ELSEVIER

Pharmaceutics

Acta Helvetiae 71 (1996) 253-258

An investigation into the transdermal delivery of nifedipine

D.M. McDaid , P.B. Deasy
Received 6 September *

1995: accepted 27 September 1995

Abstract

A systematic attempt to develop a transdermal delivery system for nifedipine is presented. Measured physicochemical properties influencing percutaneous absorption such as solubility and partition coefficient confirmed the drugs potention for such a formulation approach. However, studies involving permeation through hairless mouse skin from a range of hydrophilic and hydrophobic donor

vehicles indicated inadequate penetration. Attempts to increase the drug flux through the animal skin or a range of artificial membranes
alone or in parallel by use of the penetration enhancers sodium lauryl sulphate I% and propylene glycol 20% in a sodium carboxymethylcellulose 3% gel base failed to raise the drug flux to an acceptable level. Likewise increase in the drug thermodynamic gradient across the skin by use of mixed solvents or supersaturated drug solutions was ineffective if an aqueous receiving phase was used. Collectively the results suggest that the development of a transdermal delivery system for the chemically unmodified drug in humans is unlikely to be successful. Keywords: Transdermal drug delivery; Nifedipine

1. Introduction Existing dosage forms of nifedipine for sublingual or oral dosing are composed of tablets and capsules. In conventional form such products require dosing at least three times daily and in sustained-release form twice daily, because of high first-pass metabolism (Hermann and Morselli, 1985) and short elimination half-life (about 2 h, increasing with age) of the drug (Robertson et al., 1988). Adverse side-effects such as headache, flushing, dizziness and ankle oedema associated with plasma spikes are common (Kleinbloesem et al., 1987) and because the drug is taken for long periods in the treatment of angina pectoris, hypertension and other cardiovascular disorders, compliance problems can arise. The drug has a low daily dose

1 Corresponding author. Fax: + 353-l-2696457. Present address: Forest Laboratories Ltd., Clonshaugh
tate. Dublin 17. Ireland.

Industrial

Es-

(20-60 mg) and consequently is a candidate for design into an effective transdermal system with 3 to 5 day dosing interval, which would eliminate first-pass metabolism, ensure more uniform plasma levels, reduce side-effects and aid compliance. Other drug properties of significance to the development of a successful transdermal system include very poor aqueous solubility (11 mg I-, Kohri et al., 1987), good solubility in ethanol (Martindale, 1993) and rapid photosensitivity in either solvent under ultraviolet or daylight (Majeed et al., 1987). Hairless mouse skin was chosen as it is readily available, widely used in the literature for such studies, but is known to be more permeable than human skin (Barry, 1983). Optimization of formulation factors such as choice of base, use of penetration enhancers and selection of rate-controlling membrane should be undertaken using the animal model, to help establish if the delivery of adequate quantities of nifedipine transdermally is feasible. This study reports the results of such studies.

003 I -6865/96/$ IS.00Copyright 0 I996 Elsevier Science B.V. All rights reserved PI/ SO03 1-6X65(96)00022-2

254

D.M. McDaid, P.B. Deasy/Phamaceutica

Acta Heluetiae

71 (19961253-258

2. Experimental 2. I. Chemicals
Acetonitrile (HPLC grade, Rathbum), butanone, sodium carboxymethylcellulose, sodium lauryl sulphate (BDH Chemicals), carboxymethylethylcellulose (AQ grade, Shin-Et&, Celgard 2400 (Celanese), Cremophor RH40, macrogol 300 and 4000, yellow soft paraffin (BASF), disodium hydrogen phosphate, orthophosphoric acid, n-octanol, sodium dihydrogen phosphate, tetrahydrofuran, HPLC grade (Riedel de Haen), ethanol (HPLC grade, Lab-Scan), ethylene vinylacetate (EVA) membrane (type 987 192, 19% VA content, 2 mm thickness, type 987 292, 29% VA content, 2 mm, 3M), 1-heptane sodium sulphonate (Sigma), nifedipine BP micronized (Heterochem), polysorbate 80 (Croda), propylene glycol (May & Baker) and glass-distilled water were used. All reagents were GPR unless stated otherwise. 2.2. Analysis of samples All experiments involving nifedipine were performed under sodium light to minimize photodegradation. Analysis of drug in solution was performed where possible using a double-beam Shimadzu UV-160 spectrophotometer set at 235 nm wavelength of maximum absorbance. When interference or low drug concentration required, a HPLC method was employed using an LC-5A unit connected to a SPD-2A UV detector with output to a C-R3A chromopack integrator (Shimadzu). An injection volume of 25 ~1 was used. The column employed was an Ultratech ODS 5 p.m (250 X 4.6 mm) guarded by a precolumn (10 mm) with similar packing, and the filtered and degassed mobile phase, pumped at a rate of 1.5 ml min-, was composed of 6 mM 1-heptane sodium sulphonate, 42% v/v acetonitrile and 1% v/v tetrahydrofuran in a 0.1 M sodium dihydrogen orthophosphate buffer adjusted to pH 3.0 with 85% v/v orthophosphoric acid. The HPLC assay was validated for precision, linearity, recovery and specificity. 2.3. Determination of drug solubility

2.4. Determination

of partition

coeflcient

The partition coefficient of nifedipine was carried out in n-octanol/distilled water and n-octanol/phosphate buffer pH 7. The two phases were shaken together initially to ensure mutual saturation. An accurately weighed quantity of nifedipine was dissolved in 100 ml of the n-octanol phase and shaken at 37C for 24 h against 50 ml aqueous phase in a sealed container. The separated n-octanol phase was assayed by UV spectroscopy to determine its residual concentration and hence the amount partitioned into the aqueous phase. The partition coefficient was expressed as the concentration of drug in the n-octanol phase (% w/v) divided by the concentration in the aqueous phase.

2.5. In oitro permeation

studies

The solubility of nifedipine was determined in distilled water, n-octanol and mixed solvent. Analysis was carried out by placing an excess of the drug with 1 ml of the appropriate solvent in each of six 1.5 ml sealed polypropylene micro-vials, shaking in a water-bath at 37C for 24-36 h until equilibrium was achieved and analysing the filtrate passing through a 0.22 pm Millipore filter by UV spectroscopy or HPLC with reference to a calibration curve.

Before use in permeation studies, each artificial membrane (Celgard or EVA types) was wetted to remove entrapped air and reduce surface tension by forcing phosphate buffer of appropriate pH containing 0.05% polysorbate 80 through it, followed by rinsing and storage in the buffer before use. When the donor phase included ethanol, the artificial membrane was wetted by sonication in the medium for 1 h before rinsing and storage in fresh medium. Donor systems containing 1% w/v drug dispersions equilibrated overnight at 37C were prepared from sodium carboxymethylcellulose (NaCMC) 3% w/v in aqueous vehicles of variable pH, yellow soft paraffin, Cremophor RH40 or macrogol ointment BP. Sodium lauryl sulphate (SLS) 1% or propylene glycol (PC) 20% were substituted for water, alone or in combination, in the NaCMC gel base as penetration enhancers. Male hairless mice (6-8 weeks old), Sha-Sha strain, were sacrificed by cervical dislocation and immediately the entire skin from the abdominal area was excised. This section of skin was then cleaned of any subcutaneous fat and blood vessels, while maintaining the integrity of the viable epidermis and stratum comeum. To obtain stripped skin, the excised skin was positioned with the dermis facing outermost and the dermis was then repeatedly peeled away using cellophane tape (3M). This operation was repeated 15 times using a fresh piece of tape each time. The treated skin was used immediately after preparation so as to preserve its integrity for as long as possible. Permeation studies were carried out using diffusion cells, details of whose construction and use have been described previously (ONeill and Deasy, 1988). At various time intervals, 1 ml samples were withdrawn from the receiver compartment and analysed by UV spectroscopy or HPLC for drug content. Experiments were replicated at least four times.

D.M. McDaid, P.B. Deasy / Pharmaceutics Acta Hebetiae 71 (1996) 253-258

255

3. Results and discussion 3.1. Physicochemical properties

N2.m -tVU YSP CRH40

Nifedipine does not exist in an ionized state at physiological pH. The aqueous solubility of nifedipine was determined to be 10.1 mg 1-l and the solubility in n-octanol was 14.75 mg ml- . The observed partition coefficients of nifedipine were determined using n-octanol/water and n-octanol/phosphate buffer pH 7 at 37C and found to give log K values 2.36 and 2.49, respectively. These values are in good agreement with those reported in the literature (Diez et al., 1991). There is an optimum log K value for most compounds for partitioning through lipophilic membranes. Compounds with values below these optimum log K values do not partition readily into the stratum corneum, while compounds with higher values of log K are so lipophilic that they remain dissolved in the stratum corneum. The log K values of nifedipine auger well for effective transdermal delivery because they are of the same order as those of glyceryl trinitrate (2.05), chlordiazepoxide (2.5) and timolol base (1.91), for which systemic transdermal delivery has been demonstrated. 3.2. Permeation across hairless mouse skin

10

20

Time (hr) Fig. 2. Effect of various donor systems on the permeation of nifedipine through hairless mouse skin, where Qr is the cumulative amount of drug transfened (NaCMC is sodium carboxymethylcellulose gel, CRH40 is Cremophor RH40, MO is macrogol ointment and YSP is yellow soft paraffin).

lag time (Diez et al., 1991). Since an exposure time of 72- 120 h was envisaged for the TDS being developed, the lag time obtained from these preliminary experiments do not preclude the possibility of developing a successful TDS. 3.3. Formulation effects

The permeation profile of nifedipine across both intact and viable (stripped) hairless mouse skin is shown in Fig. 1. The calculated permeation rate was 1.Ol pg cm-* h- for intact skin and 4.06 pg cmp2 h- for viable skin with corresponding lag times of 1.94 and 0.61 h. As the stratum corneum is lipophilic in nature and represents a major barrier to permeation of most drugs, its removal by stripping has resulted in significant increase in nifedipine permeation rate and reduction in lag time. The lag time is often considered as the first limiting factor for a transderma1 delivery system (TDS) and 10% of the exposure time of the TDS is considered a reasonable upper value for the

The choice of reservoir system, possibly its pH and rate-controlling membrane will influence the transfer of drug from a TDS into skin layers. Fig. 2 shows the effect of yellow soft paraffin, Cremophor RH40, macrogol ointment BP and sodium carboxymethylcellulose (NaCMC) 3% gel on the transdermal penetration of nifedipine through hairless mouse skin, the plots confirming achievement of steady-state permeation rates after elapse of an initial lag period. The steady-state fluxes were 0.73 k 0.06, 0.88 + 0.11, 0.86 * 0.07 and 1.01 + 0.14 p_g cm-* hb for yellow soft paraffin, Cremophor, macrogol and NaCMC bases respectively (mean + 1 SD). Because maximum flux was observed with the NaCMC base, the effect of altering the pH of this vehicle was studied as shown in Fig. 3. As nifedipine exists only in the unionized form in solution over the pH examined. its

pH 5.0 pH 6.0 pH 7.0 pH 8.0

10

10

15

Time fhr) Fig. I. Permeation of nifedipine skin (mean + SD, n = 4). across intact and viable hairless mouse

Time (hrf Fig. 3. Effect of NaCMC gel pH on the steady-state nifedipine through hairless mouse skin. permeation rate of

256

D.M. McDaid, P.B. Deasy/ Pharmaceutics Acta Heluetiae 71 (1996) 253-258 Table 1 Flux and lag times of nifedipine obtained through various using mixed solvent in both donor and receiver systems
NOR.3 SLS FG SLS 8 PG

membranes

Membrane Hairless mouse skin Celgard 2400 EVA 987 192

Flux (pg cm- 1344*289 6287k372 62.1 k8.6

h-l)

Lag time (h)

I .34 + 0.29 0.21 + 0.03 0.3 I * 0.04

10

Time

(hr)

Fig. 4. Effect of various penetration enhancers on the steady-state permeation rate of nifedipine through hairless mouse skin (SLS is sodium lauryl sulphate and PC is propylene glycol).

permeation rate does not change significantly, p = 0.05 (0.89 + 0.14, 0.97 f 0.13, 1.01 + 0.14 and 0.93 + 0.11 kg cm- h-t at pH 5, 6, 7 and 8, respectively). The previous results indicated that whereas changes in formulation of the donor vehicle could affect the flux of drug through hairless mouse skin, clinically useful transdermal delivery was unlikely to be achieved without further enhancement of the effect. When SLS 1% and PG 20% as penetration enhancers were included into the NaCMC gel base (pH 7) either alone or in combination, the effect on permeation was as shown in Fig. 4. When SLS was included alone, the flux increased slightly (p > 0.05) over the control (1.14 & 0.15 compared to 1.01 + 0.14 pg cmp2 hh ). When PG was employed alone in the donor phase, the resultant flux was increased to 1.3 1 * 0.19 pg cm- h- (p > 0.5). Significant (p < 0.05) increase in flux to 1.64 + 0.23 pg cm- h- compared to the control was observed only when the penetration enhancers were used in combination. The mode of action of SLS is related to its ability to disrupt keratin resulting in increased hydration of the stratum corneum, apart from possible emulgent, solubilization and competitive binding effects (Barry, 1987). I n addition to increasing skin hydration

because of its hygroscopic property, PG probably acts mainly as a cosolvent, increasing the drug gradient in solution for percutaneous absorption (Barry, 1983). It is probable that the combination of the two enhancers produced a synergistic effect as their major modes of promoting enhancement are different. At this stage of formulation, it was apparent that the low steady-state flux of nifedipine through hairless mouse skin, would not produce therapeutic plasma levels with a TDS of convenient area if tested in humans. In order to increase the flux it was proposed to increase the concentration of the drug in the donor phase, thus increasing the thermodynamic activity. Also it was proposed to examine if it was possible to formulate an effective transdermal delivery system for nifedipine when a rate-controlling membrane was employed. Initial experiments using nifedipine in an aqueous suspension at pH 7 yielded fluxes of 1.01 k 0.14, 2.17 +0.22 and 0.36 kO.03 pg cm- hh' , with corresponding lag times of 1.94, 0.24 and 0.28 h for hairless mouse skin, Celgard 2400 and EVA 987 192, respectively. In order to achieve a theoretical daily transdermal dose of 13 mg (Diez et al., 1991), preliminary experimentation on nifedipine solubility at 37C yielded the following vehicle; ethanol:butanone:water in the ratio of l&75:37.5:43.75% v/v. The solubility of nifedipine increased over 500 fold from 10.1 f 0.82 mg 1-l in the aqueous system to 5591 + 603 mg 1-l in the mixed solvent system. The permeation profile of nifedipine through

40 -l

CelgiHM _ EVA/HM

10

15

10

, 15

Time (hr) Fig. 5. Effect of various membranes on the permeation rate of nifedipine using mixed solvent in both donor and receiver systems (EVA is EVA 987192, H. Mouse is hairless mouse skin and Celgard is Celgard 2400).

Time (hr) Fig, 6. Permeation profile of nifedipine through Celgard 2400/hairless mouse skin (Celg/HM) or EVA 987192/hairless mouse skin (EVA/HM) used in parallel with the mixed solvent system.

D.M. McDnid, P.B. Dea.~~/Pharmaceutica

Acm Helr,etiae 71 (1996) 253-258

357

Table 2 Fluxes, lag times. diffusion coefficients (D) and permeability coefficients (K,) obtained for nifedipine using EVA 987192. EVA 987292 and Celgard 2400 membranes at pH 7.4

!I;@.,,
0
5 10 15

Membrane 987192 Flux(bgcrn- hK) Lag timi (h) DX IO4 (cm \-I K,, X IO (cm s- ) 2400, EVA 0.36 0.28 3.x I .07 987292 0.66 0.32 3.73 2400 2.17 0.23 1.21

I .x

6.4I

Time (hr) Fig. 7. Permeation profiles of nifedipine 987192 and EVA 987292 at pH 7.4. through Celgard

hairless mouse skin, Celgard 2400 and EVA 987192 from a saturated donor system using the mixed solvent in both donor and receiver phases is shown in Fig. 5, with derived values shown in Table 1. The flux of nifedipine was greatly increased, using mixed solvent compared to aqueous suspension. by 1330, 2897 and 172 fold using hairless mouse skin, Celgard 2400 and EVA 987192, respectively. These resultant increases in the flux of nifedipine do not correlate directly with the increased drug solubility and could be due to interaction between the solvent system and the various membranes. The increased permeability of hairless mouse skin caused by ethanol is well documented (Knutson et al., 1990; Kurihara-Bergstrom et al., 1990). Fig. 6 illustrates the fluxes obtained when either Celgard 2400 and hairless mouse skin or EVA 987192 and hairless mouse skin are used in parallel. The Celgard combination resulted in a steady-state flux of 993 fr 62.8 compared to 51.7 k 6.81 kg cm- hh for the EVA combination using a mixed solvent receiver phase. Assuming that a suitably chosen adhesive layer in parallel would not significantly affect the overall permeability of the Celgard combination, an effective diffusional area of 0.54 cm would be required to produce therapeutic levels, compared to 10.5 cm2 for the EVA combination. Both areas would be commercially acceptable, but because in the EVA combination the artificial membrane exerts major control over the drug flux, such a device would be expected to achieve better inter- and intra-subject control when used on variable skin sites and types. However in the in-vivo situation, the drug would be expected to travel from the mixed solvent containing reservoir through the rate-controlling membrane and skin located in parallel, but into an aqueous rather than a hydroalcoholic receiving phase as discussed above. The effect of microporous Celgard 2400 and the EVA copolymer membranes 987192 and 987292 on the permeation of nifedipine in aqueous pH 7.4 donor and receiver phases is illustrated in Fig. 7 and derived physicochemical data is shown in

Table 2. The fluxes obtained using the EVA membranes show good agreement with those reported by Kondo and Sugimoto (1987) for EVA membranes containing 10% VA. The lag times for the various membranes are similar, ranging from 0.24 to 0.38 h, and hence do not exceed 10% of the envisaged exposure time. From the steady-state fluxes, it is possible to calculate the cumulative amount of nifedipine transferred in a 24 h period, assuming an exposure area of I6 cm2 (4 X 4 cm). indicating 0.14, 0.25 and 0.83 mg of nifedipine delivery through EVA 987192, 98292 and Celgard 2400, respectively. Hence all three membranes examined would not achieve therapeutic levels of nifedipine (13 mg daily) if fabricated into a TDS of reasonable area ( 1O-20 cm2 1. When the penetration enhancers SLS I% and PG 20% were used in combination in the aqueous donor vehicle at pH 7, the flux was increased only 1.5 fold for the most permeable of the membranes, Celgard 2400, corresponding to an inadequate daily dose of 1.25 mg from a 16 cm device. When ethanol 20% was incorporated into the aqueous donor phase instead of the penetration enhancers, to increase the thermodynamic activity of the drug across the membrane by raising its solubility, the tlux was increased inadequately by 1.6 fold into pH 7.4 receiver phase. In a final attempt to increase the thermodynamic gradient so as to achieve useful fluxes, supersaturated solutions of nifedipine were used in the donor phase. These saturated

0 0.0
1.0 20 3.0

4.0

Time (hr) Fig. 8. Permeation 100 mg nifedipine 2400 into pH 7.4. profile of nifedipine from a donor system containing and 300 mg CMEC I- in pH 7.4 through Celgard

258

D.M. McDaid, P.B. Deasy/ Pharmaceutics Acta Heluetiae 71 (1996) 253-258 Barry, B.W. (1987) Mode of action of penetration enhancers in human skin. J. Control. Rel. 6, 237-248. Diez, I., Colom, H., Moreno, J., Obach, R., Peraire, C. and Domenech, J. (1991) A comparative in vitro study of transdermal absorption of a series of calcium channel antagonists. J. Pharm. Sci. 80, 931-934. Hasegawa, A., Taguchi, M., Suzuki, R., Miyata, T., Nakagawa, H. and Sugimoto, I. (1988) Supersaturation mechanism of drugs from solid dispersions with enteric coating agents. Chem. Pharm. Bull. 36, 4941-4950. Hermann, P. and Morselli, P.L. (1985) Pharmacokinetics of diltiazem and other calcium channel entry blockers. Acta Pharmacol. Toxicol. 57, Suppl. 2, 10-20. Kasting, G.B., Smith, R.L. and Cooper, E.R. (1987) Effect of lipid solubility and molecular size on percutaneous absorption. In Shroot, B. and Schaefer, H. (Eds.), Skin Pharmacokinetics, Skin Pharmacology, Vol. 1, Karger, Basel, pp. 138-153. Kleinbloesem, C.H., van Brummelen, P., Danhof, M., Faber, H., Urquhart, J. and Breimer, D.D. (1987) Rate of increase in the plasma concentration of nifedipine as a major determinant of its haemodynamic effects in humans. Clin. Pharmacol. Ther. 41, 26-30. Knutson, K., Krill, S.L. and Zhang, J. (1990) Solvent-mediated alterations of the stratum corneum. J. Control. Rel. 11, 93-103. Kohri, N., Miyazaki, K., Arita, T., Shimono, H., Nomura, A. and Yasuda, H. (1987) Rel. characteristics of nifedipine sustained-Rel. granules in vitro and in healthy subjects. Chem. Pharm. Bull. 35, 2504-2509. Kondo, S. and Sugimoto, I. (1987) Enhancement of transdermal delivery by superfluous thermodynamic potential. 1. Thermodynamic analysis of nifedipine transport across the lipoidal barrier. J. Pharmacobio-Dyn. 10, 587-594. Kurihara-Bergstrom, T., Knutson, K., De Noble, L.J. and Goates, C.Y. (1990) Percutaneous absorption enhancement of an ionic molecule by ethanol-water system in human skin. Pharm. Res. 7, 762-766. Majeed, I.A., Murray, W.J., Newton, D.W., Otham, S. and Al-Turk, W.A. (1987) Spectrophotometric study of the decomposition kinetics of nifedipine. J. Pharm. Pharmacol. 39, 1044-1046. Martindale, The Extra Pharmacopoeia (1993) 30th Ed., Pharmaceutical Press, London, pp. 374-380. ONeill, C.T. and Deasy, P.B. (1988) Development and evaluation using hairless mouse skin of a transdermal timolol product. Int. J. Pharm. 48, 247-254. Robertson, D.R.C., Walker, D.G., Renwick, A.G. and George, C.F. (1988) Age-related changes in the pharmacokinetics and pharmacodynamics of nifedipine. Br. J. Clin. Pharmacol. 25, 297-305.

solutions were produced by utilizing the reported ability of carboxymethylethylcellulose (CMEC) to inhibit crystal growth from aqueous solutions of nifedipine (Hasegawa et al., 1988). The donor phase consisted of 100 mg nifedipine dissolved with the aid of 300 mg CMEC ll in pH 7.4 medium at 37C. Fig. 8 shows the permeation profile observed across Celgard 2400 membrane into pH 7.4 receiving phase. After a short lag, an initial high flux of 8.4 kg cm-2 hh was observed, corresponding to delivery still of an inadequate daily dose of 3.19 mg nifedipine from a 16 cm2 device, and followed quickly by a tail-off of drug delivery as its solubility declined below saturation. Overall the results show poor transdermal delivery of nifedipine through hairless mouse skin, which do not indicate good penetration through human skin because it is known to be more resistant to drug permeation and the penetration enhancing effects of organic solvents. Initially the partition coefficient of nifedipine augers well for effective transdermal drug delivery, however other physicochemical factors associated with the drug can be linked to the poor transdermal delivery. The very low water solubility of the compound confirms a fundamental problem in that as the steady-state delivery of the drug across a membrane is subject to Ficks laws of diffusion, the lower the aqueous solubility of the drug the lower is its delivery rate. Coupled with this is the relatively large molecular weight of the drug (346 Da) and a calculated large molecular volume of 273 i3, both of which point to poor transderma1 delivery (Kasting et al., 1987). Thus, it appears that further investigation is necessary, possibly involving drug modification, before the transdermal delivery of nifedipine in adequate therapeutic dosage for human use becomes possible.

References
Barry, B.W. (1983) Dermatological Formulations: tion. Dekker, New York, pp. 138-145. Percutaneous Absorp