INFLUENCE OF PHYSICOCHEMICAL PROPERTIES OF MODEL COMPOUNDS ON THEIR RELEASE FROM BIODEGRADABLE POLYANHYDRIDE DEVICES

A new class of fatty acid dimer-based polyanhydride, poly(fatty acid d imer-sebacic acidi (P(FAD-SA)), was reported to have desired physicoch emical and mechanical properties for a biodegradable device and believ ed to undergo pure surface erosion. The objectives of this study were to determine if P(FAD SA) undergoes pure surface erosion, and to ident ify the factors governing the release characteristics and mechanism fr om polyanhydride devices at various pHs. Three model compounds, mannit ol (small M-w, highly water soluble), inulin (large M-w, moderately wa ter soluble) and stearic acid (medium M,, extremely lipophilic) were i ncorporated at 10% w/w in disk-shaped :P(FAD-SA, 50:50 w/w) devices an d the release of the compounds studied from pH 1-9. All the three mode l compounds were released faster at alkaline pH than at acidic pH (pH 9 > pH 7.4 > pH 1-5), suggesting that the release of model compounds w as dependent on the base catalyzed erosion of polyanhydride. However, erosion was not the sole release rate controlling mechanism for these model compounds. Highly water soluble mannitol and inulin were release d rapidly at acidic pH, when erosion of polyanhydride is negligible. M annitol released faster than inulin, which was released faster than li pophilic stearic acid at all pH, suggesting that the hydrophilic/hydro phobic nature of the loaded compound influenced its release significan tly. The observed release profiles were significantly higher, and did not match the release profiles predicted from erosion rates of the man nitol and inulin devices, however they were closer for stearic acid de vices. Although inulin was released rapidly in the initial phase (40-5 0%), only 60-70% inulin was released over a 6-week period compared to 100% release of mannitol. The large molecular size of inulin may have hindered its diffusion through water-filled pores and channels observe d in the devices during release, which may have resulted in the majori ty of inulin being still trapped in the devices. Although various cont ributing factors in overall release were identified, a single release kinetic model could not explain the release profiles of all the loaded compounds under the different pH conditions. The complete release pro files could be described by first order kinetics, however the initial release profiles of all the model compounds were described very well b y zero order kinetics. The release rates of mannitol and inulin were i nfluenced by both the erosion rates and the intrinsic dissolution rate s unlike release rates of stearic acid which correlated closely with e rosion rates. This may be due to the fact that different release mecha nisms, such as diffusion, dissolution and erosion all play a significa nt role in overall release of compound from this newly developed P(FAD -SA) device. However, the contribution of each process to overall rele ase may vary as a function of pH, the nature of the compound and polym er erosion. In conclusion, the results suggest that P(FAD-SA, 50:50 w/ w) is undergoing bulk erosion rather than the desired surface erosion. If it did undergo surface erosion, release characteristics would be i ndependent of the physicochemical nature of the incorporated compound and release kinetics would be zero order from disk-shaped devices irre spective of the nature of the loaded compound. (C) 1997 Elsevier Scien ce B.V.