624 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Higher molecular weight materials also show variable penetration. Very large molecules such as proteins and polysaccharides go through very poorly, if at all. The use of DMSO as a penetration carrier indi- cated that substances which have molecular weights of 3000 or more cannot be transported into the skin (76). There are few published re- ports in which permeability coefficients are correlated with the size of penetrating molecules. Virtually nothing is known of the effect of molecular shape. Miscellaneous Physical Factors Other drug-vehicle factors of importance are the particle size of poorly suspended drugs, the viscosity of the vehicle, and the drug con- centration in the vehicle. A reduction in the particle size of fluocino- lone acetonide has been shown to enhance its penetration (77, 78). Where a drug exists in more than one crystalline form, the one with the highest thermodynamic activity would be expected to penetrate most rapidly, provided it is stable (11, 53). VEHICLES Factors in Release o[ Penetrant The literature on the influence of vehicles on skin penetration is con- fusing and often contradictory. The variety of experimental animals and methods of estimating penetration renders correlation of results tenuous. A lack of awareness of possible drug-vehicle interactions and of the func- tions of different vehicles (18, 79, 80), as well as a lack of consideration of the thermodynamics involved in the interpretation of results (11, 18, 53, 81), has added to the difficulties. In general, the emphasis in de- veloping vehicles has generally been placed on the compatibility, stability, and appearance of the product, rather than on the influence which the components in the vehicle had on enhancing or hindering the movement of the drug through the skin (82). Rothman (83) reviewed the literature on vehicles up to 1954, and Barr (84) to 1962. More recent reviews are those of Malkinson (1, 2), Vickers (21), Barrett et al. (79, 80), Busse et al. (62), Munro (85), and Sarkany and Hadgraft (86). Physiological availability of a topically applied drug depends on both the rate of release from the vehicle and the permeability through the skin. The drug, incorporated in the vehicle, should reach the skin sur- face at an adequate rate and in sufficient amounts. Earlier belief held that the primary factor influencing penetration through the skin was the

SKIN PENETRATION 625 vehicle itself. The bulk of evidence now indicates that unless an applied material is capable of passage either through the skin barrier or follicles the vehicle is of subsidiary importance. However, experimental and clinical evidence is now appearing which points the way to vehicles which may materially affect skin penetration (54, 56, 60, 62, 69-71, 87, 88). The previous physico-chemical discussion of solubility and partition coefficients, as well as in vivo (54) and in vitro (25) studies, has supported the postulate that the release of a substance will be favored by the selection of vehicles having a low affinity for the penetrant, or in which the drug is least soluble. This is consistent with the view that the rate of release is governed by the vehicle to receptor phase (stratum comeurn) partition coefficient. For a given concentration of drug in certain vehicles, the activity coefficient of the drug, and consequently the thermodynamic activity of the drug in the vehicle at that concentration, may vary by a factor as much as a thousandfold from one vehicle to the next. Solutes held firmly by the vehicle, such as when the drug forms a soluble complex with the vehicle, exhibit low activity coefficients hence, the rate of release from such drug-vehicle combinations will be slow. Solutes held "loosely" by the vehicle (less affinity of the vehicle for the drug or solute) exhibit high activity coefficients therefore, the rate of release from such drug- vehicle combinations will be faster (11). This was shown by Blank and Scheuplein (24, 25) in a study of the penetration of polar ethanol and nonpolar heptanol t/tom water and lipid solvents. Ethanol penetrates better from oils than from water, but the reverse is true for heptanol. The polar alcohol tends to stay in the polar vehicle and not be transferred to the skin, but it is transferred from oily vehicles the reverse occurs for the nonpolar alcohol. Dempski et al. (71) noted that dexamethasone was released more rapidly from gelled isopropyl myristate than t/tom petrolatum and two other nonaqueous vehicles. Their studies demonstrated that the in vitro release of a medicinal agent is a function of the degree of solubility of that agent in both the base and its surrounding media. The medicinal must be sufficiently soluble in a nonaqueous base to allow for its release into an aqueous medium but not so soluble to preferentially remain in that base. If the drug is insoluble in its vehicle, it appears that only the drug particles available at the surface of the vehicle will dissolve into an aqueous medium. If the drug is partly soluble in the vehicle, it seems to dissolve and diffuse throughout the medium as it dissolves