- Email: [email protected]
Formulation and evaluation of a testosterone transdermal spray
Formulation and Evaluation of a Testosterone Transdermal Spray ¨ THRICH,2 RICHARD H. GUY1,3 MARIE-LAURE LEICHTNAM,1,2 HERVE´ ROLLAND,2 PATRICK WU 1
School of Pharmaceutical Sciences, University of Geneva, 30, quai E. Ansermet, CH-1211 Geneva 4, Switzerland
2
Formulation Department, Technologie Servier, Drug Delivery Research, 45000 Orle´ans, France
3
Department of Pharmacy & Pharmacology, University of Bath, Claverton Down, Bath, BA2 7AY, United Kingdom
Received 10 August 2005; accepted 1 February 2006 Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.20641
ABSTRACT: The long-term goal is to develop a spray formulation for transdermal testosterone delivery, and to optimize the drug’s skin permeability. Testosterone transport from a series of ethanol/propylene glycol (PG)/water formulations was assessed in vitro across hairless rat skin, and the optimal composition determined. The formulation was then modified for delivery from a mechanical spray, and from an aerosol containing a high percentage of propellant. Drug transport was greatest from a saturated solution in 1:1:1 ethanol/PG/water (1.7 0.2 mg/cm2 h); five spray formulations were then tested, but only 1:1 ethanol/PG achieved a comparable flux. Increasing the % ethanol in the mixture increased evaporation rate but did not alter testosterone delivery. Formulation as an aerosol produced primarily unstable vehicles (phase separation, crystallization). Only 3:1 ethanol/PG remained stable, but no significant improvement in drug transport was observed (testosterone precipitated rapidly at the skin surface). The 1:1:1 ethanol/PG/water saturated solution suggested that some penetration enhancement was possible. Eliminating water to improve sprayability identified 1:1 ethanol/PG as a vehicle, which might allow transient supersaturation (and improved delivery). However, this effect was not improved by using a pressurized aerosol due to instability. Finally, testosterone fluxes were 5 to 10-fold lower than those required for useful transdermal therapy. ß 2006 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 95:1693–1702, 2006
Keywords: transport
aerosols; transdermal drug delivery; skin; formulation; passive diffusion;
INTRODUCTION Transdermal administration has developed, since its inception in the 1970s1 as an appropriate method with which to deliver drugs with low oral bioavailability.2 Testosterone is an example of such a compound, which also possesses suitable physicochemical properties for absorption across Correspondence to: Richard H. Guy 00.44.1225.384901; Fax: 00.44.1225.386114; E-mail: [email protected])
(Telephone:
Journal of Pharmaceutical Sciences, Vol. 95, 1693–1702 (2006) ß 2006 Wiley-Liss, Inc. and the American Pharmacists Association
the skin, notably a relatively low molecular weight (288 Da) and a reasonable lipophilicity (log octanol-water partition coefficient, log P ¼ 3.3).3 The transdermal ‘‘patch’’ is now a recognized dosage form, and multiple designs are available. Their cosmetic elegance is clear and their occlusive nature contributes to the provision of an appropriate drug flux across the skin by ensuring occlusion and full hydration of the skin’s principal barrier layer, the stratum corneum (SC). Patch formulations can also contain penetration enhancers to further boost the transport of the drug
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
1693
1694
LEICHTNAM ET AL.
across the skin, although these excipients, particularly under the occlusive conditions of transdermal delivery, can be quite irritating (see below). For testosterone, two types of patch have already been commercialized. Testoderm1 (Alza Corp., Palo Alto, CA), is a polymer matrix type patch, for scrotal application, containing 10 or 15 mg of drug of which 4–6 mg is delivered over a 24-h period.4,5 The site of application of this patch is designed to take advantage of the known, high permeability of scrotal skin. However, the skin does have to be shaved to ensure good adhesion and, not surprisingly, patient acceptance of this product is less than good. The second system, Androderm1 (Watson Pharma, Inc., Corona, CA), delivers 2.5 or 5 mg of testosterone per day and is designed for application to the back, abdomen, upper arms or thighs. The patch comprises a liquid drug reservoir, which includes excipients known to enhance testosterone transport across the skin. Local skin reactions are the most common adverse events reported for this patch.6 A less elegant, but nevertheless effective, alternative to a transdermal system is a conventional topical vehicle, such as a gel. Androgel1 (Solvay Pharmaceuticals, Marietta, GA), for example, is a 1% hydroalcoholic gel formulation. Such an approach can match the performance clinically of a patch, but typically with a lower incidence of skin irritation, and with good compliance. On the other hand, the formulation has to be applied over a significantly larger surface area (the control of which is obviously much less precise than that possible with a patch) and there is a significant risk of passive transfer of testosterone either to clothing or to another person.7 An alternative approach to patches and gels is the concept of a quick-drying and self-evidently non-occlusive spray. Such a system would be easy to use, well-tolerated, and allow for application of a metered dose over a fixed area of skin. For example, the idea has already been explored with isosorbide dinitrate,8 but poor bioavailability has meant that the spray must contact a large surface (200 cm2) to be effective. An improved approach is the so-called metered dose transdermal system (MDTS1, Acrux Inc., Melbourne, Australia): drug is formulated in a solution such that, postatomization, the volatile constituents evaporate rapidly leaving a highly concentrated, thin layer of drug in a residual vehicle that is rapidly taken up into the outer SC; drug is then released slowly from the reservoir created in this way over a prolonged period.9,10 Proof-of-concept has been JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
demonstrated with estradiol and granisetron, and preliminary information with respect to testosterone has appeared in the patent literature.11 To ensure dosing reproducibility, the spray is modified so that the formulation is pulverized perpendicularly to the skin surface over a controlled area that can be as small as approximately 10 cm2. Development of a delivery system of this type for testosterone may be expected to be a complicated process. The formulation is of necessity complex in that it must, in addition to more conventional excipients, contain a high level of environmentally acceptable propellant. The deposited film must then provide a transdermal flux of the drug effective for the treatment envisaged. For example, in the case of hypogonadism, the dose of testosterone needed corresponds to the delivery of 10 mg/cm2 h over a surface area of 20 cm2. The goals of the research described in this article, therefore, are to explore how the formulation of a spray (and the type of device used) impact upon testosterone permeability across the skin, and to develop formulations which can match the delivery achieved by existing patches and gels.
MATERIALS AND METHODS Materials Testosterone (Acros Organics, Noisy-le-Grand, France), ethanol ‘‘extra-neutre’’ (Sopar, MaisonAlfort, France), water for injection (Cooper, Melun, France), propylene glycol (PG; BASF, Paris, France), acetonitrile and methanol (both HPLC grade; Carlo Erba, Paris, France), phosphate buffered saline pH 7.4 (all components from Prolabo, Paris, France), HFA 134-A (Dupont De Nemours, Le Grand Sacconex, Switzerland).
Methods Testosterone Saturated Solubilities The solvents were different mixtures of ethanol, PG and/or water to which excess testosterone was added. The resulting suspensions were shaken at room temperature for more than 12 h, then filtered (pore size: 0.22 mm) prior to further examination. Saturated concentrations were determined for each solution by HPLC using the method described below. DOI 10.1002/jps
TESTOSTERONE TRANSDERMAL SPRAY
In Vitro Skin Permeation Experiments Rat skin was obtained from hairless males, aged 8 weeks (Charles River, Lyon, France). The animals were sacrificed by suffocation with CO2, and the dorsal skin was immediately excised and cleaned in normal saline. 500 mm of skin was removed with a dermatome (GA 630 Aesculap, Tuttlingen, Germany); the tissue was then wrapped in Parafilm1 and stored at 188C until use. Skin was clamped in vertical, flow-through diffusion cells with a donor surface area of 0.785 cm2 (Fig. 1). The receptor compartment was perfused (0.6 mL/min) with PBS at pH 7.4. During the 6-h experiment, the temperature of the skin was kept at 32 18C by a thermostated water jacket. The upper reservoir was filled with 250 mL of a saturated solution of testosterone and covered with Parafilm1. Samples (9 mL) were collected from receptor compartment at 0, 15, 30, 45, 60, 75, 90, 105, 120, 150, 180, 210, 240, 300, and 360 min post-dosing. Drug concentration in each sample was determined by HPLC. Each experiment was repeated 4 times. Cumulative amounts of drug (mg) penetrating per unit surface area (cm2) were graphed as a function of time (h). The in vitro skin permeation rate or steady-state flux (Jss) was calculated from the slope of the linear portion of this curve. Statistical comparisons were made using Student’s t-test with the level of significance set at p < 0.05. HPLC For isocratic chromatography, a Symmetry1 C18 column, 5 mm, 150 4.6 mm, (Waters, St Quentin en Yvelines, France) was used at 408C with a Spectra Serie P 100 HPLC-pump, a Spectra Serie AS 100 auto injector (Thermo Separation Pro-
Figure 1. Schematic representation of in vitro skin permeation experiments. DOI 10.1002/jps
1695
ducts, Les Ulis, France), and a Shimadzu SPD 10A VP absorbance detector (UV detection at 241 nm) (Shimadzu, Kyoto, Japan). A 60:40 (v/v) degassed mixture of acetonitrile/water was used as mobile phase. At a flow rate of 1 mL/min, the retention time of testosterone was 3.3 0.3 min. Unknown testosterone concentrations were calculated against known standards. Spray Device Assembling Two types of spray—respectively, mechanical and pressurized—were studied. The former are typically called non-pressurized dispensers and the dose of the active ingredient is delivered via a manual pump. Pressurized systems, also known as ‘‘aerosols,’’ comprise a canister containing a liquid under pressure from a compressed gas, which provides the driving force to expel the formulation via a valve. Unlike manual sprays, an aerosol allows the pulverization of viscous formulations (on the other hand, such vehicles pose compatibility and other problems). The mechanical sprays were prepared by filling glass containers with 10 mL of saturated drug solution. VP7/100 mL pumps (Valois, Marly le Roi, France) were crimped on glass bottles with a machine designed for that purpose (Socoge´, Colombes, France). An aluminium container, with a DF10/100 mL valve (Valois, Marly le Roi, France), held the aerosol which consisted of 8 mL of saturated drug solution. A pneumatic crimping and propellant filling machine (Pamasol P2005/20, Pamasol Willi-Ma¨der AG, Pfa¨ffikon-Switzerland) allowed 50 or 70% HFA 134a propellant to be added and the containers to be sealed. The miscibility of the mixture in HFA was visually evaluated. The two delivery systems were completed with actuators designed for pumps and valves, respectively (GP6868 actuators and actuators for Locabiotal1, Valois, Marly le Roi, France; Fig. 2).
Figure 2. Spray device assembly. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
1696
LEICHTNAM ET AL.
Figure 3. Experimental design for mixture: solvents repartition in the ternary mixtures. Each point corresponds to a formulation.
Formulation Development and Optimization It was hypothesized that an efficient formulation based upon ethanol, PG and water could be found. To do so, an experimental design was performed according to the simplex centroid model.12 Ten saturated formulations with various proportions of ethanol, PG and water were therefore prepared (Fig. 3). For each mixture, the ‘‘measured responses’’ detailed in Table 1 were determined and the results analyzed with NemrodW1 software, (R. Phan-Tan-Luu, M. Sergent, LPRAI, SARL). The measured responses included, of course, the saturation concentration of testosterone in each formulation, and the transdermal flux of the drug when applied in vitro in each corresponding mixture. In addition, the vehicles’ viscosities at 258C were determined (Ubbelohde microviscosimeter (Ic), AVS 310 type, Schott Gera¨te, Germany), as well as their surface tensions (Drop Shape Analysis G10/DSA10, Kru¨ss GmbH, Germany). Furthermore, Coomassie Blue was used to stain the formulations, which were then conditioned in a mechanical spray. A 100 mL pulverization from a fixed distance onto a filter paper was then made to evaluate the spray angle (a radians) of the deposited solution according to
a ¼ 2 x arctan(D/2d) (Fig. 4), where D is the diameter of the stain produced, and d is the distance between the actuator and the filter paper (in this case, 5 cm). In a second step, saturated testosterone solutions were selected from the results of the experimental design component and were pulverized (100 mL) directly onto hairless rat skin using the mechanical spray device described above. Immediately after dosing, the skin was mounted in the diffusion cell and the permeation of drug was evaluated as before (except that the donor compartment was not covered to simulate in vivo conditions as closely as possible). To maintain a constant distance between the point of exit of the spray from the device to the skin surface, a stainless steel funnel was attached as illustrated in Figure 2. At a distance of 5 cm, the spray produced a deposition of liquid over 19.6 cm2. To evaluate the amount of drug deposited by each pulverization, tests using cellulose filter paper (Fig. 5A) were carried out. The filter was then placed in methanol to extract the testosterone, which was subsequently quantified by HPLC. The same experiments were then performed on filters from which a central hole, of the exact area (0.785 cm2) as that of exposed skin in the diffusion cell, had been cut out (Fig. 5B). Hence, the
Table 1. Experimental Design for Mixtures: Responses Measured Responses Saturated testosterone concentration Viscosity Surface tension Spray angle Flux SD
mg/mL mm2/s mN/m Degrees mg/cm2 h
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
Figure 4. Spray angle (a) measurement system.13 DOI 10.1002/jps
TESTOSTERONE TRANSDERMAL SPRAY
1697
Figure 5. Evaluation of the amount of drug actually present on the skin during permeation experiments.
difference between the amounts of drug recovered from the two filters in the experiments illustrated in Figure 5 equaled the quantity of testosterone actually deposited on the skin.14 At the same time, the amount of drug ‘‘lost’’ on the funnel was rinsed off with methanol and subsequently quantified. It should also be mentioned that, at the end of the skin permeation experiments, the surface of the tissue was examined microscopically (Light microscope Jenavert 30-G0685, Jenoptik GmbH, Jena, Germany) for crystallization of the drug. Finally, binary ethanol/PG mixtures, containing 50, 60, 75, or 90% ethanol, were formulated into mechanical sprays or pressurized aerosol systems. Using the stainless steel funnel approach, saturated solutions were administered and drug delivery was determined. The quantity of drug deposited on the skin for each pulverization was evaluated as before.
RESULTS AND DISCUSSION Testosterone flux from a saturated ethanol/water solution was 1.2 0.2 mg/cm2 h in agreement with the literature.15 The presence of such a high ethanol content may have distinct influences on skin permeation. For example, under occlusion, both lipid extraction and fluidization may occur.16 In fact, it was noted that 100 mL of the applied formulation was lost during the 6-h experiment. Although ethanol is known to cross the skin rapidly (a steady-state flux of about 1 mg/cm2 h has been reported17), this transfer cannot account for the loss of vehicle observed. It would appear, therefore, that, despite the Parafilm1 covering the donor compartment, a significant loss of ethanol occurred by evaporation. This meant, of course, that testosterone must either crystallize out or that it adopts a metastable, supersaturated state; the latter would naturally lead to an DOI 10.1002/jps
enhanced level of transport18 and would be an obvious advantage for a spray formulation of the drug. However, crystallization was in fact observed and it was decided, as a result, to introduce PG into the formulation so as to increase viscosity and thereby, it was hypothesized, inhibit precipitation19 and enhance the stability of the supersaturated state. The experimental design adopted was chosen to ‘‘home in’’ on an optimal mixture of cosolvents such that the following criteria were satisfied: (1) the concentration of testosterone was sufficiently high that no significant depletion would occur over the course of the application period; (2) the viscosity and surface tension of the vehicle were such that sprayability and volatilization upon skin contact were balanced; (3) the angle of pulverization was minimized to avoid drug loss to the funnel upon activation of the device; and (4) drug flux was maximized. As seen in Table 2, the range of saturation concentrations of testosterone in the tested cosolvent mixtures was considerable; the solubility of the drug in ethanol is nearly 4,000 times greater than that in water. PG increases drug solubility in both ethanol and water, and also increased the viscosity of the formulations. In turn, this influenced the spray angle significantly, with smaller values of a being obtained for the vehicles containing the highest concentrations of PG. Surface tension, on the other hand, was less sensitive to the composition of the formulation. The skin permeation fluxes were first measured from saturated solutions and were found to vary over a 4-fold range; only that from a 1:1:1 ethanol/PG/water mixture was significantly different from all other measurements (Fig. 6). JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
1698
LEICHTNAM ET AL.
Table 2. Experimental Design for Mixtures: Results
EtOH 1 0 0 1 1 0 1 4 1 1
PG
Water
[Testosterone] mg/mL
Viscosity mm2/s
Surface Tension mN/m
Spray Angle (8)
Flux SD mg/cm2 h1
0 1 0 1 0 1 1 1 4 1
0 0 1 0 1 1 1 1 1 4
114.45 45.86 0.03 122.44 7.85 1.13 8.33 27.65 19.01 0.38
2.21 45.38 0.92 7.29 2.61 5.15 4.55 3.30 11.55 2.47
59.72 69.65 68.85 61.58 64.11 75.86 68.94 62.74 70.57 69.94
62.5 40.0 68.3 39.3 59.5 50.5 54.4 56.4 35.6 59.5
0.4 0.2 0.7 0.1 0.6 0.2 0.6 0.2 1.2 0.2 0.5 0.2 1.7 0.2 1.1 0.3 0.6 0.3 0.6 0.1
Interestingly, the experimental design approach identified a formulation containing 74% ethanol, 12% PG and 14% water to be optimal, a composition quite close to the 4:1:1 ethanol/PG/ water mixture, which was included in the investigation. Because of this finding, this formulation was studied further in the course of this and subsequent work, despite the fact that it underperformed the 1:1:1 mixture in terms of testosterone delivery. Theory predicts that drug flux from saturated solution should be independent of the vehicle provided that it does not alter in some way the barrier function of the skin (i.e., that there is a penetration enhancement—or retardation— effect).20 It would appear, therefore, that the 1:1:1 ethanol/PG/water combination either facilitates testosterone diffusion or alters, in a perhaps more subtle way, the solubility of the drug in the
SC. While there has been much speculation in the literature concerning the mechanism of action of PG (with or without ethanol and/or water present),21–26 the data obtained in the present study cannot reveal any further details at the molecular level. In addition, the highest flux obtained (1.7 mg/cm2 h) was considerably less than the proposed target of 10 mg/cm2 h. Consequently, a selection of the formulations were then tested following delivery from a mechanical spray device, anticipating that the evaporation process would produce at least a transient period of supersaturation and an increased delivery of the drug as a result. Five formulations (Table 3) were selected for this next phase of the investigation. The choice of vehicles was based on two criteria: (i) as only 100 mL were to be pulverized onto the skin, cosolvent mixtures in which testosterone
Figure 6. Permeation of testosterone from different ethanol/PG/water solutions (mean SD, n ¼ 4). *Significantly different from all other values (p < 0.05). JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
DOI 10.1002/jps
TESTOSTERONE TRANSDERMAL SPRAY
Table 3. Performance of the Five Formulations Selected for Testosterone Delivery from a Mechanical Spray
EtOH 1 1 4 1 1
PG
Water
Spray Angle (8)
% loss to Funnel
Flux SD mg/cm2 h
0 0 1 1 1
0 1 1 1 0
62.5 59.5 56.4 54.4 39.3
70.3 64.5 47.5 33.3 15.3
0.7 0.3 1.5 0.2 1.2 0.2 0.9 0.2 1.0 0.2
solubility was less than 1 mg/mL were eliminated to avoid problems of depletion; and (ii) highly viscous formulations (those with large proportions of PG) were not considered for practical reasons and because their evaporation rates were very slow. Even among the five selected formulations, there were important differences in the amount of testosterone eventually deposited on the target. With increasing spray angle (a), progressively greater quantities of drug were lost onto the funnel attached to the spray (Table 3). As expected, vehicles with more PG, which were more viscous, had narrower spray angles and less loss of the dose to the device. The skin fluxes achieved from the sprays are in Table 3 and Figure 7. The 1:1 ethanol/PG mixture produced the highest permeation (1.5 0.2 mg/ cm h), approximately 2.5-fold that from the equivalent saturated solution. Given that the latter demonstrated no classical enhancement effect, it may be tentatively concluded, therefore,
1699
that the spray application induced a transient supersaturated state (a phenomenon that has been alluded to in other recent work).27 The other vehicles did not show any dramatic differences, in terms of testosterone flux, compared to the saturated solutions; indeed, the 1:1:1 ethanol/PG/ water mixture was less effective as a spray implying that the apparent enhancement effect seen with the solution could not be sustained when evaporation of volatile components was allowed to occur. The third stage of the research then focussed on the optimization of the ethanol/PG balance in the vehicle and an examination of the practicality of testosterone delivery via an aerosol. Initial experiments continued with the mechanical spray, though, and it was argued that a mixture with a higher proportion of ethanol should have enhanced evaporation and (perhaps) achieve a higher degree of supersaturation. Four binary mixtures of ethanol/PG (1:1, 3:2, 3:1, and 9:1) were evaluated, therefore, and the resulting testosterone fluxes are in Figure 8. At 6 h post-dosing, testosterone crystallization was apparent for all formulations, and transport rates were rather similar (differing insignificantly by less than a factor of two). Because an earlier study had shown that a pressurized device produced finer droplets than a mechanical spray,14 it was reasonable to ask whether this might lead to a more uniform and complete coverage of the skin surface, with a concomitant improvement in evaporative loss and, as a result, transdermal transport. Aerosols
Figure 7. Permeation of testosterone from mechanical sprays containing different ratios of EtOH/PG/water (mean SD, n ¼ 4). DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
1700
LEICHTNAM ET AL.
Table 4. Physical Stability of Various Ethanol/PG Formulations of Testosterone in the Presence of HFA EtOH/PG (% v/v) 1:1 3:2 3:1
9:1
Figure 8. Permeation of testosterone from mechanical sprays containing different ratio of EtOH/PG (mean SD, n ¼ 4).
typically contain large amounts of propellant (50– 95%); given that the solubility of testosterone in the four binary mixtures of ethanol/PG varied between about 100 and 200 mg/mL (for the ethanol/PG ratios of 1:1, 3:2, 3:1, and 9:1, the saturation concentrations were 126, 176, 191, and 202 mg/mL, respectively), and that the valve capacity of the aerosol was 100 mL, the level of HFA used was set at either 50 or 70% to ensure that the theoretical amounts of drug delivered in a pulverization (6.3–10.1 mg for 50% HFA; 3.8– 6.1 mg for 70% HFA) were sufficient that skin permeation could be quantified. The next step was to visually examine the physical stability of the aerosol solutions. Although HFA is miscible with ethanol and PG,28 crystallization or phase separation occurred with all but one of the mixtures studied (Table 4). Only the 3:1 ethanol/PG mixture with 50% HFA was stable. Delivery of testosterone from this formulation as an aerosol was compared with those from a spray and a solution. Despite the wide difference in the quantity of drug deposited on the skin (0.84, 3.34, and 59.7 mg/cm2 from aerosol, spray and solution, respectively), the transport profiles were indistinguishable (Fig. 9). Crystallization on the skin was again observed for the spray and aerosol, whereas the solution (250 mL applied) did not display this problem. It seems that the soughtafter supersaturation effect was not achieved by the aerosol, therefore, even though a smaller JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
HFA
Observations
50% 70% 50% 70% 50% 70%
Immediate phase out Immediate phase out Phase out in 2–3 min Immediate phase out Stable Persistent trouble then phase out after 2 h Crystallization after about 30 min Crystallization after about 30 min
50% 70%
droplet size was possible. It is clearly necessary in further work to identify a means with which to stabilize the formulation as the ethanol evaporates (for example, by the inclusion of anti-nucleant polymers);29 also, given the fluxes achieved todate, identification of a penetration enhancer, which can boost testosterone delivery, is indicated. Two additional facts reinforce this final point: first, the animal skin model used here has a higher inherent permeability than the human barrier;30 second, although no evidence of testosterone metabolism during its percutaneous delivery was found (not surprising given that the skin had been frozen post-excision), skin biotransformation may occur in vivo, and may be expected to reduce the effective transdermal bioavailability of the hormone.
Figure 9. In vitro testosterone permeation from an ethanol/PG (3:1) saturated formulation delivered as either a spray (^), an aerosol ( ), or a solution (&). Mean SD, n ¼ 4. DOI 10.1002/jps
TESTOSTERONE TRANSDERMAL SPRAY
CONCLUSIONS Optimization of a formulation for transdermal delivery of a drug in the form of a spray is a complex process. While a carefully chosen experimental design using saturated solutions led to the identification of a 1:1:1 mixture of ethanol, PG and water as the most effective vehicle, the presence of water in a spray was found to be unnecessary. In this case, a simple 1:1 ethanol/ PG formulation performed best, and suggested the potential to achieve transient supersaturation. Aerosolization demanded the significant presence of HFA and pointed to the possible benefit of increasing the ethanol level at the expense of PG, both to realize a more practical formulation and to enhance the chances that supersaturation occurred. Unfortunately, this promise could not be demonstrated due to significant stability problems and crystallization of the drug; no significant improvement in testosterone’s steady-state flux beyond 2 mg/cm2 h could be achieved. Nevertheless, the study serves as an important starting point for further work that must address both the need for penetration enhancement and the stabilization of a supersaturated state post-aerosolization.
ACKNOWLEDGMENTS We thank the French National Association of Technical Research and the French Ministry of Research for financial support (Convention CIFRE no 675/2001).
REFERENCES 1. Chien YW. 1987. Advances in transdermal systemic medication. In: Chien YW, editor. Transdermal controlled systemic medications. New York: Marcel Dekker. pp 1–22. 2. Beckett AH. 1982. Possibilities and limitations of transdermal absorption. In: Brandau R, Lippold BH, editors. Dermal and transdermal absorption, Stuttgart: Wissenschaftliche Verlagsgesellschaft mbH. pp 154–170. 3. Mazer NA, Heiber WE, Moellmer JF, Meikle WA, Stringham JD, Sanders SW, Tollman KG, Odell WD. 1992. Enhanced transdermal delivery of testosterone: A new physiological approach for androgen replacement in hypogonadal men. J Control Release 19:347–362.
DOI 10.1002/jps
1701
4. Higuchi T, Hussain A. 1979. Device consisting of copolymer having acetoxy groups for delivery of drugs. US Patent 4,144,317. 5. Campell PS, Eckenhoff JB. 1987. Transdermal therapeutic system having improved delivery characteristic. US Patent 4,704,282. 6. Meikle AW. 1998. A permeation-enhanced nonscrotal testosterone transdermal system for the treatment of male hypogonadism. In: Nieschlag E, Behre HM, editors. Testosterone—action, deficiency, substitution, 2nd edition Berlin: SpringerVerlag. pp 389–422. 7. Schaison G, Couzinet B, Behre HM. 1998. Percutaneous dihydrotestosterone treatment. In: Nieschlag E, editor. Testosterone—action, deficiency, substitution, 2nd edition Berlin: Springer-Verlag. pp 329–348. 8. Laufen H, Leitold M. 1992. Biodisponibility and metabolism of isosorbide dinitrate from a transdermal spray. Arzneim Forsch 42:931–935. 9. Morgan TM, Parr RA, Reed BL, Finnin BC. 1998. Enhanced transdermal delivery of sex hormones in swine with a novel topical aerosol. J Pharm Sci 87: 1219–1225. 10. Morgan TM, Reed BL, Finnin BC. 1998. Enhanced skin permeation of sex hormones with novel topical spray vehicles. J Pharm Sci 87:1213– 1218. 11. Monash University. Dermal penetration enhancers and drug delivery systems involving same. Australia. Patent WO 97/29735. 12. Huisman R, Van Kamp HV, Weyland JW, Doornbos DA, Bolhuis GK, Lerk CF. 1984. Development and optimization of pharmaceutical formulations using a simplex lattice design. Pharm Weekbl Sci 6:185– 194. 13. Leichtnam ML, Rolland H, Wu¨thrich P. 2003. New aerosol transdermal drug delivery system—Effects of formulation parameters upon spray characteristics and permeation enhancement. Proc APGI, Symp Skin Formulation #43. 14. Leichtnam ML, Rolland H, Wu¨thrich P, Guy RH. 2005. Identification of penetration enhancers for testosterone transdermal delivery from spray formulations. J Controlled Release (in press). 15. Kim MK, Lee CH, Kim DD. 2000. Skin permeation of testosterone and its ester derivatives. J Pharm Pharmacol 52:369–375. 16. Hatanaka T, Shimoyama M, Sugibayashi K, Morimoto Y. 1993. Effect of vehicle on the skin permeability of drugs:polyethylene glycol 400water and ethanol-water binary solvents. J Control Release 23:247–260. 17. Berner B, Mazzenga GC, Otte JH, Steffens RJ, Juang R-H, Ebert CD. 1989. Ethanol:water mutually enhanced transdermal therapeutic system:II. Skin permeation of ethanol and nitroglycerin, J Pharm Sci 78:402–407.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
1702
LEICHTNAM ET AL.
18. Williams AC, Barry BW. 2004. Penetration enhancers. Adv Drug Del Rev 56:603–618. 19. Davis J, Hadgraft J. 1993. Supersaturated solutions as topical drug delivery systems. In: Walters K, Hadgraft J, editors. Pharmaceutical skin permeation enhancement, New York: Marcel Dekker. pp 243–267. 20. Higuchi T. 1960. Physical chemical analysis of percutaneous absorption process. J Soc Cosmetic Chemists 11:85–97. 21. Ostrenga J, Steinmetz C, Poulsen B. 1971. Significance of vehicle composition. I: relationship between topical vehicle composition, skin penetrability, and clinical efficacy. J Pharm Sci 60:1175– 1179. 22. Ostrenga J, Steinmetz C, Poulsen B, Yett S. 1971. Significance of vehicle composition. II: prediction of optimal vehicle composition. J Pharm Sci 60:1180– 1183. 23. Woodford R, Barry BW. 1982. Optimisation of bioavailability of topical steroids: Thermodynamic control. J Invest Dermatol 79:388–391. 24. Barry BW. 1983. Properties that influence percutaneous absorption. In: Barry BW, editor. Dermatological formulations, New York: Marcel Dekker. pp. 127–233.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
25. Holegaard A, Møllgaard B. 1985. Dermal drug delivery-improvement by choice of vehicle or drug derivative. J Control Release 2:111–120. 26. Panchagnula R, Salve PS, Thomas NS, Jain AK, Ramarao P. 2001. Transdermal delivery of naloxone: Effect of water, propylene glycol, ethanol and their binary combinations on permeation through rat skin. Int J Pharm 219:95– 105. 27. Wilkins NF, Klose K, Morgan TM, Reed BL, Finnin BC. 2004. Metastable pharmaceutical compositions. Patent WO200400361. 28. Rowe R, Sheskey P, Weller P, editors. 1995. CDRom handbook of pharmaceutical excipients, 4th edition. London, UK: Science and Practice, Royal Pharmaceutical Society of Great Britain. 29. Fang J-Y, Kuo C-T, Huang Y-B, Wu P-C, Tsai Y-H. 1999. Transdermal delivery of sodium nonivamide acetate from volatile vehicles: Effects of polymers. Int J Pharm 176:156–167. 30. Sekkat N, Guy RH. 2001. Biological models to study skin permeation. In: Testa B, van der Waterbeemd H, Folkers G, Guy RH, editors. Pharmacokinetic optimization in drug research: Biological, physicochemical and computational strategies, Lausanne: Wiley-VCH. pp 155–172.
DOI 10.1002/jps
School of Pharmaceutical Sciences, University of Geneva, 30, quai E. Ansermet, CH-1211 Geneva 4, Switzerland
2
Formulation Department, Technologie Servier, Drug Delivery Research, 45000 Orle´ans, France
3
Department of Pharmacy & Pharmacology, University of Bath, Claverton Down, Bath, BA2 7AY, United Kingdom
Received 10 August 2005; accepted 1 February 2006 Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.20641
ABSTRACT: The long-term goal is to develop a spray formulation for transdermal testosterone delivery, and to optimize the drug’s skin permeability. Testosterone transport from a series of ethanol/propylene glycol (PG)/water formulations was assessed in vitro across hairless rat skin, and the optimal composition determined. The formulation was then modified for delivery from a mechanical spray, and from an aerosol containing a high percentage of propellant. Drug transport was greatest from a saturated solution in 1:1:1 ethanol/PG/water (1.7 0.2 mg/cm2 h); five spray formulations were then tested, but only 1:1 ethanol/PG achieved a comparable flux. Increasing the % ethanol in the mixture increased evaporation rate but did not alter testosterone delivery. Formulation as an aerosol produced primarily unstable vehicles (phase separation, crystallization). Only 3:1 ethanol/PG remained stable, but no significant improvement in drug transport was observed (testosterone precipitated rapidly at the skin surface). The 1:1:1 ethanol/PG/water saturated solution suggested that some penetration enhancement was possible. Eliminating water to improve sprayability identified 1:1 ethanol/PG as a vehicle, which might allow transient supersaturation (and improved delivery). However, this effect was not improved by using a pressurized aerosol due to instability. Finally, testosterone fluxes were 5 to 10-fold lower than those required for useful transdermal therapy. ß 2006 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 95:1693–1702, 2006
Keywords: transport
aerosols; transdermal drug delivery; skin; formulation; passive diffusion;
INTRODUCTION Transdermal administration has developed, since its inception in the 1970s1 as an appropriate method with which to deliver drugs with low oral bioavailability.2 Testosterone is an example of such a compound, which also possesses suitable physicochemical properties for absorption across Correspondence to: Richard H. Guy 00.44.1225.384901; Fax: 00.44.1225.386114; E-mail: [email protected])
(Telephone:
Journal of Pharmaceutical Sciences, Vol. 95, 1693–1702 (2006) ß 2006 Wiley-Liss, Inc. and the American Pharmacists Association
the skin, notably a relatively low molecular weight (288 Da) and a reasonable lipophilicity (log octanol-water partition coefficient, log P ¼ 3.3).3 The transdermal ‘‘patch’’ is now a recognized dosage form, and multiple designs are available. Their cosmetic elegance is clear and their occlusive nature contributes to the provision of an appropriate drug flux across the skin by ensuring occlusion and full hydration of the skin’s principal barrier layer, the stratum corneum (SC). Patch formulations can also contain penetration enhancers to further boost the transport of the drug
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
1693
1694
LEICHTNAM ET AL.
across the skin, although these excipients, particularly under the occlusive conditions of transdermal delivery, can be quite irritating (see below). For testosterone, two types of patch have already been commercialized. Testoderm1 (Alza Corp., Palo Alto, CA), is a polymer matrix type patch, for scrotal application, containing 10 or 15 mg of drug of which 4–6 mg is delivered over a 24-h period.4,5 The site of application of this patch is designed to take advantage of the known, high permeability of scrotal skin. However, the skin does have to be shaved to ensure good adhesion and, not surprisingly, patient acceptance of this product is less than good. The second system, Androderm1 (Watson Pharma, Inc., Corona, CA), delivers 2.5 or 5 mg of testosterone per day and is designed for application to the back, abdomen, upper arms or thighs. The patch comprises a liquid drug reservoir, which includes excipients known to enhance testosterone transport across the skin. Local skin reactions are the most common adverse events reported for this patch.6 A less elegant, but nevertheless effective, alternative to a transdermal system is a conventional topical vehicle, such as a gel. Androgel1 (Solvay Pharmaceuticals, Marietta, GA), for example, is a 1% hydroalcoholic gel formulation. Such an approach can match the performance clinically of a patch, but typically with a lower incidence of skin irritation, and with good compliance. On the other hand, the formulation has to be applied over a significantly larger surface area (the control of which is obviously much less precise than that possible with a patch) and there is a significant risk of passive transfer of testosterone either to clothing or to another person.7 An alternative approach to patches and gels is the concept of a quick-drying and self-evidently non-occlusive spray. Such a system would be easy to use, well-tolerated, and allow for application of a metered dose over a fixed area of skin. For example, the idea has already been explored with isosorbide dinitrate,8 but poor bioavailability has meant that the spray must contact a large surface (200 cm2) to be effective. An improved approach is the so-called metered dose transdermal system (MDTS1, Acrux Inc., Melbourne, Australia): drug is formulated in a solution such that, postatomization, the volatile constituents evaporate rapidly leaving a highly concentrated, thin layer of drug in a residual vehicle that is rapidly taken up into the outer SC; drug is then released slowly from the reservoir created in this way over a prolonged period.9,10 Proof-of-concept has been JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
demonstrated with estradiol and granisetron, and preliminary information with respect to testosterone has appeared in the patent literature.11 To ensure dosing reproducibility, the spray is modified so that the formulation is pulverized perpendicularly to the skin surface over a controlled area that can be as small as approximately 10 cm2. Development of a delivery system of this type for testosterone may be expected to be a complicated process. The formulation is of necessity complex in that it must, in addition to more conventional excipients, contain a high level of environmentally acceptable propellant. The deposited film must then provide a transdermal flux of the drug effective for the treatment envisaged. For example, in the case of hypogonadism, the dose of testosterone needed corresponds to the delivery of 10 mg/cm2 h over a surface area of 20 cm2. The goals of the research described in this article, therefore, are to explore how the formulation of a spray (and the type of device used) impact upon testosterone permeability across the skin, and to develop formulations which can match the delivery achieved by existing patches and gels.
MATERIALS AND METHODS Materials Testosterone (Acros Organics, Noisy-le-Grand, France), ethanol ‘‘extra-neutre’’ (Sopar, MaisonAlfort, France), water for injection (Cooper, Melun, France), propylene glycol (PG; BASF, Paris, France), acetonitrile and methanol (both HPLC grade; Carlo Erba, Paris, France), phosphate buffered saline pH 7.4 (all components from Prolabo, Paris, France), HFA 134-A (Dupont De Nemours, Le Grand Sacconex, Switzerland).
Methods Testosterone Saturated Solubilities The solvents were different mixtures of ethanol, PG and/or water to which excess testosterone was added. The resulting suspensions were shaken at room temperature for more than 12 h, then filtered (pore size: 0.22 mm) prior to further examination. Saturated concentrations were determined for each solution by HPLC using the method described below. DOI 10.1002/jps
TESTOSTERONE TRANSDERMAL SPRAY
In Vitro Skin Permeation Experiments Rat skin was obtained from hairless males, aged 8 weeks (Charles River, Lyon, France). The animals were sacrificed by suffocation with CO2, and the dorsal skin was immediately excised and cleaned in normal saline. 500 mm of skin was removed with a dermatome (GA 630 Aesculap, Tuttlingen, Germany); the tissue was then wrapped in Parafilm1 and stored at 188C until use. Skin was clamped in vertical, flow-through diffusion cells with a donor surface area of 0.785 cm2 (Fig. 1). The receptor compartment was perfused (0.6 mL/min) with PBS at pH 7.4. During the 6-h experiment, the temperature of the skin was kept at 32 18C by a thermostated water jacket. The upper reservoir was filled with 250 mL of a saturated solution of testosterone and covered with Parafilm1. Samples (9 mL) were collected from receptor compartment at 0, 15, 30, 45, 60, 75, 90, 105, 120, 150, 180, 210, 240, 300, and 360 min post-dosing. Drug concentration in each sample was determined by HPLC. Each experiment was repeated 4 times. Cumulative amounts of drug (mg) penetrating per unit surface area (cm2) were graphed as a function of time (h). The in vitro skin permeation rate or steady-state flux (Jss) was calculated from the slope of the linear portion of this curve. Statistical comparisons were made using Student’s t-test with the level of significance set at p < 0.05. HPLC For isocratic chromatography, a Symmetry1 C18 column, 5 mm, 150 4.6 mm, (Waters, St Quentin en Yvelines, France) was used at 408C with a Spectra Serie P 100 HPLC-pump, a Spectra Serie AS 100 auto injector (Thermo Separation Pro-
Figure 1. Schematic representation of in vitro skin permeation experiments. DOI 10.1002/jps
1695
ducts, Les Ulis, France), and a Shimadzu SPD 10A VP absorbance detector (UV detection at 241 nm) (Shimadzu, Kyoto, Japan). A 60:40 (v/v) degassed mixture of acetonitrile/water was used as mobile phase. At a flow rate of 1 mL/min, the retention time of testosterone was 3.3 0.3 min. Unknown testosterone concentrations were calculated against known standards. Spray Device Assembling Two types of spray—respectively, mechanical and pressurized—were studied. The former are typically called non-pressurized dispensers and the dose of the active ingredient is delivered via a manual pump. Pressurized systems, also known as ‘‘aerosols,’’ comprise a canister containing a liquid under pressure from a compressed gas, which provides the driving force to expel the formulation via a valve. Unlike manual sprays, an aerosol allows the pulverization of viscous formulations (on the other hand, such vehicles pose compatibility and other problems). The mechanical sprays were prepared by filling glass containers with 10 mL of saturated drug solution. VP7/100 mL pumps (Valois, Marly le Roi, France) were crimped on glass bottles with a machine designed for that purpose (Socoge´, Colombes, France). An aluminium container, with a DF10/100 mL valve (Valois, Marly le Roi, France), held the aerosol which consisted of 8 mL of saturated drug solution. A pneumatic crimping and propellant filling machine (Pamasol P2005/20, Pamasol Willi-Ma¨der AG, Pfa¨ffikon-Switzerland) allowed 50 or 70% HFA 134a propellant to be added and the containers to be sealed. The miscibility of the mixture in HFA was visually evaluated. The two delivery systems were completed with actuators designed for pumps and valves, respectively (GP6868 actuators and actuators for Locabiotal1, Valois, Marly le Roi, France; Fig. 2).
Figure 2. Spray device assembly. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
1696
LEICHTNAM ET AL.
Figure 3. Experimental design for mixture: solvents repartition in the ternary mixtures. Each point corresponds to a formulation.
Formulation Development and Optimization It was hypothesized that an efficient formulation based upon ethanol, PG and water could be found. To do so, an experimental design was performed according to the simplex centroid model.12 Ten saturated formulations with various proportions of ethanol, PG and water were therefore prepared (Fig. 3). For each mixture, the ‘‘measured responses’’ detailed in Table 1 were determined and the results analyzed with NemrodW1 software, (R. Phan-Tan-Luu, M. Sergent, LPRAI, SARL). The measured responses included, of course, the saturation concentration of testosterone in each formulation, and the transdermal flux of the drug when applied in vitro in each corresponding mixture. In addition, the vehicles’ viscosities at 258C were determined (Ubbelohde microviscosimeter (Ic), AVS 310 type, Schott Gera¨te, Germany), as well as their surface tensions (Drop Shape Analysis G10/DSA10, Kru¨ss GmbH, Germany). Furthermore, Coomassie Blue was used to stain the formulations, which were then conditioned in a mechanical spray. A 100 mL pulverization from a fixed distance onto a filter paper was then made to evaluate the spray angle (a radians) of the deposited solution according to
a ¼ 2 x arctan(D/2d) (Fig. 4), where D is the diameter of the stain produced, and d is the distance between the actuator and the filter paper (in this case, 5 cm). In a second step, saturated testosterone solutions were selected from the results of the experimental design component and were pulverized (100 mL) directly onto hairless rat skin using the mechanical spray device described above. Immediately after dosing, the skin was mounted in the diffusion cell and the permeation of drug was evaluated as before (except that the donor compartment was not covered to simulate in vivo conditions as closely as possible). To maintain a constant distance between the point of exit of the spray from the device to the skin surface, a stainless steel funnel was attached as illustrated in Figure 2. At a distance of 5 cm, the spray produced a deposition of liquid over 19.6 cm2. To evaluate the amount of drug deposited by each pulverization, tests using cellulose filter paper (Fig. 5A) were carried out. The filter was then placed in methanol to extract the testosterone, which was subsequently quantified by HPLC. The same experiments were then performed on filters from which a central hole, of the exact area (0.785 cm2) as that of exposed skin in the diffusion cell, had been cut out (Fig. 5B). Hence, the
Table 1. Experimental Design for Mixtures: Responses Measured Responses Saturated testosterone concentration Viscosity Surface tension Spray angle Flux SD
mg/mL mm2/s mN/m Degrees mg/cm2 h
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
Figure 4. Spray angle (a) measurement system.13 DOI 10.1002/jps
TESTOSTERONE TRANSDERMAL SPRAY
1697
Figure 5. Evaluation of the amount of drug actually present on the skin during permeation experiments.
difference between the amounts of drug recovered from the two filters in the experiments illustrated in Figure 5 equaled the quantity of testosterone actually deposited on the skin.14 At the same time, the amount of drug ‘‘lost’’ on the funnel was rinsed off with methanol and subsequently quantified. It should also be mentioned that, at the end of the skin permeation experiments, the surface of the tissue was examined microscopically (Light microscope Jenavert 30-G0685, Jenoptik GmbH, Jena, Germany) for crystallization of the drug. Finally, binary ethanol/PG mixtures, containing 50, 60, 75, or 90% ethanol, were formulated into mechanical sprays or pressurized aerosol systems. Using the stainless steel funnel approach, saturated solutions were administered and drug delivery was determined. The quantity of drug deposited on the skin for each pulverization was evaluated as before.
RESULTS AND DISCUSSION Testosterone flux from a saturated ethanol/water solution was 1.2 0.2 mg/cm2 h in agreement with the literature.15 The presence of such a high ethanol content may have distinct influences on skin permeation. For example, under occlusion, both lipid extraction and fluidization may occur.16 In fact, it was noted that 100 mL of the applied formulation was lost during the 6-h experiment. Although ethanol is known to cross the skin rapidly (a steady-state flux of about 1 mg/cm2 h has been reported17), this transfer cannot account for the loss of vehicle observed. It would appear, therefore, that, despite the Parafilm1 covering the donor compartment, a significant loss of ethanol occurred by evaporation. This meant, of course, that testosterone must either crystallize out or that it adopts a metastable, supersaturated state; the latter would naturally lead to an DOI 10.1002/jps
enhanced level of transport18 and would be an obvious advantage for a spray formulation of the drug. However, crystallization was in fact observed and it was decided, as a result, to introduce PG into the formulation so as to increase viscosity and thereby, it was hypothesized, inhibit precipitation19 and enhance the stability of the supersaturated state. The experimental design adopted was chosen to ‘‘home in’’ on an optimal mixture of cosolvents such that the following criteria were satisfied: (1) the concentration of testosterone was sufficiently high that no significant depletion would occur over the course of the application period; (2) the viscosity and surface tension of the vehicle were such that sprayability and volatilization upon skin contact were balanced; (3) the angle of pulverization was minimized to avoid drug loss to the funnel upon activation of the device; and (4) drug flux was maximized. As seen in Table 2, the range of saturation concentrations of testosterone in the tested cosolvent mixtures was considerable; the solubility of the drug in ethanol is nearly 4,000 times greater than that in water. PG increases drug solubility in both ethanol and water, and also increased the viscosity of the formulations. In turn, this influenced the spray angle significantly, with smaller values of a being obtained for the vehicles containing the highest concentrations of PG. Surface tension, on the other hand, was less sensitive to the composition of the formulation. The skin permeation fluxes were first measured from saturated solutions and were found to vary over a 4-fold range; only that from a 1:1:1 ethanol/PG/water mixture was significantly different from all other measurements (Fig. 6). JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
1698
LEICHTNAM ET AL.
Table 2. Experimental Design for Mixtures: Results
EtOH 1 0 0 1 1 0 1 4 1 1
PG
Water
[Testosterone] mg/mL
Viscosity mm2/s
Surface Tension mN/m
Spray Angle (8)
Flux SD mg/cm2 h1
0 1 0 1 0 1 1 1 4 1
0 0 1 0 1 1 1 1 1 4
114.45 45.86 0.03 122.44 7.85 1.13 8.33 27.65 19.01 0.38
2.21 45.38 0.92 7.29 2.61 5.15 4.55 3.30 11.55 2.47
59.72 69.65 68.85 61.58 64.11 75.86 68.94 62.74 70.57 69.94
62.5 40.0 68.3 39.3 59.5 50.5 54.4 56.4 35.6 59.5
0.4 0.2 0.7 0.1 0.6 0.2 0.6 0.2 1.2 0.2 0.5 0.2 1.7 0.2 1.1 0.3 0.6 0.3 0.6 0.1
Interestingly, the experimental design approach identified a formulation containing 74% ethanol, 12% PG and 14% water to be optimal, a composition quite close to the 4:1:1 ethanol/PG/ water mixture, which was included in the investigation. Because of this finding, this formulation was studied further in the course of this and subsequent work, despite the fact that it underperformed the 1:1:1 mixture in terms of testosterone delivery. Theory predicts that drug flux from saturated solution should be independent of the vehicle provided that it does not alter in some way the barrier function of the skin (i.e., that there is a penetration enhancement—or retardation— effect).20 It would appear, therefore, that the 1:1:1 ethanol/PG/water combination either facilitates testosterone diffusion or alters, in a perhaps more subtle way, the solubility of the drug in the
SC. While there has been much speculation in the literature concerning the mechanism of action of PG (with or without ethanol and/or water present),21–26 the data obtained in the present study cannot reveal any further details at the molecular level. In addition, the highest flux obtained (1.7 mg/cm2 h) was considerably less than the proposed target of 10 mg/cm2 h. Consequently, a selection of the formulations were then tested following delivery from a mechanical spray device, anticipating that the evaporation process would produce at least a transient period of supersaturation and an increased delivery of the drug as a result. Five formulations (Table 3) were selected for this next phase of the investigation. The choice of vehicles was based on two criteria: (i) as only 100 mL were to be pulverized onto the skin, cosolvent mixtures in which testosterone
Figure 6. Permeation of testosterone from different ethanol/PG/water solutions (mean SD, n ¼ 4). *Significantly different from all other values (p < 0.05). JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
DOI 10.1002/jps
TESTOSTERONE TRANSDERMAL SPRAY
Table 3. Performance of the Five Formulations Selected for Testosterone Delivery from a Mechanical Spray
EtOH 1 1 4 1 1
PG
Water
Spray Angle (8)
% loss to Funnel
Flux SD mg/cm2 h
0 0 1 1 1
0 1 1 1 0
62.5 59.5 56.4 54.4 39.3
70.3 64.5 47.5 33.3 15.3
0.7 0.3 1.5 0.2 1.2 0.2 0.9 0.2 1.0 0.2
solubility was less than 1 mg/mL were eliminated to avoid problems of depletion; and (ii) highly viscous formulations (those with large proportions of PG) were not considered for practical reasons and because their evaporation rates were very slow. Even among the five selected formulations, there were important differences in the amount of testosterone eventually deposited on the target. With increasing spray angle (a), progressively greater quantities of drug were lost onto the funnel attached to the spray (Table 3). As expected, vehicles with more PG, which were more viscous, had narrower spray angles and less loss of the dose to the device. The skin fluxes achieved from the sprays are in Table 3 and Figure 7. The 1:1 ethanol/PG mixture produced the highest permeation (1.5 0.2 mg/ cm h), approximately 2.5-fold that from the equivalent saturated solution. Given that the latter demonstrated no classical enhancement effect, it may be tentatively concluded, therefore,
1699
that the spray application induced a transient supersaturated state (a phenomenon that has been alluded to in other recent work).27 The other vehicles did not show any dramatic differences, in terms of testosterone flux, compared to the saturated solutions; indeed, the 1:1:1 ethanol/PG/ water mixture was less effective as a spray implying that the apparent enhancement effect seen with the solution could not be sustained when evaporation of volatile components was allowed to occur. The third stage of the research then focussed on the optimization of the ethanol/PG balance in the vehicle and an examination of the practicality of testosterone delivery via an aerosol. Initial experiments continued with the mechanical spray, though, and it was argued that a mixture with a higher proportion of ethanol should have enhanced evaporation and (perhaps) achieve a higher degree of supersaturation. Four binary mixtures of ethanol/PG (1:1, 3:2, 3:1, and 9:1) were evaluated, therefore, and the resulting testosterone fluxes are in Figure 8. At 6 h post-dosing, testosterone crystallization was apparent for all formulations, and transport rates were rather similar (differing insignificantly by less than a factor of two). Because an earlier study had shown that a pressurized device produced finer droplets than a mechanical spray,14 it was reasonable to ask whether this might lead to a more uniform and complete coverage of the skin surface, with a concomitant improvement in evaporative loss and, as a result, transdermal transport. Aerosols
Figure 7. Permeation of testosterone from mechanical sprays containing different ratios of EtOH/PG/water (mean SD, n ¼ 4). DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
1700
LEICHTNAM ET AL.
Table 4. Physical Stability of Various Ethanol/PG Formulations of Testosterone in the Presence of HFA EtOH/PG (% v/v) 1:1 3:2 3:1
9:1
Figure 8. Permeation of testosterone from mechanical sprays containing different ratio of EtOH/PG (mean SD, n ¼ 4).
typically contain large amounts of propellant (50– 95%); given that the solubility of testosterone in the four binary mixtures of ethanol/PG varied between about 100 and 200 mg/mL (for the ethanol/PG ratios of 1:1, 3:2, 3:1, and 9:1, the saturation concentrations were 126, 176, 191, and 202 mg/mL, respectively), and that the valve capacity of the aerosol was 100 mL, the level of HFA used was set at either 50 or 70% to ensure that the theoretical amounts of drug delivered in a pulverization (6.3–10.1 mg for 50% HFA; 3.8– 6.1 mg for 70% HFA) were sufficient that skin permeation could be quantified. The next step was to visually examine the physical stability of the aerosol solutions. Although HFA is miscible with ethanol and PG,28 crystallization or phase separation occurred with all but one of the mixtures studied (Table 4). Only the 3:1 ethanol/PG mixture with 50% HFA was stable. Delivery of testosterone from this formulation as an aerosol was compared with those from a spray and a solution. Despite the wide difference in the quantity of drug deposited on the skin (0.84, 3.34, and 59.7 mg/cm2 from aerosol, spray and solution, respectively), the transport profiles were indistinguishable (Fig. 9). Crystallization on the skin was again observed for the spray and aerosol, whereas the solution (250 mL applied) did not display this problem. It seems that the soughtafter supersaturation effect was not achieved by the aerosol, therefore, even though a smaller JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
HFA
Observations
50% 70% 50% 70% 50% 70%
Immediate phase out Immediate phase out Phase out in 2–3 min Immediate phase out Stable Persistent trouble then phase out after 2 h Crystallization after about 30 min Crystallization after about 30 min
50% 70%
droplet size was possible. It is clearly necessary in further work to identify a means with which to stabilize the formulation as the ethanol evaporates (for example, by the inclusion of anti-nucleant polymers);29 also, given the fluxes achieved todate, identification of a penetration enhancer, which can boost testosterone delivery, is indicated. Two additional facts reinforce this final point: first, the animal skin model used here has a higher inherent permeability than the human barrier;30 second, although no evidence of testosterone metabolism during its percutaneous delivery was found (not surprising given that the skin had been frozen post-excision), skin biotransformation may occur in vivo, and may be expected to reduce the effective transdermal bioavailability of the hormone.
Figure 9. In vitro testosterone permeation from an ethanol/PG (3:1) saturated formulation delivered as either a spray (^), an aerosol ( ), or a solution (&). Mean SD, n ¼ 4. DOI 10.1002/jps
TESTOSTERONE TRANSDERMAL SPRAY
CONCLUSIONS Optimization of a formulation for transdermal delivery of a drug in the form of a spray is a complex process. While a carefully chosen experimental design using saturated solutions led to the identification of a 1:1:1 mixture of ethanol, PG and water as the most effective vehicle, the presence of water in a spray was found to be unnecessary. In this case, a simple 1:1 ethanol/ PG formulation performed best, and suggested the potential to achieve transient supersaturation. Aerosolization demanded the significant presence of HFA and pointed to the possible benefit of increasing the ethanol level at the expense of PG, both to realize a more practical formulation and to enhance the chances that supersaturation occurred. Unfortunately, this promise could not be demonstrated due to significant stability problems and crystallization of the drug; no significant improvement in testosterone’s steady-state flux beyond 2 mg/cm2 h could be achieved. Nevertheless, the study serves as an important starting point for further work that must address both the need for penetration enhancement and the stabilization of a supersaturated state post-aerosolization.
ACKNOWLEDGMENTS We thank the French National Association of Technical Research and the French Ministry of Research for financial support (Convention CIFRE no 675/2001).
REFERENCES 1. Chien YW. 1987. Advances in transdermal systemic medication. In: Chien YW, editor. Transdermal controlled systemic medications. New York: Marcel Dekker. pp 1–22. 2. Beckett AH. 1982. Possibilities and limitations of transdermal absorption. In: Brandau R, Lippold BH, editors. Dermal and transdermal absorption, Stuttgart: Wissenschaftliche Verlagsgesellschaft mbH. pp 154–170. 3. Mazer NA, Heiber WE, Moellmer JF, Meikle WA, Stringham JD, Sanders SW, Tollman KG, Odell WD. 1992. Enhanced transdermal delivery of testosterone: A new physiological approach for androgen replacement in hypogonadal men. J Control Release 19:347–362.
DOI 10.1002/jps
1701
4. Higuchi T, Hussain A. 1979. Device consisting of copolymer having acetoxy groups for delivery of drugs. US Patent 4,144,317. 5. Campell PS, Eckenhoff JB. 1987. Transdermal therapeutic system having improved delivery characteristic. US Patent 4,704,282. 6. Meikle AW. 1998. A permeation-enhanced nonscrotal testosterone transdermal system for the treatment of male hypogonadism. In: Nieschlag E, Behre HM, editors. Testosterone—action, deficiency, substitution, 2nd edition Berlin: SpringerVerlag. pp 389–422. 7. Schaison G, Couzinet B, Behre HM. 1998. Percutaneous dihydrotestosterone treatment. In: Nieschlag E, editor. Testosterone—action, deficiency, substitution, 2nd edition Berlin: Springer-Verlag. pp 329–348. 8. Laufen H, Leitold M. 1992. Biodisponibility and metabolism of isosorbide dinitrate from a transdermal spray. Arzneim Forsch 42:931–935. 9. Morgan TM, Parr RA, Reed BL, Finnin BC. 1998. Enhanced transdermal delivery of sex hormones in swine with a novel topical aerosol. J Pharm Sci 87: 1219–1225. 10. Morgan TM, Reed BL, Finnin BC. 1998. Enhanced skin permeation of sex hormones with novel topical spray vehicles. J Pharm Sci 87:1213– 1218. 11. Monash University. Dermal penetration enhancers and drug delivery systems involving same. Australia. Patent WO 97/29735. 12. Huisman R, Van Kamp HV, Weyland JW, Doornbos DA, Bolhuis GK, Lerk CF. 1984. Development and optimization of pharmaceutical formulations using a simplex lattice design. Pharm Weekbl Sci 6:185– 194. 13. Leichtnam ML, Rolland H, Wu¨thrich P. 2003. New aerosol transdermal drug delivery system—Effects of formulation parameters upon spray characteristics and permeation enhancement. Proc APGI, Symp Skin Formulation #43. 14. Leichtnam ML, Rolland H, Wu¨thrich P, Guy RH. 2005. Identification of penetration enhancers for testosterone transdermal delivery from spray formulations. J Controlled Release (in press). 15. Kim MK, Lee CH, Kim DD. 2000. Skin permeation of testosterone and its ester derivatives. J Pharm Pharmacol 52:369–375. 16. Hatanaka T, Shimoyama M, Sugibayashi K, Morimoto Y. 1993. Effect of vehicle on the skin permeability of drugs:polyethylene glycol 400water and ethanol-water binary solvents. J Control Release 23:247–260. 17. Berner B, Mazzenga GC, Otte JH, Steffens RJ, Juang R-H, Ebert CD. 1989. Ethanol:water mutually enhanced transdermal therapeutic system:II. Skin permeation of ethanol and nitroglycerin, J Pharm Sci 78:402–407.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
1702
LEICHTNAM ET AL.
18. Williams AC, Barry BW. 2004. Penetration enhancers. Adv Drug Del Rev 56:603–618. 19. Davis J, Hadgraft J. 1993. Supersaturated solutions as topical drug delivery systems. In: Walters K, Hadgraft J, editors. Pharmaceutical skin permeation enhancement, New York: Marcel Dekker. pp 243–267. 20. Higuchi T. 1960. Physical chemical analysis of percutaneous absorption process. J Soc Cosmetic Chemists 11:85–97. 21. Ostrenga J, Steinmetz C, Poulsen B. 1971. Significance of vehicle composition. I: relationship between topical vehicle composition, skin penetrability, and clinical efficacy. J Pharm Sci 60:1175– 1179. 22. Ostrenga J, Steinmetz C, Poulsen B, Yett S. 1971. Significance of vehicle composition. II: prediction of optimal vehicle composition. J Pharm Sci 60:1180– 1183. 23. Woodford R, Barry BW. 1982. Optimisation of bioavailability of topical steroids: Thermodynamic control. J Invest Dermatol 79:388–391. 24. Barry BW. 1983. Properties that influence percutaneous absorption. In: Barry BW, editor. Dermatological formulations, New York: Marcel Dekker. pp. 127–233.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 95, NO. 8, AUGUST 2006
25. Holegaard A, Møllgaard B. 1985. Dermal drug delivery-improvement by choice of vehicle or drug derivative. J Control Release 2:111–120. 26. Panchagnula R, Salve PS, Thomas NS, Jain AK, Ramarao P. 2001. Transdermal delivery of naloxone: Effect of water, propylene glycol, ethanol and their binary combinations on permeation through rat skin. Int J Pharm 219:95– 105. 27. Wilkins NF, Klose K, Morgan TM, Reed BL, Finnin BC. 2004. Metastable pharmaceutical compositions. Patent WO200400361. 28. Rowe R, Sheskey P, Weller P, editors. 1995. CDRom handbook of pharmaceutical excipients, 4th edition. London, UK: Science and Practice, Royal Pharmaceutical Society of Great Britain. 29. Fang J-Y, Kuo C-T, Huang Y-B, Wu P-C, Tsai Y-H. 1999. Transdermal delivery of sodium nonivamide acetate from volatile vehicles: Effects of polymers. Int J Pharm 176:156–167. 30. Sekkat N, Guy RH. 2001. Biological models to study skin permeation. In: Testa B, van der Waterbeemd H, Folkers G, Guy RH, editors. Pharmacokinetic optimization in drug research: Biological, physicochemical and computational strategies, Lausanne: Wiley-VCH. pp 155–172.
DOI 10.1002/jps