Is exercise really deleterious for the hypertensive heart?


  • Joseph R. Libonati

    1. University of Pennsylvania, Biobehavioral Health Sciences, 126B Fagin Hall, School of Nursing, 416 Curie Blvd, Room 135, Philadelphia, PA 19104, United States
    Search for more papers by this author


The recent paper by da Costa Rebelo et al. (2012) suggests that access to free running wheels deleteriously impacts cardiac phenotype in female, 1-year-old spontaneously hypertensive rats (SHR) over a 6-month follow-up. The authors emphatically concluded that ‘high physical activity in hypertensives must be considered as an important risk factor rather than a therapeutic intervention’. Given that the results of this paper conflict with multiple exercise training studies in various models of pressure overload (Scheuer et al. 1982; MacDonnell et al. 2005; Konhilas et al. 2006; Libonati & Gaughan, 2006; Reger et al. 2006; Renna et al. 2006; Kolwicz et al. 2007, 2008, 2009; Chicco et al. 2008; Garciarena et al. 2009; Libonati et al. 2011; Rossoni et al. 2011; Huang et al. 2012; Marshall et al. 2012), I feel it is imperative to further address its implications. Their paper is particularly interesting in light of reports in both humans (Oxbourough et al. 2010; O’Keefe et al. 2012) and animals (Schultz et al. 2007; Huang et al. 2009; Benito et al. 2011) showing that extreme levels of exercise can, in fact, be damaging to the heart.

da Costa Rebelo et al. reported that wheel running velocity in SHR animals was positively correlated with left ventricular (LV) mRNA of tumour growth factor (TGF)-β1, collagen III and biglycan, whereas wheel running velocity was negatively correlated with SERCA2A to Na+-Ca2+ exchange ratio. The authors also showed data for increased fibrosis with Picrosirius red, and reported that 67% of SH rats died either spontaneously or had to be killed during the study's 6-month follow-up. Wheel running in the presence of captopril (30 mg kg-1 day-1) was protective from the adverse pro-fibrotic training-induced SHR phenotype, suggesting a role for the renin-angiotensin system (RAS) in the adverse response to training.

The SHR animals in the da Costa Rebelo et al. study accrued a mean distance of 48.9 km per week (mean velocity 2.43 km h-1). This is a considerable workload for SHR's, as the total daily work performed was 6 to 7 fold higher than previous studies using shorter term, moderate intensity treadmill running programmes (MacDonnell et al. 2005; Reger et al. 2006; Kolwicz et al. 2009). One would anticipate that this high level of exertion is associated with both lower body weights and blood pressures. However, these same criteria were used to remove animals from da Costa Rebelo's study –‘Animals were removed from the study if they developed constitutive loss of body weight and blood pressure for 3 consecutive weeks’. Based on the reported data, it is unclear whether the removed animals were in heart failure, as is implied by the authors, or whether the loss of body weight and reduced blood pressure were typical responses to the extreme levels of exercise performed. One important distinguishing piece of data that was not included in the paper is whether accrued wheel distances also significantly declined at the time intervals that animals were removed from the study. These functional data in conjunction with other important clinical markers indicative of heart failure, such as dyspnoea, ascites and increased wet to dry lung weight ratio, would have greatly supported the contention of heart failure.

Given that the authors initially started out with only six animals in the SHR exercise group, and four of the six animals were removed from the study at various time intervals, the overall low animal numbers at variable time intervals is of concern in establishing the presented correlations between the acute wheel running variables (distance, time and velocity), pro-fibrotic biomarkers and calcium handling proteins. The issue is that much of the conclusion is predicated on these correlations, with one animal at the highest extreme of running velocity in the SHR trained group having a significant influence on the direction and magnitude of the relationships. Moreover, little information on Ca2+ regulation can be inferred from the SERCA2A to Na+-Ca2+ exchanger ratio, as there are parallel shifts in other Ca2+ handling proteins such as phospholamban phosphorylation with exercise training in pressure overload (MacDonnell et al. 2005).

With respect to the physiological data, there was a reduction in systolic blood pressure in SHR-trained versus SHR-sedentary animals. This finding is consistent with the literature, and might be secondary to shifts in body mass (but these data were not presented). The lower ex vivo Left ventricular developed pressure (L Dev P) and its first derivative (+dP/dt) during the Langendorff perfusions are difficult to interpret without some index of exercise-induced cardiac remodelling, such as echocardiography, histomorphometry, absolute heart/LV weight, or even LV balloon volumes. These variables would have helped support whether similar loading conditions between SHR-trained and SHR-sedentary hearts were established, particularly because four of five animals in the SHR group had LV developed pressures below 100 mmHg. These low LV developed pressure values during Langendorff perfusions may reflect heart failure as is suggested in the paper, but may also be secondary to lower preload-induced LV stretch in remodelled, exercise-trained hearts. Establishing a dynamic range by which to establish systolic performance (pressure–volume relationships, systolic elastance) would have been a stronger approach to support the contention that wheel running augmented the development of heart failure in SHR.

Despite these potential methodological concerns, the data by da Costa Rebelo et al. are enticing and need to be incorporated into the overall knowledge base regarding the effects of exercise on the hypertensive heart. Their study raises several important questions. Is there a threshold exercise volume/intensity that triggers decompensation in the hypertensive heart? Does wheel running induce different and distinct phenotypical changes in the heart relative to other forms of exercise? To what extent does the progression of disease, age and sex play in the response? The aim of this letter is to further the dialogue regarding this paper so that the results of da Costa Rebelo et al. can be placed into better context with the extant literature, which is heavily suggestive of a putative cardiac phenotype with exercise in pressure overload. Only through such scientific dialogue can we begin to translate these preclinical findings to humans.