In this study, a reevaluation of the in vivo release phases from long-release PLGA-based microspheres is presented, leading to a better characterization of the plasma concentrations/time profile. Microspheres were designed for intramuscular injection releasing a cyclic somatostatin analog over 70 days. Clinical study was performed in 64 healthy subjects receiving a subcutaneous dose of an immediate release solution as reference formulation and an intramuscular injection of microspheres as test formulation. The in vivo input curve was obtained by numerical deconvolution. Results showed that double Weibull function could not fit correctly the tri-phasic (burst, lag, and erosion) in vivo input profile typical for PLGA-based formulations, due to a change in the drug release trend in the terminal phase. Triple Weibull showed a significant improvement in the curve fitting, each term being assigned to one of the following phases: initial (burst/lag), erosion, and terminal phase of drug release. The existence of the additional terminal phase was confirmed by a mechanistic approach as well, which denoted that this phase was, most probably, a consequence of the release mechanism change from erosion to diffusion controlled. The same model demonstrated that the burst release was as well influenced by the polymer swelling, while currently existing theories state that the burst phase is mainly determined by the dissolution of immediately available drug substance and diffusion through surface related pores.