LABORATORY investigations of anesthetic actions frequently require volatile anesthetics to be dissolved in aqueous solutions, such as 0.9% NaCl or Tyrode's solution. In addition, such experiments are at times performed at temperatures other than 37 [degree sign]C, affecting anesthetic solubility. Thus, many investigators prefer to report actual anesthetic concentrations in solvent rather than volume percent of anesthetic in the carrier gas. Calculating the anesthetic concentration requires knowledge of the partition coefficient for the particular solvent at the particular temperature. However, only limited data exist for partition coefficients in aqueous solutions at temperatures other than 37 [degree sign]C, particularly for the more recently introduced volatile anesthetics desflurane and sevoflurane. Thus, we determined Ostwald partition coefficients for halothane, isoflurane, desflurane, and sevoflurane in two aqueous solutions (0.9% NaCl and Tyrode's solution) at three temperatures (4 [degree sign]C, 22 [degree sign]C, and 37 [degree sign]C). To facilitate calculation of partition coefficients at other temperatures, we also determined the correlation between temperature and anesthetic solubility, which allowed derivation of empirical equations to calculate partition coefficients for each anesthetic at temperatures between 4 [degree sign]C and 37 [degree sign]C.
Experiments on cultured cells and tissues are frequently performed in the presence of protein, which may affect anesthetic solubility in a temperature-dependent manner. To access the magnitude of this effect, we also determined the solubility of isoflurane in Tyrode's solution to which 0.1% fatty acid free bovine serum albumin (BSA; approximately the protein concentration in cell culture media) had been added.
Halothane was from Halocarbon Laboratories (River Edge, NJ); sevoflurane was from Abbott International Ltd. (Abbott Park, IL); isoflurane and desflurane were from Ohmeda Inc. (Liberty Corner, NJ). Hydrogen, helium, and air were obtained from BOC (BOC Gases, Richmond, VA). Fatty acid free BSA was purchased from ICN Pharmaceuticals Inc. (Costa Mesa, CA). All other chemicals were obtained from Sigma (St. Louis, MO).
The method of Regan and Eger was used to determine the Tyrode's-air and 0.9% NaCl-air Ostwald partition coefficient for halothane, isoflurane, sevoflurane, and desflurane at 4 [degree sign]C, 22 [degree sign]C, and 37 [degree sign]C. All materials used were preequilibrated to the temperature of study. Briefly, 50-ml glass syringes were filled with 16 ml of either 0.9% NaCl or Tyrode's solution (in mM: NaCl, 150; KCl, 5; CaCl2, 2; MgSO4, 1; dextrose, 10; HEPES, 10; with or without 0.1% fatty acid free BSA, pH adjusted to 7.4). Thirty-four milliliters of air containing 1 mM anesthetic was then added. While shaken continuously, the mixture was equilibrated for 120 min at 4 [degree sign]C, 22 [degree sign]C, or 37 [degree sign]C. The concentration of anesthetic in the gas phase over each sample was then determined by gas chromatography (Varian Aerograph model 940, Varian Analytical Instruments, Walnut Creek, CA).
After analysis of the gas phase in the syringes, aliquots of 15 ml Tyrode's solution or 0.9% NaCl were transferred anaerobically to 530-ml flasks. A portion of the air had been evacuated from each flask to produce negative pressure, which drew the aliquot into the flask through a needle piercing the Teflon stopper. A one-way stop-cock affixed to the needle was used to seal the flask. Each flask was stored at test temperature (4 [degree sign]C, 22 [degree sign]C, or 37 [degree sign]C) and continuously shaken for the ensuing 120 min, after which the concentration of the anesthetic was determined by gas chromatography.
The partition coefficient (Greek small letter lambda) was calculated using the following Equation 1: where Csis the concentration of anesthetic in the gas phase in the 50-ml syringe; Cfis the anesthetic concentration in the gas phase of the flask; Vfis the volume of the gas phase of the flask; and Vsis the volume of the aliquot of Tyrode's solution or 0.9% NaCl in the flask. Seven partition coefficient determinations were made for each anesthetic at each temperature, and the values were averaged for each temperature and anesthetic. The averaged partition coefficient values were regressed against temperature.
The partition coefficients for halothane, isoflurane, sevoflurane, and desflurane in 0.9% NaCl or Tyrode's solution at 4 [degree sign]C, 22 [degree sign]C, and 37 [degree sign]C are summarized in Table 1. For completeness, this Table alsoincludes other partition coefficients reported in the literature. Calculated partition coefficients for isoflurane in Tyrode's solution in the presence of 0.1% BSA were 3.38 +/− 0.05, 1.61 +/− 0.04, and 0.86 +/− 0.07 at 4 [degree sign]C, 22 [degree sign]C, and 37 [degree sign]C, respectively, representing an increase of approximately 50% over the values obtained in the absence of protein. We plotted our data in semi-logarithmic form (Figure 1) and performed regression analysis. In agreement with previous studies linear regression provided a close fit to the data (r2> 0.95 for each curve, Figure 1). Therefore, an empiric Equation ofthe form
[Greek small letter lambda]= 10 sup (aT+b)
can be used to calculate the Ostwald partition coefficient for any temperature between 4 [degree sign]C and 37 [degree sign]C. In this Equation Tis the temperature in [degree sign]C, and a and b are two parameters that can be found in Table 2.
Although blood-gas partition coefficients at 37 [degree sign]C are most relevant in the clinical setting, in vitro studies frequently require anesthetics to be dissolved in aqueous solutions and at a variety of temperatures. Although many investigators prefer to use anesthetic-specific partition coefficients for aqueous solutions at the temperature of study when dissolving anesthetics in such solutions during in vitro studies, these aqueous coefficients are not widely available for temperatures other than 37 [degree sign]C. This prompted us to perform the present investigation.
Our data were internally consistent when comparing the seven replications for each datapoint and externally consistent when compared with datapoints reported previously by other investigators (Table 1). Partition coefficients in electrolyte solutions were generally 10 - 15% smaller than those reported in H2O, [2,3]which may be a result of differences in osmolarity, as increasing osmolarity has been shown to be related to decreased solubility. [4,14]In agreement with this finding, we found slightly lower coefficients for Tyrode's solution (osmolarity, 341 mOsm/l) than for 0.9% NaCl (osmolarity, 308 mOsm/l).
Early studies reported changes of water-gas and oil-gas anesthetic solubility with changes in temperature. [5,16,18]In addition, if these data were plotted in semi-logarithmic form, a linear regression provided a close fit for each of the anesthetics tested (halothane, fluroxene, cyclopropene, ether, and others). Data from our study of anesthetics introduced more recently into clinical practice similarly fit closely to a linear Equation ifplotted on semi-logarithmic coordinates. Consequently the Equation derivedfrom this study can be used as an empirical method to calculate partition coefficients for the major volatile anesthetics in clinical use when dissolved in Tyrode's solution or 0.9% NaCl at temperatures between 4 [degree sign]C and 37 [degree sign]C. Previous publications in the anesthesia literature used the same empirical equation. Other equations, e.g., the Van't Hoff equation (log([Greek small letter lambda])= B/T + constant) or derivations of the Antoine equation, [3,6-8]yield essentially equivalent results.
Factors other than temperature and osmolarity affect anesthetic solubility as well. Solubility changes significantly if proteins or lipids are added to solutions. In our study 0.1% BSA increased solubility of isoflurane by a factor of [tilde operator] 50% at all temperatures tested. Part of this effect may be a result of saturable binding of anesthetics to hydrophobic protein parts, [4,19]although this cannot explain the changes in solubility seen in this and previous studies [1,2,9,18]; at clinical concentrations, only picomolar to nanomolar amounts of anesthetic bind to proteins, so that other mechanisms have to be involved as well. Therefore, our results cannot be extrapolated to solutions containing blood, other proteins or lipids, or proteins at greater concentrations (>0.1%). [10,11]
The authors thank Dr. C. Lynch III and Dr. R. M. Epstein (University of Virginia) for their insightful comments on the manuscript; and Prof. Dr. med. Hugo Van Aken (Westfalische Wilhelms-Universitat Munster, Germany) for his support.