Is there an outcome penalty linked to guideline-based class-I indications for valvular surgery? long-term analysis of patients with high gradient severe aortic stenosis

Abstract Background Current ACC/AHA guidelines recommend aortic valve replacement (AVR) as class-I triggers in patients with high-gradient severe aortic stenosis (HGSAS) in presence of symptoms and/or left ventricular dysfunction (LVEF<50%). Whether waiting for these triggers could be associated with long-term outcome penalty after AVR remains poorly studied. Objective To examine the impact of guideline-based class-I triggers on long-term post-operative survival in patients with HGSAS. Methods We used an international registry including consecutive patients operated on for HGSAS between 2000 and 2017. 2030 Patients were included in the analysis and retrospectively classified according to the guideline-based indication: no Class-I trigger (no symptoms and LVEF>50%, n=853), Symptoms with LVEF>50% (n=965), or LVEF<50% (regardless of symptoms, n=212). Survival was compared in a multivariate Cox-model and after inverse probability weighting-[IPW] for relevant variables. Finally, we explored the most sensitive left ventricular ejection fraction threshold for predicting mortality risk. Results 10-years survival was better among patients without any class-I trigger than with symptoms or LVEF<50% (67±3 vs 56±3 vs 53±7% respectively, p<0.001). After adjustment for covariates, risk of death increased significantly when patients were operated on with symptoms (HR: 1.48, [95%CI: 1.17-1.86]) or with LVEF<50% (HR: 1.47, [95%CI: 1.05-2.06]) compared to patients with no class-I trigger. Furthermore, LVEF<55% emerged as a more sensitive threshold for the prediction of post-operative mortality in comparison with LVEF<50% (see table), allowing a better separation of survival curves and indicating that patients with LVEF 50-55% are already at risk with long term consequences even after AVR. Interestingly, among asymptomatic patients with LVEF >55%, performing AVR restituted a normal life expectancy (comparable to the Belgian population of same age, see figure). Finally, the outcome penalty after AVR when waiting for symptoms or LVEF<55% was confirmed in IPW analysis (HR: 1.43, [95%CI: 1.13-1.82] and HR: 1.63, [95%CI: 1.19-2.23], respectively). Conclusions Guideline-based Class-I triggers for AVR in patients with HGSAS is associated with profound outcome consequences on long-term postoperative mortality. Our data argue that patients with HGSAS should be operated on before the onset of these triggers. Finally, our data suggest that a threshold of LVEF<55% is a stronger predictor of outcome than LVEF<50%.