Study design

The study protocol has been previously published29 and registered in ClinicalTrials.gov (Identifier: NCT04061876). This was an open-label, multicenter, randomized phase 3 controlled trial that enrolled patients with newly diagnosed aGVHD who required initial systemic immunosuppressive therapy. The enrollment started on August 25, 2019, and was completed on June 1, 2022, involving seven Chinese transplantation centers. Patients were randomly 1:1 to receive either ruxolitinib (5 mg/day) plus methylprednisolone (1 mg/kg/day) or methylprednisolone (2 mg/kg/day) alone. The randomization was stratified by pre-transplant disease status (complete remission vs. not) and aGVHD risk (intermediate vs. high). Intermediate aGVHD risk was defined by the low risk of Minnesota aGVHD Risk Score4 and intermediate biomarker risk. High aGVHD risk was defined by the high risk of Minnesota aGVHD Risk Score4 or high biomarker risk. The diagnosis and response are clinically based and not histology based. Stratified permuted block randomization lists ensured balanced stratification. Randomization was performed through an independent interactive web-based response system, with codes generated by a statistician not involved in the study or site operations. This study was open-label, and both the investigators and participants were aware of the treatment groups due to differences in the tapering strategies for CsA and methylprednisolone. However, the study staff responsible for data analysis and outcome assessments were blinded to the treatment allocations.29

Endpoints

The primary endpoint was the overall response rate (ORR) to aGVHD treatment at 28 days post-enrollment.39 ORR was defined as the percentage of patients in each group who attained either a partial response (PR) or complete response (CR) without requiring additional immunosuppressive agents. In aGVHD, CR was characterized by the complete absence of aGVHD symptoms, while PR was defined as an improvement of at least one stage in a single organ without worsening in others. No response (NR) was classified as either no improvement, worsening symptoms in any organ, or the emergence of new GVHD-related symptoms. Additionally, GVHD progression after 3 days of therapy or lack of improvement within 7 days was also considered NR.

Secondary endpoints included ORR on day 56 and the duration of response (DOR) at 6 months, non-relapse mortality (NRM), incidence of relapse, recurrent aGVHD, incidence of chronic GVHD, overall survival (OS), disease-free survival (DFS), and failure-free survival (FFS, detail information in Supplemental Methods). Secondary systemic therapy for aGVHD was assessed within 6 months after enrollment. The dose and duration of corticosteroids used were analyzed. The safety endpoints were assessed by the frequency of AEs, which were as defined according to the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE, version 4.0, Supplemental Methods).

Patients and eligibility

Inclusion criteria included: (1) age between 14 and 65 years; (2) newly diagnosed aGVHD; and (3) intermediate- or high-risk aGVHD, as determined by the Minnesota aGVHD Risk Score4 (high risk) or biomarker risk classification (intermediate- or high-risk).5,28,29,40 Exclusion criteria comprised: (1) chronic GVHD; (2) late aGVHD after donor lymphocyte infusion; (3) prior systemic immunosuppressive therapy for aGVHD; (4) contraindications to methylprednisolone; or (5) treatment with JAK inhibitor therapy after graft infusion. The study was approved by the Ethics Committee of the Chinese People’s Liberation Army General Hospital, and all participating patients provided informed consent prior to enrollment.

InterventionGVHD prophylaxis and supportive therapy

The enrolled patients had been diagnosed with hematologic malignancies and had undergone allogeneic peripheral blood stem cell transplantation. The procedures for hematopoietic stem cell mobilization, collection, conditioning regimens, and GVHD prophylaxis were conducted as previously described19,41,42,43 Most transplantation recipients received rabbit anti-thymocyte globulin (rATG), cyclosporine A (CsA), mycophenolate mofetil, and short-term methotrexate for GVHD prophylaxis (detail information in Supplemental Methods). Quantification of EBV and CMV DNA was performed by polymerase chain reaction (PCR) analysis twice weekly until 3 months after transplantation. Supportive care was provided as previously described.19

GVHD treatment

Blood samples were collected prior to the initiation of corticosteroid therapy to assess biomarker status, with results expected within 24 h after starting corticosteroids. Ruxolitinib therapy was administered within 24 h after corticosteroid treatment, based on biomarker risk. Ruxolitinib treatment was not initiated until biomarker results were available. Patients identified with low biomarker-based risk were excluded, and the treatment regimen was determined by the physician.

Ruxolitinib plus corticosteroids treatment

In the RUX/steroids combined group, patients initially received intravenous methylprednisolone (Pfizer, NY, USA) at a dosage of 1 mg/kg/day for at least 7 days. Ruxolitinib (Jakavi, Novartis, Nurnberg, Germany) was administered orally at a daily dose of 5 mg. If GVHD patients responded to treatment, achieving either PR or CR at 7 days, the dosage of methylprednisolone was gradually reduced. A recommended taper schedule was provided to discontinue methylprednisolone by 6 weeks (Supplemental Methods). Following the discontinuation of steroid therapy and no recurrence of GVHD, CsA was tapered over a 60-day period. Following the discontinuation of CsA, if there was no recurrence of GVHD, ruxolitinib was tapered over 90 days, with a total duration of approximately 6 months.

Corticosteroids treatment

In the steroids-only group, patients were initially administered methylprednisolone at a dose of 2 mg/kg/day, divided into two doses daily, for at least 7 days before starting dose reduction. The methylprednisolone dosage was gradually reduced after achieving CR and tapered over 10 weeks (Supplemental Methods). CsA was administered intravenously at 2 mg/kg twice daily, aiming for target trough levels of 150–250 ng/mL. In the steroids-only group, CsA treatment lasted for about 6 months.

Second-line therapy

In both groups, second-line therapy was initiated for patients with refractory aGVHD, which was defined as GVHD progression after 3 days of therapy, lack of improvement within 7 days, or failure to achieve CR after 14 days of treatment (see supplemental Methods and Supplementary Table 2 for more details).

Luminex assays for aGVHD biomarker measurement

An algorithm score based on ST2 and REG3α concentrations at the onset of GVHD was used to classify aGVHD patients into high-, medium-, and low-risk groups. Blood samples were collected for the measurement of aGVHD biomarkers at various time points, including before patients underwent the conditioning regimen, on days 7, 14, 28, 60, and 90 after transplantation, at the onset of aGVHD, and 3–7 days after enrollment (Supplemental Methods).

Immunophenotyping

Immune monitoring was conducted on peripheral blood samples collected on 28, 56, 180, 365, 560 and 720 days after enrollment. Antibodies against CD3, CD4, CD8, CD20 and CD56 (BD Biosciences, USA) were used to detect CD3+ T, CD4+ T, CD8+ T, B and NK cells via BD FACS Canto II (BD Biosciences, USA). Treg cells were identified by gating the population of CD3+CD4+CD25+Foxp3+ cells.

Sample size

The sample size calculation was based on the primary endpoint (ORR). In our previously phase I study involving 32 aGVHD patients treated with corticosteroids as first-line therapy, the expected ORR was 55%.19 Additionally, an ORR of 82.05% was observed in patients with aGVHD grades I–IV who received a combination of steroid and ruxolitinib as first-line therapy, including twelve patients with grade I aGVHD. Therefore, an expected ORR of 75% was established for those treated with the steroid-ruxolitinib combination. The study was designed with a 2-sided significance level α = 5% and a power of 1-β = 80%. 86 patients were required for each group as estimated using Z-Test statistics of PASS software (NCSS LCC, USA). Allowing a drop-out rate of 15%, a total of 198 patients were required (99 for each group, Fig. 1).29

Statistical analysis

All randomized participants will be analyzed for outcomes, with missing data expected to be under 10%. If needed, multiple imputations will be applied, and best-case/worst-case scenarios used for handling missing data.

Continuous data are presented as median with interquartile range (IQR) or mean and standard deviation (SD), depending on the normality of the distribution. Categorical data are described as n (%). The ORR, along with its 95% CI, was assessed using the Cochran–Mantel–Haenszel test, stratified by pre-transplant disease status and biomarker risk. The Kaplan–Meier method estimated DOR, OS, DFS, and FFS, with group differences analyzed by the log-rank test. Cumulative incidence rates of recurrent aGVHD, cGVHD, NRM, and relapse were estimated using a competing risk model and compared with the Fine and Gray test. Competing events were defined as follows: for recurrent aGVHD, relapse and death without recurrent aGVHD; for cGVHD, relapse and death without cGVHD; for relapse, death without relapse; and for NRM, relapse. For biomarker and immune reconstitution analyses, the Friedman test was used to identify statistically significant differences among time points for each biomarker. If the Friedman test was significant, Wilcoxon tests were conducted as follow-up tests, with Bonferroni correction applied to adjust the p-value for each biomarker analysis. A two-sided P < 0.05 was considered statistically significant. All analyses were performed using SPSS 22.0 software (IBM Corporation, Armonk, NY, USA) or R version 4.1.2 (www.cran.r-project.org), and all statistical analyses were based on the intend-to-treat set.