Two Myelofibrosis Patients, Two Cytopenias, Two Different JAK Inhibitors
Two patients with primary myelofibrosis, needing JAK-inhibitor therapy for the same underlying disease. The disagreement in each case is genuinely independent — not a shared debate split across two patients, but two cytopenia profiles pointing toward two different drugs.
R.F., a 61-year-old retired librarian, was diagnosed with primary myelofibrosis eighteen months ago after a routine CBC turned up a hemoglobin of 8.9 g/dL and a teardrop-cell-studded smear. She has needed two units of packed red cells roughly every six weeks since diagnosis — transfusion-dependent by any working definition — and her fatigue arrives on a schedule she can now read off a calendar, which is why she gave up the volunteer shelving shifts she used to do three mornings a week. Her spleen is palpable 9cm below the costal margin and her JAK2 V617F allele burden is high. Her platelet count is 190,000, and it matters here mostly for what it rules out: nothing in her counts constrains the dose of any JAK inhibitor on the table, so this decision gets made on the drugs' merits rather than on what her marrow will tolerate.
Anemia, then, is the whole constraint — and anemia happens to be the axis along which these two drugs genuinely differ rather than merely overlap. Ruxolitinib is the most extensively validated of them for spleen and symptom control, and its dose-limiting toxicity is myelosuppression: a drug whose principal cost is precisely the problem she already has too much of, proposed for a woman running two units in arrears every six weeks. Momelotinib adds ACVR1 inhibition to the same JAK1/2 blockade — a mechanism aimed at the anemia rather than merely tolerated alongside it — and MOMENTUM was assembled around patients who look like her: symptomatic, anemic, transfusion-dependent, not a general myelofibrosis population in which anemia happened to be present. The conventional sequence, start the established drug and switch when the predictable problem shows up, quietly assumes the trial period costs nothing. For her the currency of that trial period is units of blood.
Ruxolitinib is still the most extensively validated JAK inhibitor for spleen and symptom control in myelofibrosis, and she hasn't failed it — there's no trial failure yet to justify reaching past the first-line standard.
I'd start with momelotinib rather than wait for a switch. MOMENTUM specifically enrolled anemic, transfusion-dependent myelofibrosis patients and found momelotinib's ACVR1 inhibition reduced transfusion burden in a population that looks like her — not a general myelofibrosis cohort where anemia was incidental. Starting ruxolitinib first means starting a drug whose own dose-limiting toxicity is the exact problem she already has too much of, then switching once that's confirmed to be a problem rather than avoiding it from the start.
'Hasn't failed it yet' assumes the trial-and-failure sequence is free — for her, a trial of a myelosuppressive drug isn't a neutral first step, it's a real chance at needing more transfusions before anyone agrees to switch.
Agreed: start momelotinib as her first JAK inhibitor rather than ruxolitinib, monitoring transfusion frequency, spleen size, and symptom scores over the next three months.
G.P., a 58-year-old long-haul truck driver, was diagnosed with primary myelofibrosis four months ago during a workup for easy bruising he'd initially blamed on tight seatbelt straps. His platelet count at diagnosis was 41,000 and has drifted between 38,000 and 41,000 since — not a value trending toward a threshold but one already sitting below it. He has had two nosebleeds that took more than routine pressure to stop, no major bleeding event, and a spleen palpable 7cm below the costal margin that accounts for most of what actually bothers him day to day. His fatigue is real but secondary; the abdominal discomfort and the bruising are what he mentions first. His JAK2 V617F allele burden is moderate, which is to say the disease is not driving the platelet count down through sheer proliferative pressure — the marrow fibrosis is.
Severe thrombocytopenia changes the JAK-inhibitor calculus in the opposite direction from anemia, and it does so more sharply than a dose adjustment implies. Ruxolitinib's labeled starting-dose table runs from platelets above 200,000 down to a floor of 50,000 to under 100,000, where the starting dose is 5mg twice daily; below 50,000 the table simply stops, the manufacturer advises against use, and the dose-modification guidance is to interrupt rather than reduce. His count has sat between 38,000 and 41,000 since diagnosis, which places him underneath the label rather than at its lower edge — the question isn't what dose of ruxolitinib he should get, it's that the label declines to name one. Pacritinib was studied in exactly the population the ruxolitinib label walks away from: PERSIST-2 enrolled myelofibrosis patients with platelet counts under 100,000, including a severely thrombocytopenic subgroup at his level, at a full dose that carries no platelet-driven reduction at all.
Ruxolitinib still has the larger efficacy dataset overall, and my instinct is to attempt it at a reduced dose before reaching for a less-established agent — though I'll grant in advance that at 41,000 I'm not sure what dose I'd actually be writing.
You wouldn't be writing one, and that's the point. The labeled starting-dose table bottoms out at 50,000 to under 100,000 — 5mg twice daily — and below 50,000 it doesn't continue; the manufacturer advises against use and the hematologic-toxicity guidance is interruption, not reduction. At 41,000 a 'reduced, label-appropriate dose' isn't conservative prescribing, it's prescribing off the bottom of the table and calling it standard of care because the drug's name is familiar. PERSIST-2 enrolled patients with platelet counts under 100,000, including a severely thrombocytopenic subgroup at his level, and found pacritinib delivered spleen and symptom benefit at a full dose requiring no platelet-driven reduction.
Attempting the standard drug at a dose the trial data that established its benefit never actually tested isn't the same as giving him the standard of care — it's giving him a different, unstudied regimen under a familiar name.
Agreed: start pacritinib rather than dose-reduced ruxolitinib, with platelet count and bleeding symptoms monitored closely over the first three months.
Not agreed, carried forward from Patient A's own consultation: whether momelotinib, with its own more moderate platelet effect than ruxolitinib's, might also have been reasonable for G.P. — not raised as a live option in his consultation, since pacritinib's direct population match in PERSIST-2 settled the question before the comparison came up.