This Week’s Most Ridiculous Heart Health Headline: “Running One Marathon Can Make Your Arteries Healthier”

Study limitations

This study was conducted in healthy individuals; therefore, our findings may not apply to patients with hypertension who have stiffer arteries that may be less modifiable (40). From these data, however, those with higher SBP at baseline appeared to derive greater benefit. This study was not designed to provide structured training, but rather to observe the effects of real-world preparation for a marathon, which randomized control trials cannot address. Nevertheless, information on the intensity, frequency, and type of exercise training would have been valuable to understand further the beneficial effects on aortic stiffness. The modest change in peak VO₂ may be related to exercise training intensity or low adherence, which reflects the real world. Peak VO₂ was performed semisupine to allow concurrent echocardiography, and this may also have reduced sensitivity to changes due to running or running efficiency. We assessed only marathon finishers—plausibly, nonfinishers could have had different vascular responsiveness. The causal link of exercise to measured changes is only inferred—marathon training may lead to other lifestyle modifications (dietary, other behavioral factors), or alterations in lipid profiles and glucose metabolism, although these have not been previously associated with changes in aortic stiffness (11). We did not examine the effect of exercise on peripheral arteries or endothelial dysfunction. Although individual participants served as internal controls, there may have been run-in bias for the initial BP measurement. This appears unlikely, as BP changes would not have been age-related nor correlated with the change in separate measures (e.g., aortic stiffness) with training. Estimated aortic ages are approximations and are based on the same dataset at baseline rather than independent observations. The exercise dose-response curve here is not sampled—only training for a first-time marathon with single timepoint assessment. This area warrants further study. We measured distensibility on modulus imaging acquired at 1.5-T rather than steady-state free precession imaging. The free-breathing sequence we used achieved good temporal resolution, but may be susceptible to through-plane motion. However, this and similar sequences correlate well with breath-held cine imaging, and show similar associations with aging (18). If error was introduced into distensibility measurements related to through-plane motion, the resultant noise would minimize the effect size related to exercise training, and therefore would be unlikely to account for our key findings. PP undergoes amplification from central to more peripheral locations, typically being ∼6 mm Hg higher in the descending thoracic than the ascending aorta (20). This PP amplification is not accounted for in our analysis, because it would have involved invasive measures of aortic pressure at each location. A sensitivity analysis suggested that the likely impact of this effect on the observed changes after training would be minimal; however, we cannot completely exclude the possibility that changes in PP amplification contribute to the observed differences. Diaphragmatic descending aortic distensibility data reported here were, however, higher than expected, although there is limited published data for comparison (41). Unlike Voges et al. (41), central rather than brachial PP was used, which would explain greater distensibility, and the use of 1.5-T phase-contrast modulus may accentuate image contrast differences between 3T gradient echo sequences.