Can modern infrared analyzers replace gas chromatography to measure anesthetic vapor concentrations?

Abstract

Abstract Background Gas chromatography (GC) has often been considered the most accurate method to measure the concentration of inhaled anesthetic vapors. However, infrared (IR) gas analysis has become the clinically preferred monitoring technique because it provides continuous data, is less expensive and more practical, and is readily available. We examined the accuracy of a modern IR analyzer (M-CAiOV compact gas IR analyzer (General Electric, Helsinki, Finland) by comparing its performance with GC. Methods To examine linearity, we analyzed 3 different concentrations of 3 different agents in O2: 0.3, 0.7, and 1.2% isoflurane; 0.5, 1, and 2% sevoflurane; and 1, 3, and 6% desflurane. To examine the effect of carrier gas composition, we prepared mixtures of 1% isoflurane, 1 or 2% sevoflurane, or 6% desflurane in 100% O2 (= O2 group); 30%O2+ 70%N2O (= N2O group), 28%O2 + 66%N2O + 5%CO2 (= CO2 group), or air. To examine consistency between analyzers, four different M-CAiOV analyzers were tested. Results The IR analyzer response in O2 is linear over the concentration range studied: IR isoflurane % = -0.0256 + (1.006 * GC %), R = 0.998; IR sevoflurane % = -0.008 + (0.946 * GC %), R = 0.993; and IR desflurane % = 0.256 + (0.919 * GC %), R = 0.998. The deviation from GC calculated as (100*(IR-GC)/GC), in %) ranged from -11 to 11% for the medium and higher concentrations, and from -20 to +20% for the lowest concentrations. No carrier gas effect could be detected. Individual modules differed in their accuracy (p = 0.004), with differences between analyzers mounting up to 12% of the medium and highest concentrations and up to 25% of the lowest agent concentrations. Conclusion M-CAiOV compact gas IR analyzers are well compensated for carrier gas cross-sensitivity and are linear over the range of concentrations studied. IR and GC cannot be used interchangeably, because the deviations between GC and IR mount up to ± 20%, and because individual analyzers differ unpredictably in their performance.