As in inhaled isoflurane anesthesia, when isoflurane lipid emulsion (ILE; 8%, vol/vol) is intravenously administered, the primary elimination route is through the lungs. This study was designed to determine the minimum alveolar concentration (MAC) and the time course of washout of isoflurane for intravenously infused ILE by monitoring end-tidal isoflurane concentration.


Twelve healthy adult mongrel dogs were assigned randomly to an intravenous anesthesia group with 8% ILE or to an inhalation anesthesia group with isoflurane vapor. An up-and-down method and stimulation of tail clamping were used to determine MAC of 8% ILE by intravenous injection in the intravenous anesthesia group and MAC by the inhaled approach in the inhalation anesthesia group, respectively. Isoflurane concentration and partial pressure in end-tidal gas, femoral arterial blood, and jugular venous blood were measured simultaneously just before each tail clamping and during washout.


The induction time in the intravenous anesthesia group (105 +/- 24 s) was shorter than that in the inhalation anesthesia group (378 +/- 102 s; P < 0.01). MAC of 8% ILE by intravenous injection (1.12 +/- 0.18%) was significantly less than MAC by the inhaled approach (1.38 +/- 0.16%; P < 0.05). No significant difference was found between the two groups in the time course of washout of isoflurane.


The MAC of intravenous anesthesia with 8% ILE was less than that of inhalation anesthesia with isoflurane vapor in dogs.

THE ED50and LD50of intravenous isoflurane lipid emulsion (ILE) in mice,1the anesthetic and physiologic effects of intravenous halothane lipid emulsion in swine,2and the hemodynamics effects of intravenous infusion of halothane lipid emulsion in dogs3have been studied. However, the dog’s minimum alveolar concentration (MACiv) with intravenous infusion of ILE (8%, vol/vol) has not been investigated. As with inhaled isoflurane anesthesia, the primarily eliminative route of isoflurane by intravenous infusion of ILE is also through the lungs. Therefore, the MAC of 8% ILE by intravenous injection (MACiv-ILE8) could be determined by monitoring end-tidal isoflurane concentration. Musser et al.  2demonstrated that the MACiv(0.78%) of intravenous emulsified halothane (5%, vol/vol) was smaller than the MAC by the inhaled approach (MACinh; 1.13%) of halothane vapor in swine. Because of the similar chemicophysical properties of halothane and isoflurane, we hypothesized that the MACivof isoflurane was smaller than its MACinh. The difference of anesthetic partial pressure between alveolar gas and arterial blood has been shown to be a positive value during maintaining or deepening anesthesia by inhaled volatile anesthetic4and might be a negative value when emulsified anesthetics are infused intravenously. However, the arterial partial pressures (indicating the partial pressure in the central nervous system after equilibration) of isoflurane between the intravenous approach and inhalation should be the same, because isoflurane dissolved in Intralipid® (Huarui Pharmacy, Ltd., Wuxi, Jiangsu, China) produces anesthesia through the same mechanism as inhaled isoflurane when half of the animals are anesthetized. We designed this study to verify these predictions during intravenous and inhaled anesthesia approaches by simultaneously measuring isoflurane concentrations and partial pressures in the femoral arterial blood, jugular venous blood, and end-tidal gas.

Isoflurane was purchased from Abbott Laboratories (Queenborough, Kent, United Kingdom). Intralipid ®, 30%, was provided by Huarui Pharmacy, Ltd. On the day before the experiment, using an aseptic technique, 18.4 ml Intralipid®, 30%, and 1.6 ml liquid isoflurane were transferred into a 20-ml glass ampoule using syringes, and the ampoule was sealed using an alcohol blowtorch. Then, the ampoule was vigorously shaken on a vibrator for 15 min to solubilize isoflurane into the lipid emulsion. Sixty ampoules of 8% ILE were prepared for the study using this procedure. The ampoule was opened just before use, and the residual ILE in the opened ampoule was not reused in this study. Before this experiment, the stability of 8% ILE was investigated; no change in isoflurane concentration and no lipid droplets were found during 6 months of storage at room temperature (20°–26°C). The measured concentration of isoflurane in this lipid emulsion at 15 min after the ampoule was opened was not different compared with the isoflurane concentration before the emulsion was transferred into the ampoules, which might be due to a high isoflurane lipid/gas partition coefficient (70/52 at 20°–25°C; unpublished data from our laboratory Xiao-Lin Yang, M.D., Han-Xiang Ma, M.B., Nan-Fu Luo, M.B., Wen-Sheng Zhang, M.D., and Jin Liu, M.D., June 2004).

After obtaining approval from the Institutional Animal Care and Use Committee of Sichuan University (Chengdu, Sichuan, China), 12 healthy mongrel dogs, aged 12–18 months and weighing 9.0–12.5 kg, were randomly assigned to the intravenous anesthesia group (Giv) with 8% ILE or to the inhalation anesthesia group (Ginh) with isoflurane vapor. All dogs were allowed to socialize with people on daily basis before the study. Food, but not water, was withheld for 12 h before the start of each experiment.

After body weight had been measured, dogs were positioned in a sternal recumbency on a restraining table. Each dog was fitted for a special facemask connected to an anesthesia machine (Excel 210 SE; Datex-Ohmeda, Madison, WI) fitted with an isoflurane vaporizer; dogs initially breathed 100% oxygen before ILE or isoflurane vapor was administered. In Giv, 8% ILE was administered into the right cephalic vein at a rate of 1 ml/s, and in Ginh, 5% isoflurane was inhaled with a fresh oxygen flow rate at 4 l/min through the facemask. The tidal volume and respiratory rate were monitored using the anesthesia machine. The dog’s palpebral reflex was assessed by gently brushing the eyelashes of one eyelid with a finger every 5 s after isoflurane administration. After the palpebral reflex disappeared, an 8-French cuffed endotracheal tube was inserted. Ventilator settings were adjusted to maintain normal end-tidal carbon dioxide pressure at 35–40 mmHg. Heart rate, electrocardiograph, lingual pulse oxygen saturation, femoral arterial blood pressure, and esophageal temperature were monitored with use of a 150B3 monitor (Philips, Suzhou, China). The inspired and expiratory anesthetic concentrations and end-tidal carbon dioxide pressure were also monitored with this monitor through a sampling line placed in the tip of the cuffed endotracheal tube. The fresh oxygen flow rate was maintained between 3 and 3.5 l/min, which exceeded the minute ventilation volume to avoid isoflurane rebreathing. The left jugular vein was chosen for infusing lactated Ringer’s solution (10 ml · kg−1· h−1) and for collecting blood samples. The esophageal temperature was maintained at 36.5°–38.5°C by using heated blankets.

The up-and-down method and the tail clamping stimulation were used to determine MACiv-ILE8 and MACinh.5–7In Giv, the initial end-tidal isoflurane concentration for determining MACiv-ILE8 was maintained at 0.9% for at least 20 min by adjusting the rate of the microinfusion pump (TCI-I; Silugao High Technology Development Co., Ltd., Beijing, China). In Ginh, the initial end-tidal isoflurane concentration of 1.1% was maintained for 20 min by adjusting the isoflurane vaporizer. Before each tail clamping, 5 ml end-tidal gas, 4 ml femoral arterial blood, and 4 ml jugular venous blood were drawn simultaneously for determining isoflurane concentrations and partial pressures by using a gas chromatograph (described two paragraphs below). Then, a pair of forceps was clamped on the shaved tail approximately 5–10 centimeters from its base in intervals of three times every 15 s for 1 min. A positive response to the tail clamping was considered to be a gross purposeful muscular movement, usually of the head or other extremities. Only significant jerking or twisting motions were included, not head movements such as twitches or grimaces. Coughing, swallowing, or chewing was not considered to be a positive response.5If the tail clamping response was positive or negative, the end-tidal isoflurane concentration was increased or decreased by 20% and maintained for 20 min before subsequent clamps. The process of assessing the response was continued until a crossover of a positive response with a negative response had occurred. The concentration or partial pressure in the arterial blood, the venous blood, or the end-tidal gas to a midway point between the highest (which allowed movements) and the lowest (which prevented movements) was defined as the value of minimum arterial concentration or partial pressure, minimum venous concentration or partial pressure, or minimum alveolar concentration (MACiv-ILE8 for Givand MACinhfor Ginh) or partial pressure, respectively.

Administration of 8% ILE or isoflurane vapor was immediately stopped when the last tail clamping was finished. The anesthetic vapor analyzer continued to monitor the elimination of isoflurane from the lungs. The dog was considered awake when palpebral reflex recovered and the endotracheal tube could not be tolerated. Then, 2–4 mg midazolam, 0.05–0.1 mg fentanyl, and 10–15 mg atracurium were intravenously injected to keep the dog anesthetized and the ventilation controlled. The time from stopping administration of ILE or isoflurane vapor to just before intravenous injection of these anesthetics for the first time was defined as the awakening time. During the period from stopping intravenous ILE or inhaled isoflurane to the end-tidal isoflurane concentration on the screen reading zero, samples of end-tidal gas and femoral arterial and jugular venous blood were obtained simultaneously when the end-tidal isoflurane concentration decreased to 0.7%, 0.5%, 0.4%, 0.3%, 0.2%, 0.1%, and 0%, respectively. Samples were also obtained at the awakening time. All dogs were observed daily for 2 weeks after experimentation for redness, swelling, or thrombophlebitis at the infusion site, and for death.

Isoflurane concentration was determined by using a gas chromatograph (Agilent 4890D; Tegent Technology Ltd., Shanghai, China) equipped with a 6-m-long stainless steel column (3.2-mm internal diameter) packed with Chromosorb-P 60/80 mesh maintained at 75°C. A 17-ml/min nitrogen carrier stream flow was delivered through the column to a flame ionization detector supplied by hydrogen at 40 ml/min and by air at 200 ml/min. The output from the gas chromatography was collected by an HP 3398 GC workstation (Tegent Technology Ltd.), and peak areas were automatically integrated. The isoflurane concentration in the blood (C) and the determined value of isoflurane blood/gas partition coefficient (λb/g) were measured by the two-stage headspace equilibration method.8–10Isoflurane partial pressure in the blood (P) was calculated by the equation P = (C/λb/g) × 760 mmHg, and the end-tidal isoflurane partial pressure was equal to its percent of concentration multiplied by 760 mmHg.

Data were reported as mean ± SD. The Student t  test was used for comparison between the two groups. One-way analysis of variance was applied to the comparisons among the isoflurane concentration or partial pressure in the arterial blood, venous blood, and end-tidal gas either in Givor in Ginh. P < 0.05 was accepted as statistical significance. Linear regression analysis was used to analyze the correlation between λb/gand the volume of infused lipid emulsion. Combined data during the period of administration and washout were used for linear regression analysis to correlate isoflurane partial pressures between arterial blood and end-tidal gas at corresponding time points for both Givand Ginh. During washout, the rate constant of isoflurane elimination (β) was obtained by the slope of each linear regression equation between the logarithm of isoflurane concentration in the blood and its corresponding washout time.11The elimination half-life of isoflurane concentration in the blood (t0.5) after stopping administration of the anesthetic agent was calculated by the equation t0.5= 0.693/β. The time of isoflurane washout from the blood for 95% (t0.95) or 99% (t0.99) was calculated by the equation t = log (C0/C)/β, where C0was the isoflurane initial concentration in the blood at the beginning of the washout phrase, and the ratios of C0/C should be 100/5 and 100/1 when isoflurane was washed out to 95% and 99%, respectively. The mean values of t0.5, t0.95, and t0.99were obtained for the two groups.

In Giv, the palpebral reflex was lost at 13 ± 2 s at an injection speed of 1 ml/s of ILE, and the consumed liquid isoflurane to successful intubation was 1.8 ± 0.5 ml on average (table 1). The whole course of induction and intubation took 105 ± 24 s, which was very smooth on the first attempt in all dogs in this group. In Ginh, palpebral reflex was lost at 348 ± 90 s, and the consumed liquid isoflurane for successful intubation was 9.8 ± 2.0 ml. The whole induction and intubation course took a much longer time (378 ± 102 s) than that in Giv, and some dogs displayed induction-related complications such as struggling, swallowing, vocalizing, excitement, or coughing, which made the induction course relatively difficult. Before induction, no significant differences were found either in respiratory rate or in tidal volume between the two groups. However, just when dogs lost palpebral reflex, the respiratory rate and tidal volume in Givhad rapidly decreased to 4 ± 1 breaths/min and 60 ± 10 ml, respectively, which indicated significant respiratory depression compared with those in Ginh(8 ± 2 breaths/min and 90 ± 10 ml).

Table 2indicates that the difference of minimum arterial concentration (4.77%) versus  minimum venous concentration (4.44%) or minimum arterial partial pressure (10.53%) versus  minimum venous partial pressure (9.83%) was not significant in Giv(P > 0.05), but both minimum arterial partial pressure and minimum venous partial pressure were remarkably higher than minimum alveolar partial pressure (8.51%; P < 0.01). Similarly, in Ginh, no significant differences were found either in minimum arterial concentration (1.6%) versus  minimum venous concentration (1.46%) or in minimum arterial partial pressure (9.01%) versus  minimum venous partial pressure (8.18%), but both minimum arterial partial pressure and minimum venous partial pressure were lower than minimum alveolar partial pressure (10.09%, P < 0.01). Although no significant difference in minimum arterial partial pressures or in minimum venous partial pressures between the two groups was found, MACiv-ILE8 was lower than MACinh. The infused volume of lipid emulsion (x) in Givhad a remarkable impact on the isoflurane blood/gas partition coefficient (λb/g= y), and the regression equation was y = 0.0139x + 1.4641 (R  2= 0.883; fig. 1). A positive linear correlation existed in the isoflurane partial pressures between the arterial blood and the end-tidal gas either in Giv(fig. 2A) or in Ginh(fig. 2B). At the awakening time point, the isoflurane concentrations in both groups were in the same order—venous blood > arterial blood > end-tidal gas—and no significant difference in the arterial partial pressure, the venous partial pressure, or the end-tidal anesthetic partial pressure was found between the two groups, respectively (table 3). Also, no significant difference was found in the awakening time between the two groups. Table 4indicates that t0.5, t0.95, and t0.99were not significantly different between arterial blood and venous blood within each group or between the two groups, respectively. No fatalities occurred in the two groups during the 2-week observation after the MAC determination, and daily evaluations of the cephalic vein used for infusion showed no evidence of redness, swelling, or thrombophlebitis in Giv.

In this study, anesthesia with isoflurane administered intravenously in an emulsion was compared with an inhaled delivery of isoflurane in dogs. Compared with intravenous injection, the inhalation approach takes a longer time for a volatile agent to be delivered from the anesthetic vaporizer to alveoli, then dissolve into blood, and finally arrive in the central nervous system. Faster onset and less liquid isoflurane consumption were found in Givthan those in Ginhduring induction (table 1). This is mainly because the anesthetic agent can reach the site of action in the brain in one arm-brain circulation by intravenous administration and also partly because of the faster injection rate in this study, which imitated clinical intravenous induction. If the infusion rate for ILE were designed for mimicking the inhalation induction with isoflurane, the onset time might be longer, and liquid isoflurane consumption might be higher.

The elimination of isoflurane after intravenous infusion of ILE is mainly through the lungs, as when isoflurane vapor was inhaled. Therefore, intravenous ILE anesthesia has the characteristics of both intravenous anesthesia and inhalation anesthesia. The depth of anesthesia could be easily adjusted by changing the infusion rate of the ILE based on the monitoring of end-tidal isoflurane concentration. If low fresh gas flow is used, the eliminated isoflurane from the blood to the alveoli is reabsorbed. Then, the mode of isoflurane delivery will in part be inhalational, and the actual MACivmay be altered to a value between MACiv-ILE8 and MACinhobtained in this study. In addition, the consumed liquid isoflurane and the cost of anesthesia will be decreased.12–14Similar to the results obtained by Musser et al.  2showing that MACiv(0.78%) of emulsified halothane (5%, vol/vol) was lower than MACinh(1.13%) of halothane in swine, our current study indicated (table 2) that MACiv-ILE8 (1.12 ± 0.18%) was lower than MACinh(1.38 ± 0.16%), which corresponded with the prediction mentioned in the introduction. The difference between MACiv-ILE8 and MACinhmight mainly relate to the following factors: First, because ILE was administered intravenously and the pulmonary system is not the organ of entry of isoflurane, expired isoflurane concentration at an anesthetizing blood level would be lower and the MACiv-ILE8 would not reflect MACinhin the traditional sense. Second, the equilibration time for isoflurane concentration (or partial pressure) between alveolar gas and arterial blood was also important.4In this study, minimum alveolar partial pressure (8.51%) was lower than minimum arterial partial pressure (10.53%) in Giv, but minimum alveolar partial pressure (10.09%) was higher than minimum arterial partial pressure (9.01%) in Ginh, indicating that the partial pressures of isoflurane between alveolar gas and arterial blood had not achieved complete equilibration within 20 min. If the equilibration time were prolonged for both groups, the difference between MACiv-ILE8 and MACinhwould be smaller. Third, the infused lipid emulsion increases the gradient of isoflurane partial pressure between blood and alveolar gas. Our results demonstrated that the infused volume of lipid emulsion remarkably impacted isoflurane λb/g, as figure 1showed that approximately 1 ml Intralipid®, 30%, could increase isoflurane λb/gby 0.0139 at 37°C. In Giv, the mean infused lipid emulsion was 159.7 ± 33.9 ml, and the highest isoflurane λb/gwas 5.23 at 37°C, whereas the isoflurane λb/gwas constant at 1.4 ± 0.24 at 37°C in Ginhbecause no lipid emulsion was infused. The increased isoflurane λb/gin Givmade isoflurane much easier to be dissolved in the blood and not easier to be eliminated from the blood to the lungs. This has been demonstrated in this study by that the minimum arterial concentration and minimum venous concentration in Givwere much higher than those in Ginh.

Although MACiv-ILE8 was less than MACinh, minimum arterial partial pressures between the two groups were similar, suggesting that an equal partial pressure for isoflurane in the central nervous system and the same depth of anesthesia existed in the two groups at respective MAC values. In this study, the venous blood was drawn from the jugular vein, and no difference was found between the minimum venous partial pressure and the minimum arterial partial pressure for both groups. This implies that the isoflurane partial pressure in both jugular vein blood and arterial blood can reflect the brain tissue partial pressure. In this study, a good positive linear correlation in isoflurane partial pressures between the arterial blood and the end-tidal gas during isoflurane inhalation in Ginhwas shown again (fig. 2A).15More important, a good positive linear correlation also existed in isoflurane partial pressures between the arterial blood (possibly the central neural system) and the end-tidal gas in Giv(fig. 2B). This result indicates that the depth of anesthesia could be controlled by adjusting the infusion rate of ILE based on monitoring the end-tidal isoflurane concentration when ILE is administered intravenously in the clinic.

The awakening speed in both groups was very fast. The awakening times (approximately 8 min) were not significantly different between the two groups, and the awakening end-tidal isoflurane concentrations between the two groups exhibited no obvious differences. When the dogs awoke, the partial pressures in both groups were in the same order—venous partial pressure > arterial partial pressure > end-tidal partial pressure—which was reversed in relation to those during the MAC measurement in Ginh. The awakening isoflurane concentrations in both arterial blood and venous blood in Givwere higher than those in Ginh, which might also be due to a greater isoflurane λb/gin Givthan that in Ginh. When concentrations were transferred into partial pressure, the above comparison showed no significant difference. Therefore, the isoflurane partial pressures in the central nervous system should be close in the two groups when dogs awaken. The same awakening time and awakening blood partial pressure in both groups suggest that the awakening course in dogs has not been affected by the different isoflurane administration approaches. In addition, the time for isoflurane elimination for 50% (t0.5), 95% (t0.95), and 99% (t0.99) from the blood between the two groups had no remarkable differences, which indicates that the infused lipid emulsion had not remarkably impacted the time course for isoflurane elimination from the blood in this study.

Neither death nor local redness, swelling, or phlebitis at the infusion site occurred during 2 weeks’ observation after MAC determination, which implies that intravenous anesthesia with 8% ILE for dogs is safe. However, attention should be paid to the possible drawbacks of 8% ILE. Because of the high volatility of isoflurane, 8% ILE should be kept in a container that is absolutely sealed until immediately before use and not be reused. Also, liquid isoflurane must be completely dissolved into lipid emulsion with no lipid droplets to be found for extended periods of time to enable the delivery of consistent isoflurane concentrations.

In conclusion, compared with inhaled isoflurane vapor, intravenous 8% ILE has a smaller MAC, has faster induction of anesthesia, and uses less isoflurane. The infused lipid emulsion increases the isoflurane blood/gas partition coefficient, but the awakening time courses of the dogs are not significantly different between Givand Ginh. When 8% ILE was administered intravenously, end-tidal isoflurane concentration could be used for adjusting the infusion rate of 8% ILE, estimating the isoflurane partial pressure in arterial blood and further judging the depth of anesthesia. This conclusion indicates a good prospect for delivering 8% ILE intravenously in the clinical setting.

Eger RP, Macleod BA: Anaesthesia by intravenous emulsified isoflurane in mice. Can J Anaesth 1995; 42:173–6
Musser JB, Fontana JL, Mongan PD: The anesthetic and physiologic effects of an intravenous administration of a halothane lipid emulsion (5% vol/vol). Anesth Analg 1999; 88:671–5
Biber B, Johannesson G, Lennander O, Martner J, Sonander H, Werner O: Intravenous infusion of halothane dissolved in fat: Hemodynamics effects in dogs. Acta Anesthesiol Scand 1984; 28:385–9
Zbinden AM, Thomson DA, Westenskow DR, Frei F, Maertens J: Anaesthetic uptake and elimination: Is there a difference between halothane and isoflurane in the dog? Br J Anaesth 1988; 60:395–401
Eger EI II, Saidman LJ, Brandstater B: Minimum alveolar anesthetic concentration: A standard of anesthetic potency. Anesthesiology 1965; 26:756–63
White D: Editorial: Uses of MAC. Br J Anaesth 2003; 2:167–9
Kazama T, Ikeda K: Comparison of MAC and the rate of rise of alveolar concentration of sevoflurane with halothane and isoflurane in the dog. Anesthesiology 1988; 68:435–7
Zhou JX, Liu YQ, Liu J: The combined effect of hypothermia and crystalloid hemodilution on the blood solubility of volatile anesthetics. Chinese Med J 2002; 115:1014–9
Hu P, Zhou J-X, Liu J: Blood solubilities of volatile anesthetics in cardiac patients. J Cardiothorac Vasc Anesth 2001; 115:560–2
Yu R-G, Zhou J-X, Liu J: Prediction of volatile anaesthetic solubility in blood and priming fluids for extracorporeal circulation. Br J Anaesth 2001; 86:338–44
Wen-Quan L, Jian-Ping L: Chapter 8: One-compartment model, Biopharmaceutics and Pharmacokinetics, 2nd edition. Beijing, People’s Sanitarian Publishing Company, 2003, pp 164–225
People’s Sanitarian Publishing Company
Lockwood GG, White DC: Measuring the costs of inhaled anesthetics. Br J Anaesth 2001; 87:559–63
Bailey CR, Ruggier R, Cashman JN: Anaesthesia: Cheap at twice the price? Staff awareness, cost comparisons and recommendations for economic savings. Anaesthesia 1993; 48:906–9
Cotter SM, Petros AJ, Dore CJ, Barber ND, White DC: Low-flow anaesthesia: Practice, cost implications and acceptability. Anaesthesia 1991; 46:1009–12
Lockwood GG, Sapsed-Byrne SM, Adams S: A comparison of anaesthetic tensions in arterial blood and oxygenator exhaust gas during cardiopulmonary bypass. Anaesthesia 1999; 54:434–6