OBJECTIVE: To test the therapeutic effect of periocular corticosteroid injection on outcome, vitreous cellular infiltrate, and vitreous soluble growth-stimulating activity in an experimental model of tractional retinal detachment resulting from proliferative vitreoretinopathy. METHODS: An experimental model of proliferative vitreoretinopathy was induced in rabbits, which then were selected randomly to receive either periocular methylprednisolone acetate injection treatment or a placebo injection (control). Animals were examined and monitored to determine the degree of retinal detachment and proliferative vitreoretinopathy at weekly intervals for 4 weeks. Vitreous specimens were obtained each week and analyzed for cell number by flow cytometry and for soluble growth-stimulating activity using a bioassay. RESULTS: Eighty-six percent (12 of 14) of the rabbits receiving periocular steroid injection had retinas that remained attached, whereas only 13 percent (2 of 15) of control animals had retinas that remained attached. Periocular steroid injections were associated with a significant decrease in the vitreous soluble proliferation stimulating activity at all times during experimental proliferative vitreoretinopathy. Vitreous cellular infiltration also was reduced significantly in steroid-treated animals compared with that of control animals. Finally, the magnitude of the soluble growth-stimulating activity at day 7, before onset of proliferative vitreoretinopathy, was highly predictive of outcome 3 weeks later. CONCLUSION: Periocular steroids decreased the incidence of complicated retinal detachment caused by proliferative vitreoretinopathy in this rabbit model. Decreased vitreous cellular infiltrate and soluble proliferation stimulating activity within the vitreous microenvironment also was observed to be associated with successful outcome.