Mixed chimaerism is associated with poorer long-term failure-free survival among aplastic anaemia patients receiving HLA-matched donor transplantation

TO THE EDITOR: Allogeneic haematopoietic stem cell transplantation (allo-HSCT) from HLA-matched related donors (MRDs) has been suggested as a standard first-line choice for younger severe aplastic anaemia (SAA) patients, with a long-term overall survival (OS) of approximately 90% [1–3]. Allo-HSCT from unrelated donors (URDs) has also achieved inspiring survival outcomes [2, 4]. However, long-term failure-free survival (FFS) has not been well evaluated in the above studies. Mixed chimaerism is a controversial status in patients with nonmalignant disease after receiving allo-HSCT [5]. An increasing number of studies have indicated that mixed chimaerism is an unsettling status that leads to cytopenia or graft rejection [6–8]. Regarding the effect of MC on OS outcomes, the reported studies have demonstrated conflicting conclusions. In a cohort of 207 SAA patients, MC was associated with worse OS than full donor chimaerism (FDC) [8]. Inconsistently, earlier studies have indicated that OS was similar between MC and FDC groups [9, 10]. Nevertheless, none of the published studies have evaluated the association between mixed chimaerism and FFS. Here, we performed a retrospective analysis to assess the adverse effect of MC and clarify a predictive model for MC among SAA patients receiving HLA-matched donor transplantation. A total of 103 SAA patients were enrolled in the analysis. Genetic etiology was ruled out among all patients and they were diagnosed as “acquired aplastic anaemia”. The donor type was HLA-matched related in 79 patients and HLA-matched unrelated in 24 patients. The 10-year probabilities of OS and FFS were 87.1 ± 3.8% vs. 87.5 ± 6.8% (P= 0.965) and 76.3 ± 4.9% vs. 79.2 ± 8.3% (P= 0.836) after MRD and MUD transplants, respectively. The respective characteristics of the patients with mixed chimaerism (MC) or FDC are summarized in Table 1. Patients with MC were significantly younger than the patients with FDC (P= 0.017). The median pre-SCT ferritin level was higher in the patients who developed mixed chimaerism (P= 0.021). Regarding donor-recipient blood type, the proportion of major and minor mismatch pairs was higher (25.0% vs. 8.4%, P= 0.038) among the patients with MC than among those with FDC. The cumulative incidences of mixed chimaerism were 17.8 ± 0.2% and 25.0 ± 0.8% in the MRD and MUD cohorts, respectively (P= 0.432, Supplementary Fig 1). In the multivariate analyses, patient age (<18 vs. ≥18 years), preSCT ferritin levels (<2500 vs. ≥2500 ng/ml), donor-recipient blood type (major and minor mismatch vs. others) and donor type (MRD or MUD) were included in the analysis. Younger patient age and higher levels of pre-SCT ferritin were associated with the development of mixed chimaerism, but blood type was not. The incidence of MC was significantly higher in children (HR [95% CI] 2.575 [1.046–6.340], P= 0.040) and higher in patients with higher levels of ferritin (HR [95% CI] 3.078 [1.209–7.833], P= 0.018). The estimated 10-year OS was also not significantly different in the MC cohort (80.0 ± 8.9%) vs. the FDC cohort (88.9 ± 3.5%) (P= 0.327, Supplementary Fig 2). The patients with MC had significantly poorer probabilities of 10-year FFS than did those with FDC (35.0 ± 10.7% vs. 87.0 ± 3.9%, P < 0.001, Supplementary Fig. 3). The incidences of acute GvHD were similar between the patients in the MC and FDC cohorts (grade II-IV aGvHD of 0% vs. 9.6 ± 0.1%, P= 0.151, Supplementary Fig 4; grade III-IV aGvHD of 0% vs. 2.4 ± 0.1%, P= 0.485). The patients with FDC had a 6.3 ± 0.1% incidence of chronic GvHD, whereas none of the patients with MC experienced cGvHD (P= 0.251). The treatment and clinical course of mixed chimaerism are shown in Supplementary Fig 5. As mentioned previously, patient age and ferritin level have been identified as predictive factors for developing mixed chimaerism. In an attempt to identify those patients at high risk for developing mixed chimaerism, we further established a predictive risk model as low risk (0 factor, N= 56), intermediate risk (1 factor, N= 39) and high risk (2 factors, N= 8). As shown in Supplementary Fig 6, the respective incidences of mixed chimaerism were 5.4 ± 0.1%, 30.8 ± 0.6% and 62.5 ± 3.7% in the low-, intermediateand high-risk groups, respectively (P < 0.001). The FFS at the final follow-up was 83.6 ± 5.0%, 75.3 ± 7.4% and 37.5 ± 17.1% in the low-, intermediateand high-risk groups, respectively (P= 0.018, Supplementary Fig 7). As shown in Supplementary Fig 8, the OS didn’t differ among different groups. As reported previously, the association of mixed chimaerism and OS was inconsistent, but none of the studies have explored whether MC was related to FFS. An earlier study from Seattle showed that OS at 5 years did not differ between the MC (55%) and FDC (62%) cohorts among SAA patients [9]. Lawler et al. published their experience with serial chimeric status analysis indicated that progressive mixed chimaeras (PMCs) were at high risk of late graft rejection, and was also a poor prognostic indicator of survival [5]. Our previous study also reported that MC was associated with significantly higher secondary graft rejection rates (14.8% vs. 0.4%) and poorer OS (72.7% vs. 89.6%) than those of the FDC cohort among SAA patients [8]. The current analysis supported that SAA patients with MC had obviously poorer 10-year FFS than those with FDC. Thus, decreasing the occurrence of mixed chimaerism is necessary even in nonmalignant diseases.


TO THE EDITOR:
Allogeneic haematopoietic stem cell transplantation (allo-HSCT) from HLA-matched related donors (MRDs) has been suggested as a standard first-line choice for younger severe aplastic anaemia (SAA) patients, with a long-term overall survival (OS) of approximately 90% [1][2][3]. Allo-HSCT from unrelated donors (URDs) has also achieved inspiring survival outcomes [2,4]. However, long-term failure-free survival (FFS) has not been well evaluated in the above studies.
Mixed chimaerism is a controversial status in patients with nonmalignant disease after receiving allo-HSCT [5]. An increasing number of studies have indicated that mixed chimaerism is an unsettling status that leads to cytopenia or graft rejection [6][7][8].
Regarding the effect of MC on OS outcomes, the reported studies have demonstrated conflicting conclusions. In a cohort of 207 SAA patients, MC was associated with worse OS than full donor chimaerism (FDC) [8]. Inconsistently, earlier studies have indicated that OS was similar between MC and FDC groups [9,10]. Nevertheless, none of the published studies have evaluated the association between mixed chimaerism and FFS. Here, we performed a retrospective analysis to assess the adverse effect of MC and clarify a predictive model for MC among SAA patients receiving HLA-matched donor transplantation.
A total of 103 SAA patients were enrolled in the analysis. Genetic etiology was ruled out among all patients and they were diagnosed as "acquired aplastic anaemia". The donor type was HLA-matched related in 79 patients and HLA-matched unrelated in 24 patients. The 10-year probabilities of OS and FFS were 87.1 ± 3.8% vs. 87.5 ± 6.8% (P = 0.965) and 76.3 ± 4.9% vs. 79.2 ± 8.3% (P = 0.836) after MRD and MUD transplants, respectively.
The respective characteristics of the patients with mixed chimaerism (MC) or FDC are summarized in Table 1. Patients with MC were significantly younger than the patients with FDC (P = 0.017). The median pre-SCT ferritin level was higher in the patients who developed mixed chimaerism (P = 0.021). Regarding donor-recipient blood type, the proportion of major and minor mismatch pairs was higher (25.0% vs. 8.4%, P = 0.038) among the patients with MC than among those with FDC. The cumulative incidences of mixed chimaerism were 17.8 ± 0.2% and 25.0 ± 0.8% in the MRD and MUD cohorts, respectively (P = 0.432, Supplementary Fig 1).
In the multivariate analyses, patient age (<18 vs. ≥18 years), pre-SCT ferritin levels (<2500 vs. ≥2500 ng/ml), donor-recipient blood type (major and minor mismatch vs. others) and donor type (MRD or MUD) were included in the analysis. Younger patient age and higher levels of pre-SCT ferritin were associated with the development of mixed chimaerism, but blood type was not. The incidence of MC was significantly higher in children (HR [95% CI] 2.575 [1.046-6.340], P = 0.040) and higher in patients with higher levels of ferritin (HR [95% CI] 3.078 [1.209-7.833], P = 0.018).
The estimated 10-year OS was also not significantly different in the MC cohort (80.0 ± 8.9%) vs. the FDC cohort (88.9 ± 3.5%) (P = 0.327, Supplementary Fig 2). The patients with MC had significantly poorer probabilities of 10-year FFS than did those with FDC (35.0 ± 10.7% vs. 87.0 ± 3.9%, P < 0.001, Supplementary Fig. 3). The incidences of acute GvHD were similar between the patients in the MC and FDC cohorts (grade II-IV aGvHD of 0% vs. 9.6 ± 0.1%, P = 0.151, Supplementary Fig 4; grade III-IV aGvHD of 0% vs. 2.4 ± 0.1%, P = 0.485). The patients with FDC had a 6.3 ± 0.1% incidence of chronic GvHD, whereas none of the patients with MC experienced cGvHD (P = 0.251). The treatment and clinical course of mixed chimaerism are shown in Supplementary Fig 5. As mentioned previously, patient age and ferritin level have been identified as predictive factors for developing mixed chimaerism. In an attempt to identify those patients at high risk for developing mixed chimaerism, we further established a predictive risk model as low risk (0 factor, N = 56), intermediate risk (1 factor, N = 39) and high risk (2 factors, N = 8). As shown in Supplementary Fig 6, the respective incidences of mixed chimaerism were 5.4 ± 0.1%, 30.8 ± 0.6% and 62.5 ± 3.7% in the low-, intermediate-and high-risk groups, respectively (P < 0.001). The FFS at the final follow-up was 83.6 ± 5.0%, 75.3 ± 7.4% and 37.5 ± 17.1% in the low-, intermediate-and high-risk groups, respectively (P = 0.018, Supplementary Fig 7). As shown in Supplementary Fig 8, the OS didn't differ among different groups.
As reported previously, the association of mixed chimaerism and OS was inconsistent, but none of the studies have explored whether MC was related to FFS. An earlier study from Seattle showed that OS at 5 years did not differ between the MC (55%) and FDC (62%) cohorts among SAA patients [9]. Lawler et al. published their experience with serial chimeric status analysis indicated that progressive mixed chimaeras (PMCs) were at high risk of late graft rejection, and was also a poor prognostic indicator of survival [5]. Our previous study also reported that MC was associated with significantly higher secondary graft rejection rates (14.8% vs. 0.4%) and poorer OS (72.7% vs. 89.6%) than those of the FDC cohort among SAA patients [8]. The current analysis supported that SAA patients with MC had obviously poorer 10-year FFS than those with FDC. Thus, decreasing the occurrence of mixed chimaerism is necessary even in nonmalignant diseases.
In our current analysis, the predictive role of pre-SCT transfusions, the interval from diagnosis to transplantation, and the disease severity was not found. We observed that patient age and pre-SCT ferritin level were predictive adverse factors for the occurrence of mixed chimaerism among SAA patients who received HLA-matched donor transplantation. The association of patient age with mixed chimaerism has not been described in these studies [5][6][7][8], which included not only HLA-matched transplants. In addition, the predictive role of pre-SCT ferritin was inconsistent with previously published literature. Shenoy et al. indicated that the ferritin level with a cut-off of 1500 ng/ml could not predict the development of lateonset mixed chimaerism but potentially affected the occurrence of graft rejection [11]. It is important to note that, different from our current study, thalassemia was the disease that was focused on in Shenoy's article, and the evaluation of mixed chimaerism was at 4 years after transplantation.
The reported mixed chimaerism rates following HLA-matched donor transplantation for SAA varied from 20% to 35% with the conditioning regimen of cyclophosphamide (Cy)/antithymocyte globulin (ATG) or fludarabine (Flu)/Cy/ATG [8]. Subsequently, a novel protocol of adding 3.2 mg/kg busulfan (Bu) has been demonstrated to decrease the risk of MC from 27% (Cy/ATG, historic cohort) to 0% (Bu/Cy/ATG, study cohort) [12]. Thus, it is essential to elucidate the risk factors for developing MC and identify high-risk patients who may benefit from a more intensive regimen.
The main limitation of this study is its retrospective nature from a single center, and the predictive model needs to be further validated in a prospective cohort. However, based on a decade of experience, our analysis has shown that mixed chimaerism led to poor FFS among SAA recipients post HLA-matched donor transplantation. The predictive model is beneficial to identify the high-risk subgroup population, and the application of more intensive conditioning will further improve the prognosis among high-risk patients.  ✉ email: huangxiaojun@bjmu.edu.cn

DATA AVAILABILITY
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.