Combined RDW-CV and NT-proBNP for all-cause mortality prognostication in frail older cancer patients hospitalized with anticancer treatment-related cardiac dysfunction: a retrospective cohort study
Background and aims Frail older adults hospitalized with established anticancer treatment-related cardiac dysfunction (CTRCD)/heart failure represent a selected, high-risk cardio-oncology population. Red cell distribution width-coefficient of variation (RDW-CV) and N-terminal pro-B-type natriuretic peptide (NT-proBNP) are routinely available biomarkers, but their individual and joint associations with subsequent mortality in this setting remain incompletely characterized. We evaluated their associations with all-cause mortality and explored multiplicative interaction and prognostic discrimination. Methods This retrospective cohort study enrolled 399 patients aged ≥65 years who were hospitalized with established anticancer treatment-related heart failure/CTRCD and confirmed frailty (Fried phenotype score ≥3) between January 2018 and December 2023, with follow-up through January 2025 (median 28 months). For the reported four-group survival analyses, patients were stratified using the cohort medians of RDW-CV (14.8%) and NT-proBNP (4,562 pg/mL). Kaplan–Meier and Cox regression analyses evaluated associations with all-cause mortality. A multiplicative interaction term [RDW-CV × ln(NT-proBNP)] and restricted cubic spline analyses were exploratory. Conventional ROC analyses were also exploratory and were performed in the same derivation cohort. Results Among 399 patients (mean age 78.6 ± 7.2 years; 56.9% male), 168 all-cause deaths (42.1%) occurred during follow-up. Each 1% increment in RDW-CV was associated with an 11% higher mortality hazard after full multivariable adjustment (HR = 1.11, 95%CI: 1.04–1.18; P = 0.002), and high-versus-low NT-proBNP was associated with a 1.68-fold higher mortality hazard (HR = 1.68, 95%CI: 1.19–2.37; P = 0.003). Median-defined combined biomarker groups showed a mortality gradient from 19.4% in the dual-low group to 60.7% in the dual-high group (log-rank χ 2 = 44.86, P < 0.001). The multiplicative interaction term was statistically significant (HR = 1.04, 95%CI: 1.01–1.08; P = 0.032) and should be interpreted as exploratory statistical effect modification rather than evidence of biological synergy. In the same cohort, the combined panel had an apparent AUC of 0.762 (95%CI: 0.713–0.811), compared with 0.672 for RDW-CV alone and 0.718 for NT-proBNP alone. Conclusion In frail older cancer patients already hospitalized with established anticancer treatment-related cardiac dysfunction/heart failure, higher RDW-CV and NT-proBNP were associated with greater subsequent all-cause mortality, and the median-defined dual-high group had the poorest prognosis. The combined panel showed greater apparent discrimination than either biomarker alone in this single derivation cohort. These findings support further evaluation of combined biomarker-based prognostication but require prospective multicenter and model validation studies before ROC-derived thresholds are used as clinical decision or management triggers.
Authors
- Nana Li (ORCID: https://orcid.org/0000-0001-9933-8674)
- Jie Yang (ORCID: https://orcid.org/0000-0002-6397-6655)
- Qin Xiao
Institutions
- Jiangsu Province Hospital (CN)
- First People's Hospital of Chongqing (CN)
Publication Details
- Journal
- Frontiers in Cardiovascular Medicine
- Published
- 2026-11-17
- DOI
- https://doi.org/10.3389/fcvm.2026.1973433
- Primary Topic
- Frailty in Older Adults
- Type
- article
- Field-Weighted Citation Impact
- 0.00