The Saturated Vapor Pressure of Desflurane at Various Temperatures

The low boiling point of desflurane compared with other inhaled anesthetics in current use requires specialized, expensive delivery systems. Older, cheaper delivery systems require the anesthetic to be in liquid form. A liquid injection system could be cooled or pressurized to achieve this, but the pressure required is unknown. A draw-over vaporizer could be used if desflurane were cooled enough to reduce its saturated vapor pressure (SVP) close to that of halothane at room temperature, but the requisite temperature is also unknown. We have therefore measured the SVP of desflurane over an extended temperature range. Methods An isoteniscope [1] was constructed comprising a glass flask (nominal volume 25 mL) connected to a preformed glass tube. A reservoir of desflurane (Suprane Trademark; Ohmeda, Hatfield, UK) was injected into the glass flask using a flexible, narrow nylon tube. Air was eliminated by boiling the desflurane in the reservoir to flush the saturated vapor (SV) chamber and preformed glass tube. Cooling the dependent loop of the glass tube allowed desflurane to condense in it, trapping pure vapor in the SV chamber. The open end of the glass tube was then sealed to a length of plastic tubing. The plastic tubing was used as a flexible mercury manometer with the distal end open to the atmosphere Figure 1. The isoteniscope was completely submerged in a stirred water bath at 20 degrees C (confirmed with a National Physical Laboratories certified mercury thermometer), with the flexible connection to the mercury manometer trailing out of the bath. The desflurane menisci in the isoteniscope manometer were 5 cm apart and observed within the water bath through an angled mirror. As the temperature of the isoteniscope stabilized, the arms of the mercury manometer were adjusted constantly to keep the desflurane menisci level. When the system had equilibrated and the desflurane menisci were level, the pressure within the SV chamber was equal to that in the space connecting the desflurane and the mercury manometers and was measured by the mercury manometer. At least two measurements were made at an interval of 3 min. A mercury barometer was used to measure atmospheric pressure so the absolute pressure within the isoteniscope could be calculated.Figure 1: An isoteniscope comprising a desflurane reservoir, a preformed glass tube, and a flexible mercury manometer. The preformed glass tube contains desflurane vapor and liquid.The mercury manometer was disconnected, the desflurane within the reservoir boiled, and the above process repeated until there was no significant further decrease in SV chamber pressure at 20 degrees C. This confirmed that de-airing was complete, and the preparation was then used to measure the SVP of desflurane over an extended temperature range. The isoteniscope was refilled with desflurane and de-aired between each of six experiments. In each experiment, the SVP was measured over a range of at least 10 degrees C at intervals of at most 5 degrees C. A Microsoft Excel v4.0 spreadsheet was used to determine the best least-squares fit to the Antoine equation, with coefficient C fixed at 273.15. A second-degree polynomial was also fitted to the data using the Arcus Pro-Stat v3.27 statistical analysis program. Results The SVP measurements were made over a temperature range of 1.0 to 38.1 degrees C. A total of 115 readings were made. In every experiment, the difference between measurements made at any temperature was less than 4 mm Hg. The measured SVP and best fit polynomial are shown in Figure 2.Figure 2: Measured saturated vapor pressure (open circle). The heavy line marks the best-fit second degree polynomial, with 95% confidence intervals shown by thinner lines. The best-fit Antoine Equation isindistinguishable from the polynomial at this plot resolution.The best fit equations were as follows: Equation 1 where T is temperature (degrees C) and SVP is saturated vapor pressure (mm Hg). Discussion Provided that leaks are excluded, the main errors in this technique are from the measurement of pressure within the isoteniscope and the presence of air in the SV chamber of the isoteniscope. Leaks were easily identified by bubbling within the water bath and continued boiling of desflurane within the reservoir. The measurement errors were small. The relatively low density of desflurane means that any inaccuracy due to misalignment of the desflurane manometer levels was certainly less than 1 mm Hg (9 mm desflurane). Air within the SV chamber would cause systematic overreading of the SVP: this was avoided by repeatedly boiling the desflurane until there was no further reduction in SVP at 20 degrees C. Inspection of the six individual curves derived from the different preparations did not reveal a systematic overreading in any particular experiment. Andrews et al. [2] obtained a least-squares fit of Antoine constants to values of SVP supplied by Anaquest. These were derived from a series of vapor pressure measurements made using a gas chromatograph to analyze the composition of the headspace existing above desflurane equilibrated at various temperatures (H. G. Brittain, Director of Pharmaceutical Development, Ohmeda, personal communication). Their resulting curve differs significantly from ours, lying within our 95% confidence interval only from 21.6 degrees C to 23.2 degrees C. All our results are closer to atmospheric pressure than their calculations predict: at 0 degrees C their extrapolated estimate is 261 mm Hg, but we have measured 331 +/- 25 mm Hg, and at 38 degrees C the extrapolated estimate is 1449 mm Hg, but we have measured 1254 +/- 13 mm Hg. Such an effect could be caused by a leak in the isoteniscope or manometer or by a failure to wait long enough after changing the water bath temperature before taking pressure readings. We do not think that leaks are a likely explanation because they were easily identified when they developed; furthermore, our results are consistent across experiments between which the system was dismantled, and it is improbable that the same size of leak was created every time. Examination of our duplicate readings at each temperature does not reveal a trend of increasing pressure. We are therefore unable to identify sources of error in our work that could account for the disparity between our results and the calculated values. We have injected desflurane from an ice-cooled syringe directly into a closed breathing system [3], but it would be more elegant to maintain the desflurane in liquid form by pressurizing the delivery system. We conclude from our data that a pressure of at least 2 atmospheres is needed to be sure that heat from neighboring electrical equipment will not boil the desflurane, which would cause dangerously large quantities of the anesthetic to be discharged into the breathing system. At the other end of the temperature scale, we found that the SVP of desflurane at 0 degrees C is slightly greater than that of halothane at room temperature. We therefore expected desflurane to be usable in a draw-over vaporizer cooled in iced water and have since administered it in this way satisfactorily from a Goldman vaporizer positioned in circle. Although we stress that the safety of this technique depends crucially upon the vaporizer being kept ice cold, it does make a novel demonstration for our trainees of the application of simple physical principles to the practice of anesthesia.