Ring Sideroblasts and an EPO Level That Already Answered the Question
Her endogenous erythropoietin is already high enough to predict that more epoetin won't help. The dossier had the answer to whether to switch to luspatercept before anyone asked the question out loud.
Twice a week for the past nine years, Margaret O. has shown up at the same branch library where she worked for three decades, this time to shelve donated books rather than check them out — a routine she describes as the thing that's kept retirement from feeling like an ending. She is 77, and has lived with lower-risk myelodysplastic syndrome with ring sideroblasts for two years, driven by an SF3B1 mutation identified at diagnosis. She has been transfusion-dependent for the past four months — roughly two units every three weeks — despite escalating doses of epoetin alfa started shortly after diagnosis.
Her baseline endogenous erythropoietin level, drawn before ESA therapy began, came back at 340 mU/mL: elevated, and specifically elevated past the threshold most consistently associated with poor ESA response in this disease. That number was sitting in her chart before the first epoetin dose was ever given, a quiet prediction of futility that the months of dose escalation since have gone on to confirm rather than overturn.
The transfusion burden itself has become the thing shaping her week — scheduling infusion center visits around her volunteer shifts, the fatigue that builds in the days before each transfusion is due, the two Thursdays this month she had to skip the library entirely. MEDALIST, the trial that established luspatercept's benefit in lower-risk MDS with ring sideroblasts, enrolled transfusion-dependent patients who were refractory to ESA therapy, intolerant of it, or — if they had never been given one — carrying an endogenous erythropoietin level above 200 units per liter marking them unlikely to respond. She enters through the first of those doors, not the third: she was given epoetin and it failed. Her 340 is not what qualifies her, it is the reason the failure was predictable from the beginning, and the two are worth keeping apart, because a therapy chosen for a population a patient merely resembles carries a weaker rationale than one chosen for the population she actually belongs to.
Reading her own baseline lab back into the decision
Switch her to luspatercept. Her disease is ring-sideroblast, SF3B1-mutated MDS — the exact population MEDALIST enrolled — and her baseline endogenous EPO of 340 was already elevated before her first epoetin dose. That's not a soft impression that she's 'not responding well'; it's a specific, validated predictor that further ESA escalation was unlikely to work from the start.
MEDALIST, Fenaux et al., 2020, randomized this exact population to luspatercept versus placebo and found a real reduction in transfusion burden. Her own numbers put her squarely inside that trial's population, not at its edge.
I'll grant the EPO argument — augmenting a regimen her own baseline lab predicted would fail is not a position worth defending, and I'm not going to defend it. But I don't think that settles what to do, and I want to be careful about how her trial match is being described.
MEDALIST's endogenous-EPO threshold of 200 units per liter was the criterion for patients who had never received an ESA — it was how the trial identified people unlikely to respond before anyone tried. She isn't that patient. She had eight months of escalating epoetin and failed it, so she enters that population through the refractory door, not the EPO-threshold door. Her 340 explains why she failed; it isn't what makes her eligible. That distinction matters because the trial's own refractory arm is where her evidence actually lives.
And luspatercept is not a benign default. It carries a real hypertension signal, and MEDALIST's own population was younger and less transfusion-burdened than she is — she is running about two units every three weeks, which is over five units per eight weeks against a trial floor of two. Starting it is right. Starting it without a blood-pressure plan and a defined reassessment point is not.
You're both treating this as a question about which drug she starts. The number that will actually decide how she does over the next year is the one neither of you has mentioned: her ferritin is 680 and climbing on a transfusion burden that has been running four months.
If luspatercept works, iron chelation is a problem that solves itself and deferasirox never needs to start. If it doesn't — and roughly six in ten MEDALIST patients did not reach transfusion independence — she accumulates iron for another three months while we wait to find out, on top of what she already has. So I'd set the reassessment at three cycles rather than leave it open, and I'd draw the ferritin at that same visit rather than at some later one. That way the drug decision and the chelation decision get made off the same data, instead of the second one arriving late because the first one went well enough that nobody looked.
Agreed: discontinue epoetin alfa and start luspatercept at standard starting dose, with blood pressure monitored at each visit given the drug's known hypertension signal. Transfusion needs and ferritin drawn together at three cycles, so the response question and the chelation question are answered off the same visit.
Not agreed, and left that way: whether deferasirox should be started at that three-cycle mark if her ferritin has kept climbing but luspatercept is showing partial benefit. The Transfusion Medicine Physician would start chelation on the ferritin trend alone; the Hematologist would rather not layer a second drug's toxicity onto a regimen still being assessed, and would wait a further cycle. Neither position was overruled, and the reassessment visit was scheduled without the question being settled in advance.