Bench test of an O ₂ /CO ₂ sensor based MBW system using a lung model

Lung volume assessment using tracer gas multiple breath washout (MBW) systems needs validation under realistic conditions. We used a lung model allowing for different tidal volumes (TVs), functional residual capacities (FRCs), and respiratory rates (RRs) under BTPS conditions. We tested a nitrogen (N 2 ) MBW prototype (Exhalyzer D; Eco Medics) based on an ultrasonic flowmeter, a main-stream CO 2 , and a side-stream O 2 sensor. Linearity of O 2 and CO 2 sensors was assessed after two point calibration using a mass spectrometer (AMIS 2000; Innovision). For FRC measurements, a double chamber plexiglas lung model was filled with water, heated, and mechanically ventilated at various ranges of FRCs (900 to 4000 mL), TVs (250 to 850 mL), and RRs (30 to 15/min). N 2 MBW tests (n = 71) using 100% O 2 were done on three days. Using custom designed software (TestPoint) we synchronized gas to flow signals to preset and manually optimized settings, and calculated FRC as cumulative expired N 2 volume divided by the difference of MBW start minus end N 2 concentration. O 2 and CO 2 sensors were linear (linear regression r 2 was 0.99 for both). Using preset synchronization settings, mean difference of lung model minus measured FRCs was -7.4 mL (-0.04% of mean FRC), limits of agreement ranged from 160.4 mL to -175.2 mL (5.8% to -5.9% of mean FRC). After optimized synchronization, mean difference of FRCs was -7.7 mL (-0.25% of mean FRC), limits of agreement ranged from 50.8 mL to -66.2 mL (1.8% to -2.3% of mean FRC). This lung model seems suitable for validation of MBW systems. The new N 2 MBW system precisely measures in vitro FRCs under realistic conditions. Careful synchronization of signals is crucial for accurate FRC measurements.