Steroid-Refractory Acute GVHD: Ruxolitinib Against a Marrow That Just Cleared Engraftment
Ruxolitinib has the best randomized evidence for steroid-refractory acute GVHD, but its own trial's myelosuppression signal lands differently in a patient whose marrow only just cleared engraftment.
Diane W., a 61-year-old church choir director, underwent a matched unrelated donor transplant seven weeks ago for myelodysplastic syndrome that had begun progressing toward acute leukemia. Engraftment was slower than the team would have liked — her absolute neutrophil count crossed 500 only twelve days ago, later than typical, reflecting the same marrow that made her MDS diagnosis in the first place rather than any new problem. Two weeks ago she developed a diffuse skin rash and, more concerning, five liters of watery diarrhea a day with abdominal cramping — grade III acute GVHD involving skin and gut. Methylprednisolone at 2 mg/kg produced no improvement by day five and her stool output has if anything worsened: steroid-refractory by the standard definition.
Ruxolitinib has the strongest evidence base for exactly this situation. The REACH2 trial randomized patients with steroid-refractory acute GVHD to ruxolitinib or best available therapy and found a day-28 overall response rate of 62% versus 39% — the largest, most direct randomized signal available for any second-line agent in this setting, and the basis for its now-standard use. But that same trial's own safety data recorded grade 3-4 thrombocytopenia in roughly a third of the ruxolitinib arm, meaningfully more than with best available therapy — a real, mechanism-consistent effect of a drug that suppresses the same JAK-STAT signaling marrow progenitor cells depend on to divide. Diane's marrow only just crossed engraftment threshold, on a background of pre-transplant MDS that already made it a slow, fragile producer. A drug whose own pivotal trial shows it can push counts down further is not a neutral choice for a patient whose reserve to absorb that push is this thin. Her platelets of 58,000, though, sit above every threshold in the label's acute-GVHD dose-modification table, so nothing in that table is triggered on the day she starts — the safeguard that applies to her is not a smaller first dose but the label's requirement that her counts hold steady for three days before the dose is allowed to rise.
Her original GVHD prophylaxis was standard tacrolimus and short-course methotrexate, discontinued on schedule at day 35 with no acute GVHD until this presentation — a genuinely late onset for grade III disease, which the team noted as consistent with her slow engraftment more broadly rather than a separate, unexplained event. Stool studies sent at symptom onset were negative for infectious causes, including a specific cytomegalovirus PCR given her donor-negative, recipient-positive serostatus, ruling out the other major cause of severe diarrhea this early after transplant and leaving acute GVHD as the only explanation her presentation actually supports.
Choosing second-line therapy on day 5 of steroid failure
REACH2 is the most direct evidence we have for exactly this situation: grade III, steroid-refractory acute GVHD, randomized against best available therapy. A day-28 response rate of 62% versus 39% is a large, clinically meaningful gap, and grade III GI involvement at five liters a day carries real mortality risk the longer it goes uncontrolled. I don't think we should let a theoretical marrow concern delay starting the single most effective agent we have.
It isn't theoretical — it's in REACH2's own safety table. Grade 3-4 thrombocytopenia occurred in roughly a third of the ruxolitinib arm, clearly more than with best available therapy, which is exactly what you'd expect from a drug that inhibits the same JAK-STAT signaling marrow progenitors need to proliferate. Diane's platelets are already 58,000 and her neutrophils only crossed engraftment threshold twelve days ago, on top of an MDS marrow that was a poor producer even before conditioning.
Extracorporeal photopheresis has no myelosuppressive mechanism at all, and while its response data in steroid-refractory acute GVHD are less robust than REACH2's randomized comparison, it doesn't ask a marrow this fragile to absorb an additional hit while it's still trying to establish itself.
I think the choice doesn't have to be between REACH2's evidence and Diane's marrow reserve — but I want to be precise about what the label actually gives us, because it isn't a lower starting dose. The platelet-count-based starting dose people remember belongs to the myelofibrosis indication. For steroid-refractory acute GVHD the Jakafi label sets a flat starting dose of 5 mg twice daily, which is already the lowest dose level there is.
What the label does build in is the escalation gate, and that is the part that answers your concern directly: the dose only goes up to 10 mg twice daily after at least three days, and only if her ANC and platelet count have not fallen by 50% or more from day one. Diane's marrow gets to vote before she is exposed to any more drug than the minimum. If it fails that test we simply hold at 5 mg twice daily, and the label's own dose-modification table takes her down to 5 mg once daily if clinically significant thrombocytopenia develops on treatment. That lets us use the agent with the strongest response data without pretending her marrow risk doesn't exist — and without inventing a dose reduction the label doesn't actually describe.
Agreed: ruxolitinib at the labeled acute-GVHD starting dose of 5 mg twice daily, with the escalation to 10 mg twice daily held back unless her counts clear the label's own three-day criterion. CBC every other day for the first two weeks, and a pre-specified threshold (platelets below 20,000 or ANC falling below 500 again) that triggers an immediate switch to extracorporeal photopheresis rather than a wait-and-see approach.
By day 14, her GI output had fallen to under a liter a day and her skin rash was resolving; platelets held at 41,000 without needing transfusion. Her counts had cleared the three-day escalation criterion, so she had gone up to 10 mg twice daily on day four — which the team noted was the point: the label had let her marrow decide, rather than the team guessing at a dose in advance.