Buccal Absorption of Fentanyl Is pH-Dependent in Dogs 

The study was approved by the Institutional Animal Care and Use Committee of the University of Utah. Six mongrel dogs, weighing 22–32 kg, served as the study population. Each dog was studied seven times, once with intravenous fentanyl and six times with fentanyl in pH-buffered solutions that were administered through the buccal mucosa. For each dog, studies were separated by at least 1 week and not more than 2 weeks.

After a 12-h fast, each dog was sedated with intramuscular Telazol (7 mg/kg). A catheter was inserted in a forearm vein, and Ringer's lactate was infused at 2 ml/kg/h to maintain fluid homeostasis. Anesthesia was induced with 10 mg/kg thiopental. After tracheal intubation, the lungs were mechanically ventilated. Anesthesia was maintained with halothane in oxygen such that the mean arterial pressure was maintained within 20% of preanesthesia values. A 20-G catheter was inserted percutaneously into the femoral artery for blood sampling. Arterial P^CO2was maintained between 35 and 40 mmHg by ventilator adjustment after frequent analysis of arterial blood gases.

Each dog was given 100 micro gram/kg fentanyl (in the form of citrate salt in solution) by intravenous injection over 30 s. Arterial blood samples were taken as follows: At baseline before fentanyl administration; at 1, 3, 5, 7, 10, 15, 20, 30, 45, 60, 90, and 120 min; and hourly thereafter for 12 h. Blood samples were injected into preheparinized glass tubes and immediately placed on ice. Plasma was separated from erythrocytes with a refrigerated centrifuge, placed in polypropylene tubes, and frozen at -20 degrees Celsius until the time of assay.

A specially constructed Teflon cell was used to deliver the fentanyl solution in all transmucosal studies. This trapezoidally shaped cell was clamped to the dog's buccal mucosa such that the mucosa formed the base of the cell and was directly exposed to the fentanyl solution placed into the cell (Figure 1). The tension on the clamp was adjusted (by direct vision) so that no solution leaked underneath and out the walls of the cell. The surface area of the mucosa at the base of the cell was 18 cm².

Fentanyl solutions of pH 6.6, 7.2, and 7.7 were made from dry fentanyl citrate powder (Mallinckrodt, Glen Falls, NY) and phosphate buffer to a concentration of 1 mg/ml. At the start of each transmucosal study, 4 ml of one of the above fentanyl solutions was placed into the cell clamped to the dog's buccal mucosa. Thus, the dose^transmucosalof 4 ml *symbol* 1 mg/ml *symbol* 0.64 (factor used to convert fentanyl citrate (molecular weight 528.60) to fentanyl base (molecular weight 336.46) was 2.55 mg. After 60 min. the residual solution was removed, the buccal mucosa at the base of the cell was rinsed with saline, and the clamp was removed from the mucosa and taken out of the dog's mouth. Arterial samples were taken at baseline before placement of the buccal cell; at 3, 6, 10, 15, 20, 30, 40, 50, 60, 65, 70, 75, 80, 85, 90, 100, 110, and 120 min; and every hour thereafter for 8 h. To calculate permeability coefficients, 0.1-ml aliquots were taken from the fentanyl solution in the buccal cell at 0 and 60 min and assayed for fentanyl.

Fentanyl concentrations were determined by radioimmunoassay using the technique described by Schuttler and White. [11] The assay was sensitive to 0.2 ng/ml with a coefficient of variation of 3.8% at 0.2 ng/ml and 4.7% at 5.0 ng/ml.

For the transmucosal delivery portion of the study, the peak plasma concentration (C^max) and its time of occurrence (T^max) were noted from the plasma concentration versus time profile. The terminal slope for each curve was determined using log-linear regression of the terminal portion of each curve, which was identified visually. The terminal elimination half-life of fentanyl was calculated from this terminal slope (In 2/terminal slope). The area under the plasma fentanyl versus time curve after intravenous and transmucosal administration (AUC sub iv and AUC^transmucosal, respectively) was calculated from the time of administration of fentanyl to the last measurable plasma concentration using the linear trapezoid method. [12] Extrapolation of the AUC from the time of the last measurable fentanyl concentration to infinity was calculated by dividing the last plasma concentration by the first-order rate constant of the terminal phase of the profile. The sum of these two components was the estimate of the total AUC.

The dose-adjusted bioavailability for each transmucosal study was calculated according to the formula: Equation 1.

The permeability coefficient for each transmucosal study was calculated by: Equation 2where A = the area of mucosa (18 cm²) exposed to fentanyl solution, t = the duration of drug delivery (3600 s), and c = the average concentration of fentanyl in the delivery solution at 0 and 60 min.

Several parameters derived from the pharmacokinetic analysis, C^max, bioavailability, and permeability coefficient, were considered variables for a two-stage analysis. The programs 11) and 5V from BMDP Statistical Software (1990 DEC/ULTRIX version, Los Angeles, CA) were used for analysis. The effect of pH on the variables C^max, bioavailability, and permeability coefficient was modeled as an unbalanced, repeated measures multivariate analysis of variance with two within factors, pH (three levels, 6.6, 7.2, and 7.7) and replicates (two levels, two trials at each pH); interactions between pH and replicates were included in the model; and pH was treated as a linear contrast. As there was no improvement in Akaike's information criterion with higher order variance structures, a compound symmetry variance structure (equal variance and equal correlation) was assumed. Estimation of the parameters used the restricted maximum likelihood algorithm. Statistical significance was asserted for P < 0.05.

(Figure 2) illustrates the plasma fentanyl versus time relationship that developed after fentanyl solutions of pH 6.6, 7.2, and 7.7 were placed into the cell adherent to the dog's buccal mucosa. Plasma fentanyl became detectable by 6 min at all three pHs. Absorption of fentanyl from the buccal solution at pH 7.7 resulted in the most rapid increase in plasma fentanyl and the highest peak fentanyl concentration (C^max). Figure 3displays the decay of fentanyl from the plasma to demonstrate the similarity in terminal half-life (mean plus/minus SD) that existed with all three pHs (pH 7.7, 205 plus/minus 89 min; pH 7.2, 205 plus/minus 65 min; pH 6.6, 196 plus/minus 48 min). The mean terminal half-life after intravenous administration was 244 plus/minus 68 min and not statistically different from the terminal half-lives determined after buccal administration.

The mean C^maxat pH 7.7 was nearly three times that of the mean C^maxat pH 6.6 (Figure 4). Similar relationships (increasing values with increasing pHs) were found for bioavailability and permeability coefficient (Figure 4). In all buccal studies regardless of pH, T^maxoccurred within 10 min of removal of the fentanyl solutions from the buccal mucosa (Figure 2).

The results from our study indicate that fentanyl can be absorbed through the oral mucosa more rapidly and to a greater extent by increasing the pH of the delivery solution applied to the oral mucosa. This finding is in agreement with the classic studies of Walton, [10] who found that drug passage across biologic membranes is dependent on both the pH and lipid solubility of the drug. [10] In theory, a low degree of ionization and a high oil-water partition coefficient favors maximal transmucosal absorption. Fentanyl, a weak base with a pKa of 8.2 becomes more unionized as the surrounding pH is increased. Thus, fentanyl is 2.45% unionized at pH 6.6, 9.1% unionized at pH 7.2, and 24% unionized at pH 7.7. At first, we attempted to choose a wider range vehicle pH to test our hypothesis. However, a pH less than 6.6 does not significantly affect the degree of ionization. In addition, it is impossible to dissolve 1 mg/ml fentanyl into an aqueous solution at a pH greater than 7.7. Fortunately, the spread from pH 6.6 to 7.7, spanning a tenfold difference in unionized fentanyl concentrations, was clearly sufficient to prove the hypothesis.

Although the unionized concentration of fentanyl increased tenfold over the pH span tested, the permeability coefficient only increased approximately five-fold. There are two possible explanations for this discrepancy. First, boundary layer control may exist. Boundary layer control occurs when the permeability coefficient of the mucosa is so high that the overall permeation process is rate-limited by diffusion of the drug in the layer of the vehicle solution adjacent to the mucosa. [9] This effect is only significant when the mucosal permeability coefficient is high, which is the case here. Indeed, the transmucosal permeability coefficient of fentanyl obtained in this study is among the highest permeability coefficients reported in the literature (i.e., tritiated water 0.26–0.50 x 10 sup -4 cm/s, flurbiprofen 2.7 x 10 sup -4 cm/s). [9,13] Alternatively, increasing ionization and, thus, the lipid solubility of fentanyl may not equally increase transmembrane permeability because biologic membranes have both lipid and aqueous domains. In other words, some fentanyl absorption may occur through aqueous channels as well as lipophilic pathways. This biphasic relationship has been demonstrated in both spinal meninges [14] and arteries [15] and is likely to apply to the buccal mucosa.

Weinburg et al. found that the sublingual absorption of methadone, an opioid with a pKa of 9, was improved from 35% to 75% by increasing the pH of the carrier solution from 6.5 to 8.5. [16] However, changing the pH to improve transmucosal absorption is not effective with all opioids. In Weinburg et al.'s study, the sublingual absorption of hydromorphone was not changed by a higher pH. [16] This finding was probably due to the limited lipid solubility of both the ionized and unionized forms of hydromorphone.

The terminal half-life of fentanyl after intravenous administration was similar to the terminal half-life after transmucosal administration. This result implies that there is no long-term storage of fentanyl in the buccal mucosa after transmucosal delivery. This is in contrast to transdermal fentanyl delivery, wherein the terminal elimination half-life is markedly prolonged because of the continued absorption of fentanyl from a depot in the skin long after the delivery system is removed. [17].

The dog was used to investigate transmucosal fentanyl absorption because the histology and drug transport properties of the dog's buccal mucosa are similar to those of human buccal mucosa. Both flurbuprofen and diclofenac sodium are absorbed with similar permeability coefficients through canine and human buccal mucosa. [9].*

Can the results of our study be related to clinical practice? The variability in fentanyl absorption from oral transmucosal fentanyl citrate (OTFC) may result, in part from variations in mouth pH. [4] The pH of saliva normally varies from 6.5 to 7.5. [18] Administration of OTFC after exposure of the oral mucosa to a pH-raising antacid or a pH-lowering citrus drink may result in exaggerated or reduced effects, respectively. Fentanyl is administered transmucosally by two methods: by sucking on a lozenge of OTFC' (a solid dosage form) or with a nasally administered metered-dose inhaler of the commercially available solution. [3] In both routes of administration, however, there is a 20–30-min lag from administration to peak clinical effect. Faster fentanyl absorption and faster onset of clinical effect may occur by increasing the pH of the saliva with an NAHCO⁴tablet (for oral transmucosal administration) or buffering the commercially available fentanyl solution to a higher pH (for nasal administration). The clinical impact of these ideas remains to be evaluated.

*Ebert C., John V, Beall R, Rosenzweig K: Transbuccal absorption of diclofenac sodium in a dog model, Controlled Release Technology. Symposium Series. Edited by Lee PI, Good WR, Atlanta, American Cancer Society, 1987, p 348.