Abstract
We previously reported a risk score that predicted mortality in patients with chronic graft-versus-host disease (CGVHD) after hematopoietic stem cell transplant (HCT) between 1995–2004 and reported to the Center for International Blood and Marrow Transplant Registry (CIBMTR). We sought to validate this risk score in an independent CIBMTR cohort of 1128 patients with CGVHD transplanted between 2005–2007 using the same inclusion criteria and risk-score calculations. According to the sum of the overall risk score (range 1 to 12), patients were assigned to 4 risk-groups (RGs): RG1 (0–2), RG2 (3–6), RG3 (7–8) and RG4 (9–10). RG3 and 4 were combined as RG4 comprised only 1% of the total cohort. Cumulative incidences of non relapse mortality (NRM) and probability of overall survival (OS) were significantly different between each RG (all p<0.01). NRM and OS at five years after CGVHD for each RG were 17% and 72% in RG1, 26% and 53% in RG2, and 44% and 25% in RG 3, respectively (all p<0.01). Our study validates the prognostic value of the CIBMTR CGVHD RGs for OS and NRM in a contemporary transplant population. The CIBMTR CGVHD RGs can be used to predict major outcomes, tailor treatment planning, and enrollment in clinical trials.
Introduction
The beneficial anti-leukemic effect of chronic GVHD (CGVHD) is offset by an increased risk of late non relapse mortality (NRM)[1]. Several clinical risk factors have been identified as predictors of outcomes in patients with CGVHD[2–17]. We previously identified 10 variables significantly associated with both NRM and survival in 5343 patients with CGVHD transplanted between 1995 and 2004[18]. We then developed a CGVHD CIBMTR risk score using these variables. Variables included in the score were age, prior acute GVHD, time from hematopoietic stem cell transplant (HCT) to CGVHD, donor type, disease status at transplant, GVHD prophylaxis, gender mismatch, serum total bilirubin, Karnofsky performance status (KPS) and platelet count at time of diagnosis of CGVHD. Six risk-groups (RG) were identified with powerful discriminative ability to predict NRM and overall survival.
In the current study we sought to validate this risk score in an independent dataset of 1128 patients with CGVHD transplanted between 2005 and 2007 using the same inclusion criteria and risk-score calculations.
Methods
The Center for International Blood and Marrow Transplant Research (CIBMTR) is a research organization formed in 2004 as an affiliation of the International Bone Marrow Transplant Registry (IBMTR), the Autologous Blood and Marrow Transplant Registry (ABMTR), and the National Marrow Donor Program (NMDP). The CIBMTR is a voluntary organization involving more than 500 transplantation centers that have collaborated to share patient data and conduct scientific studies. The quality and compliance of data submission are monitored by computerized checks for errors, physician reviews, and on-site audits.
Study Population
The study population was selected using similar criteria as reported in the original cohort[18]. We included all patients who received a first allogeneic HCT from a related or unrelated donor (URD) including umbilical cord blood for acute myelogenous leukemia, acute lymphoblastic leukemia, chronic myelogenous leukemia, or myelodysplastic syndrome between 2005 and 2007; were diagnosed with CGVHD within one year of transplantation; and reported to the Center for International Blood and Marrow Transplant Research (CIBMTR).
All surviving recipients who received transplantations from URDs included in this analysis were retrospectively contacted and provided informed consent for participation in the National Marrow Donor Program (NMDP) research program. Informed consent for retrospective data analysis was waived by the NMDP institutional review board for all deceased patients. Surviving patients who did not provide signed informed consent to allow analysis of their clinical data were excluded. To adjust for potential bias introduced by the exclusion of nonconsenting surviving patients, a corrective action plan modeling process randomly excluded approximately the same percentage of deceased patients using a biased coin randomization with exclusion probabilities based on characteristics associated with not providing consent for use of data in survivors. The study cohort included a total of 1128 HCT recipients with CGVHD.
Definitions
CGVHD was diagnosed according to standard CIBMTR criteria[9, 19] which includes all patients with clinical criteria of CGVHD[19] with or without positive histology, irrespective of time of onset of symptoms. Data required to generate an National Institutes of Health (NIH) score[20] were not prospectively collected by the CIBMTR during the study period. We defined reduced intensity/nonmyeloablative regimens (RIC/ NMA) based on standard definitions [21]. HLA- matching was defined based on algorithm described by Weisdorf et al[22]. Disease stage at HCT was defined as early, intermediate or advanced based on pre-defined criteria[18]. Overall survival was estimated from onset of CGVHD. Death from any cause was treated as the event. Surviving patients were censored at the date of last contact. NRM was defined as death in continuous remission. The event was summarized by the cumulative incidence estimate with relapse as the competing risk.
Study Endpoints and Statistical Analysis
The primary endpoint of this study was to validate the previous CGVHD risk categorization[18] for NRM and overall survival using a more recent cohort of CIBMTR-registered subjects. The score used has been previously reported and is shown in Supplemental Table 1[18]. In brief, 10 variables identified to be significantly associated with overall survival and NRM (recipient age at transplant, prior acute GVHD, time to onset of CGVHD, serum bilirubin at CGVHD onset, KPS at CGVHD onset, presence of thrombocytopenia at CGVHD onset, donor type, disease status at transplantation, GVHD prophylaxis, and gender mismatch) were used to build the risk score. Each variable was assigned a variable specific risk score. Variable specific risk scores were summed for each patient to assign an overall risk score, which was then categorized into risk groups (Table 1a and Supplemental Table). Variables related to patient, disease and transplant characteristics were described using descriptive statistics. Continuous variables were reported as medians with ranges, while categorical variables were reported as absolute numbers and proportions. Cumulative incidence for NRM was calculated treating disease progression/relapse as the competing risk[23]. Overall survival was calculated using Kaplan-Meier estimates, and 95% confidence intervals (CIs) were calculated using the variance derived from the formula of Greenwood[24]. Time to event was calculated from the date of diagnosis of CGVHD. We used the log-rank test to compare the differences between groups in the time-to-event analysis. All P values were 2-sided. Patient-, disease-, HCT-, and CGVHD- related variables and the assigned risk group were included in the multivariate analyses using a stepwise forward selection technique. P≤ 0.01 was the criteria for inclusion in the final models[25].
Table 1.
| a. Characteristic of the validation and training cohorts including the 10 variables used to build the CIBMTR Risk Score[18] | ||||
|---|---|---|---|---|
| Characteristics | Overall Risk Score* |
Validation cohort N (%) |
Training cohort N (%) |
P |
| Number of patients | 1128 | 5343 | ||
| Variables associated with both OS and NRM included in CIBMTR CGVHD score[18] | ||||
| 1. Age at transplant, years | <0.01 | |||
| ≤29 | 0 | 402 (36) | 2022 (38) | |
| 30–59 | 1 | 594 (53) | 3161 (59) | |
| ≥60 | 2 | 132 (12) | 160 (3) | |
| 2. Donor recipient gender pair | <0.01 | |||
| Female to male | 1 | 222 (20) | 1378 (26) | |
| Other combination | 0 | 906 (80) | 3965 (74) | |
| 3. Donor recipient HLA match | <0.01 | |||
| Identical sibling/ Well- or Partially-matched Unrelated Donor (URD) | 0 | 1006 (89) | 4289 (80) | |
| Other related / Mismatched URD | 1 | 122 (11) | 758 (14) | |
| 4. Disease status at transplant | <0.01 | |||
| Early | 0 | 900 (80) | 2956 (55) | |
| Intermediate | 1 | 39 (3) | 1367 (26) | |
| Advanced | 3 | 189 (17) | 1020 (19) | |
| 5. GVHD prophylaxis regimen | <0.01 | |||
| Cyclosporine ± methotrexate ± other | 0 | 391 (35) | 3341 (63) | |
| Tacrolimus ± methotrexate ± other or T-depletion | 1 | 737 (65) | 2002 (37) | |
| 6. Prior acute graft-versus-host disease (GVHD) | 0.61 | |||
| No | 0 | 311 (28) | 1433 (27) | |
| Yes | 1 | 817 (72) | 3910 (73) | |
| 7. Time from transplant to CGVHD onset | <0.01 | |||
| ≥ 5 months | 0 | 550 (49) | 2317 (43) | |
| < 5 months | 1 | 578 (51) | 3026 (57) | |
| 8. KPS at the time of CGVHD onset | <0.01 | |||
| ≥ 80% | 0 | 714 (63) | 3007 (56) | |
| < 80% | 1 | 414 (37) | 1570 (29) | |
| 9. Total bilirubin at CGVHD onset | <0.01 | |||
| < 2 mg/dl | 0 | 1059 (94) | 3910 (73) | |
| ≥ 2 mg/dl | 1 | 69 (6) | 925 (17) | |
| 10. Platelet count at time of CGVHD onset | 0.48 | |||
| ≥ 100 × 109/L | 0 | 735 (65) | 2650 (50) | |
| < 100 × 109/L | 1 | 393 (35) | 1921 (36) | |
| Risk group ‡ |
Overall score‡ |
Proportion of patients | ||
|
Validation cohort |
Training cohort |
|||
| 1 | 0–2 | 15 | 2.5 | |
| 2 | 3–6 | 73 | 48 | |
| 3 | 7–8 | 11 | 18 | |
| 4 | 9–10 | 1 | 28 | |
| 5 | 11 | 0 | 2.5 | |
| 6 | ≥12 | 0 | 1 | |
| b. Other characteristics of the validation and training cohorts[18] not included in the 10 variables | |||
|---|---|---|---|
| Proportion of patients | P value |
||
| Validation cohort (n = 1,128) |
Training cohort (n = 5,343) |
||
| Year of transplant | 2005–2007 | 1995–2004 | |
| Median age (range), years | 41 (<1–74) | 36 (<1–72) | < 0.01 |
| Diagnosis at transplant | < 0.01 | ||
| Acute myeloid leukemia (AML) | 53% | 34% | |
| Acute lymphocytic leukemia (ALL) | 27% | 21% | |
| Chronic myeloid leukemia (CML) | 10% | 34% | |
| Myelodysplastic syndrome (MDS) | 10% | 11% | |
| CMV serological status pretransplant | 0.79 | ||
| Donor and recipient seronegative | 28% | 29% | |
| Donor and recipient seropositive | 67% | 66% | |
| Missing | 5% | 5% | |
| Graft source | < 0.01 | ||
| Bone marrow | 25% | 62% | |
| Peripheral blood stem cells | 67% | 36% | |
| Umbilical cord blood | 8% | 2% | |
| Conditioning regimen intensity | < 0.01 | ||
| Myeloablative | 66% | 89% | |
| Type of CGVHD onset | < 0.01 | ||
| Progressive | 28% | 43% | |
| Interrupted | 40% | 27% | |
| De novo | 28% | 27% | |
| Missing/not collected on prior forms | 4% | 3% | |
| Maximum grade of chronic GVHD | < 0.01 | ||
| Extensive | 73% | 68% | |
| Limited | 27% | 32% | |
|
Follow-up of survivors, months, median (range)* |
63 (2–92) | 73 (3–168) | < 0.01 |
Disease risk status is categorized as follows: Early = AML or ALL and in first complete remission (CR); CML in 1st chronic phase (CP) and MDS (refractory anemia [RA] or RA with ringed sideroblasts. Intermediate = AML/ALL (≥CR2); CML in accelerated phase or CP ≥2. Advanced = AML/ALL (relapse or primary induction failure, CML in blast phase; MDS (RA with excess of blast or in transformation [RAEB/RAEB-t].
Risk group represents the sum of overall risk score assigned to each of the above 10 variables associated with both NRM and OS in the Cox model as previously described[18]. In the validation study, there were only 4 patients with a score value of 10 and no patient with an overall score value above 10.
Follow-up is defined from date of CGVHD onset to date of last contact at the time of the analysis
Results
The study population included 1128 HCT recipients with CGVHD. Table 1a and 1b compares the demographic, transplant and CGVHD characteristics of the validation cohort (HCT between 2005–2007) to the training cohort (HCT between 1995–2004)[18]. As shown in Table 1b, patients in the validation cohort (2005–2007) were older than in the training cohort (1995–2004), were more likely to receive a HCT from a HLA-matched sibling or well-matched/ partially matched URD and have a transplant for early disease status compared to the training cohort [1]. Cyclosporine (CSA) and methotrexate based GVHD prophylaxis and a gender mismatched HCT (i.e. male recipient from female donor versus other) was less frequent in the validation cohort. Amongst CGVHD characteristics, the validation cohort HCT recipients were less likely to be diagnosed early (< 5 months) after HCT, and to have a serum bilirubin of > 2mg/ dL.
When assigning risk groups, the proportion of patients assigned to the lower risk groups (RG1 and RG2) were significantly higher in the validation cohort as compared to the training cohort (Table 1a). Surprisingly, in the validation cohort, there were only 4 patients with a score value of 10 and none with a score value over 10. Since only 12% of the patients (11% in RG3 and 1% in RG4) were assigned RG3 or 4, for testing the risk score, RG3 and 4 were combined; hence only three categories were tested: RG1, RG2 and RG3. Table 2 displays the distribution of risk score variables amongst patients assigned to each RG. As shown, the most common variables amongst patients assigned to RG1 (> 20% of patients belonged to this category) were prior acute GVHD, GVHD prophylaxis other than CSA+ methotrexate ± other and age > 29 years. Amongst patients assigned to RG2, in addition to the above variables, shorter time from transplant to CGVHD and thrombocytopenia were also frequent. Amongst patients assigned to RG3 or 4, all variables except other related and mismatched URD and KPS < 80 were seen at a higher frequency.
Table 2.
Comparison of risk score characteristics between assigned risk groups
| Characteristics | cGVHD CIBMTR Risk Group | ||
|---|---|---|---|
| 1 | 2 | 3 | |
| Number of patients | 165 | 824 | 139 |
| Age at transplant, years | |||
| ≤ 29 | 122 (74) | 263 (32) | 17 (12) |
| 30 – 59 | 43 (26) | 472 (57) | 79 (57) |
| ≥ 60 | 0 | 89 (11) | 43 (31) |
| Disease risk at transplant | |||
| Early | 163 (99) | 689 (84) | 48 (35) |
| Intermediate | 2 (1) | 33 (4) | 4 (3) |
| Advanced | 0 | 102 (12) | 87 (63) |
| Donor-recipient HLA match | |||
| Identical sibling/ Well- or partially-matched URD | 154 (93) | 736 (89) | 116 (83) |
| Other related / Mismatched URD | 11 (7) | 88 (11) | 23 (17) |
| Donor/ Recipient gender pair | |||
| Female to male | 12 (7) | 168 (20) | 42 (30) |
| Other combinations | 153 (93) | 656 (80) | 97 (70) |
| GVHD prophylaxis | |||
| Cyclosporine ± methotrexate ± Other | 130 (79) | 243 (29) | 18 (13) |
| Tacrolimus ± methotrexate ± Other)/ T-cell depletion | 35 (21) | 581 (71) | 121 (87) |
| Prior acute GVHD | |||
| Yes | 119 (72) | 585 (71) | 113 (81) |
| No | 46 (28) | 239 (29) | 26 (19) |
| Time from transplant to chronic GVHD diagnosis | |||
| < 5 months | 22 (13) | 446 (54) | 110 (79) |
| ≥ 5 months | 143 (87) | 378 (46) | 29 (21) |
| Platelet count at cGVHD onset, × 109/L | |||
| < 100 | 6 (4) | 287 (35) | 100 (72) |
| ≥ 100 | 159 (96) | 537 (65) | 39 (28) |
| Total serum bilirubin at cGVHD onset, mg/dL | |||
| ≥ 2 | 0 | 36 (4) | 33 (24) |
| < 2 | 165 (100) | 788 (96) | 106 (76) |
| Karnofsky score at cGVHD onset | |||
| < 80 | 7 (4) | 305 (37) | 102 (73) |
| ≥ 80 | 158 (96) | 519 (63) | 37 (27) |
| Follow-up of survivors, months, median (range)1 | 62 (2–73) | 64 (3–92) | 63 (4–71) |
Abbreviations: GVHD: graft-versus-host disease; URD: unrelated donor;
follow up is measured from onset of CGVHD
Amongst other characteristics not included in building the CGVHD risk score (Table 1b), patients in the validation cohort were less frequently transplanted for CML, had a higher proportion of patients receiving peripheral blood stem cells (PBSC) or umbilical cord blood (UCB) as the stem cell graft source and had higher proportion of patients receiving a NMA or RIC regimens. Amongst CGVHD characteristics, progressive onset was less frequent as compared to the training cohort (Table 1b).
We tested the discriminatory ability of the risk score to predict overall survival and NRM. Figure 1 shows the cumulative incidence of overall survival and NRM according to CGVHD risk groups. The probability of overall survival in RG1 was 72% (95% CI: 64–79%), RG2 was 53% (95% CI: 50–57%), and RG3 was 25% (95% CI: 18–33%) at 5 years after the diagnosis of CGVHD. The cumulative incidence of NRM was 17% (95% CI: 11–23%) in RG1, 26% (95% CI: 23–29%) in RG2 and 44% (95% CI: 35–52%) in RG3 at 5 years after diagnosis of CGVHD, validating the predictive ability of the risk score in predicting both overall survival and NRM.
Figure 1.
Overall Survival (A) and Non-relapse mortality (B) among patients with CGVHD according to risk score: All estimates are estimated from the date of onset of CGVHD
In multiple regression analysis of NRM (Table 4a), after adjusting for CMV serostatus and type of CGVHD onset, the risk group was independently associated with NRM (overall p <0.0001). Relative risk of NRM was 1.67 (95% CI: 1.1–2.54, p 0.015) and 3.98 (95% CI: 2.5–6.35, p<0.0001) times higher in RG2 and RG3, respectively as compared to RG1. Pairwise comparison also revealed lower relative risk of NRM in RG2 as compared to RG3 (RR: 0.42, 95% CI: 0.32–0.56, p <0.0001), demonstrating the ability of the score to differentiate all three levels of risk group from one another.
In multiple regression analysis of overall survival (Table 4b), after adjusting for disease and type of CGVHD onset, the risk group was independently associated with overall mortality (overall p < 0.0001). The relative risk of death was 1.87 (95% CI 1.37–2.55, p< 0.0001) and 4.57 (95% CI 3.18–6.58, p< 0.0001) times in RG2 and RG3 as compared to RG1. Pairwise comparison revealed lower relative risk of overall mortality in RG2 as compared to RG3 (RR: 0.41, 95% CI: 0.32–0.52, p <0.0001), confirming the discriminative ability of the score to stratify the three levels of risk groups from one another.
While disease relapse was not the focus of the study, the CGVHD score was evaluated and had no association with risk of relapse in the validation cohort (data not shown).
Discussion
Results of the current analysis validated the original CIBMTR risk score in predicting differences in overall survival and NRM in an independent and more contemporaneous cohort of patients with CGVHD. In contrast to our training cohort, some difference was observed in applying the risk score to validation cohort of patients. For instance, the proportion of patients in the validation cohort assigned to risk groups 4 to 6 was very low, due to only 4 patients (1%) with overall score of 9 or 10 and no patient (0%) with overall score above 10, as compared to training cohort[18](Table 1a). When evaluating the differences amongst the variables used to build the risk score between the two cohorts, the baseline characteristics of the two populations are uniquely different and appear mostly responsible for higher risk scores in the training cohort (Table 2). Though the current cohort had a higher proportion of older patients and patients receiving GVHD prophylaxis with agents other than CSA+ methotrexate + other, they had less frequent HCT from a mismatched or other related donor, HCT in intermediate or advanced disease status and gender mismatched HCT. CGVHD characteristics revealed that early diagnosis and higher bilirubin was also less frequent. While the reasons for the very low proportion of patients assigned to risk group 4 to 6 is not completely understood, similar findings have been reported recently in a study that analyzed the CIBMTR risk score in patients with chronic GVHD using the NIH criteria at two individual centers[26]. In the Inamoto, Kim et al. study, the proportion of patient represented in the risk group 4 was only ≤ 1% and 0% representation for the RG5 and 6, which are similar to the findings or our validation cohort.
The disease groups selected in this study population reflect the training cohort selection criteria. Applicability of the risk score in non-malignant diseases, particularly in pediatric patients has not been tested.
Though CGVHD classification using NIH consensus criteria is not available in the CIBMTR data base, the validation cohort includes patients with NIH chronic GVHD diagnosis, which may account for less frequent diagnosis of early CGVHD (persistent or late acute misclassified as CGVHD)[16, 27, 28]. Despite this, we observed good discriminatory ability of the risk score in predicting survival and NRM. Moreover, as mentioned above, in the study by Inamoto, Kim et al[26], the application of the CIBMTR risk score performed well in predicting differences in OS also in contemporary patients treated for NIH defined CGVHD (Supplemental Figure 1).
We also compared overall survival and NRM within the respective risk groups, between the two cohorts (data not shown). Patients in the validation cohort had a lower probability of overall survival at 5 years after the diagnosis of CGVHD as compared to the training cohort, within the three/ four levels of risk groups (90.8% versus 72% in RG1, 67.1% versus 53% in RG2, and 50.5 (RG3) and 40.2% (RG4) versus 25% (RG3) in the training and validation cohorts, respectively). The cumulative incidence of NRM was similarly lower in the training cohort for RG1 (4.6% versus 17%), RG2 (20.4% versus 26%) and RG3 (33.2% in the training cohort), RG4 (43.2% in the training cohort) versus 44% (RG3, in the validation cohort). This could be accounted for by the fact that patients in the validation cohort were significantly older, accounting for the higher mortality within the early RGs. Also, differences in the cohort amongst variables not contributing to the risk score (table 1b), but predictive of NRM and overall survival (tables 4a and 4b), such as higher proportion of patients with AML and lower proportion with CML in the validation cohort may account for the differences seen.
Prior studies have identified several factors predicting overall survival and NRM including thrombocytopenia, greater than 50% skin involvement, lower gastro-intestinal involvement, progressive onset of CGVHD, HLA mismatch, hyperbilirubinemia, lower KPS and increasing age in patients with CGVHD diagnosed by the traditional criteria[16, 17] [2–15]. Single institution retrospective studies have also identified similar risk factors predicting major outcomes in patients with CGVHD using the NIH criteria[16, 17, 27, 28]. Among several prospective studies using NIH consensus criteria, results from the CGVHD consortium reported similar risk factors predicting overall survival and NRM such as thrombocytopenia, KPS of < 80, but also identified new factors such as NIH overlap subtype of CGVHD and severe NIH global severity [29].
A recent analysis through the CGVHD consortium reported no impact of prior acute GVHD on overall survival or NRM after CGVHD[30]. This is in contrast to our current report and may reflect differences in the population as identified by the NIH Consensus Criteria, excluding patients with late acute, persistent or recurrent acute GVHD. Also, the follow-up of the consortium study is shorter at approximately 20 months as compared to 63 months in the current study.
As indicated above, a recent analysis applied the CIBMTR risk score to NIH CGVHD including a cohort of 376 patients treated at two individual centers[26]. Overall survival was well stratified according to the risk score at both centers, and NRM was stratified according to risk score at one center, confirming that the CIBMTR risk score performs well in predicting differences in overall survival in contemporary patients treated for NIH CGVHD. Factors accounting for the center-specific difference in mortality and the CIBMTR cohort, particularly for patients in risk group 3, remains to be determined (Supplemental Figure 1). A dedicated long-term follow-up program may have contributed to the better survival of patients in this high risk category. Supporting this hypothesis, a recent CIBMTR survey found a superior survival for patients transplanted at centers that have reported having a dedicated long-term follow-up clinic (Navneet Majhail, personal communication).
To conclude, our study validates the prognostic value of the CIBMTR CGVHD RGs for OS and NRM in a recent transplant population. The CIBMTR CGVHD RGs can be used to predict major outcomes, treatment planning and enrollment in clinical trials.
Supplementary Material
Table 3.
| a. Variables independently predictive of non relapse mortality in the validation cohort – results of the multivariable analysis | ||||
|---|---|---|---|---|
| Variable | Relative risk (95% CI) |
P value |
Overall P value |
N |
| Risk group score | ||||
| 1 | 1 | <.0001 | 165 | |
| 2 | 1.67 (1.10 – 2.54) | 0.0155 | 824 | |
| 3 | 3.98 (2.50 – 6.35) | <.0001 | 139 | |
| CMV status pretransplant | ||||
| Donor/recipient seronegative | 1 | 0.018 | 319 | |
| Others combination | 1.39 (1.06 – 1.83) | 0.018 | 809 | |
| Chronic GVHD onset | ||||
| Progressive | 1 | 0.002 | 313 | |
| Interrupted | 0.62 (0.48 – 0.82) | 0.0006 | 448 | |
| De novo | 0.54 (0.40 – 0.74) | 0.0001 | 313 | |
| Missing | 0.56 (0.31 – 0.99) | 0.0477 | 54 | |
| Contrast | ||||
| Risk group 2 vs. 3 and 4 | 0.42 (0.32 – 0.56) | <.0001 | ||
| Interrupted vs. De novo | 1.14 (0.83 – 1.57) | 0.4102 | ||
| b: Variables independently predictive of overall survival in the validation cohort– results of the multivariable analysis | ||||
|---|---|---|---|---|
| Variable | Relative risk (95% CI) |
P value | Overall P value |
N |
| Risk group score | ||||
| 1 | 1 | <.0001 | 165 | |
| 2 | 1.87 (1.37 – 2.55) | <.0001 | 824 | |
| 3 | 4.57 (3.18 – 6.58) | <.0001 | 139 | |
| Disease | ||||
| AML | 1 | 0.0001 | 599 | |
| ALL | 0.93 (0.75 – 1.14) | 0.4552 | 301 | |
| CML | 0.66 (0.48 – 0.91) | 0.0111 | 115 | |
| MDS | 0.508 (0.37 – 0.71) | <.0001 | 113 | |
| Chronic GVHD onset | ||||
| Progressive | 1 | 0.002 | 313 | |
| Interrupted | 0.78 (0.64 – 0.96) | 0.0183 | 448 | |
| De novo | 0.71 (0.56 – 0.89) | 0.0031 | 313 | |
| Missing | 0.48 (0.29 – 0.79) | 0.0038 | 54 | |
| Contrast | ||||
| Risk group 2 vs. 3 and 4 | 0.41 (0.32 – 0.52) | <.0001 | ||
| ALL vs. CML | 1.41 (0.995 – 1.99) | 0.0533 | ||
| ALL vs. MDS | 1.82 (1.28 – 2.60) | 0.0009 | ||
| CML vs. MDS | 1.30 (0.84 – 2.00) | 0.2416 | ||
| Interrupted vs. De novo | 1.11 (0.89 – 1.39) | 0.3634 | ||
Abbreviations: AML: acute myeloid leukemia, ALL: acute lymphoid leukemia, CML: chronic myeloid leukemia, MDS: myelodysplastic syndrome
Acknowledgements
The CIBMTR is supported by Public Health Service Grant/Cooperative Agreement U24-CA076518 from the National Cancer Institute (NCI), the National Heart, Lung and Blood Institute (NHLBI) and the National Institute of Allergy and Infectious Diseases (NIAID); a Grant/Cooperative Agreement 5U10HL069294 from NHLBI and NCI; a contract HHSH250201200016C with Health Resources and Services Administration (HRSA/DHHS); two Grants N00014-13-1-0039 and N00014-14-1-0028 from the Office of Naval Research; and grants from *Actinium Pharmaceuticals; Allos Therapeutics, Inc.; *Amgen, Inc.; Anonymous donation to the Medical College of Wisconsin; Ariad; Be the Match Foundation; *Blue Cross and Blue Shield Association; *Celgene Corporation; Chimerix, Inc.; Fred Hutchinson Cancer Research Center; Fresenius-Biotech North America, Inc.; *Gamida Cell Teva Joint Venture Ltd.; Genentech, Inc.;*Gentium SpA; Genzyme Corporation; GlaxoSmithKline; Health Research, Inc. Roswell Park Cancer Institute; HistoGenetics, Inc.; Incyte Corporation; Jeff Gordon Children’s Foundation; Kiadis Pharma; The Leukemia & Lymphoma Society; Medac GmbH; The Medical College of Wisconsin; Merck & Co, Inc.; Millennium: The Takeda Oncology Co.; *Milliman USA, Inc.; *Miltenyi Biotec, Inc.; National Marrow Donor Program; Onyx Pharmaceuticals; Optum Healthcare Solutions, Inc.; Osiris Therapeutics, Inc.; Otsuka America Pharmaceutical, Inc.; Perkin Elmer, Inc.; *Remedy Informatics; *Sanofi US; Seattle Genetics; Sigma-Tau Pharmaceuticals; Soligenix, Inc.; St. Baldrick’s Foundation; StemCyte, A Global Cord Blood Therapeutics Co.; Stemsoft Software, Inc.; Swedish Orphan Biovitrum; *Tarix Pharmaceuticals; *TerumoBCT; *Teva Neuroscience, Inc.; *THERAKOS, Inc.; University of Minnesota; University of Utah; and *Wellpoint, Inc. The views expressed in this article do not reflect the official policy or position of the National Institute of Health, the Department of the Navy, the Department of Defense, Health Resources and Services Administration (HRSA) or any other agency of the U.S. Government.
*Corporate Members
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Financial Disclosure Statement: The authors have no financial disclosures.
References
- 1.Socie G, Stone JV, Wingard JR, et al. Long-term survival and late deaths after allogeneic bone marrow transplantation. Late Effects Working Committee of the International Bone Marrow Transplant Registry. The New England journal of medicine. 1999;341:14–21. doi: 10.1056/NEJM199907013410103. [DOI] [PubMed] [Google Scholar]
- 2.Akpek G, Lee SJ, Flowers ME, et al. Performance of a new clinical grading system for chronic graft-versus-host disease: a multicenter study. Blood. 2003;102:802–809. doi: 10.1182/blood-2002-10-3141. [DOI] [PubMed] [Google Scholar]
- 3.Akpek G, Zahurak ML, Piantadosi S, et al. Development of a prognostic model for grading chronic graft-versus-host disease. Blood. 2001;97:1219–1226. doi: 10.1182/blood.v97.5.1219. [DOI] [PubMed] [Google Scholar]
- 4.Arora M, Burns LJ, Davies SM, et al. Chronic graft-versus-host disease: a prospective cohort study. Biol Blood Marrow Transplant. 2003;9:38–45. doi: 10.1053/bbmt.2003.50003. [DOI] [PubMed] [Google Scholar]
- 5.Arora M, Nagaraj S, Wagner JE, et al. Chronic graft-versus-host disease (cGVHD) following unrelated donor hematopoietic stem cell transplantation (HSCT): higher response rate in recipients of unrelated donor (URD) umbilical cord blood (UCB) Biol Blood Marrow Transplant. 2007;13:1145–1152. doi: 10.1016/j.bbmt.2007.06.004. [DOI] [PubMed] [Google Scholar]
- 6.Jacobsohn DA, Arora M, Klein JP, et al. Risk factors associated with increased nonrelapse mortality and with poor overall survival in children with chronic graft-versus-host disease. Blood. 2011;118:4472–4479. doi: 10.1182/blood-2011-04-349068. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Koc S, Leisenring W, Flowers ME, et al. Therapy for chronic graft-versus-host disease: a randomized trial comparing cyclosporine plus prednisone versus prednisone alone. Blood. 2002;100:48–51. doi: 10.1182/blood.v100.1.48. [DOI] [PubMed] [Google Scholar]
- 8.Lee JH, Choi SJ, Kim S, et al. Graft-versus-host disease (GVHD)-specific survival and duration of systemic immunosuppressive treatment in patients who developed chronic GVHD following allogeneic haematopoietic cell transplantation. British journal of haematology. 2003;122:637–644. doi: 10.1046/j.1365-2141.2003.04472.x. [DOI] [PubMed] [Google Scholar]
- 9.Lee SJ, Klein JP, Barrett AJ, et al. Severity of chronic graft-versus-host disease: association with treatment-related mortality and relapse. Blood. 2002;100:406–414. doi: 10.1182/blood.v100.2.406. [DOI] [PubMed] [Google Scholar]
- 10.Martin PJ, Storer BE, Carpenter PA, et al. Comparison of short-term response and long-term outcomes after initial systemic treatment of chronic graft-versus-host disease. Biol Blood Marrow Transplant. 2011;17:124–132. doi: 10.1016/j.bbmt.2010.06.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ochs LA, Miller WJ, Filipovich AH, et al. Predictive factors for chronic graft-versus- host disease after histocompatible sibling donor bone marrow transplantation. Bone marrow transplantation. 1994;13:455–460. [PubMed] [Google Scholar]
- 12.Stewart BL, Storer B, Storek J, et al. Duration of immunosuppressive treatment for chronic graft-versus-host disease. Blood. 2004;104:3501–3506. doi: 10.1182/blood-2004-01-0200. [DOI] [PubMed] [Google Scholar]
- 13.Storb R, Prentice RL, Sullivan KM, et al. Predictive factors in chronic graft-versus-host disease in patients with aplastic anemia treated by marrow transplantation from HLA-identical siblings. Annals of internal medicine. 1983;98:461–466. doi: 10.7326/0003-4819-98-4-461. [DOI] [PubMed] [Google Scholar]
- 14.Sullivan KM, Witherspoon RP, Storb R, et al. Prednisone and azathioprine compared with prednisone and placebo for treatment of chronic graft-v-host disease: prognostic influence of prolonged thrombocytopenia after allogeneic marrow transplantation. Blood. 1988;72:546–554. [PubMed] [Google Scholar]
- 15.Wingard JR, Piantadosi S, Vogelsang GB, et al. Predictors of death from chronic graft-versus-host disease after bone marrow transplantation. Blood. 1989;74:1428–1435. [PubMed] [Google Scholar]
- 16.Vigorito AC, Campregher PV, Storer BE, et al. Evaluation of NIH consensus criteria for classification of late acute and chronic GVHD. Blood. 2009;114:702–708. doi: 10.1182/blood-2009-03-208983. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Flowers ME, Inamoto Y, Carpenter PA, et al. Comparative analysis of risk factors for acute graft-versus-host disease and for chronic graft-versus-host disease according to National Institutes of Health consensus criteria. Blood. 2011;117:3214–3219. doi: 10.1182/blood-2010-08-302109. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Arora M, Klein JP, Weisdorf DJ, et al. Chronic GVHD risk score: a Center for International Blood and Marrow Transplant Research analysis. Blood. 2011;117:6714–6720. doi: 10.1182/blood-2010-12-323824. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Shulman HM, Sullivan KM, Weiden PL, et al. Chronic graft-versus-host syndrome in man. A long-term clinicopathologic study of 20 Seattle patients. The American journal of medicine. 1980;69:204–217. doi: 10.1016/0002-9343(80)90380-0. [DOI] [PubMed] [Google Scholar]
- 20.Filipovich AH, Weisdorf D, Pavletic S, et al. National Institutes of Health consensus development project on criteria for clinical trials in chronic graft-versus-host disease: I. Diagnosis and staging working group report. Biol Blood Marrow Transplant. 2005;11:945–956. doi: 10.1016/j.bbmt.2005.09.004. [DOI] [PubMed] [Google Scholar]
- 21.Eapen M, Logan BR, Confer DL, et al. Peripheral blood grafts from unrelated donors are associated with increased acute and chronic graft-versus-host disease without improved survival. Biol Blood Marrow Transplant. 2007;13:1461–1468. doi: 10.1016/j.bbmt.2007.08.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Weisdorf D, Spellman S, Haagenson M, et al. Classification of HLA-matching for retrospective analysis of unrelated donor transplantation: revised definitions to predict survival. Biol Blood Marrow Transplant. 2008;14:748–758. doi: 10.1016/j.bbmt.2008.04.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Gooley TA, Leisenring W, Crowley J, Storer BE. Estimation of failure probabilities in the presence of competing risks: new representations of old estimators. Statistics in medicine. 1999;18:695–706. doi: 10.1002/(sici)1097-0258(19990330)18:6<695::aid-sim60>3.0.co;2-o. [DOI] [PubMed] [Google Scholar]
- 24.Kaplan EL MP. Nonparametric estimationfrom incomplete observation. J Am Stat Assoc. 1958;53(282):457–481. [Google Scholar]
- 25.D C. Regression models and life tables. J R Stat Soc B. 1972;34(2):187–220. [Google Scholar]
- 26.Inamoto Y, Kim DD, Storer BE, et al. Application of CIBMTR risk score to NIH chronic GVHD at individual centers. Blood. 2014;123:453–455. doi: 10.1182/blood-2013-11-536581. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Jagasia M, Giglia J, Chinratanalab W, et al. Incidence and outcome of chronic graft-versus- host disease using National Institutes of Health consensus criteria. Biol Blood Marrow Transplant. 2007;13:1207–1215. doi: 10.1016/j.bbmt.2007.07.001. [DOI] [PubMed] [Google Scholar]
- 28.Arora M, Nagaraj S, Witte J, et al. New classification of chronic GVHD: added clarity from the consensus diagnoses. Bone marrow transplantation. 2009;43:149–153. doi: 10.1038/bmt.2008.305. [DOI] [PubMed] [Google Scholar]
- 29.Pidala J, Vogelsang G, Martin P, et al. Overlap subtype of chronic graft-versus-host disease is associated with an adverse prognosis, functional impairment, and inferior patient-reported outcomes: a Chronic Graft-versus-Host Disease Consortium study. Haematologica. 2012;97:451–458. doi: 10.3324/haematol.2011.055186. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Arora M, Pidala J, Cutler CS, et al. Impact of prior acute GVHD on chronic GVHD outcomes: a chronic graft versus host disease consortium study. Leukemia. 2013;27:1196–1201. doi: 10.1038/leu.2012.292. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.