Impact of Cytopenias in Patients Treated with Ruxolitinib Versus Best Available Therapy for Steroid-Refractory Acute Graft-Versus-Host Disease: A REACH2 Post Hoc Analysis

Introduction: The oral Janus kinase (JAK)1/JAK2 inhibitor ruxolitinib (RUX) is approved for steroid‐refractory acute graft‐versus‐host disease (SR‐aGVHD). RUX was associated with higher response rates vs best available therapy (BAT) in the randomized, open‐label phase 3 REACH2 trial in patients (pts) aged ≥12 years with grade II–IV SR‐aGVHD. Outcomes of REACH2 pts who developed clinically relevant cytopenias need further validation. Objective: This REACH2 post hoc analysis assessed the effect of cytopenias on treatment response. Methods: Laboratory values at baseline, weekly through Week (Wk) 8, and every 4 wks through Wk 24 were assessed. Clinically relevant low blood counts for this SR‐aGVHD population were defined as platelet count (PLT) <30×109/L, hemoglobin (Hgb) <8 g/dL, white blood cell count (WBC) <5×109/L, and absolute neutrophil count (ANC) <1×109/L. Overall response rate (ORR [complete response + partial response]) at Day 28 and durable response at Day 56 among Day 28 responders were assessed in PLT, Hgb, WBC, ANC, and any cytopenia subgroups defined by the presence (low) or absence (not low) of ≥1 low PLT, Hgb, WBC, or ANC measure between baseline and Wk 4. All pts received RUX at a starting dose of 10 mg twice daily (bid) with dose modifications allowed per clinician discretion for safety/ efficacy per protocol. Odds ratios and 95% CIs were calculated with the stratified Cochran‐Mantel‐Haenszel test. Results: Of 309 randomized pts, 302 were included in the safety analysis (RUX, n=152; BAT, n=150). Baseline demographics and disease characteristics were balanced between treatment groups in the randomized population. Mean PLT and Hgb were similar between treatment groups through Wk 24, decreasing over the first 4–8 wks of treatment and recovering thereafter (Figure 1). WBC and ANC similarly decreased through Wk 8, remaining lower in the RUX vs BAT group at Wk 24. Median RUX dose intensity within the first 4 wks of treatment was at or near 20 mg/d in all subgroups both with or without cytopenias (range between subgroups, 18.5–20.0 mg/d). Day 28 ORR was higher with RUX vs BAT regardless of the presence or absence of low PLT, Hgb, WBC, or ANC levels (Figure 2). Among Day 28 responders, durable responses at Day 56 were observed across all subgroups. Median duration of response was numerically longer with RUX vs BAT for all PLT and Hgb subgroups as well as the low WBC and not low ANC subgroups (Figure 3).(Figure Presented)(Figure Presented)(Figure Presented) Conclusions: RUX and BAT were associated with similar reductions in blood counts early in treatment. This finding highlights the potential myelosuppressive effects of aGVHD development vs the impact of treatment on blood counts. Regardless of the presence or absence of cytopenias early in treatment for SR‐aGVHD, most pts were able to be treated at initially prescribed RUX dose of 10 mg bid. Furthermore, RUX was associated with higher Day 28 response rates than BAT, with durable responses observed at Day 56