Juntendo University: Physical Function Emerges as Crucial Survival Predictor in Elderly Heart Failure Patients

Japan's Growing Heart Failure Challenge Among the Elderly

Japan faces a mounting heart failure (HF) crisis as its population ages rapidly. With over 30 percent of citizens aged 65 or older by 2025, HF has become the leading cause of hospitalization for those over 65, affecting an estimated 1.28 million patients in 2020 and projected to rise further. 52 57 Elderly patients often experience poor prognosis post-hospitalization, with 1-year mortality rates around 16-30 percent. Traditional risk models focusing on cardiac biomarkers frequently overlook non-cardiac factors like physical function, frailty, and nutrition, which are modifiable and critical in Japan's super-aging society.

This context underscores the significance of recent research from Juntendo University Graduate School of Medicine and Faculty of Health Science, highlighting how assessing physical function at discharge can transform survival predictions and care strategies for elderly HF patients.

Breakthrough from Juntendo: Machine Learning Model Pinpoints Physical Function's Role

Led by Professor Tetsuya Takahashi and Assistant Professor Kanji Yamada, a team at Juntendo University developed a novel machine learning model using data from the J-Proof HF registry—a nationwide prospective cohort of 9,700 patients aged 65 or older hospitalized for HF across 96 Japanese institutions from December 2020 to March 2022. 63 The study, published February 3, 2026, in The Lancet Regional Health – Western Pacific, employed eXtreme Gradient Boosting (XGBoost) algorithms to predict 1-year all-cause mortality. 62

Key innovation: Integrating direct physical function measures like the Barthel Index (BI, assessing activities of daily living), Short Physical Performance Battery (SPPB, evaluating balance, gait, and strength), gait speed, and handgrip strength. These outperformed traditional cardiac variables in prognostic importance, with seven of the top 20 predictors being function-related. The full XGBoost model achieved an area under the curve (AUC) of 0.76, stratifying patients into low (2.5% mortality), intermediate (11.5%), and high-risk (35.5%) groups.

Unpacking the J-Proof HF Registry and Study Design

The J-Proof HF (Japanese Physical Frailty in Patients with Heart Failure) registry captures comprehensive data on elderly HF patients prescribed rehabilitation, excluding those bedridden pre-admission or with acute coronary events. Of 9,700 participants (median age 83), 16.5% died within one year post-discharge. Predictors spanned 77 variables: demographics, labs (e.g., BNP, albumin), echocardiography, and functional tests at discharge.

XGBoost's leave-one-site-out validation ensured robustness across institutions. The parsimonious Top-20 model retained accuracy (AUC 0.76) while simplifying to key variables, emphasizing physical function's dominance via SHAP (SHapley Additive exPlanations) analysis. As Dr. Yamada noted, "Performance-based assessments like BI and SPPB offer greater reproducibility and directly capture functional limitations." 63

Physical Function Metrics: Grip Strength, Gait Speed, and Beyond

Physical function encompasses muscle strength (handgrip), mobility (gait speed, SPPB), and daily activities (BI). In elderly HF patients, sarcopenia—age-related muscle loss—affects up to 50 percent, exacerbating frailty. Juntendo's model revealed these as top mortality predictors, rivaling ejection fraction or BNP levels.

• Grip Strength: Simple handheld dynamometer test; low values (<26kg men, <16kg women) signal 2-3x higher risk.
• Gait Speed: Timed 4-6m walk; <0.8m/s indicates frailty.
• SPPB: Composite score (0-12); scores <8 predict poor outcomes.
• BI: 0-100 scale; discharge scores <85 correlate with rehospitalization.

These low-cost, bedside tests enable rapid risk stratification, vital in resource-strapped Japanese hospitals. 62

Superior Performance Over Traditional Models

Existing scores like AHEAD (AUC 0.60) and BIOSTAT-CHF compact (AUC 0.61) underperform in East Asian elderly, ignoring non-cardiac factors. Juntendo's Top-20 XGBoost improved net reclassification index (NRI 21-24%) and decision curve analysis (DCA), offering better clinical utility. In high-risk groups, it identified 35.5% mortality vs. 20-30% by benchmarks.

This Japan-specific model addresses demographic nuances: higher frailty prevalence due to longevity and multimorbidity.Read the full Lancet study.

Photo by James Pere on Unsplash

Clinical Implications: From Prediction to Prevention

The model empowers physicians to prioritize rehab for high-risk patients, potentially reducing mortality via targeted interventions. Dr. Yamada emphasized: "Physical function at discharge rivals traditional factors; modifiable through rehabilitation." In Japan, where only 33% of HF patients receive inpatient rehab, scaling cardiac rehabilitation could cut rehospitalizations by 10-20%. 63

Stakeholder views: Japanese Society of Cardiovascular Physical Therapy (registry sponsor) advocates routine function screening. Policymakers eye integration into national HF guidelines (JCS/JHFS 2025).Explore research assistant roles in cardiac rehab.

Juntendo's Research Ecosystem: GLIS and CFS Advances

Juntendo leads in geriatric cardiology. The GLIS (Global Leadership Initiative on Sarcopenia) model, validated in 891 FRAGILE-HF patients, outperforms AWGS2019; sarcopenic patients faced 3.4x mortality risk. 61 Similarly, Clinical Frailty Scale (CFS) predicts 2-year mortality (HR 1.42 per point) in 3,905 JROADHF-NEXT patients. 60

Pre-admission exercise habits link to better post-discharge function, per another Juntendo study. These converge: frailty screening via CFS/GLIS, ML prognosis, rehab emphasis.

Sarcopenia, Frailty, and Multimorbidity in Japanese Elderly HF

Sarcopenia prevalence: 40-50% in HF elderly. Overlapping cardiovascular-kidney-metabolic (CKM) conditions worsen function by 20-30%. Juntendo's CKM-HF analysis showed compounded physical decline. Cultural factors: Japan's diet, activity levels influence baseline function, but urbanization reduces exercise.

• Benefits of intervention: Rehab boosts SPPB by 2-3 points, cutting mortality 15%.
• Risks ignored: Rehospitalization drops function 10%.

Solutions: Multidisciplinary teams (cardiologists, PTs, nutritionists). For med students, Juntendo offers specialized training.Discover university jobs in Japan.

Rehabilitation Strategies and Actionable Insights

Step-by-step rehab: 1) Assess at admission/discharge; 2) Tailor resistance (grip training), aerobic (walking); 3) Monitor weekly SPPB; 4) Outpatient follow-up. Examples: Juntendo trials show pre-hab exercise halves mortality risk. Community programs like Japan's long-term care insurance support home rehab.

Real-world case: 85-year-old Tokyo patient improved BI from 60 to 90 post-rehab, avoiding rehospitalization. Future: App-based XGBoost tool for risk input.

Future Outlook: Global Impact from Japanese Innovation

Juntendo plans multicenter validation beyond Japan. Aligns with WHO aging strategies. Implications for higher ed: Boosts demand for PT, cardiology researchers. AcademicJobs.com lists openings at Juntendo and peers.Browse higher ed jobs in cardiac research University jobs worldwide.

Balanced view: While promising, ML needs diverse data; ongoing J-Proof expansions address this. Positive: Empowers proactive care in aging nations.

Photo by Damon Lam on Unsplash

Careers in Geriatric Cardiology: Opportunities at Juntendo and Beyond

Juntendo's Sport and Health Science Graduate School trains next-gen researchers. Roles: Postdocs in HF registries, faculty in rehab sciences. Japan invests ¥557 billion in higher ed 2026, prioritizing health sciences.Career advice for lecturers Rate professors.

Actionable: Pursue certifications in geriatric assessment; network via Japanese Circulation Society.