Cardiology Department, Lille Catholic Hospitals, Lille Catholic University, 59160 Lomme, France.
Biostatistics Department-Delegations for Clinical Research and Innovation, Lille Catholic Hospitals, Lille Catholic University, 59160 Lille, France.
Arch Cardiovasc Dis. 2021 Mar;114(3):197-210. doi: 10.1016/j.acvd.2020.07.004. Epub 2021 Jan 8.
Despite having an indication for cardiac resynchronization therapy according to current guidelines, patients with heart failure with reduced ejection fraction who receive cardiac resynchronization therapy do not consistently derive benefit from it.
To determine whether unsupervised clustering analysis (phenomapping) can identify distinct phenogroups of patients with differential outcomes among cardiac resynchronization therapy recipients from routine clinical practice.
We used unsupervised hierarchical cluster analysis of phenotypic data after data reduction (55 clinical, biological and echocardiographic variables) to define new phenogroups among 328 patients with heart failure with reduced ejection fraction from routine clinical practice enrolled before cardiac resynchronization therapy. Clinical outcomes and cardiac resynchronization therapy response rate were studied according to phenogroups.
Although all patients met the recommended criteria for cardiac resynchronization therapy implantation, phenomapping analysis classified study participants into four phenogroups that differed distinctively in clinical, biological, electrocardiographic and echocardiographic characteristics and outcomes. Patients from phenogroups 1 and 2 had the most improved outcome in terms of mortality, associated with cardiac resynchronization therapy response rates of 81% and 78%, respectively. In contrast, patients from phenogroups 3 and 4 had cardiac resynchronization therapy response rates of 39% and 59%, respectively, and the worst outcome, with a considerably increased risk of mortality compared with patients from phenogroup 1 (hazard ratio 3.23, 95% confidence interval 1.9-5.5 and hazard ratio 2.49, 95% confidence interval 1.38-4.50, respectively).
Among patients with heart failure with reduced ejection fraction with an indication for cardiac resynchronization therapy from routine clinical practice, phenomapping identifies subgroups of patients with differential clinical, biological and echocardiographic features strongly linked to divergent outcomes and responses to cardiac resynchronization therapy. This approach may help to identify patients who will derive most benefit from cardiac resynchronization therapy in "individualized" clinical practice.
尽管根据现行指南，射血分数降低的心力衰竭患者有心脏再同步治疗的适应证，但接受心脏再同步治疗的患者并不总能从中获益。
确定非监督聚类分析（表型映射）是否可以识别常规临床实践中接受心脏再同步治疗的患者中具有不同结局的不同表型组。
我们使用数据简化后的表型数据（55 个临床、生物学和超声心动图变量）进行无监督层次聚类分析，以定义来自常规临床实践的 328 例射血分数降低的心力衰竭患者中接受心脏再同步治疗的新表型组。根据表型组研究临床结局和心脏再同步治疗反应率。
尽管所有患者均符合心脏再同步治疗植入的推荐标准，但表型映射分析将研究参与者分为四个表型组，这些组在临床、生物学、心电图和超声心动图特征和结局方面存在明显差异。表型组 1 和 2 的患者死亡率改善最明显，心脏再同步治疗反应率分别为 81%和 78%。相比之下，表型组 3 和 4 的患者心脏再同步治疗反应率分别为 39%和 59%，且预后最差，与表型组 1 的患者相比，死亡率明显升高（危险比分别为 3.23，95%置信区间为 1.9-5.5 和 2.49，95%置信区间为 1.38-4.50）。
在常规临床实践中具有心脏再同步治疗适应证的射血分数降低的心力衰竭患者中，表型映射可识别具有不同临床、生物学和超声心动图特征的亚组患者，这些特征与不同结局和心脏再同步治疗反应密切相关。这种方法可能有助于在“个体化”临床实践中确定最能从心脏再同步治疗中获益的患者。
