A Retrospective Comparison of Tacrolimus versus Cyclosporine with Methotrexate for Immunosuppression after Allogeneic Hematopoietic Cell Transplantation with Mobilized Blood Cells

Abstract

This retrospective study was carried out to compare results with tacrolimus versus cyclosporine in combination with methotrexate for immunosuppression after allogeneic hematopoietic cell transplantation (HCT) with granulocyte colony-stimulating factor-mobilized blood cells. The cohort included 456 consecutive patients who received first allogeneic T-cell replete HCT with mobilized blood cells from related or unrelated donors after high-intensity conditioning for treatment of hematological malignancies. Study endpoints included grades II–IV acute graft-versus-host disease (GVHD), grades III–IV acute GVHD, chronic GVHD, end of treatment for chronic GVHD, overall mortality, disease-free survival, recurrent malignancy and nonrelapse mortality. Adjusted multivariate Cox regression analysis showed no statistically significant differences between tacrolimus and cyclosporine for any of the endpoints tested. Although the size of the cohort is not sufficient to exclude clinically meaningful differences in outcomes, these results support the continued use of cyclosporine at centers that have not adopted tacrolimus as the standard of care after HCT with mobilized blood cells after high-intensity conditioning regimens. A larger registry study should be carried out to provide more definitive information comparing outcomes with the two calcineurin inhibitors.

INTRODUCTION

Graft-versus-host disease (GVHD) is a frequent immunologic complication after allogeneic hematopoietic cell transplantation (HCT). Cyclosporine or tacrolimus in combination with other agents represent widely accepted standards of care as immunosuppressive regimens after HCT. Results of open-label randomized prospective phase III studies have indicated that the risk of grades II–IV acute GVHD after bone marrow transplantation with related or unrelated donors was lower with the use of tacrolimus than with cyclosporine, in combination with methotrexate [1-3]. Overall survival appeared similar between the groups [1, 3], although an imbalanced distribution of risk factors confounded the interpretation of results in one of the studies [2, 4].

The source of cells used for HCT has changed since the 1990s, when the trials comparing tacrolimus and cyclosporine were done. Granulocyte colony-stimulating factor (G-CSF)-mobilized blood cells are now used more frequently than bone marrow cells for HCT in adults with high-intensity conditioning [5, 6]. The different composition of grafts between bone marrow and mobilized blood cells might influence the relative efficacy of the two calcineurin inhibitors. Outcomes with the two calcineurin inhibitors after HCT with G-CSF-mobilized blood cells after high-intensity conditioning have not been compared previously. In addition, recent practices such as the lower targeted concentration of tacrolimus and the use of ursodeoxycholic acid to prevent hepatic complications could affect the relative efficacy of the two calcineurin inhibitors.

METHODS

Patients

The study cohort included 456 consecutive patients who received first allogeneic T-cell replete HCT with G-CSF-mobilized blood cells from related or unrelated donors after high-intensity conditioning for treatment of hematological malignancies between 7/1/2003 and 12/31/2009 at our center. High-intensity (i.e., “myeloablative”) conditioning regimens included fractionated total body irradiation at ≥12.0 Gy or busulfan at 4 mg/kg/day or 130 mg/m²/day for 4 consecutive days. Only patients who received either tacrolimus or cyclosporine in combination with methotrexate for GVHD prophylaxis were included. Patients who received antithymocyte globulin to prevent GVHD were excluded. Patients signed consent forms allowing medical records to be used for research, and the IRB approved the study.

Post-transplant Management

All patients received ursodeoxycholic acid from 2 weeks before conditioning until 90 days after HCT to prevent hepatic complications, and all patients received immunosuppression with either tacrolimus or cyclosporine in combination with methotrexate after HCT. Methotrexate was administered at 15 mg/m² on day 1 and at 10 mg/m² on days 3, 6 and 11 after HCT. Between 2003 and 2005, both tacrolimus and cyclosporine were used in parallel depending on the particular protocol. Thereafter, tacrolimus was the standard adopted for most protocols at our center. Cyclosporine dosing was adjusted to maintain whole blood trough concentrations at 150 to 450 ng/mL, as measured by immunoassay. Tacrolimus dosing was adjusted to maintain concentrations at 5 to 15 ng/mL, as measured by a liquid chromatography-tandem mass spectrometry assay for the parent compound.

Study Endpoints

Study endpoints included grades II–IV acute GVHD, grades III–IV acute GVHD, chronic GVHD, end of systemic treatment for chronic GVHD (i.e., duration of treatment for patients with chronic GVHD), overall mortality, disease-free survival, recurrent malignancy and nonrelapse mortality. Acute GVHD was graded according to previously described criteria [7], and chronic GVHD was diagnosed according to NIH criteria [8]. Grade II GVHD was subdivided according to the extent of liver and skin involvement [9]. Most patients with grade IIa GVHD had upper gastrointestinal symptoms, with or without stage 1 or stage 2 skin involvement (i.e., <50% body surface affected by rash) and no liver involvement. Grade IIb GVHD included patients with stage 3 skin disease (≥50% body surface affected by rash) or stage 1 liver involvement.

Statistical Analysis

Kaplan-Meier estimates were used for overall survival and disease-free survival, and cumulative incidence estimates were used for other endpoints [10]. Multivariate Cox regression models were used to evaluate hazard ratios for results with tacrolimus as compared to cyclosporine. The models were adjusted for patient age, donor type, recipient and donor gender combination, disease type, disease risk category, use of total body irradiation in the conditioning regimen, and year of HCT. The analysis was carried out as of July, 2010.

RESULTS

Patient Characteristics

The median age of patients was 47 years (range, 1 to 66 years). Of the 456 patients, 232 (51%) had HLA-identical related donors, 187 (41%) had HLA-matched unrelated donors, and 37 (8%) had HLA-mismatched related or unrelated donors. The distribution of diagnoses at transplant, HLA-matching and donor type, and year of transplantation showed statistically significant differences according to type of immunosuppression (Table 1).

Table 1.

Patient Characteristics According to Type of Immunosuppression

Characteristic	Tacrolimus (n = 292)	Cyclosporine (n = 164)	P-value
Patient age (y), median (range)	47 (1-66)	46 (9-64)	.60
Diagnosis at transplant, no. (%)			<.0001
Acute myeloid leukemia	125 (43)	59 (36)
Acute lymphoblastic leukemia	51 (17)	21 (13)
Chronic myeloid leukemia	16 (5)	31 (19)
Myelodysplastic syndromes or myeloproliferative neoplasms	86 (29)	40 (24)
Other (CLL, HD, NHL, MM)	14 (5)	13 (8)
Disease risk category at transplant*, no. (%)			.31
Low risk	168 (58)	86 (52)
Intermediate risk	14 (5)	13 (8)
High risk	110 (38)	65 (40)
Donor / patient gender, no. (%)			.49
Female / male	78 (27)	39 (24)
Other	214 (73)	125 (76)
High intensity conditioning regimens, no. (%)			.98
with total body irradiation (≥1200 cGy)	100 (34)	56 (34)
without total body irradiation	192 (66)	108 (66)
HLA-matching and donor type, no. (%)			<.0001
HLA-identical related	147 (50)	85 (52)
HLA-matched unrelated	135 (46)	52 (32)
HLA-mismatched	10 (3)	27 (16)
Year of transplantation, no. (%)			<.0001
2003-2004	2 (1)	132 (80)
2005-2006	118 (40)	26 (16)
2007-2009	172 (59)	6 (4)

CLL indicates chronic lymphocytic leukemia; HD, Hodgkin lymphoma; NHL, non-Hodgkin lymphoma; MM, multiple myeloma; HLA, human leukocyte antigen.

^*The low-risk category included chronic myeloid leukemia in chronic phase, acute leukemia in remission, and refractory anemia without excess blasts. The intermediate-risk category included CLL, HD, NHL and MM. The high-risk category included all other diseases and stages.

Transplant Outcomes

The median duration of follow-up among survivors was 45 months (range, 3 to 85 months) after HCT. Cumulative incidence rates of grades II–IV, IIb–IV and III–IV acute GVHD at 120 days were 71%, 55% and 11%, respectively, and the cumulative incidence of chronic GVHD at 3 years after HCT was 44%. The cumulative incidence of discontinued systemic treatment for chronic GVHD at 2 years after the onset of treatment for chronic GVHD was 16%. The remaining patients with chronic GVHD either continued treatment for more than 2 years or died or had recurrent malignancy during continued treatment within 2 years after the onset of the disease. Overall survival and disease-free survival rates at 3 years after HCT were 62% and 55%, respectively. The cumulative incidence rates of recurrent malignancy and nonrelapse mortality at 3 years after HCT were 28% and 18%, respectively.

Multivariate Analysis of Outcomes with Tacrolimus As Compared to Cyclosporine

Adjusted multivariate Cox regression analysis showed no statistically significant differences between tacrolimus and cyclosporine for any of the endpoints tested (Table 2). Further analysis did not identify any subgroup according to age, diagnosis, disease risk, recipient and donor gender combination, conditioning regimen, or donor type in which the risk of grades II–IV GVHD was lower with tacrolimus than with cyclosporine (data not shown).

Table 2.

Multivariate Analysis of Outcomes with Tacrolimus As Compared to Cyclosporine

Endpoint	No. of Events	Hazard Ratio* (95% CI)	P-value
Grades II–IV acute GVHD	325	1.12 (0.80-1.55)	.52
Grades IIb–IV acute GVHD†	158	0.81 (0.50-1.33)	.41
Grades III–IV acute GVHD	51	1.78 (0.75-4.24)	.19
Chronic GVHD	190	0.89 (0.58-1.38)	.61
End of systemic treatment for chronic GVHD	48	0.83 (0.29-2.42)	.73
Overall mortality	177	0.96 (0.62-1.49)	.86
Disease-free survival	211	1.09 (0.73-1.64)	.67
Recurrent malignancy	125	1.36 (0.80-2.30)	.25
Nonrelapse mortality	86	0.82 (0.43-1.57)	.54

CI indicates confidence interval; GVHD, graft-versus-host disease.

^*The models were adjusted for patient age, donor type, recipient and donor gender combination, disease type, disease risk category, use of total body irradiation in the conditioning regimen, and year of transplantation.

^†See Methods for definition of grade IIb GVHD.

DISCUSSION

Our retrospective results showing that tacrolimus offered no statistically significant advantage over cyclosporine for preventing grades II–IV acute GVHD contrast with results from 3 previous randomized prospective studies all showing that the risk of acute GVHD after bone marrow transplantation was lower with the use of tacrolimus as compared to cyclosporine [1-3]. At least two explanations might account for the different results. First, the higher risks of acute and chronic GVHD after HCT with G-CSF-mobilized blood cells as compared to bone marrow might influence the relative efficacy of the two calcineurin inhibitors [11-14]. Second, the targeted whole blood concentration of tacrolimus used at our center was lower than the concentrations evaluated in the prospective trials (10-40, 10-30 and 20-25 ng/ml, respectively) [1-3], but other studies have suggested that tacrolimus levels >20 ng/ml increased the risk of toxicity without decreasing the risk of GVHD [15, 16].

Our results showing no statistically significant differences with the two calcineurin inhibitors in overall survival, disease-free survival and recurrent malignancy after mobilized blood cell transplantation are consistent with results from previous studies of bone marrow transplantation [1, 4]. A randomized prospective Japanese study showed an increased incidence of recurrent malignancy with tacrolimus after bone marrow transplantation from related donors [3], but this effect was not observed in a large retrospective Japanese study [17] or in a randomized prospective U.S. study [2].

The retrospective Japanese study showed that the risk of nonrelapse mortality after bone marrow transplantation from unrelated donors was lower with tacrolimus than with cyclosporine [17]. In that study, the lower risk of nonrelapse mortality in the tacrolimus group was associated with a lower risk of acute GVHD and higher survival compared to the cyclosporine group. The Japanese results differ from those observed in our retrospective study, suggesting that the type of graft or the ethnicity of patients might influence the relative efficacy of calcineurin inhibitors.

The routine use of prophylactic ursodeoxycholic acid started at our center in 2003 [18, 19] and has decreased the incidence of acute GVHD and hepatic complications (data not shown). The choice of calcineurin inhibitor was changed to tacrolimus for virtually all transplant protocols with high-intensity conditioning at our center in early 2005. While we do not believe that any other transplant practices have changed during the study period, we cannot completely exclude effects of unknown confounders even if the multivariate analysis is adjusted for year of transplantation. The incidence of grades II–IV acute GVHD at our center is higher than reported at other centers, reflecting a high sensitivity for diagnosing upper gastrointestinal GVHD [20]. Even so, we believe that our results showing no difference in the risk of grades II–IV acute GVHD are likely to be representative of experience at other centers, since we found no statistically significant advantage for tacrolimus in preventing more severe grades IIb–IV or III–IV GVHD.

Our results support the continued use of cyclosporine at centers that have not adopted tacrolimus as the standard of care after HCT with mobilized blood cells after high-intensity conditioning regimens, but the size of our cohort is not sufficient to exclude clinically meaningful differences in outcomes with the two calcineurin inhibitors. A larger registry study should be carried out to provide more definitive information comparing outcomes with the two calcineurin inhibitors.