Abstract 19525: Identifying Novel Phenotypes of Heart Failure Using Cluster Analysis of Clinical Variables

    loading  Checking for direct PDF access through Ovid

Abstract

Background: Heart failure (HF) is a heterogeneous syndrome. Although classification based on physical findings, biomarkers, echocardiogram or right-side catheterization are proposed, optimal subgrouping is yet to be determined.

Objectives: We aimed to evaluate whether heat map and cluster analysis of clinical variables could identify distinct HF categories.

Methods: Detailed clinical data of 303 consecutive HF patients were evaluated by heat map and hierarchical cluster analysis of 29 continuous variables, including age, vital signs, laboratory data, echocardiogram and right-side heart catheterization. Kaplan-Meier analysis and Cox proportional hazards regression analysis were performed to estimate the association between the clusters and clinical outcomes.

Results: Mean age was 70 ± 12 years old, 73% were male and mean EF was 45 ± 19 %. The heat map showed heterogeneity among study population (Figure 1). Gap analysis demonstrated 2 as the optimal number of clusters, and hierarchical cluster analysis divided the patients into cluster 1 (206 patients) and cluster 2 (97 patients). Hypertension and dyslipidemia were less common in cluster 1 than in cluster 2. Cluster 1 had lower BMI (22 ± 4 kg/m2 vs. 25 ± 5 kg/m2, p < 0.01), worse renal function (creatinine; 1.03 [0.82, 1.37] mg/dL vs. 0.93 [0.77, 1.17] mg/dL, p < 0.01), lower hemoglobin (12.5 ± 2.1 g/dL vs. 13.3 ± 2.1 g/dL, p < 0.01), higher BNP (461 [206, 730] pg/ml vs. 453 [242, 497] pg/ml, p < 0.01) and lower lymphocyte count (1104 ± 348 count/ml vs. 2189 ± 575 count/ml, p < 0.01) than cluster 2. Echocardiographic and right-side catheterization data were not different between the groups. The composite of death or HF hospitalization occurred in 23% of cluster 1 and in 9.3% of cluster 2 (HR 2.75, 95% CI 1.35-5.60, p < 0.01, Figure 2).

Conclusion: Heatmap using clinical variables showed heterogeneity of HF. Hierarchical cluster analysis revealed high-risk HF patients with impaired renal function and malnutrition.

Related Topics

    loading  Loading Related Articles