentrations, and lactate threshold in the experimental group. The treadmill protocol consisted of two discontinuous phases. Submaximal blood lactate concentrations and lactate threshold were assessed in the first phase. Once blood lactate levels exceed lactate threshold the treadmill was stopped and participants rested for a period of 15 min. The speed of running that immediately preceded lactate threshold was compared before and after the marathon training. Following 15 min of passive recovery, VO_2max was measured in the second phase of the exercise test.

The IPAQ was used to measure time (min per day) spent performing moderate activity and vigorous activity. The IPAQ was used to determine if physical activity was different between the groups and if within group activity was different between the beginning and the end of the study.

Prior to hypothesis testing, all dependent variables were tested for normality using the Shapiro-Wilkes method. Variables that deviated significantly (p < 0.05) from the normal distribution (e.g., experimental pre-training carotid IMT) and the corresponding variables (e.g., control pre-training carotid IMT as well as control and experimental post-training carotid IMT) were normalized with log transformation (Log ^10). Two by two repeated measures ANOVA with one within subject factor (pre and post measurements) and one between subject factor (experimental and control groups) were used to identify significant interactions between the groups. When appropriate (time x group interaction; p < .05), a paired T-test was used to determine whether or not the dependent arterial dimensions changed over time within either group and an independent T-test was used to assess between group differences. T-test results were adjusted with the Bonferroni correction for multiple comparisons. A single researcher collected the ultrasound data. To assess researcher reliability, test-retest analysis (paired T-test; Pearson product moment correlation and coefficient of variation) was carried out on samples collected from an independent set of participants on two different occasions. Two-tailed alpha levels were used for all analyses and all data are presented as mean ± standard deviation.

Initial testing combined moderate and vigorous activity values for the experimental and control groups were 361 ± 210 min/week and 31 ± 56 min/week (between group p = .000), respectively, and final testing values were 393 ± 136 min/week and 35 ± 69 min/week (between group p = .000). When the moderate and vigorous activity components were evaluated separately, the experimental group transitioned to spending significantly more time in vigorous exercise than moderate activity by the end of the study (pre moderate: 114 ± 123 vs. post moderate: 70 ± 83; p < 05; pre vigorous: 247 ± 102 vs. post vigorous: 323 ± 113; p = .038).

In the experimental group, VO_2max (ml/kg/min; L/min) did not differ between initial and final measurements. Running speed at lactate threshold increased significantly (p = .008) and lactate concentrations at a submaximal running speed decreased significantly (p = .001) (Table 2).

Initial testing carotid arterial measurements were not different between the groups. Likewise, initial- and final-carotid artery TWT, IMT, LD, and W:L were not different within either group (Table 3).

Initial testing data shows that the experimental group had significantly larger brachial LD (p = .002) than the control group: no other initial-testing between group differences were found in terms of brachial TWT, IMT or W:L. There were no significant changes in the brachial artery measurements within the control group. Marathon training decreased brachial artery TWT (p = .007) and increased LD (p = .015). These combined changes resulted in a significant reduction in W:L (p = .000). Allowing for the Bonferroni statistical correction, IMT of the brachial artery tended to decrease in the experimental group (p = 0.032) (Table 3).

Initial testing data shows that the experimental group had significantly larger popliteal LD (p = .007) than the control group; no other initial-testing between group differences were found in terms of popliteal TWT, IMT or W:L. There were no significant changes in the popliteal artery measurements within the control group. Marathon training decreased popliteal TWT (p = .005) and increased LD (p = .007). There was also a significant reduction in W:L (p = .000). IMT of the popliteal artery did not change in the experimental group (p = .260) (Table 3).

This is the first study to simultaneously assess changes in TWT, W:L, IMT and LD in trained limbs (popliteal arteries), non-trained limbs (brachial arteries) and the carotid arteries after marathon training. As hypothesized, the experimental group experienced significant changes in TWT, W:L, and LD in the brachial artery and popliteal arteries after training and brachial IMT trended towards significance. IPAQ data showed that the experimental group was involved in significantly more moderate and vigorous activity than the control group before (p = .000) and after training (p = .000). This data also showed that vigorous activity increased significantly within the experimental group by the end of the training program (p = .038). Although VO_2max did not change in the experimental group, a 5 % increase in running speed at lactate threshold (p = .008) and lower lactate concentrations at a constant running speed (p = .001) was observed (Table 2). These data establish that the activity levels between the two groups were significantly different before and after training and that the marathon training elicited additional physiological adaptations.

The experimental group had significantly larger LD in the brachial (p = .002) and popliteal (p = .007) arteries than the control group during initial testing, whereas no differences were observed in the carotid LD. Initial testing TWT, W: L and IMT measurements were not different between the groups in any of the arteries (Table 3). The LD data is consistent with cross-sectional data presented by Rowley et al. [23] who found that LD in the brachial and femoral arteries of athletes was greater than inactive age-matched controls, but that carotid LD was not different. In contrast to the present study, these researchers [23] also reported that arterial TWT was lower in athletes when compared to controls. There are several possible explanations for this discrepancy.

Rowley et al. [23] proposed that exercise induced alterations in arterial dimensions are influenced by different mechanisms, with increases in LD caused by local factors (i.e. shear induced NO production) and reductions in TWT mediated by systemic mechanisms (e.g., arterial pressure). Similarly, Thijssen et al. [29, 32] found that in the weeks following spinal cord injury induced paralysis that the associated reduction in femoral LD stabilized after three weeks while TWT had not changed at three weeks but was significantly increased after 24 weeks. The authors concluded that changes in LD and TWT following spinal cord injury were on different time courses and that they may be influenced by different mechanisms [32]. Given the potential variability of the physical activity within the experimental group prior to initiating the marathon training program as well as the data suggesting that arterial adaptations may be dependent on the time course of the stimulation, it is not surprising that LD was found to be different between the two groups while TWT was not. Other evidence shows that TWT increases with age [18, 33] and that the increases in TWT are proportionally greater than age-related changes in LD [33]. These data further suggest that younger individuals with thinner vessel walls may experience a proportionally greater increase in LD compared to reductions TWT with training.

The between group differences found in the IPAQ and arterial dimension data provides evidence of a comparatively mild effect of physical activity (i.e., greater brachial and popliteal LD in the experimental group) that existed prior to beginning participation in the training intervention. In combination with the training effect seen in the experimental group, as discussed below, the initial between group differences suggest that arterial adaptations may be dose-dependent.

Consistent with previous research we observed no training adaptations in the carotid artery [22‐25] but found significant remodeling in the popliteal artery (trained limb) of the experimental group [4, 8, 23‐25, 27] (Table 3). Popliteal adaptations included reductions in TWT (-10.4 %) and W: L (-15.0 %) as well as an increase in LD (8.0 %). No change was found in the popliteal IMT in the experimental group.

Similar to Green et al. [8], but in contrast to others [4, 8, 25, 27, 28], we found that brachial artery, or non-trained limb, dimensions were significantly altered by the 12 weeks of training (Table 3). Reductions were seen in TWT (-15.2 %) and W: L (-17.2 %), while LD increased (5.6 %) and IMT trended toward a significant reduction after allowing for the Bonferroni statistical correction (-14.3 %; p = 0.032). These results indicate that arterial adaptations to exercise training can be manifested in the non-trained limb (e.g. brachial artery) as well as the trained limb (e.g. popliteal artery). The data further suggest that adaptations may be dependent on training volume. This interpretation is supported by the evidence that the experimental group had greater LD in the brachial and popliteal arteries than the control group at the beginning of the study combined with the fact that LD further increased when vigorous training volume was increased. Our observations of increased LD in the brachial and popliteal arteries of the experimental group are consistent with data reported elsewhere [8]. Green et al. [8], for example, found that brachial LD increased by 4 % and popliteal LD increased by 12 %. Thijssen et al. [28] found that femoral LD increased by 7 % with training. The reduction in TWT and W: L in the brachial and popliteal arteries observed in this study are also consistent with Green et al. [8] who reported reductions in brachial TWT (14 %) and W: L (17 %) as well as reductions in popliteal TWT (10 %) and W: L (14 %).

Several mechanisms potentially contribute to arterial changes with training [27]. Included among these are, the endothelial derived NO-dilator system, exercise induced variations in systemic blood pressure as well as adaptations to oxidative stress and localized inflammation.

In the context of the present study, the role of NO is of particular interest given its responsiveness to variations in blood flow. Specifically, as blood flow increases, there is an increase in the shear stress acting on the vessel. With the increases in arterial blood flow that accompanies exercise, shear stress stimulates activity of the NO-dilator system, including the up-regulation of endothelial-derived NO synthase, to buffer the increased shear [16]. Although the precise mechanisms have yet to be elucidated, the increased activity of the NO-dilator system is associated with the modulation of platelet, macrophage and endothelial derived growth factors that favors increased LD and reductions in arterial TWT. This arterial remodeling continues until the shear stress is normalized and the NO activity returns back to baseline levels [16].

The simultaneous assessment of the carotid, brachial and popliteal arteries as performed in this study combined with blood flow and endothelial function data from other research provides a basis for explaining the potential role of shear stress and the NO-dilator system in arterial remodeling. There is clear evidence that blood flow to different regions of the body is varied during exercise [1, 2] and that it is further influenced by exercise mode [3] and intensity [7, 26]. While no data was found regarding blood flow to the legs and arms during running, Calbet et al. [3] showed that leg-only cross-country skiing lead to a second to eight-fold increase in femoral blood flow and a two to three-fold increase in subclavian blood flow. Other studies show that when the arms are immobilized during leg exercise, blood flow in the brachial artery increases as much as four-fold [7, 26] and that the higher flow is adequate to stimulate the NO-dilator system [9, 17]. In contrast, blood flow to the carotid artery during exercise increases to a lesser degree. Hellstrom et al. [10] found that the maximum change in blood flow to the head during exercise is 30-40 % higher than resting values. Tanaka et al. [26] found that blood flow to the active limbs during exercise increases in proportion to exercise intensity, while Goto et al. [7] showed that blood flow within a range of intensities is needed to stimulate the NO-dilator system and potentially alter arterial dimensions. These researchers reported that moderate intensity exercise (50 % Vo_2max) is within the range required to enhance endothelial function in the non-training limb [7]. On the other hand, they found that low intensity exercise (25 % Vo_2max) failed to stimulate the NO-dilator system, and any stimulatory effect associated with exercise at higher intensities (75 + % Vo_2max) may be countered by an inflammatory response [7].

Given that leg exercise substantially increases blood flow in the legs and arms [3, 9], and that flow to the arms is adequate to invoke endothelial-derived dilation [9], it is plausible that the brachial and popliteal adaptations in LD, TWT, W: L found in the experimental group were mediated by the NO-dilator system. Although some researchers [4, 28] report that leg exercise is not associated with alterations in brachial artery dimensions, we propose that the adaptations found in the present study were a function of repeated bouts of moderate to high intensity exercise that were performed during 12 weeks of training. This exercise volume and intensity may have exceeded that used in studies that failed to show brachial artery adaptations to lower limb exercise training [4, 28]. Moreover, the fact that LD, but not TWT, was different between the two groups prior to training and that LD, TWT and W: L changed in the experimental group supports this position.

Despite the significant body of evidence indicating the NO-dilator system’s role in regulating arterial anatomy, the cyclical change in blood pressure that accompanies exercise training has not been ruled out as a mechanism responsible for these changes. Chronic elevations in transluminal pressure against the arterial wall, as seen in hypertension, activates pro-atherogenic endothelial cell phenotypes known to contribute to arterial wall thickening [6]. Exercise training, however, is associated with transient and cyclical increases in pressure; and, it has been proposed that the regulation of pro-atherogenic and anti-atherogenic genes is favorably altered by these cyclical variations in pressure [20]. Testing this theory is confounded by the fact that the exercise stimulus leading to increases in pressure are also associated with increases in blood flow and shear stress [20]. In one study, however, Rowley et al. [22] suggested that differences in brachial and femoral TWT between elite squash players and sedentary participants combined with a lack of any difference between arterial TWT in the dominant and non-dominant arms of squash players supports the systemic mechanism theory. Data from the study by Rowley et al. [22] support the need for more research designed to better explain the role that blood pressure plays in arterial remodeling.

An intriguing finding in the present study was that following the marathon training brachial IMT in the experimental group, when statistically adjusted for multiple comparisons, trended toward significance (p = 0.032) while popliteal IMT remained unchanged (p = 0.260). This result is in contrast with data presented by others [4, 8, 25] showing that IMT decreased in the trained limb as a result of exercise training. It is possible that the discrepancy is due to localized inflammation in the lower extremities associated with a relatively high volume of training (i.e., 36 total miles, inclusive of an 18 mile run) that occurred during the week prior to data collection. In response to this training load, the participants may have experienced sufficient micro-trauma in the legs to stimulate localized inflammation. This explanation is supported by limited data [7, 19] and constitutes an area of research that has been identified as one in need of further investigation [30].

An alternate explanation for the lack of change in popliteal IMT may be related to variations in oscillatory pattern of blood flow in limbs during exercise. Thijssen et al. [31] and Tinken et al. [32] demonstrated that the NO-dilator activity is directly related to ratio of antegrade to retrograde flow; and vessel occlusion reduces the ratio in favor of retrograde flow and diminished NO-dilator activity. The reduction in NO-dilator activity is proportional to the degree of occlusion, and even a mild occlusion (i.e., cuff pressure at 25 mmHg) is adequate to stimulate a significant change in the ratio and NO-dilator activity. Based on this data, it seems plausible that during the loading phase of the running stride, occlusion of the conduit arteries in the leg would favor a reduction in the antegrade to retrograde flow and perhaps attenuate NO-dilator activity and any related effect on IMT.

A limitation of the study is that the initial training before the marathon class began was unsupervised. Therefore it is not clear as to what type or amount of training was being done initially. Another limitation of the study is that due to the unique loading of the arms and legs during running, the results are not generalizable to cycling (static loading of the arms) or activities that place a greater load on the arms (cross country skiing, circuit training).