Hexane and carbon disulfide.

IN AN EARLIER STUDY (Barr et al., 1981) we explored the pathogenesis of soft-tissue necrosis induced in guinea-pigs by the con-crolled extravascular administration of cytotoxic drugs. This system provides a model of the phenomenon which may complicate the inadvertent perivenous injection of cancer chemotherapeutic agents in clinical practice. The current investigation was undertaken in order to determine whether the appearance of such lesions could be prevented by secondary prophylaxis. To this end a variety of materials was instilled into the target tissues following the injection of the injurious drugs. Hartley guinea-pigs (Camm Labs, Wayne, New Jersey) were housed and prepared for study as described previously (Barr et al., 1981); three animals were used for each study point. Vincristine was administered intradermally (i.d.) in doses of 2-5 or 5 0 jtg in 03 ml. Adriamycin was injected s.c. at various concentrations in volumes of 1 ml. Agents used for secondary prophylaxis were isotonic phosphate-buffered saline (PBS, pH 7.4), hydro-cortisone (at several concentrations), in-domethacin (2.8 x 1O-7M) and sodium bi-carbonate (8.4%). Indomethacin was prepared for injection by dissolving the contents of one capsule (25 mg) in 5 ml of aqueous dimethyl sulphoxide at 37°C over several hours, and diluting this solution further with water. These agents were administered i.d. in volumes of 1 ml and s.c. in volumes of 5 ml at the sites of prior injection of cytotoxic drugs. Such intervention was used either immediately or 24 h after drug injection. Positive controls were animals receiving vincristine or Adriamycin alone. Negative controls were obtained by injecting animals with the intervention agents alone. Secondary pro-phylaxis was attempted only by the same route as the original cytotoxic injection. The animals were examined daily for evidence of local injury at the sites of injection. Such reactions were graded in severity according to a simple scale, shown in Table I (Barr et al., 1981). The results are summarized in Tables I and II.

IN AN EARLIER STUDY (Barr et al., 1981) we explored the pathogenesis of soft-tissue necrosis induced in guinea-pigs by the concrolled extravascular administration of cytotoxic drugs. This system provides a model of the phenomenon which may complicate the inadvertent perivenous injection of cancer chemotherapeutic agents in clinical practice. The current investigation was undertaken in order to determine whether the appearance of such lesions could be prevented by secondary prophylaxis. To this end a variety of materials was instilled into the target tissues following the injection of the injurious drugs.
Hartley guinea-pigs (Camm Labs, Wayne, New Jersey) were housed and prepared for study as described previously (Barr et al., 1981); three animals were used for each study point. Vincristine was administered intradermally (i.d.) in doses of 2-5 or 5 0 jtg in 03 ml. Adriamycin was injected s.c. at various concentrations in volumes of 1 ml. Agents used for secondary prophylaxis were isotonic phosphatebuffered saline (PBS, pH 7.4), hydrocortisone (at several concentrations), indomethacin (2.8 x 1O-7M) and sodium bicarbonate (8.4%). Indomethacin was prepared for injection by dissolving the contents of one capsule (25 mg) in 5 ml of aqueous dimethyl sulphoxide at 37°C over several hours, and diluting this solution further with water. These agents were administered i.d. in volumes of 1 ml and s.c. in volumes of 5 ml at the sites of prior injection of cytotoxic drugs. Such intervention was used either immediately or 24 h after drug injection. Positive controls were animals receiving vincristine or Adriamycin alone. Negative controls were obtained by injecting animals with the intervention agents alone. Secondary prophylaxis was attempted only by the same route as the original cytotoxic injection. The animals were examined daily for evidence of local injury at the sites of injection. Such reactions were graded in severity according to a simple scale, shown in Table I (Barr et al., 1981).
The results are summarized in Tables   I and II. Positive controls At doses of 2*5 and 5 ,g vincristine evoked lesions in all control animals within 48 h of i.d. administration. With the lower   )ccurred con-consistent findings. After the higher dose rhree mg of of vincristine, immediate secondary prons uniformly, phylaxis failed to prevent the appearance _nt resolution. of lesions, but these proceeded to heal did not cause spontaneously within 1 week. Delayed s.c. injection. intervention was without measurable effect. Again the outcome was observed in all animals. condary pro-(b) Adriamycin-no beneficial effect was {hen injected observed with any of the intervention aethacin and agents, even when used immediately after .cited florid administration of the drug. administered In this initial study of secondary prophylaxis in the management of softtissue necrosis, vincristine and Adriamycin were chosen as the injurious drugs, since Lydrocortisone most reported cases of this necrosis have trance of softbeen associated with the use of these two f injection of agents. Moreover, as determined by our ,r either agent earlier study (Barr et al., 1981), vincristine tely after the evokes such injury only after i.d. injection, is delayed for whereas Adriamycin also causes tissue inhibited the damage after s.c. administration. The is. These were timing of intervention was selected to resemble common events in clinical practice, namely recognition by the individual administering the drug that extravasation has occurred during the injection procedure or, in the absence of such an observation, the complaint by the patient on the following day of discomfort at the site of injection. PBS was deemed to be a simple, isotonic diluent, hydrocortisone an antiinflammatory agent, indomethacin an inhibitor of prostaglandin synthesis, which may play a role in some of these druginduced reactions (Giri et al., 1975), and sodium bicarbonate a means of inhibiting the binding of Adriamycin to nucleic acids (Wilson et al., 1976), for it has been suggested that it is the DNA-Adriamycin complex rather than the drug itself which is harmful in this fashion (Zweig & Kabokow, 1978).
From the current data it appears that the locally injurious effects of i.d. vincristine can be prevented or diminished by prompt dilution of the drug within the skin. Even delayed intervention with hydrocortisone may help in amelioration. Evidently the efficacy of such manoeuvres will be related to the amount of the drug which extravasates.
Our experimental findings with Adriamycin are disappointing, and seem to conflict with anecdotal clinical experience (Zweig & Kabokow, 1978) including our own. Furthermore, recent studies have suggested that sodium bicarbonate may be of value (Bartkowski-Dodds & Daniels, 1980). However, it should be noted that soft-tissue necrosis has followed the extravasation of this material in clinical practice (Gaze, 1978). Certainly, local instillation of hydrocortisone does not appear to be of clinical benefit (Reilly et al., 1977). Clearly, further studies of secondary prophylaxis are required, and the role of intervention agents in the prevention of lesions induced by other drugs remains to be investigated.