Monitoring respiration using the pressure sensors in a dialysis machine

Objective: Although respiratory problems are common among patients with end-stage renal disease, respiration is not continuously monitored during dialysis. The purpose of the present study is to investigate the feasibility of monitoring respiration using the pressure sensors of the dialysis machine. Approach: Respiration induces variations in the blood pressure that propagates to the extracorporeal circuit of the dialysis machine. However, the magnitude of these variations are very small compared to pressure variations induced by the dialysis machine. We propose a new method, which involves adaptive template subtraction and peak conditioned spectral averaging, to estimate respiration rate from the pressure sensor signals. Using this method, an estimate of the respiration rate is obtained every 5th second provided that the signal quality is sufficient. The method is evaluated for continuous monitoring of respiration rate in nine dialysis treatment sessions. Main results: The median absolute deviation between the estimated respiration rate from the pressure sensor signals and a reference capnography recording was 0.02 Hz (1.3 breaths per min). Significance: Our results suggest that continuous monitoring of respiration using the pressure sensors of the dialysis machine is feasible. The main advantage with such monitoring is that no additional sensors are required which may cause patient discomfort.