Abstract 040: Bedscales: Non-Contact Home Health Monitor For Longitudinal Characterization Of Periodic Breathing In Chronic Heart Failure

Abstract

Periodic breathing (PB), including its subtype Cheyne-Stokes respiration, is a pattern of sleep-disordered breathing marked by cyclic modulations in tidal volume and respiratory rate. PB is commonly identified in patients with chronic heart failure (CHF), and portends a worse prognosis. Due to the impracticality of existing monitoring methods, no longitudinal data exist on whether individual patients may experience prognostically significant fluctuations in PB. Here, we present a novel tool enabling long-term in-home investigation of PB. By passively capturing respirations during sleep via mechanical sensors placed under the legs of a bed, the need for patient compliance is circumvented. PB events were identified by computing an amplitude modulation index, and features including respiratory rate, cycle length, and hyperpnea duration were extracted. To demonstrate viability for longitudinal studies, devices were installed in the homes of 25 CHF patients to continuously record respirations, 9 of whom were found to exhibit PB. In an illustrative case, ~15,000 cycles of PB were captured over 3 months in a patient discharged after hospitalization for heart failure with reduced ejection fraction. Bedscales documented the patient’s worsening tachypnea accompanied by changes in PB parameters (shortened cycle and hyperpnea durations). A chest X-ray showed evidence of pneumonia and antibiotics were prescribed, after which tachypnea improved and trends of PB parameters reversed. This case highlights the potential utility of Bedscales in detecting subtle deviations from an individual’s baseline respiratory signatures; future work is needed to assess the predictive power of these parameters. Taken together, these results demonstrate the feasibility of Bedscales as a low-cost, scalable tool for longitudinal population-level studies of PB in relation to chronic diseases such as CHF. Future research could examine whether such data might predict exacerbations and direct early intervention to prevent rehospitalization.