Objectives: To assess the accuracy of the expired tidal volumes (VT(E)) displayed by one of the most frequently used ventilators that measures exhaled volume at the expiratory valve.
Design: Prospective study.
Setting: The intensive care units of a pediatric tertiary referral center in London, UK.
Patients: A total of 56 intubated children aged between 3 wks and 16.6 yrs who were clinically stable and ventilated with a Servo 300 ventilator.
Interventions: The CO2SMO Plus respiratory monitor, which measures flow at the airway opening, was validated using calibrated syringes and appropriate tracheal tubes and connections. Simultaneous in vivo recordings of VT(E) from the Servo 300 and CO2SMO Plus were compared before (displayed Servo VT(E)) and after (effective Servo VT(E)) compensating for ventilator circuit compliance.
Measurements and main results: The in vitro accuracy of the CO2SMO Plus was within +/-5% over a wide range of volumes and measurement conditions. The displayed Servo 300 VT(E) overestimated the true VT(E) by between 2% and 91%. The magnitude of error varied within and between children, according to pressure change (peak inspiratory pressure minus positive end-expiratory pressure), VT(E), and circuit size. Mean (sd) error was 32% (20%) in 40 children with displayed Servo VT(E) of <160 mL and 18% (6%) in 16 subjects with displayed Servo VT(E) of >/=160 mL. After correcting for gas compression, effective VT(E) from the Servo 300 underestimated the true VT(E) by up to 64% in the smallest infants but continued to overestimate by as much as 29% in older children.
Conclusions: The accuracy of tidal volume values is crucially dependent on the site of measurement. Unless measured at the airway opening, displayed values are an inconsistent and misleading indicator of the true volumes delivered.