JMIR Publications

ABSTRACT

Background:

Excessive fluid accumulation in the lungs leading to pulmonary congestion (PC) is a major contributor to heart failure (HF) deterioration and often necessitates emergency hospitalization. Early detection of PC followed by prompt decongestive therapy could serve as a key strategy to prevent acute decompensation and reduce hospital admissions. Home-based monitoring of PC-related respiratory abnormalities through wearable devices may provide a promising approach for early detection of HF exacerbation. However, the feasibility of using wearable technology to detect specific respiratory biomarkers of PC remains unclear.

Objective:

We aimed to evaluate the feasibility of using wearable devices to monitor respiratory data in hospitalized patients with HF, and to determine whether respiratory signals collected by these devices can distinguish HF patients with PC.

Methods:

This was a single-center, observational, exploratory study that enrolled hospitalized HF patients without severe lung diseases or indications for intensive care/mechanical ventilation, all of whom wore the designated wearable device for ≥24 hours starting within 24 hours after admission, though only nighttime respiratory data were analyzed. Breathing patterns were quantified across three dimensions: respiratory cycle, amplitude, and multiscale entropy (MSE). Within the first 24 hours post-admission, patients underwent comprehensive evaluations including vital signs, laboratory tests, echocardiography, and PC assessment via 28-zone lung ultrasound (LUS), with PC defined as >5 B-lines.

Results:

The study enrolled 62 HF patients between May 2021 and November 2022, including 44 with PC. Compared to non-PC patients, those with PC demonstrated significantly prolonged mean expiratory time (TE_mean: 2.17 ± 0.43 s vs. 1.94 ± 0.34 s, P = 0.033), elevated expiratory time ratio (TE_ratio_mean: 59.12 ± 2.94% vs. 56.40 ± 3.36%, P = 0.006), and higher MSE values in respiratory amplitude (RA_area_1_5: 4.20 ± 1.52 vs. 2.93 ± 1.03, P < 0.001; RA_area_6_20: 8.86 ± 3.14 vs. 12.28 ± 4.84, P = 0.002). Logistic regression identified TE_ratio_mean, RA_area_1_5, and RA_area_6_20 as significant predictors of PC (P < 0.05). After adjusting for clinical confounders, both RA_area_1_5 and RA_area_6_20 remained independently associated with PC. Receiver operating characteristic (ROC) analysis revealed that RA_area_1_5 had the largest area under the curve (AUC) of 0.754 (95% CI: 0.632–0.876, P = 0.002), with 65.9% sensitivity and 73.3% specificity. For multivariate logistic regression model constructed using combined parameters of DBP, logNT-proBNP, NYHA IV, TE_ratio_mean, RA_area_1_5, and RA_area_6_20, the AUC was 0.910 (95% CI: 0.837～0.984, P<0.001), sensitivity of 77.3%, and specificity of 88.9%.

Conclusions:

In this small-scale exploratory study, wearable-based MSE analysis successfully distinguished between hospitalized HF patients with and without PC, as evidenced by prolonged expiratory phases and increased respiratory amplitude complexity in the PC group.