Metabolic Dysfunction in Aortic Stenosis – A Key Piece of the Pathophysiological Puzzle

Aortic stenosis (AS) places an increased load on the myocardium, which according to Laplace’s law, must increase its thickness to maintain normal wall tension. Indeed, an increase in left ventricular wall thickness has been regarded as one of the hallmarks of well-compensated aortic stenosis. However, myocardial function is not normal in the hypertrophied myocardium of patients with severe AS. It has been known for many years that LV strain is reduced in these patients, even when LV ejection fraction is completely normal.¹ The myocardium in patients with severe AS also shows evidence of metabolic dysfunction,² lipid accumulation and fibrosis.³ This has been demonstrated directly using magnetic resonance imaging and spectroscopy,² and indirectly through metabolic profiling of the serum.⁴ It has not been clear, however, when these changes begin to develop, whether they are seen early in the course of AS, and whether they are fully reversible. The ability of transcatheter-based techniques to replace the aortic valve with low morbidity and mortality,⁵ makes the need to answer these questions highly compelling. The study by Monga and colleagues in this issue of the journal is, therefore, an extremely important and welcomed study.⁶

The investigators studied 74 subjects with AS, ranging from mild to severe disease. All subjects underwent cine MRI, including tagged cine images for LV strain calculation, T1 mapping, late gadolinium enhancement and MR spectroscopy. The spectroscopy protocol included P31 spectroscopy to measure high energy phosphates, specifically the PCr/ATP ratio, and proton spectroscopy to measure myocardial lipid content. In addition, 13 healthy volunteers underwent the same protocol. The subjects were analyzed in groups based on the severity of their AS, quartiles of LV wall thickness, quartiles of AV gradient (AVG) and quartiles of total LV gradient (AVG + systolic blood pressure). The key finding of the study is that a reduction in the PCr/ATP ratio, indicative of metabolic dysfunction, occurs early in the course of AS when the degree of LV hypertrophy is still fairly mild.⁶ Likewise, lipid accumulation in the myocardium was found to develop at a similarly early stage. The authors findings suggest that the myocardium may begin to suffer from metabolic dysfunction and toxicity even in subjects with mild-moderate aortic stenosis.

Monga and colleagues are to be congratulated on this study. P31 spectroscopy is technically challenging to perform and their group is one of the few with extensive experience with the technique. The study of 74 subjects with MR spectroscopy is also noteworthy. Several aspects of the study, however, merit some reflection and discussion. The most striking differences in both PCr/ATP and lipid content were seen between subjects in the first and second quartiles of both LV thickness and AVG. Many of the subjects in the first quartiles were healthy volunteers and, consequently, significant differences in age and gender were present between the first quartile and the remaining quartiles in the study.⁶ We cannot rule out, therefore, that some of the differences seen in PCr/ATP and lipid content between quartile 1 and quartile 2 were mediated by differences in age and gender. Future studies will require age and gender-matched healthy controls. Other appealing possibilities would include the investigation of younger subjects with AS related to bicuspid aortic valve anatomy and older patients with nonrheumatic calcific disease. This would allow the impact of AVG and LV thickness on myocardial energetics to be compared in patients with different ages and metabolic milieus.

The parameters measured in this study, such as LV thickness and AVG, are continuous variables. While discretizing these continuous ranges into quartiles helps underscore the point that energetic dysfunction occurs earlier in the disease course, it is also useful to examine the correlations between LV wall thickness, PCr/ATP and myocardial lipid content over their continuous ranges, as shown in Figure 6 of the manuscript supplement.⁶ The correlations, even when significant, are fairly modest and seem to be driven strongly by values at the low-mid portion of the range, where LV wall thickness ranged from 5–12mm. There seems to be little correlation between LV wall thickness and the PCr/ATP ratio with wall thickness values > 12mm. A similar phenomenon is seen when the data are divided into quartiles: Large differences are present between quartile 1 and 2 and much smaller differences thereafter.⁶ This suggests that the transition to metabolic dysfunction in aortic stenosis is non-linear and occurs, as the authors emphasize, early in the disease process. Definitive confirmation of this observation, however, will require comparison with carefully matched healthy controls as alluded to above.

The development of LV hypertrophy (LVH) in AS has been presumed to be a mechanical response to the increased intracavitary pressure in the LV, as described by the Law of Laplace. Monga and colleagues, however, raise the intriguing possibility that metabolic dysfunction and the ensuing lipid accumulation it causes may also contribute to the development LVH. This is certainly plausible but will require extensive additional study. For instance, the evolution of LVH in patients with similar gradients across the aortic valve who are/are not taking the metabolic modulator ranolazine could be compared. Similar experiments could also be performed in large animals under highly controlled conditions. Basic science studies will need to be performed to determine whether the reduction in free fatty acid metabolism and the ensuing lipid accumulation are sufficient to exacerbate LVH or whether other mechanisms are at play as well.

Preliminary observations in small animal models suggest that pressure overload may induce changes in metabolically active tissues outside the heart, such as brown adipose tissue.⁷ The metabolic derangement in AS may, therefore, not be isolated to the heart, but human studies will be needed to establish this. We hope that Monga and colleagues, as well as other groups with expertise in the imaging of cardiac energetics, will continue to perform P31-spectroscopy studies. The technical challenges of P31-spectroscopy, however, will create a barrier to its use in large multicenter clinical studies and population registries of AS. One appealing mechanism to address this will be to correlate metabolic changes by MR spectroscopy with changes in serum metabolites in small hypothesis-generating studies. These hypotheses will then be able to be tested at scale by biomarker and/or metabolomic profiling of the serum, which is feasible in large multicenter studies.^8,9

As the authors highlight, pressure overload and heart failure have been associated with a metabolic shift away from fatty acid oxidation and with mitochondrial dysfunction.¹⁰ We and others have demonstrated circulating metabolic derangements, presumed to indicate altered fatty acid oxidation that develops early in the disease course of aortic stenosis and reverses with alleviation of left ventricular unloading with transcatheter aortic valve replacement.⁴ Moreover, distinct circulating metabolomic signatures associate with indices of left ventricular volume, systolic and diastolic function, and LV hypertrophy.⁸ While these associations and the reversal of signatures of myocardial energetics with alleviation of LV afterload suggest a cardiac source for identified circulating metabolite signatures, Monga and colleagues leverage MR spectroscopy to extend these observations and clearly demonstrate in vivo altered myocardial energetics in early stages of severe aortic stenosis.⁶

For many years the management of AS was purely reactive, with surgical aortic valve replacement (AVR) being triggered by the development of symptoms. In the current era, the guidelines for AVR have become more proactive with reductions in EF, extremely high gradients and marked elevations in BNP all being indications for AVR in the absence of symptoms.¹¹ The study by Monga and colleagues suggests that the approach to AVR may need to evolve further to be driven by the development of subclinical damage to the myocardium and potentially by circulating biomarkers of myocardial stress, injury, or altered energy metabolism.⁹ Moreover, growing evidence implicating energy metabolism in the pathobiology of aortic stenosis introduces opportunities for pharmacologic intervention before or after valve intervention to potentially preserve cardiac metabolism in the face of increased afterload and to bolster cardiac recovery and improve clinical outcomes after AVR. While such developments may bring the timing of AVR forward in many patients, the low morbidity and mortality of transcatheter-AVR (TAVR) has made earlier intervention more palatable. If we are to arrive at the point where TAVR is used to prevent myocardial damage, rather than react to it, the natural history of AS will need to be exquisitely characterized. The study by Monga and colleagues makes an important contribution to this endeavor and we commend the investigators for this important study,⁶ which significantly enhances our understanding of the natural history of aortic stenosis.