Introduction

Propofol is a general anesthetic with a rapid onset of action and a short recovery period. It is often used together with remifentanil, an opioid analgesic, to induce and maintain anesthesia or sedation. Propofol is formulated as an oil-in-water emulsion because of its low aqueous solubility (Table 1).

Table 1:

Molecular structure and relevant physico-chemical properties of propofol and remifentanil.

Property	Propofol	Remifentanil hydrochloride
Structure
Molecular weight	178.3 g/mol	412.9 g/mol
Aqueous solubility	124 mg/l	591 mg/l
LogP (octanol/water)	3.79	1.75
pKa	11.1	7.1

It is a common clinical practice that remifentanil and propofol are co-administered via the same intravenous (i.v.) catheter line. Co-administration of drugs predisposes to potential compatibility issues. Adverse consequences range from i.v. line obstruction to life-threatening embolism. In the case of parenteral emulsions, such as propofol, an incompatibility reaction could manifest as destabilisation of the emulsion. This involves an increase in the diameter size of the propofol oil droplets that ultimately leads to coalescence and phase separation. An early sign of emulsion destabilisation is an increase in droplet size in the large diameter tail of the size distribution [1]. Droplets that reach a diameter of 5 μm or more have an increased tendency to become entrapped in small capillaries, for instance those found in the lungs or the brain, and they may cause embolism [2, 3, 4, 5]. Entrapment in the liver, the reticuloendothelial system and several other organs has also been reported [6, 7]. In addition, incompatibilities have been suspected of having a deleterious immunomodulating effect [8, 9].

The available propofol formulations differ in lipid and excipient composition (Table 2). Our hypothesis is that different propofol formulations may behave differently upon contact with remifentanil.

In clinical practice, infusion flow rates are constantly adjusted to achieve the desired clinical response. This creates an ever-changing relationship between the drugs that are co-administered via the same i.v. line. It is therefore important to examine these drugs in mixing ratios that mimic the relationships that might occur in the i.v. line during infusion. To the best of our knowledge, there is a lack of thorough compatibility studies comparing propofol formulations after mixing with remifentanil [10, 11, 12]. In the case of propofol, an incompatibility can be revealed by a change in the emulsion droplet size. An increase in the content of large sized oil droplets (>5 µm) may become critical (exceed the acceptable safety limit) long before it is possible to visually detect emulsion destabilisation.

The aim of this study was to determine and compare the emulsion stability of the three propofol formulations currently used in Norwegian hospitals (Propolipid^® (P-MCT1), Propofol-Lipuro^® (P-MCT2) and Diprivan (P-LCT)) when administered together with remifentanil in the same i.v. line.

Material and methods

Results

The background noise in all tested centrifuge tubes and containers both filled with Milli-Q water were very low: 1.0 ± 0.9 particles/mL measured immediately after filling and 2.5 ± 2.4 particles/mL measured after 4 hours. These values were well below the limit of background noise of 50 particles/mL, and demonstrate that the droplets analysed in this study had minimal risk of being influenced by particle contaminants originating from the water or the containers.

PFAT5

All propofol control samples (unmixed propofol), showed PFAT5 values below the 0.05 limit both immediately and after 4 hours (Table 3). This indicated that all three propofol formulations were stable throughout the test period.

In all samples of propofol 10 mg/mL where remifentanil was added in abundance, PFAT5 values were above the recommended 0.05 limit 4 hours after mixing (long contact time); this happened immediately after mixing in samples of propofol 20 mg/mL irrespective of the mixing ratios of remifentanil and propofol (Table 3).

Principal component analysis (PCA) showed that PFAT5 displayed a positive correlation with remifentanil content in the mixing ratio (high value of ratio and high PFAT5), time after mixing (high value of time and high PFAT5) and negative correlation with pH-values of the sample (high value of pH located on the opposite side of origo from PFAT5) (Figure 1(A)). Propofol concentration was somewhat correlated with PFAT5, but to a

Figure 1:

Results from a principal component analysis (PCA) of PFAT5 (percentage of fat residing in globules larger than 5 µm) for all mixtures of remifentanil with the various propofol formulations. (A) correlation loading plot showing the correlation between the factors. (B) score plot showing the projections of the samples.

lesser extent than the other factors as shown by the loading located much closer to origo on the first two principal components. The distribution of samples in the score plot showed three distinct groups along the first principal component (PC1). The group with the highest positive values on PC1 (right end of the x-axis) displayed a positive correlation with PFAT5, whereas the group with the highest negative values (left end of the x-axis) displayed a negative correlation with PFAT5. This group displayed a positive correlation with the pH values of the samples. It should be noted that all propofol formulations are present in all three groups; however, an overrepresentation of P-MCT2 (green color in Figure 1(B)) might be seen in the group with high PFAT5-values.

The influence of each variable on PFAT5 was analysed by PLS regression (Figure 2). The regression coefficients from PLS models of all factors containing all samples (full matrix) showed that propofol concentration, mixing ratio, mixing time and pH had a significant influence on PFAT5 (Figure 2(A)). High propofol concentration (20 mg/mL), long contact time (4 h after mixing) and low pH-values (high amounts of remifentanil) were positively correlated with high PFAT5 values.

Another interesting finding was that P-LCT showed a negative correlation with PFAT5 values whereas P-MCT2 showed a positive one. No significant correlation between P-MCT1 and PFAT5 values was identified.

Since long contact time between remifentanil and propofol seemed to have a detrimental effect on the PFAT5, it was interesting to investigate the data from the immediate time point separately. Therefore, data obtained 4 h after mixing were left out and a

Figure 2:

Balanced and weighted (BW) regression coefficients from partial least squares (PLS1) regression analyses of investigated variables on the measured PFAT5 (percentage of fat residing in globules larger than 5 µm) in mixtures of propofol and remifentanil. Significant factors are those with error bars not crossing zero. Propofol formulation (P-MCT1 = Propolipid, P-MCT2 = Propofol-Lipuro, P-LCT = Diprivan), P conc = propofol concentration, Ratio = mixing ratio of remifentanil and propofol, Time = time after mixing propofol with remifentanil, determined pH. (A) Full data matrix. (B) All formulations but only data immediate after mixing, i. e. excluding measurements of PFAT5 4 hours after mixing. (C) All formulations of 10 mg/mL propofol. (D) All formulations of 10 mg/mL propofol immediate after mixing i. e. excluding measurements of PFAT5 4 hours after mixing.

new PLS model was made for the PFAT5 measured immediately after mixing. The regression coefficients showed the same correlations with PFAT5 as before (Figure 2(B)), even though most of the PFAT5 values for immediate samples were within the 0.05 limit (Table 3). Due to the identified correlations of high propofol concentration and long contact time with high PFAT5, PLS analyses were performed on the data set of propofol 10 mg/mL separately (Figure 2(C)), and propofol 10 mg/mL immediately after mixing separately (Figure 2(D)). Similar trends of the regression coefficients were again seen to correlate to high PFAT5.

Discussion

Propofol is very slightly soluble in water, and is, therefore formulated as an oil-in-water emulsion. Injectable emulsions are composed of oil droplets in water stabilized by one or more emulsifiers. The droplets have a charged surface securing the necessary repulsion between the droplets, thereby, avoiding coalescence and phase separation [18]. Emulsions are sensitive to physicochemical changes, such as changes in pH, temperature or variation in electrolyte contents (e. g. oppositely charged counter ions). In a destabilised emulsion, an increase in droplet size will occur [14, 19, 20]. Since intravenous infusion of enlarged droplets encompass potentially serious risks, droplet stability is essential for injectable emulsions [2, 3, 4, 5, 6, 7, 8, 9].

Knowledge on propofol emulsion stability when co-administered with remifentanil is sparse. None of the few studies that have been conducted is able to show conclusive evidence about the compatibility between propofol and remifentanil or lack of such [10, 11, 12]. No studies have analysed both 10 mg/mL and 20 mg/mL propofol even though both concentrations are used regularly in the clinical setting. Neither have a direct comparison of different propofol formulations been carried out.

To capture possible changes in emulsion stability we used different analysis methods, including dynamic light scattering, measurements of pH and zeta potential, and analyses of size and number of large oil droplets using light obscuration followed by calculation of PFAT5. Even though significant differences in MDD could sometimes be identified between the control and mixed samples, despite all being well below 500 nm, this should not be directly interpreted as incompatibility. The mean droplet size is not the best parameter to assess changes in stability due to droplet growth [21]. As mentioned above, the first signs of droplet growth appears in the large diameter tail of the size distribution, and a considerable amount of large droplets will have to be formed before it can be recognized as an increased MDD. In the current study, MDD and PI were measured to keep an eye on the bulk properties, which was useful since small differences could be observed between different batches of the same formulation, but PFAT5 is the parameter that provides the information on the large diameter tail region. Defined as the percentage of fat droplets larger than 5 µm, increased PFAT5 values suggest destabilisation of emulsion. According to USP <729> PFAT5 values should lie below 0.05 (i. e. 5 % of droplets below 5 µm in size) to ensure emulsion stability, from an infusion safety perspective [1].

A slight decrease in pH-values was observed for most samples after 4 hours contact time. It is known from literature that lipid emulsions, when exposed to air, may undergo oxidative processes releasing free fatty acids and lowering pH [22]. Steger and Mühlebach showed that emulsions based on LCT and LCT/MCT degrade at different rates under light exposure at room temperature [22]. Therefore, lipid phase stability should not be forgotten when evaluating emulsion stability.

The experimental area was spanned out with the aim of making the results as clinically relevant as possible. The maintenance dose of propofol is 5–150 μg/kg/min and that of remifentanil is 0.008–0.25 μg/kg/min. This is in concordance with local guidelines at Oslo University Hospital, Norway [23], the American College of Critical Care Medicine’s revised guideline “Clinical practice guidelines for the management of Pain, Agitation and Delirium (PAD)” [24], and information provided in the section General anesthesia: maintenance, in UpToDate [25]. This gives a wide range of possibilities of mixing ratios when adjusting the dose according to the response of the patient. To also accommodate for the possible scenario of bolus doses, the mixing ratios spanned from 20+1 to 1+20 for remifentanil: propofol. This was discussed with local clinicians, taking anesthesiology both during surgery and at intensive care into account.

The infusion rates of the drugs are determined by the desired level of sedation and the contact time of the drugs in the i.v. line is therefore difficult to predict. We therefore, analysed all samples immediately after mixing and 4 hours after mixing remifentanil and propofol. We did not identify any values of PFAT5 above the recommended limit of 0.05 immediately after mixing with remifentanil for all three formulations at a propofol concentration of 10 mg/mL, but elevated PFAT5 values were seen in most samples after 4 hours (Table 3). This is indicative of emulsion destabilisation at long contact times, which can lead to adverse implications for the patient during surgery that last longer or in cases where the infusion rates are very slow, e. g. for children. In fact, the infusion rates in paediatric patients might be much longer with contact times of the drugs exceeding the experimental conditions covered by this study.

For all three formulations, the high propofol concentration (20 mg/mL) was more prone to the formation of larger droplets after mixing with remifentanil (Table 3). This phenomenon was even more marked after 4 hours, but also significant immediately after mixing (Figure 2). The direct correlation of PFAT5 with propofol concentration, and samples where remifentanil was in abundance and pH was low (i. e. high ratio of remifentanil) indicated that these factors influence the stability of propofol emulsions. Gersonde et al. also reported high PFAT5 values for two MCT-based propofol formulations (Propofol-Lipuro and Propofol MCT) at 20 mg/mL upon mixing with remifentanil [11]. They measured the samples 24 hour after mixing. Nemec et al. found an aggregate formation immediately after mixing equal parts remifentanil either with Propofol LCT 2 % (20 mg/mL) or Propofol MCT 1 % (10 mg/mL) and recommended not to mix remifentanil directly with propofol [10].

Our regression models suggested that the different propofol formulations responded differently to mixing with remifentanil. It seemed like the P-MCT2 formulation (Propofol-Lipuro) was positively correlated with PFAT5 and associated with high values of PFAT5 immediately after mixing with remifentanil and also after 4 hours whereas this was not the case for the other MCT-formulation (Propolipid) or the LCT-formulation (Diprivan) (Figure 2(A–D)). This deviates from the reports from the manufacturer B.Braun, as they did not see any sign of emulsion destabilisation in their analysis of their MCT-formulation (Propofol-Lipuro 1 %) mixed with remifentanil performed after one hour storage. Their mixing ratios were 10+1, 1+1 and 1+10 [26]. For P-LCT we identified high PFAT5 values for both concentrations but only 4 hours after mixing with remifentanil (Table 3). Humbert-Delaloye et al. reported no sign of emulsion destabilisation when visually inspecting Disoprivan (LCT) 20 mg/mL mixed with remifentanil in glucose (50 mg/mL) in the concentration 0.12 mg/mL [12], but visual examination is not an accurate method to identify early signs of emulsion destablilisation . Even though our findings suggest that different propofol formulations can react differently upon mixing with remifentanil, the difference in the lipid phase alone (MCT vs LCT) could not explain these differences; the complete formulation and/or the manufacturing process are the most likely determining factors.

We did not see any high values of PFAT5 in the mixing ratio of 1+20, i. e. propofol in abundance, neither immediately nor 4 hours after mixing. In intensive care units, propofol 20 mg/mL is frequently used, and at this concentration, we found high PFAT5 values for all three propofol formulations, both immediately and after 4 hours, in mixing ratios 1+1 and 10+1, that is when remifentanil was present in equal parts or in abundance. This demonstrates that high propofol concentrations are more sensitive to reactions with remifentanil and caution is advised when remifentanil is present in equal parts or in abundance.

Stewart et al. concluded that from a chemical stability perspective of the drugs, it was safe to mix remifentanil and propofol in a polypropylene syringe and store for up to 36 hours, but mixed in the same polyvinylchloride bag the mixture should be used within 3 hours [27]. Our results showed that from a physical compatibility perspective it is not safe to store remifentanil mixed with propofol in the same container, such as premixing in a syringe, since both long contact time and mixing ratios of remifentanil>propofol increased the risk of emulsion instability. In fact, the importance of always including an emulsion stability test in compatibility studies involving propofol was emphasized by Cherin and Smiler in a letter to the editor as a response to the study of Stewart et al. [28]. If premixing remifentanil and propofol is desired, propofol 10 mg/mL should be used, the ratio of remifentanil to propofol should be dominated by propofol, and the mixture should be utilised within 4 hours, as we lack stability data beyond 4 hours contact time.

Our findings need to be interpreted with the following limitations and strengths in mind: The samples had to be strongly diluted in order to be analysed. This may have affected the physicochemical properties of the emulsions since flocculates/aggregates that otherwise would gradually coalesce and destabilise the emulsion may have been broken up by the dilution process. However, the control samples of propofol alone were diluted to the same extent. Another limitation is that by performing sample preparation and analysis in the lab environment and not in a classified clean room, the risk of particle contamination from the surroundings as well as from the equipment could not be entirely eliminated, although particle content in the MilliQ-water and main containers were checked. Particles, if in the right size range, could be interpreted as emulsion droplets. The clinical relevance of a rigid cut off limit of 5 µm could be discussed [29]. Oil droplets are more flexible and malleable than solid rigid particles, such as precipitates of crystals, and might be better tolerated by the human body. Furthermore, the 5 % threshold of the acceptable amount of enlarged droplets can also be discussed, and it might be worth mentioning that the PFAT5 monography of the USP does not have an equivalent in the European Pharmacopeia. Emulsion stability was only analysed at two time points, immediately after mixing and 4 hours after mixing, and there will be clinical scenarios that were not accounted for in this study. Infusion rates, especially for the paediatric population, could exceed 4 hours, and as shown, extended contact times could lead to emulsion destabilisation. Last, but not least, our study was set up to simulate a Y-site administration of remifentanil and propofol and could not be directly converted to the dynamic situation when drugs are co-administered via the same i.v. line.

Conclusions

Differences in emulsion stability depending on the propofol formulation were identified under the most extreme test conditions in our study (long contact times and excess of remifentanil in the mixing ratio). The differences could not be explained by type of propofol lipid phase alone. Nevertheless, all tested propofol formulations were found to be stable after mixing with remifentanil in ratios where propofol was in abundance. Our results indicate that co-administration of remifentanil and propofol 20 mg/mL in the same i.v. line when remifentanil is present in equal parts or in abundance should preferably be avoided. Based on the results after 4 hours contact time, i. e. development of larger droplets and a drop in pH, premixing of propofol and remifentanil in syringes or as admixtures should, if possible, be avoided. Since destabilisation of the emulsion may increase the risk of large fat droplet infusion, we recommend taking differences in emulsion stability between propofol formulations into account when switching between different propofol products.