This work investigated different approaches to modify the release from PLGA matrices. The drug solid state in PLGA was analyzed with respect to its physical stability and impact on in-vitro drug release. Furthermore, the effect of drug particle size and the addition of a cationic excipient on the in-vitro drug release from PLGA was examined. First, the interactions between PLGA and the drugs ibuprofen, risperidone and dexamethasone were characterized. By assessing the solubility of the drugs in PLGA polymers of different grade, dexamethasone showed a lower solubility than risperidone and with up to 25 % ibuprofen exhibited the highest solubility. The examination of Hansen solubility parameters (HSPs) confirmed that the extend of interactions with PLGA was crucial for the solubility differences. The HSPs of ibuprofen and PLGA were more similar than those of dexamethasone and PLGA, leading to weaker interactions. The similarity of the risperidone HSPs to PLGA ranged in between dexamethasone and ibuprofen, which was in good accordance to the results from solubility testing. The similarity was assessed by calculating the HSP distance, Ra. The solid state stability of solid solutions and solid dispersions of the respective drugs in PLGA was characterized at ambient conditions, at 40 °C and 75 % RH and during in-vitro drug release. Films with solid solutions of all three drugs were stored under ambient conditions and dexamethasone and risperidone films remained stable over 3 month, while ibuprofen crystallized within 14 days. Polarized light microscopy and DSC suggested that ibuprofen formed a supersaturated solution in PLGA, leading to a time dependent crystallization. Storage of the films at 40 °C and 75 % RH gave a deeper insight into the crystallization mechanism of ibuprofen, denoting water uptake and the subsequent depression of glass transition temeprature as the main crystallization triggers. It is assumed that a pH decrease in the matrix contributed as well to the crystallization. The drug concentration in the matrix and polymer characteristics were able to control the crystallization. A higher level of supersaturation led to a faster crystallization, while a slower water uptake and a higher polymer viscosity slowed down the crystallization, because the nucleation of a drug crystal depended on the diffusion rate of drug molecules. A solid solution of risperidone remained stable at 40 °C and 75 % RH, but a risperidone dispersion started to dissolve into the polymer during incubation, due to a significant decrease of pH and the plasticizing effect of water. The opposite case than ibuprofen was observed here: A faster water uptake, decreasing the glass transition temperature and thus the polymer viscosity, led to a faster dissolution, since this process was likewise diffusion dependent. A higher drug loading led to a longer dissolution time, while polymer grades with a lower polymer viscosity and faster water uptake asset led to a faster dissolution. Additionally, the availability of carboxylic groups in the polymer influenced the dissolution rate, as it was shown that this process depended on the decrease in matrix pH. Dexamethasone solid solutions and dispersions did not show any solid state change during incubation at 40 °C and 75 % RH. With the help of HSPs the change in interaction between PLGA and drugs during incubation and associated water uptake was assessed and the calculated changes in Ra supported the observed drug behavior in PLGA. The in-vitro release studies showed that crystallization decelerated ibuprofen release, when the crystallization rate exceeded drug release rate. The observation of solid solutions at different ibuprofen loadings revealed multiple aspects: A high level of supersaturation led to a crystallization and thus a slower release. Likewise, drug loadings below the saturation solubility slowed down the terminal release phase from PLGA. Since the drug showed a higher solubility in the polymer than in the aqueous medium (factor 50), the drug only released by partitioning between PLGA and medium, leading to a slow terminal release rate. Risperidone solid dispersions were pre-incubated at 40 °C and 75 % RH to obtain solid solutions and their release was compared to untreated solid dispersions. Polarized light microscopic observations of the solid dispersion during release confirmed that the drug first dissolved into the PLGA film and released from the dissolved state into the medium. "Skipping" this dissolution (by pre-incubation) led to a faster release. Transferring these insights from PLGA films to implants supported the former findings. Solid solutions and solid dispersions of ibuprofen in PLGA were prepared by hot melt extrusion, controlling the solid state by extrusion temperature. The crystallization of ibuprofen during incubation at 40 °C and 75 % RH and in-vitro release was proven and quantified per DSC for different PLGA grades. The release studies confirmed that a fast crystallization decelerated the release to the level of a solid dispersion. A PLGA grade (503 H) with low water uptake and high polymer viscosity was able to stabilize the supersaturated solution and led to a faster release than from a solid dispersion of equal composition. Solid dispersions of risperidone were extruded directly, while solid solutions of risperidone were prepared by casting the respective composition and extruding dried film. The dissolution of dispersed risperidone into the polymer at 40 °C and 75 % RH was confirmed in implants and consequently the release from solid solution was faster than from a solid dispersion. In a second part, the effect of drug particle size was investigated with risperidone and dexamethasone, exhibiting different release characteristics. The dexamethasone particle size showed only slight effects on in-vitro release from films and implants. The polymer absorbed only low amounts of water in presence of dexamethasone. The low water amount limited the release, as the drug required water to dissolve inside the matrix before being released. The initial burst decreased with decreasing particle size, since the drug was better distributed in the PLGA matrix. In films with 20 % drug loading and implants with 50 % drug loading, a smaller particle size led to a faster terminal release phase due to a faster dissolution inside the implant. In risperidone loaded films the discovered release mechanism was confirmed by solid dispersions with different particle sizes: the drug dissolved into the polymer prior to release. At 50 % drug loading, a smaller risperidone particle size led to a faster release, since the drug dissolved faster into the PLGA matrix. However, this effect was not observed at a lower drug loading (20 %), since the dissolution into the polymer occurred very rapid, showing only minor differences among the particle sizes. Overall, the dexamethasone and risperidone release from PLGA was not very susceptible to particle size changes. Only high drug loadings revealed, that despite their different release mechanisms, the drug release profiles could be influenced by the drug particle size. Finally, the interactions between the cationic lipid, stearylamine, PLGA and the ionizable drugs ibuprofen and risperidone were investigated. A conductometric titration of stearylamine and lactic acid suggested ionic interactions between the amine groups of the cationic lipid and the carboxylic groups of the PLGA degradation product. Polarized light micrographs of casted and molten stearylamine-PLGA mixtures showed a homogeneous distribution of stearylamine in the polymer and DSC thermograms further supported an interaction potential between both components. The characterization of degrading PLGA-stearylamine implants without drug addition showed, that stearylamine retarded the degradation and erosion of PLGA, which was attributed to their ionic interactions. Implants with risperidone and ibuprofen loading and stearylamine addition were hot melt extruded. Hot stage polarized light microscopy revealed, that implants with risperidone were present as solid dispersion, while ibuprofen was molecularly dispersed in the PLGA stearylamine matrix. When stearylamine was added to risperidone-PLGA implants, the release was faster than without stearylamine, since the basic drug exhibited a cationic charge at pH 7.4, which led to electrostatic repulsion by the cationic lipid. The acidic drug ibuprofen showed the opposite behavior: its anionic charge at pH 7.4 interacted with stearylamine, leading to a delayed release compared to implants without stearylamine addition. Negative charges of ibuprofen compete with the carboxylic groups of degrading PLGA for the interaction with amine groups of stearylamine. With increasing amount of carboxylic groups, ibuprofen is displaced from the interaction an can be released. The interactions between ibuprofen and stearylamine were confirmed by the DSC thermogram of their physical mixture. In conclusion, the drug release profile of PLGA matrices was modified by a change in drug solid state or particle size and by addition of a cationic lipid. However, the effect of particle size was small compared to the other two approaches. In turn, controlling the solid state of a drug in PLGA is more elaborate and the physical stability of the formulation may be greater by varying only the drug particle size. The characterization of drug-PLGA interactions provided a deeper insight into the formation of solid solutions with PLGA and the release mechanisms from drug loaded PLGA matrices. The addition of a cationic excipients showed great impact on the release profile. Nevertheless, the data base for biocompatability and biodegradability of stearylamine is rather weak and thus investigation should be extended to other cationic excipients than stearylamine.
