A wrist-worn, machine-learning device analyzing venous waveforms — the NIVAHF device — produced a non-invasive pulmonary capillary wedge pressure estimate (NIVA Score) that fell significantly after intravenous diuresis in both 11 hospitalized heart failure patients (13 paired measurements, mean fluid removal -2.1 ± 1.0 L; P = 0.04) and a controlled porcine model (5 animals, 24 paired measurements; P < 0.01), with directional concordance to urine output in every animal. Exploratorily, discharge NIVA Score predicted 30-day readmission with an AUC of 0.85.
Heart failure readmission rates remain stubbornly near 20–25% at 30 days globally, largely because clinicians lack reliable outpatient tools to detect residual congestion before it triggers decompensation. Invasive hemodynamic monitoring via pulmonary artery catheters remains the gold standard but is impractical for serial use. Non-invasive surrogates — BNP, point-of-care lung ultrasound, impedance cardiography — each carry significant limitations in sensitivity or feasibility. A validated wrist wearable closing this gap would be clinically transformative.
However, important caution is warranted. The human cohort involves only 11 patients, and the 0.85 AUC readmission finding carries a confidence interval spanning 0.575–1.00, reflecting profound uncertainty. The animal model, while mechanistically controlled, uses crystalloid-induced volume overload — an imperfect proxy for chronic heart failure pathophysiology. As a preprint not yet peer-reviewed, these results may change substantially after independent scrutiny. This is best characterized as a promising feasibility signal, not practice-changing evidence, pending the authors' planned adequately powered prospective trial.
