SINGLE-BREATH CO2 ANALYSIS - DESCRIPTION AND VALIDATION OF A METHOD

Objectives: To evaluate the performance of a newly developed single br eath CO2 analysis station in measuring the airway deadspace in a lung model (study 1), and then to quantify the bias and precision of the ph ysiologic deadspace measurement in a surfactant-depleted animal model (study 2). Design: A prospective bench validation of a new technique o f airway deadspace measurement using a criterion standard (study 1); a prospective, animal cohort study comparing a new technique of physiol ogic deadspace measurement with a reference method (Bohr-Enghoff metho d) (study 2). Setting: A bench laboratory and animal laboratory in a u niversity-affiliated medical center. Subjects: A lung model (study 1), and adult sheep with induced surfactant deficiency (saline ravage) (s tudy 2). Methods: The single breath CO2 analysis station consists of a mainstream capnometer, a variable orifice pneumotachometer, a signal processor, and computer software with capability for both on and off l ine data analysis. Study 1: We evaluated the accuracy of the airway de adspace calculation using a plexiglass lung model. The capnometer and pneumotachometer were placed at the ventilator Y-piece with polyvinyl chloride tubing added to simulate increased airway deadspace. Segments of tubing were sequentially removed during each testing session to si mulate decreasing deadspace. The calculated airway deadspace was deriv ed from the single breath CO2 plot and compared with the actual tubing volume using least-squares linear regression and paired t-tests. Stud y 2: The accuracy of the physiologic deadspace measurement was examine d in a saline-lavaged animal model by comparing the physiologic deadsp ace calculated from the single breath CO2 analysis station with values obtained using the Enghoff modification of the Bohr equation: deadspa ce/tidal volume ratio = (PaCO2 - mixed expired PCO2)/PaCO2. Measuremen ts and Main Results: Study 1: Thirty six measurements of calculated ai rway deadspace were made and compared with actual circuit deadspace du ring four different testing conditions. Measured airway deadspace corr elated significantly with actual circuit deadspace (r(2) = .99). The p roportional error of the method was -0.8% with a 95% confidence interv al from -3.6% to 1.9%. Study 2: A total of 27 pairs of measurements in four different animals were available for analysis. The derived physi ologic deadspace/tidal volume ratio significantly correlated with the value obtained using the Bohr Enghoff method (r(2) = .84). The bias an d precision of our physiologic deadspace calculation were .02 and .02, respectively, and the mean percent difference for the physiologic dea dspace calculated from the single breath CO2 analysis station was 2.4% . Conclusions: Our initial experience with the single breath CO2 analy sis station indicates that this device can reliably provide on line ev aluation of the single-breath CO2 waveform. In particular, estimation of the airway and physiologic deadspace under a variety of testing con ditions was consistently within 5% of actual values. We feel that with further application and refinement of the technique, single breath CO 2 analysis may provide a noninvasive, on-line monitor of changes in pu lmonary blood flow.