Primary graft failure following allogeneic hematopoietic stem cell transplantation: risk factors, treatment and outcomes

ABSTRACT

Objectives

Graft failure (GF) is an intractable complication of transplantation, which can severely affect the efficacy of the graft; however, the characteristics, incidence, and risk factors of primary GF have not been well described. This study aimed to analyze the risk factors and outcomes of primary GF to swiftly identify high-risk patients for GF.

Methods

We performed a case-control study with a case-control ratio of 1:4 with 869 patients who underwent allogeneic hematopoietic stem cell transplantation (allo-HSCT) between January 2015 and December 2019 at our center.

Results

Nineteen (2.19%) patients experienced primary poor graft function (PGF), while eleven (1.27%) patients developed primary graft rejection (GR). Univariate and multivariate logistic analyses identified two independent risk factors for primary PGF: splenomegaly [P = 0.030; odds ratio (OR), 3.486; 95% confidence interval (CI), 1.139 to 13.109], and donor type [non-matched sibling donor (non-MSD)] (P = 0.018; OR, 4.475; 95% CI, 1.289 to 15.537). However, only donor type (non-MSD) was statistically significant (P = 0.020; OR, 19.432; 95% CI, 1.595 to 236.691) for primary GR. The overall survival was significantly lower in the primary PGF (P = 0.001) and GR group (P = 0.000), respectively, compared to the control group.

Conclusion

GF can significantly affect the overall survival of patients who underwent allo-HSCT, despite its considerably low incidence. A human leukocyte antigen-matched sibling donor should be the first choice for patients undergoing allo-HSCT for the prevention of GF. Moreover, splenomegaly is an independent risk factor for PGF, and caution must be exercised while treating such patients.

KEYWORDS:

• Allogeneic hematopoietic stem cell transplantation
• poor graft function
• graft rejection
• risk factors
• treatment

Introduction

Allogeneic hematopoietic stem cell transplantation (allo-HSCT) is a curative treatment for patients with malignant or nonmalignant hematopoietic diseases. However, several intractable complications occurring after transplantation, including graft failure (GF), can severely affect the efficacy of the graft. The failure to achieve a threshold absolute neutrophil count (ANC) of 0.5 × 10⁹/L for three consecutive days by day 28 after HSCT is the relatively well-established criterion for GF. However, the relationship between poor graft function and GF varies among different sudies [1–5].

Recent studies [6–8] defined poor graft function (PGF) as at least bilinear severe cytopenia, which is characterized by the following: (1) ANC⩽0.5 × 10⁹/L; (2) platelet count ⩽20 × 10⁹/L; and (3) hemoglobin ⩽70 g/L for at least 3 consecutive days beyond day +28 after HSCT or the requirement of transfusion with full chimerism and a hypoplastic-aplastic bone marrow, while severe graft versus host disease (GVHD) and relapse were excluded. In this study, we classified GF according to the chimerism status: PGF referred to full chimerism, whereas graft rejection (GR) represented mixed chimerism or complete recipient chimerism [4,6,9,10]. GF is a life-threatening complication of HSCT, and the survival rate of patients with GF is considerably lower than that in patients with complete engraftment [3,11]. Moreover, it is also important to distinguish between primary GF and secondary GF, as the prognosis of primary GF is more dismal. Primary GF is defined as incomplete engraftment, whereas secondary GF indicates loss of initial engraftment. However, the pathogenesis of primary GF has yet to be elucidated, while the treatment options are very limited. Recent studies suggested that cell dose, human leukocyte antigen (HLA) disparity, blood-type mismatch, high serum ferritin level, splenomegaly, GVHD, and cytomegalovirus (CMV) infection may be related to primary GF [7,9,10,12,13].

Therefore, we performed a case–control study to analyze the risk factors and outcomes in patients with primary GF in order to identify those with a high risk of developing GF as early as possible and consequently ensure better management.

Patients and methods

Patients and definitions

A total of 869 patients who received allo-HSCT at the Hematopoietic Stem Cell Transplantation Center of Blood Diseases Hospital, Chinese Academy of Medical Sciences, between January 2015 and December 2019 were reviewed retrospectively. Nineteen of the 869 patients met the criteria for primary PGF and eleven were diagnosed with primary GR. A matched control group was selected using the case-pair method to analyze the risk factors based on the following criteria: (1) sex, (2) month in which the transplantation was received (±2 months), (3) age (± 5 years), and (4) a case-control ratio of 1:4. The final date of follow-up was June 30, 2020 for patients who did not experience any untoward events. All patients or their legal representatives provided written informed consent before transplantation. This study adhered to the Declaration of Helsinki and was approved by the Ethics Review Committee of our center.

GF was defined as the failure to achieve a threshold absolute neutrophil count (ANC) of 0.5 × 10⁹/L for three consecutive days by day 28 after HSCT. And GF was classified into PGF and GR according to the chimerism status: PGF referred to full chimerism, whereas graft rejection (GR) represented mixed chimerism or complete recipient chimerism.

The Mount Sinai Acute GVHD International Consortium (MAGIC) criteria was used to diagnose and grade acute GVHD (aGVHD) [14]. CMV (cytomegalovirus) -DNA was detected by plasma sample using real-time PCR and CMV infection was defined as >1000 copies/mL.

In this study, splenomegaly was defined as splenic thickness >4 cm or craniocaudal length >12 cm, as identified by ultrasonography [7].

Transplantation regimen

All patients undergoing allo-HSCT underwent myeloablative pre-conditioning. However, the pre-conditioning regimen was heterogenous for patients with different underlying diseases.

The principal regimen for patients diagnosed with acute myeloid leukemia (AML)/myelodysplastic syndrome (MDS)/myeloproliferative neoplasms (MPN) included busulfan and cyclophosphamide (Bu + Cy), in addition to optional fludarabine (Flu), cytarabine (Ara-c), or antithymocyte globulin (ATG). The regimen for patients with aplastic anemia (AA) included Flu + ATG + Cy, with optional administration of Bu. Patients with lymphoma/acute lymphocytic leukemia (ALL) received total body irradiation (TBI)/melphalan (Mel) + Cy, with the optional administration of Flu, Ara-c, and ATG.

All transplant recipients received cyclosporine A or FK506 and short-term methotrexate, in addition to the optional administration of mycophenolate mofetil, for the prophylaxis of GVHD.

Statistical analysis

Data were analyzed using the IBM software, SPSS statistics 25 and GraphPad Prism 8. The chi-squared test and Fisher’s exact test were used for categorical variables and descriptive statistics were used for continuous variables to compare the incidence of GF using univariate analysis. Parameters with P-values < 0.10 were regarded as potential risk factors and further analyzed using multivariate logistic regression. The Kaplan-Meier method was used to estimate the survival curves and differences were compared using the log-rank test. Two-sided P-values < 0.05 were considered statistically significant.

Results

Incidence and characteristics

A total of 869 patients who underwent allo-HSCT between January 1, 2015 and December 31, 2019 were enrolled in the incidence cohort. Nineteen (2.19%) of the 869 patients experienced primary PGF, while eleven (1.27%) developed primary GR.

The characteristics of patients with primary PGF and primary GR are listed in Table 1. Fourteen of the 19 patients with primary PGF were men and 5 were women, whereas 6 of the 11 patients with primary GR were men and 5 were women. The median ages of the patients with primary PGF and GR were 37 years (13–56) and 27 years (8–59), respectively. Thirteen patients in the primary PGF group were diagnosed with acute leukemia, 4 with MDS, 1 with lymphoma and 1 patient was diagnosed with MPN. Three patients in the primary GR group were diagnosed with acute leukemia, 3 with MDS, 3 with aplastic anemia, and 2 patients were diagnosed with MPN.

Table 1. Characteristics of primary PGF (A) and GR (B) patients.

A.
Number	Gender	Age	Disease	Donor type	Risk stratification	Disease state	MRD	splenomegaly	SF (ng/ml)	MNC (*10^8/kg)	CD34^+ (*10^6/kg)
1	Male	29	AML	MUD	Standard	PR	Positive	No	912.70	6.80	5.70
2	Male	45	AML	HRD	High	CR	Positive	No	1805.00	8.02	4.89
3	Male	39	AML	HRD	Standard	NR	Positive	Yes	1027.50	8.89	3.83
4	Female	41	MDS	MSD	Standard	/	/	No	1664.00	11.02	2.09
5	Male	24	ALL	HRD	Standard	CR	Negative	No	1138.30	7.27	3.35
6	Male	38	ALL	HRD	Standard	CR	Positive	No	793.30	7.44	4.58
7	Male	35	ALL	HRD	High	CR	Negative	Yes	913.50	13.45	2.13
8	Male	39	MDS	HRD	High	/	Positive	Yes	424.70	18.57	2.44
9	Male	13	MPN	HRD	Low	/	Positive	No	62.90	7.46	2.85
10	Male	56	MDS	HRD	High	/	Positive	No	1883.00	10.64	1.50
11	Male	37	AML	MSD	High	PR	Positive	Yes	1434.60	10.00	3.00
12	Male	35	AML	MSD	High	PR	Positive	Yes	2080.00	9.00	2.52
13	Female	33	AML	HRD	High	NR	Positive	Yes	943.00	8.74	3.23
14	Female	39	AML	HRD	High	CR	Negative	Yes	28.30	19.13	5.02
15	Male	53	Lymphoma	MUD	High	/	Positive	No	1922.00	6.24	2.99
16	Male	21	MDS	MSD	Standard	/	/	Yes	477.50	8.00	3.60
17	Female	27	AML	HRD	Standard	CR	Negative	No	1855.00	12.00	2.28
18	Female	39	AML	HRD	High	CR	Negative	Yes	384.10	8.00	4.40
19	Male	28	AML	MSD	Standard	CR	Negative	Yes	2469.00	7.89	2.38
B.
Number	Gender	Age	Disease	Donor type	Risk stratification	Disease state	MRD	splenomegaly	SF (ng/ml)	MNC (*10^8/kg)	CD34^+ (*10^6/kg)
1	Male	8	AA	HRD	/	/	/	No	304.50	10.00	2.10
2	Male	8	AA	HRD	/	/	/	No	465.40	6.79	3.51
3	Male	27	MDS	HRD	High	/	Positive	Yes	1414.70	13.46	2.02
4	Female	53	MPN	HRD	/	/	/	Yes	78.80	8.00	3.20
5	Male	26	MDS	HRD	Standard	/	/	No	4946.00	7.99	7.63
6	Male	17	ALL	MUD	Standard	CR	Negative	Yes	903.40	7.43	4.31
7	Female	31	ALL	MUD	High	CR	Negative	No	912.00	8.00	9.52
8	Female	45	MDS	HRD	Standard	CR	Negative	No	130.00	8.00	3.20
9	Female	59	MPN	HRD	/	/	/	Yes	180.80	4.35	3.83
10	Female	44	AML	MSD	High	NR	Positive	Yes	2039.00	7.77	3.11
11	Male	18	AA	HRD	/	/	/	No	911.10	8.00	2.48

MRD indicates minimal residual disease; SF, serum ferritin; MNC, mononuclear cell; AML, acute myeloid leukemia; ALL, acute lymphocytic leukemia; MDS, myelodysplastic syndrome; MPN, myeloproliferative neoplasm; AA, aplastic anemia; CR, complete remission; PR, partial remission; NR, No remission; HRD, HLA-haploidentical related donor; MUD, matched unrelated donor; MSD, matched sibling donor.

Risk factors

A control group with a 4:1 ratio for primary PGF and GR was used to identify the risk factors for PGF and GR, respectively. Univariate analysis revealed that the time from diagnosis to transplantation (>6 months), splenomegaly, SF (serum ferritin) level (>1000 ng/mL), donor type [non-matched sibling donor (non-MSD)], and blood-type-mismatch were potential risk factors (P < 0.10) for primary PGF (Table 2A). Univariate analysis for primary GR revealed different potential risk factors, including acute GVHD occurring in 30 days after transplantation, disease, and donor type (non-MSD) (Table 2B). Multivariate logistic analysis identified two independent risk factors for primary PGF, i.e. splenomegaly [P = 0.030; odds ratio (OR), 3.486; 95% confidence interval (CI), 1.139 to 13.109], and donor type (non-MSD) (P = 0.018; OR, 4.475; 95% CI, 1.289 to 15.537) (Table 3A). However, only donor type (non-MSD) was a statistically significant factor (P = 0.020; OR, 19.432; 95% CI, 1.595 to 236.691) associated with primary GR (Table 3B).

Table 2. Univariate analysis of risk factors for primary PGF (A) and GR (B).

A
Risk factors		PGF	Non-PGF	P
Disease	AL	13	57	0.671
	MDS	4	16
	Lymphoma	1	2
	MPN	1	1
Splenomegaly	Yes	10	18	0.013
	No	9	58
Donor type	Non-MSD	14	24	0.001
	MSD	5	52
Gender-mismatch	Yes	7	41	0.182
	No	12	35
Bloodtype-mismatch	Yes	13	26	0.007
	No	6	50
TBI in pre-condition	Yes	5	16	0.853
	No	14	60
GVHD prophylaxis	FK506	7	41	0.182
	CSA	12	35
SF(>1000 ng/ml)	Yes	10	23	0.080
	No	9	51
Time before transplantation	>6months	13	30	0.023
	≤6months	6	46
CMV infection in 30d	Yes	1	1	0.858
	No	18	75
aGVHD in 30d	Yes	7	14	0.155
	No	12	62
MNC (*10^8/kg)	Median (range)	8.74 (6.24–19.13)	8.59 (4.85–20.83)	0.791
CD34^+(*10^6/kg)	Median (range)	3.00 (1.5–5.7)	3.06 (1.80–7.50)	0.952
Donor age(years)	Median (range)	33 (14–60)	36 (8–60)	0.739
B
Risk factors		GR	non-GR	P
Disease	AL	3	27	0.083
	MDS	3	8
	AA	3	8
	MPN	2	1
Splenomegaly	Yes	5	8	0.132
	No	6	36
Donor type	Non-MSD	10	17	0.002
	MSD	1	27
Gender-mismatch	Yes	5	28	0.449
	No	6	16
Bloodtype-mismatch	Yes	3	16	0.832
	No	8	28
TBI in pre-condition	Yes	2	11	0.937
	No	9	33
GVHD prophylaxis	FK506	5	13	0.518
	CSA	6	31
SF(>1000 ng/ml)	Yes	3	16	0.832
	No	8	28
Time before transplantation	>6m	4	20	0.838
	≤6m	7	24
CMV infection in 30d	Yes	3	3	0.160
	No	8	41
aGVHD in 30d	Yes	5	7	0.087
	No	6	37
MNC (*10^8/kg)	Median (range)	8.00 (4.35–13.46)	8.00 (4.58–13.20)	0.413
CD34^+(*10^6/kg)	Median (range)	3.20 (2.02–9.52)	3.30 (2.09–7.24)	0.949
Donor age(years)	Median (range)	29 (15–53)	35.5 (11–56)	0.312

PGF indicates poor graft function; GR, graft failure; GVHD, graft versus host disease; aGVHD, acute GVHD; AL, acute leukemia; MDS, myelodysplastic syndrome; MPN, myeloproliferative neoplasm; MSD, matched sibling donor; TBI, total body irradiation; SF, serum ferritin; CMV, cytomegalovirus; MNC, mononuclear cell.

Table 3. Multivariate logistic analysis of risk factors for primary PGF (A) and GR (B).

A
Potential risk factors	P	OR	95% CI
Splenomegaly	0.030	3.486	1.139	13.109
Donor type(non-MSD)	0.018	4.475	1.289	15.537
Bloodtype-mismatch	0.132
SF(>1000 ng/ml)	0.153
Time before transplantation (>6months)	0.096
B
Potential risk factors	P	OR	95% CI
Donor type(non-MSD)	0.020	19.432	1.595	236.691
aGVHD in 30d	0.544
Disease	0.161

MSD indicates matched sibling donor; aGVHD, acute GVHD; SF, serum ferritin.

Treatment and outcomes

All patients with decreased blood cell counts received supportive treatment, including transfusion and administration of growth factors. Moreover, 2 patients from the primary PGF group received donor stem cell infusion and 1 patient received mesenchymal stromal cell infusion. The median follow-up period was 18.6 months (1–47 months), during which 11 patients (58%) died (1 died of relapse, while the others died of infection or severe GVHD). Eight patients in the primary GR group received donor stem cell infusion and 3 patients underwent a second transplantation; the donor was changed in 2 of these 3 patients because of complete rejection of the former graft. The median follow-up period was 16.1 months (3–65 months), and 7 patients (63.6%) died (1 died of relapse, while the others died of infection or severe GVHD). The overall survival was significantly lower in the primary PGF (P = 0.001) and GR group (P = 0.000), respectively (Figure 1A and B), compared to the control group.

Figure 1. Survival analysis of patients with PGF (A) and GR (B).

Discussion

GF is a rare complication of allo-HSCT, which is characterized by high mortality. Previous studies [3,6,13] reported that the incidence rate of primary GR and poor graft function ranged from 3.8% to 5.6%, and 5% to 27%, respectively. The incidence rate of primary PGF and GR in this study was 2.19% and 1.27%, respectively. As all patients treated in our center received a myeloablative pre-conditioning regimen, the prevalence of GF cannot be compared with that of previous studies.

It is imperative to identify the possible risk factors and take measures to prevent PGF and GR, since the overall survival of patients with primary PGF and GR was significantly lower than that in the control group (P = 0.001 and P = 0.000, respectively).

GR is thought to result from the recipient’s immune response against donor hematopoietic cells, which is mediated by the residual host immunity. Underlying disease, HLA disparity, low dose of infused CD34⁺ cells, conditioning regimen, major ABO blood group incompatibility, stem cell source, etc. have been reported to be amongst the risk factors associated with GR [15–18]. Mattsson J et al. revealed that the incidence rate of GR was 0.1% in patients who received grafts from HLA-identical siblings, whereas it rose to 5% in those who received HLA-mismatched grafts [16]. In our study, non-MSD was an independent risk factor, consistent with the findings of previous studies.

According to previous studies, the risk factors for PGF include low dose of infused CD34⁺ cells, CMV infection, acute GVHD, donor-specific antibody, iron overload, splenomegaly etc [6,7,19,20]. The current study identified donor type (non-MSD) and splenomegaly as independent risk factors for PGF.

Splenomegaly was an independent risk factor for primary PGF in our study, concurrent with recent studies [5,7,21]. The entrapment and destruction of hematopoietic cells by the enlarged spleen could be a possible mechanism underlying the development of PGF. Shimomura et al. [21] suggested that splenic enlargement (>320 cm³) was linked to low neutrophil and platelet engraftment after allo-HSCT in patients with AML and MDS. Moreover, the persistence of significant splenomegaly (⩾10 cm under the costal margin) on day +30 following HSCT was associated with a higher incidence of PGF (33% versus 12%; P = 0.05) in patients with myelofibrosis [5]. Akpek et al. [22] suggested that splenomegaly would not delay engraftment if a CD34⁺ cell dose >5.7 × 10⁶/kg was transplanted, indicating that higher doses of CD34⁺ cells should be considered for patients with an enlarged spleen. However, in our study, the CD34⁺ cell dose did not differ significantly between the PGF and good graft function group, possibly because the majority of CD34⁺ cell doses administered in our center were lower than 5.7 × 10⁶/kg and few patients can attain this dose. Another independent risk factor was donor type (non-MSD), which was also an independent risk factor for GR, suggesting that an HLA-matched sibling donor should be the first choice for the prevention of GF.

Treatment for patients with GF traditionally primarily consists of supportive treatment, donor cell infusion, and second allo-HSCT. New strategies include CD34⁺ stem cell boosts, mesenchymal stem cells, thrombopoietin receptor agonists, and drugs that can improve the impaired bone marrow microenvironment [6,7,23–27]. Although the prognosis of GF has improved owing to the advancement in treatment methods, it is still a lethal complication of allo-HSCT.

In conclusion, this study demonstrated that primary GF was a rare but severe complication of allo-HSCT. Our study provided some clinically relevant indications, despite limitations such as its retrospective design and small sample size (due to the low incidence of primary GF). Selecting an HLA-matched sibling donor is a better option for patients undergoing allo-HSCT in order to prevent GF, although other types of donor may confer other benefits. Moreover, splenomegaly is a risk factor for PGF, and clinicians should take measures to manage splenomegaly before transplantation. It is vital to explore effective treatment strategies since the prognosis of primary GF is still dismal.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the CAMS Innovation Fund for Medical Sciences (CIFMS) [grant numbers 2021-1-I2M-017 and 2021-I2M-C&T-B-080].