Discussion
The conditioning regimen used for our children with acute leukemia was initially developed for adult transplant recipients. It included Bu (targeted at an AUC of 3750μmol*min) and Flu (dosed at 250mg/m2 total), to which ATG was added to reduce the incidence of GvHD.13,16 TBI at 400cGy was included in the protocol to reduce the incidence of relapse and was noted to significantly improve outcomes.17 Our protocol uses a higher Flu dose compared with other centers using similar regimens.
Survival outcomes, relapse incidence and non-relapse mortality in our cohort were similar between the subjects who had TDM and those who did not. The number of subjects in our cohort played an important role in the results not reaching significance levels. Previous studies that have compared Bu based conditioning to TBI regimens in children with malignancies have presented outcomes favouring TBI based conditioning.8,18 This has mainly been due to higher rates of relapse and non-relapse mortality rates in the Bu groups. However, these earlier studies have not consistently reported measurements of Bu serum levels.8,18,47
Recently, the results of the international FORUM study (For Omitting Radiation Under Majority age), a multicenter randomised trial comparing a fractionated TBI/Etoposide regimen to various chemotherapy-based conditioning (CHC) regimens has shown conclusively that a TBI based conditioning regimen is superior for children above four years of age with ALL.19 In that study individual institutions were allowed to choose between Bu and Treosulfan (Treo) as a backbone for their CHC regimens. A potential drawback to this flexibility is the introduction of heterogeneity.51 Therefore, despite an overall equivalence between the Bu and Treo based regimens, data does not allow a direct comparison across individual CHCs. Furthermore, Bu serum level targeting was allowed but not mandated by the FORUM study and no analysis is done to compare outcomes between the centers that used serum targeting of Bu and those that did not.51Additionally, although cranial radiation was allowed for central nervous system positive disease, none of the CHC regimens in the FORUM study included TBI. For these reasons, a direct comparison of outcomes with the regimen used in the current study is not appropriate.
In 2007, Chaudhury et al published an abstract reporting their findings using a regimen incorporating both Bu and TBI.20Although the targets for Bu TDM were similar (4000 μmol *min/day), an Important difference in their preparation regimen was their use of lower cumulative doses of Flu (150 mg/m2 over five days). They reported a one-year OS and PFS of 63% and 56% respectively. A recent article by Rosoff et al further expanding analysis of the same cohort, using Bu-Flu-ATG and 400cGy TBI regimen, reported a five-year OS of 37% in their ALL cohort consisting of only nine children (one of whom had positive MRD).21 In both studies by Chaudhury and Rosoff the most important reason for failure was high rates of relapse, seen in 29% and 28% of subjects respectively.20,21 We speculate that the lower cumulative incidence of relapse in our cohort may potentially be related to the higher doses of Flu. Evidence for this has recently been published in both transplant and Chimeric Antigen Receptor T-cell (CAR-T) settings with studies showing that both lower and very high serum Flu levels result in inferior outcomes driven by increased relapse rates and toxicity respectively.52,53
Although the rates of aGvHD reported in this cohort are higher than those in other studies, which quote incidences closer to 30%,18,21,23 they are similar to results reported by Chaudhury et al who report their severe aGvHD rate as being 32%.20 A striking similarity between both studies is the use of PBSC products, utilised in close to 90% of transplants in the current study and in 75% of cases in the study by Chaudhury. PBSC products have been shown to be associated with increased aGvHD in multiple studies although the difference has not been consistently significant.24,25,48 Six (20%) of the donor-recipient pairs in our study were female donors to male recipients. Pulsipher et al, have shown that the presence of aGvHD is related to a lower risk of relapse,50 which perhaps contributed to the EFS in the current study.
The cumulative incidence of cGvHD in our study was comparable with most other studies. 18,20,29 It was lower than rates reported by Rosoff et al and those in a report by Modi et al (which used a similar conditioning regimen for adults with AML).21,23 The reason for the low rates of cGvHD in our study despite a higher incidence of aGvHD may be explained by our method of giving ATG.30,31 At our institution we follow a regimen similar to one recommended by Bacigalupo et al.31 In their 2001 paper, they suggested all patients receive ATG at 7.5mg/kg as this dose was not associated with increased risk of infections and was shown to reduce the incidence of cGvHD.
The incidence and timing of viral infections seen in our cohort were similar to those reported in literature32-34,36 with the exception of CMV which appeared to occur earlier than some studies have described (median of 24 days in our study compared with 33 to 41 days in some studies).34,54 There was no effect on outcomes which is contrary to findings in other publications,32,35 with the exception of delayed platelet engraftment dates in those subjects receiving platelet infusions for hematuria secondary to BK virus induced cystitis Bacterial blood stream infection incidence, timing and the organisms grown were also comparable to other reports.37-41 The incidence of mucositis was 100% in this cohort which is in keeping with other reports using a similar regimen.21
The incidence of SOS in studies using conditioning regimens including Bu and Flu compared to other combinations have had mixed results with some studies reporting a reduction in SOS42-44 and others showing no difference.45 In our cohort the cumulative incidence of SOS at d100 was 15% (three patients). Although this rate appears higher compared with other studies, it may not be significant owing to the small number of patients developing SOS.20,46 In the adult study by Kabriaei et al, six patients in their cohort of 107 patients developed SOS resulting in a prevalence of 6%.46 Similarly, Chaudhury et al reported a SOS prevalence of 10% (three cases).6 It is notable that SOS was cited as a direct cause of death in only one case in each of the two studies. The univariate analysis in our cohort revealed no significant correlation of the occurrence of SOS with mortality.
The limitations to our study include firstly, that this is a retrospective chart review and therefore dependant on the quality of data collected and stored. Data pertaining to SOS were missing for our earlier subjects. Secondly, the number of patients, once split into the two cohorts, was limited and this may have contributed to the lack of significance when comparing the outcomes following Bu pharmacokinetic targeting. Furthermore, the comparison with a historic cohort introduces some inconsistencies as there is a possibility that improvements in outcomes over the years were in fact due to an unidentified confounder and not the practice of targeting Bu serum levels. It should be noted however, that no significant changes to supportive care were introduced during the period of data collection. Related to this, the combination during analysis of those children who did not have any Bu monitoring with the group that were monitored and were out of range may have introduced a bias, as the earlier group is a heterogenous one and may contain some children who were within acceptable serum levels. Finally, we did not differentiate between T-cell and B-cell ALL which introduces heterogeneity in the cohort analysed. Despite these limitations, the paper documents outcomes following transplant in children with ALL using a unique and previously unpublished regimen of Bu-ATG and lower dose TBI regimen combined with a higher dose of Flu. Although previous studies have utilized lower dose radiation in this setting, those cohorts have been of limited size compared with the current report. We have demonstrated similar outcomes to other conditioning regimens when serum Bu levels are targeted suggesting this may be a substitute to consider when lowering radiation dose is desirable.