O063 - SPLENECTOMY DOESN'T PRECLUDE SUBSEQUENT ALLOGENEIC HEMATOPOIETIC CELL TRANSPLANTATION IN MYELOFIBROSIS PATIENTS: A FRENCH NATIONWIDE STUDY USING A MULTISTATE MODEL (ID 1044)
- J. Bossard
- J. Beuscart
- M. Robin
- M. Mohty
- F. Barraco
- P. Chevallier
- M. Rubio
- A. Charbonnier
- D. Blaise
- J. Bay
- C. Botella-Garcia
- G. Damaj
- S. Maury
- P. Ceballos
- T. Cluzeau
- J. Cornillon
- M. Meunier
- C. Orvain
- T. Marchand
- F. Garnier
- A. Duhamel
- J. Kiladjian
- I. Yakoub-Agha
Background: Pretransplant splenectomy may improve allogeneic hematopoietic cell transplant (allo-HCT) outcomes but is associated with substantial morbidity and mortality that may delay or cancel a project of transplant. This study aims at determining whether pre-transplant splenectomy precludes subsequent allo-HCT, in myelofibrosis (MF) patients waiting for a transplant.
Methods: This study included all MF patients who were candidate to first allo-HCT from an unrelated donor in France, between January 1st, 2008 and January 1st, 2017, using the French registry of bone marrow transplantation (RFGM, Registre France Greffe de Moelle). With the Promise database, we identified transplanted patients, along with data regarding pretransplant splenectomy. For non-transplanted patients, local centers provided data from medical files. We excluded patients splenectomized before the initiation of unrelated donor search (ie, at registration).
We applied a multi-state model including the four following states:
1) “RFGM registration”, (ie, all patients at the time of initiation of unrelated donor search );
2) “Splenectomy”, (ie, splenectomized patients);
3) “Death before allo-HCT”; and 4) “Allo-HCT”.
All patients started in state
1) and could move to the intermediate state 2), or to one of the absorbing states 3) and 4). Similarly, patients could move from state 2) to state 3) or 4). We used Cox models with splenectomy as a time-varying variable and a clock-reset timescale to evaluate the association between splenectomy and subsequent HSCT or death.
Results: We recruited and analyzed 530 patients from 57 centers in France. Median age was 59 years old (interquartile range [IQR], 53 to 63 years old) and 310 (58.5%) were men. Median follow-up was 6 years. Eighty-one patients were splenectomized after registration, of whom 65 underwent subsequent allo-HCT and 9 died. Stacked probabilities of being in each state as a function of time are represented with the Aalen-Johansen estimator (Figure). At each timepoint, the distance between 2 adjacent curves represent the probability of being in the corresponding state. For instance; two years after registration on the RFGM, the estimated probabilities were: 28.6% for being alive, neither splenectomized nor transplanted; 2.1% for being alive, splenectomized and not transplanted; 45.7% for being transplanted without previous splenectomy; 12.1% for being dead without previous splenectomy; 10.6% for being splenectomized and transplanted; and 0.9% for being splenectomized and dead.
[Aalen-Johansen estimation of state occupancy probabilities.]
Splenectomized patients had a higher probability of being transplanted in the first 4 months after splenectomy, in comparison with non-splenectomized patients (unadjusted HR [Hazard Ratio], 7.2; 95% CI, 5.1 to 10.3), but not afterwards (HR, 1.2; 95% CI, 0.7 to 2.0). We found no significant association between splenectomy and death without allo-HCT (unadjusted HR, 1.6; 95% CI, 0.8 to 3.1).
Conclusions: When indicated, splenectomy can be performed among MF patients without precluding subsequent allo-HCT. However, its impact on post-transplant outcomes must be clarified before recommending larger indications for it.
Disclosure: This work was funded by a grant from the Agence Régionale de Santé Hauts-de-France (“Bourse année recherche”) and by the patients' association ALTE-SMP.
Authors declare no conflict of interest.