DRUG DISPOSITION
Clin. Pharmacokinet. 23 (4): 253-266. 1992 0312-5963(92(00 I0-0253($07.00(0 © Adis International Limited. All rights reserved. CPK1220
Percutaneous Absorption of Drugs Ronald C. Wester and Howard I. Maibach Department of Dermatology. University of California School of Medicine, San Francisco, California
Contents 253 254 254 256 257 257 258 259 260 261
262 263 264 265
Summary
Summary I. Methodology 1.1 In Vivo Methodology 1.2 In Vitro Methodology 1.3 Animal Models 2. Individual and Regional Variation 3. Race 4. Single Versus Multiple Topical Administration 5. Diseased Skin 6. Skin Decontamination 7. Dose-Response Relationships 8. Accountability (Mass Balance) 9. Percutaneous Absorption and Drug Development 10. Steps to Percutaneous Absorption
The skin is an evolutionary masterpiece of living tissue which is the final control unit for determining the local and systemic availability of any drug which must pass into and through it. In vivo in humans, many factors will affect the absorption of drugs. These include individual biological variation and may be influenced by race. The skin site of the body will also influence percutaneous absorption. Generally, those body parts exposed to the open environment (and to cosmetics, drugs and hazardous toxic substances) are most affected. Treating patients may involve single daily drug treatment or multiple daily administration. Finally, the body will be washed (normal daily process or when there is concern about skin decontamination) and this will influence percutaneous absorption. The vehicle of a drug will affect release of drug to skin. On skin, the interrelationships of this form of administration involve drug concentration, surface area exposed, frequency and time of exposure. These interrelationships determine percutaneous absorption. Accounting for all the drug administered is desirable in controlled studies. The bioavailability of the drug then is assessed in relationship to its efficacy and toxicity in drug development. There are methods, both quantitative and qualitative, in vitro and in vivo, for studying percutaneous absorption of drugs. Animal models are substituted for humans to determine percutaneous absorption. Each of these methods thus becomes a factor in determining percutaneous absorption because they predict absorption in humans. The relevance of these predictions to humans in vivo is of intense research interest. The most relevant determination of percutaneous absorption of a drug in humans is when
Clin. Pharmacokinet. 23 (4) 1992
254
the drug in its approved formulation is applied in vivo to humans in the intended clinical situation. Deviation from this scenario involves the introduction of variables which may alter percutaneous absorption.
In vivo percutaneous absorption is a complex biological process. The skin is a multilayered biomembrane with particular absorption characteristics. If the skin were a simple membrane, absorption parameters could easily be measured and would be fairly constant, provided there was no change in the chemistry of the membrane. However, skin is a dynamic, living tissue and as such its absorption parameters are susceptible to constant change. Many factors and skin conditions can rapidly change the absorption parameters. Additionally, since skin is a living tissue it will change through its own growth patterns and this change will also be influenced by many factors. When dealing with percutaneous absorption, the skin should not be regarded as an inert membrane. Instead, the skin should be viewed as a dynamic, living biomembrane with unique properties. We review some of the more relevant factors to influence the percutaneous absorption of drugs.
1. Methodology There are various methods to determine percutaneous absorption. Since not all methods are equal they can give different results. Thus, a major factor affecting percutaneous absorption is methodology. 1.1 In Vivo Methodology
1.1.1 Systemic Bioavailability (Blood and Excreta) Percutaneous absorption in vivo is usually determined by the indirect method of measuring radioactivity in excreta following topical application of the labelled compound. In human studies, plasma concentrations of the compound are extremely low following topical application and are often below the assay detection limit, so it is necessary to use tracer methodology. The labelled
compound, usually carbon-l 4 or tritium, is applied to the skin. The total amount of radioactivity excreted in urine (or urine plus faeces) is then determined. The amount of radioactivity retained in the body or excreted by some route not assayed (C02, sweat) is corrected by determining the amount of radioactivity excreted following parenteral administration. This final amount of radioactivity is then expressed as the percentage of the applied dose that was absorbed. The equation used to determine percutaneous absorption is: Percutaneous absorption = (total radioactivity following topical administration/total radioactivity following parenteral administration) X 100 Determination of percutaneous absorption from urinary radioactivity does not account for metabolism by the skin. The radioactivity in urine is a mixture of the parent compound and metabolites. Plasma radioactivity can be measured and the percutaneous absorption determined by the ratio of the areas under the plasma concentration-time curves following topical and intravenous administration. Radioactivity in blood and excreta can include both the applied compound and metabolites. If the metabolism by the skin is extensive and different from that of other systemic tissues then this method is not valid, because the pharmacokinetics of the metabolites can be different from that of the parent compound. However, in practice, this method has given results similar to those obtained from urinary excretion (Wester et al. 1983). The only way to determine the absolute bioavailability of a topically applied compound is to measure the compound by specific assay in blood or urine after topical and intravenous administration. This is difficult, since plasma concentrations after topical administration are often very low. However, as more sensitive assays are developed, estimates of absolute topical bioavailability will
Percutaneous Drug Absorption
become a reality. A comparison of these methods was performed by using [14C]nitroglycerin in Rhesus monkeys (Wester et al. 1983). The difference between the estimate of absolute bioavailability (56.6%) and that of 14C (72.7 to 77.2%) is the percentage of compound metabolised in the skin as the compound was being absorbed. For nitroglycerin, this is about 20%.
1.1.2 Surface Disappearance Another approach used to determine in vivo percutaneous absorption is to measure the loss of radioactive material from the surface as it penetrates the skin. Recovery of an ointment or solution following skin application is difficult because total recovery from the skin is never assured. With topical application of a transdermal delivery device, the total unit can be removed from the skin and the residual amount of drug in the device can be determined. The difference between the applied and the residual dose is assumed to be the amount of drug absorbed. We must be aware that the skin may act as a reservoir for unabsorbed material.
1.1.3 Biological Response Another in vivo method of estimating absorption is to use a biological or pharmacological response (McKenzie & Stoughton 1962). Here, a biological assay is substituted for a chemical assay and absorption is estimated. An obvious disadvantage to the use of biological response is that it is only good for compounds that will elicit an easily measurable response. An example of a biological response would be the vasoconstrictor assay when the balancing effect of I compound is compared with that of a known compound. This method is perhaps more qualitative than quantitative. Other qualitative methods of estimating in vivo percutaneous absorption include whole body autoradiography and fluorescence. Whole body autoradiography will give an overall picture of dermal absorption followed by the involvement of other body tissues with the absorbed compound.
255
1.1.4 Skin Stripping: Short Term Exposure The stripping method determines the concentration of chemical in the stratum corneum at the end of a short application period (30 min) and by linear extrapolation predicts the percutaneous absorption of that chemical for longer application periods. The chemical is applied to the skin of animals or humans and after the 30 min application time the stratum corneum is removed by successive cellophane tape application and removal. The tape strippings are assayed for chemical content. Rougier and co-workers (1983, 1986) established a linear relationship between this stratum corneum reservoir content and percutaneous absorption using the standard urinary excretion method. The major advantage of this method are, first, the elimination of urinary (and faecal) excretion to determine absorption and, secondly, the applicability to nonradiolabelled determination of percutaneous absorption because the skin strippings contain adequate chemical concentrations for nonlabelled assay methodology. This is an exciting new system for which more research is needed to establish limitations. 1.1.5 Skin Flaps This method involves surgically isolating a section of skin on an animal such that the blood supply is singular and this singular source can be used to collect chemical in blood as the chemical absorbs through skin. The isolated skin section can be used to study percutaneous absorption while it is intact on the animal in vivo or the skin section with its intact blood vessels can be mounted in an in vitro perfusion system to study percutaneous absorption. The isolated perfused porcine skin flap (IPPSF) is surgically created on a pig and then the viable flap with intact blood supply can be mounted in an in vitro perfusion system. The absorption of chemicals through skin and metabolism within the skin can be determined by assay of the blood vessel perfusate. The IPPSF model offers advantages of being an alternative to an in vitro animal model and that metabolism of chemicals penetrating the skin can be determined (Riviere et al. 1986, 1987). The skin sandwich flap (SSF) is an island flap that
256
has split-thickness (dermatoned) skin grafted to its subcutaneous surface directly under the superficial epigastric vasculature. In this setting, the dermis of the donor skin and subcutaneous tissue of the host flap grow together sandwiching the vessels supplying the flap (the superficial epigastric vessels). Two additional steps allow this sandwich to be converted to an island sandwich flap, which is isolated on its vasculature and transferred to the rat's back by a series of surgical procedures. The juncture on the femoral vessels supplying and draining the flap can be readily visualised with an incision in the groin and is accomplished routinely. The exposed vein draining the flap tolerates multiple venipunctures. The SSF can be constructed with either human, pig or rat skin as the donor skin (Pershing & Krueger 1989). 1.2 In Vitro Methodology
1.2.1 In Vitro Diffusion Cell The most commonly used in vitro technique involves placing a piece of excised skin in a diffusion chamber, applying a radioactive compound to I side of the skin and then assaying for radioactivity in the collection vessel on the other side (Bucks et al. 1985b). Excised human or animal skin may be used and the skin can be wholly intact or separated into epidermis or dermis. Artificial membranes can be used in place of skin to measure diffusion pharmacokinetics. The advantages of the standard in vitro technique are that the method is easy to use and the results are obtained quickly. The disadvantage is that the fluid in the collection bath which bathes the skin is buffered saline, which may be appropriate for studying hydrophilic compounds but is not suitable for hydrophobic compounds. Absorption of triclocarban in a standard static system in vitro was 0.12 ± 0.05% of the applied dose through human adult abdominal skin. In contrast, in humans in vivo the absorption was 7.0 ± 2.8%. The discrepancy appeared to be due primarily to the insolubility oftriclocarban in the small volume of saline used in the reservoir of the static system. By changing to a continuous flow system, in which the volume of saline was greatly increased, the sol-
Clin. Pharmacokinet. 23 (4) 1992
ubility of triclocarban was no longer the limiting factor in absorption and the extent of absorption in vitro approached that of absorption in vivo (Wester et al. 1985). 1.2.2 Powdered Human Stratum Corneum This is an in vitro model that uses the partition coefficient of the chemical contaminant in water or other vehicle with that of powdered human stratum corneum. Adult foot calluses are ground with dry ice and freeze-dried to form a powder. That portion of the powder that passed through a 40mesh but not an 80-mesh sieve is used. The chemical (radio labelled) as a solutio~ in 1.5ml water or other vehicle is mixed with 1.5mg powdered human stratum corneum and the mixture is allowed to sit for 30 min. The mixture is then centrifuged and the proportions of chemical bound to human stratum corneum and that remaining in water are determined by scintillation counting or some other analytical method (Wester et al. 1987). The capacity of skin and soil for cadmium was studied with binding studies. Cadmium chloride in water (116 ppb) was partitioned against Img of soil and against I mg of powdered human stratum corneum. Table I shows the percentage of the dose in water and in matter (soil or powdered human stratum corneum). Soil has a relatively higher affinity for cadmium than does stratum corneum. This correlates with data in studies where the skin absorption is greater from water than from soil (soil binding capacity relative to skin).
Table I. The partitioning of cadmium chloride (1 ppb) between water 1.0ml and powdered human stratum corneum 1.0mg and between water and soil 1.0mg over 30 min, followed by centrifugation. N = 3 for each measurement
Test substance
% Dose
Water Stratum corneum Total
68.6 ± 5.6 33.2 ± 3.8 101.8 ± 3.3
Water Soil Total
9.3 ± 1.4 82.5 ± 1.0 91.8 ± 1.8
257
Percutaneous Drug Absorption
Table II. In vivo percutaneous absorption in Rhesus monkey and human, and in rat and human Compound
Difference
% Dose absorbed animal
human
Rhesus monkey Benzoic acid Cortisone DDT Diethyl maleate 2,4-Dinitrochlorobenzene Hydrocortisone Nitrobenzene Retinoic acid Testosterone
60 5 19 68 52 3 4 2 18
± ± ± ± ± ± ± ± ±
8 3 9 7 4 1 1 1 10
43 3 10 54 54 2 2 1 13
± ± ± ± ± ± ± ± ±
16 2 4 7 6 2 1 0.2 3
0 x2 x2 0 0 0 x2 x2 0
Rat Butter yellow Caffeine Cortisone Haloprin Lindane N-Acetylcystein Parathion Testosterone Trichlorocarbanilide
48 53 25 96 31 4 95 47 16
± ± ± ± ± ± ± ± ±
2 12 4 14 10 4 3 3 9
22 48 4 11 9 2 46 13 7
± ± ± ± ± ± ± ± ±
5 21 2 4 4 2 5 3 3
x2 0 x6 x9 x3 x2 x2 x2 x2
1.3 Animal Models
The ideal way to determine the percutaneous absorption of a compound in humans is to do the actual study in humans. Mechanisms and parameters of percutaneous absorption elucidated in vivo with human skin are most relevant to the clinical situation. However, many compounds are potentially too toxic to test in vivo in humans and so their percutaneous absorption must be tested in animals. Likewise, until more complete animal-tohuman validation studies become available not all investigators will have access to human volunteers. Mechanism studies and studies of factors affecting absorption must, therefore, be explored using animals and in vitro techniques. There are 2 basic criteria to judge whether an animal model is good. First, does the animal model give the same percutaneous absorption as that in humans? If this is not possible, then the animal model should be consistently different from that in humans. The Rhesus monkey is a good animal model, while an ex-
ample of a poor animal model is the rat (table II). Generally, Rhesus monkey and miniature pig are good animal models for human absorption. The smaller laboratory animals (rabbit, rat, mouse) are not good animal models (Wester & Maibach 1985).
2. Individual and Regional Variation Percutaneous absorption in humans and animals in vivo shows individual differences and regional variation. Feldmann and Maibach (1967) were the first to show this in human volunteers and the concept has been shown to be true for animals (Wester et al. 1980a). In vivo percutaneous absorption variation has been ascribed to a multiplicity of events such as skin thickness, blood flow, lipid content, number of hair follicles, etc. Some of these factors are changed when skin is removed from the body and placed in a diffusion cell. Table III shows the differences when skin from the abdomen and thigh of different individuals are tested for their ability to absorb a test compound in 5
258
Clin. Pharmacokinet. 23 (4) 1992
Table III. Individual variation in in vitro percutaneous absorption, as exemplified by mean (± SO) percentage recovery of the test compound from human skin, a surface wash and the buffered saline receptor fluid Skin source
Skin
Surface wash
Receptor fluid
Total recovery
68y White male abdomen 69y White male thigh 33y White male abdomen
5.0 ± 2.4 3.2 ± 2.8 2.6 ± 1.0
85.7 ± 7.8 83.7 ± 9.5 72.1 ± 12.3
2.5 ± 4.5 0.3 ± 0.2 4.4 ± 5.0
93.2 ± 6.0 87.7 ± 8.1 79.0 ± 12.5
different formulations in vitro. The buffered saline receptor fluid accumulation in the 3 skin sources shows that the barrier properties of the second human skin source were such that no skin absorption occurred. A formulation comparison with only that human skin source would have provided completely negative data. Therefore, especially with in vitro percutaneous absorption it is best to use a selection of human skin sources (just as an in vivo study would use many individuals). 3. Race
Differences in skin permeability exist between patients and often therapy needs to be adapted to these differences. Racial differences in skin exist and the literature suggests that skin from people of difference races may have different responses to various chemical stimuli. Weigan and Gaylor (1974) using dinitrochlorobenzene found decreased minimal perceptible erythema in Blacks compared with Caucasians. Gean et al. (1989) suggest that racial differences in response to topical methyl nicotinate exist, but that perception of these distinctions may depend upon the method of measurement. Racial differences in physicochemical properties of the skin are also reported. In Black vs Caucasian skin, there is increased resistance to tape stripping (Weigan & Gaylor 1974) [see section 1.1.4], increased electrical resistance (Johnson & Corah 1960) and increased lipid content (Reinertson & Wheatley 1959). Wedig and Maibach (1981) reported that in vivo percutaneous penetration of dipyrithione was less in Blacks than in Whites. Aside from Black vs Caucasian skin, the statistical analysis included such variables as different skin sites (forehead, scalp,
forearm), different vehicles (methanol, cream, shampoo) and different skin treatments (untreated, stripped, hair clipped). Each of the variables themselves - site, vehicle, treatment - will affect percutaneous absorption (Wester & Maibach 1983). If the reported data of Wedig and Maibach (1981) are examined between Blacks and Caucasian only within the same site, vehicle and treatment, there are no differences in percutaneous absorption between Blacks and Caucasians. This agrees with the recent study of Wester et al. (1990), who examined the percutaneous absorption of 3 test chemicals to determine if racial differences exist. Figure I shows that for the chemicals (benzoic acid, caffeine and acetylsalicylic acid) there was no difference in in vivo percutaneous absorption in Caucasian, Asian and Black individuals. Racial differences in biological responses to those chemicals were not addressed by this study. In vivo, percutaneous absorption of diflorasone di30
o Caucasian o Black • Asian
~ !!.... 20
~ 0
CI)
.0
'" Q)
'"0
10
0
Benzoic acid
Caffeine
AspIrin
Fig. 1. Influence of race on in vivo percutaneous absorption in humans.
259
Percutaneous Drug Absorption
acetate was the same in Black and White study participants (Wickrema et at. 1978). However, Berardesca and Maibach (1987) and Wilson and coworkers (1988) reported increased transepidermal water loss (TEWL) in Blacks and Hispanics. A study by Williams et at. (1991) suggest racial differences in nitroglycerin absorption after transdermal application. Black study participants had lower drug concentrations than Caucasians and Asians. A pharmacological response depends on the bioavailability (percutaneous absorption in topical therapy) of a chemical and the inherent activity of the chemical once it is in a biological system (Wester & Maibach 1983). Berardesca et at. (1991) found decreased vasodilation after nicotinate exposure in Blacks compared with Caucasians and an altered hyperaemia reaction after constrictive stimuli. Conventional thought maintains that Black skin resists chemical irritation better than White skin. Experimental support for the statement does exist, but most is based on a visual end-point measurement (i.e. erythema assessment). This end-point has been questioned because erythema is more difficult to detect in Black skin; hence, differential irritancy response between Blacks and Whites may be less pronounced than expected. In fact, objective measurements of skin blood flow by laser Doppler velocimetry showed no differences between Black and Caucasian individuals in methyl nicotinate-induced vasodilation (Guy et at. 1985). Recently, Gean et at. (1989) demonstrated that Asian skin did not differ from either Black or White skin in response to ethyl nicotinate. On the basis of their results they recognised that it was impossible to make a case for differential permeability of different skin types unless it is argued that poor percutaneous absorption in I racial group is exactly compensated by increased microvasculature sensitivity, which seems implausible. Percutaneous absorption studies show different patterns of penetration depending on the molecules under investigation. Thus, in vitro no differences were recorded in water permeation through White or Black skin (Bronaugh et at. 1986) whereas permeation of fluocinolone acetonide was higher through normal-appearing White skin than through
o Theoretical
0.5
Observed N
04
8
0;
.:; 0.3
~
:0
~ 0.2 >
'"o iIi 0.1 Single dose
Single dose x 3
Triple therapy
Triple therapy
Fig. 2. Hydrocortisone skin delivery in vivo in humans with triple therapy. Three daily doses of hydrocortisone in solvent vehicle deliver more drug than predicted from daily dose.
Black skin (excised from legs amputated for tumours or gangrene) [Stoughton 1969].
4. Single Versus Multiple Topical Administration Topical application of hydrocortisone and other corticosteroids frequently uses repeated, rather than single, applications of drug to the skin. It is commonly assumed that multiple applications of hydrocortisone effectively increase its bioavailability and absorption. A long term multiple-dose study in Rhesus monkeys by Wester et at. (l980b) indicated that this was true. However, short term experiments in the Rhesus monkey by Wester et at. (1977) and long term pharmacokinetic assays by Bucks et at. (l985a) showed no increase in hydrocortisone absorption following multiple administration. An investigation was designed to determine if mUltiple-dose therapy would increase drug bioavailability in human skin (Melendres et at. 1990). Percutaneous absorption of hydrocortisone was measured in 6 healthy adult male volunteers. The study compared a single topical dose to treatment with multiple topical doses (I vs 3 applications) in the same day. [14C]Hydrocortisone in acetone was applied to 2.5 cm 2 of ventral forearm skin and pro-
260
Clin. Pharmacokinel. 23 (4) 1992
tected with a nonocclusive polypropylene chamber. The amount of l4C measured in urine collected over 7 days was used to determine hydrocortisone absorption. The treatments, performed 2 to 3 weeks apart, each used adjacent sites on the same individuals. A single dose of hydrocortisone 13.33 ~g/cm2 delivered a mass of 0.056 ~g/cm2. Three serial doses of 13.33 ~g/cm2 (total 40 ~g/cm2) were expected to deliver 0.168 ~g/cm2 with or without soap-and-water washing between doses, but the observed amount of hydrocortisone delivered significantly exceeded the expected mass absorbed. This indicates that multiple administration resulted in a significant increase in bioavailability (fig. 2). It is postulated that increased vehicle application and washing dissolved and mobilised previously administered hydrocortisone and increased bioavailability.
Removal
o Soap and water
100
Water only
80
~
'"'"0
60
"0
'0 c:
.9
13
40
~
u.
20
0
First wash
Second wash
Third wash
Recovery
80
5. Diseased Skin
J~
Jt-
F-'-
F-'-
The assumption has been that percutaneous absorption is enhanced in diseased/damaged skin and that the ability of the skin to protect against intrusion by chemicals is impaired. The picture of skin floodgates opening and chemicals pouring in is certainly not warranted (except perhaps in patients with severe burns). The data suggest that diseased
Co
60
~
'"'"0 "0 '0
Co Co
Co
40
c:
.9
13
~
u.
20
o
80
Total
o
0.5
3
6
24
TIme (h)
Fig. 4. (a) Glyphosate, a water-soluble compound, is easily removed from Rhesus monkey skin in vivo with both water only and soap and water. (b) Time-response in vivo recovery of glyphosate from Rhesus monkey skin.
Time (h)
Fig. 3. With time, isofenphos disappears from human skin in vivo. This is due to a combination of isofenphos volatility and removal by rubbing from clothing.
skin can retain barrier properties and that differences will exist for different drugs and for different disease conditions. In psoriasis, there are definite formulation effects and treatment (e.g. salicylic acid scrub) probably will affect drug delivery into the skin. However, even for a compound such as hydrocortisone, for which skin absorption is low in diseased and damaged skin, enhanced absorption
Percutaneous Drug Absorption
261
will occur in severe conditions. For more potent corticosteroids, the risk of systemic adverse effects such as adrenal suppression remains real. However, data for percutaneous absorption in diseased human skin is limited. More research in this area is warranted (Wester & Maibach 1992).
Removal
80
o Soap and
water Water only
60
6. Skin Decontamination "0
A time-recovery study was done to determine the ability of soap and water wash to remove isofenphos and to determine evaporation of this agent from human skin (Wester et al. 1992). Figure 3 shows that at time zero (approximately 5 to 10 min needed to administer the dose and then wash 4 subjects) a total of 61.4 ± 10.4% of the applied dose could be recovered from the study participants. Recovery decreased slightly to 54.1 ± 12.9% at Ih and 54.9 ± 14.2% by 4h and recovery was down to 30.3 ± 11.1 % at 8h. Recovery at 24h was minimal; 0.53 ± 0.17%. Recovery at 24h (0.75 ± 0.93%) was also minimal. Most of the wash recovery occurred with the first soap application. Many compounds exhibit little skin surface recovery after 24h of skin application time, despite vigorous washing and tape stripping. In the case of isofenphos, chemical volatility of the agent was the answer. In other situations, clothing will rub off the topically applied chemical. Other compounds such as salicylic acid have the ability to be retained in skin. The in vivo percutaneous absorption of salicylic acid in humans is 6.5 ± 5.0% of a dose after 24h skin application. The 24h postapplication skin wash is able to recover 53.4 ± 6.3%, most with the first soap application. There are 2 components to skin washing in the recovery of drug. One is the physical rubbing and removal of the skin surface. The second component is the solvent action of soap and water. Glyphosate is a water-soluble compound. Its removal from skin with water alone or soap and water is the same (fig. 4) [Wester et al. 1991b]. In contrast, alachlor is lipid soluble. Therefore, more alachlor can be removed from skin with soap present, rather than water alone (fig. 5). Alachlor partitions into powdered human stratum corneum from its ve-
40
c .Q
U
u:'" 20
0
First wash
Third wash
wash
Recovery
100
80 ~ L Ql
Clin. Pharmacokinet. 23 (4): 253-266. 1992 0312-5963(92(00 I0-0253($07.00(0 © Adis International Limited. All rights reserved. CPK1220
Percutaneous Absorption of Drugs Ronald C. Wester and Howard I. Maibach Department of Dermatology. University of California School of Medicine, San Francisco, California
Contents 253 254 254 256 257 257 258 259 260 261
262 263 264 265
Summary
Summary I. Methodology 1.1 In Vivo Methodology 1.2 In Vitro Methodology 1.3 Animal Models 2. Individual and Regional Variation 3. Race 4. Single Versus Multiple Topical Administration 5. Diseased Skin 6. Skin Decontamination 7. Dose-Response Relationships 8. Accountability (Mass Balance) 9. Percutaneous Absorption and Drug Development 10. Steps to Percutaneous Absorption
The skin is an evolutionary masterpiece of living tissue which is the final control unit for determining the local and systemic availability of any drug which must pass into and through it. In vivo in humans, many factors will affect the absorption of drugs. These include individual biological variation and may be influenced by race. The skin site of the body will also influence percutaneous absorption. Generally, those body parts exposed to the open environment (and to cosmetics, drugs and hazardous toxic substances) are most affected. Treating patients may involve single daily drug treatment or multiple daily administration. Finally, the body will be washed (normal daily process or when there is concern about skin decontamination) and this will influence percutaneous absorption. The vehicle of a drug will affect release of drug to skin. On skin, the interrelationships of this form of administration involve drug concentration, surface area exposed, frequency and time of exposure. These interrelationships determine percutaneous absorption. Accounting for all the drug administered is desirable in controlled studies. The bioavailability of the drug then is assessed in relationship to its efficacy and toxicity in drug development. There are methods, both quantitative and qualitative, in vitro and in vivo, for studying percutaneous absorption of drugs. Animal models are substituted for humans to determine percutaneous absorption. Each of these methods thus becomes a factor in determining percutaneous absorption because they predict absorption in humans. The relevance of these predictions to humans in vivo is of intense research interest. The most relevant determination of percutaneous absorption of a drug in humans is when
Clin. Pharmacokinet. 23 (4) 1992
254
the drug in its approved formulation is applied in vivo to humans in the intended clinical situation. Deviation from this scenario involves the introduction of variables which may alter percutaneous absorption.
In vivo percutaneous absorption is a complex biological process. The skin is a multilayered biomembrane with particular absorption characteristics. If the skin were a simple membrane, absorption parameters could easily be measured and would be fairly constant, provided there was no change in the chemistry of the membrane. However, skin is a dynamic, living tissue and as such its absorption parameters are susceptible to constant change. Many factors and skin conditions can rapidly change the absorption parameters. Additionally, since skin is a living tissue it will change through its own growth patterns and this change will also be influenced by many factors. When dealing with percutaneous absorption, the skin should not be regarded as an inert membrane. Instead, the skin should be viewed as a dynamic, living biomembrane with unique properties. We review some of the more relevant factors to influence the percutaneous absorption of drugs.
1. Methodology There are various methods to determine percutaneous absorption. Since not all methods are equal they can give different results. Thus, a major factor affecting percutaneous absorption is methodology. 1.1 In Vivo Methodology
1.1.1 Systemic Bioavailability (Blood and Excreta) Percutaneous absorption in vivo is usually determined by the indirect method of measuring radioactivity in excreta following topical application of the labelled compound. In human studies, plasma concentrations of the compound are extremely low following topical application and are often below the assay detection limit, so it is necessary to use tracer methodology. The labelled
compound, usually carbon-l 4 or tritium, is applied to the skin. The total amount of radioactivity excreted in urine (or urine plus faeces) is then determined. The amount of radioactivity retained in the body or excreted by some route not assayed (C02, sweat) is corrected by determining the amount of radioactivity excreted following parenteral administration. This final amount of radioactivity is then expressed as the percentage of the applied dose that was absorbed. The equation used to determine percutaneous absorption is: Percutaneous absorption = (total radioactivity following topical administration/total radioactivity following parenteral administration) X 100 Determination of percutaneous absorption from urinary radioactivity does not account for metabolism by the skin. The radioactivity in urine is a mixture of the parent compound and metabolites. Plasma radioactivity can be measured and the percutaneous absorption determined by the ratio of the areas under the plasma concentration-time curves following topical and intravenous administration. Radioactivity in blood and excreta can include both the applied compound and metabolites. If the metabolism by the skin is extensive and different from that of other systemic tissues then this method is not valid, because the pharmacokinetics of the metabolites can be different from that of the parent compound. However, in practice, this method has given results similar to those obtained from urinary excretion (Wester et al. 1983). The only way to determine the absolute bioavailability of a topically applied compound is to measure the compound by specific assay in blood or urine after topical and intravenous administration. This is difficult, since plasma concentrations after topical administration are often very low. However, as more sensitive assays are developed, estimates of absolute topical bioavailability will
Percutaneous Drug Absorption
become a reality. A comparison of these methods was performed by using [14C]nitroglycerin in Rhesus monkeys (Wester et al. 1983). The difference between the estimate of absolute bioavailability (56.6%) and that of 14C (72.7 to 77.2%) is the percentage of compound metabolised in the skin as the compound was being absorbed. For nitroglycerin, this is about 20%.
1.1.2 Surface Disappearance Another approach used to determine in vivo percutaneous absorption is to measure the loss of radioactive material from the surface as it penetrates the skin. Recovery of an ointment or solution following skin application is difficult because total recovery from the skin is never assured. With topical application of a transdermal delivery device, the total unit can be removed from the skin and the residual amount of drug in the device can be determined. The difference between the applied and the residual dose is assumed to be the amount of drug absorbed. We must be aware that the skin may act as a reservoir for unabsorbed material.
1.1.3 Biological Response Another in vivo method of estimating absorption is to use a biological or pharmacological response (McKenzie & Stoughton 1962). Here, a biological assay is substituted for a chemical assay and absorption is estimated. An obvious disadvantage to the use of biological response is that it is only good for compounds that will elicit an easily measurable response. An example of a biological response would be the vasoconstrictor assay when the balancing effect of I compound is compared with that of a known compound. This method is perhaps more qualitative than quantitative. Other qualitative methods of estimating in vivo percutaneous absorption include whole body autoradiography and fluorescence. Whole body autoradiography will give an overall picture of dermal absorption followed by the involvement of other body tissues with the absorbed compound.
255
1.1.4 Skin Stripping: Short Term Exposure The stripping method determines the concentration of chemical in the stratum corneum at the end of a short application period (30 min) and by linear extrapolation predicts the percutaneous absorption of that chemical for longer application periods. The chemical is applied to the skin of animals or humans and after the 30 min application time the stratum corneum is removed by successive cellophane tape application and removal. The tape strippings are assayed for chemical content. Rougier and co-workers (1983, 1986) established a linear relationship between this stratum corneum reservoir content and percutaneous absorption using the standard urinary excretion method. The major advantage of this method are, first, the elimination of urinary (and faecal) excretion to determine absorption and, secondly, the applicability to nonradiolabelled determination of percutaneous absorption because the skin strippings contain adequate chemical concentrations for nonlabelled assay methodology. This is an exciting new system for which more research is needed to establish limitations. 1.1.5 Skin Flaps This method involves surgically isolating a section of skin on an animal such that the blood supply is singular and this singular source can be used to collect chemical in blood as the chemical absorbs through skin. The isolated skin section can be used to study percutaneous absorption while it is intact on the animal in vivo or the skin section with its intact blood vessels can be mounted in an in vitro perfusion system to study percutaneous absorption. The isolated perfused porcine skin flap (IPPSF) is surgically created on a pig and then the viable flap with intact blood supply can be mounted in an in vitro perfusion system. The absorption of chemicals through skin and metabolism within the skin can be determined by assay of the blood vessel perfusate. The IPPSF model offers advantages of being an alternative to an in vitro animal model and that metabolism of chemicals penetrating the skin can be determined (Riviere et al. 1986, 1987). The skin sandwich flap (SSF) is an island flap that
256
has split-thickness (dermatoned) skin grafted to its subcutaneous surface directly under the superficial epigastric vasculature. In this setting, the dermis of the donor skin and subcutaneous tissue of the host flap grow together sandwiching the vessels supplying the flap (the superficial epigastric vessels). Two additional steps allow this sandwich to be converted to an island sandwich flap, which is isolated on its vasculature and transferred to the rat's back by a series of surgical procedures. The juncture on the femoral vessels supplying and draining the flap can be readily visualised with an incision in the groin and is accomplished routinely. The exposed vein draining the flap tolerates multiple venipunctures. The SSF can be constructed with either human, pig or rat skin as the donor skin (Pershing & Krueger 1989). 1.2 In Vitro Methodology
1.2.1 In Vitro Diffusion Cell The most commonly used in vitro technique involves placing a piece of excised skin in a diffusion chamber, applying a radioactive compound to I side of the skin and then assaying for radioactivity in the collection vessel on the other side (Bucks et al. 1985b). Excised human or animal skin may be used and the skin can be wholly intact or separated into epidermis or dermis. Artificial membranes can be used in place of skin to measure diffusion pharmacokinetics. The advantages of the standard in vitro technique are that the method is easy to use and the results are obtained quickly. The disadvantage is that the fluid in the collection bath which bathes the skin is buffered saline, which may be appropriate for studying hydrophilic compounds but is not suitable for hydrophobic compounds. Absorption of triclocarban in a standard static system in vitro was 0.12 ± 0.05% of the applied dose through human adult abdominal skin. In contrast, in humans in vivo the absorption was 7.0 ± 2.8%. The discrepancy appeared to be due primarily to the insolubility oftriclocarban in the small volume of saline used in the reservoir of the static system. By changing to a continuous flow system, in which the volume of saline was greatly increased, the sol-
Clin. Pharmacokinet. 23 (4) 1992
ubility of triclocarban was no longer the limiting factor in absorption and the extent of absorption in vitro approached that of absorption in vivo (Wester et al. 1985). 1.2.2 Powdered Human Stratum Corneum This is an in vitro model that uses the partition coefficient of the chemical contaminant in water or other vehicle with that of powdered human stratum corneum. Adult foot calluses are ground with dry ice and freeze-dried to form a powder. That portion of the powder that passed through a 40mesh but not an 80-mesh sieve is used. The chemical (radio labelled) as a solutio~ in 1.5ml water or other vehicle is mixed with 1.5mg powdered human stratum corneum and the mixture is allowed to sit for 30 min. The mixture is then centrifuged and the proportions of chemical bound to human stratum corneum and that remaining in water are determined by scintillation counting or some other analytical method (Wester et al. 1987). The capacity of skin and soil for cadmium was studied with binding studies. Cadmium chloride in water (116 ppb) was partitioned against Img of soil and against I mg of powdered human stratum corneum. Table I shows the percentage of the dose in water and in matter (soil or powdered human stratum corneum). Soil has a relatively higher affinity for cadmium than does stratum corneum. This correlates with data in studies where the skin absorption is greater from water than from soil (soil binding capacity relative to skin).
Table I. The partitioning of cadmium chloride (1 ppb) between water 1.0ml and powdered human stratum corneum 1.0mg and between water and soil 1.0mg over 30 min, followed by centrifugation. N = 3 for each measurement
Test substance
% Dose
Water Stratum corneum Total
68.6 ± 5.6 33.2 ± 3.8 101.8 ± 3.3
Water Soil Total
9.3 ± 1.4 82.5 ± 1.0 91.8 ± 1.8
257
Percutaneous Drug Absorption
Table II. In vivo percutaneous absorption in Rhesus monkey and human, and in rat and human Compound
Difference
% Dose absorbed animal
human
Rhesus monkey Benzoic acid Cortisone DDT Diethyl maleate 2,4-Dinitrochlorobenzene Hydrocortisone Nitrobenzene Retinoic acid Testosterone
60 5 19 68 52 3 4 2 18
± ± ± ± ± ± ± ± ±
8 3 9 7 4 1 1 1 10
43 3 10 54 54 2 2 1 13
± ± ± ± ± ± ± ± ±
16 2 4 7 6 2 1 0.2 3
0 x2 x2 0 0 0 x2 x2 0
Rat Butter yellow Caffeine Cortisone Haloprin Lindane N-Acetylcystein Parathion Testosterone Trichlorocarbanilide
48 53 25 96 31 4 95 47 16
± ± ± ± ± ± ± ± ±
2 12 4 14 10 4 3 3 9
22 48 4 11 9 2 46 13 7
± ± ± ± ± ± ± ± ±
5 21 2 4 4 2 5 3 3
x2 0 x6 x9 x3 x2 x2 x2 x2
1.3 Animal Models
The ideal way to determine the percutaneous absorption of a compound in humans is to do the actual study in humans. Mechanisms and parameters of percutaneous absorption elucidated in vivo with human skin are most relevant to the clinical situation. However, many compounds are potentially too toxic to test in vivo in humans and so their percutaneous absorption must be tested in animals. Likewise, until more complete animal-tohuman validation studies become available not all investigators will have access to human volunteers. Mechanism studies and studies of factors affecting absorption must, therefore, be explored using animals and in vitro techniques. There are 2 basic criteria to judge whether an animal model is good. First, does the animal model give the same percutaneous absorption as that in humans? If this is not possible, then the animal model should be consistently different from that in humans. The Rhesus monkey is a good animal model, while an ex-
ample of a poor animal model is the rat (table II). Generally, Rhesus monkey and miniature pig are good animal models for human absorption. The smaller laboratory animals (rabbit, rat, mouse) are not good animal models (Wester & Maibach 1985).
2. Individual and Regional Variation Percutaneous absorption in humans and animals in vivo shows individual differences and regional variation. Feldmann and Maibach (1967) were the first to show this in human volunteers and the concept has been shown to be true for animals (Wester et al. 1980a). In vivo percutaneous absorption variation has been ascribed to a multiplicity of events such as skin thickness, blood flow, lipid content, number of hair follicles, etc. Some of these factors are changed when skin is removed from the body and placed in a diffusion cell. Table III shows the differences when skin from the abdomen and thigh of different individuals are tested for their ability to absorb a test compound in 5
258
Clin. Pharmacokinet. 23 (4) 1992
Table III. Individual variation in in vitro percutaneous absorption, as exemplified by mean (± SO) percentage recovery of the test compound from human skin, a surface wash and the buffered saline receptor fluid Skin source
Skin
Surface wash
Receptor fluid
Total recovery
68y White male abdomen 69y White male thigh 33y White male abdomen
5.0 ± 2.4 3.2 ± 2.8 2.6 ± 1.0
85.7 ± 7.8 83.7 ± 9.5 72.1 ± 12.3
2.5 ± 4.5 0.3 ± 0.2 4.4 ± 5.0
93.2 ± 6.0 87.7 ± 8.1 79.0 ± 12.5
different formulations in vitro. The buffered saline receptor fluid accumulation in the 3 skin sources shows that the barrier properties of the second human skin source were such that no skin absorption occurred. A formulation comparison with only that human skin source would have provided completely negative data. Therefore, especially with in vitro percutaneous absorption it is best to use a selection of human skin sources (just as an in vivo study would use many individuals). 3. Race
Differences in skin permeability exist between patients and often therapy needs to be adapted to these differences. Racial differences in skin exist and the literature suggests that skin from people of difference races may have different responses to various chemical stimuli. Weigan and Gaylor (1974) using dinitrochlorobenzene found decreased minimal perceptible erythema in Blacks compared with Caucasians. Gean et al. (1989) suggest that racial differences in response to topical methyl nicotinate exist, but that perception of these distinctions may depend upon the method of measurement. Racial differences in physicochemical properties of the skin are also reported. In Black vs Caucasian skin, there is increased resistance to tape stripping (Weigan & Gaylor 1974) [see section 1.1.4], increased electrical resistance (Johnson & Corah 1960) and increased lipid content (Reinertson & Wheatley 1959). Wedig and Maibach (1981) reported that in vivo percutaneous penetration of dipyrithione was less in Blacks than in Whites. Aside from Black vs Caucasian skin, the statistical analysis included such variables as different skin sites (forehead, scalp,
forearm), different vehicles (methanol, cream, shampoo) and different skin treatments (untreated, stripped, hair clipped). Each of the variables themselves - site, vehicle, treatment - will affect percutaneous absorption (Wester & Maibach 1983). If the reported data of Wedig and Maibach (1981) are examined between Blacks and Caucasian only within the same site, vehicle and treatment, there are no differences in percutaneous absorption between Blacks and Caucasians. This agrees with the recent study of Wester et al. (1990), who examined the percutaneous absorption of 3 test chemicals to determine if racial differences exist. Figure I shows that for the chemicals (benzoic acid, caffeine and acetylsalicylic acid) there was no difference in in vivo percutaneous absorption in Caucasian, Asian and Black individuals. Racial differences in biological responses to those chemicals were not addressed by this study. In vivo, percutaneous absorption of diflorasone di30
o Caucasian o Black • Asian
~ !!.... 20
~ 0
CI)
.0
'" Q)
'"0
10
0
Benzoic acid
Caffeine
AspIrin
Fig. 1. Influence of race on in vivo percutaneous absorption in humans.
259
Percutaneous Drug Absorption
acetate was the same in Black and White study participants (Wickrema et at. 1978). However, Berardesca and Maibach (1987) and Wilson and coworkers (1988) reported increased transepidermal water loss (TEWL) in Blacks and Hispanics. A study by Williams et at. (1991) suggest racial differences in nitroglycerin absorption after transdermal application. Black study participants had lower drug concentrations than Caucasians and Asians. A pharmacological response depends on the bioavailability (percutaneous absorption in topical therapy) of a chemical and the inherent activity of the chemical once it is in a biological system (Wester & Maibach 1983). Berardesca et at. (1991) found decreased vasodilation after nicotinate exposure in Blacks compared with Caucasians and an altered hyperaemia reaction after constrictive stimuli. Conventional thought maintains that Black skin resists chemical irritation better than White skin. Experimental support for the statement does exist, but most is based on a visual end-point measurement (i.e. erythema assessment). This end-point has been questioned because erythema is more difficult to detect in Black skin; hence, differential irritancy response between Blacks and Whites may be less pronounced than expected. In fact, objective measurements of skin blood flow by laser Doppler velocimetry showed no differences between Black and Caucasian individuals in methyl nicotinate-induced vasodilation (Guy et at. 1985). Recently, Gean et at. (1989) demonstrated that Asian skin did not differ from either Black or White skin in response to ethyl nicotinate. On the basis of their results they recognised that it was impossible to make a case for differential permeability of different skin types unless it is argued that poor percutaneous absorption in I racial group is exactly compensated by increased microvasculature sensitivity, which seems implausible. Percutaneous absorption studies show different patterns of penetration depending on the molecules under investigation. Thus, in vitro no differences were recorded in water permeation through White or Black skin (Bronaugh et at. 1986) whereas permeation of fluocinolone acetonide was higher through normal-appearing White skin than through
o Theoretical
0.5
Observed N
04
8
0;
.:; 0.3
~
:0
~ 0.2 >
'"o iIi 0.1 Single dose
Single dose x 3
Triple therapy
Triple therapy
Fig. 2. Hydrocortisone skin delivery in vivo in humans with triple therapy. Three daily doses of hydrocortisone in solvent vehicle deliver more drug than predicted from daily dose.
Black skin (excised from legs amputated for tumours or gangrene) [Stoughton 1969].
4. Single Versus Multiple Topical Administration Topical application of hydrocortisone and other corticosteroids frequently uses repeated, rather than single, applications of drug to the skin. It is commonly assumed that multiple applications of hydrocortisone effectively increase its bioavailability and absorption. A long term multiple-dose study in Rhesus monkeys by Wester et at. (l980b) indicated that this was true. However, short term experiments in the Rhesus monkey by Wester et at. (1977) and long term pharmacokinetic assays by Bucks et at. (l985a) showed no increase in hydrocortisone absorption following multiple administration. An investigation was designed to determine if mUltiple-dose therapy would increase drug bioavailability in human skin (Melendres et at. 1990). Percutaneous absorption of hydrocortisone was measured in 6 healthy adult male volunteers. The study compared a single topical dose to treatment with multiple topical doses (I vs 3 applications) in the same day. [14C]Hydrocortisone in acetone was applied to 2.5 cm 2 of ventral forearm skin and pro-
260
Clin. Pharmacokinel. 23 (4) 1992
tected with a nonocclusive polypropylene chamber. The amount of l4C measured in urine collected over 7 days was used to determine hydrocortisone absorption. The treatments, performed 2 to 3 weeks apart, each used adjacent sites on the same individuals. A single dose of hydrocortisone 13.33 ~g/cm2 delivered a mass of 0.056 ~g/cm2. Three serial doses of 13.33 ~g/cm2 (total 40 ~g/cm2) were expected to deliver 0.168 ~g/cm2 with or without soap-and-water washing between doses, but the observed amount of hydrocortisone delivered significantly exceeded the expected mass absorbed. This indicates that multiple administration resulted in a significant increase in bioavailability (fig. 2). It is postulated that increased vehicle application and washing dissolved and mobilised previously administered hydrocortisone and increased bioavailability.
Removal
o Soap and water
100
Water only
80
~
'"'"0
60
"0
'0 c:
.9
13
40
~
u.
20
0
First wash
Second wash
Third wash
Recovery
80
5. Diseased Skin
J~
Jt-
F-'-
F-'-
The assumption has been that percutaneous absorption is enhanced in diseased/damaged skin and that the ability of the skin to protect against intrusion by chemicals is impaired. The picture of skin floodgates opening and chemicals pouring in is certainly not warranted (except perhaps in patients with severe burns). The data suggest that diseased
Co
60
~
'"'"0 "0 '0
Co Co
Co
40
c:
.9
13
~
u.
20
o
80
Total
o
0.5
3
6
24
TIme (h)
Fig. 4. (a) Glyphosate, a water-soluble compound, is easily removed from Rhesus monkey skin in vivo with both water only and soap and water. (b) Time-response in vivo recovery of glyphosate from Rhesus monkey skin.
Time (h)
Fig. 3. With time, isofenphos disappears from human skin in vivo. This is due to a combination of isofenphos volatility and removal by rubbing from clothing.
skin can retain barrier properties and that differences will exist for different drugs and for different disease conditions. In psoriasis, there are definite formulation effects and treatment (e.g. salicylic acid scrub) probably will affect drug delivery into the skin. However, even for a compound such as hydrocortisone, for which skin absorption is low in diseased and damaged skin, enhanced absorption
Percutaneous Drug Absorption
261
will occur in severe conditions. For more potent corticosteroids, the risk of systemic adverse effects such as adrenal suppression remains real. However, data for percutaneous absorption in diseased human skin is limited. More research in this area is warranted (Wester & Maibach 1992).
Removal
80
o Soap and
water Water only
60
6. Skin Decontamination "0
A time-recovery study was done to determine the ability of soap and water wash to remove isofenphos and to determine evaporation of this agent from human skin (Wester et al. 1992). Figure 3 shows that at time zero (approximately 5 to 10 min needed to administer the dose and then wash 4 subjects) a total of 61.4 ± 10.4% of the applied dose could be recovered from the study participants. Recovery decreased slightly to 54.1 ± 12.9% at Ih and 54.9 ± 14.2% by 4h and recovery was down to 30.3 ± 11.1 % at 8h. Recovery at 24h was minimal; 0.53 ± 0.17%. Recovery at 24h (0.75 ± 0.93%) was also minimal. Most of the wash recovery occurred with the first soap application. Many compounds exhibit little skin surface recovery after 24h of skin application time, despite vigorous washing and tape stripping. In the case of isofenphos, chemical volatility of the agent was the answer. In other situations, clothing will rub off the topically applied chemical. Other compounds such as salicylic acid have the ability to be retained in skin. The in vivo percutaneous absorption of salicylic acid in humans is 6.5 ± 5.0% of a dose after 24h skin application. The 24h postapplication skin wash is able to recover 53.4 ± 6.3%, most with the first soap application. There are 2 components to skin washing in the recovery of drug. One is the physical rubbing and removal of the skin surface. The second component is the solvent action of soap and water. Glyphosate is a water-soluble compound. Its removal from skin with water alone or soap and water is the same (fig. 4) [Wester et al. 1991b]. In contrast, alachlor is lipid soluble. Therefore, more alachlor can be removed from skin with soap present, rather than water alone (fig. 5). Alachlor partitions into powdered human stratum corneum from its ve-
40
c .Q
U
u:'" 20
0
First wash
Third wash
wash
Recovery
100
80 ~ L Ql
