Modulated drug release using iontophoresis through heterogeneous cation-exchange membranes. 2. Influence of cation-exchanger content on membrane resistance and characteristic times.

An implantable drug delivery method using iontophoresis through cation-selective membranes was further developed. Heterogeneous cation-exchange membranes (HCMs) were prepared by mixing conductive sulfonated polystyrene beads into a nonconductive silicone rubber matrix. The membrane resistivity and lag time to steady-state transport of two salts, (+/-)-phenylpropanolamine hydrochloride (PPA) and NaCl, were evaluated during constant current iontophoresis at 37 degrees C as a function of the resin content in the HCMs. A continuous decline in membrane resistivity was observed as fractional resin content (l) was increased over the entire usable region (l = 0.29-0.52), a characteristic that could be described by a percolation scaling law (for an infinite lattice, 3-D geometry). Morphological analysis of the membranes before and after swelling strongly suggested that the conducting clusters of resin beads form during the swelling period prior to use. The response time to steady-state transport of PPA into NaCl during a 40 microA constant current (0.27 cm2) was found to increase with increasing l, but not without decreasing the permselectivity of the HCMs for the drug cation. The lag time effect could be explained in terms of an increasing number of fixed charge groups in the membrane available for transport (mfcA), which was derived from a macroscopic mass balance model. The values of mfcA were also found to be related to the characteristic time of diffusion in a homogeneous transport projection of the HCM (or an effective medium), an essential parameter for future non-steady-state simulations. The characteristic time of diffusion was found to be invariant with changing resin content, suggesting that the membranes are fairly nontortuous (ca. seven beads thick). By assuming that the thickness of the HCM approaches the thickness of its homogeneous projection, an expression was derived to predict lag time to steady-state PPA transport requiring resistance measurements only (provided that the resin capacity is known). There was excellent agreement between the theoretical and experimental lag time to steady-state transport of PPA (r = 0.96, p < 0.001), further implicating the role of membrane resistance in the bi-ionic system. These modeling approaches have already found utility in iontophoretic implant design for prevention of cardiac arrhythmias and may be valuable in future non-steady-state analysis to further develop on-line detection-implant response technology.