Pulse-distribution analysis of 1-minute radial pulse signals discriminated prefrailty in older adults with an AUC of 0.70, outperforming conventional machine learning models (AUC ≤0.63).
Observational (n=112)
Does spectral pulse analysis of radial pulse signals accurately identify prefrailty in community-dwelling older adults?
A 1-minute noninvasive radial pulse spectral analysis can identify prefrailty in older adults with 70% accuracy, offering a novel biomarker linking vascular dysfunction to geriatric decline.
Effect estimate: AUC 0.70
Absolute Event Rate: 0.7% vs 0.63%
The identification of prefrailty, a reversible state of vulnerability in older adults, remains a clinical challenge. Arterial stiffness—a hallmark of vascular aging—has been implicated in frailty, suggesting shared pathophysiology. This link between prefrailty and early vascular dysfunction provides a rationale for exploring arterial pulse signals. This study investigated whether spectral pulse analysis could yield objective biomarkers for prefrailty. We analyzed 1-minute radial pulse signals from 112 community-dwelling older adults (49 prefrail, 63 robust). Forty harmonic indices were extracted and evaluated using two classification approaches: conventional machine learning (ML) algorithms and a purpose-developed pulse-distribution analysis (PDA). Harmonic analysis revealed significant alterations in the pulse waveform of prefrail individuals, characterized by reduced amplitude proportions and increased variability coefficients. While conventional ML models demonstrated limited discrimination (AUC ≤0.63), the PDA method achieved an AUC of 0.70. This performance was consistent across subgroups defined by hypertension, prediabetes, or sex (AUC range: 0.68–0.74), with minimal confounding effects from age, blood pressure, or glucose levels (R 2 < 0.03). These findings reveal a novel pulse-derived harmonic signature as a biomarker for prefrailty, indicating that subtle arterial functional alterations detectable via spectral analysis are associated with early physical vulnerability and may help bridge vascular pathology with geriatric decline. The discriminative performance (AUC = 0.70) is competitive with existing tools, while the 1-minute, noninvasive protocol establishes a favorable balance between accuracy and clinical feasibility. Future validation could support its use for early identification and triage of high-risk older adults in clinical practice. • Pulse analysis detects prefrailty with 70% accuracy. • Vascular changes visible in early frailty stages • Minimal interference from age, BP, or glucose levels • Noninvasive 1-minute screening for early intervention • Classification performance comparable to existing techniques
Wu et al. (Wed,) conducted a observational in Prefrailty (n=112). Pulse-distribution analysis (PDA) of radial pulse signals vs. Conventional machine learning (ML) algorithms was evaluated on Discrimination of prefrailty (Area Under the Curve) (AUC 0.70). Pulse-distribution analysis of 1-minute radial pulse signals discriminated prefrailty in older adults with an AUC of 0.70, outperforming conventional machine learning models (AUC ≤0.63).