This is the first study to use CMR to investigate both the magnitude of the hypertrophic response and the different morphological patterns of remodelling and hypertrophy observed in patients with aortic stenosis. We have demonstrated that the degree and pattern of hypertrophy is independent of the severity of valve stenosis severity. We observed six distinct patterns of left ventricular adaption, which in line with previous echocardiographic definitions included normal left ventricular structure, concentric remodeling, and concentric hypertrophy. However in contrast to previous classifications, we also describe asymmetric remodeling and asymmetric hypertrophy as common structural variants.
Magnitude of the hypertrophic response
The magnitude of left ventricular hypertrophy varied widely but was unrelated to the severity of aortic stenosis, such that patients with severe valve narrowing were found to have normal ventricular structure whilst patients with moderate disease often had extensive hypertrophy. This observation is consistent with the findings of several previous echocardiographic studies [18–20], but has not previously been reported using CMR.
The apparent disconnect between processes in the valve and myocardium might explain the marked heterogeneity between the severity of valve stenosis and symptom onset, and is important because an advanced hypertrophic response is associated with an adverse prognosis in a range of cardiac conditions including aortic stenosis [3–5, 21]. Recent data has suggested that this might reflect its association with mid-wall fibrosis: an independent predictor of all-cause mortality in patients with moderate and severe disease . We therefore believe that when considering overall aortic stenosis severity, attention should be paid not only to the degree of valve narrowing but also to the hypertrophic response accompanying it.
Consistent with previous studies, male gender was the only variable associated with an increased left ventricular mass [23, 24] although an apparent trend was observed with concomitant hypertension. The latter was also seen in an analysis of the Simvastatin and Ezetimibe in Aortic Stenosis (SEAS) trial where hypertension predicted increased left ventricular mass in patients with aortic stenosis independent of other known confounders . Strict blood pressure control may therefore provide an important clinical means of blunting the hypertrophic response.
Importantly the lack of correlation between valve stenosis and the hypertrophic response persisted in subgroup analyses of our male, female and normotensive patients indicating that gender and blood pressure cannot in themselves explain this observation. Genetic factors were not investigated in this study and are likely to play an important role. These are known to modulate the magnitude of the hypertrophic response to a number of physiological and pathological triggers [26–28]. These include aortic stenosis, in which polymorphisms of the angiotensin-converting enzyme I/D genotype have been associated with different degrees of wall thickening and hypertrophy , as well as the regression of these processes following valve replacement .
Asymmetric remodeling and hypertrophy
An asymmetric pattern of wall thickening was observed in 27% of our cohort, being particularly prevalent amongst the elderly and in those with hypertension. This confirms the echocardiographic findings of Tuseth and colleagues who described asymmetric wall thickening in 22% of patients with aortic stenosis, and also observed an association with co-existent hypertension . Importantly our study demonstrated asymmetric wall thickening in a sixth of normotensive patients with aortic stenosis, confirming that it is not only a response related to high blood pressure.
Asymmetric wall thickening was most frequently observed in the septum at the basal and mid-cavity levels with a mean of 17 mm and a maximum of 22 mm. Current guidelines recommend that a diagnosis of hypertrophic cardiomyopathy be considered if regional wall thickness exceeds 15 mm, with an intermediate area existing between 13 and 15 mm . Our data therefore suggest considerable overlap in the appearances of asymmetric wall thickening in these two conditions and underlines the fact that the morphological diagnosis of hypertrophic cardiomyopathy may be challenging or impossible in the context of an increased afterload. Furthermore it is plausible that specific genotypes, related to those causing hypertrophic cardiomyopathy, may predispose to an asymmetric rather than concentric remodeling response. In line with this theory, patients with hypertension and asymmetric thickening have a higher familial incidence of hypertrophic cardiomyopathy and more myocardial disarray .
Left ventricular decompensation
Eccentric hypertrophy has been included in previous definitions of remodeling and hypertrophy in aortic stenosis but this was not observed in our study . The strict selection criteria employed in our population excluded patients with conditions such as aortic and mitral regurgitation and ischemic heart disease that in clinical practice will result in a composite form of left ventricular adaption and perhaps explain the eccentric phenotype observed in previous cohorts. Instead we observed a form of left ventricular decompensation characterized by impaired systolic function and left ventricular dilatation. This form of remodeling can be considered as the end-stage of the hypertrophic process in which the left ventricle has failed in the face of an increased afterload. Given that a reduction in ejection fraction is a powerful prognostic marker in aortic stenosis, this pattern of LV adaption is likely to be associated with an increased mortality . Interestingly not all patients with a reduction in ejection fraction conformed to this phenotype: 41% had either concentric or asymmetric hypertrophy. It is likely that these patients are in the early stages of decompensation without having yet proceeded to left ventricular dilatation. They represent an important group to identify because prompt surgery might avoid a further deterioration in ejection fraction.
Patterns of remodeling and aortic stenosis
It has been proposed that in response to an increased afterload patients progress from a normal ventricle to LV remodeling, and then hypertrophy before decompensating and developing heart failure. However in our study, there was no clear correlation between the severity of aortic valve narrowing and the different patterns of LV adaption, except that all patients with LV dilatation had severe disease. This may simply reflect the relatively small sample size, however there are two alternative explanations. Firstly patients transition through the various stages of the remodeling process at different rates, much in the same way that they display variation in the magnitude of the hypertrophic response. Secondly patients do not in fact progress with time from LV remodeling to hypertrophy but instead these represent two distinct and independent pathways of adaption. Prospective longitudinal studies are required to address this issue.
Thirty-seven patients from our cohort had co-existent mild to moderate hypertension. Given that hypertension and aortic stenosis commonly co-exist, it was not desirable to exclude all patients with hypertension from the study, as this would have affected the generalizability of our findings. However, sub-group analysis allowed us to examine how co-existent hypertension might modulate the effect of valve narrowing. This demonstrated that hypertension was associated with an apparent increase in the LV mass index and an increased proportion of patients with asymmetric wall thickening. Importantly however the lack of correlation between the LV mass index and aortic stenosis severity remained and was independent of the effects of blood pressure.
In our institution local guidelines recommend CMR for all patients with severe aortic stenosis. However patients with moderate disease were referred at the discretion of their clinician and therefore there may have been some referral bias in this group. In addition, we have not examined patients with mild disease nor the influence of the duration of aortic stenosis. The latter is almost impossible to adjudicate because the majority of patients will have subclinical disease for many years before a murmur is detected and the diagnosis established.
Multi-centre longitudinal trials are required in an unbiased population to confirm our findings and to provide prognostic information. In the era of transcatheter aortic valve implantation (TAVI), where earlier treatment may have benefit, it will also be important to assess if the different patterns of remodeling and hypertrophy show variable potential for reverse remodeling following intervention.