LB065 - CALF CIRCUMFERENCE CUTOFF VALUES FOR MORTALITY PREDICTION DEPEND ON PHYSICAL FUNCTION IN HOSPITALIZED PATIENTS WITH HEART FAILURE

LB065

CALF CIRCUMFERENCE CUTOFF VALUES FOR MORTALITY PREDICTION DEPEND ON PHYSICAL FUNCTION IN HOSPITALIZED PATIENTS WITH HEART FAILURE

Y. Iida^1,2,*, M. Kato², Y. Kono², K. Kamiya², M. Saitoh², K. Sakurada², M. Taya², K. Iwata², Y. Funami², T. Morisawa², E. Nakatani², T. Takahashi²

¹     Faculty of Health and Medical Sciences, Aichi Shukutoku University, Nagakute, ²Committee of the J-Proof HF Registry, Japanese Society of Cardiovascular Physical Therapy, Tokyo, Japan

Rationale: Prognostic stratification tools are essential for clinical nutrition and rehabilitation, especially in older or frail hospitalized patients. The Short Physical Performance Battery (SPPB) and calf circumference (CC) are surrogate indicators for sarcopenia and functional status. However, in patients with heart failure, the accuracy of CC is often compromised by peripheral edema, and optimal cutoff values for mortality prediction have not been established. This study aimed to develop and validate a simple risk stratification model for 1-year mortality using decision tree analysis based on SPPB and CC in hospitalized older patients with heart failure.

Methods: This secondary analysis utilized data from the Japanese PT multi-center Registry of Older Frail patients with Heart Failure (J-Proof HF), a prospective multicenter cohort study of patients aged ≥65 years who received physical therapy during hospitalization between July 2020 and March 2022. A total of 7,362 patients were included. Participants were chronologically divided to avoid cohort contamination and assess external validity into a training cohort (July 2020–July 2021, n=4,417) and a validation cohort (August 2021–March 2022, n=2,945). In the training cohort, logistic regression identified SPPB and CC as factors associated with 1-year mortality. Cutoff values for these variables were determined via decision tree analysis to stratify patients into three risk groups: low, intermediate, and high. The resulting classification rule was applied to the validation cohort, where 1-year survival was analyzed using Kaplan–Meier curves and confirmed through multivariable Cox proportional hazards regression.

Results: Decision tree analysis identified SPPB and CC thresholds that stratified patients into six terminal nodes with 1-year mortality rates ranging from 6.1% to 34.9%. These nodes were reorganized into three clinically meaningful groups: low risk (SPPB ≥10 and CC ≥25.9), intermediate risk (SPPB ≥10 and CC <25.9, or SPPB 7–9 and CC ≥25.3), and high risk (SPPB 7–9 and CC <25.3, or SPPB <7). In the validation cohort, 1-year survival was significantly different across the three groups (Log-rank p < 0.001), and all pairwise comparisons remained significant after Bonferroni correction. Multivariable Cox regression showed that, compared to the low-risk group, the intermediate- and high-risk groups had significantly higher hazards of 1-year mortality (HR 1.75 and 3.39, respectively; both p < 0.001), independent of significant clinical covariates.

Conclusion: A straightforward combination of SPPB and CC enables robust and practical stratification of 1-year mortality risk in hospitalized older patients with heart failure. This model offers an accessible and reliable method for risk stratification, contributing to personalized care planning in real-world hospital settings.

Disclosure of Interest: None declared