`Received July 9, 1996
`Accepted November 5, 1996
`
`© 1997 J. Pharm. Pharmacol.
`
`In-vitro Permeability of the Human Nail and of a Keratin
`Membrane from Bovine Hooves: Penetration of Chloramphenicol
`from Lipophilic Vehicles and a Nail Lacquer
`
`DIRK MERTIN AND BERNHARD C. LIPPOLD
`
`Department of Pharmaceutical Technology, Heinrich-Heine-University, Universitiitsstrasse 1, D-40225
`Diisseldorf, Germany
`
`Abstract
`
`Lipophilic vehicles and especially nail lacquers are more appropriate for topical application on the nail than
`aqueous systems because of their better adhesion. This work has, therefore, studied the penetration through the
`human nail plate of the model compound chloramphenicol from the lipophilic vehicles medium chain
`triglycefides and n-octanol and from a lacquer based on quaternary poly(methyl methacrylates) (Eudragit
`RL). The results were compared with data obtained with a keratin membrane from bovine hooves.
`If the swelling of the nail plate or the hoof membrane is not altered by use of lipophilic vehicles, the
`maximum flux of the drug is independent of its solubility in the vehicle and is the same as that from a saturated
`aqueous solution. These vehicles are not able to enter the hydrophilie keratin membrane because of their non-
`polar character and so cannot change the solubility of the penetrating substance in the barrier. If the
`concentration of the drug in the nail lacquer is sufficiently high, the maximum flux through both barriers
`equals that from aqueous vehicles or even exceeds it because of the formation of a supersaturated system.
`Penetration through the nail plate follows first order kinetics after a lag-time of 400 h. The course of penetration
`through the hoof membrane is initially membrane-controlled and later becomes a matrix-controlled process
`because of the membrane’s greater permeability. Chloramphenicol is dissolved in the lacquer up to a
`concentration of 31%. The relative release rates from these solution matrices are independent of the drug
`concentration but they decrease on changing to a suspension matrix.
`These results show that drug flux is independent of the character of the vehicle and that penetration of the
`drug is initially membrane-controlled and changes to being matrix-controlled as the drug content of the lacquer
`decreases.
`
`The nail plate and the bovine hoof membrane behave like
`hydrophilic gel membranes rather than lipophilic partition
`membranes (Mertin & Lippold 1997). The maximum flux of a
`drug through both barriers is primarily a function of its water-
`solubility. Because aqueous solutions are not important in the
`topical therapy of nail infections, due to their insufficient
`adhesion, lipophilic vehicles or nail lacquers were investigated
`to determine whether the flux from these reached the max-
`imum obtained from aqueous vehicles.
`Studies of the penetration of the antifungal agents ciclopirox
`(Hiinel & Ritter 1990; Nolting & Seebacher 1993) and amor-
`olfine (Polak & Zaug 1990; Franz 1992; Polak 1992) show that
`active drug concentrations are obtained in the whole nail plate
`after a few days. Little is, however, known about the rela-
`tionship between flux and concentration in the lacquer. The
`influence of the release on the kinetics of nail penetration have,
`moreover, not yet been described.
`According to Fick’s law (eqn 1) the penetration rate from an
`aqueous solution at sink conditions is directly proportional to
`the drug concentration in the barrier on the donor side (Car))
`(Mertin & Lippold 1997):
`
`dM/dt = DBACBo/hB
`
`(1)
`
`in which dM/dt is the amount penetrating per unit time, Dn the
`
`Correspondence: B. C. Lippold, Department of Pharmaceutical
`Technology, Heinrich-Heine-University, Universit~tsstrasse 1, D-
`40225 Diisseldorf, Germany.
`
`effective diffusion coefficient in the barrier, A the diffusion
`area, and ha the thickness of the barrier. For a suspended
`substance in a given vehicle, the saturation concentration
`(C~BD) forms on the donor side owing to partition. Then the
`maximum concentration gradient causes the maximum flux:
`
`J~x = dM.~,IdtA, = DBCsBD/hB
`
`(2)
`
`As long as the vehicle does not change the barrier (e.g. by
`deswelling), the maximum flux from a suspension is inde-
`pendent of the vehicle (Lippold 1984). The swelling of
`hydrophilic gel membranes should be unchanged in contact
`with lipophilic vehicles as long as they also stay in contact
`with an aqueous solution. Simulating the swelling of a living
`nail, which is ventrally supplied by the richly vasculated nail
`bed, by using an aqueous solution as acceptor and a lipophilic
`vehicle as donor, the fiux from a saturated solution should
`equal the maximum flux from an aqueous suspension assuming
`an identical extent of swelling.
`To test this hypothesis, the model compound chlor-
`amphenicol was used because it is relatively highly soluble in
`water, which causes sufficiently high fluxes, and it is analyti-
`cally easy to determine, through both the nail plate and the
`hoof membrane. Its molecular size is, moreover, in the range of
`most antimycotics and the results can, therefore, be transferred
`to these drugs. Differences between the solubilities of chlor-
`amphenic01 in pH 7.4 phosphate buffer on the one hand and
`in medium-chain triglycerides and n-octanol on the other
`
`i
`
`e
`
`9
`
`?
`
`)
`
`t
`
`0
`
`it
`
`ARGENTUM EX1022
`
`Page 1
`
`
`
`242
`
`DIRK MERTIN AND BERNHARD C. LIPPOLD
`
`Table 1, Solubility of chloramphenicol in pH 7.4 phosphate buffer,
`n-octanol and medium-chain triglycefides at 32°C.
`
`Vehicle
`
`Solubility (mg L-*)
`
`Studies with different drug concentrations in the polymer
`(from 2.2 to 47-6%) should show whether penetration from the
`lacquer is matrix- or membrane-controlled.
`
`Phosphate buffer, pH 7.4
`n-Octanol
`Medium-chain triglycerides
`
`N=3, mean + s.d.
`
`4520 4-64
`23210 4- 767
`2350 4- 13
`
`(Table 1) seem to be large enough to indicate a possible
`influence of drug solubility in the vehicle on the maximum flux
`through the barrier.
`Results obtained using lipophilic liquids should be trans-
`ferable to nail lacquers, assuming that membrane diffusion,
`and not release from the polymer, is the rate-limiting step. For
`both barriers, however, it must be investigated whether the
`penetration rate is controlled by the permeability of the barrier
`as well as the release of drug from the lacquer. The liberation
`of a substance which is suspended or dissolved in a nail lacquer
`should follow kinetics typical of a matrix system. From Fick’s
`first law Higuchi (1961) developed an equation for the release
`of a suspended drug from a matrix (sink conditions):
`
`Q = A~/(D~frCs(2C0 - Cs)t)
`
`(3)
`
`Where Q is the amount of drug released at time t, A is the
`release area, D~ff is the effective diffusion coefficient in the
`matrix, Cs is the solubility of the drug in the matrix, and Co is
`the initial concentration of the drug in the matrix.
`On the premise that Co>> Cs equation 3 can be reduced to:
`
`Q = A~/(2De~CsC0t)
`
`(4)
`
`Higuchi also deduced an equation describing the course of
`liberation of a drug which is completely dissolved in the matrix
`(up to 30% release, sink conditions):
`
`Q = 2AC0~!(D~irt/n)
`
`Transformation of this equation leads to:
`
`Q/Q0 = (2A/VL)~/(D~fft/n)
`
`(5)
`
`(6)
`
`in which Qo is the initial amount of drug in the matrix and VL
`the matrix volume. Equation 6 shows that the relative release
`rate (Q/Qo) is, in contrast with equation 4, independent of the
`amount of drug incorporated and so enables distinction
`between solution and suspension matrix.
`The release rate can deviate from the ideal ~/t kinetic,
`especially at the beginning of the process, if the drug has to
`penetrate an adherent membrane or aqueous layer after leaving
`the matrix. Roseman & Higuchi (1970) described the course
`of penetration from such systems by combining equations 1
`and 4.
`This work has investigated the penetration of chlor-
`amphenicol from lacquers based on quaternary poly(methyl
`methacrylates) with dibutyl sebacate as a plasticizer through
`the nail plate and the hoof membrane. Eudragit RL was used
`because of its ten-fold higher permeability in comparison with
`Eudragit RS (Lehmann 1989). Because previous results have
`shown the permeability characteristics of both barriers to be
`similar (Mertin & Lippold 1997), most of the investigation was
`performed with the hoof membrane.
`
`Materials and Methods
`
`Chemicals
`A phosphate buffered saline solution, pH 7-4, was used as
`acceptor. Chloramphenicol was obtained from Caesar & Lor-
`entz (Hilden, Germany), medium-chain triglycerides (Miglyol
`812) from HiJls AG (Witten, Germany), n-octanol and
`methanol from J. T. Baker (Deventer, Netherlands), Euclragit
`RL PO from Ri3hm GmbH (Darmstadt, Germany) and dibutyl
`sebacate (Rilanit DBS) from Henkel KGaA (Diisseldorf, Ger-
`many). HPLC-grade methanol (chromasolv methanol) is a
`product of Riedel-de-Hafin (Seelze, Germany).
`
`Penetration studies
`The modified Franz diffusion cells, the preparation of the nails
`and the hoof membranes and the performance of the penetra-
`tion studies have been described in an earlier publication
`(Mertin & Lippold 1997). For experiments with lipophilic
`liquid vehicles, chloramphenicol was used in a suspended form
`with its maximum thermodynamic activity. The formation of a
`saturated solution was guaranteed by stirring at 32°C for 48 h.
`Despite occasional very long penetration times no visual
`degradation of the nails was observed.
`
`Analytical conditions
`The HPLC method differs from that described earlier (Mertin
`& Lippold 1997) in one aspect: the mobile phase acetonitrile-
`water (3: 1) was pumped at flow rates ranging from 1.0 to
`1.25 mL min- I.
`
`Composition and application of the lacquer solution
`The effect of concentration was examined by varying the
`amount of chloramphenicol between 0.5 and 20% of the lac-
`quer solution-equivalent to between 2.2 and 47.6% of the dry
`lacquer. The formulations were: Eudragit RL PO, 20-0%;
`dibutyl sebacate, 2.0%; chloramphenicol, 0-5, 5-0, 10-0 and
`20-0%; and methanol to 100%.
`The swollen membrane was fixed in the empty diffusion cell
`and dried under ambient conditions for 2 h. A 200-pm film
`resulted after application of the lacquer solution (500 I.tL on to
`about 2.5 cm2 hoof membrane and 120pL on to 0.64cm2 nail
`plate), initial drying with warm air for a period of 30 min and
`final drying at room temperature for 24 h. The filling of the
`acceptor compartment started the experiment.
`
`Results and Discussion
`
`Penetration from lipophilic liquids
`Phosphate buffer pH 7-4, n-octanol and medium-chain trigly-
`cerides were used as donors (medium-chain wiglyeerides only
`in experiments with hoof membrane). Table 2 shows max-
`imum fluxes (Jm~ (10001ira)) from the different vehicles,
`standardized to a barrier thickness of 1000 ttm corresponding
`to the average thickness of the big-toe nail.
`
`Page 2
`
`
`
`PERMEABILITY OF NAIL AND KERATIN MEMBRANE
`
`243
`
`Table 2. Maximum flux of ehloramphenicol, standardized to a barrier thickness of 1000 gm
`(Jma~(1000 lam)), from different vehicles through hoof membrane and nail plate at 32°C.
`
`Vehicle
`
`Maximum flux of chloramphenieol
`(mgem-2s-I)
`
`Hoof membrane
`
`Nail plate
`
`Phosphate buffer, pH 7.4
`n-Octanol
`Medium-chain triglycerides
`
`4.074- 1-18 × 10-6
`3.40+0.68 × 10-6
`4.064- 1-00 × 10-6
`
`8.21 4-2-11 × 10-7
`9.134-0.63 × 10-7
`n.d.*
`
`*Not determined. N = 3 or 4, mean 4- s.d.
`
`O
`
`G.
`
`t)
`
`O
`
`!
`
`Fluxes through the hoof membrane are forty-fold those
`through the nail plate, confirming the different permeability of
`the barriers (Mertin & Lippold 1997). It is, however, more
`interesting that the vehicle has no influence on the maximum
`flux. There is no significant difference (P=0-05) between the
`fluxes from the various vehicles through both barriers. Because
`the fluxes from lipophilic vehicles are equal to those from
`aqueous saturated solutions, the assumption that the flux is
`independent of the character of the vehicle is completely
`confirmed. Obviously, a saturated solution and, therefore, the
`maximum concentration gradient forms on the donor side of
`the water-swollen membrane owing to distribution. Neither
`medium-chain triglycerides nor n-octanol have significant
`influence on keratin swelling or the solubility of chlor-
`amphenicol in the membrane. It is of practical significance that
`the therapeutically desired maximum flux is reached as soon as
`the drug is present at its maximum thermodynamic activity, i.e.
`the saturated state. Although low solubility can be used to save
`drugs, very low solubilities lead to emptying effects, i.e. the
`flux cannot be maintained over the whole period of application.
`This result is of great importance in respect of drug penetration
`from nail lacquers.
`
`Penetration from nail lacquers
`Kinetics of penetration. Figs 1 and 2 illustrate the concentra-
`tion-dependence of the diffusion of chloramphenicol from
`Eudragit RL lacquers through the hoof membrane. Plotting
`the amount penetrated against t gives linear relationships after
`a lag-time of a few hours; this is typical of matrix control (Fig.
`1). As expected for a solution matrix, the rate of penetration of
`chloramphenicol increases with the concentration of the drug
`in the matrix between 2.2 and 18.5% and so the relative release
`rates (i.e. the amount penetrated relative to the total amount in
`the lacquer) remain constant. Increasing the concentration in
`the lacquer to 47.6% has no effect on the penetration rate,
`however, and so the relative rates decrease. This proves that,
`except for the lacquer containing 47.6% chloramphenicol, all
`systems are solution matrices as the relative release rates are
`independent of the amount of drug incorporated, in accordance
`with equation 6. Because the relative release rate decreases by
`half for the 47.6% lacquer, this, therefore, can be characterized
`as a suspension matrix. Fig. 1 does not enable distinction
`between matrix- and membrane-controlled processes. For
`membrane-controlled release from a solution matrix, first-
`order kinetics are expected. The plot of the amount of drug
`remaining in the lacquer (logarithmic scale) against time (Fig.
`
`50’
`
`A
`
`~ 40-
`
`30.
`
`E= 20.
`o
`e
`’<10.
`
`0
`
`o i
`
`Time 1/2 (h1/2)
`
`FIG. 1. Percentage penetration of ch]oramphenicol from Eudragit RL
`lacquers containing different concentrations, CL, of drug through the
`hoof membrane at 32°C (n =4, mean 4-s.d.). Chloramphenicol con-
`centration: ¯ 2.2%, [] 18.5%, ¯ 31.3%, ¯ 47.6%.
`
`1 oo
`
`o 90 8o
`
`70
`
`60
`
`50
`
`lb
`
`2’0
`
`Time (h)
`
`’ 5O
`
`FIG. 2. Penetration of chloramphenicol from Eudragit RL lacquers
`containing different concentrations, CL, of drug, through the hoof
`membrane at 32°C, plotted as the amount of drug remaining in the
`lacquer (Qo - Q; n = 4, mean 4- s.d.). Chloramphenicol concentration:
`¯ 2.2%, [] 18.5%, ¯ 31.3%, ¯ 47.6%.
`
`2) shows, because all curves are flattening, that membrane
`control does not occur over the whole period.
`The ideal ~t kinetics follow after the expiry of the lag-time
`because of the initially predominant membrane control. Delayed
`drug release from silicone matrices through aqueous adherent
`layers results in similar penetration profiles (Haleblian et al
`1971; Roseman 1972).
`
`Page 3
`
`
`
`244
`
`DIRK MERTIN AND BERNHARD C. LIPPOLD
`
`Release exponent, n, for the penetration kinetics of ch!oramphenicol from Eudragit RL lacquers
`Table 3.
`through the hoof membrane at different drug concentrations.
`
`CL (%) Calculated according to equation 7* Calculated according to equation 8I"
`
`Exponent (n)
`
`r~
`
`Exponent (n)
`
`tt,s (11)
`
`2.2
`18-5
`31.3
`47.6
`
`0.67 -4- 0.04
`0.734-0-05
`0.70 ± 0.02
`0.64 -I- 0-03
`
`0.9990
`0.9917
`0.9995
`0.9995
`
`0.56 4- 0-03
`0.504-0.05
`0.62 -t: 0.03
`0.65 -4- 0.04
`
`1.55 4- 0.27
`3.28+0.19
`1-45 4- 0.28
`0.09 4- 0.18
`
`0.9999
`0.9985
`0.9999
`0.9999
`
`*Q/Qo = kt~- tQ/Qo = k(t - tlag)n. ZCorrelation coefficient of the regression line. N = 4, mean 4- s.d.
`
`The course of the drug release from a dosage form can be
`expressed by a semi-empirical function (Peppas 1985):
`
`Q/Q0 = kt~
`
`(7)
`
`Where Q/Qo has the same meaning as in equations 3 and 6, k is
`the release-rate coefficient, and n is an exponent which
`describes the kinetics. If release is delayed, neglecting the lag-
`time can lead to incorrect conclusions about the penetration
`kinetics. In this circumstance equation 8 offers a better
`approach:
`
`Q/Q0 = k(t - tlag)n
`
`(8)
`
`The release exponent (n) and the lag-time can be determined
`by a computer-aided, iterative method. Here the lag-time is
`established by a progressive shift of the experimental curve to
`the left parallel to the abscissa, beginning with data from 5 h
`onwards, inserting in the logarithmic form of equation 7 and
`subsequent linear regression to obtain the best curve fit
`(Lindner 1994). Table 3 shows the liberation parameters
`determined directly with equation 7 and iteratively with
`equation 8.
`Neglecting the lag-phase normally leads to overestimation
`of the release exponent n and a worse fit of the calculated curve
`to the experimental data; this is reflected in a lower correlation
`coefficient. With the exception of the lacquer containing
`47.6% chloramphenicol, the lag-time ranges from 1-4 to 3.3 h
`and agrees with the values determined graphically from Fig. 1.
`The iterative exponents (0-50 to 0.62, Table 3) are in the range
`expected for pure matrix release (n =0.5) and confirm the
`visual assessment of the profiles. The exponent of the 47.6%
`lacquer is significantly different from unity which is expected
`for completely membrane-controlled release from a saturated
`vehicle. These results, on the other hand, correspond with the
`assumption of initial membrane control changing to matrix
`
`control as the drug content of the lacquer decreases.
`
`Fluxes from nail lacquers compared with liquid vehicles
`The profile between the first and the sixth hours after the start
`of the experiment was evaluated to determine the fluxes
`through the hoof membrane. Here the penetration rate is
`highest and there is an approximately linear relationship
`between the amount penetrated and time. This behaviour, not
`typical of matrix control, corresponds to the initial membrane-
`control. The expected first order kinetics (only dissolved drug
`in the lacquer) results in a more or less linear course for an
`
`For saturated solutions (real steady-state), zero order kinetics
`prevail, which causes a linear increase of the concentration in
`the acceptor. Not before a sufficiently large emptying zone of
`the drug has developed, causing also a decrease in con-
`centration in the membrane, does the process become matrix-
`controlled.
`As the maximum flux is, in addition to antifungal power, the
`most important parameter for predicting the therapeutic effi-
`cacy of anfimycotics, the penetration of chloramphenicol
`through the nail in man was investigated with a single for-
`mulation (31-3%). Fig. 3 shows the low penetration rate and
`large lag-time in comparison with the thinner hoof membrane.
`The lag-time (about 400h) is significantly longer than for
`penetration from an aqueous suspension (about 200h).
`Because both fluxes do not differ from each other in the
`steady-state (Table 4), the distinction cannot be explained by
`the different diffusion coefficients. Possibly the initially dry
`nail plate is slowly hydrated under the occlusive lacquer after
`contact with the acceptor medium, whereupon the partition
`equilibrium between polymer and nail plate and, therefore, the
`formation of the maximum concentration gradient, is delayed.
`The flux was calculated from the steady-state values between
`t=670 and 940h. Because of the lower penetration rate it
`should stay constant over a longer period than for the hoof
`membrane, for which the emptying area in the matrix widens
`after a short time and then the course of diffusion corresponds
`to classical ~/t-kinetics. Thus, the hoof membrane has to be
`
`20.
`
`15
`
`-,~
`
`g
`
`0.20
`
`0-15
`
`0.10
`
`E
`
`0
`
`0-05
`
`0
`
`200 400 600
`
`Time (h)
`
`0-00
`800 1000
`
`FIG. 3. Penetration of chloramphenicol (CL = 31.3%) from Eudragit
`RL lacquer through the hoof membrane (thickness, dB, = 104 pm) and
`the nail plate (ds=953pm) at 32°C (n=4, mean_-h s.d.). QH and QN
`&re, respectively, the amounts of drug penetrated through hoof mem-
`~-’~ne (O) and nail nlate (@).
`
`Page 4
`
`
`
`. Table 4. Flux of chloramphenicol through the hoof membrane and
`the nail plate at 32°C from Eudragit RL lacquers containing different
`concentrations of drug, CL, and from a saturated solution in pH 7-4
`phosphate buffer.
`
`Concentration in the
`lacquer (%)
`
`Maximum flux of ehloramphenicol
`(J(1000 btm); mgcm-2s- I)
`
`Hoof membrane
`
`Nail plate
`
`2.2
`18.5
`31.3
`47.6
`Saturated solution in
`phosphate buffer 5.32+1.62 × 10-6
`
`3-104-0.39 x 10 7
`4.344-1.27 × 10-6
`7.264-1.75 × 10-6
`8-114-2.16 x 10-6
`
`n.d.*
`8.024-1.81 × I0-8
`n.d.
`n.d.
`
`8.214-2-ll x 10-s
`
`*Not determined. Results are standardized to dB=10001.tm.
`n=3-7, mean + s.d.
`
`used with caution as a model for studying controlled-release
`systems for application on nails.
`The fluxes of chlnrarnphenicol, standardized to a barrier
`thickness of 1000 p.m are in the same range as the maximum
`fluxes from aqueous suspensions (Table 4). Firstly, it is sur-
`prising that the flux through the hoof membrane from the more
`highly concentrated lacquers (for the 47.6% lacquer it is even
`statistically significant, P = 0-05) is greater than that from an
`aqueous suspension. This can only be explained by the
`assumption of the formation of a supersaturated solution in the
`barrier. For the 31.3% Eudragit RL lacquer the development of
`a thermodynamically unstable supersaturated solution could be
`proved by polarization microscopy. This state seems also to be
`formed on the donor side of the hoof membrane, because of the
`distribution equilibrium, and remains until crystallization starts
`or the concentration falls below the solubility as a result of the
`emptying of the matrix. As the flux from the 18.5% lacquer is
`not significantly lower than the maximum flux from water, it
`has to be assumed that the solubility of chlorampbenicol in the
`poly(methyl rnethaerylate) lacquer is in the same range. These
`findings are confirmed by the results of the nail plate--the flux
`from the 18.5% lacquer equals the maximum flux from water.
`The lacquer presented, consisting of a highly permeable
`quaternary poly(methyl methacrylate) (Eudragit RL) and
`dibutyl sebacate as a plasticizer, is, therefore, a suitable dosage
`form for achieving high drug fluxes through the nail plate and
`the hoof membrane. By addition of a sufficiently high con-
`centration of drug it is possible to achieve penetration rates
`which correspond to those from saturated liquid vehicles
`
`.ATh .... ~E
`
`243
`
`(water or non-aqueous solvents) or even exceed those owing to
`the temporary formation of a supersaturated system. Because
`of the low permeability of the nail plate the release rate of the
`lacquer is not important in this instance. The rapid develop-
`ment of the partition equilibrium between lacquer and barrier
`is, however, significant and so high-swelling polymers
`(Eudragit RL) have to be preferred. It must, however, be
`considered that the dried lacquer remains water-insoluble when
`it becomes more hydrophilic-otherwise it would be removed
`by washing and, therefore, the application intervals have to he
`shortened. On the other hand, the occlusivity of the nail lac-
`quer is of some importance; this is probably increased for
`poorly swelling polymers with low water-vapour permeability.
`
`References
`
`Franz, T. J. (1992) Absorption of amorolfine through human nail.
`Dermatology 184 (Suppl. 1): 18-20
`Haleblian, J., Runkel, R., Miiller, N., Christopherson, J., Ng, K. (1971)
`Steroid release from silicone eI2-stomer containing excess drug in
`suspension. J. Pharm. Sci. 60:541-545
`H~nel, H., Ritter, W. (1990) Formulation. In: Ryley, J. F. (ed.)
`Chemotherapy of Fungal Diseases (Handbook of Experimental
`Pharmacology, Vol. 96). Springer, Berlin, pp 251-278
`Higuchi, T. (1961) Rate of release of medicaments from ointment
`bases containing drugs in suspension. J. Pharm. Sci. 50:874--875
`Lehmann, K. (1997) Chemistry and application properties of poly-
`methacrylate coating systems. In: McGinity, J. W. (ed.) Aqueous
`Polymeric Coatings for Pharmaceutical Dosage Forms. Marcel
`Dekker, New York, pp 101-176
`Lindner, W. D., M0ckel, J. E., Lippold, B. C. (I996) Controlled release
`of drugs from hydrocolloid embeddings. Pharmazie 51:263-272
`Lippold, B. C. (1984) Biopharmazie. Wissenschafdiche Verlagsge-
`sellschaft, Stuttgart, pp 106-108
`Mertin, D., Lippold, B. C. (1997) In-vitro permeability of the human
`nail and of a keratin membrane from bovine hooves: influence of the
`partition coefficient octanol/water and the water solubility of drugs
`on their permeability and maximum flux. J. Pharm. Pharmacol. 49:
`30-34
`Nolting, S., Seebacher, C. (1993) Ciclopiroxolamin - Wegweiser
`topischer Mykose-Therapie, Universit~tsverlag Jena, .lena
`Peppas, N. A. (1985) Analysis of Fickian and non-Fickian drug release
`from polymers. Pharm. Acta Helv. 60:110-111
`Polak, A. (1992) Kinetics of amorolfine in human nails. Mycoses 36:
`101-103
`Polak, A., Zaug, M. (1990) Amorolfine. In: Ryley, L F. (ed.) Che-
`motherapy of Fungal Disease (Handbook of Experimental Pharma-
`cology, Vol. 96). Springer, Berlin, k’P 505-52l
`Roseman, T. J. (1972) Release of steroids from a silicone polymer. J.
`Pharm. Sci. 61:46-50
`Roseman, T. J., Higuchi, W. I. (1970) Release of medroxyprogesterone
`acetate from a silicone polymer. J. Pharm. Sei. 59:353-357
`
`¢).
`
`O
`
`o
`
`lI.
`
`o
`
`i ~,ii
`
`:i
`
`Page 5
`
`