A murine (BALB/c) skin toxicity model was used to evaluate various possible antagonists to vinca alkaloid-induced skin ulceration. Reproducible dose-response relationships were developed for vinblastine (VBL) and vindesine (VDS). With vincristine (VCR) only about 70% of mice developed dose-dependent ulceration. On an equal weight basis, VCR proved to be significantly more toxic than either VBL or VDS (P less than .05 by Student's t-test). Effective local intradermal antidotes to VBL, VDS, and VCR included hyaluronidase, normal saline, and calcium leucovorin (P less than .05 by the Student's Newman-Keuls multiple range test). Mild, topical skin heating significantly reduced VCR ulceration. In contrast, diphenhydramine and sodium bicarbonate were ineffective as local antidotes. Topical skin cooling, however, significantly increased vinca-induced skin ulcers for VBL, VDS, and VCR (P less than .05). Hydrocortisone, vitamin A topical cream, and isoproterenol increased skin toxicity. [3H]VBL was given intradermally to follow the drug's pharmacokinetic disposition from the skin and adherent panniculus carnosus muscle. [3H]VBL exhibited two phases of elimination: a rapid early phase [half-life (t 1/2) of approximately equal to 30 min] and a prolonged terminal phase (t 1/2 of approximately equal to 17 hr). The application of heat increased the distributive, early phase by 0.5-2.5 hours and did not enhance the terminal elimination of the drug from skin. Intradermal hyaluronidase significantly reduced the area under the ulceration multiplied by the time curve to one-seventh the control value, the peak [3H]VBL skin concentration to one-half the control value and the terminal [3H]VBL t 1/2 in skin to one-third the control level (P less than .05 by Student's t-test). These results show hyaluronidase to be an effective antidote for vinca-induced skin ulceration. Local glucocorticosteroids and topical cooling are definitely contraindicated in the management of inadvertent vinca alkaloid extravasations.