Aerosolized pentamidine as pneumocystis prophylaxis after bone marrow transplantation is inferior to other regimens and is associated with decreased survival and an increased risk of other infections

Pneumocystis carinii pneumonia (PCP) is a life-threatening but preventable infection that may occur after bone marrow transplantation (BMT). Although various prophylactic regimens have been used in this setting to prevent act ive infection, their efficacy, toxicity profile, and impact on outcomes are poorly described in this patient group. We undertook a retrospective cohor t study in which we reviewed the records of 451 adult patients who underwen t BMT for hematologic malignancies, aplastic anemia, or myelodysplasia over a 7-year period at the Brigham and Women's Hospital. Post-BMT PCP prophyla xis consisted of aerosolized pentamidine (AP) 150 mg every 2 weeks or 300 m g per month, trimethoprim/sulfamethoxazole (TMP/SMX) 160/800 mg orally b.i. d. 3 times per week, or dapsone 100 mg orally each day. Prophylaxis was con tinued for 1 year post-BMT in all patients when clinically feasible. One hu ndred twenty-one patients were unevaluable because of death or relapse <60 days after BMT (n = 89), loss to follow-up upon hospital discharge (n = 20) , or other reasons (n = 12). Three eligible patients did not receive any pr ophylaxis and were not further evaluated. Of the 327 patients analyzed, 133 underwent autologous BMT, 4 syngeneic BMT, 159 related allogeneic BMT, and 31 unrelated allogeneic BMT. Graft-versus-host disease prophylaxis in the 190 patients receiving allogeneic BMT consisted of T-cell depletion with an ti-CD5 and complement in 58 patients and cyclosporine/methotrexate or FK506 with or without steroids in 132 patients. Eight of 327 (2.4%) documented P CP cases were identified, 0 of 105 in patients receiving only TMP/SMX. Four cases occurred in patients receiving only AP (4/44, 9.1%; odds ratio [OR] relative to TMP/SMX 23.4, 95% confidence interval [CI] 1.2, 445.2); 1 in pa tients receiving only dapsone (1/31, 3.2%; OR not significant); 2 in patien ts receiving more than 1 prophylactic regimen (2/147 1.4%; OR not significa nt); and 1 >1 year post-BMT in a patient who was off PCP prophylaxis. Altho ugh the patients receiving only AP had a significantly lower probability of treatment-related toxicity than those receiving TMP/SMX (OR 0.19 [95% CI 0 .04, 0.85]), the probability of their acquiring other serious non-PCP infec tions was increased (OR 2.2 [95% CI 1.0, 4.6]), and the probability of thei r dying by 1 year post-BMT was significantly higher (OR 5.2 [95% CI 2.4, 26 .6]), even when adjusted for variables such as type of BMT (autologous vers us allogeneic; high versus low risk) and sex. Although AP is associated wit h fewer toxicities, the data show that it is inferior to TMP/SMX in prevent ing PCP in the post-BMT setting and is associated with an increased risk of other infections and a higher mortality at 1 year after BMT.