Polymer degradation induced drug precipitation in PLGA implants - Why less is sometimes more

Nifedipine and nicardipine loaded PLGA extrudates have a great potential to prevent cerebral vasospasms after subarachnoid hemorrhage or surgical clipping of aneurysm. A constant release over approx. two weeks is desired. Although in vivo studies on humans have been reported, there is limited knowledge about the release kinetics and the underlying mechanisms. Therefore, nifedipine and nicardipine loaded PLGA implants with different drug loads were manufactured by extrusion and investigated. In addition to the measurements of the release kinetics, GPC, DSC, X-ray diffraction and light microscopic investigations were performed for a detailed characterization. The water uptake and polymer erosion studies showed an initial lag phase of 5-7 days and an acceleration of both processes thereafter. With 5% loaded implants a higher drug release compared to 10% drug loaded polymers could be achieved and not only the relative amount of drug release (% of loaded drug), but surprisingly also the absolute amount of the released drug increased. The drugs were initially in an amorphous state. For nifedipine, formation of drug crystals with time has been observed by light microscopy and X-ray diffraction. The analysis of the drug content in the degrading polymer showed a very large increase from 10% to about 20% (nifedipine) and over 50% (nicardipine). In contrast, no or only a moderate increase of the drug content occurred for initially 5% loaded polymer implants. We postulate that water penetration and polymer degradation induced changes of the microenvironment lead to supersaturated systems. A supersaturated state is faster reached for polymers with higher drug load and therefore, drug precipitation takes place at earlier time points. As a result, drug release might be incomplete for poorly soluble drugs and paradoxically, the total amount of drug release might be higher for systems with a lower drug load. Drug release is initially controlled by the PLGA matrix, but later by the dissolution kinetics of the precipitated drug which are very slow for poorly soluble drugs according to the Noyes-Whitney equation.