Objectives: To assess differences between the rates of acute graft-versus-host disease (aGVHD), relapse, non-relapse mortality (NRM) and overall survival (OS) at low (4–6 ng/ml) versus standard (8 – 12 ng/ml) tacrolimus blood levels. The central hypothesis was that a lower level of tacrolimus would allow a greater graft-versus-leukemia effect.

Methods: Prospective, Phase II/III trial, for patients with hematologic malignancies undergoing matched related (6/6) sibling donor transplant. Patients were randomized to one of 2 arms: Study arm: tacrolimus level kept at 5 ng/ml +/− 20% (LL, low level group) and Control arm: tacrolimus level kept at 10 ng/ml +/− 20% (SL, “standard” level group). Both arms started GVHD prophylaxis on day -2 pre-transplant with IV tacrolimus plus methotrexate at 5 mg/m2 on days 1, 3, 6 and 11. All patients were on study for 3 months post transplant and started tacrolimus taper at 3 months, with the aim to be off immunosuppressive therapy at 4 months if possible.

Statistical Methods: Kaplan-Meier estimates and logrank tests were used to assess and compare the rates of aGVHD, NRM and OS and GVHD-free survival between the tacrolimus LL and SL groups. The effects of tapering tacrolimus blood level were assessed by fitting a piecewise hazard within each treatment group that changed at the times tapering was started and completed in each patient.

Results: A total of 46 patients (22 females, 24 males) with MDS/AML (n= 44), CML (n= 1) and NHL (n= 1), were randomized to either the LL (n= 25) or SL (n= 21) group. Nineteen were in CR/chronic phase (LL 11, SL 8), 25 had refractory disease or untreated relapse (LL 12, SL 13) and 2 patients in the LL group were untreated. The incidence of acute GVHD grade 2–4 was 48% (12/25) in the LL group compared to 19% (4/21) in the SL group (p= 0.01). Four patients developed grade 3–4 acute GVHD, and all of them were in the LL group. The relapse rates were 40% (10/25) in the LL group and 24% (5/21) in the SL group (p= 0.35). Non-relapse mortality did not differ between the two groups (LL 4/25 SL 3/21, p= 0.73) but relapse-related mortality was significantly higher in the LL group (6/25), compared with the SL group (0/21, p=0.01). GVHD-free survival at 6 months was significantly lower in the LL group (23%, 95% CI 0.09–0.59) compared to the SL group (50%, 95% CI 0.3–0.83). When failure was defined as aGVHD or death, the SL group had a relative risk (RR) of 0.36 compared to the LL group (p= 0.02). When failure was defined as relapse or death, the RR of failure in the SL group was 0.30 compared to the LL group (p= 0.002). For both types of failure, the timing of immunosuppression taper did not significantly affect the risk of any type of failure.

Conclusions: In this prospective, randomized study the use of lower blood levels of tacrolimus was associated with a substantially worse outcome, including higher rates of aGVHD, relapse and death. Based on these findings and our previous retrospective analysis (Biol Blood Marrow Transplant. Feb 11:108–14), we recommend a blood level of tacrolimus between 8–12 ng/dl during the first 3 months post transplantation for the prophylaxis of aGVHD. The finding of a higher relapse rate in the LL group deserves further investigation.

Author notes

Corresponding author