Incidence and Factors Associated with Graft-Versus-Host Disease in the First Year After Allogeneic Peripheral Blood Stem Cell Transplantation

Allogeneic peripheral blood stem cell transplantation (allo-SCT) is an established treatment for controlling or curing a variety of hematologic disorders and some metabolic and autoimmune diseases and has been increasingly performed in Colombia and Latin America in recent years.^[1] Colombia ranks fourth in Latin America in terms of the number of transplants performed on the basis of the Global Transplant Activity Survey.^[2] Compared with nonmatched donors, transplants from human leukocyte antigen (HLA) matched-related donors (MRDs) (identical donors) have been associated with more favorable outcomes in terms of overall survival, incidence of relapse, and nonrelapse mortality.^[3,4] Haploidentical donors offer patients without a matched donor the opportunity to receive potentially curative treatment and have become a more affordable source in Latin America than have matched unrelated donors, resulting in increased use in recent years.^[2,3,5]

Graft-versus-host disease (GVHD) remains one of the major complications of allo-SCT, causing significant morbidity and accounting for up to 20% of transplant-related mortality.^[4,5] Acute and chronic GVHD incidence ranges from 49–62% and 30–70%, respectively.^[4–7] GVHD is a clinical condition caused by transplanted donor cells reacting to histocompatibility antigens expressed in the recipient’s tissue. The prevention, early detection, and treatment of GVHD results in better outcomes.^[5–7]

Some risk factors for acute and chronic GVHD, including HLA mismatch, donor and recipient sex, donor parity, recipient age, high-dose total body irradiation, intensity of the conditioning regimen, ABO compatibility, donor transfusions, and previous acute GVHD, among others, have been described.^[8–11]

This study aimed to compare the incidence of acute and chronic GVHD between haploidentical and identical allo-SCT recipients and evaluate potential factors associated with the development of GVHD in the first year after transplant in a cohort of patients from a transplant center in Colombia.

Our retrospective cohort study included adult patients with malignant and nonmalignant hematologic pathologies who received allo-SCT between 2014 and 2022 at a stem cell transplant center in Bogotá, Colombia. This study was approved by the Ethics Committee of the Pontificia Universidad Javeriana and Hospital Universitario San Ignacio, Bogotá, Colombia (record number FM-CIE-1266-22). Informed consent was not needed.

The inclusion criteria were patients older than 18 years of age who were diagnosed with a hematologic disorder, including acute or chronic leukemia, lymphoma, myelodysplastic syndrome, aplastic anemia, or myelofibrosis, and who underwent allo-SCT. Patients who did not meet the institutional protocol for 12-month posttransplant follow-up between March 2014 and June 2022 were excluded.

Data were collected retrospectively from electronic medical records. An electronic database was created in Redcap (Research Electronic Data Capture) to record the clinical variables of interest.

The primary outcome was the incidence of GVHD. Based on the Transplant Group Task Force Position Statement,^[7] GVHD was classified as acute or chronic according to its clinical manifestations. Acute involvement affects the skin, gastrointestinal tract (GIT), and liver. Cutaneous manifestations include maculopapular rash, generalized peeling, and, in some cases, blistering. Gastrointestinal involvement may be high or low; high gastrointestinal involvement is usually associated with anorexia, vomiting, nausea, and dyspepsia; low gastrointestinal involvement is accompanied by diarrhea, severe abdominal pain with or without ileus, or bloody stools. Liver involvement is mainly diagnosed by elevated liver function tests.^[7] Chronic GVHD is defined on the basis of specific findings or distinguishing features, in conjunction with additional confirmatory tests, with involvement of at least one of the following organs: mouth, eyes, skin and appendages, esophagus, muscles, fascia, joints, lungs, or genitals.^[7,8,10,11]

To score the severity of acute GVHD, we used the Mount Sinai Acute GVHD International Consortium criteria, which are considered the most updated and accurate criteria for diagnosing and scoring the severity of acute GVHD.^[6] Each affected organ is given a score that correlates with its severity, from O “mild” to 4 “severe,” and the sum of the systems affected correlates with worse outcomes, with O being the “mildest scenario” and IV “the worst.”^[7]

For the scoring of chronic GVHD, we used the 2014 NIH diagnostic criteria. Both functional organ impairment and patient symptoms were considered, with a range of 0 (absent) to 3 (severe) for each affected organ. The final summary severity score for chronic GVHD is “mild” if two organs are scored as one, “severe” if any organ is scored as 3, and “moderate” for all other possible combinations.^[7]

A descriptive analysis of the sociodemographic and clinical variables, such as sex, age, body mass index, hematologic diagnosis, conditioning regimen, GVHD prophylaxis, and infectious complications, was performed. Additionally, information about relevant outcomes, such as toxicity complications associated with chemotherapy, relapse, and mortality at 1-year posttransplant was obtained. Measures of central tendency and dispersion were calculated for quantitative variables. The means and standard deviations were used for normally distributed variables, and the medians and interquartile ranges were used for nonnormally distributed variables. Normal distribution was assessed via the Kolmogorov–Smirnov test at the 5% significance level (p < 0.05).

Comparisons between groups (haploidentical and MRD l transplants) were made via the Student’s t-test for normally distributed continuous variables and the Mann–Whitney U test for nonnormally distributed continuous variables. The χ² test or Fisher’s test was used for categorical variables, depending on the sample size.

Univariate and multivariate logistic regression models were used to assess risk factors associated with the presence of acute and chronic GVHD. Variables that had a p-value < 0.2 and those of clinical relevance were included in the multivariate model. The selection of variables for the final model was performed via a backward stepwise methodology. Two-tailed analysis was used to determine statistical significance, with a p-value <0 .05 defined as significant. The analysis was performed with R software version 4.4.1.

A total of 152 patients were enrolled, including 108 patients with an MRD (identical donor) transplant and 44 patients who transplanted from a haploidentical donor. The median age was 45 years, and 80 (41%) patients were women. The most common diagnosis for transplant was acute myeloid leukemia (37.5%) followed by acute lymphoblastic leukemia (36.2%) and aplastic anemia (7.9%), as described in Table 1.

Regarding the intensity of the conditioning regimen, 80 (74%) of MRD recipients received myeloablative conditioning (MAC), whereas 28 (25.9%) received reduced-intensity conditioning (RIC). Among the haploidentical donor transplants, 21 (47.7%) received MAC + total body irradiation 2 Gy (TBI 2 Gy), whereas 23 (52.3%) received RIC + TBI 2 Gy (Table 1).

The interquartile range (IQR) platelet and neutrophil engraftment days were 12 (IQR: 11–14) versus 13 (IQR: 12–19) and 15 (IQR: 13–16) versus 15 (IQR 12–19), respectively, between the MRD and and haploidentical donor transplant (HDT) group respectively. The main type of infection occurring 1 month after transplantation was of unknown origin in 49 patients (54.4%) in the MRD group and 14 (35.9%) in the HDT group (p = 0.08). The second most common origin was bacteremia, which was present in 13 (33.3%) and 12 (13.3%) patients with HDT and MRD group, respectively (p = 0.01). Mucositis was present in 31 (70.4%) and 75 (69.4%) of the HDT and MRD groups, respectively (p = 1). Neutropenic fever was the most common complication at 1 month after transplantation and was present in 42 (95.4%) HDT recipients and 87 (80.5%) MRD recipients (p = 0.06).

In the MRD group, 78 (72%) of the patients received cyclosporine plus methotrexate, whereas 30 (27.7%) received cyclophosphamide (Cyclo) plus calcineurin inhibitor (CNI) in combination with mycophenolate (MFM). In contrast, 44 (100%) patients in the HDT group received Cyclo + CNI + MFM (Table 2).

Intensive care unit (ICU) admission 1 month after transplantation was greater in the HDT group, corresponding to nine patients (20.4%) and five patients (4.6%) in the MRD group (p < 0.05). Mortality at 1-year posttransplant was greater in the HDT group than in the MRD group (23 [54.8%] vs. 30 [29.7%] patients; p < 0.05). There were no statistically significant differences in relapse at 1 year between the MRD (n = 28; 28%) and HDT (n = 4; 19%) (p = 0.5). (Table 1).

The incidence of acute GVHD at 1-year posttransplant was greater in the HDT group (28/44 [63.6%] vs. 39/108 [36.1%]; p < 0.05 in MRD group). The severity of acute GVHD was mild (grade II) in the HDT group, with 9 (32%) patients versus 5 (13.8%) in the MRD group. The severity of acute GVHD (grades III–IV) was similar in both groups but tended to be slightly greater in the MRD group, with 7 (19.4%) patients versus 5 (17%) in the HDT group. The main organ involved in acute GVHD was the skin in 69% of patients, the GIT in 19%, and the liver in 13%. The incidence of chronic GVHD at 1-year posttransplant was 36 of 108 (33.3%) patients in the MRD group and 8 of 44 (18.2%) in the HDT group (p = 0.09). The severity of chronic GVHD was mild in both types of transplantation, and no cases of severe GVHD were observed at the 1-year follow-up. The main organs involved in chronic GVHD were the skin, liver (47%), and GIT (5%) in MRD group versus the skin (37%), liver (27%), and GIT (25%) in the HDT group (Table 2, Figs. 1–4).

The results of the univariate and multivariate analyses are presented in Tables 3 and 4. Our results revealed that female sex (OR 2.34; 95% CI, 0.93–6.1) and age older than 50 years (OR 2.1; 95% CI, 0.81–5.71) were associated with the development of acute GVHD, whereas relapse was associated with a reduced likelihood of acute GVHD (OR 0.41; 95% CI, 0.13–1.16). In contrast, haploidentical transplant type (OR 0.22; 95% CI, 0.05–0.75) and relapse (OR 0.16; 95% CI, 0.04–0.56) were associated with a reduced likelihood of chronic GVHD.

The hematopoietic cell transplant-specific comorbidity index (HCT-CI) was employed to ascertain the mortality risk in patients undergoing allogeneic HCT. Patients were classified into the following three risk groups: Group 0 (low risk), Group 1–2 (intermediate risk), and Group 3 or more (high risk). Among all patients with acute GVHD, 11% exhibited a low-risk score, 7.3% demonstrated an intermediate-risk score, and 2.7% displayed a high-risk score.

An international consensus has established a mean fluorescence intensity (MFI) of more than 1000 as the cut-off for donor-specific antibodies (DSA) positivity to define the presence of anti-HLA DSA in haplo-HLA donor recipients. An MFI of 1000–3000 is classified as low positive, 3000–5000 as intermediate positive, and more than 5000 as high positive. An MFI of less than 1000 is considered negative. In our cohort, 80% of donors exhibited no anti-DSA reactivity, while 18% demonstrated low positivity.

Our work provides relevant findings concerning the incidence and risk factors for acute and chronic GVHD in a Colombian bone marrow reference center. A total of 152 patients were analyzed, including 108 MRD recipients and 44 HDT recipients. The median age was 45 years in both groups. Most recipients were male, and the most common diagnosis was myeloid leukemia. Patients who received HDT had a greater incidence of acute GVHD, infections, and ICU admission than those who received MRD transplant (identical donor). The factors associated with the development of acute GVHD were female sex and age older than 50 years, whereas relapse was associated with a reduced likelihood of acute GVHD. In contrast, the factors associated with a reduced likelihood of chronic GVHD were HDT and relapse.

In terms of the basal characteristics of our cohort, the mean age of our patients was lower than that reported in previous studies,^[4] and most transplants were performed in men, with acute myeloid leukemia being the most common indication for transplantation, which is consistent with data published in other cohorts.^[4,11,12] The risk of infection is determined by several factors, including the conditioning regimen, the type of transplant, and the use of GVHD prophylaxis.^[11–13]

The conditioning regimen was defined as MAC when the total dose of busulfan was greater than 8 mg/kg or greater than 6.4 mg/kg when administered orally or intravenously, respectively. All other regimens were classified as RIC. The MAC regimen was used more frequently in patients who underwent MRD transplants (74%) than in those who received haploidentical donor transplants (47%). Myeloablative regimens employ cytotoxic chemotherapies that result in prolonged neutropenia, increased mucositis, and an elevated risk of organ toxicity.^[14] In our study, both transplant groups received myeloablative regimens, with a greater proportion in the MRD group than in the HDT group, which could explain the significant toxicity observed in these patients.

In recipients of allo-SCT, cytomegalovirus (CMV) reactivation causes significant morbidity. GVHD prophylaxis with posttransplant cyclophosphamide (PTCy) is associated with an increased risk of CMV infection.^[14] In our study, we had more CMV reactivations in the HDT group, in which the majority of the patients received PTcy, which is in line with the current literature.^[14] Additionally, the type of transplant appears to be a stronger risk factor for infection. Haploidentical transplants take longer to engraft, are predisposed to infections and complications, and usually have higher mortality rates that are not associated with relapse.^[4,5,12] In our study, the HDT group had a longer engraftment time and more complications related to toxicity, such as infections, mucositis, febrile neutropenia, and ICU admission, with higher posttransplant mortality after 1 year in the HDT group, which is consistent with the current literature.^[4,5,12]

Among the most common infectious complications are bacteremia and septicemia of unknown focus, which are more common in recipients of MRD group than in those of HDT group. This is different from what is reported in the literature.^[5,12–14] A probable explanation is the use of cyclophosphamide posttransplantation as GVHD prophylaxis in both transplant groups, which may increase the time of aplasia and neutrophil and platelet engraftment, adding to the toxicity of chemotherapy with an increased incidence of infections in the two transplant groups using this therapy as prophylaxis.

Various integrated scoring systems are available to estimate mortality risk in patients undergoing allo-SCT. An elevated HCT-CI score has been identified as a potential risk factor for severe (grades 3–4) acute GVHD (aGVHD).^[15–18] However, our findings did not support the hypothesis that patients at higher pretransplant risk would experience more severe aGVHD, a result that differs from the literature.^[18]

In haploidentical stem cell transplant, there is a demonstrated association between anti-HLA antibodies DSA and a 10-fold increased risk of graft failure.^[19] In our cohort, 18% of transplants were performed with weakly positive anti-DSA due to the lack of fully negative anti-DSA. However, the sample size was insufficient to permit estimation of differences.

Regarding mortality within the first year after transplantation, a higher incidence was observed in the HDT group, although no statistically significant differences were identified. It is postulated that an extended aplasia period, a prolonged graft time, an elevated risk of infection, a greater prevalence of mucositis, and an extended stay in the ICU may be associated with an increased mortality rate. However, further methodologic designs are required to evaluate the underlying causes of mortality.

We found that the incidence of acute GVHD at 1-year posttransplant was greater in the HDT group than in the MRD group, but the severity of acute GVHD was similar in both groups. HLA mismatch is one of the major risk factors for the presence of acute GVHD and classically describes a higher incidence of acute GVHD in HDT versus MRD transplants, similar to other cohorts.^[20] On the other hand, the incidence of chronic GVHD at 1 year posttransplant tended to be greater in the MRD group than in the HDT group. One of the explanations for the lower incidence of chronic GVHD and greater incidence of acute GVHD may be due to greater T-lymphocyte alloreactivity with HLA disparity, which decreases by using posttransplant cyclophosphamide as a prophylactic strategy in all patients in the HDT group.

In the context of haploidentical transplant, prophylaxis works by depleting rapidly proliferating alloreactive T cells and preserving slowly dividing memory and regulatory T cells, leading to less chronic GVHD, improved immune tolerance, and reconstitution.^[21–26] In our study, we observed milder acute GVHD in the HDT group than in the MRD group and a tendency toward more severe acute GVHD in the MRD group. Concerning the severity of chronic GVHD, no severe GVHD occurred in our cohort. This finding may be explained by posttransplant cyclophosphamide activity in dissimulating alloreactive T lymphocytes and a better response to immunosuppression in the presence of aGVHD.^[27]

In terms of organ involvement, in acute GVHD, the main involvement was cutaneous, followed by gastrointestinal and hepatic; in chronic GVHD, the main involvement was cutaneous, followed by hepatic and gastrointestinal involvement, similar to cohorts reported in the literature.^[9]

HDT with posttransplant cyclophosphamide (PTCy) as a prophylactic strategy for GVHD has been shown to provide lower rates of chronic GVHD.^[4,13,23] In haploidentical donor transplants, alloreactive T lymphocytes are activated in response to HLA mismatch. posttransplant cyclophosphamide stops this activation and regulates the immune system.^[4,22] In our study, haploidentical transplant was associated with a lower risk of chronic GVHD, according to multivariate analysis, which aligns with the current literature.^[4,13,27] Patients with severe chronic GVHD are less likely to relapse, but this condition is associated with greater complications and mortality.^[28–30] Our results demonstrated that relapse was associated with a reduced likelihood of chronic GVHD, which is consistent with the current literature.^[24–32]

Donor sex and parity are known risk factors for acute GVHD, with male donors typically being associated with lower rates of acute GVHD.^[6,18,28] Female recipients of male donors are known to have more acute and chronic GVHD.^[28] The relationship with the female sex is probably related to previous sensitization through pregnancy and the detection of minor antigens on chromosome Y from male donors to female recipients. Multivariate analysis revealed that female sex was associated with the development of acute GVHD, which is consistent with previously reported findings.^[6,8,28,32]

The occurrence of mild and moderate acute GVHD has been associated with a reduced risk of posttransplant relapse, known as the graft-versus-tumor effect, which helps prevent relapse. In contrast, severe acute GVHD has been associated with increased mortality and relapse. Our results demonstrated that relapse was associated with a reduced likelihood of acute GVHD, which aligns with other studies.^[32]

The age of donors older than 50 years have been reported to be a significant risk factor for both mild and severe acute GVHD.^[6,11,23] In our study, we found that an age older than 50 years was associated with the development of acute GVHD, similar to other reports.^[6,11,23]

We did not find an association between the pretransplant status of a positive minimal residual disease in leukemia patients and the presentation of acute or chronic GVHD. The lack of more accurate molecular tests may lead to underreporting of minimal residual disease in patients with leukemia.

During the follow-up period, a continuous, reliable dataset was created based on the adherence to the institutional protocol carried out by professionals, which will be the basis for future research at our center. However, as a single data source, it is not possible to extrapolate these data to the general population, and more comparable studies are needed. The retrospective nature of the study may have led to information bias. Other methodologic designs could be used to evaluate the causes of mortality and relapse-free survival in this population.

The incidence of acute GVHD was greater in HDT recipients than in MRD (identical donor) recipients, and no differences were found for chronic GVHD between the two comparisons. Our results demonstrated that female sex and age older than 50 years were associated with developing acute GVHD, whereas relapse was associated with a reduced likelihood of acute GVHD. In contrast, haploidentical transplant type and relapse were associated with a reduced likelihood of chronic GVHD.