Risk Factors for Transplant Outcomes in Children and Adolescents with Non-Malignant Diseases Following Allogeneic Hematopoietic Stem Cell Transplantation

Background

DESIGN OF THE STUDY:

Treatment outcome was analyzed in a retrospective study on children with non-malignant diseases treated with allogeneic HCT in 10 consecutive years from 2006 to 2015 in 4 pediatric centers in Poland. Data were retrieved from institutional databases. All patients gave their consent for the individual transplant center to store the data. Approval for the study was given by the institutional committee. This study was a retrospective analysis of patient data. Given the retrospective nature of the study, the requirement for obtaining informed consent from each patient was waived.

To compare the outcome, patients were analyzed according to the diagnosis, age at transplant, donor type, stem cell source, conditioning regimens, transplanted CD34+ cells dose, and pediatric TRM score. This has been published by Matthes-Martin et al. and is defined as a sum of points depending on 3 risk factors: donor type, age at transplantation, and disease status (0 points for matched sibling donor, age ≤10 years at the time of HCT, and all non-malignant diseases; 1 point for alternative donors, age >10 years, and advanced malignant disease at the time of HCT) [5]. In non-malignant patients, such stratification results in 3 risk groups: low risk (score 0), intermediate risk (score 1), and high risk (score 3).

Material and Methods

TRANSPLANT CHARACTERISTICS:

Matched donors were specified as 10/10 and 9/10 match by HLA high-resolution typing. Conditioning regimens were in line with current EBMT recommendations [4]. Conditioning was considered myeloablative (MAC) when full-dose busulfan (4 mg/kg/day for 4 days) was used, and it was considered non-myeloablative (non-MAC) when fludarabine, cyclophosphamide, low-dose TBI (1–2 Gy), low-dose busulfan (1 mg/kg/day for 2–4 days), or ATG alone were used. Reduced-toxicity conditioning (RTC) was based on treosulfan (10–14 g/m2 for 3 days, age-dependent). The diagnosis and management of GvHD, as well as supportive care and infections prophylaxis, were provided according to EBMT guidelines and center procedures.

DEFINITIONS:

The main outcomes analyzed were transplant-related mortality (TRM) and overall survival (OS). TRM was defined as death resulting from any cause connected with transplant procedure. For estimating OS, death from any cause was considered an event. Surviving patients were censored at last follow-up.

STATISTICAL METHODS:

Statistical analysis was performed using SPSS IBM Statistics (Version 24) and R package version 3.4.1 (survival library version 2.41–3). Non-parametric tests used were Pearson’s chi-square test and the chi-square test with simulating p values, which are insensitive to small numbers, and the Kruskal-Wallis test was used for group comparison. OS was estimated using Kaplan-Meier method and log-rank tests. Multivariate analysis of OS was performed with a cause-specific proportional hazard model (Cox model). Statistical significance was considered P<0.05.

Results

TRANSPLANT-RELATED MORTALITY:

Depending on the donor type, TRM occurred more frequently in MUD and MMD HCT compared to MSD transplant recipients (p=0.02). TRM was more frequent in PB recipients compared to in BM recipients (p=0.004), and in patients who received >5×10⁶/kg CD34 in comparison to those who were given ≤5×10⁶/kg (p<0.0001).

TRM occurred more frequently in SCID patients compared to other diagnoses, but the difference was at the border of significance (p=0.055). SCID patients in most cases (61.1%) received >5×106/kg CD34+ cells. There was no statistically significant difference in the occurrence of TRM depending on age at transplant, conditioning type, and pediatric TRM score. Risk factors for transplant-related mortality are shown in Table 3.

GvHD was diagnosed in 38 patients (20.8%): acute GvHD in 30 and chronic in 8. There was no statistically significant difference in acute and chronic GvHD occurrence depending on diagnosis, stem cell source, donor type, and CD34+ cells dose. Acute GvHD occurred more frequently in children age 10 years and younger at time of HCT: 4/110 (21.8%) children ≤10 years vs. 6/73 (8.2%) >10 years (p=0.01). None of the patients age >10 years developed cGvHD, and it was diagnosed in 8/110 children age ≤10 years (p=0.015).

SURVIVAL:

Overall survival (OS) in the entire group of patients was 85% at the median observation time of 3.66 years, as well as at the maximum (11.23 years) observation time. OS differed significantly according to underlying disease, and was 95% in CGD patients, 92% in IBMFS patients, 91% in IEM patients; 89% in SAA patients; 75% in non-SCID PID patients, and 65% in SCID patients (p=0.022).

Significantly lower survival was observed in patients transplanted from MMD (67%) and MUD (81%) compared to those transplanted from MSD (96%) (p=0.02). After excluding 6 patients transplanted from MMD, the OS was also significantly lower in patients transplanted from MUD compared to MSD (p=0.012). OS was significantly higher in patients receiving BM as a graft source compared to those who received PB (89.6% and 69.8%, respectively) (p=0. 0007). CB transplants were excluded from the analysis due to the small number of such patients. Statistically higher OS was observed in patients receiving ≤5×10⁶/kg CD34+ cells (92%) compared to those receiving >5×10⁶/kg (68%) (p<0.001). OS did not differ according to sex and age at HCT or according to conditioning type and TRM score stratification.

RISK FACTOR ANALYSIS:

In univariate analysis, the probability of TRM was significantly lower: HR 0.19 for MSD recipients (95% CI 0.04–0.80) (p=0.02), HR 0.33 in patients receiving BM as stem cell source (95% CI 0.16–0.70) (p=0.004), and HR 4.53 in patients receiving ≤5×106/kg CD34+ cells (95% CI 2.07–9.9) (p=0.0001). For patients diagnosed with SCID, predicted TRM was HR 3.84, which was higher than in patients with other diagnoses (95% CI 1.33–11.06) (p=0.01). Prediction of TRM events according to diagnosis, donor type, stem cell source, and CD34+ cells dose is presented in Figures 1 and 2. Multivariate analysis showed that 3 factors affect TRM: (1) stem cell source, with lower probability of TRM in BM recipients: HR 0.44 (95% CI 0.2–0.96) (p=0.04); (2) diagnosis, with higher probability of TRM in SCID: HR 6.02 (95% CI 1.34–26.95) (p=0.02); and (3) CD34+ cell dose, with higher probability of TRM in recipients of >5×106 138 cells/kg: HR 4.53 (95% CI 2.07–9.9) (p=0.0001) (Table 4).

CAUSES OF DEATHS:

There were 156 patients still alive at the end of the study (85.2%), and 27 patients (14.8%) had died, all due to transplant-related complications. The causes of death with patient details are presented in Table 5.

Discussion

Patients with non-malignant diseases treated with HCT constitute a heterogenous group in terms of diagnoses. In our cohort of patients, TRM was observed more frequently in patients with SCID, and these children had the lowest survival. According to Gennery et al., significantly lower survival was observed in SCID patients transplanted from unrelated and mismatched related donors (69% and 66%, respectively) compared to matched related donors (90%) [6]. In the same study, the absence of respiratory impairment, or viral infection before transplantation, were associated with better prognosis on multivariate analysis [6]. According to an analysis of SCID patients by Heimall et al., the average survival for patients who underwent HCT was about 70% at 3 years after transplantation [7]. This may be the consequence of prolonged immunosuppression and previous life-threatening infections, observed especially in severely immunocompromised patients. In another study of this same group, 2-year overall survival in SCID patients was lower in those transplanted with active infection [8]. In our analysis, 6/18 patients (33.3%) with SCID had active infection at the time of HCT, and 2 patients with SCID died due to infections. Infection was the most common cause of death in the whole analyzed cohort.

In our patients, TRM occurred more frequently in PB recipients compared to BM recipients, with significantly higher OS in patients receiving BM. Bacigalupo et al. reported that PB as the stem cell source was a negative predictor of survival in patients transplanted for SAA [9]. Studies by Eapen et al. on SAA patients showed that stem cell source was an important risk factor for survival, and BM was the preferred graft source in SAA patients, with lower mortality after BM compared to PB [10,11]. Several analyses show that PB stem cells and higher CD34+ dose tend to be associated with increased mortality and lower survival due to GvHD [12,13]. In our analysis, CD34+ cells dose, stem cell source, and underlying disease had no effect on acute and chronic GvHD, although CD34+ cells dose >5×106/kg was associated with higher incidence of TRM and lower OS. Many studies found that higher CD34+ cell dose improved engraftment and had no effect on severe GvHD and TRM in children [14–16], but other studies found a correlation between low CD34+ dose and worse survival [17,18]. According to Remberger et al., in patients treated for malignancies, those given a very high CD34+ cell dose (≥11×106/kg) had decreased rates of survival [19]; stem cell dose was an important risk factor for survival when connected with graft source. Patients who received PB containing <11.0×106/kg CD34+ cells had significantly better OS compared with those receiving BM, regardless of cell dose. The study found that PB CD34+ cell dose ≥11×106/kg was correlated with reduced OS and increased mortality rate [19].

In our cohort, univariate analysis showed that OS was significantly lower in children transplanted from MUD compared to MSD. Several authors showed that patients transplanted from unrelated donors have lower survival and higher TRM [17–20]. Locatelli et al. reported an estimated OS of 87% in patients with FA when the donor was an HLA-identical sibling vs. 40% in MUD transplants [21]. Roy et al. reported outcomes of 61 transplanted BDA patients, with 76% 3-year OS after MSD HSCT, and 39% OS after MUD HSCT [20]. In SAA patients, the OS was 89% in our cohort, which is in line with other studies, in which OS varied from 91% to 86% after MSD HCT. In patients after failed immunosuppressive therapy (IST) treated with MUD HCT, OS was lower but still improving [22–24].

Recommended conditioning regimens have changed in recent years and now differ. In our cohort, conditioning type had no effect on TRM and OS. The results an analysis by Burroughs et al. of patients with non-malignant diseases indicate that RTC conditioning is effective and improves survival due to its lower toxicity [25]. However, Al Mulla et al. showed that reduced-toxicity conditioning was not significantly less toxic than myeloablative in pediatric and adolescent patients undergoing allogeneic HCT [26]. In our study, pediatric TRM score had no effect on TRM and OS, probably because the type of non-malignant disease was not included in the TRM stratification, and according to our analysis, the underlying disease affected survival after HCT.

This study has some limitations. The population was not homogeneous, as the underlying disease varied among the analyzed groups. The retrospective nature of the study, as well as the extended time of data collection and hence the changes of guidelines and standards of care during that time, could also have affected the results. Nevertheless, our results should be of value in investigating the causes of TRM in children with non-malignant diseases undergoing allogeneic HCT. Further research is needed to assess the effects of different risk factors on HCT outcome. In addition, prospective studies would allow gathering larger and more homogeneous populations of patients.

Predicting the outcome of HCT based on pretransplant factors is important for clinical practice. Despite the introduction of procedures that reduce the risk of TRM, it is necessary to assess such a risk individually before making a transplant decision.

Conclusions

In our study, SCID diagnosis, PB as stem cell source, and CD34+ dose higher than 5×106/kg were the most important risk factors for TRM in patients with non-malignant diseases treated with HCT, and they all appear to affect OS.