Work

Proteomics-based clustering outperforms clinical clustering in identifying heart failure patient groups with distinct outcomes

Year: 2025

Type: preprint

Abstract: Background. Clustering of heart failure (HF) patients typically relies on clinical characteristics, which may not reflect underlying pathophysiology relevant for personalised medicine. We aimed to identify plasma protein profiles of HF patients with reduced ejection fraction (HFrEF). Methods. Using latent class analysis, we derived clusters based on 1) clinical characteristics, and 2) proteomics (SomaScan) from 379 HFrEF patients. Survival analysis assessed associations with major cardiovascular (CV) events (a composite of HF hospitalization, CV death, heart transplantation, and left ventricular assist device implantation), HF hospitalization, CV death, and all-cause mortality. To aid clinical understanding, we identified differentially expressed proteins and explored their druggability. Results. Clustering on clinical characteristics identified three patient clusters that did not differ in disease progression. Proteomics-based clustering identified three clusters associated with disease progression. Cluster 1 included younger patients with fewer comorbidities, whereas cluster 3 consisted of older patients with more atrial fibrillation and renal failure. Cluster 2 had intermediate values for most characteristics; medication use was similar across clusters. Over three years of follow-up, cluster 1 had few events. Compared to this cluster, cluster 2 had increased risk of major CV events (HR 2.31, 95%CI 1.23; 4.36) and HF hospitalization (HR 2.30, 95%CI 1.10; 4.78), but not of death. Cluster 3 had high event rates with HRs of 5.84 (major CV events), 6.50 (HF hospitalization), 8.58 (CV death), and 5.07 (all-cause mortality). Results were externally validated in 511 HFrEF patients. Twelve proteins were differentially expressed, including druggable targets CD2, GDF-15, ABO, IGFBP-1, IGFBP-2, and RNase1. Conclusions. Proteomics-based HFrEF clustering identified three clusters associated with distinct disease outcomes, undetected using clinical characteristics.

Source: medRxiv

DOI: 10.1101/2025.08.19.25333956