Introduction
Tramadol is a centrally acting analgesic with weak opioid agonist and monoamine reuptake inhibitor activity1 and with good oral bioavailability in cats.2 Building on previous evidence supporting the analgesic efficacy of tramadol in cats,3–6 a recent placebo-controlled, blinded crossover study7 reveals that oral administration of 2 mg of tramadol/kg every 12 hours to geriatric cats optimized activity counts as determined with a collar-mounted activity monitor and optimized mobility as determined with an owner questionnaire; however, 3 of 24 cats were withdrawn from the study for refusing to take the medication. Tramadol has a bitter taste, and coupled with the challenge of restraining uncooperative cats so the drug can be administered orally, its oral administration can be difficult.
Many in vivo studies7–19 include descriptions of the transdermal absorption of various drugs in cats, with only a few that yield clinically relevant or consistent plasma concentrations. By the Lipinski rule of 5, a drug molecule is expected to have good oral absorption if it has a molecular weight of < 500 Da, < 5 hydrogen bond donors, < 10 hydrogen bond acceptors, and good lipophilicity reflected by an octanol-water partition coefficient of > 5.20 Drugs delivered transdermally adhere to more stringent rules, typically having a molecular weight < 335 Da, < 2 hydrogen bond donors, < 5 hydrogen bond acceptors, and an octanol-water partition coefficient of between 0 and 5, such that the ideal transdermal drug is not too lipophilic or hydrophilic.21 Because tramadol's physiochemical properties are within these parameters (molecular weight, 263.37 Da; hydrogen bond donor, n = 1; hydrogen bond acceptors, 3; octanol-water partition coefficient, 1.36),22 good transdermal bioavailability is expected.
Ex vivo transdermal absorption of tramadol through the skin of the inner pinnae of cats has been demonstrated by use of a Franz skin finite dose model.23 In this model, absorption is a ratio of the drug mass recovered from a receptor solution in contact with the underside of a piece of mounted skin relative to the original mass of the drug that was applied to the skin surface. Mean ± SD percentage absorption is significantly higher for tramadol in a proprietary transdermal basea than for tramadol in a traditional pluronic lecithin organogel penetration enhancer (100.4 ± 3.2% vs 90.3 ± 14.7%, respectively). For both formulations, peak flux is achieved between 1 and 3 hours after the time of application over a 48-hour study period. These results suggested that the proprietary transdermal basea and pluronic lecithin organogel may facilitate in vivo transdermal absorption.
A commercial compounded transdermal formulation (gel) of tramadolb has been developed as a low-stress alternative to oral administration of tablets to cats; however, reports that describe in vivo transdermal absorption of tramadol in cats are lacking. Because it is a compounded medication, the commercial entity is not required to demonstrate the product's efficacy and safety. Therefore, the objectives of the study reported here were to determine whether detectable and clinically relevant plasma concentrations of tramadol and its metabolites could be achieved after application of 1 dose of the commercial compounded transdermal tramadol gel product to healthy cats, describe the pharmacokinetics of transdermal tramadol, and determine the concentration of tramadol in the transdermal gel product. Transdermal application of the gel was expected to result in measurable and clinically relevant plasma concentrations of tramadol and its metabolites, and gel tramadol concentrations were expected to be within 90% to 110% of the label concentration.
Materials and Methods
Animals
The study protocol was approved by the Institutional Animal Care and Use Committee at Kansas State University. Eight healthy client-owned cats were recruited from the Kansas State University Veterinary Health Center for voluntary participation in the study. Written owner-informed consent was obtained prior to enrollment. Cats were determined to be healthy on the basis of history and physical examination (performed on day 1); cats did not receive any medications except for parasite preventives within 7 days of enrollment. Environmental conditions and access to water were maintained throughout the study. Food was maintained ad libitum except for a 12-hour fasting period before sedation for IV catheter placement. Cats were observed for behavioral changes and other visible drug effects throughout the study. History, physical examination findings, treatments, and clinical observations were recorded for each cat in the health center medical record as well as stored in a restricted-access, cloud-based application.
Sampling catheter placement
On day 1, each cat was sedated with 1 dose each of ketamine hydrochloridec (5 mg/kg, IM) and butorphanol tartrated (0.2 mg/kg, IM), but a second dose of each was administered if necessary on the basis of a cat's temperament. With the aseptic technique, a 20-gauge, 45-cm through-the-needle central venous cathetere was placed in one of the medial saphenous veins, with the tip advanced 35 cm in the vein and secured with medical tape. Cats were fitted with a rigid Elizabethan collar extending past the ear tips and nosethat was worn for the remainder of the study. Cats were allowed to recover and acclimate to their environment for at least 12 hours before receiving tramadol.
Positive control for oral formulation of tramadol
By use of a random number table, 1 cat was assigned to be a positive control, having received tramadol orally rather than transdermally. The next day (time 0), this cat received 2 mg of tramadol hydrochloride/kg, PO, once. A 50-mg tablet of tramadol hydrochloridef was pulverized, and the appropriate dose (milligram-per-kilogram basis) was transferred to a pharmaceutical-grade gel capsuleg by a pharmacist at the Veterinary Health Center.
Application of transdermal formulation of tramadol
Seven proprietary metered-dose applicators that contained 200 mg of tramadol/mLb (compounded August 6, 2020; lot No. 111-02369708) in a proprietary basea were preordered for each of the 7 remaining cats. The applicators were stored at room temperature (20°C) per the compounding pharmacy's instructions and administered prior to the beyond use date (January 28, 2021). On day 2 of the study (time 0), 1 metered dose that contained an estimated 10 mg of tramadol was applied to the left inner pinna of each cat. According to the compounding pharmacy's instructions, the pen-shaped applicator was first primed by rotating the end of the applicator until a small bead of gel appeared at the tip, then by rotating the end until the left- and right-facing arrows were aligned, and lastly by wiping the tip clean. After priming, 1 dose was applied as recommended by the compounding pharmacy: first, by twice rotating the end of the applicator 360° (2 full twists of the applicator were intended to deliver 0.05 mL of gel that contained 10 mg of tramadol), and second, without preparing or cleaning the skin, the applicator tip was used to spread and gently rub the metered dose of 0.05 mL over the entire concave skin surface of the left inner pinna.
Blood sampling
Two milliliters of blood was collected from each cat through the IV catheter immediately before tramadol administration (time 0) and at 30 and 60 minutes and 2, 4, 8, 12, and 24 hours after administration (PO for 1 cat and transdermal for 7 cats) by use of a standard 3-syringe sampling technique as follows: 1 mL of blood was collected into a syringe with heparinized saline (0.9% NaCl) solution (5 U of heparin/mL). Another syringe without heparinized saline solution was used to collect a 2-mL blood sample. The contents of the first syringe were returned to the patient through the IV catheter, and then a third syringe containing 2 mL of heparinized saline solution (5 U of heparin/mL) was administered through the IV catheter. Each 2-mL blood sample was transferred to a blood collection tube that contained heparin. The sample then underwent centrifugation at 3,250 X g for 10 minutes, and the plasma was retrieved, transferred to a cryogenic storage vial, and stored at–80°C until analysis.
Tramadol assay
Plasma concentrations of tramadol and its metabolites, TM1 (active) and TM2 (inactive), were determined with high-performance liquid chromatographyh with mass spectrometry.i For the latter, the scan range was 265.17 to 58.98 m/z (qualifying ion to quantifying ion) for tramadol, 251.13 to 58.98 m/z for TMI, and 251.13 to 44.98 m/z for TM2; for all analytes, the scan range for the internal standardj was 269.17 to 59.03 m/z. The retention times for tramadol, TM1, TM2, and the internal standard were 1.01, 0.91, 1.01, and 1.01 minutes, respectively. The mobile-phase composition started at 95% of a solution of deionized water with 0.1% formic acid and 5% of a solution of acetonitrile with 0.1% formic acid and progressed linearly to 5% of the solution of deionized water with 0.1% formic acid over 1.8 minutes; however, the mobile-phase composition was kept constant at 1 minute for 0.1 minutes and then resumed linear progression to 5% of the solution of deionized water with 0.1% formic acid until the end time of 1.8 minutes was attained. Separation was achieved with a columnk maintained at 40°C.
Plasma samples from the 8 treated cats, a plasma sample from an untreated cat for quality control, and a plasma sample from an untreated cat that was mixed with the internal standard were processed identically. Fifty microliters of plasma was added to a phospholipid removal plate,l to which 150 µL of acetonitrile with 1% formic acid that had 25 ng of the internal standard/mL was forcefully added thereafter to precipitate the plasma proteins. Positive pressure was applied and the eluate collected, of which 10 µL was transferred into the mass spectrometer's ion source. The standard curve was linear between 1 and 500 ng/mL for tramadol and TM1 and from 5 to 500 ng/mL for TM2. Standard curves were accepted if the coefficient of determination was at least 0.99 and the measured concentrations of the analytes were within 15% of their actual concentrations. Quality control samples were included in triplicate for concentrations of 5, 50, and 500 ng/mL for each analyte, and the replicate was accepted if at least 4 of the 6 quality control samples had concentrations that were within 15% of the actual concentrations. From the measured concentrations of 5 replicates for each concentration of 5, 50, and 500 ng/mL, the assay's accuracy for tramadol, TM1, and TM2 were determined to be 95%, 98%, and 93%, respectively, and the assay's precision (coefficient of variation) was determined to be 5%, 6%, and 13%, respectively.
Transdermal tramadol concentration testing
The concentration of tramadol in the compounded gel was determined in 50 µL (10 mg) and 100 µL (20 mg) dispensed from each applicator (n = 7 [14 total samples]). The gel was mixed with methanol to achieve a total volume of 1 mL, sonicated for 5 minutes in a water bath, vortexed for 10 seconds, sonicated again for 5 minutes in a water bath, and vortexed again for 10 seconds. The resultant suspension was serially diluted in methanol to concentrations of 500 ng/mL for the 10-mg dispensed dose and 1,000 ng/mL for the 20-mg dispensed dose. The final dilution (100 µL) was further diluted in an injection vial that contained 900 µL of 0.1% formic acid in water and 500 ng/mL of the internal standard. A standard curve of tramadol in methanol was made similarly for both samples per applicator with a range of 10 to 5,000 ng/mL, and the results were accepted if the coefficient of determination was at least 0.99 and the measured concentrations were within 15% of the actual concentrations. The quality control samples were made in triplicate at 100, 500, and 5,000 ng of tramadol/mL. On the basis of the results for the quality control samples, the assay's accuracy and precision were determined to be 100% and 2%, respectively. The injection volume was 5 µL.
Results
Eight healthy client-owned cats (spayed female, n = 4; neutered male, 4) were recruited for the study; cats were 1 to 7 years of age (median age, 3 years) and weighed 3.9 to 6.7 kg (median body weight, 4.7 kg). All cats completed the study; no vomiting, diarrhea, skin reactions, and behavioral changes were observed for any cat following tramadol administration. Each cat had mild swelling distal to the site of the sampling catheter at the medial saphenous vein that lessened after the catheters were removed at the end of the study.
In the cat that only received the oral formulation of tramadol at 2 mg/kg, the parent drug was detectable in plasma at all times from 30 minutes to 24 hours and reached its maximum concentration (261.4 ng/mL) at 4 hours. Active TM1 became detectable starting at 60 minutes and reached its maximum concentration (27.3 ng/mL) at 12 hours; inactive TM2 became detectable starting at 2 hours and reached its maximum concentration (51.9 ng/mL) at 8 hours.
Final doses on a milligram-per-kilogram basis for the cats that received an estimated 10-mg dose of the transdermal formulation of tramadol ranged from 1.5 to 2.6 mg/kg (median, 2.1 mg/kg). After application of the transdermal formulation, plasma tramadol concentrations in 49% (24/49) of the samples were below the assay's limit of detection of 1 ng/mL. In 2 of 7 cats, tramadol was undetectable in the plasma at all times. In the plasma of 5 cats that had detectable tramadol, maximum plasma tramadol concentrations were detected at 24 hours and ranged from 2.0 to 4.3 ng/mL (median, 2.2 ng/mL; Figure 1). At 24 hours, only minimal plasma concentrations of TM1 were detected in 3 of 49 samples, and plasma concentrations of TM2 were not detected in any samples. Because plasma tramadol concentrations were near or less than the assay's limit of detection, characterization of the pharmacokinetics for tramadol following a single dose of the transdermal formulation was not attempted.
Plasma tramadol concentrations (ng/mL) collected from 30 minutes through 24 hours following oral administration of a single dose (2 mg/kg; black line) to 1 cat or following application of a single dose (approx 10 mg/cat [range, 1.5 to 2.6 mg/kg]; range of actual doses, 1.9 to 3.2 mg/kg) of a transdermal formulation to the inner pinna of the ear of 5 cats (gray lines). Data were not included for 2 cats that received the transdermal formulation because plasma tramadol concentrations were below the assay's (high-performance liquid–chromatography-mass spectrometry) limit of detection of 1 ng/mL.
Citation: American Journal of Veterinary Research 82, 10; 10.2460/ajvr.82.10.840
The concentration of tramadol in the gel exceeded the labeled dose in 12 of 13 samples, with concentrations outside of the 90% to 110% range, which is the United States Pharmacopeia standard for compounded drugs (Table 1). One sample was lost during processing and therefore not included in analysis. After correcting for the actual concentrations of tramadol in the gel, actual doses for the 7 cats ranged from 1.9 to 3.2 mg/kg (median, 2.8 mg/kg).
Tramadol concentration expressed as a percentage of the estimated dose, 10 or 20 mg, dispensed as gel by 7 proprietary metered-dose applicators.
| Applicator* | Percentage of estimated 10-mg dose (%) | Percentage of estimated 20-mg dose (%) |
|---|---|---|
| 1 | — | 118 |
| 2 | 125 | 131 |
| 4 | 139 | 108 |
| 5 | 130 | 120 |
| 6 | 127 | 126 |
| 7 | 136 | 136 |
| 8 | 126 | 115 |
Applicator number corresponds to the number assigned to each cat.
— = Sample lost and therefore sample was not analyzed.
Discussion
In the present study, plasma tramadol concentrations were undetectable or low following 1 dose of a transdermal formulation of tramadol dispensed by a proprietary metered-dose applicator (median estimated dose, 2.1 mg/kg; median actual dose, 2.8 mg/kg). In those cats with detectable plasma tramadol concentrations 4 hours after tramadol application, concentrations were > 100X lower than those attained after oral administration of tramadol. A single 2-mg/kg dose for both formulations of tramadol was expected to yield clinically relevant plasma tramadol concentrations on the basis of previous studies.2, 7 In 1 study,7 activity counts for geriatric cats with osteoarthritis were significantly higher with administration of 2 mg of tramadol/kg, PO, every 12 hours, and were not with a dosage of 1 or 4 mg/kg, PO, every 12 hours, compared with placebo.7 To the authors’ knowledge, plasma tramadol concentrations following 1 or more 2-mg/kg doses of the oral formulation of tramadol have not been reported. In another study,2 1 dose of 5.2 mg of tramadol/kg, PO, resulted in a maximum plasma concentration of 914 ± 232 ng/mL. Based on results that a single 2-mg/kg dose of tramadol, PO, yielded a maximum plasma concentration of 261.4 ng/mL in the present study was not surprising.
Analysis of the tramadol concentration in the gel dispensed from the metered-dose applicators showed that each applicator delivered an amount that exceeded the estimated dispensed dose by up to 39%. Yet, a 39% higher dose (up to 3.2 mg/kg applied transdermally) is well within the margin of safety for tramadol. This difference may be attributed to a formulation error, such that the tramadol concentration was too high, or to a delivery error, such that the applicator dispensed too great a volume. In a future study, determining whether the error was due to a formulation or delivery error may be determined by aspirating a standard volume or weight of gel directly from an applicator and analyzing the gel's tramadol content. In the present study, however, a dispensed dose was elected for analysis because that is the deliverable clinical dose. The undetectable or low plasma tramadol concentrations following tramadol application was unlikely attributable to a formulation or delivery error in the present study.
By use of the modified Lipinski rule of 5 model, tramadol's physicochemical properties predict its suitability for transdermal delivery to people. Generated by a review of 17 successful FDA-approved transdermal formulations of drugs for people, the proposed threshold values are a molecular weight of < 335 Da, ≤ 2 hydrogen bond donors, ≤ 5 hydrogen bond acceptors, and an octanol-water partition coefficient between 0 and 5.20, 21 Tramadol meets these criteria because it has a molecular weight of 263.37 Da, 1 hydrogen bond donor, 3 hydrogen bond acceptors, and an octanol-water partition coefficient of 1.36.22 However, a limitation of both the modified and original rule of 5 models is that they are based on the permeability of the skin of people, not cats. Species differences in the thickness of the stratum corneum, viscosity of the intercellular space, and local tissue vascularity likely influence transdermal drug delivery.
Factors that may account for the marked differences in the transdermal absorption of tramadol among experiments with the skin of cats include the influence of propylene glycol as a penetration enhancer,23 the effects of freezing and thawing on the permeability of the skin harvested from the cadavers of cats, or the possibility of phase I metabolism (reduction, oxidation, and hydrolysis) and phase II metabolism (conjugation of functional groups to the drug) of tramadol during its transit through the viable epidermis of cats.24 If phase I metabolism existed, however, plasma TM1 and TM2 should have been detected, unless the metabolic enzymes are different in the skin versus in the liver of cats.
A limitation of the present study was the use of a transdermal dose equivalent to the oral dose. Previous studies14, 17 reveal that equivalent doses of transdermal and oral formulations of some drugs may not yield equivalent systemic concentrations. For example, 1 transdermal dose of 5 mg of glipizide, compared with 1 oral dose of 5 mg, administered to cats results in only 20% relative absorption.14 Also in cats, twice-daily transdermal dosages of 9 mg of phenobarbital/kg yield therapeutic serum phenobarbital concentrations that are similar to those achieved with twice-daily oral dosages of 3 mg of phenobarbital/kg.17 This dosage difference was speculated to be because of differences in transdermal bioavailability.17
Because the tramadol concentration gradient between the gel (10 mg/0.05 mL [200,000,000 ng/mL]) and the plasma (2.2 ng/mL) was already large such that transdermal absorption was saturated and application of a gel with a higher tramadol concentration or of more gel to the skin of the inner pinna was not expected to change the concentration gradient. Similarly, more frequent applications were unlikely to result in drug accumulation. Because transdermal drug absorption is directly related to the surface area of application, increasing the plasma concentration from 4.3 ng/mL (the maximum plasma concentration following the transdermal dose in the present study) to 261.4 ng/mL, (the maximum plasma concentration following the oral dose) would require application of a sufficient volume of gel to cover an area of skin > 60X the size of the pinna (261.4/4.3 = 60.8), which would be clinically impractical.
In conclusion, plasma tramadol concentrations in cats were undetectable or low following application of 1 dose of a transdermal formulation (median dose, 2.8 mg of tramadol/kg), despite that the amount of tramadol in each metered dose of the gel product consistently exceeded that of the expected amount. Further research is necessary to determine the reasons for these results. A larger volume of the gel product could be applied onto a larger surface area than the inner pinna of the ear or, more practically, alternative vehicles for transdermal drug delivery could be evaluated.
Acknowledgments
No external funding was used in this study. The authors declare that there were no conflicts of interest.
The authors thank Landa Colvin-Marion for formulation of the oral tramadol and Kara Smith for assistance with procurement of supplies.
Abbreviations
| TM1 | Tramadol metabolite 1 (O-desmethyltramadol) |
| TM2 | Tramadol metabolite 2 (N-desmethyltramadol) |
Footnotes
Lipoderm, PCCA Inc, Houston, Tex.
Tramadol Twist-a-Dose Transdermal Gel, Wedgewood Pharmacy, Swedesboro, NJ.
Ketaset, Fort Dodge Animal Health, Fort Dodge, Iowa.
Torbugesic, Zoetis Inc, Kalamazoo, Mich.
LL2045 long line catheter, Mila International Inc, Florence, Ky.
Sun Pharmaceutical Industries Inc, Princeton, NJ.
Fagron Inc, Saint Paul, Minn.
Acquity UPLC H-Class UPLC, Waters Corp, Milford, Mass.
TQD, Waters Corp, Milford, Mass.
Tramadol C13d2, Cerilliant, Round Rock, Tex.
Acquity UPLC HSS T3 (1.8 mm), Waters Corp, Milford, Mass.
Ostro Pass-through Sample Preparation Plates, Waters Corp, Milford, Mass.
References
- 2. ↑
Pypendop BH, Ilkiw JE. Pharmacokinetics of tramadol, and its metabolite O-desmethyl-tramadol, in cats. J Vet Pharmacol Ther 2008;31:52–59.
- 3. ↑
Pypendop BH, Siao KT, Ilkiw JE. Effects of tramadol hydrochloride on the thermal threshold in cats. Am J Vet Res 2009;70:1465–1470
- 4.
Cagnardi P, Villa R, Zonca A, et al. Pharmacokinetics, intraoperative effect and postoperative analgesia of tramadol in cats. Res Vet Sci 2011;90:503–509
- 5.
Basiri B, Cheng CH, Rahman NA. Analgesic efficacy of pre-operative tramadol in combination with acepromazine in cats undergoing ovariohysterectomy. Pak Vet J 2014;34:403–405.
- 6. ↑
Evangelista MC, Silva RA, Cardozo LB, et al. Comparison of preoperative tramadol and pethidine on postoperative pain in cats undergoing ovariohysterectomy. BMC Vet Res 2014;10:252.
- 7. ↑
Guedes AGP, Meadows JM, Pypendop BH, et al. Evaluation of tramadol for treatment of osteoarthritis in geriatric cats. J Am Vet Med Assoc 2018;252:565–571.
- 8.
Sleeper MM, O’Donnell P, Fitzgerald C, et al. Pharmacokinetics of furosemide after intravenous, oral and transdermal administration to cats. J Feline Med Surg 2019;21:882–886.
- 9.
Barnoski J, Lee-Fowler TM, Boothe DM, et al. Serum theophylline after multiple dosing with transdermal gels in cats. J Feline Med Surg 2019;21:329–334.
- 10.
Hill KE, Mills PC, Jones BR, et al. Percutaneous absorption of methimazole: an in vitro study of the absorption pharmacokinetics for two different vehicles. J Vet Pharmacol Ther 2015;38:581–589.
- 11.
Benson KK, Zajic LB, Morgan PK, et al. Drug exposure and clinical effect of transdermal mirtazapine in healthy young cats: a pilot study. J Feline Med Surg 2017;19:998–1006.
- 12.
Hoffman SB, Yoder AR, Trepanier LA. Bioavailability of transdermal methimazole in a pluronic lecithin organogel (PLO) in healthy cats. J Vet Pharmacol Ther 2002;25:189–193.
- 13.
Helms SR. Treatment of feline hypertension with transdermal amlodipine: a pilot study. J Am Anim Hosp Assoc 2007;43:149–156.
- 14. ↑
Bennett N, Papich MG, Hoenig M, et al. Evaluation of transdermal application of glipizide in a pluronic lecithin gel to healthy cats. Am J Vet Res 2005;66:581–588.
- 15.
Zajic LB, Herndon AK, Sieberg LG, et al. Assessment of absorption of transdermal ondansetron in normal research cats. J Feline Med Surg 2017;19:1245–1248.
- 16.
Eichstadt LR, Corriveau LA, Moore GE, et al. Absorption of transdermal fluoxetine compounded in a lipoderm base compared to oral fluoxetine in client-owned cats. Int J Pharm Compd 2017;21:242–246.
- 17. ↑
Delamaide Gasper JA, Barnes Heller HL, Robertson M, et al. Therapeutic serum phenobarbital concentrations obtained using chronic transdermal administration of phenobarbital in healthy cats. J Feline Med Surg 2015;17:359–363.
- 18.
Mealey KL, Peck KE, Bennett BS, et al. Systemic absorption of amitriptyline and buspirone after oral and transdermal administration to healthy cats. J Vet Intern Med 2004;18:43–46.
- 19. ↑
Lee DD, Papich MG, Hardie EM. Comparison of pharmacokinetics of fentanyl after intravenous and transdermal administration in cats. Am J Vet Res 2000;61:672–677.
- 20. ↑
Lipinski CA, Lombardo F, Dominy BW, et al. Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Adv Drug Deliv Rev 2001;46:3–26.
- 21. ↑
Choy YB, Prausnitz MR. The rule of five for non-oral routes of drug delivery: ophthalmic, inhalation and transdermal. Pharm Res 2011;28:943–948.
- 22. ↑
National Center for Biotechnology Information. PubChem compound summary for CID 63013, tramadol hydrochloride. Available at: pubchem.ncbi.nlm.nih.gov/compound/Tramadol-hydrochloride. Accessed Dec 13, 2020.
- 23. ↑
Bassani ASBD, Simmons C, Phan H. In vitro characterization of the percutaneous absorption of tramadol into inner ear domestic feline skin using the Franz skin finite dose model. Vet Med Anim Sci 2015;3.
