This study aimed at developing a controlled drug delivery system for recombinant human insulin-like growth factor-I (IGF-I) for localized delivery in bone healing. IGF-I was microencapsulated into an end-group uncapped 14 kDa poly(D,L-lactide-co-glycolide) 50:50 (PLGA 50:50) by solvent extraction from a W(1)/O/W(2) dispersion. Prior to encapsulation, IGF-I was exposed to ultrasonication in a water/dichloromethane dispersion, and its stability tested in the presence and absence of various excipients in the W(1) phase. HPLC and RIA were used for the assessment of IGF-I stability. Microencapsulated IGF-I was tested again for its structural intactness and also for in vitro release from various formulations containing appropriate co-encapsulated excipients. A specific fat cell assay was used to determine the biological activity of released IGF-I. Moderate ultrasonic treatment of aqueous IGF-I/dichloromethane mixtures caused approx. 50% IGF-I degradation. However, IGF-I was fully protected when bovine serum albumin, succinylated gelatin or poly(ethyleneglycol) were added to the aqueous IGF-I. Co-encapsulation of these excipients protected efficiently the protein upon microencapsulation. IGF-I release from microsphere formulations was sustained for up to 13 days featuring a moderately pulsatile pattern, depending on the microsphere composition. Typically, the amounts of IGF-I released within the first 24 h (burst) and during the second release pulse were in the order of 20 and 40%, respectively, of the total dose. The biological activity of released IGF-I was confirmed at selected time-points by the fat cell assay. In conclusion, the developed microspheres proved to be suitable to release biologically intact IGF-I over up to 13 days, a time-period considered to be relevant to promote bone fracture healing.