Long-Term Follow-Up of IV Busulfan (Bu) with Fludarabine (Flu) vs IV Bu with Cyclophosphamide (Cy) as Pre (Allogeneic) Transplant Conditioning Therapy for AML/MDS.

We previously demonstrated early safety and efficacy of fixed-dose IV BuFlu. Here we compare IV BuFlu and IV BuCy2, with long follow-up in a large number of pts. Since pts were treated on consecutive, not randomized programs, we used pts receiving Flu-Melphalan (M) in a program spanning the time of the IV Bu-based studies to estimate/correct for this non-randomized “period” effect.

Methods: 293 pts received Bu 130 mg/m2, Flu 40 mg/m2 (after April 2001; n=148); Bu 0.8 mg/kg × 16, Cy 60 mg/kg × 2 (before April 2001; n=67); Flu 30mg/m2 × 4, M 140–180 mg/m2 (FM; before and after April 2001; n=78). Patients received FM mainly based on higher age and coorbidities. Graft-vs-host dx (GVHD) prophylaxis was tacrolimus-based. Bayesian log-normal regression models were used to assess covariate- and treatment effects on survival, non-relapse mortality (NRM) and time to progressive disease (TPD). Covariates and survival model are shown in the table. The model accounts for the “period” effect, the effects of Bu-vs-FM, and Flu-vs-Cy within IV Bu pts. Patient and treatment characteristics were (BuFlu, BuCy2 and FM): median age 46 (19–66), 39 (13–64) and 54 (22–74) years. At HSCT 47%, 48% and 21% of the patients were in CR1-3. Donors were HLA-identical siblings in 53%, 79% and 41%. 30% of the Bu groups and 47% of FM had poor prognosis cytogenetics. Results: Median follow-up in months (BuFlu, BuCy2 and FM; alive pts): 30 (9–56; n=80; all but 2 pts followed for at least 1 yr), 69 (24–102; n=23) and 48 (17–89, n=25). 100-day NRM for pts in CR at HSCT was 2% and 9% after Bu-Flu and BuCy2, while 1 yr-NRM was 7% and 18%. BuFlu produced significantly better EFS (P=0.04) and overall NRM (P=0.02) than BuCy2, but TPD was similar (P=0.34). 2-yr actuarial survival (CR pts): 75% (BuFlu) vs. 50% (BuCy), P=0.01.

Conclusion: After accounting for covariate and treatment period effects, there was a significant benefit to being treated with BuFlu compared to BuCy2 in spite of a higher median age and proportion of MUDs. The benefit was most evident for CR-pts due to a significant reduction in NRM, without loss of antileukemic activity due to exchanging Cy with Flu.

Fitted Bayesian log-normal overall survival model

Variable .	Mean .	SD .	Posterior 95% credible interval 2.5% 97.5%* .		Probability of beneficial effect .
*1=beneficial effect; 0=low likelihood of beneficial effect
Cytogenetics	−0.85	0.4	−1.652	−0.045	0.017
Disease status (CR vs other)	1.12	0.55	0.019	2.15	0.975
Donor (sibling vs other)	0.606	0.237	0.149	1.056	0.992
Age	−0.022	0.010	−0.041	−0.003	0.017
Circulating blasts (0 vs >0)	0.951	0.268	0.454	1.463	1
Platelet count	0.002	0.001	0.000	0.004	0.967
Period effect	−0.914	0.458	−1.795	0.041	0.029
IV Bu (vs FM)	−0.513	0.481	−1.455	0.420	0.148
BuFlu (vs Bucy2)	1.476	0.544	0.406	2.555	0.995

Variable .	Mean .	SD .	Posterior 95% credible interval 2.5% 97.5%* .		Probability of beneficial effect .
*1=beneficial effect; 0=low likelihood of beneficial effect
Cytogenetics	−0.85	0.4	−1.652	−0.045	0.017
Disease status (CR vs other)	1.12	0.55	0.019	2.15	0.975
Donor (sibling vs other)	0.606	0.237	0.149	1.056	0.992
Age	−0.022	0.010	−0.041	−0.003	0.017
Circulating blasts (0 vs >0)	0.951	0.268	0.454	1.463	1
Platelet count	0.002	0.001	0.000	0.004	0.967
Period effect	−0.914	0.458	−1.795	0.041	0.029
IV Bu (vs FM)	−0.513	0.481	−1.455	0.420	0.148
BuFlu (vs Bucy2)	1.476	0.544	0.406	2.555	0.995

View Large

View largeDownload slide

Survival – patients in CR at transplant

View largeDownload slide

Survival – patients in CR at transplant

Close modal