Steroid-Sparing in Hypereosinophilic Syndrome: A Trial That Almost Fits Her
A single patient who needs off steroids and doesn't quite meet the entry criteria of the trial that would otherwise settle her next drug. The disagreement is how much an almost is worth.
Renata M. taught middle school science for thirty-one years, retired two years ago, and has spent the time since trying to get her hypereosinophilic syndrome onto a dose of prednisone she can actually live with. She hasn't found one. Every attempt to taper below 25 milligrams daily has brought her eosinophil count back up within weeks, and the steroid itself has cost her more than the disease has this year — a new diagnosis of type 2 diabetes and, six months ago, a vertebral compression fracture from bone density that's now solidly in osteoporotic range. Her HES has tested negative for the FIP1L1-PDGFRA fusion, which takes imatinib off the table entirely; what she needs is a steroid-sparing agent, and the modern, best-evidenced answer for FIP1L1-PDGFRA-negative disease is mepolizumab, the anti-IL-5 antibody that Roufosse and colleagues' 2020 phase III trial found meaningfully reduced flares against placebo.
The complication is that Renata doesn't quite fit inside that trial. Roufosse's study enrolled patients who'd had two or more flares in the preceding twelve months with a screening eosinophil count of at least 1000 per microliter; Renata has had one clear flare this year, and her count on her current dose sits at 780. Neither number disqualifies her from the drug working, and the trial's own data speak to one of them: Rothenberg and colleagues' post-hoc analysis of that same study found mepolizumab reduced flares irrespective of baseline eosinophil count. Its lowest subgroup still sat above 1000, so it argues her 780 is unlikely to matter without reaching down to it, and her flare count has no comparable analysis behind it at all. Prescribing mepolizumab for her remains an extrapolation past the studied population — but one the trial's own internal structure partly supports, which is different from a guess. The alternative, hydroxyurea, has decades of real-world use across the entire HES spectrum with no flare-count threshold attached to it at all, precisely because it predates the era of enrollment-criteria trials this specific. Renata's own fracture history sharpens what's actually at stake in choosing wrong. Hydroxyurea's main real-world liability is myelosuppression, and a woman who's already lost bone density fast enough to fracture a vertebra at 58 is not someone with much reserve to spare if a second toxicity gets layered on top of the first. She's said plainly, in her own words, that she'd rather try "almost the right drug" than go back to a medication she already associates with the two complications that have defined her year — a preference the team is taking seriously rather than treating as incidental to the pharmacology.
A trial that almost describes her, and what to do with the gap
I'd start mepolizumab. Her disease is FIP1L1-PDGFRA-negative, IL-5-dependent eosinophilia, exactly what the drug targets — and Rothenberg's post-hoc of Roufosse's own trial found the flare reduction held irrespective of baseline eosinophil count. That's not me waving at a mechanism; it's the trial's own data saying the 1000 threshold wasn't doing the work everyone assumes it was.
I'd rather use hydroxyurea. It's been the steroid-sparing option across the whole non-fusion HES spectrum for decades, with no flare-count threshold attached to it at all — that's not because it's less rigorously studied, it's because it predates trials built with this specific kind of enrollment gate.
I don’t disagree that her biology fits mepolizumab’s mechanism. I’m pointing out that "fits the mechanism" and "matches the population the drug was actually tested in" aren’t the same claim, and she’s genuinely on the wrong side of the second one.
Both of you are right about different things. She is an extrapolation past mepolizumab's trial population — that should be said plainly, not treated as a technicality. But hydroxyurea carries its own real cost for her specifically: myelosuppression risk in a woman who's already had a vertebral fracture and new diabetes from steroid toxicity is not a small consideration.
Given her actual comorbidities, I'd try mepolizumab first, watched closely, with hydroxyurea held in reserve if it fails — not because the trial-fit question resolves in its favor, but because her specific risk profile makes hydroxyurea's known toxicity the harder one to accept right now.
Agreed: start mepolizumab with a defined steroid taper attempt over the following three months, and monitor eosinophil counts and flare frequency closely given her extrapolated trial fit.
Not agreed, and named as a real open question rather than settled by starting the drug:
The hematologist would move to hydroxyurea next, arguing her steroid-toxicity profile has already been weighed once and shouldn't override a real efficacy failure a second time.
The allergist would want a second anti-IL-5 attempt at a shorter interval before conceding the class doesn't work for her, given how far outside the trial's own population she started.
The extrapolation is being tried, not assumed to work. What happens if it doesn't is left open, deliberately, rather than pre-decided today.