Objectives: Graft-versus-host disease is still one of the most important complications after hematopoietic stem cell transplantation. The risk factors remain unclear, with effects of graft-versus-host disease on survival varying among different centers. We aimed to determine risk factors that may affect development of graft-versus-host disease and the corresponding patient survival rates at a single pediatric hemato-poietic stem cell transplant unit.
Methods: Our study included 104 of 118 pediatric patients who underwent allogeneic hematopoietic stem cell transplant at our institute between 2005 and 2018. Patient characteristics, clinical information, pretransplant and posttransplant factors, and laboratory parameters were obtained from the database.
Results: Acute graft-versus-host disease was seen in 19 pediatric patients. Chronic graft-versus-host disease, which was seen in 13 of our pediatric study patients, occurred more often in those with peripheral blood stem cell than in those with bone marrow transplant (odds ratio of 9.969; 95% CI, 1.040-95.547; P = .046). Female donor-to-male recipient transplant was significantly associated with incidence of chronic graft-versus-host disease (odds ratio of 8.51; 95% CI, 1.323-54.843; P = .024). Later neutrophil engraftment was associated with incidence of acute graft-versus-host disease (odds ratio of 1.107; 95% CI, 1.012-1.212; P = .02).
Conclusions: Although there are some known risk factors for graft-versus-host disease in adult patients, little is known about risk factors in children. In our comprehensive study in pediatric patients, we found that peripheral blood stem cell transplant, female-to-male transplant, and later neutrophil engraftment were associated with incidence of graft-versus-host disease. Although peripheral blood as a source of stem cells and female-to-male transplant are known risk factors, later neutrophil engraftment was a new finding as a possible risk factor for acute graft-versus-host disease. This finding requires further verification in future prospective studies.
Key words : Children, Neutrophil engraftment, Stem cell source
Graft-versus-host disease (GVHD) is still one of the most common complications despite continued improvements in outcomes after hematopoietic stem cell transplantation (HSCT). It remains as a major determinant of early morbidity and mortality.1,2 Clinically significant acute GVHD (aGVHD) has been shown to effect from 30% to 50% of allogeneic HSCT recipients.3 Furthermore, chronic GVHD (cGVHD) has been shown in 30% to 70% of HSCT recipients who survived 100 days.4 The most common risk factors for aGVHD include use of HLA mismatched donors and HLA matched unrelated donors (MUD), older patient age, female donor-to-male recipient transplants, prior alloimmunization of the donor, and type of GVHD prophylaxis.5
Manifestations of cGVHD include a wide spectrum of clinical findings, including skin, pulmonary, oral mucosal, and eye involvement.6,7 It is still unclear why GVHD develops in severe forms in some patients but not in others. Proper management of GVHD could be performed by identifying the individual risk factors for GVHD.8 Although some risk factors associated with the development of GVHD have been determined,1,2,5 the incidence and risk factors of aGVHD still vary among different clinical centers.
There have been differing reports on survival rates in pediatric HSCT recipients, with differences depending on factors like underlying diseases and presence and grade of GVHD. Horn and associates9 reported that nonmalignant patients with aGVHD grade III/IV had significantly lower 3-year survival rates (52.9%) than those without aGVHD (90.1%) or those with aGVHD grade I/II (98.1%). Patients without cGVHD and those with limited and extensive cGVHD had 3-year survival rates of 88.9%, 91.7%, and 84.8%, respectively.9 In a report from Inagaki and colleagues, the probability of disease-free survival of pediatric patients with cGVHD after allogeneic HSCT at 5, 10, and 15 years was 69.2%, 66.6%, and 58.5%, respectively.10
Because data on GVHD risk factors in children are limited, we aimed to investigate any pretransplant and posttransplant risk factors that could have an impact on the development of GVHD among pediatric allogeneic HSCT recipients and the survival rates of these patients with GVHD at a single pediatric HSCT unit.
Materials and Methods
This study was approved by our local ethics committee in 2017 in accordance with the Declaration of Helsinki.
Acute GVHD was classified based on clinical criteria,11 and cGVHD was evaluated according to National Institutes of Health criteria and classified as classic cGVHD or overlap syndrome. The scoring of severity of cGVHD was done for each involved organ using the National Institutes of Health criteria and global severity.6,12
Neutrophil engraftment is most commonly defined as the first of 3 consecutive days of achieving a sustained peripheral blood (PB) neutrophil count of >500 × 106/L.13 Platelet engraftment is usually defined as independence from platelet transfusion for at least 7 days with a platelet count of more than >20 × 109/L.14
There were 118 patients who underwent allogeneic HSCT for malignant or nonmalignant disease at our center between 2005 and 2018. The patient characteristics, clinical information, pretransplant and posttransplant factors, and laboratory parameters were retrieved from medical records. All patients were engrafted in our study group. There were 14 patients who died early after HSCT, before evaluation for cGVHD. Eleven of these patients died as a result of primary disease progression (78.6%), 2 died as a result of HSCT complication (14.3%; 1 with veno-oclusive disease and 1 with pulmonary embolism), and 1 patient died as a result of infectious complication (7.1%). Patients who died before evaluation of cGVHD within 3 months after HSCT were excluded from our statistical analyses.
Patients with grade ≥II GVHD were included in the aGVHD group; all other patients, including grade I GVHD, who had engraftment were accepted as controls (no aGVHD group). With these conside-rations, there were 104 patients included in the statistical analyses.
Until 2017 at our center, HSCT was performed from only matched sibling donors (MSD) and matched related donors (MRD). After HLA genotyping was assessed, a compatible donor was accepted when there was 6/6 compability for MSDs and 9/10 and 10/10 compability for MRDs. Since 2017, MUD HSCT procedures with 9/10 or 10/10 compability have also been carried out.
Preparative regimen selection was based on disease diagnosis, disease status, recipient age, and comorbidities at the discretion of the treating physician. Patients were given busulphan or total body irradiation-based myeloablative therapy or fludarabine plus cyclophosphamide with or without antithymocyte globulin (ATG)-based nonmye-loablative therapy,15 according to the diagnosis. Graft-versus-host disease prophylaxis included cyclosporine for MSD,16,17 cyclosporine plus short-term methotrexate for MRD, and cyclosporine plus short-term methotrexate and ATG for MUD. Prophylactic defibrotide was given to patients who were under 7 years old and according to the diagnosis, as described elsewhere.18
In general, systemic treatment with steroids (1-2 mg/kg) is initiated for aGVHD of grade II or more and for newly diagnosed cGVHD in our center. For patients with steroid resistance, second-line therapies were administered according to the site and grade of GVHD as well as the availability of the therapy.19
Statistical analysis was performed using the Statistical Package for Social Sciences software (version 22.0). Categorical variables are presented as numbers and percentages, whereas numerical variables are presented as means ± standard deviation or median (minimum to maximum) according to distribution of normality. Demographic and clinical characteristics of patients were analyzed with the use of chi-square and Mann-Whitney U tests. These tests were used to analyze the relationship between each clinical characteristic and aGVHD and cGVHD individually. Regression analysis was used to evaluate the factors that influence the development of aGVHD or cGVHD together to eliminate confounding factors. The factors that were tested in regression models were donor age, donor sex, body mass index (BMI; calculated as weight in kilograms divided by height in meters squared), presence of sinusoidal obstruction syndrome, total parenteral nutrition (TPN), graft manipulation, absolute neutrophil and lymphocyte counts (on the day of HSCT), serum albumin level on the day of HSCT, and BK viremia. Survival analysis was done with the use of Kaplan-Meier estimates. P < .05 was considered as statistically significant.
There were 118 pediatric patients who underwent allogeneic HSCT (45 female, 73 male). Median age at transplant was 7.82 years (range, 0.29-17.49 years). Among all 118 patients, there were 40 patients with nonmalignant diseases, including thalassemia (n = 19), bone marrow failure (n = 16), and metabolic diseases(n = 5; including 2 with adrenoleukodystrophia, 1 with mucopolysaccharidosis, 1 with osteopetrosis, and 1 with Wolman disease). There were 78 patients with malignancies at diagnosis, including 35 with acute lymphoblastic leukemia, 25 with acute myeloblastic leukemia, 3 with myelodysplastic syndromes, 1 with hemophagocytic lymphohistiocytosis, 9 with juvenile myelomonocytic leukemia, and 5 with other malignancies (Table 1). Fourteen of the 118 patients died before evaluation for GVHD. Median follow up was 30.5 months (range, 0.39-151 months).
Acute GVHD was seen in 19 of our pediatric patients, and cGVHD was seen in 13 of our pediatric patients. Acute GVHD grades were as follows: 12% with grade I, 64.0% with grade II, 16.0% with grade III, and 8% with grade IV. The most common organ involvement was skin (60.0%), followed by liver (4.0%) and intestinal (16.0%) involvement. There were also mixed (20.0%) organ involvements (ie, having any 2 of the 3 organ involvements mentioned above). Chronic GVHD grades were as follows: 46.1% with mild, 38.5% with moderate, and 15.4% with severe grade.
Median recipient age was 8.33 years (range, 0.92-17.49 years) for the study group. Younger recipient age was related to higher aGVHD grade(P = .04). We found no significant association between recipient age and presence of cGVHD (P = .98; Table 1).
Median donor age was 14 years (0.5-47 years) for the study group. We found a significant association between older donor age and higher cGVHD grade (P = .01). However, we found no significant association between donor age and presence of aGVHD (P = .29; Table 1).
Type of transplant
There were 77 HSCTs from MSD, 17 from MRD, 9 from MUD, and 1 from haploidentical donors (Table 1). We observed significantly more patients with cGVHD in the MRD group than in the MSD group (P = .01). However, there was no association between type of transplant and presence of aGVHD (Table 1).
There were 77 patients who had cyclosporine-based GVHD prophylaxis, 17 patients who had cyclosporine plus methotrexate, and 7 patients who had cyclosporine plus methotrexate plus ATG-based GVHD prophylaxis. Incidence of cGVHD was significantly higher in the cyclosporine plus methotrexate plus ATG group than in either the cyclosporine plus methotrexate or the cyclosporine groups (P = .01). However, there was no association between type of prophylaxis and incidence of aGVHD (Table 1).
Stem cell source
Stem cell source was PB for 42 patients and bone marrow for 62 patients. There were more patients with cGVHD in the PB group than in the bone marrow group (P = .004; Table 1).
Median neutrophil engraftment was on day 17 (range, day 11-47) for patients with aGVHD and on day 15 (range, day 10-33) for patients without aGVHD, with median neutrophil engraftment occurring later in patients with aGVHD than in patients without aGVHD (P = .048). Median neutrophil engraftment was on day 14 (range, day 12-21) for patients with cGVHD and day 16 (range, day 10-47) for patients without cGVHD; time of neutrophil engraftment did not differ between those with and without cGVHD (P = .17; Table 1).
Median thrombocyte engraftment day was on day 22 (range, day 6-108) for patients with aGVHD and day 20 (range, day 12-64) for patients without aGVHD. There were no significant differences between patients with and without aGVHD with regard to median thrombocyte engraftment day (P = .39; Table 1).
Female donor-to-male recipient transplantation
There were 26 HSCT procedures that were performed from a female donor to a male recipient. Female-to-male transplant did not affect the development of aGVHD (P = .30); however, it seemed to be a risk factor for cGVHD. In patients with cGHVD, 53.8% had female-to-male transplant procedures (P = .01; Table 1).
In multiregression analysis, later neutrophil engraftment day (odds ratio [OR] = 1.107; 95% CI, 1.012-1.212; P = .02) was related to incidence of aGVHD (Table 2). Our multiregression analysis also showed that PB stem cell transplant procedures were significantly associated with incidence of cGVHD (OR = 9.969; 95% CI, 1.040-95.547; P = .046) and female-to-male transplant was also significantly associated with incidence of cGVHD (OR = 8.51; 95% CI, 1.323-54.843; P = .024) (Table 2).
Patients with grade I GVHD were excluded from our survival analysis.
One-year and 5-year overall survival rates in patients with aGVHD were 61.1% and 55.0%, respectively. In patients with cGVHD, 1-year, 3-year, and 5-year overall survival rates were 92.3%, 83.1%, and 71.2%, respectively. There was no statistical difference in survival rates between patients with aGVHD versus patients with cGVHD (P = .08; Figure 1, top).
Prior history of aGVHD was observed in 6 patients (46.1%) who had cGVHD. Prior aGVHD was significantly related to higher mortality in cGVHD patients, with 1-year, 3-year, and 5-year overall survival rates in patients with cGVHD and history of prior aGVHD of 100.0%, 62.5%, and 31.0%, respectively. In contrast, the 5-year overall survival rate of patients with cGVHD without prior aGVHD was 100% (P = .02; Figure 1, bottom).
Graft-versus-host disease is one of the most common and most lethal complications of HSCT.20 Studies on risk factors of GVHD have increased during the past decades, but there are still few studies in children.2 Because of the significant variations in practices, underlying diagnosis, conditioning regimens, GVHD prophylaxis, and other patient characteristics, discrepancies are still present among the risk factors of GVHD in the literature.21 In this study, we aimed to determine the risk factors of GVHD in pediatric patients seen at our center. These data are important to predict the pediatric patients who may be at increased risk of GVHD and to build strategies to manage GVHD for each center individually.
Receiving an allograft from a female donor has been already shown to be related to an increased risk of GVHD. This is a well-known issue due to the antibodies that could be present in females.22 We also observed that female-to-male transplant was significantly related to cGVHD.
Lee and colleagues stated that younger recipient age, acute leukemia versus other malignancies, and unrelated donors were related to aGVHD in their adult study.23 A large study by Flowers and colleagues showed that HLA mismatched or unrelated donors, the use of total body irradiation, and a female donor for a male recipient were significantly associated with an increased risk of aGVHD in adult and pediatric patients who underwent allogeneic HSCT after myeloablative conditioning.24 In both of these studies, PB source was not found to be a risk factor for GVHD. In our univariate analyses, we found that PB stem cell as a source, MRD and MUD transplants, and older age of donors were other significant risk factors for cGVHD.
Our multivariate analysis showed that PB stem cells as a source and female-to-male transplant were associated with higher incidence of cGVHD and that later neutrophil engraftment day was associated with higher incidence of aGVHD. Although PB transplant and female-to-male transplant are well known risk factors, later neutrophil engraftment day may be a new finding. Periengraftment clinical problems like capillary leakage, hepatic and renal dysfunction, and fever are known to be associated with cellular and cytokine interactions.25 Both the so-called cytokine storm resulting from the toxicity of conditioning regimens and T-cell activation leading to GVHD may have acted together with neutrophil recovery. However, the mechanism needs to be defined more clearly in future studies.
With regard to the effect of PB cells on aGVHD, previous studies have shown inconsistent findings.26,27 One study reported that use of PB as the source of stem cells was an important risk factor for aGVHD with the impact of the intensity of conditioning and the use of total body irradiation.20 Interestingly, we found that PB was a risk factor for cGVHD; however, there was no difference between presence or absence of aGVHD versus PB stem cell transplant.
Studies have shown that low or high BMI may be related to poor HSCT outcomes.28,29 Paviglianiti and colleagues stated that patients with BMI of less than the fifth percentile had a higher incidence of grade II to IV aGVHD compared with patients with normal BMI.29 In addition, BMI ≥35 was found to be a risk factor for increased mortality after allogeneic HSCT. In a study from Asslan and colleagues, a pretransplant high BMI tended to be associated with an increased risk of grade II to IV aGVHD (P = .07).30 Our study showed no significant differences between normal BMI and abnormal BMI groups in terms of development of GVHD (P = .8). There were also no significant differences in terms of BMI and incidence of aGVHD or cGVHD (P = .59, P = .34).
Clinicians use TPN as an adjunctive therapy during transplant in up to 92% of patients, and it seems to improve long-term survival in transplant recipients. Neither the best time to start TPN nor the best composition of nutritional substrates and supplements is clear.31 In some studies, it is estimated that enteral nutrition could provide benefits over TPN, including reducing the incidence of aGVHD and faster platelet engraftment.32,33 According to a study by Evans and colleagues, incidences of gastrointestinal GVHD of any stage and aGVHD greater than grade I were increased in patients who had received TPN.32 One note from our study was no significant difference was observed in incidence of aGVHD (P = .15) or cGVHD (P = .98) versus TPN.
Storb and colleagues reported that there were significant associations between high lymphocyte count just before HSCT and reduced risk of relapse and overall mortality but no association with risk of GVHD or nonrelapse mortality.34 Similarly, we found that there was no relation between aGVHD or cGVHD and lymphocyte counts immediately before HSCT (on the day of HSCT). We also did not find an association between neutrophil count and GVHD, although a slightly slower neutrophil increment was significantly associated with aGVHD, as discussed above.
Our analysis showed no difference in survival rates between the aGVHD and cGVHD groups. However, prior history of aGVHD was significantly related to higher mortality in cGVHD patients. Because the number of patients was low, our aim was not primarily to analyze this finding; however, we included survival rates to show patient outcomes for our center.
There are certain risk factors for aGVHD and cGVHD that have been mostly revealed in adult patients; however, published results in pediatric patients are scarce. Although our study had limitations, such as the small number of patients and the retrospective design, this is still a comprehensive study performed in pediatric patients. Similarly to that reported previously, PB as a stem cell source and female-to-male transplant were found to be related to cGVHD occurrence. However, we also observed that later neutrophil engraftment was associated with incidence of aGVHD. Although this finding should be confirmed with new prospective studies, patients who have later neutrophil engraftment should be monitored more carefully for aGVHD and may be given more intensive prophylaxis.
DOI : 10.6002/ect.2021.0157
From the 1Department of Pediatric Infectious Disease, the 2Department of Pediatrics, the 3Department of Pediatric Hematology and the 4Department of Public Health, Gazi University, School of Medicine, Ankara, Turkey
Acknowledgements: The authors have not received any funding or grants in support of the presented research or for the preparation of this work and have no declarations of potential conflicts of interest.
Corresponding author: Tu?ba Bedir Demirda?, Gazi University, School of Medicine, Department of Pediatric Infectious Disease, Ankara, Turkey
Table 1. Comparison of Patients With and Without Acute or Chronic Graft-Versus-Host Disease
Table 2. Multivariate Logistic Regression Analyses of Variables to Predict Graft-Versus-Host Disease in Pediatric Hematopoietic Stem Cell Transplant Recipients
Figure 1. Survival Analysis