Patients and methods

SCT procedure

A review was conducted of the medical records of patients who received allogeneic SCTs from September 1998 to October 2002, including 25 patients (27.8%) who received a bone marrow transplant and 65 patients (72.2%) who received a peripheral blood stem cell transplant (PBSCT). Myeloablative conditioning was used in 66 cases (73.3%) and nonmyeloablative conditioning in 24 cases (26.7%). The conditioning regimens were as follows: BuCy in 47 cases, a fludarabine-based regimen in 17 cases, cyclophosphamide plus ATG in 10 cases, a TBI-based regimen in two cases, and other nonmyeloablative protocols in two cases (thymic irradiation/Cy/ATG, 1; TBI 200 cGy, 1). Cyclophosphamide (50 mg/kg once daily intravenous (i.v.) for 4 days) plus ATG (equine antithymocyte immunoglobulin, Pharmacia and Upjohn, Kalamazoo, MI, USA; 30 mg/kg once daily i.v. for 3 days) was used in patients with aplastic anemia (AA), and ATG (10 mg/kg once daily i.v. for 4 days within 2 days of the transplant) was used for transplants from unrelated or partially mismatched donors. Peripheral blood stem cells (PBSCs) were mobilized with a target of more than 4 10⁶/kg of CD34 with G-CSF alone, a G/GM-CSF combination, or sequential injections from donors of a matched sibling (n=71), partially mismatched sibling (n=6), or unrelated donors (n=13), respectively. aGVHD prophylaxis consisted of cyclosporin A and methotrexate. Cyclosporin A was started at a dose of 5 mg/kg/day (continuous i.v.) on day -1 and reduced to a dose of 2.5 mg/kg/day i.v. on day +7, then changed to a dose of 3 m/kg (twice daily p.o.) when tolerable. Methotrexate treatment was 15 mg/m² on day 1, and 10 mg/m² on days 3, 6, and 11. The last dose of methotrexate was omitted when mucositis over grade 3 or renal impairment was observed. Infection prophylaxis consisted of ciprofloxacin (250 mg twice daily p.o.)/metronidazole (500 mg thrice daily p.o.)/fluconazole (100 mg once daily p.o.) beginning with the initiation of conditioning and acyclovir (600 mg twice daily p.o.) from day -1. Cotrimoxazole was started after engraftment. Ursodeoxycholinic acid was used for veno-occlusive disease (VOD) prophylaxis beginning with the initiation of conditioning. CMV antigenemia tests and chimerism studies using variable numbers of tandem repeats based on a PCR with peripheral blood or bone marrow (BM) mononuclear cells were conducted according to protocol guidelines. Engraftment was confirmed by peripheral blood counts (myeloid: peripheral absolute neutrophil count more than 0.5 10⁹/l; megakaryocyte: peripheral platelet count more than 20 10⁹/l for 3 consecutive days without requiring transfusion). When febrile neutropenia developed, ceftazidime and tobramycin were added after obtaining blood cultures. Vancomycin and amphotericin B were used according to protocol guidelines to exclude other infectious causes of neutropenic fevers.

Diagnosis of haGVHD

HaGVHD was diagnosed according to the following criteria: (1) unexplained fever of greater than 38.3°C of body temperature on two occasions for more than 3 days before engraftment with the lack of resolution after a minimum of 3 days of treatment with antibiotics and antifungal agents including amphotericin B, (2) rapid development of skin rash before engraftment with the exclusion of other causes, such as drug eruption, (3) hepatic dysfunction, especially alkaline phosphatase and bilirubin, before engraftment with the exclusion of other causes, such as VOD, drug-induced hepatitis or right-sided heart failure, and (4) development of mucoid, greenish diarrhea with more than 5 stools/day with the exclusion of mucositis. The diagnosis of haGVHD was established based on (1) plus one out of (2)–(4) with or without histologic confirmation and supported by the resolution of the above clinical findings with steroids. Fever associated with mucositis was not included as a criterion, while a new fever after the complete resolution of mucositis was included. Efforts to exclude other causes of hepatic abnormalities were pursued using abdominal ultrasonography, discontinuation of hepatotoxic drugs, and 2D echocardiography. Daily oral examinations for mucositis and stool bacteriologic studies were also performed. Cutaneous manifestations were evaluated by the dermatologist and a skin biopsy was conducted if indicated. Two pathologists conducted the review of skin pathology. The diagnosis of GVHD was made based on the modified Lerner's aGVHD grading system.¹⁰ The histological diagnosis was occasionally difficult when a grade 1 or 2 aGVHD was present. In these cases, the diagnosis was made based on the interpretation of subtle changes and the clinical decision of hematologist.

Corticosteroid therapy

Immediate high-dose corticosteroids (methyl-prednisolone 3–5 mg/kg i.v. for 5 days, then tapered by a dose of 0.2 mg/kg/2 days) were administered to all patients who met the above diagnostic criteria after February 2000, whereas prior to this, supportive measures or a late administration of high- or conventional-dose corticosteroids were given. The early trial group was defined as those patients who started corticosteroid therapy within 7 days after the initiation of haGHVD, while the late trial group started therapy after 7 days.

Definition and statistical analysis

In the results, E–n indicates the onset date of clinical signs before engraftment (eg E-3 indicates 3 days before engraftment). The clinical characteristics between the haGVHD+ and haGVHD- groups were compared using Fisher's exact test and Student's t-test. To identify the relevance of haGVHD with aGVHD and cGVHD, a McNemar test was used. To determine the risk factors for haGVHD, a likelihood ratio test for logistic regression was used. The analysis of the haGVHD risk factors was performed using several parameters, including the age and sex of the recipient and donor, sex mismatch and ABO/Rh mismatch, MNC and CD34 cell dose, matched sibling donors vs alternative donor, including unrelated matched or partially mismatched sibling donors, diagnosis of AA vs non-AA, advanced disease status (more than CR2, relapse or primary refractoriness at transplant) vs standard risk disease status, regimen using ATG vs non-ATG regimen, PBSC as the stem cell source vs BM, and nonmyeloablative regimen vs myeloablative regimen. The overall survival (OS), disease-free survival (DFS), and probability of relapse were calculated from the date of transplant to the date of death, date of relapse or any cause of death, and date of relapse, respectively. The survival curves were estimated using the method of Kaplan and Meier. A log-rank test was used for the survival analysis. The multivariate analysis for survival was analyzed using Cox's regression model. A cutoff P-value of 0.05 was adopted for all statistical analyses. The statistical data were obtained using the SPSS software package (SPSS 10.0 Inc., Chicago, IL, USA).

Results

Transplant outcomes

Among 90 allogeneic stem cell transplants, 34 cases (36.7%) met the haGVHD criteria. The clinical characteristics and transplant outcomes of the haGVHD+ and haGVHD- groups are shown in Table 1. In the haGVHD+ group, the infused MNC and CD34+ cell dose were 5.520.78 10⁸/kg and 10.071.89 10⁶/kg, respectively. Engraftment was achieved in 31 of the 34 cases (91.2%), and the median duration to myeloid and platelet engraftment was 17 days (range, 9–24) and 21 days (range, 10–42) in these 31 patients, respectively. In the haGVHD- group, the infused MNC and CD34+ cell dose were 7.150.65 10⁸/kg and 11.171.86 10⁶/kg, respectively. The median duration to myeloid and platelet engraftment was 16 days (range, 0–30) and 16 days (range, 0–161), respectively.

Table 1 - Patients characteristics and transplant outcomes.

Full table

Clinical manifestation of haGVHD

The clinical manifestations of the haGVHD+ group are summarized in Table 2. The nadir occurred on day 5 (range, 4–7). The first manifestation of haGVHD was an unexplained fever in all cases with a median onset of day 5 post transplant, followed by diarrhea in 27 cases (79.4%) at day 6, hepatic dysfunction (increased AST, ALT, ALP, or bilirubin) in 20 cases (58.8%) at day 10, and skin rashes with erythematous or purpura-like lesions in 21 cases (61.8%) at day 11. The diagnosis of haGVHD was confirmed at a median of 8 days post transplant (range, 5–21). Among the 34 patients diagnosed with haGVHD, 22 biopsies were performed on 18 patients at the time of diagnosis. In 17 cases, a pathologic finding of GVHD was documented. The most common biopsy site was the skin (n=16), followed by the gastrointestinal tract (n=4) and liver (n=2). Among the 16 patients who did not undergo a pathological evaluation, 7 experienced early transplant-related mortality associated with GVHD. The incidence of haGVHD (P=0.010) was significantly lower after PBSCT (20/66, 30.3%) than after BMT (14/24, 58.3%). A relatively high number of patients (12/19, 63.1%) received BM cells with alternative donor transplantation when compared with matched sibling transplantation.

Table 2 - Clinical manifestations of hyperacute GVHD.

Full table

Therapeutic outcomes in patients with haGVHD

Among the 34 patients diagnosed with haGVHD, 22 patients (64.7%) received high-dose corticosteroids beginning at a median of 17 days post transplant (range, 9–40). The median interval between the diagnosis of haGVHD and the initiation of steroids was 6 days (range, 1–33). Among the 22 treated patients, 14 patients (63.6%) received steroids within 7 days after diagnosis, while the administration of corticosteroids to the remaining eight patients (36.4%) was delayed by more than 7 days after the onset of fever due to a hesitant diagnosis.

A prompt response to steroids was documented in all 22 patients (100%) treated with corticosteroids. The fever was resolved within 24 h in 18 cases (81.8%), the diarrhea within 2 days in 15 cases (64.7%), the hepatic dysfunction within 3 days in 14 cases (63.6%), and the skin lesions within 24 h in 13 cases (60.0%). However, in some cases, the responses were not sustained when the steroids were tapered. As such, the recurrence of haGVHD was noted in 11 cases (50%), based on the recurrence of fever, diarrhea, hepatic dysfunction, and skin lesions in 31.8, 36.4, 36.4, and 22.7% of the cases, respectively. When a recurrence was documented, several investigational protocols were attempted, including the reinitiation of high-dose steroids, FK-506, pulse ATG therapy, daclizumab, thalidomide, or clofazimine.

Those patients that achieved a sustained response to the corticosteroids exhibited a superior survival trend compared to the patients with a nonsustained response. Early intervention with corticosteroids within 7 days after the diagnosis of haGVHD was also associated with better survival with a weak statistical significance (P=0.1165).

The median follow-up duration was 379 days (range, 9–1609) for the overall group and 216 days (range, 10–1518) for the haGVHD+ group. Among the 34 patients diagnosed with haGVHD, 32 cases (94.2%) developed aGVHD (grade 2: 14 (43.8%); grade 3: 13 (40.6%); and grade 4: 5 (15.6%)). Among the 32 patients with aGVHD, 17 patients progressed to aGVHD without the complete resolution of haGVHD, while 15 patients developed aGVHD after a complete resolution of the initial events. Cytomegalovirus antigenemia was detected in 20 patients (60.6%) of 33 evaluable cases with the onset between day +22 and day +73. Chronic GVHD (cGVHD) developed in 21 (91.3%) of 23 evaluable patients with 13 extensive cGVHDs (60.9%) and 11 progressive cGVHDs (47.8%).

Risk factors for haGVHD

In the univariate analysis, the use of an alternative donor (P=0.010) and diagnosis of AA (P=0.012) were both identified as risk factors (Table 1). However, in the multivariate analysis, the use of an alternative donor (P=0.02, hazard ratio 4.714 (95% C.I., 1.453–15.299)) was the only significant risk factor for the development of haGVHD. Other risk factors, including nonmyeloablative conditioning, the stem cell source, MNC dose, and CD34+cell dose, were not found to be associated with the development of haGVHD.

Relevance of haGVHD with aGVHD and cGVHD

The association of haGVHD with aGVHD or cGVHD is shown in Table 1. HaGVHD was associated with aGVHD over grade 2 (P<0.001) or extensive cGVHD (P=0.007). The incidence of severe aGVHD (grade 3) in the haGVHD+ and haGVHD- groups was 56.2 and 16.7%, respectively (P<0.001). The incidence of extensive cGVHD in the haGVHD+ and haGVHD- groups was 60.9% and 26.3%, respectively (P=0.007).

Survival analysis

The 1- and 3-year OS were estimated as 56.28.9 and 49.210.2% in the haGVHD+ group, respectively. The 1- and 3-year DFS, and probability of relapse after 1 and 3 years were estimated as 52.69.1 and 32.510.5, and 8.76.0 and 21.513.1% in the haGVHD+ group, respectively. A comparative survival analysis of the OS and DFS did not reveal any difference between the haGVHD+ and haGVHD- groups, although an increasing trend of early deaths was observed in the haGVHD+ group (Figure 1). Meanwhile, the haGVHD+ group showed a lower probability of relapse than the haGVHD- group (P=0.0017). The causes of death within 90 days of transplant were GVHD (n=3), infection (n=3), and refractoriness of the primary disease (n=1), while those after 90 days were GVHD (n=4), infection (n=3), and HUS/TTP (n=1). For the overall group, the risk factors revealed by a multivariate analysis using Cox's hazard ratio regression model were advanced disease (P=0.026) for the OS, advanced disease (P=0.001) and cGVHD (P<0.001) for the DFS, and advanced disease (P<0.001), cGVHD (P<0.001), and haGVHD (P=0.005) for the probability of relapse (Table 3).

Figure 1.

Comparison of survival between haGVHD+ and haGVHD- groups. The 3-year OS (a), DFS (b), and probability of relapse rate (c) were 49.2, 39.5, and 21.7% in the haGVHD+ group and 39.9, 28.1, and 60.7% in the haGVHD- group, respectively.

Full figure and legend (157K)

Table 3 - Prognostic factors for OS, DFS, and probability of relapse using Cox's regression model.

Full table

Discussion

Immunological phenomena occurring during pre-engraftment periods do not necessarily describe a set of clinical abnormalities that fits a precise disease entity. Such phenomena can be transient, yet in some cases, they can result in devastating outcomes, including progressive severe GVHD, sustained marrow suppression, or graft rejection, even with intensive immunosuppressive treatment.^1,11

Sullivan et al¹ raised the issue of haGVHD in patients who received allogeneic BMT without immunosuppressive treatment. Tanaka et al⁸ also described haGVHD in allogeneic PBSCT following a reduced-intensity conditioning regimen th at occurred several days before engraftment (7–14 days after transplant). Plus, Colvin et al⁵ reported a 'haploimmunostorm' syndrome following reduced-intensity conditioning haploidentical PBSCT. Hyperpyrexia and malaise developed as early as 4 h after cell transfusion with a median time of 14 h to the onset of symptoms and signs. A skin rash occurred in 40% of patients, diarrhea in 20%, and transient elevation of liver enzymes in 40% within 6–24 h after infusion. As such, these observations can provide some insight into the mechanism of haGVHD. For example, HLA disparity can evoke an immunologic reaction in an allogeneic setting when the immunogenecity toward the graft-versus-host (GVH) direction or host-versus-graft (HVG) direction is suppressed partially.

The overall incidence of haGVHD (37.8%) in the current study seemed to be similar to that reported by other investigators in an allogeneic setting, including a 26.9% incidence reported by Choi et al⁶ and 28% incidence reported by Nurnberger et al.¹² Spitzer et al² noted the incidence of immunological phenomena in up to 75% of patients undergoing nonmyeloablative PBSCT. The reason for this high incidence may have been because a nonmyeloablative regimen cannot completely suppress a HVG or GVH reaction; however, PBSCs can further increase the immunological reaction via high T cell activity related to the graft compared to BM. Contrary to expectations, the current study noted a decreased incidence of haGVHD in patients undergoing PBSCT compared to patients undergoing BMT both in family matched and unrelated settings. Similarly, a lower incidence of severe aGVHD (grade 3) after PBSCT (P<0.05) than after BMT was also observed in an unrelated setting in the study by Elmaagacli et al.¹³ Donor pretreatment with G-CSF may have certain advantageous effects on the occurrence of haGVHD, as has been shown in murine models.^14,15 Yet, further investigation is still needed to explain this unexpected result.

In the current study, the usage of an alternative donor was found to be a risk factor for haGVHD in a multivariate analysis (P=0.02). The increased incidence of haGVHD in SCT using alternative donors can be explained by HLA mismatching. Especially in an unrelated setting, the mismatching of HLA-C, HLA-DP, or HLA-DQ can induce a more profound immune reaction post transplant. The diagnosis of non-AA (P=0.012) was also identified as another possible risk factor for haGVHD in a univariate analysis. The rarity of haGVHD in AA patients can be partially explained by the use of ATG for conditioning. ATG can suppress the function of the host and donor T cells, as such, the reduction of both blood progenitor and antigen-presenting cells in the marrow of patients with AA can decrease the immune reaction, thereby decreasing the incidence of haGVHD.

High-dose corticosteroids (3–5 mg/kg/day i.v.) for 5 days were found to be very effective in treating haGVHD, leading to an abrupt resolution of the fever in the majority of patients in the current study. Some patients who exhibited an excellent initial response to corticosteroids had a recurrence of haGVHD when the steroids were stopped or rapidly tapered, leading to disappointing clinical outcomes of early mortality, prolonged BM suppression, or graft rejection. According to Spitzer's report,² half of the patients who achieved complete resolution of ES with steroids experienced a recurrence of ES or the development of aGVHD when the steroids were tapered. In the current study, three patients who were not treated with corticosteroids failed to engraft with persistent clinical features of haGVHD and were finally lost. Accordingly, when haGVHD is suspected without a definite focus of infection or fever after allogeneic SCT, early intervention with corticosteroids would appear to improve transplantation outcomes though aiding in the differentiation of other conditions or minimizing the risk of rejection.

The current study revealed a close association between haGVHD and aGVHD (P<0.001) and extensive cGVHD (P=0.007). Choi et al⁶ also reported that the development of aGVHD and cGVHD was associated with the manifestations of haGVHD during the peri-engraftment period after allogeneic BMT, and that the higher incidence of cGVHD in the PECA positive group led to a lower possibility of relapse. In the current study, although the OS and DFS of the haGVHD group did not differ significantly from that of the non-haGVHD group, the haGVHD group exhibited a lower probability of relapse, while the OS curve revealed a decreasing survival benefit before day 100 after transplant, yet without statistical significance. There was a distinct difference in the causes of death between the haGVHD+ and haGVHD- groups (Table 1): 46.7% died due to GVHD in the haGVHD+ group, while 57.1% died due to recurrence in the haGVHD- group. Therefore, these results emphasize the seriousness of careful monitoring for GVHD.

In conclusion, the use of an alternative donor was found to be a risk factor for haGVHD in a multivariate analysis. The haGVHD was found to be associated with a higher incidence of aGVHD or cGVHD, and with a decreased probability of relapse (P=0.0017), yet not with a decrease in the OS or DFS.

References

1. Sullivan KM, Deeg HJ, Sanders J et al. Hyperacute graft-versus-host disease in patients not given immunosuppression after allogeneic marrow transplantation. Blood 1986; 67: 1172–1175. | PubMed | ISI | ChemPort |
2. Spitzer TR. Engraftment syndrome following hematopoietic stem cell transplantation. Bone Marrow Transplant 2001; 27: 893–898. | Article | PubMed | ISI | ChemPort |
3. Lee CK, Gingrich RD, Hohl RJ et al. Engraftment syndrome in autologous bone marrow and peripheral stem cell transplantation. Bone Marrow Transplant 1995; 16: 175–182. | Article | PubMed | ISI | ChemPort |
4. Cahill RA, Spitzer TR, Mazumder A. Marrow engraftment and clinical manifestations of capillary leak syndrome. Bone Marrow Transplant 1996; 18: 177–184. | PubMed | ISI | ChemPort |
5. Colvin GA, Lambert J-F, Lum LG et al. The Haplo-Immunostorm Syndrome: a new clinical entity observed in minimal immunotherapy with outpatient haploidentical bone marrow transplantation. Blood 2002; 100: 411a. | ISI |
6. Choi SJ, Lee KH, Lee JH et al. Peri-engraftment clinical abnormalities following allogeneic hematopoietic cell transplantation: a retrospective review of 216 patients. Bone Marrow Transplant 2003; 32: 809–813. | Article | PubMed | ISI |
7. Huang X, Chen Y, Guo N et al. Hyperacute graft versus host disease after allo-stem cell transplantation, analysis of 118 cases. Zhonghua Yi Xue Za Zhi 2002; 82: 511–514. | PubMed |
8. Tanaka Y, Kami M, Ogawa S et al. Hyperacute graft-versus-host disease and NKT cells. Am J Hematol 2000; 63: 60–61. | Article | PubMed | ISI | ChemPort |
9. Maiolino A, Biasoli I, Lima J et al. Engraftment syndrome following autologous hematopoietic stem cell transplantation: definition of diagnostic criteria. Bone Marrow Transplant 2003; 31: 393–397. | Article | PubMed | ISI | ChemPort |
10. Horn TD. Acute cutaneous eruptions after marrow ablation: roses by other names? J Cutan Pathol 1994; 21: 385–392. | Article | PubMed | ISI | ChemPort |
11. Takatsuka H, Takemoto Y, Yamada S et al. Complications after bone marrow transplantation are manifestations of systemic inflammatory response syndrome. Bone Marrow Transplant 2000; 26: 419–426. | Article | PubMed | ISI | ChemPort |
12. Nurnberger W, Willers R, Burdach S et al. Risk factors for capillary leakage syndrome after bone marrow transplantation. Ann Hematol 1997; 74: 221–224. | Article | PubMed | ISI | ChemPort |
13. Elmaagacli AH, Basoglu S, Peceny R et al. Improved disease-free-survival after transplantation of peripheral blood stem cells as compared with bone marrow from HLA-identical unrelated donors in patients with first chronic phase chronic myeloid leukemia. Blood 2002; 99: 1130–1135. | PubMed | ISI | ChemPort |
14. Pan L, Teshima T, Hill GR et al. Granulocyte colony-stimulating factor-mobilized allogeneic stem cell transplantation maintains graft-versus-leukemia effects through a perforin-dependent pathway while preventing graft-versus-host disease. Blood 1999; 93: 4071–4078. | PubMed | ISI | ChemPort |
15. Reddy V, Hill GR, Pan L et al. G-CSF modulates cytokine profile of dendritic cells and decreases acute graft-versus-host disease through effects on the donor rather than the recipient. Transplantation 2000; 69: 691–693. | Article | PubMed | ISI | ChemPort |

Acknowledgements

We thank Dr Thomas R Spitzer (Department of Hematology/Oncology, Massachusetts General Hospital, Harvard Medical School, Boston, USA) for his careful review and critical advice for this paper. The great work of our nursing staff is also gratefully acknowledged.