Individualized treatment responses to cardiac resynchronization therapy: a cluster analysis of the Swedish heart failure registry
R Laborante, A Uijl, C Basile, L Benson, P Gatti, B N Beer, V Valente, D D'amario, C C Cabrera, R Scorza, G SavareseAbstract
Background
In real-world clinical practice, patients with heart failure (HF) eligible for cardiac resynchronization therapy(CRT) are highly heterogeneous with respect to comorbidities, sex, and etiology. It is not well investigated whether these factors define subgroups where CRT implantation might lead to more or less benefit.
Purpose
Aim was to assess whether CRT effectiveness was consistent in patient clusters defined considering comorbidities, sex, and etiology and derived by using latent class analysis(LCA).
Methods
Patients in the Swedish HF Registry who met the 2021 ESC HF guidelines recommendation for CRT (i.e., reduced ejection fraction, QRS duration ≥150ms, and on optimal medical therapy) between 2014 and 2022 were included. LCA was used to identify clusters based on sex, etiology, and various comorbidities, including arterial hypertension(HTN), cancer, diabetes, stroke, peripheral artery disease(PAD), atrial fibrillation(AF), and chronic obstructive pulmonary disease. The primary endpoint was the composite of time to first HF hospitalization(HFH) or cardiovascular(CV) death. Secondary endpoints were individual components of the primary outcome, and all-cause death. Heterogeneity in the association between CRT use and outcomes according to cluster membership was assessed by using Cox regression models with overlap weighting to balance baseline clinical and laboratory characteristics, and including an interaction term between CRT and cluster membership.
Results
Out of 3530 patients with HF and an indication for CRT[median age 75 years, 77% males], 25% received a CRT[median age 73 years, 77% males]. Five distinct patient clusters were identified: Cluster 1(30%) included patients(74% males) with non-ischemic HF(62%), HTN(100%), a large proportion with AF(68%), and low prevalence of other comorbidities(i.e., 64% with <3 comorbidities); Cluster 2(17%) included patients with high comorbidity burden, i.e. ≥3 comorbidities—most commonly HTN(100%), AF(91%), diabetes(61%), stroke(39%), PAD(25%), and cancer(24%)—and were almost exclusively males(98%) and with ischemic HF(99%); Cluster 3(17%) included approximately one third females(32%), with ischemic HF(100%), HTN(96%), and no history of AF(95%) and of cancer(100%); Cluster 4(19%) included males(99%) with ischemic HF(91%), AF(57%), and low prevalence of other comorbidities(only 12% with ≥3 comorbidities); and Cluster 5(18%) included predominantly females(52%) with non-ischemic HF(89%), and low comorbidity burden(only 1% with ≥3 comorbidities)(Figure 1). Across the different clusters, CRT use was consistently associated with lower adjusted risks of CV death/first HFH, its individual components, and all-cause death(Figure 2).
Conclusions
The use of CRT was associated with reduced mortality/morbidity regardless of cluster membership, i.e., a heterogeneous population of patients who met criteria for the device. Our findings call for broader implementation of CRT in clinical practice.For image description, please refer to the figure legend and surrounding text.For image description, please refer to the figure legend and surrounding text.