Myocardial blood flow in aortic stenosis: predictors of recovery after aortic valve replacement

Background Aortic stenosis (AS) commonly leads to myocardial hypoperfusion in the absence of obstructive epicardial coronary disease, a potential driver of irreversible myocardial scarring, associated with cardiac decompensation and death.(1,2) Long-term prognosis following aortic valve replacement (AVR) remains poor and the determinants of reverse remodelling remain poorly understood. Cardiovascular magnetic resonance (CMR) can determine myocardial remodelling, quantification of scar burden and rest/stress myocardial blood flow (MBF). Purpose Using a multi-modal approach, we aimed to demonstrate the reversal of myocardial hypoperfusion and fibrosis burden one year following aortic valve replacement for AS and to understand the predictors of post-AVR myocardial perfusion. Methods Patients with severe symptomatic AS referred for AVR were recruited and underwent deep phenotyping (CMR including adenosine stress perfusion and multiparametric mapping; echocardiography; and cardiac biomarkers [NT-proBNP/HsTnT]) at baseline and 1-year following AVR. CMR quantified extracellular volume fraction (ECV%), late gadolinium enhancement (LGE), rest and stress MBF using automated in-line pixel-wise imaging. Multivariable linear regression models were fitted to assess baseline predictors of 1-year global stress MBF. Model fit and parsimony was assessed using Akaike’s Information Criterion (AIC). Results 71 patients (age 69.6±8.6years, 70.4% male, Vmax 4.2±0.4 m/s, AVA 0.82±0.22cm2/m2) underwent paired stress perfusion imaging. At a median follow-up of 389 days post AVR (55 [77.5%] surgical AVR / 16 [22.5%] TAVI), indexed left ventricular (LV) volume (-6%), mass (-20%), diffuse and focal fibrosis significantly regressed and global stress MBF improved significantly (1.78±0.49 to 2.30±0.64 ml/g/min, p<0.001) (Table 1, Figure 1). The greatest improvement in perfusion was seen in the subendo- than subepicardium (+0.55±0.64 vs +0.49±0.81 ml/g/min). On multivariable linear regression, baseline LV mass (b -0.13 [-0.024 to -0.002]; p=0.022) and baseline matrix volume (b -0.003 [-0.007 to 0.000]; p=0.049) were predictors of global stress MBF at 1-year post AVR in their respective models, along with age, sex, and concomitant hypertension, independently of BMI, the presence of coronary artery disease, AS severity and baseline global stress MBF. Both models demonstrated similar model fit and parsimony (AIC 121.1 vs. 122.7). Conclusions 1 year post AVR for severe AS, left ventricular mass regression is accompanied not only by improvement of stress myocardial perfusion but also regression in focal and diffuse fibrosis. Improvement in stress myocardial blood is independently predicted by the degree of myocardial hypertrophy and diffuse fibrosis. These therefore not only provide better post AVR outcome prediction but are also promising targets for post AVR pharmacological intervention.