Risk and prognostic factors for acute GVHD based on NIH ... - Nature

Bone Marrow Transplantation (2013) 48, 587–592 & 2013 Macmillan Publishers Limited All rights reserved 0268-3369/13 www.nature.com/bmt

ORIGINAL ARTICLE

Risk and prognostic factors for acute GVHD based on NIH consensus criteria S-E Lee, B-S Cho, J-H Kim, J-H Yoon, S-H Shin, S-A Yahng, K-S Eom, Y-J Kim, H-J Kim, S Lee, C-K Min, S-G Cho, D-W Kim, J-W Lee, W-S Min and C-W Park To investigate the risk factors for acute GVHD (aGVHD), based on NIH consensus criteria (NCC), we evaluated 775 patients who underwent allogeneic transplantation. Of them, 346 patients developed aGVHD by NCC, in whom we also analyzed factors affecting aGVHD-specific survival. The cumulative incidence of aGVHD was 44.7%, consisting of classic aGVHD (n ¼ 320) and late-onset (n ¼ 26). Multivariate analyses revealed that younger age (P ¼ 0.015), unrelated donors (P ¼ 0.004) and acute leukemia compared with other hematologic malignancies (P ¼ 0.005) were significant risk factors for aGVHD, whereas PBSCs showed no association (P ¼ 0.720). Multivariate analyses, with only aGVHD patients, revealed that late-onset aGVHD had superior aGVHD-specific survival to classic aGVHD (P ¼ 0.044), and identified the association of visceral organ involvement (P ¼ 0.002), severity of aGVHD at onset (P ¼ 0.035) and advanced disease status (Po0.001) with inferior aGVHD-specific survival. In conclusion, this study demonstrates the risk and prognostic factors for aGVHD by NCC with some differences with the previous reports that were based on old criteria. The difference in the risk factors according to different criteria will give insights about the pathophysiology of GVHD. The better prognosis of late-onset aGVHD than of classic aGVHD raises the necessity for prospective trials with a large cohort focusing on the onset time. Bone Marrow Transplantation (2013) 48, 587–592; doi:10.1038/bmt.2012.187; published online 24 September 2012 Keywords: acute GVHD; NIH consensus criteria; risk factor; prognostic factor

INTRODUCTION The diverse risk factors for acute GVHD (aGVHD) have been described over the last three decades, based on traditional diagnoses of aGVHD as the time of onset; within 100 days following allo-SCT.1–7 However, on the basis of recent advances in allo-SCT practice, including reduced-intensity conditioning (RIC) and donor lymphocyte infusion, the National Institutes of Health consensus criteria (NCC) for GVHD emphasize the manifestations of GVHD, instead of the time of onset after allo-SCT.8 According to the new classification, patients with acute features of GVHD after day 100 and with chronic features of GVHD within day 100 are reclassified as each type of GVHD by NCC. Thus, risk factors for aGVHD need to be re-evaluated on the basis of the new criteria. Flowers et al.9 first evaluated the risk factors for grades II-IV aGVHD and chronic GVHD (cGVHD) by NCC, and showed some discrepancies with the previously reported risk factors by day 100 definitions. However, to confirm risk factors of aGVHD by NCC, the evaluation of another cohort is needed. Moreover, the prognostic factors, including characteristics of aGVHD itself, affecting aGVHD-specific survival in a cohort of aGVHD defined by NCC need to be evaluated, but, to our knowledge, there have been no trials yet. Thus, we retrospectively reclassified historically defined GVHD by NCC, and have evaluated the prognostic factors as well as risk factors for newly defined aGVHD.

MATERIALS AND METHODS Patient selection This study included 775 adult patients, who underwent allo-SCT from sibling and unrelated donors for hematologic malignancies between January 2002 and December 2008 at our institution. The analysis of risk factors for aGVHD was performed in all patients (the first cohort). Of them, 346 patients with aGVHD by NCC (the second cohort) were analyzed to assess the prognostic factors for aGVHD. In this study, aGVHD with grade II-IV were considered as aGVHD. This retrospective study was approved by the Institutional Review Board of the Seoul St Mary’s Hospital of The Catholic University of Korea.

Transplant procedures Patients received either a myeloablative conditioning regimen (MAC; TBI/CY, TBI/BU, TBI/cytarabine/melphalan or BU/CY; n ¼ 540) or RIC (fludarabine and BU with or without TBI 400 cGy or fludarabine and melphalan; n ¼ 235). MAC given in this cohort consisted of mainly TBIbased regimens (93%) compared with RIC (32%). Overall, 483 patients received BM and 247 received granulocyte-CSF (G-CSF)-mobilized PBSCs. Forty-five patients received G-CSF-primed unmanipulated BM plus CD34 þ enriched PBSCs, because of the risk of graft failure. HLA matching was based on serologic typing for HLA-A, -B, and -C antigens and molecular typing for HLA DRB1 in 2002, and molecular typing for HLA-A, -B, -C and DRB1 from 2003 to 2008. Antithymocyte globulin (ATG, IMTIX-Sangstat, Lyon, France; 2.5 mg/kg) was added as a part of conditioning to reduce GVHD in patients who received not only transplants from mismatched

Department of Internal Medicine, Division of Hematology, Catholic Blood and Marrow Transplantation Center, Seoul St Mary’s Hospital, College of Medicine, The Catholic University of Korea, Seoul, Korea. Correspondence: Professor B-S Cho, Department of Hematology, Catholic Blood and Marrow Transplantation Center, Seoul St Mary’s Hospital, College of Medicine, The Catholic University of Korea, 505, Banpo-dong, Seocho-gu, Seoul 137-701, Korea. E-mail: [email protected] Received 23 May 2012; revised 23 July 2012; accepted 13 August 2012; published online 24 September 2012

Risk and prognostic factors of aGVHD by NCC S-E Lee et al

588 unrelated donors, but also PBSCs from some matched unrelated donors. GVHD prophylaxis was attempted by administration of calcineurin inhibitor (CYA for the majority of related transplants and tacrolimus for all unrelated transplants) plus short course MTX. The calcineurin inhibitor dose was tapered gradually starting on day 100–120 after allo-SCT in the absence of aGVHD. The other general transplantation procedures were performed as described previously.10,11

exception of ATG use and type of diagnosis. Late-onset aGVHD developed only in patients not receiving ATG, and in a higher proportion of malignancies other than acute leukemia. GVHD characteristics between the two groups differed in that patients with late-onset aGVHD were more frequently involved in liver with or without skin (Table 1).

Diagnosis, management and reclassification of GVHD

Cumulative incidence of aGVHD with grade II–IV Among the 775 patients, 346 patients were reclassified as having aGVHD by NCC, and 320 were from 362 patients with historically defined aGVHD (within day 100). Other 42 of the 362 patients, who developed chronic features of GVHD before day 100, were not considered as aGVHD by NCC. On the other hand, 26 patients were reclassified as late-onset aGVHD by NCC from 274 patients, who developed GVHD over 100 days after SCT (Supplementary Figure S1). The cumulative incidence of aGVHD by NCC was 44.7% (95% confidence interval (CI), 41.2–48.2). In detail, the cumulative incidences of classic aGVHD and late-onset aGVHD were 41.2% (95% CI, 37.7–44.7) and 3.6% (95% CI, 2.4–5.1), respectively (Figure 1).

The historical diagnosis and grade of GVHD was made, as previously described by consensus criteria.12,13 The primary treatment for historically defined aGVHD mainly consisted of methylprednisolone (2 mg/kg) or an equivalent dose of prednisone, and the treatments of steroid-refractory aGVHD were variable, as described in a previous report.14 The treatment of cGVHD also varied, but followed general guidelines as described previously.15 All patients, historically defined aGVHD with grade II-IV or cGVHD, were reclassified according to the NCC,8 in which all events of aGVHD by NCC (classic aGVHD and late-onset aGVHD) were included in the second cohort for prognostic factor analysis of aGVHD. In this study, among patients with overlap syndrome, patients who developed both acute and chronic features simultaneously without any history of GVHD were considered to have de novo overlap syndrome. They were considered to have cGVHD, and were not included in the analyses of the risk and prognostic factors for aGVHD. However, patients with de novo aGVHD, who later developed chronic features with persistent former acute features (n ¼ 28) or who developed overlap syndrome after the resolution of aGVHD (n ¼ 7), were included in the events of aGVHD by NCC at the onset of acute features.

Statistical analysis Main purpose of this study is to investigate the risk and prognostic factors for aGVHD by the NCC that occurred at any time after SCT. In the first cohort, the cumulative incidence was used to estimate the probability of aGVHD, treating non-relapse death and relapse as respective competing risks. Covariates having a P value of o0.1 for association with aGVHD by NCC in the univariate analyses were added to a Cox proportional hazards regression models. However, type of GVHD prophylaxis and use of TBI were considered as the confounding factors against donor type and conditioning intensity, respectively, because the kind of calcineurin inhibitors was mainly determined by donor type (related, mainly CYA; unrelated, all tacrolimus) and MAC consisted of mostly TBI-based conditioning regimens. We focused on the effect of donor type and conditioning intensity. In the second cohort, to identify prognostic factors of aGVHD by NCC, the Cox proportional hazards regression models were used for aGVHD-specific survival. aGVHD-specific survival was defined as the interval from the onset of aGVHD to death because of aGVHD-related complications without a relapse of the underlying malignancies after the onset of aGVHD. GVHD-related death occurring after onset of cGVHD by NCC was not considered as aGVHD-related death. Covariates having a P value of o0.1 for association with aGVHD by NCC in the univariate analyses were added to a final multivariate model. The software used was SAS (SAS Institute, Inc., Cary, NC, USA), and information on the required data format for these analyses can be found in http://www.biomedcentral.com/content/supplementary/cc6852-S1.pdf. Cumulative incidence functions of aGVHD were computed with R functions from competing risk analysis libraries (R software, version 2.9.1; R foundation for Statistical Computing, Vienna, Austria).

RESULTS Patients and transplantation characteristics Table 1 lists the demographic information of all patients. There were 426 male and 349 female patients with a median age of 36 years (range, 15–68). Patients with various hematologic malignancies were transplanted from sibling and unrelated donors. According to the degree of HLA match,16 unrelated donors were divided into the following three groups: well matched, partially matched and mismatched. At the time of the transplant, 153 patients (20%) had advanced disease features. Patients- and transplantation-related characteristics between the patients with classic aGVHD and late-onset aGVHD were similar, with the Bone Marrow Transplantation (2013) 587 – 592

Risk factors for the occurrence of aGVHD by NCC Table 2 lists the results of multivariate analyses for risk factors affecting cumulative incidences of aGVHD and its subtypes (classic aGVHD and late-onset aGVHD). Of the potential risk factors for the occurrence of aGVHD on univariate analyses (Supplementary Table S1), younger age of patients and acute leukemia were significant factors that showed an association with the occurrence of aGVHD, and more intense conditioning had a tendency for a higher risk for aGVHD, whereas PBSCs were not associated with aGVHD. Unrelated donors had a significantly higher risk for aGVHD than matched sibling donors. The risk factors for classic aGVHD were similar to those for total aGVHD, while only the use of a female donor for a male recipient showed a tendency for higher risk of late-onset aGVHD. We performed separate multivariate analyses; including the use of TBI instead of conditioning intensity also showed similar results, with a trend for higher risk for aGVHD by NCC in the use of TBI (Supplementary Table S2). In addition, to analyze homogenous populations, we performed multivariate analyses with separated cohorts according to donor type (sibling donor versus unrelated donor), diagnosis (acute leukemia versus others) and conditioning intensity (MAC versus RIC). The significance of patient’s age, acute leukemia and unrelated donor transplant were held within each model, suggesting the independency of these factors as a predictor for aGVHD (Supplementary Table S3). Prognostic factors affecting aGVHD-specific survival With a median follow-up of 39 months (range, 34.4–43.6) after the onset of aGVHD, the 3-year aGVHD-specific survival was 81.2% (95% CI, 76.6–86.1). To evaluate the prognostic factors for aGVHDspecific survival in the second cohort, consisting of 346 patients with aGVHD by NCC, univariate and multivariate analyses were performed. Characteristics of aGVHD at the onset were also included in these analyses. All aGVHD involving only skin had grade II at the time of onset, while some proportion of aGVHD involving visceral organs had grade III–IV (24%), as well as grade II (76%) at the time of onset. Thus, considering the close relationship between organ involvement and severity of aGVHD at the time of onset, each factor was separately included in multivariate models. Table 3 lists the results of multivariate analyses for aGVHD-specific survival. Of the factors that affected aGVHD-specific survival on univariate analyses (Supplementary Table S4), multivariate analyses revealed that the advanced disease status at transplant was significantly associated with inferior aGVHD-specific survival, and increasing age of patients had a tendency for a lower survival & 2013 Macmillan Publishers Limited


589 Table 1.

Patients and transplantation characteristics Pc

All patients (n ¼ 775) (%)

Classic aGVHD (n ¼ 320) (%)

Late-onset aGVHD (n ¼ 26) (%)

Age of patient (years), median (range) Sex of patient, M/F Sex of donor, M/F

36 (15–68) 426 (55)/349 (45) 474 (61)/301 (39)

33 (15–60) 176 (55)/144 (45) 205 (64)/115 (36)

36 (16–53) 19 (73)/7 (27) 15 (58)/11 (42)

0.602 0.076 0.532

Sex pair (donor–recipient) Female to male/othersa

169 (22)/606 (78)

64 (20)/256 (80)

9 (35)/17 (65)

0.072

594 (77); 367 (47)/227 (29) 181 (3); 82 (11)/68 (9)/31 (4)

275 (86); 150 (47)/125 (39) 45 (14); 20 (6)/22 (7)/3 (1)

17 (65); 8 (31)/9 (35) 9 (35); 7 (27)/2 (8)/0 (0)

622 (80)/153 (20)

256 (80)/64 (20)

24 (92)/2 (8)

492 (63) 283 (37) 152 (20)/100 (13)/31 (4)

176 (55) 144 (45) 70 (22)/58 (18)/16 (5)

17 (65) 9 (35) 6 (23)/1 (4)/2 (8)

385 (50)/390 (50)

150 (47)/170 (53)

10 (39)/16 (61)

0.402

483 (62)/247 (32)/45 (6)

221 (69)/86 (27)/13 (4)

18 (69)/5 (19)/3 (12)

0.202

Conditioning regimen TBI-based/non-TBI-based

583 (75)/192 (25)

269 (84)/51 (16)

20 (77)/6 (23)

0.407

ATG given as conditioning Yes/no

115 (15)/660 (85)

58 (18)/262 (82)

0 (0)/26 (100)

0.026

Conditioning intensity MAC/RIC CD34 þ cells ( 106/kg), median (range) CD3 þ cells ( 107/kg), median (range)

540 (70)/235 (30) 3.9 (0.3–49.0) 4.5 (0.04–109.70)

259 (81)/61 (20) 4.1 (0.1–34.4) 4.6 (0.04–72.7)

19 (73)/7 (27) 3.1 (1.0–11.0) 3.8 (0.2–47.3)

0.352 0.441 0.329

GVHD prophylaxis CYA-based/FK506-based

450 (58)/325 (42)

166 (52)/154 (48)

15 (58)/11 (42)

0.597

Organ involvement Skin/gut/liver Gut þ skin/liver þ skin Gut þ liver/skin þ gut þ liver UG

— — — —

141 (44)/51 (16)/23 (7) 42 (13)/22 (7) 6 (2)/13 (4) 22 (7)

3 (12)/0 (0)/17 (65) 1 (4)/5 (19) 0 (0)/0 (0) 0 (0)

Initial aGVHD grade II/III-IV

—

275 (86)/45 (14)

22 (85)/4 (15)

Parameters

Diagnosis Acute leukemia; AML/ALL Others; CML/MDS/PM Pre-SCT disease status Standard/advancedb Donor type Sibling Unrelated WM/PM/MM ABO match Match/mismatch Sources of graft BM/PBSC/BM þ PBSC

0.004

0.110 0.311

o0.001

0.827

Abbreviations: aGVHD ¼ acute GVHD; ATG ¼ antithymocyte globulin; F ¼ female; M ¼ male; MAC ¼ myeloablative conditioning; MDS ¼ myelodysplastic syndrome; MM ¼ mismatched; PM ¼ plasma cell myeloma; PM ¼ partially-matched; RIC ¼ reduced-intensity conditioning; UG ¼ upper gut; WM ¼ wellmatched; y ¼ year-old. aOthers included male to male (n ¼ 257), female to female (n ¼ 132) and male to female (n ¼ 217). bAdvanced disease status at transplantation includes acute leukemia beyond the first remission, CML beyond the first chronic phase, high-risk myelodysplastic syndrome (international prognostic scoring system Xintermediate-2), plasma cell myeloma with chemoresistance. cValues for categorical variables were analyzed using a w2 or Fisher’s exact test to compare the characteristics of three subcategories of GVHD. Independent-samples T test were utilized to compare the continuous variables.

in analysis with organ involvement at the time of onset. Lateonset aGVHD had a significantly lower risk or a lower tendency for aGVHD-specific deaths, which was compared with that of classic aGVHD in multivariate analyses, including the initial aGVHD grade or organ involvement, respectively. aGVHD involving skin only and the lower grade of aGVHD at onset were significantly associated with a higher GVHD-specific survival, respectively. DISCUSSION Our main interest was to evaluate the risk and prognostic factors for the newly classified aGVHD by NCC. Of the 362 patients & 2013 Macmillan Publishers Limited

with historically defined aGVHD, 42 patients with cGVHD before day 100 were excluded, while 26 patients with late-onset aGVHD were added to aGVHD by NCC. Younger age of patients, unrelated donors and acute leukemia compared with other hematologic malignancies were revealed as significant risk factors for aGVHD by NCC, while the source of graft showed no association. To evaluate prognostic factors for aGVHD-specific survival, we performed further analyses including covariates associated with the characteristics of aGVHD at onset, as well as patients- and transplants-related characteristics. Using these analyses, we found that late-onset aGVHD had better survival than classic aGVHD, and identified the association of visceral Bone Marrow Transplantation (2013) 587 – 592


590 1.0 Cumulative lncidence of aGVHD

Cumulative lncidence of aGVHD

1.0 0.8 0.6 44.7% (95%Cl, 41.2-48.2) 0.4 0.2 0.0

Classic aGVHD Late onset aGVHD

0.8 0.6

41.2% (95%CI, 37.7-44.7) 0.4 0.2

3.6% (95%CI, 2.4-5.1)

0.0 0

500

1000 1500 2000 Days after SCT

2500

0

500

1000 1500 2000 Days after SCT

2500

Figure 1. Cumulative incidence of aGVHD by NCC. The cumulative incidence of aGVHD is 44.7%, which includes 41.2% of classic aGVHD and 3.6% of late-onset.

Table 2.

Potential risk factors for aGVHD by NCC

Multivariate variables

Total aGVHD (n ¼ 346)

N

Classic aGVHD (n ¼ 320)

Late-onset aGVHD (n ¼ 26)

RR (95% CI)

P

RR (95% CI)

P

RR (95% CI) —

Patient’s age (years)

775

0.99 (0.98–1.00)

0.015

0.99 (0.98–1.00)

0.026

Donor type Sibling Unrelated

492 283

1 1.43 (1.12–1.83)

0.004

1 1.35 (1.03–1.78)

0.029

Sources of graft BM PBSC BM þ PBSC

483 247 45

1 0.95 (0.71–1.27) 0.98 (0.59–1.64)

0.720 0.950

1 0.95 (0.70–1.28) 0.89 (0.50–1.57)

0.720 0.680

— — —

Diagnosis Acute leukemiaa Othersb

594 181

1 0.62 (0.44–0.86)

0.005

1 0.54 (0.37–0.78)

0.001

— —

Conditioning intensity MAC RIC

540 235

1 0.73 (0.51–1.03)

0.076

1 0.72 (0.50–1.04)

0.083

— —

ATG given as conditioning Yes No

115 660

— —

1 0.76 (0.54–1.09)

0.130

— —

Sex pair Female to male Others

169 606

— —

— — a

P

—

1 0.49 (0.22–1.12)

0.091

b

Abbreviations: CI ¼ confidence interval; N ¼ number; RR ¼ relative risk. Acute leukemia included AML (n ¼ 367) and ALL (n ¼ 227). Others included CML (n ¼ 82), MDS (n ¼ 68) and PM (n ¼ 31).

organ involvement and severity of aGVHD at onset with inferior aGVHD-specific survival. To our knowledge, there have been no trials to investigate prognostic factors in patients with aGVHD defined by NCC. Recently, a large study by Flowers et al.9 has been conducted to evaluate the risk factors for aGVHD by NCC. In their study, HLAmismatched or unrelated donors, the use of TBI and a female donor for a male recipient were significantly associated with an increased risk of aGVHD in adults and pediatric patients, who underwent allo-SCT after MAC. In contrast, diagnosis of CML was significantly associated with a decreased risk of aGVHD, and the use of ATG as a part of conditioning regimen had a tendency for lower risk of aGVHD. The current study included only adult patients, who received RIC, as well as MAC, which differed from aforementioned study. We also observed a higher risk for aGVHD Bone Marrow Transplantation (2013) 587 – 592

by NCC in unrelated donors and a lower risk in CML than acute leukemia, which strongly suggests the importance of the types of donor and disease as risk factors for aGVHD by NCC. Our data also demonstrated that MAC had a trend for higher risk of aGVHD by NCC than RIC. These results are consistent to the previous reports, which showed the higher risk of MAC for aGVHD by old criteria.3,7 Younger age of patients was associated with higher risk for aGVHD by NCC in this study, which is in contrast with the no association with patient’s age found in the study by Flower et al.,9 which used NCC. This discrepancy may be caused by the distinctive characteristic of our cohort, including RIC and MAC, in which younger patients were more frequently transplanted after MAC than older patients. Regarding the association of patient’s age with aGVHD, the results of previous studies by old criteria, the majority of which included only MAC, have been & 2013 Macmillan Publishers Limited


591 Table 3.

Potential prognostic factors affecting aGVHD-specific survival

Multivariate variables

N

aGVHD-specific survival (including organ involvement)

aGVHD-specific survival (including initial grade)

RR (95% CI)

P

RR (95% CI)

P

Patient’s age (year)

346

1.03 (1.00–1.05)

0.080

1.02 (1.00–1.05)

0.110

Donor type Sibling Unrelated

193 153

1 1.35 (0.78–2.33)

0.280

1 1.34 (0.76–2.34)

0.310

Pre-SCT disease status Standard Advanced

280 66

1 3.08 (1.74–5.47)

0.001

1 2.87 (1.61–5.12)

o0.001

aGVHD type by NCC Classic aGVHD Late-onset aGVHD

320 26

1 0.18 (0.03–1.22)

0.078

1 0.13 (0.02–0.94)

0.044

Organ involvement Skin only Visceral organs

144 202

1 2.75 (1.46–5.18)

0.002

— —

Initial aGVHD grade II III–IV

297 49

— —

— —

1 2.28 (1.06–4.89)

conflicting; some showed an association with older age,1,2,4,7,17 whereas others no association.18,19 The diverse risk factors for aGVHD were reported to be associated with an increased risk of aGVHD by old criteria, which included cGVHD occurring earlier than day 100 and did not include late-onset aGVHD,1–7 despite the existence of conflicting data. Particularly regarding the effect of PBSCs on aGVHD, previous studies using old criteria have shown discrepant findings.6,7 In the recent report from the Center for International Blood and Marrow Transplant Research,18 type of graft, PBSCs, was one of the predominant determinants for aGVHD risk with the intensity of conditioning and the use of TBI. However, studies using NCC, including ours and the study by Flowers et al.9 demonstrated that the occurrence of aGVHD was not affected by the type of graft sources in common. Indeed, contrasting the controversies in the role of PBSCs in aGVHD by old criteria,6,18,20,21 PBSCs is a well-known risk factor for cGVHD by not only old criteria,6,20,21 but also NCC.9 Thus, in previous reports, using old criteria for aGVHD, the proportion of inclusion of cGVHD developed before day 100 in each study may affect the results related with the association of PBSCs with aGVHD. These conflicting effects of PBSCs on aGVHD by NCC with old criteria, not like cGVHD, suggest the limitation of arbitrary criteria of day 100, but also support the basic theory of NCC that aGVHD is a different disease entity with cGVHD. Prospective trials, with a large cohort observing the occurrence of GVHD by NCC, are needed to give an answer to unsolved issues, regarding risk factors for each type of GVHD. Another main objective of this study is to evaluate the prognostic factors for aGVHD-specific survival in patients with aGVHD by NCC. The detrimental effects of advanced disease status at transplantation, visceral organ involvement and the severity of aGVHD by NCC on aGVHD-specific survival were similar in our previous report of aGVHD by old criteria.11,14 Of note, we found that late-onset aGVHD had superior aGVHD-specific survival to classic aGVHD, and more frequent involvement of the liver without any difference in the severity of grade at onset. This suggests that the onset time of acute features by NCC may affect the prognosis as well. We also showed that late-onset aGVHD developed in a higher proportion of other hematologic malignancies than acute leukemia, and did not occur in patients receiving ATG as a part of conditioning. & 2013 Macmillan Publishers Limited

—

0.035

However, we could not reach a conclusion on these issues owing to the small number of late-onset aGVHD. Further trials to evaluate the prognosis and characteristics of aGVHD by NCC according to the onset time are needed. In conclusion, despite limitations by its retrospective nature and recording bias, this study clearly demonstrates the risk and prognostic factors for aGVHD by NCC, with some differences from previous reports that have been based on the old criteria. In situations including the predominant determinants of aGVHD risk in this study, further efforts to reduce the occurrence of aGVHD are needed to optimize outcomes after allo-SCT. Particularly, contrasting to the well-known association of PBSCs with cGVHD, the irrelevance of PBSCs to aGVHD that has been observed in studies based on NCC, including the current study, suggests that aGVHD is a different disease entity compared with cGVHD by NCC. Ongoing research in fields such as biomarkers and genetic profiling, to further characterize the difference in pathogenesis of each type of GVHD, needs to be performed in a homogeneous cohort implicating the risk factors for GVHD. Despite the small number of late-onset aGVHD, the better prognosis of late-onset aGVHD than classic aGVHD raises the necessity for prospective multicenter trials with a large cohort to evaluate the prognostic implication of onset time of acute features by NCC. CONFLICT OF INTEREST The authors declare no conflict of interest.

ACKNOWLEDGEMENTS This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (20110026368).

REFERENCES 1 Gale RP, Bortin MM, van Bekkum DW, Biggs JC, Dicke KA, Gluckman E et al. Risk factors for acute graft-versus-host disease. Br J Haematol 1987; 67: 397–406. 2 Weisdorf D, Hakke R, Blazar B, Miller W, McGlave P, Ramsay N et al. Risk factors for acute graft-versus-host disease in histocompatible donor bone marrow transplantation. Transplantation 1991; 51: 1197–1203.

Bone Marrow Transplantation (2013) 587 – 592


592 3 Nash RA, Pepe MS, Storb R, Longton G, Pettinger M, Anasetti C et al. Acute graft-versus-host disease: analysis of risk factors after allogeneic marrow transplantation and prophylaxis with cyclosporine and methotrexate. Blood 1992; 80: 1838–1845. 4 Kollman C, Howe CW, Anasetti C, Antin JH, Davies SM, Filipovich AH et al. Donor characteristics as risk factors in recipients after transplantation of bone marrow from unrelated donors: the effect of donor age. Blood 2001; 98: 2043–2051. 5 Lee SJ, Klein J, Haagenson M, Baxter-Lowe LA, Confer DL, Eapen M et al. Highresolution donor-recipient HLA matching contributes to the success of unrelated donor marrow transplantation. Blood 2007; 110: 4576–4583. 6 Eapen M, Logan BR, Confer DL, Haagenson M, Wagner JE, Weisdorf DJ 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. 7 Hahn T, McCarthy Jr PL, Zhang MJ, Wang D, Arora M, Frangoul H et al. Risk factors for acute graft-versus-host disease after human leukocyte antigenidentical sibling transplants for adults with leukemia. J Clin Oncol 2008; 26: 5728–5734. 8 Filipovich AH, Weisdorf D, Pavletic S, Socie G, Wingard JR, Lee SJ 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. 9 Flowers ME, Inamoto Y, Carpenter PA, Lee SJ, Kiem HP, Petersdorf EW 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. 10 Min CK, Kim SY, Lee MJ, Eom KS, Kim YJ, Kim HJ et al. Vascular endothelial growth factor (VEGF) is associated with reduced severity of acute graft-versus-host disease and nonrelapse mortality after allogeneic stem cell transplantation. Bone marrow transplant 2006; 38: 149–156. 11 Cho BS, Lee SE, Song HH, Lee JH, Yahng SA, Eom KS et al. Graft-versus-tumor effect according to type of graft-versus-host disease defined by National Institutes

12

13 14

15 16

17

18

19

20

21

of Health Consensus Criteria and Associated Outcomes. Biol Blood Marrow Transplant 2012; 18: 1136–1143. Przepiorka D, Weisdorf D, Martin P, Klingemann HG, Beatty P, Hows J et al. 1994 Consensus Conference on acute GVHD grading. Bone marrow transplant 1995; 15: 825–828. Lee SJ, Vogelsang G, Flowers ME. Chronic graft-versus-host disease. Biol Blood Marrow Transplant 2003; 9: 215–233. Cho BS, Min CK, Eom KS, Kim YJ, Kim HJ, Lee S et al. Clinical impact of thrombotic microangiopathy on the outcome of patients with acute graft-versus-host disease after allogeneic hematopoietic stem cell transplantation. Bone marrow transplant 2008; 41: 813–820. Cho BS, Min CK, Eom KS, Kim YJ, Kim HJ, Lee S et al. Feasibility of NIH consensus criteria for chronic graft-versus-host disease. Leukemia 2009; 23: 78–84. Weisdorf D, Spellman S, Haagenson M, Horowitz M, Lee S, Anasetti C 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. Eisner MD, August CS. Impact of donor and recipient characteristics on the development of acute and chronic graft-versus-host disease following pediatric bone marrow transplantation. Bone marrow transplant 1995; 15: 663–668. Jagasia M, Arora M, Flowers ME, Chao NJ, McCarthy PL, Cutler CS et al. Risk factors for acute GVHD and survival after hematopoietic cell transplantation. Blood 2012; 119: 296–307. Przepiorka D, Smith TL, Folloder J, Khouri I, Ueno NT, Mehra R et al. Risk factors for acute graft-versus-host disease after allogeneic blood stem cell transplantation. Blood 1999; 94: 1465–1470. Korbling M, Anderlini P. Peripheral blood stem cell versus bone marrow allotransplantation: does the source of hematopoietic stem cells matter? Blood 2001; 98: 2900–2908. Cutler C, Giri S, Jeyapalan S, Paniagua D, Viswanathan A, Antin JH. Acute and chronic graft-versus-host disease after allogeneic peripheral-blood stem-cell and bone marrow transplantation: a meta-analysis. J Clin Oncol 2001; 19: 3685–3691.

Supplementary Information accompanies the paper on Bone Marrow Transplantation website (http://www.nature.com/bmt)

Bone Marrow Transplantation (2013) 587 – 592

& 2013 Macmillan Publishers Limited