Portal de Periódicos da CAPES

Human Skin Penetration of Flufenamic Acid: In Vivo/In Vitro Correlation (Deeper Skin Layers) for Skin Samples from the Same Subject

2002; Elsevier BV; Volume: 118; Issue: 3 Linguagem: Inglês

10.1046/j.0022-202x.2001.01688.x

1523-1747

Heike Wagner, Claus‐Michael Lehr, Ulrich F. Schaefer, Karl‐Heinz Kostka,

Contact Dermatitis and Allergies

Previously, the interest in in vivo/in vitro correlations in the dermal field of research has increased steadily. Unfortunately, in most cases the skin from different human donors was taken for in vivo and in vitro experiments, which led to problems concerning the interindividual variability of the skin. Therefore, we established a methodology to utilize the same skin for both sets of data. In time dependency, drug amounts in the stratum corneum and the deeper skin layers were determined from eight donors using the same skin area for in vivo and the corresponding in vitro tests. Penetration experiments were carried out with the lipophilic drug flufenamic acid dissolved in wool alcohols ointment as the model formulation, which was administered to the skin under “infinite dose” conditions. At different time points prior to starting the surgery, the drug preparation was applied topically on the edges of the skin area, which was planned for excision using Finn chambers. After anesthetizing the patient and disinfecting the operation area, the incubated skin pieces were cut off first and immediately frozen to limit further drug diffusion. In vitro experiments were performed on the remaining skin flap, using two different test systems, a penetration and a permeation model. At the end of all experiments (in vivo and in vitro) the skin specimens were segmented horizontally and the drug was extracted and quantified. The in vivo and in vitro drug amounts in the stratum corneum and the deeper skin layers, respectively, were compared. The inevitable use of unknown volumes of disinfectant in vivo (medical reasons) might be the reason why a correlation failed for the stratum corneum. Nevertheless, for both in vitro test systems a direct linear correlation was found for the deeper skin layers, which showed slopes of a = 3.2272 ± 0.3933 (penetration model vs in vivo) and a = 1.7776 ± 0. 1926 (permeation model vs in vivo). This difference demonstrates the varying influence of the test systems and represents a factor about which in vivo and in vitro data are shifted against each other. As far as the model drug flufenamic acid is concerned, this methodology represents a tool to predict drug penetration into the deeper skin layers in vivo after carrying out corresponding in vitro experiments. Therefore, the potential is given to reduce the number of in vivo experiments, the risk for the volunteers, and the costs for the development of new drug preparations. Previously, the interest in in vivo/in vitro correlations in the dermal field of research has increased steadily. Unfortunately, in most cases the skin from different human donors was taken for in vivo and in vitro experiments, which led to problems concerning the interindividual variability of the skin. Therefore, we established a methodology to utilize the same skin for both sets of data. In time dependency, drug amounts in the stratum corneum and the deeper skin layers were determined from eight donors using the same skin area for in vivo and the corresponding in vitro tests. Penetration experiments were carried out with the lipophilic drug flufenamic acid dissolved in wool alcohols ointment as the model formulation, which was administered to the skin under “infinite dose” conditions. At different time points prior to starting the surgery, the drug preparation was applied topically on the edges of the skin area, which was planned for excision using Finn chambers. After anesthetizing the patient and disinfecting the operation area, the incubated skin pieces were cut off first and immediately frozen to limit further drug diffusion. In vitro experiments were performed on the remaining skin flap, using two different test systems, a penetration and a permeation model. At the end of all experiments (in vivo and in vitro) the skin specimens were segmented horizontally and the drug was extracted and quantified. The in vivo and in vitro drug amounts in the stratum corneum and the deeper skin layers, respectively, were compared. The inevitable use of unknown volumes of disinfectant in vivo (medical reasons) might be the reason why a correlation failed for the stratum corneum. Nevertheless, for both in vitro test systems a direct linear correlation was found for the deeper skin layers, which showed slopes of a = 3.2272 ± 0.3933 (penetration model vs in vivo) and a = 1.7776 ± 0. 1926 (permeation model vs in vivo). This difference demonstrates the varying influence of the test systems and represents a factor about which in vivo and in vitro data are shifted against each other. As far as the model drug flufenamic acid is concerned, this methodology represents a tool to predict drug penetration into the deeper skin layers in vivo after carrying out corresponding in vitro experiments. Therefore, the potential is given to reduce the number of in vivo experiments, the risk for the volunteers, and the costs for the development of new drug preparations. deeper skin layers stratum corneum Previously, many scientists have been interested in correlating human in vivo and in vitro data in the field of dermatology and (trans)dermal drug delivery, especially to test the possibility of decreasing the number of in vivo studies during formulation development. Only a limited number of in vivo test methods are suitable for application on humans, however, e.g., the tape stripping technique, used in addressing dermatopharmacokinetic questions (Albery and Hadgraft, 1979Albery W.J. Hadgraft J. Percutaneous absorption: in vivo experiments.J Pharm Pharmacol. 1979; 31: 140-147Crossref PubMed Scopus (165) Google Scholar;Rougier et al., 1983Rougier A. Dupius D. Lotte C. Roguet R. Schaefer H. In vivo correlation between stratum corneum reservoir function and percutaneous absorption.J Invest Dermatol. 1983; 81: 275-278Crossref PubMed Scopus (159) Google Scholar;Pershing et al., 1992Pershing L.K. Silver B.S. Krueger G.G. Shah V.P. Skelley J.P. Feasibility of measuring the bioavailability of topical betamethasone dipropionate in commercial formulations using drug content in skin and a skin blanching bioassay.Pharm Res. 1992; 9: 45-51Crossref PubMed Scopus (92) Google Scholar) or the determination of blood and urine levels, utilized to assess systemic bioavailability for transdermal delivery (Rougier et al., 1983Rougier A. Dupius D. Lotte C. Roguet R. Schaefer H. In vivo correlation between stratum corneum reservoir function and percutaneous absorption.J Invest Dermatol. 1983; 81: 275-278Crossref PubMed Scopus (159) Google Scholar;Guy et al., 1986Guy R.H. Carlstroem E.M. Bucks D.A.W. Hinz R.S. Percutaneous penetration of nicotinates: In vivo and in vitro measurements.J Pharm Sci. 1986; 75: 968-972Crossref PubMed Scopus (40) Google Scholar).Zesch and Schaefer, 1975Zesch A. Schaefer H. Penetrationskinetik von radiomarkiertem Hydrocortison aus verschiedenen Salbengrundlagen in die menschliche Haut. II. In vivo.Arch Dermatol Forsch. 1975; 252: 245-256Crossref PubMed Scopus (37) Google Scholar carried out in vivo studies and explored the amount of penetrated drug in different skin layers. Unfortunately, corresponding in vitro data of the same skin were not available. In a previous study (Wagner et al., 2000Wagner H. Kostka K.H. Lehr C.-M. Schaefer U.F. Drug distribution in human skin using two different in vitro test systems. Comparison with in vivo data.Pharm Res. 2000; 17: 1475-1481Crossref PubMed Scopus (105) Google Scholar) we reported on an in vivo/in vitro correlation for drug penetration data in the stratum corneum (SC). In the present study we aimed to demonstrate in the deeper skin layers (DSL) that in vivo data may also be predicted from in vitro experiments. To overcome the problem of interindividual and site variability of lipid composition, pH, and water content within the SC, which are known to influence significantly the penetration and distribution of drugs, we introduced an experimental design that allowed the realization of in vitro and in vivo experiments (SC and DSL) by using the skin of the same donor. With such sets of data an in vivo/in vitro correlation for human DSL should be tested. This information appears to be pivotal for the interpretation of in vitro results and may eventually lead to a reduction of the number of in vivo human studies necessary for the development of dermopharmaceutical drug products. The following chemicals and equipment were used: flufenamic acid in the crystalline modification I (Kali-Chemie Pharma, Hannover, Germany); wool alcohols ointment and Multifilm kristall-klar (Beiersdorf, Hamburg, Germany); Ringer solution, Sørensen phosphate buffer pH 7.4, McIlvaine citric acid-phosphate buffer pH 2.2 and NaOH (all components from Merck, Darmstadt, Germany); Plastibase (Heyden GmbH, München, Germany); Braunoderm (Braun Melsungen AG, Melsungen, Germany); methanol (Baker, Deventer, the Netherlands); Finn chambers (Hermal, Reinbek, Germany); Franz diffusion cell type 4G-01-00-20 (Perme Gear, Riegelsville, PA); cryomicrotome HR Mark II, model 1978 (SLEE, Mainz, Germany); isocratic high-performance liquid chromatography consisting of a 655 A 40 autosampler, L 4250 detector, L 6220 pump, 6000 K data interface and 5 µm LiChrospher 100/RP-18 column/12.5 cm × 4 mm (Merck-Hitachi, Darmstadt, Germany). Flufenamic acid (0.9% wt/wt) dissolved in wool alcohols ointment (WAO, German Pharmacopoeia 1999) was taken as the drug preparation under “infinite dose” conditions (Franz et al., 1993Franz T.J. Lehman P.A. Franz S.F. et al.Percutaneous penetration of N-Nitrosodiethanolamine through human skin (in vitro): Comparison of finite and infinite dose applications from cosmetic vehicles.Fundam Appl Toxicol. 1993; 21: 213-221Crossref PubMed Scopus (38) Google Scholar) applying 300 mg per cm2 (in vitro) and 34 mg per cm2 (in vivo). In contrast to the normal use of semisolid dosage forms, infinite dosing was chosen due to the following reasons: (i) to minimize the influence of the dose (standardized kinetic conditions); (ii) to achieve higher drug amounts within the skin (no problems with the detection limit); and (iii) to have a simple application procedure (recommendation of the ethics committee to reduce the stress of the patients) In the crystalline modification I, flufenamic acid has a melting point of 132.8°C and a pKa of 3.9 (Alignente and De Caprariis, 1982Alignente E. De Caprariis P. Flufenamic acid.in: K Florey Analytical Profiles of Drug Substances. 11. Academic Press, New York1982: 313-346Google Scholar). Its solubility in water is limited, but can be increased with lipophilic solvents as well as with alkaline solutions. This improvement of the solubility was made use of by extracting the drug out of the skin samples with NaOH. At 32°C, 0.9% flufenamic acid is completely dissolved in the ointment base (Wild, 1988Wild T. Einfluß der physikochemischen Eigenschaften von Arzneistoffen und Vehikeln auf die Permeabilitaet der menschlichen Hornschicht.Dissertation Saarbruecken. 1988Google Scholar), which was verified by light microscopic inspection. The ointment base consisted of lipophilic components only; therefore, flufenamic acid was undissociated within the preparation. The in vivo study was approved from the ethics committee of the “Aerztekammer des Saarlandes” and the Caritas-Traegergesellschaft Trier e.V. Eight subjects undergoing abdominal plastic surgery participated in the study. They were in good health and had no medical history of any dermatologic disease. For in vivo experiments, two Finn chambers (15 mm in diameter and 1 mm in depth) filled with the drug preparation (about 60 mg) were applied at marked edges of the skin area, which was planned for excision (Figure 1) at two different time points before the beginning of the surgery. As far as possible, incubation times between 30 and 60 min (chamber 1) as well as 60 and 180 min (chamber 2) were maintained. At the start of the surgery, directly after anesthetizing the patient, the Finn chambers were removed and the remaining ointment was wiped off with a piece of cotton in the direction of the edges of the skin flap to prevent any contamination of the middle area of the skin. The whole operation area was disinfected using Braunoderm, an isopropanolic solution of polyvinylpyrrolidone-iod. The incubated skin biopsies were cut off, their subcutaneous fatty tissue was removed and the skin pieces were frozen at -80°C to limit drug diffusion inside the skin. Afterwards, the rest of the skin flap was excised, the subcutaneous fatty tissue was also removed and pieces 10 cm × 10 cm in size were wrapped in aluminum foil, collected in an impermeable polyethylene bag and stored for a maximum of 3 mo in a freezer at -26°C until use. For in vitro experiments, skin disks were punched out, thawed, cleaned with cotton and transferred into the respective test system: the Saarbruecken penetration model (SB-M) or the Franz diffusion cell (FD-C). The SB-M (Figure 2a) is a model to study the rate and the extent of drugs that penetrated into the SC and the DSL. The drug preparation was filled in a 2 mm deep cavity of a Teflon punch and applied on the surface of the skin, which lay in a Teflon bloc on a filter paper soaked with Ringer solution to prevent water loss from the skin during the experimental time. The Teflon punch was charged with 500 g for 2 min to improve the contact between the drug preparation and the skin. Afterwards, the Teflon punch was fixed in its position and the gap between the two Teflon parts was sealed with Plastibase. The whole arrangement was transferred into a plastic box and placed in a water bath of 32°C for the respective time interval. The filter paper beneath the skin was always tested for its content of flufenamic acid. Using this model, the skin for itself is the only acceptor for the penetrating drug. The FD-C (Figure 2b) is a model that was originally developed to study drug permeation through the skin. In this case, the skin was sandwiched between two compartments: the donor compartment holds the drug preparation. To minimize side diffusion of the drug the drug formulation was applied in the present test series using a plastic cylinder. In addition to the skin, a Sørenson phosphate buffer solution pH 7.4, mixed with a magnetic stirring bar set at 500 r.p.m., served as a supplemental acceptor compartment. To improve the reproducibility of the experiments, the skin was always prehydrated with the basolateral acceptor medium for 30 min before the application of the drug formulation. The temperature of the whole apparatus was regulated using a water jacket flowed by circulating water of 32°C. One sample of the acceptor medium was always taken before and at the end of the incubation time to have a blind value of the acceptor medium and to check whether flufenamic acid had permeated through the skin during the investigated time intervals. The in vitro experiments lasted as long as the corresponding in vivo ones from the same skin flap. According to the in vivo design, the skin was also treated with Braunoderm directly after the removal of the ointment using the liquid of four times spraying in each case. Owing to the limited availability of the skin specimens in some cases, only two in vitro experiments could be carried out for both test systems and all investigated incubation times. All incubated skin biopsies were investigated in the same way, no matter whether the experiment was done in vivo or in vitro. First, the skin was transferred into a so-called stripping apparatus (Borchert, 1994Borchert D. Methoden zur Untersuchung der simultanen Penetration von Arzneistoffen und Vehikelbestandteilen aus Salben in exzidierter Humanhaut.Dissertation Saarbruecken. 1994Google Scholar) where it was mounted on cork disks using small pins. After covering the skin with a Teflon mask that released only an area of 15 mm in diameter, the skin was subsequently stripped with 20 pieces of adhesive tape (size = 15 × 19 mm), applying a weight of 2 kg for 10 s on the surface of the strip in each case and removing the strip rapidly afterwards. This methodology resulted in SC layers of nearly the same thickness (Borchert, 1994Borchert D. Methoden zur Untersuchung der simultanen Penetration von Arzneistoffen und Vehikelbestandteilen aus Salben in exzidierter Humanhaut.Dissertation Saarbruecken. 1994Google Scholar;Pershing et al., 1994Pershing L.K. Corlett J. Jorgensen C. In vivo pharmacokinetics and pharmacodynamics of topical ketoconazole and miconazole in human stratum corneum.Antimicrob Agents Chemother. 1994; 38: 90-95Crossref PubMed Scopus (48) Google Scholar;Theobald, 1998Theobald F. In vitro Methoden zur biopharmazeutischen Qualitaetspruefung von Dermatika unter Beruecksichtigung der Lipidzusammensetzung des Stratum corneum.Dissertation Saarbruecken. 1998Google Scholar) throughout the whole procedure. The first tape strip was always discarded because of potential contamination by the residual drug on the surface of the skin (Howes et al., 1996Howes D. Guy R. Hadgraft J. et al.Methods for assessing percutaneous absorption. The report and recommendations of ECVAM Workshop 13.Alternatives to Laboratory Animals. 1996; 24: 81-106Google Scholar). The rest of the tape strips was combined in five pools according to the following scheme: one, three, four, five, and six strips. Second, the skin was rapidly frozen in a stream of expanding carbon dioxide, and a specimen with a diameter of 13 mm was taken out of the stripped area and transferred into a cryomicrotome. The skin was cut into surface parallel sections (Zesch and Schaefer, 1974Zesch A. Schaefer H. Penetration kinetics of four drugs in the human skin.Acta Derm Venereol (Stockh). 1974; 54: 91-98PubMed Google Scholar) of 25 µm thickness and collected in 12 centrifuge tubes according to the following scheme: incomplete cuts, 4 × 2 × 25 µm, 4 × 4 × 25 µm, 2 × 8 × 25 µm, and residual skin rest. Using a cross-section of the skin of each subject, the SC thickness' was determined by a light microscopic procedure. The number of tape strips was linearly correlated with the thickness of the SC. The thickness of the incomplete cuts and the skin rests was calculated over their weights relating to a standard cut with known weight and thickness. The flufenamic acid was extracted from the adhesive tape, the skin cuts and the filter paper with 1.5 ml 0.05 M NaOH by shaking for 2 h at room temperature and centrifugation. For details concerning the high-performance liquid chromatography procedure seeWagner et al., 2000Wagner H. Kostka K.H. Lehr C.-M. Schaefer U.F. Drug distribution in human skin using two different in vitro test systems. Comparison with in vivo data.Pharm Res. 2000; 17: 1475-1481Crossref PubMed Scopus (105) Google Scholar. Table I illustrates the flufenamic acid amounts that penetrated into the SC (Table Ia) and the DSL (Table Ib) under in vivo and in vitro conditions. Each pair of data represents varying experimental times caused by the routine in the hospital as well as emergency operations or unexpected problems with the operation before the “in vivo” experiment. The planned incubation intervals could not be maintained in all cases.Table IValues of the drug amounts penetrated into the SC and the DSL in vivo (n = 1) as well as in vitro (n = 2) (ng per cm2) in dependence on the incubation time (min) and the test systemSubjectTime (min)In vivo (ng per cm2)In vitro SB-M (ng per cm2)In vitro FD-C (ng per cm2)(a) Stratum corneum (SC)A95776.6794.7–1049.7607.3–698.9185664.31087.2–1400.7977.6–1095.9B65742.51259.1–1303.7641.0–769.5155452.41121.8–1287.0748.2–832.6C (+ 75 min)48286.8175.5–437.6333.4–365.6108261.3277.8–329.1280.3–344.3D35621.3458.9–533.4310.5–758.8190424.7680.1–728.2611.9–717.1E (+ 60 min)80396.6602.0–803.7268.6–414.6170531.7614.9–868.8762.4–859.7F75391.1511.5–765.91088.0–1088.9165458.11024.9–1489.1733.2–845.2G (+ 40 min)45525.7686.3–818.7393.5–432.21351406.0511.6–735.3570.5–661.1H100496.2488.9–494.5433.2–479.2190200.8574.4–911.6424.8–442.9(b) Deeper skin layers (DSL)A950809.6–1270.10–0185626.41659.2–1920.9610.6–777.8B65427.6660.9–983.3117.7–135.11551191.03028.4–4034.21580.9–2408.3C (+ 75 min)48930.11407.3–1548.5625.4–634.51081156.02667.3–3266.51378.0–2033.6D35276.41197.2–1287.80–01901074.14394.6–5146.12470.8–2877.3E (+ 60 min)8076.5559.3–676.30–0170581.02976.5–3027.8654.6–953.9F75674.81274.2–2141.61110.9–1655.91652015.05652.5–7001.93175.7–3287.4G (+ 40 min)45435.3727.0–740.939.5–57.71351054.53779.8–4401.51242.2–2048.6H1001080.92689.2–3676.51153.2–1339.41901860.75985.3–8623.82664.1–3225.8 Open table in a new tab In contrast to in vivo data obtained earlier in our laboratory using the tape stripping technique without any further disinfection of the incubated skin area (Wagner et al., 2000Wagner H. Kostka K.H. Lehr C.-M. Schaefer U.F. Drug distribution in human skin using two different in vitro test systems. Comparison with in vivo data.Pharm Res. 2000; 17: 1475-1481Crossref PubMed Scopus (105) Google Scholar), in five of eight subjects the in vivo drug amounts present in the SC after the longer incubation time were lower than the ones after the shorter incubation time. Using the in vitro test systems, such circumstances were only found once for the SB-M and the FD-C, respectively. In the other in vitro investigations, the drug amounts became higher with increasing incubation times, or were more or less equal to each other (considering the relatively high span width), indicating that steady-state conditions might be reached within the SC. To examine whether the use of the disinfectant was responsible for the obtained data, in vitro experiments with and without disinfectant were carried out. The results are shown in Figure 3 demonstrating two very different drug concentration–SC depth profiles that did not leave any doubt about the influence of the Braunoderm solution. In the DSL (Table Ib) statistically significant differences could be noticed at six data points between the two in vitro models (Student's t test; p < 0.05). In all subjects, the flufenamic acid amounts detected in the FD-C were lower than those observed using the SB-M. Concerning the drug amounts in the DSL in vivo, the values for all time points were also lower than those observed with the SB-M. In contrast to these results the drug amounts achieved with the FD-C after the shorter time intervals were in five of eight cases even lower than the ones achieved under in vivo conditions. No flufenamic acid was found in either the filter paper beneath the skin incubated in the SB-M nor in the Sørensen phosphate buffer pH 7.4 used as an additional acceptor medium in the FD-C. The main difference between the in vivo and in vitro experiments is the absence of perfusion within the DSL in the in vitro test series. The blood circulation still present under in vivo conditions causes a high tissue clearance by the blood stream and prevents the increase of higher flufenamic acid amounts not only in the SC, but also in the DSL. Using the FD-C, the tissue clearance may be simulated by the buffer system beneath the skin in the acceptor compartment. But the experimental conditions are nevertheless different from in vivo ones, because the drug has to permeate through the whole skin before an effect comparable with the transport with the blood stream takes place. Therefore, the diffusion distance is much longer. Besides, the diffusion of the drug may be affected by the diffusion of the water from the aqueous part of the buffer solution in the acceptor compartment. Both processes run in opposite directions and possibly hinder each other. In addition, the buffer substances change the pH within the DSL to a more neutral one (formerly frozen skin and skin incubated in the SB-M: pH ≈ 8.1; skin incubated in the FD-C using the Sørensen phosphate buffer pH 7.4 as the acceptor medium: pH ≈ 7.4) and, therefore, lead to a decreased solubility of the penetrating drug (unpublished data). These events can be the reason why the drug amounts in the FD-C are always lower than the ones in the SB-M, although a higher water content in the tissue and especially in the SC leads to reduced barrier functions of the SC and, therefore, higher drug amounts within the skin should be obtained (Wiedmann, 1988Wiedmann T.S. Influence of hydration on epidermal tissue.J Pharm Sci. 1988; 77: 1037-1041Crossref PubMed Scopus (12) Google Scholar;Alonso et al., 1996Alonso A. Meirelles N.C. Yushmanov V.E. Tabak M. Water increases the fluidity of intercellular membranes of stratum corneum. Correlation with water permeability, elastic and electrical resistance properties.J Invest Dermatol. 1996; 106: 1064-1069Crossref PubMed Scopus (107) Google Scholar). In the SB-M, unphysiologic hydration of the tissue is avoided; however, the experimental design of this model leads to restricted incubation times, because only the tissue itself acts as an acceptor for the penetrating drug. In this case, no further clearance arises. Concerning the absence of drug in the filter paper beneath the skin in the SB-M as well as in the buffer solution in the FD-C, these data indicate that sink conditions were maintained for all time intervals investigated. Apart from the mentioned reasons for the different results it has to be kept in mind that some problems concerning the course of the in vivo experiments had to be overcome, which also influenced the drug amount present in the skin and had to be simulated in an appropriate way under in vitro conditions. In two cases due to high traffic volume or a late notification from the hospital the arrival at the hospital was delayed. The freezing of the skin biopsies was not punctual after the excision and led to delays of 40–75 min (Table I). Therefore, diffusion processes could not be stopped immediately. Subsequently, there was a lower drug amount in the SC. An additional problem was the single omission of the operating team to remove the subcutaneous fatty tissue of the incubated skin biopsies. The freezing of these biopsies was slower and penetration of the drug within the skin specimens could not be completely avoided. As mentioned in the Results concerning the in vitro experiments, the differences between the drug amounts detected in the SC after two different incubation times for one skin flap were sometimes low, or the longer incubation times led to less drug amount than the shorter incubation time. The hypothesis was put forward that quasi steady-state conditions in the SC might be reached. Investigations concerning the calculation of quasi steady-state conditions within the SC had been carried out previously (Wagner et al., 2000Wagner H. Kostka K.H. Lehr C.-M. Schaefer U.F. Drug distribution in human skin using two different in vitro test systems. Comparison with in vivo data.Pharm Res. 2000; 17: 1475-1481Crossref PubMed Scopus (105) Google Scholar) leading to time intervals from 6.6 ± 3.0 up to 27.0 ± 14.4 min, after which half of the quasi steady-state drug amount in the SC was reached. These time intervals were relatively short and might explain this phenomenon. Concerning the SC an in vivo/in vitro correlation failed using the represented methodology, although previous studies demonstrated that a correlation could be achieved (Wagner et al., 2000Wagner H. Kostka K.H. Lehr C.-M. Schaefer U.F. Drug distribution in human skin using two different in vitro test systems. Comparison with in vivo data.Pharm Res. 2000; 17: 1475-1481Crossref PubMed Scopus (105) Google Scholar). The main difference of both studies is the application of disinfectant after the removal of the incubated drug preparation. As illustrated in Figure 3 the disinfectant has a severe effect on the drug amount present in the SC and the drug concentration–SC depth profile. Within the first two-thirds of the SC the drug amount was reduced by 80%. For the DSL, the described influence of the disinfectant was considered negligible as it was within the variation of the data (detailed data not shown). Therefore, an in vivo/in vitro correlation can be demonstrated for the DSL shown in Figure 4. A direct linear correlation between in vivo and in vitro data was found for both in vitro test systems leading to correlation coefficients of r = 0.910 (SB-M vs in vivo; slope = 3.2272 ± 0.3935) and r = 0.927 (FD-C vs in vivo; slope = 1.7776 ± 0.1926) in which both slopes are statistically significantly different from each other (Student's t test; p < 0.01). This difference illustrates on the one hand the varying influence of the in vitro test systems on the drug penetration within the skin. On the other hand, the calculated slopes represent a factor about which in vivo and in vitro data are shifted against each other. The described experimental design succeeded in establishing an in vivo/in vitro correlation in the dermal field of research using human skin from the same subject for in vivo as well as in vitro experiments. This correlation was convincing for the DSL, whereas earlier data for the SC could not be confirmed. Reasons for the failure of the correlation within the SC could be attributed to the influence of the disinfectant. The results obtained with both in vitro test systems correlate in a satisfactory way with the in vivo data, but the in vitro measurements overpredict the drug amounts in the DSL (SB-M > FD-C). Nevertheless, this methodology could permit the prediction of drug amounts penetrating into the DSL under in vivo conditions after different incubation times as far as the lipophilic model drug flufenamic acid is concerned. Besides, it should have the potential to decrease the number of in vivo experiments followed by a reduction of the risks for volunteers and costs for the development of new drug preparations.