Propofol, a highly lipophilic anesthetic, is formulated in a lipid emulsion for intravenous use. Propofol has brisk onset and offset of effect after rapid administration and retains rapid offset characteristics after long-term administration. The authors tried to determine whether the emulsion vehicle is requisite for propofol's evanescent effect-time profile.
The time course of sedation and electroencephalographic (EEG) effect after propofol administration was measured in three studies in rats instrumented. In study 1, propofol was infused in either emulsion or lipid-free vehicle (n = 12), in a repeated measures cross-over design. In study 2, propofol in lipid-free vehicle was infused with or without simultaneous infusion of drug-free lipid emulsion (n = 6) in a repeated measures cross-over design. In study 3, propofol was infused in either emulsion (n = 5) or lipid-free vehicle (n = 5) to EEG burst suppression.
In study 1, relative to the emulsion formulation, propofol administered at equivalent doses in lipid-free vehicle resulted in a longer time to effect onset (1.4 +/- 0.2 vs. 0.5 +/- 0.1 min, EEG) and a trend for delayed anesthetic recovery (26.8 +/- 9.4 vs. 17 +/- 3.5 min, EEG; 26.1 +/- 8.8 vs. 16.8 +/- 3.3 min, sleep). In study 2, coadministration of drug-free emulsion with propofol did not alter the time course of effect. In study 3, more than twice the dose of propofol was required to achieve EEG burst suppression with the lipid-free formulation. Two animals died after administration of propofol to EEG burst suppression with the lipid-free formulation; no deaths occurred in the emulsion group.
The incorporation of propofol in emulsion reduces dose requirements and produces rapid onset and recovery of anesthetic effect.
Propofol (2,6-diisopropyl-phenol), is a highly lipophilic (log octanol-water partition coefficient: 4.33 ) anesthetic agent that is commercially formulated in a lipid emulsion (Diprivan; Zeneca Pharmaceuticals, Wilmington, DE) for intravenous use. Brisk onset and offset of effect occurs after rapid intravenous administration of induction doses of propofol. Minimal accumulation of propofol is observed after long-term infusions. Correspondingly, propofol demonstrates only moderate increases in context-sensitive half-times after infusions are terminated.
The rationale for the formulation of intravenous drugs in emulsions is usually driven by the biopharmaceutical concern of limited aqueous solubility. [2–5] Incorporation of lipid-soluble drugs in lipid emulsions has resulted in a decrease in hypersensitivity reactions, irritation, and pain on injection. [2,5–9] Lipid emulsions have been shown to modify the disposition of cyclosporin,  d-alpha-tocopherol,  and other compounds. Although the pharmacokinetics can be altered by emulsion formulation, the influence of lipid emulsion on the pharmacodynamics and the time course of drug action has not been systematically and rigorously examined. Clinical consequences of this pharmaceutical manipulation are not apparent. In this investigation, we tested whether lipid emulsion is required for propofol's evanescent effect-time profile. The time course of pharmacologic effect was examined in statistically efficient three-way, repeated-measures designs that allowed us to rapidly examine and establish whether emulsion formulation alters the time course of effect and potentially influences propofol's clinical behavior. Specifically we performed three studies. In study 1 we examined the time course of the electroencephalographic (EEG) effect of propofol administered in emulsion versus lipid-free formulation. The time course of effect was examined using these two formulations after administering equivalent and equieffective (transient burst suppression) propofol doses. In study 2, we tested whether the drug has to be intimately associated with the lipid emulsion to elicit an evanescent effect-time profile. This was performed by simultaneous infusion of propofol in the lipid-free formulation and the blank (drug-free) emulsion through separate venous access sites. In study 3, we examined the consequences of infusion of propofol in emulsion versus lipid-free formulation to a fixed dose requirement endpoint (transient burst suppression).
Emulsion Formulation. The emulsion formulation consisted of a 1:4 dilution of the commercial preparation (1 part 10 mg/ml Diprivan to 3 parts 5% dextrose USP to 0.02 parts ethanol). The final propofol concentration was 2.5 mg/ml. This formulation was infused at a rate of 2.5–3.2 ml [center dot] min sup -1 [center dot] kg sup -1. The minute amount of added ethanol achieved the same alcohol content as the lipid-free, in situ-prepared formulation.
Lipid-free Formulation. Due to propofol's limited solubility, the administration of propofol in a lipid-free vehicle required the development of a unique administration system using two syringe pumps (Harvard model 22). This lipid-free formulation is prepared in situ during the infusion process. Propofol oil (97% pure; Aldrich Chemical Co., Milwaukee, WI) was diluted with ethanol (approximately 1:5) and placed in a 250-micro liter gas-tight syringe on the first pump. A carrier solution, which was similar in composition to the diluted aqueous phase of the commercial emulsion preparation (one part 22.5 mg/ml glycerol in water to three parts 5% dextrose USP), was placed in a 5-ml gas-tight syringe on the second pump. This carrier solution did not contain ethanol. The two solutions, propofol in ethanol and the carrier solution, were simultaneously infused into a preprimed 22G mixing tee tube. The outflow of the mixing tee tube was connected to the rat's jugular cannula by a short piece of PE50 tubing. The flow rates of the component solutions were approximately 60 micro liter [center dot] min sup -1 [center dot] kg sup -1 for the propofol in ethanol solution (pump 1) and 2.5–3.2 ml [center dot] min sup -1 [center dot] kg sup -1 for the carrier solution (pump 2). Depending on the study treatment arm, the flow rate of pump 1 was adjusted to achieve propofol infusion rates of 6.25, 8, or 10 mg [center dot] min sup -1 [center dot] kg sup -1.
A model using rats fitted with instruments for long-term studies to assess EEG effect [12,13] was used to quantitate the time course of central nervous system depression resulting from intravenous administration of propofol in emulsion and lipid-free formulation. The details of animal instrumentation have been reported previously.  Five electrodes were permanently implanted in the skull in male Wistar rats weighing 548 (+/- 109) g under halothane anesthesia. Seven days after surgical implantation of EEG electrodes, intravenous (jugular; left, right, or both) PE50 catheters were implanted for drug and vehicle infusion under halothane anesthesia. Animals were allowed to recover for at least 24 h from halothane anesthesia before study.
Execution of Studies
On the day of the study, an EEG recording cable was connected to the mating socket (previously implanted in the skull) of a conscious rat confined in an open plastic animal cage (59 L x 39 H x 20 W cm, VWR Scientific, Bridgeport, NJ). PE50 tubing was used to connect the syringe(s) containing the drug, the vehicle, or both to the indwelling venous catheters. Animals were not restrained during the study. Propofol in emulsion or lipid-free vehicle was infused through the left jugular vein. In all treatments, ethanol was infused at a rate of 50 micro liter [center dot] min sup -1 [center dot] kg sup -1. This dose of ethanol resulted in ethanol concentrations of 1.5–2% in the infusate.
Heart rate was assessed using a model 7 polygraph and model 7DA electrocardiography preamplifier (Grass Instruments, Quincy, MA) from subcutaneous needles inserted into the fore paws and left hind leg shortly after loss of righting reflex. Previous studies in our laboratory have correlated ventilatory depression (arterial pressure of carbon dioxide greater than 50 mmHg, arterial pH less than 7.3) with bradycardia (fewer than 250 beats/min) in nonventilated rats receiving high-dose infusions of propofol. Therefore, in this study, when heart rate decreased to fewer than 250 beats/min, ventilation was assisted by mask using a custom-designed rodent face mask and a rodent-positive pressure ventilator (70 cycles/min, 20 cm H2O; Small Animal Ventilator, Harvard Apparatus Ltd., Natal, MA). Ventilation was continued until heart rate returned to 350 beats/min. Rectal temperature was maintained at 37.5 [degree sign] Celsius using heating pads and heat lamps as required.
Study Designs. Figure 1shows the study designs and treatment protocols for the three studies that were performed.
Study 1. The primary objective of this study was to compare the time course of EEG effect after administration of equivalent (treatment A vs. B) and equieffective (with respect to EEG burst suppression; treatment A vs. C) doses of propofol in emulsion vs. lipid-free formulations. Twelve rats were infused with propofol in a balanced repeated-measures, three-way crossover Latin square design. All treatments were administered at least 24 h apart. The three treatments consisted of infusions of propofol at a rate of (A) 6.25 mg [center dot] min sup -1 [center dot] kg sup -1 (propofol incorporated into emulsion), (B) 6.25 mg [center dot] min sup -1 [center dot] kg sup -1 (propofol in lipid-free formulation), and (C) 10 mg [center dot] min sup -1 [center dot] kg sup -1 (propofol in lipid-free formulation). Preliminary studies indicated that treatments A and C were equieffective with respect to the EEG nadir effect intensity. In all treatments, propofol was infused over 2 min or less if 5 s of isoelectric EEG (0 waves per second) was observed. In each treatment, the same volume of the fluid was infused. The small doses of ethanol incorporated in the formulations could potentially produce EEG or behavioral effects. To test this hypothesis, each rat received an additional treatment (treatment D) of ethanol in lipid-free vehicle (no propofol) 24 h after completion of the Latin square. In all treatments (A-D), ethanol was infused at a rate of 50 micro liter [center dot] min sup -1 [center dot] kg sup -1 for no more than 2 min.
Once anesthesia was induced (loss of spontaneous, unprovoked righting reflex), animals were positioned on their side on a 37.5 [degree sign] Celsius heating pad. Sleep duration was determined as time from induction of anesthesia (loss of righting reflex) until the time animals could spontaneously place all four paws of their feet on the pad (return of righting reflex). Animals were not stimulated to evoke their righting reflex response.  Peak EEG activity at onset and offset of effect is usually associated with onset and offset of anesthetic sleep. We compared the time to peak EEG activity and time between EEG activation and reactivation with the behavioral measures.
Study 2. The primary objective of this study was to test the effect of simultaneous infusion of drug-free lipid emulsion or drug-free lipid emulsion preload on EEG activity and animal behavior. Intralipid (Kabi Pharmacia, Clayton, NC)(10%) has the same composition as Diprivan, with the exception of the presence of propofol. Six rats were infused with propofol in a balanced repeated-measures, three-way crossover Latin square design. All treatments were administered at least 48 h apart. In all three treatments (E, F, and G), propofol in a lipid-free formulation was infused at 8 mg [center dot] min sup -1 [center dot] kg sup -1. For treatments F and G, the animals received an additional simultaneous infusion of drug-free 2.5% lipid emulsion (one part 10% Intralipid to three parts 5% dextrose USP) at 3.2 ml [center dot] min sup -1 [center dot] kg sup -1 for 2 min. This infusion rate (3.2 ml [center dot] min sup -1 [center dot] kg sup -1) resulted in the animals receiving the same volume of lipid as the diluted Diprivan formulation. This drug-free emulsion was infused into a separate indwelling venous catheter (right jugular vein). For treatment G, the animals received a drug-free emulsion preload (4 ml of 10% Intralipid) 30 min before propofol administration. Propofol was subsequently administered as in treatment F. In all treatments, ethanol was infused at a rate of 50 micro liter [center dot] min sup -1 [center dot] kg sup -1 for 2 min.
Study 3. The primary objective of this study was to examine the consequences of infusion of propofol to a fixed dose requirement endpoint in emulsion versus lipid-free formulations. Ten rats were infused with propofol at 6.25 mg [center dot] min sup -1 [center dot] kg sup -1 in emulsion (treatment H) or lipid-free formulation (treatment I) to the onset of an equieffective endpoint (5 s isoelectric EEG [0 waves/s]; burst suppression).  Ventilatory support was not provided for these animals.
Electroencephalographic Recording and Aperiodic Analysis
A four-lead EEG  was recorded on a model 79 EEG physiograph (Grass Instruments, Quincy, MA) with 100 Hz high pass, 0.3 Hz low pass, and 60 Hz notch filter with a gain of 15 micro V/mm. On-line EEG analysis was performed using an aperiodic EEG analyzer (LifeScan; Neurometrics, San Diego, CA). Aperiodic EEG analysis  continuously characterized the time course of the drug effect. The total number of waves/s (w/s) from 0.5 to 30 Hz was used as a measure of propofol drug effect. The details of EEG recording have been reported previously. [12,13]
EEG Processing and Data Analysis
Our EEG processing software reports EEG effect (w/s) every second. Time-EEG effect data were averaged every 5 s during the first hour from the start of the study. Because EEG effect changed less rapidly during the course of the study, the time interval between reports was decreased to one time-effect pair every 30 s. A total of at least 500 time-EEG effect data pairs were collected for each treatment in each rat.
Propofol's concentration-EEG effect relation, as with most general anesthetics, is biphasic. At low concentrations EEG activates, and at higher concentrations EEG activity decreases. Complete depression of EEG activity (burst suppression) occurs at very high concentrations. Because of the large amount of processed EEG data obtained in each treatment in each rat, the shape of the effect-time profile was characterized by the use of a spline fit to the data. These splines  perform well as interpolating functions in the sense that they go through the data and do not require the specification of a potentially biased model for the time course of effect. To objectively quantitate the time course of pharmacologic effect, descriptors were calculated from the least-squares fit of the spline to the data. Figure 2shows the descriptors used to quantify the EEG-time data.
Three-way analysis of variance was used to detect treatment, subject, and period effects. Tukey's pairwise comparison test was used to determine the difference between means at P < 0.05.
(Figure 3) presents the time course of EEG effect in a typical rat receiving a 6.25 mg [center dot] min sup -1 [center dot] kg sup -1 infusion of propofol in emulsion. The left panel shows the time course of the processed EEG data (w/s) with the spline fit, and the right panels shows the pattern (in 5-s epochs) of raw EEG activity. In the absence of any drug, the baseline EEG activity is characterized with high-frequency-low-amplitude activity (pattern 1). After the start of the infusion of propofol, EEG initially activates (high-frequency-high-amplitude activity: pattern 2), which is followed by a decrease in activity (low-frequency-high-amplitude activity: pattern 3). Continued infusion leads to periods of isoelectricity (pattern 4). Further infusion leads to complete ablation of EEG activity (pattern 5). After termination of infusion, the EEG activity progresses back through these patterns as the animal emerges from anesthesia. Propofol formulation (i.e., in emulsion or lipid-free vehicle) did not produce any qualitative change in structure of the raw (unprocessed) EEG signals. However, the time course of the processed EEG measure (waves per second) was affected by formulation. The propofol dose required to achieve each morphologic pattern shown in Figure 3depended on the formulation.
In the Latin square portion of the study, animals received propofol infusion for 2 min or until 5 s of isoelectricity (0 w/s) was observed. The mean infusion time in treatment A was 1.8 min (range, 1.3–2 min; 9 of the 12 animals had less than 2 min of infusion) and was 2 min in all animals for treatments B and C. Figure 4shows the time course of EEG effect in a typical rat for the three propofol treatments. Qualitatively, the time course of EEG activity for propofol in emulsion treatment (treatment A; top panel) is characterized by sharp peaks and troughs, indicating a transient effect-time profile. Propofol administered in lipid-free formulation exhibits a slower effect-time profile (middle and bottom panels). The maximal drug effect assessed as EEG activity at nadir was lower (i.e., higher number of waves per second) for treatment B (middle panel) compared with treatments A and C (top and bottom panels). The solid line in Figure 4represents the best spline fit for each data set.
For clarity, only the spline interpolations of the time-effect data for all treatments in 12 animals are presented in Figure 5. This figure qualitatively illustrates the consistency in the shape of the time versus effect data among the study animals for each treatment arm. For the emulsion formulation (treatment A; top panel), the time course of EEG activity for all animals is characterized by a rapid onset of peak EEG activation (less than 1 min), a sharp reversal of nadir effect, and an early offset of effect in the region of peak reactivation occurring within 15–30 min of infusion. When the same dose (as treatment A) of propofol was administered in lipid-free formulation (treatment B), onset of peak EEG activity was slower (1–2 min). The maximal effect intensity as measured by EEG nadir activity was lower (i.e., higher number of w/s) for treatment B compared with treatment A. The effect versus time trajectory around the nadir effect appeared more blunted, with the nadir effect being more variable in both time to and intensity of effect. The time to recovery of peak EEG activity was delayed (15–40 min) and clearly more variable. One animal in this treatment group had a higher maximal effect with an extended recovery time (more than 50 min).
When a higher dose of propofol was administered in lipid-free formulation (treatment C, Figure 5), as expected the time to onset of peak EEG activation effect was accelerated (compared with treatment B). Most animals achieved transient burst suppression (which in this spline representation is less than 3 w/s), but the effect versus time trajectory around the nadir effect still appeared blunted and the time to achieve nadir effect was variable and delayed (compared with treatment A). The time to recovery of peak EEG activity was delayed (30–60 min) and highly variable.
(Table 1) quantifies the changes in profiles presented in Figure 5using the descriptors of the time course of EEG effect of propofol for the Latin square portion of Study 1. The transient nature of the time course of EEG activity for the emulsion treatment is supported by shorter time taken to reach peak activation (0.5 min), nadir (1.8 min), and peak reactivation (17 min). Times to peak activation, nadir, and peak reactivation were longer (Table 1) for treatments B and C compared with treatment A.
Baseline effect and peak EEG activity at onset (Table 1) were not significantly different across formulations. Peak EEG activity at offset (peak reactivation) was significantly lower for treatment C. The maximal drug effect as assessed by EEG activity at nadir was similar for treatments A (1.4 +/- 1.2 w/s) and C (2.2 +/- 1.8 w/s) but significantly lower for treatment B (5.8 +/- 2.6 w/s). Time to peak EEG activity at onset and baseline effect (pre-nadir) were significantly different across treatments. The time to maximal drug effect (time to EEG nadir) was significantly shorter for the emulsion formulation. Even the higher dose of propofol in the lipid-free formulation could not accelerate the time to maximal (nadir) effect (treatment C vs. B). Time to baseline effect (post-nadir) and peak EEG activity at offset were significantly longer for treatment C compared with treatments A and B.
Behavioral sleep time was compared with EEG sleep time. The time to onset (loss of righting reflex) and offset (return of righting reflex) of anesthetic sleep were highly correlated (r2> 0.98) to time to peak EEG activity at onset (time to peak activation) and offset (time to peak reactivation) of effect, respectively. Figure 6shows the high correlation between EEG effect and behavioral effects.
The duration of sleep (difference of time to peak EEG activity at onset and offset of effect or loss and return of righting reflex) was significantly longer (Table 1) for treatment C (52.4 +/- 16.2 min) compared with treatments A (16.8 +/- 3.3 min) and B (26.1 +/- 8.8 min). Table 1also shows the high correlation between the time to loss and return of righting reflex and the time to peak activation and reactivation, respectively.
Control treatment with ethanol (treatment D) did not produce any change in EEG activity either by direct examination of the EEG traces or by the reported processed effect.
(Figure 7) shows the mean processed EEG data for the time course of EEG effect of propofol for the three treatments (E, F, and G). Table 2reports the descriptors of the time course of the EEG effect of propofol. No significant difference across treatments was seen for any of the descriptors. Peak EEG activity at onset and offset of effect and baseline effect were similar to the values from Study 1. Intralipid preload for treatment G had no effect on EEG activity. There appears, however, to be a trend for shorter time to reactivation (28.7 min;Table 2) and higher nadir EEG activity (5.7 w/s;Table 2) in treatment G.
By combining the results of studies 1 and 2, we found that the incorporation of propofol in emulsion results in a more transient time course of effect. Propofol administered without emulsion showed a slower and more sluggish recovery, as characterized by the descriptors of time course of effect (Figure 3, Figure 4, Figure 5and Table 1and Table 2). In 5 of the 30 (17%) treatments of the lipid-free formulation, the animals received mechanical ventilation based on our ventilation protocol. Ventilatory support was not required in any animal receiving propofol in emulsion treatment.
Animals infused with the lipid-free formulation (treatment 1) to burst suppression (equieffective endpoint) showed high mortality rates (two of five animals died). No deaths occurred in any animal infused to burst suppression with emulsion formulation (treatment H). Due to the delay in onset of effect in animals infused with the lipid-free formulation, infusion time to EEG burst suppression (Table 3) extended from 2.1 min for the emulsion formulation to 4.8 min for the lipid-free formulation. This resulted in a dose requirement increase (Table 3) from 13 mg/kg (emulsion formulation) to 29 mg/kg (lipid-free formulation).
For the investigation of single anesthetic agents, EEG provides a continuous measure of anesthetic effect intensity that can reflect clinical behavior. Recently Gustafsson et al.  correlated various behavioral measures with EEG activity in this rodent model for thiopental. Hung et al.  observed a similar correlation for thiopental in patients classified as American Society of Anesthesiologists physical status I and II who were having surgery. For our propofol studies, the time to peak EEG activity at onset and offset of effect were highly correlated with behavioral effects (e.g., loss and return of righting reflex). This shows that for propofol, like thiopental,  aperiodic analysis is predictive of clinical (behavioral) state and may serve as a good, objective surrogate measure of other clinically observable anesthetic effects. More importantly, EEG effect measures allows quantitation of effect in a continuous graded manner, revealing a wider range of drug effects.
Because of solubility considerations, propofol concentration in the lipid-free formulation was 2.5 mg/ml. A 1:4 dilution of the commercial vehicle ensured that each treatment arm received the same total volume of fluid (studies 1 and 2). In addition, the 1:4 dilution allows sufficient flow rates through implanted cannulas and minimizes the influence of catheter dead volumes on our observations. In separate experiments, we have shown that dilution of the commercial preparation in this ratio does not influence the time course of EEG activity and is within the recommended dilution guidelines contained in the Diprivan package insert.
Our studies indicate that emulsion plays a significant role in onset, intensity, and duration of pharmacologic effect of propofol. Time to peak EEG activity at onset, loss of righting reflex, and maximal effect (EEG nadir) were significantly shorter for the emulsion formulation (Table 1). Thus emulsion formulation accelerates time to onset and maximal effect. Maximal effect (EEG nadir activity) was lower (i.e., higher EEG w/s) for the lipid-free formulation compared with the emulsion formulation (treatment A vs. B;Table 1) for equivalent doses of propofol. Duration of anesthetic effect (sleep time) tended to be shorter for the emulsion formulation (treatments A vs. B;Table 1) for equivalent doses of propofol. Therefore it appears that the emulsion formulation may be responsible for imparting many of the clinical characteristics that have led to the wide use of this anesthetic.
Simultaneous infusion of drug-free lipid emulsion or drug-free lipid emulsion preload produces minimal change in the time course of EEG activity or animal behavior (study 2). Thus it appears that the evanescent effect-time profile of propofol does not arise from the simple presence of lipid in systemic circulation, but rather the drug needs to be incorporated in the emulsion to produce desirable clinical characteristics of rapid onset and recovery. However, for treatment G (study 2), the duration of sleep tended to be slightly shorter and maximal effect tended to be lower. This last observation suggests that infused propofol may be associated, to some extent, with circulating emulsion particles that are initially void of drug. However, it is clearly evident that this treatment does not mimic the behavior of the emulsion formulation.
Emulsion formulation decreases the dose of propofol required to produce EEG burst suppression. Both treatments A and C of study 1 produced EEG burst suppression even though the dose of propofol in treatment C was 60% larger. In treatment B, the same dose of propofol was administered as in treatment A, but administration in the lipid-free vehicle failed to produce EEG burst suppression. Dose requirement determined by administering drug to an EEG burst suppression endpoint was increased more than two times for the lipid-free formulation compared with the emulsion formulation (study 3).
It should be noted that we could be mislead by examining single endpoints such as sleep time when comparing dose response relationships. Using traditional measures of sleep times (study 1) as the sole endpoint, treatments A and B (equivalent doses of propofol) were not significantly different, whereas sleep time was longer for treatment C (consistent with a larger dose of propofol). Sleep time has been used previously to test potency of anesthetic agents.  This practice may be misleading given our results using a continuous measure of effect, which can fingerprint the complete shape of the time-effect relation.
The emulsion-induced evanescent effect-time profile cannot be explained by a simple shift in propofol potency (concentration vs. effect relationships) because it does not appear that the effect-time profile resulting from treatment A (study 1) could be reproduced by simply altering the dose of propofol in the lipid free vehicle. These two formulations produce intrinsically different time course signatures. For example, the high-dose, lipid-free formulation (treatment C, study 1) could not produce the time to nadir effect seen in treatment A even though this dose was equi-effective with respect to nadir effect intensity. Therefore, it appears that emulsion alters propofol disposition and possibly the intrinsic uptake and distribution of propofol to tissues and the effect site. This is further supported by the observation that the time to achieve maximal effect intensity (time to nadir effect) depended on formulation but not on dose (treatment B vs. C; study 1), suggesting that a formulation-induced shift in the time constants controlling propofol disposition has occurred.  Another possibility is a formulation-induced alteration in the shape of the concentration vs. effect relation.
Emulsion formulation enhances the safety (study 3) of propofol. The clinical doses of propofol necessary for anesthetic induction generally produce transient EEG burst suppression. In our administration paradigm, animals infused to burst suppression showed high mortality rates for the lipid-free formulation. This was possibly due to physiologic decompensation, which may have occurred during the extended time spent at the deepest levels of anesthesia. To administer safely propofol in a lipid-free formulation, the continued accumulation of effect that occurs after dosing must be anticipated and the dose potentially reduced. Thus propofol takes on inertial characteristics of traditional lipid-soluble anesthetics when administered in the lipid-free formulation. It may not be possible to achieve profound anesthesia safely using the lipid-free formulation of propofol.
What are the potential mechanisms for this emulsion effect? Emulsions can restrict lipid-soluble drugs within the vascular space and decrease the volume of distribution. Incorporation of cyclosporin  in lipid emulsion produced a 70% decrease in volume of distribution. Emulsions can also accelerate removal of lipid-soluble drugs by two primary mechanism. First, by restricting the drug to the central vascular space, emulsions increase the amount of drug available for clearance by the eliminating organs. Second, emulsion particles are subject to phagocytosis by the reticuloendothelial system.  In our second study, drug-free lipid emulsion was concomitantly administered with the lipid-free formulation to determine whether it is possible to recover the evanescent properties of the emulsion formulation. Because drug-free lipid emulsion could not accelerate the time course of effect (study 2; treatment E vs. treatments F and G), it does not appear that propofol and the infused emulsion particles associate significantly in vivo. From the results of this study, we cannot determine whether reticuloendothelial system uptake is potentially important for the elimination of propofol incorporated in emulsion.
An alternative explanation of the propofol-emulsion effect could be that emulsion particles may diminish the extent of propofol first-pass pulmonary uptake, accelerate pulmonary transit after intravenous administration, or both. Initial peak plasma concentrations of lipid-soluble drugs such as fentanyl are significantly reduced through pulmonary uptake mechanisms. [22–24] Enhanced pulmonary uptake after administration of propofol in the lipid-free formulation is consistent with the observed reduction in nadir effect intensity and prolongation in time to reach this maximal effect.
It appears that the emulsion formulation is more than a pharmaceutically elegant means of solublizing the anesthetic. It is essential for propofol's desirable clinical characteristics. These initial studies were designed to screen for the possibility that the lipid potentially alters the time course of pharmacologic effect. Pharmacokinetic and pharmacodynamic studies are needed to understand the mechanisms of this interaction.