66 JOURNAL OF COSMETIC SCIENCE numerous advantages. Acrylic, cellulosic, block copolymers and, more recently, polyes­ ters are among the polymers that received most of the attention from investigators ( 1-S). In semisolid preparations, these polymers usually increase the viscosity of the system (2,6) without effect on the rate of delivery of actives (7). Other polymeric systems studied used polymer microparticles that contained the drug and were capable of re­ leasing it over an extended period of time. Won (8) introduced porous solid microspheres into which the drug could be incorporated. Mathiowitz et al. (9) presented non­ bioerodable and erodable microspheres that were capable of reducing the release rate of actives. Additional patents reference incorporation of cationic polymers (acrylates) and skin-depositing polyurethanes. These polymers are reported to deposit the active drug salicylic acid onto the skin surface or to target penetration into the epidermis from cleansers and emulsions (10,11). The use of the polyurethane polymer polyethylene glycol-8/SMDI copolymer (polyol­ prepolymer-15) in controlling the delivery of salicylic acid and lactic acid from topical preparations was recently studied by Fares and Zatz (12). The effect of this polyurethane polymer, polyolprepolymer-15, on permeation was measured in vitro using flow-through diffusion cells and dermatomed pig skin. Skin uptake was also evaluated over time using tape-stripping and tissue analysis. The polymer decreased the flux of salicylic acid through pig skin but did not affect the delivery of lactic acid. The polymer increased the overall deposition of salicylic acid in the stratum corneum but did not change the levels of salicylic acid in the viable skin significantly. Skin uptake of lactic acid was not affected by the presence of the polymer. Based on dialysis and cloud point measurements, it was found that polyolprepolymer-15 reduced the activity of salicylic acid in the vehicle via binding, leading to a decrease in permeation. The binding mechanism accounts for the effect of polyolprepolymer-15 on the solutes investigated salicylic acid was found to bind to the polymer but lactic acid did not. Because of binding, the thermodynamic activity of salicylic acid is reduced in the presence of the polymer and the stratum corneum/vehicle partition coefficient is reduced. As a consequence, the transfer rate into stratum corneum is lower than for a control system without polymer, which also results in a lower rate of passage through the skin. Another strategy frequently used to control delivery of active compounds to skin is entrapment in surfactant micelles. For topical treatment products this strategy can become problematic since surfactants are used for cleansing but are generally too irri­ tating to be in contact with skin for an extended period of time. Several patents have recently been issued that show the use of surfactant complexes to control delivery of actives (13,14) while maintaining the gentleness of formulations. This technology refers to the formulation of surfactant complexes that produce milder formulations and en­ hance deposition of salicylic acid onto the skin surface layers. In the first technology (13), an anionic-amine oxide complex was carefully preformed to end up with a charge density of zero, meaning complexation is complete. Thus the system no longer is comprised of anionics or zwitterionics rather, the complex is "pseudononionic." The irritation potential of the pseudononionic will likely be as mild as typical nonionics. The critical micelle concentration would be very low, thus producing a large micelle reservoir to trap the drug. The irritating anionic moiety is tied up in the complex. A similar technology strategy incorporates complexes of anionics with traditional cationic surfac­ tants (14). The result is also a pseudononionic complex with similar consequences, providing complexation is complete.

TARGETED DELIVERY OF SALICYLIC ACID 67 Polyolprepolymer-15®, is a hydroxyl-terminated block copolymer of 1,1"-methylene­ bis-[ 4,isocyanatocyclohexane} and 8 moles of ethylene oxide, which makes the polymer soluble in water. The average molecular weight of the polymer is 1,800. This paper reports on the effect of this polymer on the delivery of salicylic acid into skin from hydroalcoholic solutions typically used in acne treatment formulations. It also studies the effect of other components of the formulation, namely surfactants and salts, on salicylic acid delivery and the interactions between these components to control drug delivery and formulation mildness. The relationship of controlled drug delivery to the irritancy of these hydroalcoholic solutions will be investigated. Studies in this report were thus done to extend our understanding of the relationship of the controlled delivery phenomena of the polyurethane polymer and surfactant type and solution behavior to the irritation potential of salicylic acid hydroalcoholic solutions. MATERIALS Materials include salicylic acid (SA), phosphate-buffered saline (called PBS, Sigma, St. Louis, MO), scintillation fluid, glacial acetic acid (Fisher Scientific, Fair Lawn, NJ), polyethylene glycol-8/SMDI copolymer (polyolprepolymer-15, Bertek, Inc., Foster City, CA), isoceteth-20 (ICI, Wilmington, DE), [1 4 C}SA-56.l mCi/mmol (NEN Products, Boston), and skin-digesting fluid (Solvable, Packard Instrument Company, Inc., Mer­ iden, CT). METHODS Briefly, the in vitro penetration method used is to place human cadaver skin in a diffusion chamber, apply the drug on top, and measure how much drug goes into the receptor, a buffered receptor solution, at various time points. The skin is separated into its various layers (i.e., epidermis, dermis, and receptor fluid) and the drug content is measured in the various layers. Preparation of the skin. Fresh, excised, human skin was obtained from cadavers. Upon receipt, the skin was washed gently with 1 % (v/v) aqueous dishwashing liquid, rinsed with distilled water, and patted dry with a paper towel. A 250-300-µm-thick layer of the skin was prepared with a Padgett Electrodermatome (Padgett Dermatome, Division of Kansas City Assemblage Co., Kansas City, MO). The dermatomed skin was refrig­ erated until used. Two hours before each experiment, the skin was placed at room temperature to equilibrate. Circular pieces of the dermatomed skin (about 12 mm in diameter) were cut with a brass punch and placed epidermis-side up on the diffusion cells. Penetration method. The skin discs, 12-mm in diameter, were mounted on flow-through diffusion cells according to Bronaugh (15 ). The diffusion cells (Bronaugh design, Crown Glass Co.) were clamped, and the receptor fluid, phosphate-buffered saline (PBS) con­ taining 1.5% Oleth and 0.01 % sodium azide, was pumped through, as per Bronaugh (15). Unless otherwise indicated, a clinically relevant dose (5 mg/cm2 ) of a sample of the formula was dispensed and spread evenly on a 0.64-cm2 area of the skin surface using a glass rod or micropipette. The cells' temperature was maintained at 3 7 ° C throughout the experiment using a water bath/circulator (Haake, Paramus, NJ). Fraction collection