Extravasation of vesicant antineoplastic agents such as doxorubicin into the skin or subcutaneous tissues may result in loss of the full thickness of the skin or underlying structures. Several treatment methods have been advocated but none has demonstrated any superiority to the others. The authors designed a controlled animal study in 88 rats to test three methods of early treatment of extravasation of the vesicant antineoplastic agent doxorubicin. The first step of the study included 48 Sprague-Dawley rats. All animals received intradermal injections of 1 mg doxorubicin superficially to the panniculus carnosus in the dorsum. The rats were then divided into four groups of 12 rats each, as follows: group 1, no treatment; group 2, immediate intradermal injection of 0.1 ml saline to the same site; group 3, immediate intradermal injection of 10 mug granulocyte macrophage- colony stimulating factor (GM-CSF) in 0.1 ml saline to the same site; group 4, immediate intradermal injection of 10 mug granulocyte-colony stimulating factor (G-CSF) in 0.1 ml saline to the same site. During the next 6 weeks the rats were observed for the development of necrosis. Ulcers developed and reached maximum size two weeks after the injections. The largest ulcers according to area were observed in group I and the mean value was 21.25 mm(2) (p < 0.05). Although wound areas were significantly smaller in the saline group than in the control group and the mean value was 7.58 mm(2) (p < 0.05), the smallest lesions were observed in groups 3 and 4, and the mean values were 1.08 mm(2) and 0.83 mm(2) respectively (p < 0.05). There was statistically no difference with regard to mean ulcer area between groups 3 and 4. During the second step of the experiment, the remaining 40 Sprague-Dawley rats were used. Groups containing 10 rats each were designed similarly after all animals received intradermal injections of I ring doxorubicin into the back. On the 10th day after the injection, the entire area of the ulcer together with the underlying panniculus carnosus was excised for pathological examination and for determination of glucose 6-phosphate dehydrogenase (G6PD) activity. On microscopic examination, the extravasated ulcer consisted of a large area of ischemic necrosis. There was marked damage to small blood vessels in the form of fibrinoid necrosis and vasculitis. Injured vessel counts were higher in the control group (group 1; p < 0.05). No difference was observed in G6PD activity between the groups. The authors conclude that both saline and tissue growth factors (GM-CSF and G-CSF) are useful for the early treatment of doxorubicin extravasation; however, GM-CSF and G-CSF are more beneficial.