International Journal of Hematology

, 94:193

Pharmacokinetics-based optimal dose prediction of donor source-dependent response to mycophenolate mofetil in unrelated hematopoietic cell transplantation

Authors

  • Kanako Wakahashi
    • Division of Hematology, Department of MedicineKobe University Graduate School of Medicine
  • Motohiro Yamamori
    • Department of Clinical Pharmacy, School of Pharmacy and Pharmaceutical SciencesMukogawa Women’s University
    • Division of Hematology, Department of MedicineKobe University Graduate School of Medicine
  • Shinichi Ishii
    • Division of Hematology, Department of MedicineKobe University Graduate School of Medicine
  • Shinichirou Nishikawa
    • Division of Hematology, Department of MedicineKobe University Graduate School of Medicine
  • Manabu Shimoyama
    • Division of Hematology, Department of MedicineKobe University Graduate School of Medicine
  • Hiroki Kawano
    • Division of Hematology, Department of MedicineKobe University Graduate School of Medicine
  • Yuko Kawano
    • Division of Hematology, Department of MedicineKobe University Graduate School of Medicine
  • Yuriko Kawamori
    • Division of Hematology, Department of MedicineKobe University Graduate School of Medicine
  • Akiko Sada
    • Division of Hematology, Department of MedicineKobe University Graduate School of Medicine
  • Toshimitsu Matsui
    • Division of Hematology, Department of MedicineKobe University Graduate School of Medicine
  • Yoshio Katayama
    • Division of Hematology, Department of MedicineKobe University Graduate School of Medicine
Original Article

DOI: 10.1007/s12185-011-0888-6

Cite this article as:
Wakahashi, K., Yamamori, M., Minagawa, K. et al. Int J Hematol (2011) 94: 193. doi:10.1007/s12185-011-0888-6

Abstract

Mycophenolate mofetil (MMF) has been widely used for prophylaxis against graft-versus-host disease (GVHD) following allogeneic hematopoietic stem cell transplantation (allo-SCT). However, no clear advantage over methotrexate has been reported, other than reduced incidence of mucositis. We speculated that the wide inter-individual variation of plasma mycophenolic acid (MPA) levels veiled the benefits of MMF. Data from 36 unrelated allogeneic bone marrow (allo-BMT) and cord blood transplantation (CBT) were analyzed retrospectively based on MPA area under the curve (AUC0–24h). In allo-BMT, high AUC0–24h (>30 μg h/ml) resulted in no incidence of grade II–IV acute/extensive chronic GVHD and tended to show higher overall and disease-free survival, lower relapse rates, and non-relapse mortality. In CBT, AUC0–24h less than 30 μg h/ml was sufficient for low incidence of acute/chronic GVHD and high survival. Strong correlation between AUC0–24h and C2h, plasma MPA concentration at 2 h after administration was observed. Single point assessment of C2h was shown to provide a useful surrogate of AUC0–24h to predict GVHD incidence. The results of this study suggest that individualized MMF dosing in a donor source-dependent fashion may be important for maximizing the benefit of MMF in allo-SCT.

Keywords

MMFGVHD prophylaxisDrug monitoringUnrelated donor

1 Introduction

Graft-versus-host disease (GVHD) is a major cause of morbidity and mortality after allogeneic hematopoietic stem cell transplantation (allo-SCT). The combination of cyclosporine (CsA) or tacrolimus (FK506) with short-course methotrexate (MTX) after allo-SCT has been the gold standard for GVHD prophylaxis for more than 20 years [1, 2]. Common adverse effects of MTX in allo-SCT are mucositis and delayed neutrophil engraftment, both of which can be sometimes hardly of negligible clinical concern.

Mycophenolate mofetil (MMF) has recently been used as a new immunosuppressive agent for solid organ transplantation [3] and, subsequently, it is an alternative of MTX in allo-SCT for acute GVHD prophylaxis [414]. It is now well known that MMF does not cause mucositis and facilitates engraftment possibly attributed to its non-cytotoxicity, which avoids the major problems associated with MTX [7, 13, 15, 16].

On the basis of real-time pharmacokinetic monitoring, we have recently demonstrated that every 8-h administration of MMF after allo-SCT may maintain higher plasma levels of mycophenolic acid (MPA), the active metabolite of MMF, than that with every 12 h even with the same daily dose [15]. However, it is unclear whether higher plasma MPA levels actually result in more favorable outcome.

Few studies have proposed that there is a slight difference or none at all between CsA/MTX and CsA/MMF in terms of acute and chronic GVHD, relapse rate, and overall survival (OS) [8, 10, 12, 14, 17]. Pharmacokinetic studies in organ transplantation, as well as in allo-SCT, have demonstrated the wide inter-individual variations in plasma MPA and its pharmacokinetic parameter, the area under the curve (AUC) [15]. Therefore, it is possible that a certain population, with higher AUC for instance, in the same daily dose of MMF might display better clinical outcomes compared to that with MTX. To fully utilize the benefits of MMF, it is essential to isolate such a favorable population and to individualize treatment of other populations.

This study is a retrospective evaluation of the potential efficacy of the individualization depending on the levels of plasma MPA-AUC for hours 0–24 (AUC0–24h) in unrelated allogeneic bone marrow (allo-BMT) and cord blood transplantation (CBT).

2 Materials and methods

2.1 Patient characteristics on analysis of transplant outcome of unrelated allo-SCT in association with MPA-AUC

Between May 2005 and January 2009, a total of 50 allo-SCT patients [8 related peripheral blood stem cell transplantation, 2 related BMT, 17 unrelated BMT, and 23 unrelated CBT (single unit)] were enrolled in the MPA pharmacokinetic studies at the Kobe University Hospital. Eligibility requirements included age younger than 69 years and Eastern Cooperative Oncology Group (ECOG) performance status of two or lower. MMF was preferentially administered to patients based on one or more of the following criteria: those who underwent unrelated CBT, those who were expected to have severe infection from oral mucositis as a result of MTX use or MRSA carrier patients, or those who had high-risk disease features, including acute leukemia beyond first remission, high-risk myelodysplastic syndrome (refractory anemia with excess blasts), chronic myelogenous leukemia with accelerated phase, and chemotherapy-resistant hematological malignancies. Excluded from these protocols were patients who had serum creatinine of greater than 2.0 mg/dl, abnormal liver function with total bilirubin of greater than 2.0 mg/dl, or cardiac ejection fraction of less than 50%. Patients were also excluded if serology was positive for human immunodeficiency virus, or if uncontrolled infections were present. The Institutional Review Board at Kobe University Hospital approved the study protocols. Written informed consent was obtained from all patients.

Among these patients, all unrelated BMT and CBT cases, except inappropriate cases for the analyses of survival rates (one graft failure rescued by subsequent transplant, one flared acute GVHD due to uncontrollable compliance to FK506 in unrelated BMT, two graft failures rescued by subsequent transplant, or the spontaneous recovery of autologous hematopoiesis in CBT), relapse rate (one aplastic anemia), and incidences of GVHD (one early death due to sinusoidal obstructive syndrome in unrelated BMT and one early death before the engraftment due to infection in unrelated CBT, in addition to the excluded patients for the analysis of survival) were evaluated for the transplant outcome [acute and chronic GVHD, OS, disease-free survival (DFS), relapse rate, and non-relapse mortality] in association with the level of MPA-AUC0–24h and day 16 C2h (peak plasma MPA concentration at 2 h after MMF administration). The characteristics of these patients are summarized in Table 1.
Table 1

Patient characteristics, toxicities, and causes of death

 

AUC < 30 (N = 15)

AUC > 30 (N = 21)

Median age, years (range)

43 (33–60)

50 (20–66)

Gender (male:female)

10:5

10:11

Diagnosis

 AML

3

2

 ALL

4

7

 CML

1

1

 MDS

6

8

 NHL

1

2

 AA

0

1

Risk category

 BMT

  Standard

4

2

  High

6

3

 CBT

  Standard

3

9

  High

2

7

Conditioning regimen (no. of TBI containing)

 Myeloablative

12 (11)

11 (10)

 Non-myeloablative

3 (2)

10 (8)

Engraftment (median)

 BMT

Day 11

Day 11

 CBT

Day 20

Day 21

Toxicities

 BMT

  Diarrhea

6/10 (60%)

2/5 (40%)

  CMV-Ag

7/10 (70%)

3/5 (60%)

 CBT

  Diarrhea

4/5 (80%)

9/16 (56%)

  CMV-Ag

2/5 (40%)

12/16 (75%)

Causes of death

 BMT

  Relapse

2

0

  GVHD

2

0

  IP

2

0

  Viral infection

1

0

  SOS

0

1

 CBT

  Relapse

0

8

  Bacterial infection

0

1

  Viral infection

1

2

  SOS

1

0

  Others

0

1

AUC area under curve 0–24 h, AML acute myelogenous leukemia, ALL acute lymphoblastic leukemia, CML chronic myelogenous leukemia, MDS myelodysplastic syndrome, NHL non-Hodgkin lymphoma, AA aplastic anemia, IP interstitial pneumonitis, SOS sinusoidal obstructive syndrome, CMV-Ag cytomegalovirus antigenemia, Engraftment neutrophil engraftment

2.2 Allogeneic donors

Bone marrow from unrelated donors were 5/6 or 6/6 HLA-matched by DNA typing. Cord blood transplants were 3/6–6/6 HLA-matched and contained minimal cell count of 1.8 × 107 nucleated cells/kg recipient body weight before freezing.

2.3 Conditioning regimens and post-grafting immunosuppression

Conditioning regimens were diverse, wherein 23 of 36 patients received cyclophosphamide-based myeloablative conditioning regimens and the remaining 13 patients received fludarabine-based reduced-intensity conditioning regimens (Table 1). No one received a conditioning regimen containing T-cell depleting agents such as anti-thymocyte globulin. Thirty-one of 36 patients received total body irradiation (TBI). For immunosuppressive therapy, all patients received FK506 plus MMF except for one patient who received CsA until the onset of acute GVHD at day 20. FK506 at a daily dose of 0.03 mg/kg was administered by continuous intravenous infusion from day −1, and then shifted to oral formulation when tolerated. Doses were adjusted to maintain FK506 trough concentrations between 5 and 15 ng/ml. MMF was given orally 4–6 h after allo-SCT on day 0; then, on the succeeding days, it was given every 12 h (15–25 mg/kg/dose including dose-escalating cases up to a total daily dose of 3 g/day) or 8 h (fixed 1000 mg/dose, 3 g/day). MMF was generally discontinued on day 30, but in some patients it was continued beyond day 30 and tapered on an individual basis depending on the risk of relapse and the appearance of GVHD [18].

2.4 Pharmacokinetic analysis

Blood samples were collected at 0, 1, 2, 4, 8, and/or 12 h after the morning dose of MMF on days 2, 9, and 16 after allo-SCT in ethylenediaminetetraacetic acid (EDTA) tubes. The samples were centrifuged for 10 min at 4°C, and the plasma was stored at −80°C.

Total MPA levels were quantified by reverse-phase high-performance liquid chromatography (HPLC) with UV detection (Shimadzu, Kyoto, Japan). Non-compartmental analyses of total MPA concentration–time data were conducted to estimate the AUC0–12h (twice daily) and AUC0–8h (thrice daily) exactly as previously described [15]. The AUC0–24h of the group given twice daily were calculated by 2/3 × (the sum of AUC0–12h on days 2, 9, and 16), and those of the group given thrice daily were calculated by 3/3 × (the sum of AUC0–8h on days 2, 9, and 16).

A linear correlation of C2h on days 2, 9, and 16 after allo-SCT and AUC0–24h was assessed for the available data from all 50 patients. Similar analysis was performed with the data of C2h on day 16 and utilized for the evaluation of clinical outcome. As for clinical outcome, C2h on day 2 or day 9 in two cases was substituted for C2h on day 16 because of non-available data for various reasons.

2.5 Evaluation of toxicities, GVHD, and engraftment

MMF has been reported to be associated with gastrointestinal toxicity. Diarrhea of grade III or more, according to the National Cancer Institute Common Toxicity Criteria for Adverse Events version 3.0, was assessed, within 30 days after allo-SCT for descriptive purposes. All patients were also monitored for cytomegalovirus (CMV) reactivation on a weekly basis through measurement of CMV pp65 antigen. Preemptive therapy with ganciclovir was given when one or more antigen-positive cells per slide were observed. Acute and chronic GVHD were graded according to the consensus grading criteria [19, 20]. Chronic GVHD was assessed among patients who survived more than 100 days post-transplant. Neutrophil engraftment was defined as occurring on the first of two consecutive days after allo-SCT with neutrophil counts of more than 0.5 × 109/l.

2.6 Statistical analysis

The Mann–Whitney’s U test was used to compare the AUC0–24h between the groups of twice and thrice daily MMF administration schedule. The data for survival, relapse, non-relapse mortality and GVHD measured from the date of transplantation were examined using the Kaplan–Meier method and then compared using the log rank test. Correlation of AUC0–24h and C2h was assessed by Spearman’s correlation coefficient test. Significance was set at p < 0.05.

3 Results

3.1 Pharmacokinetics of MPA

The concentration–time profiles for plasma total MPA levels of the patients who received 3 g/day MMF twice or thrice daily are shown in Fig. 1. The patients who received 3 g/day divided twice daily were those enrolled in the dose-escalating study as described in Sect. 2. According to the pharmacokinetics on days 2 and 9 in each individual, the daily dose had been escalated to 3 g/day until day 14 [15]. Therefore, the analysis was performed using the data on day 16 post-transplant to compare the two groups, twice and thrice daily with the same daily dose. There was a markedly wide inter-individual variation of MPA levels (Fig. 1a, b). However, the patients with relatively high MPA level at 0 or 2 h (peak value) maintained high concentration along with the time, and the ones with low levels persisted through the monitoring period. The low intra-patient variability of plasma MPA levels was confirmed exactly in allo-SCT, as previously reported in the organ transplantation [21]. As shown in Fig. 1c, the AUC0–24h in the thrice-daily group showed a trend toward higher values than the twice-daily group [mean ± SEM: 27.52 ± 3.30 μg h/ml (n = 10) in twice vs. 44.80 ± 4.88 μg h/ml (n = 31) in thrice, p = 0.050]. Because the median AUC0–24h of all patients in both groups was 30.4 μg h/ml (n = 41), this study analyzed the incidence of GVHD, survival rate, relapse rate and non-relapse mortality in association with AUC0–24h higher or lower than 30 μg h/ml.
https://static-content.springer.com/image/art%3A10.1007%2Fs12185-011-0888-6/MediaObjects/12185_2011_888_Fig1_HTML.gif
Fig. 1

Pharmacokinetic results of MPA. Concentration–time profiles of total MPA on day 16 post-transplant in the groups given a 3 g/day twice a day (every 12 h, n = 10) and b 3 g/day three times a day (every 8 h, n = 31). c AUC0–24h in each group. Bar median

3.2 Correlation between AUC0–24h and C2h

The best correlation between AUC0–24h and C2h of MPA on days 2, 9, and 16 was observed (r2 = 0.657, p < 0.0001, Fig. 2a), suggesting that the value of C2h, which can be obtained by a single point assessment, may be a useful surrogate of AUC0–24h. According to the linear correlation, C2h 2.5 μg/ml represented AUC0–24h 30 μg h/ml (Fig. 2a). As a single assessment point of C2h, we picked up day 16 because it is about the date of engraftment and the initiation of acute GVHD after allo-SCT. Indeed, strong correlation between AUC0–24h and day 16 C2h was observed (r2 = 0.641, p < 0.0001), and again day 16 C2h 2.5 μg/ml represented AUC0–24h 30 μg h/ml (Fig. 2b). Thus, we analyzed clinical outcome depending not only on AUC0–24h, but also on day 16 C2h as a candidate of simplified surrogate of AUC0–24h.
https://static-content.springer.com/image/art%3A10.1007%2Fs12185-011-0888-6/MediaObjects/12185_2011_888_Fig2_HTML.gif
Fig. 2

Correlation between MPA-ACU0–24h and MPA concentration of 2 h (C2h). Data of C2h on a days 2, 9, and 16, and b day 16 only

3.3 Acute and chronic GVHD

Grades II–IV acute GVHD occurred in 10 out of 34 evaluable patients (29.4%) with onset between 21 and 57 days in unrelated allo-SCT. Subjects with AUC0–24h > 30 μg h/ml had a significantly lower cumulative incidence than those with AUC0–24h < 30 μg h/ml (15.8 vs. 46.7%, respectively, p < 0.05, Fig. 3a). This difference mainly originated from unrelated allo-BMT (0 vs. 60%, respectively, p = 0.07, Fig. 3b). In CBT, cumulative incidence of acute GVHD was low (20%) independent of AUC0–24h level (Fig. 3c). These results were reproducible by the analysis of day 16 C2h with more clear significance (Fig. 3a–c).
https://static-content.springer.com/image/art%3A10.1007%2Fs12185-011-0888-6/MediaObjects/12185_2011_888_Fig3_HTML.gif
Fig. 3

Cumulative incidence of acute GVHD. Cumulative incidence of grade II–IV acute GVHD in comparison between patients with AUC0–24h < 30 μg h/ml and AUC0–24h > 30 μg h/ml (left column), day 16 C2h < 2.5 μg/ml, and day 16 C2h > 2.5 μg/ml (right column). a Total unrelated allo-SCT patients, b unrelated allo-BMT, and c unrelated CBT

Extensive chronic GVHD occurred in 4 (13.3%) out of 30 evaluable patients in unrelated allo-SCT. Subjects with AUC0–24h > 30 μg h/ml had a significantly lower cumulative incidence than those with AUC0–24h < 30 μg h/ml (0 vs. 30.8%, respectively, p < 0.05, Fig. 4a). This difference mainly originated from unrelated allo-BMT (0 vs. 44.4%, respectively, p = 0.12, Fig. 4b). In CBT, no extensive chronic GVHD occurred regardless of AUC0–24h level (Fig. 4c). These results were reproducible by the analysis of day 16 C2h (Fig. 4a–c).
https://static-content.springer.com/image/art%3A10.1007%2Fs12185-011-0888-6/MediaObjects/12185_2011_888_Fig4_HTML.gif
Fig. 4

Cumulative incidence of chronic GVHD. Cumulative incidence of extensive chronic GVHD in comparison between patients with AUC0–24h < 30 μg h/ml and AUC0–24h > 30 μg h/ml (left column), day 16 C2h < 2.5 μg/ml, and day 16 C2h > 2.5 μg/ml (right column). a Total unrelated allo-SCT patients, b unrelated allo-BMT, and c unrelated CBT

3.4 OS, DFS, relapse rate, and non-relapse mortality

In unrelated allo-BMT, there was a trend toward better OS and DFS at 3 years in the subjects with AUC0–24h > 30 μg h/ml compared to those with AUC0–24h < 30 μg h/ml (80%; 95% CI 44.9–100 vs. 40%; 95% CI 9.6–70.4%, respectively) with a median follow-up of 646 days (range 22–1633 days), although none of these data reached statistical significance (p = 0.243, Table 2). Relapse rate was low in both groups (0% in AUC0–24h > 30 μg h/ml and 22.2% in AUC0–24h < 30 μg h/ml). Non-relapse mortality in the subjects with AUC0–24h > 30 μg h/ml tended to be lower compared to those with AUC0–24h < 30 μg h/ml. Similar results were obtained from the analysis of day 16 C2h.
Table 2

Univariate analysis of pharmacological variates affecting transplantation outcome

 

Patients, n

Median follow-up days (range)

% at 3 years after transplantation (95% CI)

n

AUC < 30 μg h/ml

n

AUC > 30 μg h/ml

p

n

C2h < 2.5 μg/ml

n

C2h > 2.5 μg/ml

p

OS

 BMT

15

646 (22–1633)

10

40 (9.6–70.4)

5

80 (44.9–100)

 

9

44.4 (13.6–71.9)

6

66.7 (19.5–90.4)

 
       

0.243

    

0.360

 CBT

21

277 (11–1442)

5

80 (44.9–100)

16

30 (6.9–53.1)

 

7

28.6 (4.1–61.1)

14

49.0 (21.6–71.7)

 
       

0.054

    

0.172

DFS

 BMT

15

646 (22–1633)

10

40 (9.6–70.4)

5

80 (44.9–100)

 

9

44.4 (13.6–71.9)

6

66.7 (19.5–90.4)

 
       

0.243

    

0.360

 CBT

21

179 (11–1442)

5

80 (44.9–100)

16

18.8 (0–37.9)

 

7

28.6 (4.1–61.1)

14

35.7 (13.0–59.4)

 
       

0.046

    

0.290

Relapse

 BMT

15

646 (22–1633)

10

22.2 (6.1–63.5)

5

0

 

9

12.5 (1.9–61.3)

6

20 (3.1–79.6)

 
       

0.331

    

0.774

 CBT

20

188 (50–1442)

5

0

15

72 (46.0–92.8)

 

6

58.3 (23.3–94.4)

14

52.4 (27.6–81.8)

 
       

0.031

    

0.429

NRM

 BMT

15

646 (22–1633)

10

48.6 (21.4–84)

5

20 (3.1–79.6)

 

9

49.2 (21.9–84.3)

6

16.7 (2.5–72.7)

 
       

0.440

    

0.220

 CBT

21

277 (11–1442)

5

20 (3.1–79.6)

16

25.5 (10.4–54.6)

 

7

28.6 (8.0–74.2)

14

21.4 (7.2–52.8)

 
       

0.816

    

0.607

p values were calculated with Kaplan–Meier method

Values in italics represent statistically significant difference (p < 0.05)

Surprisingly, these results appeared to be completely adverse in CBT. The subjects with AUC0–24h > 30 μg h/ml showed remarkably lower OS (30%; 95% CI 6.9–53.1%, p = 0.054) and DFS (18.8%; 95% CI 0–37.9%, p < 0.05) due mainly to high incidence of relapse (72%, p < 0.05, Table 2), whereas the group with AUC0–24h < 30 μg h/ml, although the number of patients was small, displayed continuous high OS and DFS (80%; 95% CI 44.9–100%) with no relapse. Day 16 C2h, however, failed to extract this high-risk population of relapse-related mortality. Non-relapse mortality was low in CBT regardless of AUC0–24h or day 16 C2h. The causes of death are summarized in Table 1.

3.5 Toxicities and neutrophil engraftment

No patient developed grade III or IV oral mucositis in all subjects given MMF. Major gastrointestinal toxicity was diarrhea. The incidence of grade III or more diarrhea was 58.3% (21/36), but there was no significant difference between the groups with high and low AUC0–24h (Table 1). It might be that the incidence of grade III or more diarrhea was proportionately higher (60% in BMT and 80% in CBT) in the subjects with AUC0–24h < 30 μg h/ml. It was supposedly due to myeloablative conditioning regimen, which was employed to more patients (12/15, 80%) with AUC0–24h < 30 μg h/ml compared to those with AUC0–24h > 30 μg h/ml (11/21, 52%). For most cases, diarrhea was transient and curable without the withdrawal of MMF.

Although relatively high incidence of CMV reactivation was observed in both high and low AUC groups (Table 1), most cases were controllable with preemptive therapy. Death of CMV pneumonia occurred in two (13.3%) of the subjects with AUC0–24h < 30 μg h/ml and in two (9.5%) of those with AUC0–24h > 30 μg h/ml (Table 1). Higher AUC0–24h did not facilitate neutrophil engraftment (Table 1).

4 Discussion

This study discovered a certain population in which MMF would be highly beneficial in unrelated allo-SCT because of wide inter-patient variability of plasma MPA levels even with the same MMF exposure. The benefit is predictable by the pharmacokinetic parameters during the early days after transplantation. Thus, we propose that the MMF dosing strategies in unrelated allo-SCT should be changed depending on donor source and individual monitoring of AUC0–24h or C2h.

4.1 Unrelated BMT

The incidence of acute GVHD in patients with AUC0–24h < 30 μg h/ml (Fig. 3b) was similar to or even higher than the data of Japan Marrow Donor Program 2007 (40 ± 2%, 3944 HLA-matched unrelated BMTs, wherein MTX was used in the vast majority of these patients), whereas no GVHD was observed in patients with AUC0–24h > 30 μg h/ml (Fig. 3b). In unrelated BMT, MPA-AUC0–24h should be maintained at more than 30 μg h/ml to surpass MTX for acute GVHD prophylaxis. Most acute GVHD progressed to chronic GVHD (Fig. 4b), which implies the importance of suppression of prior acute GVHD by high AUC0–24h to avoid subsequent chronic GVHD. Importantly, the high AUC0–24h did not increase the risk of relapse (Table 2) or one of the possible adverse effects of MMF, which is CMV infection. Moreover, the greater AUC did not increase the incidence or grade of diarrhea or delay neutrophil engraftment. Thus, the diarrhea could not be mainly caused by administration of MMF. It was difficult to interpret why the incidence of relapse was lower in patients with higher AUC after unrelated BMT (Table 2). Indeed, no grade II–IV acute GVHD and chronic extensive GVHD occurred in this group except for one early death. Nevertheless, grade I skin GVHD and limited chronic GVHD occurred in most of the patients. Therefore, higher AUC in BMT might suppress severe GVHD, but not immunological effect per se (i.e., graft-versus-leukemia effect: GVL). There was a trend toward higher survival in patients given a greater dose of AUC (Table 2). On the other hand, the major causes of death in patients given low AUC0–24h other than relapse were GVHD and interstitial pneumonitis, both of which were thought to be due to immunological reaction and might be avoidable with higher AUC0–24h. Collectively, MMF is particularly useful in unrelated allo-BMT patients if AUC0–24h is higher than 30 μg h/ml. In lower AUC0–24h patients, it would be a reasonable strategy to increase AUC0–24h to expect low risk of acute and chronic GVHD resulting in high incidence of survival with no increased risk of relapse. Alternatively, additional immunosuppressant(s), such as steroids, should be considered for preemptive therapy against GVHD.

4.2 Unrelated CBT

We have reported the usefulness of MMF for CBT as a substitute for MTX for avoidance of graft failure [7]. In addition, the incidence of acute and chronic GVHD was very low even in the low AUC0–24h group (Figs. 3c, 4c). Thus, there is no doubt that MMF is quite useful to avoid early complications after CBT. Patients with AUC0–24h > 30 μg h/ml showed lower survival due to higher incidence of relapse (Table 2). However, there was no difference in survival and relapse rate if patients were divided into two groups, lower and higher C2h at day 16 (Table 2). It is important to examine whether the excessive immunosuppression by MMF might result in the inhibition of graft-versus-leukemia effects in CBT or not. In contrast to unrelated allo-BMT, AUC0–24h lower than 30 μg h/ml is good enough for GVHD prophylaxis and this condition may even be critical to avoid the relapse in CBT. Dose reduction may be required in some cases with high AUC0–24h. In this study, all CBTs were performed with single unit cord blood. Because the incidence of acute GVHD is relatively high in double unit CBT [22], the optimal strategy of MMF administration might be close to that for unrelated allo-BMT.

The incidence of CMV pneumonia might be unexpectedly high despite the preemptive administration of anti-CMV agents; three were CBT cases. A previous report also demonstrated higher unbound MPA concentration steady state (Css) associated with CMV reactivation [13]. Therefore, we should carefully deal with CMV antigenemia-positive patients, especially in a CBT setting. Appropriate tapering of MMF after engraftment and relatively intense preemptive treatment could be required.

Recently, it has been pointed out that the Kaplan–Meier method, which treats the competing events as censored at the time they occurred, was biased when these incidences were relatively high [23]. Therefore, we also reassessed acute and extensive chronic GVHD, and relapse and non-relapse mortality by competing risk analysis using Gray’s method (statistical software R version 2.13.0). The incidences of acute and extensive chronic GVHD analyzed by Gray’s method are almost identical to those of Kaplan–Meier method. In relapse and non-relapse mortality, the incidences by Gray’s method are relatively low but similar to those by Kaplan–Meier method, except for relapse rate according to AUC0–24h in CBT. Subjects with AUC0–24h < 30 μg h/ml and AUC0–24h > 30 μg h/ml were 0 and 56.7% (95% CI 25.4–79.1%), respectively. Although the incidence was relatively low compared to the result of the analysis using Kaplan–Meier method, the statistical significance was preserved (p = 0.046).

Pharmacokinetics monitoring to calculate AUC0–24h is a heavy burden on the patient, is time-consuming and not practical. This study demonstrated that a single point monitoring of MPA concentration at 2 h after administration, C2h, is a good substitute of AUC0–24h, particularly for the assessment of acute and chronic GVHD (Figs. 2, 3, 4). The day on which to check the MPA C2h after allo-SCT may not be necessarily rigid because AUCs on days 2, 9, and 16 were similar in each individual because of the relatively low intra-patient variability [15, 24]. Although day 16 C2h might be much too simplified to extract the high-risk population of relapse in CBT, this study raises the possibility of significantly improving allo-SCT outcome with an individualized MMF administration strategy by checking this simple parameter.

As a strategy to keep the AUC high in each individual, the increase in dosing frequency rather than dose escalation is likely effective in allo-SCT patients (Fig. 1c) [15]. MPA levels in allo-SCT are lower compared with those in the transplantation of solid organs [25]. Gastrointestinal damage induced by prior chemotherapy, preparative regimens, and gut GVHD has been speculated to result in inadequate absorption. However, it was unlikely because MPA/AUC levels by intravenous administration of MMF were as low as those by oral MMF in allo-SCT (data not shown, our pilot study). Recently, de Winter et al. [25] have reported that this difference might originate from different MPA clearance (higher efficiency with much wider inter-individual variations) depending on plasma albumin concentrations and predose concentrations of calcineurin inhibitors. It may be important to consider these factors in cases that need higher AUC0–24h, for example changing the dose of calcineurin inhibitors to adjust the MPA levels. The present study is a retrospective one composed of small sample number and heterogeneous patient characteristics. Thus, prospective randomized study of individualized administration is necessary to confirm the veiled usefulness of MMF.

Acknowledgments

This work was supported, in part, by the Grants-in-Aid for Scientific Research from the Ministry of Health, Welfare, and Labor in Japan.

Copyright information

© The Japanese Society of Hematology 2011