Percutaneous absorption of alkanoic acids II: Application of regular solution theory

The permeability coefficient, Kp, of pure unbranched alkanoic acids (C2-C7) applied to isolated porcine skin, reached a maximum in the solubility parameter (delta 2) range of 9.7-10 cal1 /2/cm3/2. When these and other penetrants were delivered from a solvent vehicle, the following linear relationships could be demonstrated: (a) between log Kp and the molar attraction constant of the penetrant [delta 2v2 or (-Ev)1/2] for six unbranched and six branched acids delivered from 1 M solution in n-heptane; (b) between Kp and the partial molal volume difference in n-heptane (-v2-v02) for the unbranched acids; and (c) between Kp and (-v2-v02) for propionic acid delivered from 1 M solutions in nine solvents having delta 1 values in the range 7.4-12.7 cal1 /2/cm3/2. Drug penetrability in a given series could be assessed from knowledge of the excess free energy of the penetrant in the delivery system used.     

Correlation of aqueous and lipid solubilities with flux for prodrugs of 5-fluorouracil, theophylline, and 6-mercaptopurine: A Potts-Guy approach

The Potts and Guy equation that has been used to predict permeability coefficients for molecules being delivered from aqueous vehicles has been transformed to accommodate lipid vehicles that are less polar than skin, and polar vehicles that are less polar than water. Solubilities in pH 4.0 aqueous buffer (SAQ), solubilities in isopropyl myristate (SIPM), and molecular weights (MW) of prodrugs of 5-fluorouracil (5-FU), theophylline (Th), and 6-mercaptopurine (6-MP) have been regressed against their fluxes from suspensions in IPM (JM). Seven series (n = 39) of alkylcarbonyloxymethyl (ACOM), alkyloxycarbonyl (AOC), alkylcarbonyl (AC), and alkylaminocarbonyl (AAC) prodrugs were used to determine the best fit to the transformed Potts and Guy equation (eq 6): log JM = x + y log SIPM + (1 - y)log SAQ - z MW. The estimated values for x, y, and z were -0. 193, +0.525, and +0.00364, respectively, with r 2 = 0.945 for n = 39. Inclusion of a miscellaneous series comprised of the parent drugs and a branched alkyl chain prodrug gave an equally good fit only if 6-MP was excluded from the analysis. The best performer (largest JM) in each series was usually correctly identified. The values for x, y, and z were consistent with values obtained by Potts and Guy, but the inclusion of the (l - y)log SAQ term in eq 6 and the value for y, shows that water solubility is almost as important as lipid solubility in predicting flux. There were no significant changes in predicted log JM or xi for each series if their log JM or xi were calculated using y and z coefficients obtained for solutions to eq 6 from which the data for the series had been excluded. This suggests that the data from all the series is homogeneous. Data from Kasting, Smith, and Cooper for SIPM, SPG, and MW of unrelated molecules were regressed against their fluxes from propylene glycol (PG) using eq 7: log JM = x + y log SIPM + (1 - y) log SPG - z MW. The estimated values for x, y, and z were -1.673, +0.599, and +0.00595, respectively, with r 2 = 0.852 for n = 28. These values for x, y, and z are also consistent with those previously reported by Potts and Guy, and, together with the results for fluxes from IPM, show the general utility of the transformed Potts and Guy equation in predicting flux from vehicles other than water and in showing the importance of solubility in a polar solvent as well as a nonpolar solvent in predicting flux.     

Effects of vehicles and prodrug properties and their interactions on the delivery of 6-mercaptopurine through skin: bisacyloxymethyl-6-mercaptopurine prodrugs

A series of S6,9-bisacyloxymethyl-6-mercaptopurine (6,9-bis-6-MP) prodrug derivatives was synthesized and characterized. The solubilities of the derivatives in solvents (vehicles), which exhibited a wide range of polarities from water to oleic acid, were measured. The abilities of the prodrugs to deliver 6-mercaptopurine (6-MP) from the vehicles have also been determined, and experimental fluxes and permeability coefficients (Kp) have been calculated for a large number of prodrug: vehicle combinations. Generally the best prodrugs of the series in terms of delivering 6-MP, regardless of the vehicle, were the first two members--the bisacetyl- and the bispropionyloxymethyl-6-mercaptopurine prodrugs. This result has been attributed mainly to the increased water solubility of these two prodrugs compared with that of 6-MP and the other prodrugs, since all of the prodrugs are much more lipid soluble than 6-MP. For three vehicles--isopropyl myristate, propylene glycol, and water--there was a good correlation between log experimental Kp for the delivery of 6-MP by the prodrugs from those vehicles and the theoretical solubility parameters of the prodrugs. The stabilities of the bisacetyl-(2), bisproprionyl-(3), and bisbutyryloxymethyl-6-mercaptopurine (4) derivatives were determined in buffer and in buffer containing enzymes leached from the dermis. Prodrug 2 was more stable than 3 or 4 in the buffer containing the enzymes, while 4 was more stable than 2 or 3 in the plain buffer.     

Percutaneous absorption of physostigmine: optimization of delivery from a binary solvent by thermodynamic control

A series of binary vehicles was used to deliver physostigmine across dermatomed human skin. The vehicles consisted of isopropyl myristate (IPM) and isopropyl alcohol (IPA) mixed in various volume fractions. The kinetics of penetration is conveniently considered as the sum total of two contributing effects: a "push" process resulting from the excess free energy (delta EG) of the penetrant in the donor vehicle, and a "pull" process resulting from the effect of IPA and IPM on the skin barrier. The inverse ratio of the solubility of the drug in a given vehicle to that in pure IPA was used to estimate the relative delta EG, hence the relative "push" effect. The solubility of physostigmine was highest in pure IPA (delta = 11.5), lowest in pure IPM (delta = 8.5), and intermediate in their various mixtures. But the permeability coefficient (Kp) of physostigmine was highest when delivered from a 1:9 (v/v) solution of IPA:IPM and a calculated delta y = 8.8. A further increase in the volume fraction of IPA caused an opposite decrease in the Kp values of physostigmine. The "steady-state" flux (Jss) of IPA from the same vehicle was lowest at a volume fraction of 1:9 and highest at one of 1:1 IPA:IPM. Thus, the maximal physostigmine penetration enhancing effect of IPA occurs at the lowest flux of IPA found in the present series. This indicates that the "pull" process ascribed to the presence of IPA in the barrier membrane is not important enough to outweigh the decrease in delta EG of physostigmine following an increase in the volume fraction of IPA in the donor vehicle, or that an excess of IPA in the barrier is not conducive to further enhancement of physostigmine diffusivity across the barrier. Optimized percutaneous delivery of physostigmine is possible by thermodynamic control of the penetration process.     

Viscosity of pharmacopeial multimolecular ointment vehicles and pharmaceutical availability of a model therapeutic agent

Viscosity was tested of basic ointment vehicles such as: white petrolatum, yellow petrolatum, anhydrous lanolin and eucerin produced by different manufacturers. Ointment vehicles of definite type differ significantly in rheological parameters. In the same group of products, the experimentally determined viscosity value of some vehicles is two-fold (petrolatum, anhydrous lanolin) or even three-fold (eucerin) higher than that of others. On the basis of rheological tests, using Einstein-Smoluchowski equation (D = kT/ 6pireta), theoretical coefficient was calculated of a model therapeutic agent--salicylic acid diffusion (-log chi2(i) = 1,22) from the tested vehicles to the external compartment. The obtained results were related to the performed in vitro measurements of the rate of salicylic acid release from the above mentioned ointment vehicles to model acceptor fluid. High correlation was observed between theoretical values of diffusion coefficients calculated on the basis of viscosity measurements and tested experimentally pharmaceutical availability of salicylic acid. It was confirmed by describing this dependence with regression equations of high correlation coefficients (r > or = 0,9667). Marked disproportions between rheological parameters of the vehicles of definite type produced by individual manufacturers are the cause of differences in pharmaceutical availability of therapeutic agents contained in these vehicles.     

Chromatographic indexes on immobilized artificial membranes for the prediction of transdermal transport of drugs

A set of 12 drugs, consisting of structurally unrelated neutral, basic, acidic and amphoteric compounds, was examined by high performance liquid chromatrography (HPLC) on a model of fluid membrane bilayers, the immobilized artificial membrane (IAM) column. The logarithms of chromatographic capacity factors extrapolated to 100% aqueous phase at pH 5.5 (log kw) were measured and compared to the n-octanol/water partition coefficients (log P). The scale derived from the IAM system was different from the lipophilicity scale expressed by the log P, due to the peculiar capability of phospholipids to well accommodate the ionized form of some molecules and show additive or repulsive extra-interactions when particular structural motifs on the molecule are present. The relationship between log P and log kw previously obtained for compounds interacting on IAM phase by a uniquely lipophilicity-based mechanism, allowed us to calculate, from log P, the values of log kw expected for the drugs considered. These values were subtracted from the log kw experimentally determined and the differences were assumed to quantify the amount of extra-interactions (hydrogen bond and electrostatic interactions) with phospholipids (delta log kw). The coefficients of permeability through the human skin (Kp) for the compounds considered did not correlate with either log kw or log P values. However, the Kp values correlated well with the delta log kw values indicating that the higher the ability of a molecule to cross the skin barrier, the lower its component of interaction with phospholipids not accounted for by lipophilicity-based interactions.     

Thermodynamics of cosolvent action: phenacetin, salicylic acid and probenecid

The solubility of phenacetin, salicylic acid, and probenecid in ethanol-water and ethanol-ethyl acetate mixtures at several temperatures (15-40 degrees C) was measured. The solubility profiles are related to medium polarity changes. The apparent thermodynamic magnitudes and enthalpy-entropy relationships are related to the cosolvent action. Salicylic acid and probenecid show a single peak against the solubility parameter delta(1) of both solvent mixtures, at 40% (delta(1) = 21.70 MPa(1/2)) and 30% (delta(1) = 20.91 MPa(1/2)) ethanol in ethyl acetate, respectively. Phenacetin displays two peaks at 60% ethanol in ethyl acetate (23.30 MPa(1/2)) and 90% ethanol in water (delta(1) = 28.64 MPa(1/2)). The apparent enthalpies of solution display a maximum at 30% (phenacetin and salicylic acid) and 40% (probenecid) ethanol in water, respectively. Two different mechanisms, entropy at low ethanol ratios, and enthalpy at high ethanol ratios control the solubility enhancement in the aqueous mixture. In the nonaqueous mixture (ethanol-ethyl acetate) enthalpy is the driving force throughout the whole solvent composition for salicylic acid and phenacetin. For probenecid, the dominant mechanism shifts from entropy to enthalpy as the ethanol in ethyl acetate concentration increases. The enthalpy-entropy compensation plots corroborate the different mechanisms involved in the solubility enhancement by cosolvents.     

The effect of vehicles and pressure sensitive adhesives on the percutaneous absorption of quercetin through the hairless mouse skin

To investigate the feasibility of developing a new quercetin transdermal system, a preformulation study was carried out. Therefore, the effects of vehicles and pressure-sensitive adhesives (PSA) on the in vitro permeation of quercetin across dorsal hairless mouse skin were studied. Among vehicles used, propylene glycol monocaprylate (PGMC) and propylene glycol monolaurate were found to have relatively high permeation flux from solution formulation (i.e., the permeation fluxes were 17.25 +/- 1.96 and 9.60 +/- 3.87 microg/cm2/h, respectively). The release rate from PSA formulations followed a matrix-controlled diffusion model and was mainly affected by the amount of PSA and drug loaded. The overall permeation fluxes from PSA formulations were less than 0.30 microg/cm2/h, which were significantly lower compared to those obtained from solution formulations. The lower permeation fluxes may be due to the decrease of solubility and diffusivity of quercetin in the PSA layer, considering the fact that the highest flux of 0.26 microg/cm2/h was obtained with the addition of 0.2% butylated hydroxyanisole in PGMC-diethylene glycol monoethyl ether co-solvents (80-85 : 15-20, v/v). Taken together, these observations indicate that improvement in the solubility and diffusivity of quercetin is necessary to realize fully the clinically applicable transdermal delivery system for the drug.     

Determination of partial solubility parameters of lactose by gas-solid chromatography

On the basis of the Snyder/Karger-Hansen interaction model, where delta EA = Vi(delta di delta dj + delta pi delta pi + delta hi delta nj), the partial solubility parameters of a solid used as the stationary phase may be determined through gas-solid chromatography by null-injection of solutes with known solubility parameters. Using n-decane, acetonitrile, and 1-propanol as molecular probes, the values found for unhydrated lactose were 9.6, 12.8, 11.3, and 19.5 (cal1/2/cm3/2) for delta d, delta p, delta h, and delta t, respectively; relative standard errors were better than 3%. The choice and the minimum number of the best molecular probes were determined by optimization of the experimental matrix according to the D-criterion, which permits considerable reduction of experimental time yet enhances total precision.     

Extended Hildebrand solubility approach: p-hydroxybenzoic acid in mixtures of dioxane and water

The extended Hildebrand solubility approach was used to reproduce the solubilities of p-hydroxybenzoic acid in a dioxane-water system. The solubility parameter of p-hydroxybenzoic acid was determined and found to be approximately 15 (cal/cm3)1/2. Residual plots (scattergrams) were used in conjunction with R2, F, and standard deviation values to determine whether a quadratic, cubic, quartic, or higher degree polynomial was required in the calculations. The earlier iteration method for back-calculations of solubilities was replaced by the more reliable root-finder method. The solubility profile of p-hydroxybenzoic acid in dioxane-water mixtures did not follow a log linear relationship even in the ranges where the solubility parameters of the water-cosolvent mixture might be expected to produce a straight-line function, as observed in other studies.     

Influence of membrane-solvent-solute interactions on solute permeation in model membranes

The interaction of the components of topical formulations with the skin is an important consideration for effective drug delivery and efficacy. The relative importance of solubility parameters and other solvent properties on membrane diffusion processes has not been fully elucidated in the literature. In this paper, the effect of different vehicles on the permeation of caffeine, salicylic acid and benzoic acid through silicone membranes was evaluated. Polydimethylsiloxane membranes were used as model membranes for comparing the release characteristics of saturated solutions of model permeants because of their homogeneity and uniformity. Log P (octanol-water partition coefficient) and solubility parameter values were calculated for the compounds under study. In vitro diffusion studies indicated that the permeation profiles of all solutes showed a similar pattern. The permeation rates of benzoic acid and salicylic acid through silicone membrane from saturated solutions were higher than those for caffeine reflecting the more lipophilic nature of these compounds in comparison with caffeine. Solvent uptake studies confirmed that the vehicles that were highly sorbed by the membrane altered its properties and hence the flux. Vehicles that were not sorbed by the membrane showed similar steady-state fluxes for the model drugs. This suggests that the diffusion process is mainly influenced by the interactions between the vehicles and the membrane. Solubility parameter alone cannot explain the interactions between the membrane and the vehicles in all cases. Rather, it is likely that membrane flux reflects a combination of different solvent and solute characteristics, such as size, shape and charge distribution.     

Alkylation activity and molecular properties of two antineoplastic agents that utilise indometacin and a conjugate of aspirin with 2-aminonicotinic acid to transport nitrogen mustard groups

BACKGROUND: Nitrogen mustard (N-mustard) compounds are considered important anticancer drugs. Various transporting agents have been utilised to carry N-mustard groups including coumarins, amides, polyaromatic molecules and cycloalkyl structures. N-mustards act as bifunctional alkylating agents that induce cross-linking within DNA strands and cytotoxic activity. Compounds that transport the N-mustard group in vivo can also express drug-likeness that can have advantages in clinical application. This study presents data on two anticancer drugs with N-mustard groups covalently attached to NSAIDs. METHODS: Two alkylating compounds were synthesised by covalently attaching a single N-mustard group to 2-2-acetoxybenzoylaminonicotinic acid (for compound I) and indometacin (for compound II). Molecular properties such as aqueous solubility, 1-octanol/water partitioning coefficient (log Kow), molar volume, polar surface area, 1-octanol/water partitioning at pH values other than human blood (log D) and the dermal permeability coefficient (Kp) were determined. The rate-order of reaction and rate constant of alkylation were determined by reacting compound I and compound II with a target compound having a primary amine group in buffered aqueous solution at blood pH 7.4 and 37 masculineC, and monitoring absorbance at 400nm. RESULTS AND C onclusion: Compounds I and II were stable at room temperature, soluble in water and effectively alkylated a nucleophilic primary amine target at physiological temperature and pH. The water solubility of compound I was considerably greater than that of compound II. Both compounds showed second-order rate order of alkylation and effectively alkylated a nucleophilic target under aqueous physiological conditions of pH 7.4 and 37 masculineC. Kp values for compounds I and II were determined to be 0.000786 cm/h and 0.024 cm/h, respectively. Both compound I and compound II had zero percent ionisation at pH 7.4, and compound I showed zero violations of the Rule of 5. Log P values for compounds I and II were 3.27 and 5.08, respectively. This study describes the benefits of antineoplastic agents with NSAID substituents that provide favourable pharmacological properties.     

Transport of drugs through human erythrocyte membrane. Structure-activity relationship of benzoic acid and its derivatives between membrane transport and partition coefficient]

The transport properties of benzoic acid and its eighteen derivatives such as o-, m- or p-hydroxybenzoic acid (o-, m- or p-HBA), aminobenzoic acid (o-, m- or p-ABA), toluic acid (o-, m- or p-TA), fluorobenzoic acid (o-, m- or p-FBA), chlorobenzoic acid (o-, m- or p-CBA) and bromobenzoic acid (o-, m- or p-BBA) through human erythrocyte membranes were examined. The drugs having a hydrophilic ortho-substituent and those having a hydrophobic meta- or para-substituent showed higher transport. The ratio of free drugs in erythrocytes, fuR, did not relate to the partition coefficient (P). However, fuM (the ratio of free drugs in the plasma) and Kp (the ratio of partition between erythrocytes and plasma) related to the P: fuM = -0.3128 x log P + 0.9727 (R2 = 0.8722***), Kp = -0.2558 x log P + 0.8642 (R2 = 0.8413***). In p-nitrophenol-glycosides, the fuM and the Kp that were predicted from these equations were compatible with the experimental results. It was suggested from these results that the fuM and the Kp may be predictable from the P.     

Percutaneous absorption of 5 glycol ethers through human skin in vitro.

Absorption across full thickness human skin was evaluated in vitro for five selected glycol ethers. Skin membranes were settled on static diffusion cells and both neat and 50% water diluted glycol ethers were applied on the donor chamber for 8 h. The amount of glycol ethers permeated into the receptor fluid was measured by gas chromatograph equipped with flame ionization detector (GC-FID). For neat solvents, permeation coefficient Kp ranged from 0.06 to 0.83 cm h(-1) 10(-3) respectively for DEGBEA and EGMEA while for 50% v/v diluted glycol ethers it varied from 0.08 to 1.81 cm h(-1) 10(-3) respectively for DPGME and EGMEA. These experiments show a statistically significant (Student's t-test, P <0.05) increase in permeation coefficients from neat to 50% water diluted glycol ethers and the same trend can be observed in fluxes and lag times. Only DPGME show an opposite behaviour. These results confirm the good ability of these solvents of permeating the skin and show that they could represent a risk for their potential dermal absorption both for workers and for occasional exposures, since the average lag time is 1.57 h.     

Transdermal delivery of antiparkinsonian agent, benztropine. I. Effect of vehicles on skin permeation

The influence of pH and various lipophilic and hydrophilic vehicles on the epidermal permeation of benztropine (BZ) free base and its mesylate salt were studied in vitro using the hairless mouse (HLM) and human cadaver (HC) skin membranes. The pH-partition behavior of BZ base (pK(a)=10) was examined using n-octanol and Britton-Robinson buffers over the pH range of 5-12. Unexpectedly, the ionized species of BZ yielded a high partition coefficient (log K(octanol/water)=2. 14), which was reflected by its relatively high skin permeability (P=1.6x10(-2)cm h(-1)). BZ base delivered from a lipophilic vehicle with a solubility parameter range of 7.1-10.3 (cal cm(3))(1/2) exhibited a significantly enhanced rate of permeation as compared to that attained from a hydrophilic vehicle of solubility parameter range between 12.5-23.4 (cal cm(3))(1/2). Among the neat solvents examined, a lipophilic carrier, isopropyl myristate (IPM) provided the most enhancing effect on the permeation of BZ base. In addition, the neat IPM carrier offered the maximum BZ base flux of 150 microg per cm(2) h(-1) across HC skin, which was approximately 16 times greater than the target delivery rate of BZ from a 10-cm(2) device. In comparison, BZ base exhibited a 2-60 times greater flux than BZ mesylate when delivered from the neat solvents. However, interestingly enough, the binary cosolvents consisting of IPM and short-chain alkanols such as ethanol (EtOH), isopropanol (iPrOH), and tertiary butanol (tBtOH), in particular a 2:8 combination, produced a marked synergistic enhancement of BZ flux from the mesylate salt, whereas a retarding effect was noticed for the permeation of BZ base. The enhancement potency for the BZ mesylate permeation increased linearly with the carbon number of the branched alcohols present in the binary mixtures. A tBtOH-IPM (2:8) combination produced the highest BZ flux from the mesylate salt, i.e. , 2016 mg per cm(2) h(-1), which was 100-fold greater than from water and 44-540-fold greater than the individual neat solvents, respectively. The observed permeation enhancement of BZ mesylate by the alkanol-IPM mixtures was probably as a result of a combination of decreasing barrier ability of the stratum corneum by the binary vehicles and moderately partitioning BZ mesylate through the viable epidermis/dermis.     

Molecular structure-based prediction of human abdominal skin permeability coefficients for several organic compounds

Mechanistically based predictions of skin permeability coefficients (Kp) derived solely on the basis of molecular structure information of organic compounds have not been reported previously. The objective of the present study was to predict the human abdominal Kp of structurally unrelated organic compounds using a mechanistic equation that takes into account compound-specific and species-specific determinants of Kp and relates the compound-specific determinants directly to molecular structure information. The method consisted of obtaining predictions based on a modified form of the conventional equation used to describe Kp (= P x D/L, where P, D, and L represent the stratum corneum:water partition coefficient, diffusion coefficient, and path length of diffusion). The value of P was predicted from the n-octanol:water partition coefficient (Po:w) computed from molecular structure information as well as the lipid, protein, and water contents in stratum corneum. The value of D was predicted according to the conventional Einstein-Stoke equation using the molar volume (Vmolar) of compounds computed from molecular structure information as well as data on viscosity of the stratum corneum components. Finally, the value of L was estimated in accordance with the characteristics of the diffusion pathways in the stratum corneum (i.e., transcellular and intercellular pathways). The average ratio of predicted Kp to the corresponding mean experimental values obtained from the literature was 1.25 (SD = 0.68, r = .95) for 47 structurally unrelated organic compounds (volatile organics and drugs), which were characterized by the presence of less than 2 hydrogen bonding groups and Vmolar < 231 cm3/mol. The present study is the first initiative that permits prediction of the human Kp of organic compounds by using molecular structure information as the sole chemical-specific input in a mechanistic equation.     

Hildebrand solubility parameter to predict drug release from hydroxypropyl methylcellulose gels

An equation including the Hildebrand solubility parameter δ of the drugs is used for the first time to model drug release from hydroxypropyl methylcellulose (HPMC) gels: l nM = -21.578 + 2.102 δ-0.037 δ(2)+0.48 ln t + 1.028 ln C(i) (r(2) = 0.94 for a total of 286 cases). The experimentally determined release data of six drugs having different polarity (caffeine, theophylline, paracetamol, salicylic acid, naproxen and diclofenac) at several initial concentrations C(i) were included in the equation. In general, the amount of drug delivered is linear at the first 5-6h of the release profiles and the zero order constants K(o) increase as the solubility parameter of the drugs become larger. The Peppas exponential law M/M(∞) = Kt(n) is applicable to larger fractional release, until 67-87% (48-51 h) for the less polar drugs (diclofenac and naproxen, lower δ values) and more than 80% (26-28 h) for the more polar drugs (higher δ values, theophylline, salicylic acid, caffeine and paracetamol). The Peppas release rate (lnK) shows a parabolic relationship with the drug solubility parameter. The diffusional exponent n varies between 0.40 and 0.58 indicating that drug release is mainly controlled by diffusion. An extended form of the Peppas equation is also tested for each drug including all the initial concentrations: lnM = a + b ln t + c ln C(i) (r(2) = 0.88-0.94). The logarithm of the octanol-water partition coefficients can also be used in combination with the drug concentrations.     

ATR-FTIR spectroscopic investigations on the effect of solvents on the permeation of benzoic acid and salicylic acid through silicone membranes

The effect of a series of alcohols on the permeation of salicylic acid (SA) and benzoic acid (BA) through silicone membrane was evaluated, using Franz-type diffusion cells. Although permeants were applied at the same thermodynamic activity in all vehicles, the resulting fluxes were found to differ significantly. This was a consequence of the interactions between the vehicles and the membrane. The interactions between the vehicles and the membrane were further investigated using ATR-FTIR spectroscopy. With this technique, it was possible to identify two different diffusion processes when the membrane was pre-treated with buffer, whereas one single diffusion process was observed when the membrane was pre-soaked with the vehicle. The technique was successfully used to deconvolute the relative magnitude of partition and diffusion in the permeation process. It was shown that the permeation of both acids was affected by the effect of the vehicles on the diffusion coefficient and the partition coefficient in the silicone membrane. The solubility of the drug in the impregnated membrane was found to be proportional to the saturated solubility in the vehicle used to treat the membrane. The solubility of BA in the impregnated silicone membrane was twice that of SA.