PROPHYLAXIS FOR ACUTE GRAFT-VERSUS-HOST DISEASE FOLLOWING UNRELATED DONOR BONE-MARROW TRANSPLANTATION

Despite the use of conventional chemoprophylaxis regimens, patients re ceiving unrelated-donor BMT are at high risk of developing severe acut e GVHD. We evaluated a prophylactic regimen combining CsA, MTX and ant i-CD5-ricin A chain immunotoxin (H65-RTA) in 31 patients; pentoxifylli ne was also given to reduce the anticipated nephrotoxicity of CsA. In most cases, planned doses of CsA, MTX and H65-RTA were given (i.e. to 77%, 77% and 93% of patients, respectively). Although fluid retention requiring diuretic therapy was frequent, only 1 patient had a >10% une xplained increase in body weight during the first 21 days post-BMT. Al so, while significant increase of the baseline serum creatinine was no ted in 7 patients, none required dialysis. One patient suffered a reve rsible allergic reaction to the immunotoxin; no other side effects att ributable to this regimen were observed. All but 2 patients engrafted (1 died of fungemia on d + 19 and the other had persistent leukemia) a nd no late graft failures were observed. Seventeen patients developed acute GVHD grade greater than or equal to II (probability, 58% [95% CI 41-76%]); 7 had grade greater than or equal to III (probability, 24% [95% CI 12-43%]). In the 27 patients who achieved stable engraftment a nd have survived beyond d + 100, the 3-year probability of developing chronic GVHD was 66% (95% CI 48-84%). As of the last follow-up prior t o 01 May 1994, 13 patients are alive in CR and one in relapse; 9 of th ese patients are off all immunosuppressives and well. Four other patie nts relapsed and died, and 13 died of other transplant-related causes. Event-free survival (EFS) is 46% (95% CI 27-63%) at 3 yrs, with a med ian followup of 33 mos (range 19-42) in survivors. This regimen may di minish the severity of acute GVHD without a high incidence of graft re jection. The relatively low relapse rate observed to date suggests the persistence of an allogeneic antitumor effect. These preliminary resu lts require confirmation.