Dermal Penetration of Fentanyl: Inter- and Intraindividual Variations

Transcription

1 C Pharmacology & Toxicology 2003, 93, Printed in Denmark. All rights reserved Copyright C ISSN Dermal Penetration of Fentanyl: Inter- and Intraindividual Variations Rikke H. Larsen 1, Flemming Nielsen 1, Jens A. Sørensen 2 and Jesper B. Nielsen 1 1 Environmental Medicine, University of Southern Denmark, and 2 Department of Plastic Surgery, University Hospital of Odense, Denmark (Received May 21, 2003; Accepted September 9, 2003) Abstract: Fentanyl is a potent synthetic opioid that is increasingly being used in transdermal drug delivery systems. The target organ concentration of a drug administered dermally will depend on the rate of dermal absorption and the systemic elimination. We have studied the intra- and interindividual variation in dermal penetration of fentanyl in an in vitro model (static diffusion cells) with human skin, and compared the absorption of fentanyl from an aqueous solution with absorption from a commercial patch. The intraindividual variation in dermal penetration of fentanyl in aqueous solution was limited (18%) and no differences in penetration characteristics were observed between breast and abdominal skin. The interindividual variation in dermal penetration of fentanyl was extensive, with maximal fluxes ranging from ng/ cm 2 /hr following application of an infinite dose of fentanyl to the donor chamber. Use of transdermal drug delivery systems (patches) reduced the inter-individual variation. The permeability coefficients after application of fentanyl in aqueous solution and through patches were identical ( cm/hr). One person had a higher than average penetration rate following patch application, which may indicate that the human skin and not the patch barrier was the rate-determining factor for the other individuals included in this study. Fentanyl is a potent synthetic opioid agonist. During recent years, fentanyl administered transdermally has gained appreciation in the postoperative treatment and as an important drug in palliative treatment for late stage cancer and other patients with chronic pain (Cherny et al. 1995; Simpson et al. 1997; Dellemijn et al. 1998; Collins et al. 1999; Jakobsson & Strang 1999; Ringe et al. 2002). Fentanyl has a low molecular weight of 336 g/mol, which is required for transdermal use ( 1000 g/mol). It has an adequate solubility in oil as well as in water given by a high lipophilicity (octanol/water coefficient of 717). These properties make it 1000 times more permeable than morphine (octanol/water coefficientω0.7) (Roy & Flynn 1988). It has also been reported to be times as potent as morphine (Grond et al. 2000). Thus, the physicochemical properties of fentanyl makes it highly qualified as a transdermal drug and means that a smaller amount has to cross the skin barrier to be effective and only a small skin area has to be occluded (Roy & Flynn 1989). Fentanyl has a bioavailability close to 1 (0.92) which means that the drug when administered transdermal is neither significantly degraded by the skin s micro-flora nor by the cutaneous enzyme metabolism (Varvel et al. 1989; Roy et al. 1994). This makes an experimental model using in vitro penetration suitable for a study of dermal penetration of fentanyl (Roy et al. 1994). Fentanyl is rapidly eliminated from the plasma by met- Author for correspondence: Jesper B. Nielsen, Environmental Medicine, University of Southern Denmark, Winsløwparken 17, DK-5000 Odense C, Denmark (fax π , jbnielsen/health.sdu.dk). abolism in the liver. It undergoes phase I metabolism, predominantly by oxidative N-dealkylation (McClain & Hug 1980). Fentanyl is known as a short-acting drug with a short half-life. Thus, when fentanyl is given through intravenous injections it is almost totally (98.6%) eliminated from the blood stream (central compartment) within only 60 min. The terminal elimination phase, however, is very slow with an average half-life of 219 min. (range 141 min. 853 min.) (McClain & Hug 1980; Reilly et al. 1985). Transdermal patches deliver a certain amount of drug per hour. The amount of drug is rate-limited and poisoning of the patient should not be possible if the patch is undamaged and placed correctly. However, there have been reports of patients suffering from a fatal overdose due to high serum concentrations of fentanyl achieved from a transdermal delivery patch. These reports describe severe hypoventilation and death due to transdermal delivery of fentanyl (Grond et al. 2000). Whether this is caused by an accidental or intentional high exposure (Roberge et al. 2000) or due to the inherent heterogeneity of man is not known. This heterogeneity may render some individuals more susceptible than others due to unexpected high dermal absorption rate or a poor capacity to metabolise the drug. Drugs applied on the skin must initially penetrate the avascular and lipophilic structure (stratum corneum) and continue through the more aqueous lower layers (lower epidermis and dermis) to the blood vessels. A compound that is only lipophilic will easily cross the stratum corneum but the penetration rate will decrease as it reaches to the hydrophilic layers below, and the diffusion of the drug will be slower. This process is called a reservoir effect. A drug that is soluble in the lipophilic layer as well as in the more aqueous structures,

2 DERMAL PENETRATION OF FENTANYL 245 and at the same time has a small molecular size, has the best permeability through the skin barrier (Guy et al. 1987). Experimental evidence demonstrate that dermal absorption will mainly depend on the uppermost skin layer, the stratum corneum (Roy & Flynn 1990), but multiple features such as type of chemical, application site and skin condition may also affect dermal absorption (Bronaugh & Maibach 1985; Varvel et al. 1989; Roy et al. 1994). The aims of this study were 1. To assess the inter-individual as well as intra-individual variation in skin absorption of fentanyl, 2. to compare penetration across abdominal and breast skin from the same individual, and 3. to describe differences between absorption of an aqueous fentanyl solution and fentanyl absorption from a transdermal delivery system. Materials and Methods Fentanyl was either used as a citrate salt dissolved in water (50 mg/ ml) or as patches delivering 25 mg fentanyl/hr (Durogesic A ; Jansson- Cilag). Experimental model. The Franz diffusion cell (Franz 1975) was used to study the absorption of fentanyl. The system that has previously been used by the laboratory (Nielsen 2000; Nielsen & Nielsen 2000) consisted of two half-cells where the upper cell compartment represents the donor chamber and the lower the receptor chamber. The skin sample dividing the two cells was placed on a metal grid and held in place by a clamp, which at the same time kept the half-cells together. The cells were kept at a constant temperature (32 æ) in a water bath with individual magnetic stirring of the receptor chamber. The median diffusion area was 2.12 cm 2 /cell and the receptor chamber volume 17.2 ml. Skin. Human skin was obtained from the Department of Plastic Surgery at the University Hospital in Odense. Skin was sampled from women (18 60 years old) that underwent breast reconstruction or abdominal plastics. The donors were given complete anonymity with registration of age of the female donors only, and the study was approved by the regional ethics committee. Skin samples were kept at ª20 æ for periods not exceeding 12 months. This has proven to keep the skin fresh and give no significant change in the water permeability (Bronaugh et al. 1986). The skin was allowed to thaw at room temperature for one hour before removal of subcutaneous fat. The skin was cut into suitable pieces to fit the static diffusion cells adapted from Southwell et al. (1984). Full-thickness skin with an average thickness close to 1 mm was used. Experimental set-up. When skin samples had been mounted between the donor and the receptor chamber, the receptor chamber was filled with phosphate buffer (0.05 M Na 2 HPO 4,2H 2 O (Merck, Germany); phω7.4). The barrier integrity of the skin was evaluated after 1 hr of equilibration by capacitance measurements. Skin samples with a high capacitance are unable to act as a capacitor, which means that the skin is damaged. The integrity of the barriers was evaluated before as well as after the experimental period. Cells with a capacitance above 55 nf were excluded from the experiment. Experiments began when buffer in the donor chamber was replaced by 1 ml fentanyl solution (50 mg/ml) or when Durogesic patch delivering 25 mg/hr was placed between the two half-cells on top of the skin membrane in the donor chamber. Thus, the area available for penetration was identical in all experiments. At specified time intervals throughout the experiment, the buffer solution in the receptor chamber was sampled using a 2 ml syringe. The sample volume was replaced by an equal volume of new phosphate buffer. The samples from the receptor chamber were then analyzed for fentanyl. Fig. 1. Cumulative absorption of fentanyl (50 mg/ml in water) during 48 hr. Six diffusion cells set up with human full thickness breast skin from the same person. The donor chambers were covered with Parafilm A throughout the experiment to avoid evaporation and to resemble the humidity under a transdermal delivery system. Analysis of fentanyl. The fentanyl concentration in the receptor fluid was assayed by HPLC. The chromatographic system was produced by Kontron (Bio-tek instruments, Milano, Italy) and consisted of a HPLC 420 pump, a HPLC 360 auto sampler and a HPLC 430 UVdetector. System control and data handling was performed on a personal computer equipped with Kontron MT 450 software. Detection was performed at 205 nm, an absorbance range of and a response time of 2 sec. Separation was performed on a LiChrospher 60 RP-selectB (5 mm) mm column equipped with a RP-selectB (5 mm) 4 4 guard column (Merck, Darmstadt, Germany). The mobile phase consisted of 0.23% perchloric acid in Milli-Q water: acetonitrile (60:40 vol/vol) (Lambropoulos et al. 1999) and was used with a flow of 1.5 ml/min. All samples were analysed in duplicates, and the average concentration reported. The limit of detection was 10 ng/ml. Standard solutions were prepared in water. Standard curves were produced each day of analysis and covered a range from 0 to 5000 ng/ml. Correlation coefficients of the standard curves were All organic solvents were purchased from Riedel-de Haën and were of CHROMASOLV A purity (Sigma-Aldrich Laborchemikalien, Seelze, Germany), perchloric acid was o pro-analysis quality and was provided from Merck (Darmstadt, Germany). Data was plotted as the cumulative amount of drug collected in the receptor compartment as a function of time. The absorption rate was determined by linear regression showing the cumulative amount of fentanyl absorbed per cm 2 of skin sample per time. The permeability coefficient (kp) was calculated by dividing the absorption rate by the applied concentration. Fig. 2. Absorption of fentanyl (ng/cm 2 ) during 48 hr across breast skin from 21 women.

3 246 RIKKE H. LARSEN ET AL. Fig. 3. Cumulative amount (%) of fentanyl absorbed across breast skin (white column) and abdominal skin (grey column) in three different individuals. Fig. 4. Median cumulative absorption of fentanyl across breast skin from transdermal delivery systems (nω8). Results Intra-individual variation. The dermal penetration of fentanyl across breast skin after 48 hr varied between 1400 ng/cm 2 and 2400 ng/cm 2 (fig. 1). Likewise, the intraindividual variation in maximal flux obtained between 24 and 48 hr was 18% (45 ng/cm 2 /hr to 72 ng/cm 2 /hr). Interindividual variation. The cumulative absorption of fentanyl in 21 individuals after an experimental period of 48 hr varied between approximately 500 ng/cm 2 and 4250 ng/cm 2 (fig. 2). Some individuals had a slow and almost negligible absorption, while others had a fast penetration within the same period. The majority of the individuals absorbed between 1400 ng/cm 2 and 2600 ng/cm 2 of fentanyl within 48 hr. However, two individuals (10%) absorbed around 4250 ng/cm 2 within 48 hr and three (15%) absorbed only around 600 ng/cm 2 in the same period (fig. 2). The average flux between 24 and 48 hr was 61 ng/cm 2 /hr with a standard deviation close to 24 ng/ cm 2 /hr. The large standard deviation is mainly due to the two individuals with an average flux of 105 ng/cm 2 /hr and the three donors with very low penetration rates (approximately 21 ng/cm 2 /hr). Differences between breast and abdominal skin. One person had the highest penetration of fentanyl across abdominal skin, whereas the other two had the highest penetration across breast skin (fig. 3). The variation in ab- sorption of fentanyl between the individuals was small with the lowest value being 12.8% and the highest 19.7%. In the three donors included in this experiment, we did not observe any differences regarding flux or permeability coefficients of fentanyl (table 1). Patients undergoing abdominal as well as breast receonstruction during one surgical process are rare, and results were based on only three individuals with 2 3 samples from each skin area. It is therefore not possible to make meaningful statistical comparisons between absorption through breast and abdominal skin. Interindividual patch absorption. The cells containing the delivery system with patches demonstrated a high rate of fentanyl absorption. The individual absorption varied between 56 mg/cm 2 and 87 mg/cm 2 in 72 hr (fig. 4). Seven out of eight individuals absorbed between 56 mg/cm 2 and 69 mg/cm 2, whereas one person had a 30% larger absorption. The amount of absorbed drug was 7.4% at 48 hr and 11.5% at 72 hr with an almost identical flux of 1300 ng/cm 2 /hr and a permeability coefficient of cm/ hr (table 2). Discussion Intraindividual penetration. Variation in results obtained from a single cell will reflect experimental variation as well as variations in penetration Table 1. Flux and permeability coefficients for dermal penetration of fentanyl across breast and abdominal skin from three donors. An infinite dose of 50 mg fentanyl in 1 ml was applied to the donor chamber and the experimental period was 72 hr. Flux Permeability coefficient (ng/cm 2 /hr) (cm/hr) Breast Abdomen Breast Abdomen Donor Donor Donor Median Table 2. Absorption, flux, and permeability coefficients for dermal penetration of fentanyl across breast skin after 48 and 72 hr exposure (nω 8). Fentanyl was applied as a commercial patch (Durogesic) on top of the skin. Results are given as medians with range. Exposure Permeability time Absorption Flux coefficient (hr) (%) (ng/cm 2 /hr) (cm/hr) ( ) ( ) ( ) ( ) ( ) ( )

4 DERMAL PENETRATION OF FENTANYL 247 characteristics of the donor skin. The initial study on intraindividual variation using 6 diffusion cells with skin from the same donor demonstrates an acceptable and low intraexperimental as well as intraindividual variation, which is a prerequisite for a valid discussion of interindividual differences regarding dermal penetration of fentanyl. The experiment comparing absorption across abdominal versus breast skin from the same individuals was intended to give information on the importance of the application site. The results demonstrate limited variation between sites with no particular pattern. Two of the three individuals have a higher penetration across breast skin and one have a higher penetration across abdominal skin. The small variation seen between sites could mean that there is no real significant variation. In accordance with our observations on penetration of fentanyl through full-thickness skin, an experiment using dermatomed skin samples from cadavers (Roy & Flynn 1990) showed no significant difference between the permeability of fentanyl across abdominal, thigh or breast skin. The only site that had lower permeability was the sole of the foot (Roy & Flynn 1990). Interindividual penetration Previous published data indicate that there is a considerable interindividual variation in drug penetration (Feldmann & Maibach 1967; Southwell et al. 1984; Egger et al. 1998). In an in vitro study using abdominal cadaver skin, Southwell et al. (1984) observed a coefficient of variation close to 70%. Likewise, Nielsen & Nielsen (2000) found standard deviations in the region of 40 67% when they used the same method and laboratory as used in the present study with non-cadaverous skin to examine the penetration of a range of pesticides. Considering fentanyl, a permeability coefficient varying 4 5 times through skin sections from different cadavers has been described (Roy & Flynn 1990). Thus, our results on the interindividual variation in penetration of fentanyl across human skin are slightly lower than observations from previous studies. The interesting observation is, however, that a few individuals have very high or very low penetration rates, respectively. This means that at identical doses they will absorb approximately two times as much fentanyl as the average person whereas another subpopulation will absorb only one third as much as the average person. The clinical consequence is that the individuals with low penetration rates will not reach a therapeutic serum concentration unless the dermal application is increased 3 times, whereas a small group unintentionally may obtain too high serum concentrations of fentanyl. The clinical consequence regarding the risk of overdosing could, however, be minimized through the use of fentanyl patches with a barrier limiting the dosage (mg fentanyl per hour) reaching the skin. Patch application versus aqueous solution. A comparison of results from the two experiments with the 21 individuals given fentanyl in aqueous solution and the 8 individuals given the fentanyl patch shows that the amount absorbed from the patch is much larger then from the solution. The absorption rate was 1300 ng/cm 2 /hr delivered from the patch and 61ng/cm 2 /hr in the solution. However, the amount of fentanyl in the patch experiment was considerable higher than in the other experiments, which apart from the small sample size make a direct comparison of amounts of dermally transferred fentanyl difficult. Considering amount absorbed in percentages we see that 8.6% of the solution is absorbed, which is close to the 7.4% of the fentanyl in the patch. The best way to compare the absorption profile was to compare the permeability coefficients. The median permeability coefficient of the patch was cm/hr, which was similar to the permeability coefficient of the aqueous fentanyl solution, and comparable to previous data using an identical experimental setup except for the skin samples, which were dermatomed skin from cadavers (Roy & Flynn 1989; Roy et al. 1996). Thus, the use of full-thickness skin instead of dermatomed skin changed these penetration characteristics to a limited degree only. However, lag-times were higher in our experimental setup, and the amount of fentanyl present in the skin membrane acting as a reservoir would also be expected to be larger than in experiments using dermatomed skin. The delivery system consists of a rate-controlling membrane to make sure the drug is delivered at a certain rate that should decrease the interindividual variation in absorption and eliminate the risk of patients receiving an overdose. The interesting observation was that the permeability coefficient was the same among the 21 individuals given the fentanyl solution and the 8 individuals given the fentanyl patch. This observation from our in vitro model indicates that the skin may actually be the rate-limiting factor. Graphs of the absorption of fentanyl with time indicate that the majority of observations lie within a small interval but that there are some exceptions that are positioned either above or below the others. If the rate-limiting membrane in the patch was set to slow the absorption down and by that control the delivery, we would expect all individuals to absorb equal amounts and only slow absorbers would show variation from this by absorbing less. Fast absorbers would be slowed down and held below the ratelimit. However, we observe one person, who absorbs more than the rest. That does not comply with the patch membrane being rate-limiting, unless the rate limit of the patch has not yet been reached or the patch was in some way damaged. If the skin determines the absorption limit in most individuals, a controlling membrane is only necessary in case the fast absorbers use the delivery system. To avoid increased absorption by these persons and to control the absorption better, the delivery rate of the membrane should be as close to the average absorption profile as possible. Individuals that have a higher dermal penetration than the average would then not receive an unintentional high dose. Individuals that are slow absorbers could, by wearing an increased number of patches, reach the same serum concentration of fentanyl and thereby clinical effect as the average person.

5 248 RIKKE H. LARSEN ET AL. Acknowledgements We are indebted to the donors as well as staff from the Department for Plastic Surgery at Odense University Hospital. The project was funded by The Danish Medical Research Council and E. Danielsen and wife s foundation. References Bronaugh, R. L. & H. I. Maibach: Percutaneous absorption. Marcel Dekker Inc., USA, Bronaugh, R. L., R. F. Stewart & M. Simon: Methods for in vitro percutaneous absorption studies. VII: Use of excised human skin. J. Pharm. Sci. 1986, 75, Cherny, N. J., V. Chang, G. Frager, J. M. Ingham, P. J. Tiseo, B. Popp, R. K. Portenoy & K. M. Foley: Opioid pharmacotherapy in the management of cancer pain: a survey of strategies used by pain physicians for the selection of analgesic drugs and routes of administration. Cancer 1995, 76, Collins, J. J., I. J. Dunkel, S. K. Gupta, C. E. Inturrisi, J. Lapin, L. N. Palmer, S. M. Weinstein & R. K. Portenoy: Transdermal fentanyl in children with cancer pain: feasibility, tolerability, and pharmacokinetic correlates. J. Pediatr. 1999, 134, Dellemijn, P. L., H. van Duijn & J. A. Vanneste: Prolonged treatment with transdermal fentanyl in neuropathic pain. J. Pain Symptom. Manage. 1998, 16, Egger, C. M., T. Duke, J. Archer & P. H. Cribb: Comparison of plasma fentanyl concentrations by using three transdermal fentanyl patch sizes in dogs. Vet. Surg. 1998, 27, Feldmann, R. J. & H. I. Maibach: Regional variation in percutaneous penetration of 14C cortisol in man. J. Invest. Dermatol. 1967, 48, Franz, T. J.: Percutaneous absorption on the relevance of in vitro data. J. Invest. Dermatol. 1975, 64, Grond, S., L. Radbruch & K. A. Lehmann: Clinical pharmacokinetics of transdermal opioids: focus on transdermal fentanyl. Clin. Pharmacokinet. 2000, 38, Guy, R. H., J. Hadgraft & D. A. Bucks: Transdermal drug delivery and cutaneous metabolism. Xenobiotica 1987, 17, Jakobsson, M. & P. Strang: Fentanyl patches for the treatment of pain in dying cancer patients. Anticancer Res. 1999, 19, Lambropoulos, J., G. A. Spanos, N. V. Lazaridis, T. S. Ingallinera & V. K. Rodriguez: Development and validation of an HPLC assay for fentanyl and related substances in fentanyl citrate injection. J. Pharm. Biomed. Anal. 1999, 20, McClain, D. A. & C. C. Hug: Intravenous fentanyl kinetics. Clin. Pharmacol. Therap. 1980, 28, Nielsen, J. B. & F. Nielsen: Dermal in vitro penetration of methiocarb, paclobutrazol, and pirimicarb. Occup. Environ. Med. 2000, 57, Nielsen, J. B.: Effects of four detergents on the in-vitro barrier function of human skin. Int. J. Occup. Environ. Health 2000, 6, Reilly, C. S., A. J. Wood & M. Wood: Variability of fentanyl pharmacokinetics in man. Computer predicted plasma concentrations for three intravenous dosage regimens. Anaesthesia 1985, 40, Ringe, J. D., H. Faber, O. Bock, S. Valentine, D. Felsenberg, M. Pfeifer, H. W. Minne & S. Schwalen: Transdermal fentanyl for the treatment of back pain caused by vertebral osteoporosis. Rheumatol. Int. 2002, 22, Roberge, R. J., E. P. Krenzelok & R. Mrvos: Transdermal drug delivery system exposure outcomes. J. Emerg. Med. 2000, 18, Roy, S. D. & G. L. Flynn: Solubility and related physicochemical properties of narcotic analgesics. Pharm. Res. 1988, 5, Roy, S. D. & G. L. Flynn: Transdermal delivery of narcotic analgesics: comparative permeabilities of narcotic analgesics through human cadaver skin. Pharm. Res. 1989, 6, Roy, S. D. & G. L. Flynn: Transdermal delivery of narcotic analgesics: ph, anatomical, and subject influences on cutaneous permeability of fentanyl and sufentanil. Pharm. Res. 1990, 7, Roy, S. D., M. Gutierrez, G. L. Flynn & G. W. Cleary: Controlled transdermal delivery of fentanyl: Characterizations of pressuresensitive adhesives for matrix patch design. J. Pharm. Sci. 1996, 85, Roy, S. D., S. Y. Hou, S. L. Witham & G. L. Flynn: Transdermal delivery of narcotic analgesics: comparative metabolism and permeability of human cadaver skin and hairless mouse skin. J. Pharm. Sci. 1994, 83, Simpson, R. K., E. A. Edmondson, C. F. Constant & C. Collier: Transdermal fentanyl as treatment for chronic low back pain. J. Pain Symptom. Manage. 1997, 14, Southwell, D., B. W. Barry & R. Woodford: Variations in permeability of human skin within and between specimens. Int. J. Pharm 1984, 18, Varvel, J. R., S. L. Shafer, S. S. Hwang, P. A. Coen & D. R. Stanski, D. R.: Absorption characteristics of transdermally administered fentanyl. Anesthesiology 1989, 70,