Objectives: There is a continuing risk of production of toxic levels of carbon monoxide (CO) as a result of interaction of volatile anesthetics and desiccated strong base carbon dioxide absorbents like soda lime. The aim of this study is to establish the reliability of detection of CO levels by an electrochemical carbon monoxide sensor compared to gas chromatography. Methods: Completely desiccated sodalime was conducted through a circle anesthesia system connected to an artificial lung. For different rates of CO production, a low flow anesthesia with a oxygen/nitrous oxide mixture was maintained using five volatile anesthetics. For quantification of CO production, a portable gas chromatograph (GC) was connected to this setup, as well as a Bedfont EC40 electrochemical carbon monoxide sensor (ES) with a claimed reliable sensitivity of 0-200 parts per million (ppm) and a maximum detection range of more than 5500 ppm. To assess the agreement between the GC and ES measurements the intra class correlation coefficient (ICC) and the 95% limits of agreement were calculated. Bland and Altman scatterplots were made to visualize the difference between measurements. Results: For concentrations up to 200 ppm, no significant differences between the GC and ES mean CO measurements were found in the halothane experiments. However CO was not accurately measured at every moment during these experiments by the ES. For concentrations above 200 ppm the results of the two instruments differed significantly. The ES malfunctioned when exposed to sevoflurane and desiccated sodalime. Conclusions: From these data we conclude that the ES can only be used as an indicator of CO production. When this sensor is used with sevoflurane and desiccated sodalime it is not capable of normal operation. The use of a strong base free carbon dioxide absorbent is therefore recommended.