Long‐term results of the NF‐08‐TM protocol in stem cell transplant for patients with thalassemia major: A multi‐center large‐sample study

To the Editor: Hematopoietic stem cell transplant (HSCT) can cure thalassemia major (TM), but only 30% of the patients have a matched sibling donor. Alternative donors, including unrelated donors, cord blood, and parental donors, have continuously improved the results of HSCT in leukemia in the last decades. Therefore, HSCT from alternative donors were gradually applied to thalassemia. Our previous encouraging single-center results of the "the NF-08-TM" protocol in unrelated donor transplant in TM patients have published in Blood in 2012.1 Considering the limitations of our single-center study, we subsequently carried out a multi-center collaboration using the NF-08-TM protocol.1 Here we report the long-term results from the multi-center study. A total of 486 consecutive HSCT from four pediatric centers in China between July 2011 and June 2016 using the NF-08-TM protocol were analyzed. The ratio of male to female was 309 to 177 with a median age of 6 years (range, 2-23 years). Most patients were classified in group two according to the criteria of the NF-08-TM protocol. Patients were classified into two groups according to human leukocyte antigen (HLA-A,HLA-B, HLA-C and HLA-DRB1) matching, that is,10/10 matched group (n = 428, 88.1%) and 9/10 (n = 58, 11.9%). One hundred and nine-three patients underwent sibling donor HSCT (190 matches and three mismatches); 212 unrelated donor HSCT (173 matches and 39 mismatches), 35 parent donor HSCT (19 matches and 16 mismatches) and 46 sibling cord blood transplants (all matched). Except for cord blood transplant, peripheral blood stem cells were used as a stem cell source in the current cohort. A constant mononuclear cell dose (8.0 × 108/kg) was given to all patients regardless of CD3+ and CD34+ cell dose. All 497 patients received the NF-08-TM protocol, but the target of busulfan plasma steady-state concentration were 500-800 ug/L instead of 300-600 ug/L and doses were adjusted based on PK parameters. Sirolimus and tacrolimus were more often used since 2014. Among the 486 HSCTs with the median follow-up time of 56 months (range,28-93 months), overall survival (OS), thalassemia-free survival (TFS), graft rejection (GR) and transplant related mortality (TRM) were 94.7%, 93.3%, 2.8% and 5.3% (Figure 1), respectively. The corresponding outcomes for sibling donor HSCT (n = 193) were 97.4%, 97.4%, 1% and 2.4% (Figure S1A, see supplement, same below); for unrelated donor HSCT (n = 212) were 92%, 88.9%, 4.9% and 8% (Figure S1B); for parent donor HSCT (n = 35) were 94.3%, 94.3%, 0% and 5.7% (Figure S1C); and for cord blood transplant (n = 46) were 97.8%, 95.3%, 2.6% and 2.2% (Figure S1D), respectively. The results in mismatched unrelated donor HSCT were worse than matched sibling donor HSCT. The corresponding OS, TFS and TRM for mismatched unrelated donor HSCT vs (vs) matched sibling donor HSCT were 84.6% vs 97.4%(P = .001), 82.1% vs 97.4% (P = <.001) (Figure S2A) and 15.4% vs 2.6% (P = .001) (Figure S2B). The OS, TFS, TRM and GR for matched and mismatched unrelated donor HSCT were 93.6% vs 84.6% (P = .058, Figure S2C), 90.4% vs 82.1% (P = .089, Figure S2D), 4.4% vs 15.4% (P = .058, Figure S2E), and 4.8% vs 5.3% (P = .744, Figure S2F), respectively. The incidence of grade II-IV acute graft vs host disease (aGVHD), grade III-IV aGVHD, mild chronic GVHD (cGVHD), and moderate/severe cGVHD in the current cohort was 11.1%, 5.1%, 4.2%, and 1.7%, respectively. When compared with matched sibling donor HSCT, grade II-IV GVHD in matched unrelated donor HSCT was more common (18% vs 5.4%, P = .008) (Figure S3A).Matched unrelated donor HSCT also had more aGVHD than matched parent donor HSCT (23.6% vs 0%, P = .046) (Figure S3B).When compared with matched sibling donor HSCT, the incidence of grade II-IV aGVHD and moderate/severe cGVHD was higher in mismatched unrelated donor HSCT (5.4% vs 25.9%, P = .000) (Figure S3C) and 3.9% vs 18.4%, P = .000) (Figure S3D), respectively. More moderate/severe cGVHD occurred in mismatched unrelated donor HSCT than matched unrelated donor HSCT (18.4% vs 6.9%, P = .018).The cumulative incidence of hemorrhagic cystitis and venous occlusive disease in the entire cohort was 9.1% and 1.4%, respectively, and no statistically difference was found in both by donor-groups. In the current multi-center large-sample study, we observed the consistent outcomes with those published in 2012 and the more than 90% long-term TFS when using the NF-08-TM protocol in HSCT for TM patients, regardless of donor family relationships. Such results are encouraging, particularly in countries with limited resources per capita, insurance deficiency, blood-product shortage, or high prevalence of irregular transfusion and chelation due to economic reasons. Another notable fact is that most of the patients in the current study might be classified as unfavorable class II-III with Pesaro classification because of irregular iron-chelation, even though we did not routinely perform a liver biopsy. Our results compared favorably to those reported by Shenoy (2-year TFS of 81.3% with unrelated donor bone marrow transplants).2 In a recent international study of related and unrelated donor transplantation, the NF-08-TM protocol has also been shown to be one of the best protocols in TM HSCT.3 However, the NF-08-TM protocol was not designed with randomized control groups; the exact reasons for its superior results are unknown. Some specific features of the NF-08-TM protocol are worth considering: (a) reduced dose cyclophosphamide (Cy, only two days instead of four days); (b) Cy given prior to busulfan (Bu, Cy-Bu) to reduce hepatotoxicity; (c) avoided giving ATG and Cy on the same day which may precipitate heart failure due to ATG inducing high fever, and Cy requiring aggressive hydration and inherent cardiotoxicity; (d) reduced dose Bu (three days instead of four days) which was supplemented by reduced dose TT (10 vs 30 mg/kg); and (e) a relatively high but constant dose of MNC (8 × 108/kg) to avoid excessive CD3+ cell infusion, particularly when CD34+ cell dosage was low or underestimated. In our current study, the VOD incidence was only 1.4% and no patient died of VOD. It is remarkable that there were many patients with irregular chelation and poor transfusion in our cohort. Recent studies showed that Cy-Bu as myeloablative conditioning was associated with less hepatotoxicity, severe VOD, and mortality when compared with Bu-Cy.4 So, Cy is, to a major extent, eliminated through 4-hydroxylation by cytochrome P450 (CYP) family. Some Cy metabolites are detoxified through conjugation with glutathione (GSH). When Bu is detoxified, it also needs and reduces GSH. Reverse order of usual Cy following Bu (Bu-Cy) to Bu following Cy (Cy-Bu) in mice and human studies showed that Cy-Bu could reduce the incidence of VOD and TRM through reduced damage to the liver.5 Furthermore, it is known that Cy is a marrow cell mobilizing agent, and it may induce lymphocyte proliferation by negative feedback after lymphodepletion. In Cy-Bu conditioning, therefore, more lymphocytes and stem cells can be mobilized into a cell-proliferation cycle in the blood and consequently enhance elimination by Bu and Flu administered thereafter. We added TT to the NF-08-TM protocol in replacement of 1/4 dose of Bu because of its potent myeloablative and immunosuppressive effect. Thus BU-TT-CY is an effective preparative regimen for T- cell-depleted allo-HSCT in genetic diseases, providing sustained engraftment, with a relatively low incidence of both VOD and pulmonary toxicity. The NF-08-TM protocol is associated with low GR incidence; despite WBC-depleted blood infusion was unusual before 2010 in China. The study mentioned above from other Chinese centers showed a high GR incidence (17/50, 34%) when the NF-08-TM protocol was not used.6 It is known that Cy is a marrow cell mobilizing agent and may induce lymphocytes proliferation by negative feedback7 after lymphodepletion. In Cy-Bu conditioning, therefore, more lymphocytes and stem cells can be mobilized into the cell-proliferation cycle in the blood, and consequently enhance elimination by Bu and Flu administered thereafter. A relatively high and constant-dose of MNC (8 × 108/kg) was infused regardless of CD34+ and CD3+ doses, which may be an easy and practical way to overcome the high risk of rejection in thalassemic HSCT and to limit the possibility of excessive T cell dose, particularly when the CD34 cell dose is on the low side. It is worth mentioning that all patients received peripheral blood stem cells to overcome high rejection risk. Although this stem cell resource has more T cells and more GVHD tendency, it did not result in high GVHD in the current study when compared with those reported in the literature. We thought that the use of proximal ATG and a reduced-toxicity regimen may lead to in vivo T cell depletion and may limit tissue damage that predisposes inflammation and alloantigen presentation. No significant difference between mismatched and matched unrelated donor transplants was found. Still, we observed that mismatched unrelated donor transplant had worse OS, TFS, and TRM, when compared with matched sibling transplant, which could be due to two risk factors (ie, unrelated relationship plus mismatch) existed together. In summary, this is a long-term, large-sample multi-center study with alternative donor transplants for TM patients in the literature. The current study demonstrated that the NF-08-TM protocol had improved outcomes in alternative donor transplants with excellent OS and TFS for all groups regardless of donor family relationships. Mismatched unrelated donor transplants had poor results when comparing matched sibling donor transplants for TM patients. The authors declare no conflict of interest with the content of this paper. Appendix S1 Supporting information Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.