PREDICTING INJECTION SITE MUSCLE DAMAGE .1. EVALUATION OF IMMEDIATE-RELEASE PARENTERAL FORMULATIONS IN ANIMAL-MODELS

Purpose. The current animal model generally accepted by the pharmaceut ical industry and the FDA for assessment of muscle damage following in tramuscular injection (IM) is the rabbit lesion volume model (RbLV). H owever, this model is resource intensive. The goal of this study was t o find a resource sparing alternative to the rabbit lesion model for a ssessing injection site toleration in IM formulation screening. Method s. Short term animal model alternatives to RbLV for evaluating IM form ulations were examined. In addition to RbLV, myeloperoxidase (MPO), p- nitrophenyl N-acetyl-beta-glucosaminide (NA beta G) and/or plasma crea tine phosphokinase (CK) activities were determined in rabbits (Rb) and rats (Rt) after injection of formulations (digoxin, azithromycin and danofloxacin). The edema from these formulations 24 hr after subcutane ous injection into the rat footpad (RFE) was also determined. Results. MPO and NA beta G were not considered very useful as biochemical pred ictors of muscle damage for these formulations. Histology generally co rrelated with RbLV values. Compared to saline, RbLV was marked for all formulations within 1-3 days of injection. After day 3, lesions quick ly resolved, and no significant differences were found. For these form ulations, all CR animal models and RFE were generally predictive of Rb LV. A formulation with RtCK > 1000 U/L or RbCK > 3000 U/L, was predict ed to be poorly, tolerated. Conclusions. Due to ease, number of animal s, time and intrinsic mechanism, we concluded that for most formulatio ns, 2 and 4 hr RtCK data alone should be reasonably predictive of musc le damage.