A Guideline That Says “May Help” and a Cohort That Says It Didn't: Steroids in Acute Exacerbation of IPF
A man's known IPF has acutely worsened, and the same guideline recommending steroids as the only real option also recommends against the immunosuppression that made his underlying disease worse — the disagreement is about what a weak recommendation actually licenses.
W.H., a 74-year-old man, has spent the two years since his IPF diagnosis mostly at home with his wife of fifty-one years, who has quietly become the one keeping track of his oxygen tank changes and pill organizer, on top of a marriage that had never before required either of them to manage a chronic illness. He has been on nintedanib since diagnosis, tolerating it reasonably well aside from occasional loose stools — an antifibrotic, not an immunosuppressant, which is a distinction nobody has had reason to care about until this week. Two years in, he has never taken a drug from the class that turns out to matter most in what follows. Over the past week his breathing has worsened sharply — from needing 2 liters of oxygen with exertion to requiring 6 liters at rest — and a new CT shows bilateral ground-glass opacity superimposed on his known basal honeycombing, a pattern distinct from his baseline scan obtained three months ago. A respiratory viral panel is negative, procalcitonin is low, CT angiography shows no pulmonary embolism, and his BNP is unremarkable — no infection, no clot, no heart failure to explain this. By the criteria he meets, this is an acute exacerbation of his underlying disease, not a new one.
The 2022 ATS/ERS/JRS/ALAT guideline gives a weak recommendation, based on very low-quality evidence, that corticosteroids may be used in acute exacerbation of IPF — the same document that, for stable IPF, recommends against the immunosuppression PANTHER-IPF showed actively increased mortality and hospitalization. Papiris and colleagues published the most-cited counterweight, and it does not say what it is usually quoted as saying. Every patient in that series was managed the same way — immunosuppression stopped, best supportive care, empiric antimicrobials, no steroids for the exacerbation. Half survived. The split that produced the striking numbers was by what patients had been on beforehand: three of twelve with a prior immunosuppression history survived, against nine of twelve with none. So it is a finding about arriving at an exacerbation already immunosuppressed, not a head-to-head of steroids against no steroids. That distinction places W.H. squarely in the more favorable group — nintedanib is an antifibrotic, not an immunosuppressant, and he has taken nothing else. A 2026 systematic review of thirty-two studies is the closer fit to the question actually being asked, and its signal is about dose rather than yes-or-no: ninety-day mortality ran to 54% among patients given methylprednisolone at 1000mg daily against 39% at 500 to 1000mg. Neither body of evidence settles anything; between them they establish that the argument here is over how much steroid and for how long, not whether the drug is categorically the safer default.
A weak recommendation, cutting both ways
The 2022 guideline gives a weak recommendation for corticosteroids in acute exacerbation of IPF, and with no other therapy shown to do anything, I want to give pulse methylprednisolone. A weak, low-quality recommendation is still the only recommendation we have — and I'd point out that Papiris's own protocol reserved its best results for patients who arrived without prior immunosuppression, which is W.H. exactly.
You're right that the guideline does endorse trying it. But PANTHER-IPF showed real, measured harm from this same drug class in this same disease's stable phase, and the 2026 systematic review of thirty-two studies found ninety-day mortality of 54% at 1000mg/day methylprednisolone against 39% at 500 to 1000mg — if we give this, the dose is not a detail.
A weak recommendation based on very low-quality evidence isn't a reason to give a drug — it's an honest admission nobody actually knows, which cuts against giving it just as much as for it.
Neither of those numbers is strong enough to dictate an extreme, so I'd propose a bounded trial at the lower end of the pulse range rather than the top of it: a defined three-day course of methylprednisolone at 500 to 1000mg daily, with an explicit reassessment point, not an open-ended course at 1000. W.H. himself has told his family he wants a real attempt made, not prolonged escalation regardless of what happens — that should shape the boundary we set, not just the evidence.
Agreed: a defined three-day course of methylprednisolone at 500 to 1000mg daily, with an explicit family meeting and reassessment on day four regardless of trajectory, rather than an open-ended course.
Not agreed: what “not working” should trigger at that reassessment — the pulmonologist would stop all immunosuppression at that point regardless of small improvements; the critical care physician would want to see at least a week's trend before calling it a failure.