VEXAS with a Clonal Marrow: When the Bone Disease and the Blood Disease Want Different Drugs
One clone is behind both the anemia and the inflammation, but the two treatments best suited to each half of that picture aren’t the same drug — and neither works fast enough for today’s pain.
Three seemingly unconnected problems sent Walter H., a 68-year-old retired postal inspector, through a year of specialist visits before anyone recognized them as one disease rather than three: recurrent fevers with painful, red, swollen ears that a first rheumatologist diagnosed as relapsing polychondritis; a macrocytic anemia that has required transfusion every three to four weeks despite a normal iron and B12 workup; and crops of tender, red skin nodules his dermatologist biopsied and called a neutrophilic dermatosis without a clear unifying cause. A bone marrow biopsy, ordered once the pattern of fever, cartilage inflammation, and transfusion-dependent anemia was recognized as a cluster rather than three separate diagnoses, showed characteristic cytoplasmic vacuoles in his myeloid and erythroid precursors, and genetic testing confirmed a somatic UBA1 mutation — the molecular finding that unified all three problems under a single diagnosis of VEXAS syndrome, first described only in 2020 and still being actively characterized.
The same marrow biopsy also identified a low-grade myelodysplastic clone — not an incidental second diagnosis, but very likely the same somatically mutated hematopoietic population responsible for both his inflammatory symptoms and his anemia, since UBA1 mutations in VEXAS arise in myeloid progenitor cells and frequently co-occur with or evolve toward a myelodysplastic phenotype. His most recent transfusion was eleven days ago, and today he is describing his right ear as “the worst it's been,” tender enough that he winces when his hearing aid touches it — an acute flare of the chondritis that is affecting his daily function right now, independent of and on a faster timescale than the underlying clonal disease driving it. His transfusion interval, roughly six weeks apart a year ago, has shortened to three to four weeks — a trend his hematologist has flagged as evidence the marrow's dysfunction is not stable.
Two real, evidence-supported treatment approaches exist for VEXAS, and each targets a different layer of a disease that is, underneath both his marrow findings and his ears, ultimately one clone. Ruxolitinib showed the strongest responses in Heiblig and colleagues' 2022 retrospective multicenter comparison of JAK inhibitors — 83 percent responding at three months against 18 percent on the other agents in the class — primarily addressing symptoms like today's chondritis flare, without directly targeting the clone. Azacitidine paired with tocilizumab, in Hadjadj and colleagues' 2024 FRENVEX registry analysis, improved transfusion dependence and marrow findings, but acted more slowly on ear pain and fevers. Neither has been directly compared to the other in a trial, and both carry real risks — clotting with ruxolitinib, cytopenia with azacitidine — which is why the team cannot simply pick whichever drug sounds most comprehensive without addressing the pain he is in today.
One clone, two treatments, and today’s pain
The largest retrospective comparison of JAK inhibitors in VEXAS, Heiblig and colleagues, found ruxolitinib produced meaningfully better response rates than other agents in the class — including for the inflammatory and chondritis features that are his most disabling problem right now. I'd start there.
I'd frame this differently. His myelodysplastic clone isn't a second diagnosis sitting next to VEXAS — it's very likely the same UBA1-mutant population driving both his inflammation and his transfusion-dependent anemia. Azacitidine targets that clone directly, and a recent multicenter cohort, Hadjadj and colleagues, found azacitidine combined with tocilizumab outperformed JAK inhibitors and anakinra on outcomes that specifically included transfusion burden.
Ruxolitinib may quiet the cytokine output of that clone. It doesn't touch the clone itself. If we're choosing a drug meant to address the disease at its actual source rather than its downstream signal, azacitidine is the more mechanistically complete answer.
You're both arguing about the right long-term disease-modifying agent, and that's a real disagreement worth having. But neither ruxolitinib nor azacitidine is going to touch the ear pain he's in today — azacitidine especially can take multiple cycles to show a hematologic effect, and even ruxolitinib's inflammatory benefit isn't instantaneous.
I'd start a short corticosteroid bridge now, specifically for the acute flare, regardless of which of you turns out to be right about the longer-term choice. That decision doesn't need to be made under the pressure of his current pain, and it shouldn't be.
A short corticosteroid bridge was started immediately for the acute chondritis flare, with azacitidine and tocilizumab initiated as his primary disease-modifying regimen once the acute symptom was addressed — sequencing both voices' concerns rather than resolving them as a single simultaneous decision.
Not agreed: whether ruxolitinib should have been the first disease-modifying agent instead. The Rheumatologist maintains that his most disabling symptom argued for the drug with the strongest specific evidence for inflammatory and chondritis control, and would revisit ruxolitinib if azacitidine's hematologic benefit, which can take several cycles to appear, does not also bring adequate relief of his ear symptoms in the meantime. The Hematologist holds that treating the clone directly was the more fundamental choice regardless of timeline. Both agreed to reassess jointly once azacitidine has had a fair, multi-cycle trial, rather than judge it prematurely against a JAK inhibitor's faster inflammatory response.