Phenotyping patients with aortic stenosis using cluster analysis to determine mortality and suitability for transcatheter aortic valve replacement. (3rd October 2022)

• Record Type: Journal Article
• Authors: Sen, J; Pires, D; De Sa, A; Ascher, D; Wahi, S; Marwick, T
• Is Part Of: European heart journal. Volume 43(2022)Supplement 2
• Publication Date: 2022-10-03
• Subjects: Cardiology -- Periodicals; Heart -- Diseases -- Periodicals; 616.12005
• DOI: 10.1093/eurheartj/ehac544.1627
• Languages: English
• ISSNs: 0195-668X

Abstract

Abstract: Background: Current classification of aortic stenosis (AS) is based on guideline-recommended echocardiographic criteria. Heterogeneity of patients with AS is increasingly recognised. Clinical and demographic factors in addition to echocardiographic parameters can determine those who may derive the greatest benefit from transcatheter aortic valvular replacement (TAVR) and influence patient outcomes. Purpose: Our study aims to define distinct AS echocardiographic and clinical phenotypes and to accurately identify patients most likely to die or benefit from TAVR. Methods: Patients diagnosed with at least mild AS between 2009 and 2021 (pre-TAVR) from a multicentre echocardiographic database at a quaternary referral centre included. Unsupervised clustering analysis was performed using K-means, partitioning around medoids, density-based spatial clustering, hierarchical clustering algorithms on 56 demographic, echocardiographic and clinical variables. Associations between AS clusters and clinical outcomes (all-cause mortality, cardiovascular death, AS-related death), and effect of TAVR on clinical outcomes were assessed using Cox proportional hazards models. Results: Four AS clusters were identified amongst 2, 456 patients with median follow up of 4.7 years (median age: 77 years, male: 66%). Cluster 1 (n=542) had the lowest aortic valvular area (AVA, mean 0.89 cm 2 ), highest peak velocity (Vmax) (4.3 m/s), mean gradient (45 mmHg), and the most bicuspid valves (12.7%). Cluster 2 (n=827) had 50% women, mostly in sinus rhythm and less severe echo findings. Cluster 3 (n=592) had predominantly males (85%) with a mean AVA of 1.65 cm 2 and the most cardiovascular risk factors (hypertension, diabetes, hyperlipidaemia, stable angina, acute coronary syndrome, and atherosclerosis). Cluster 4 (n=495) had the highest left atrial size (mean 32 cm 2 ), the most atrial fibrillation (82%), heart failure (80%), rheumatic heart disease (26%) and chronic kidney disease (55.8%), but only moderate AS (AVA 1.3 cm 2, Vmax 3 m/s). All-cause mortality was highest in Cluster 4 (Hazard ratio (HR) 1.57, 95% CI: 1.33–1.85) and AS-related death was highest in Cluster 1 (HR 3.96, 95% CI: 2.61–5.99, Figure 1A). TAVR reduced AS-related death in only Cluster 1 (HR 0.22, 95% CI: 0.05–0.88, p=0.033, Figure 1B). Conclusions: We demonstrated that phenotypic classification via a combination of demographics, echocardiography and comorbidities can significantly improve management of AS. This personalised approach can be implemented to identify patients most likely to die and most likely to benefit from TAVR. Funding Acknowledgement: Type of funding sources: Foundation. Main funding source(s): National Heart Foundation of Australia (ID: 102578)National Health and Medical Research Council of Australia (ID: 1191044)