Progressive Fibrosis, No Settled Diagnosis: What Does “Despite Management” Require?
A patient with unnamed autoimmune-feature lung disease has real, physiologic decline but no failed treatment behind her — the disagreement is over what a diagnostic criterion written around a treatment trial actually requires when no treatment was ever tried.
L.M., a 58-year-old woman, has spent eleven years as a hospice chaplain, sitting at bedsides during the last weeks of other people's lives — a fact impossible for her pulmonologist to ignore once the conversation turned to her own prognosis. She was referred nine months ago for a persistent cough and mild exertional dyspnea; a high-resolution CT showed a UIP-predominant pattern with faint ground-glass overlay suggesting an NSIP component, and her only positive serology was a titer of ANA 1:320 — not enough, alone or with anything else found, to meet criteria for any named connective tissue disease. The multidisciplinary panel settled on interstitial pneumonia with autoimmune features, a label that describes what she doesn't have as much as what she does. She was watched rather than treated, on the reasoning that a single antibody titer wasn't grounds to start immunosuppression for a disease that hadn't been named.
Nine months later, the watching produced an answer nobody wanted: her FVC has fallen from 72% to 65% predicted, and her DLCO has dropped alongside it — real, physiologic progression, not statistical noise. The 2022 ATS/ERS/JRS/ALAT criteria for progressive pulmonary fibrosis ask for exactly this kind of decline occurring despite appropriate management of the underlying disease, and nintedanib's approval for that entity rests on the INBUILD trial's own progressive-fibrosing population. The problem is that L.M. was never actually given any management to fail — no immunosuppression was ever tried, because no defined disease existed to direct it at. Her UIP-predominant pattern is also the reason PANTHER-IPF still matters here: in idiopathic pulmonary fibrosis, that same triple-therapy immunosuppression regimen was stopped early after it raised mortality and hospitalization rather than lowering them, precisely because UIP tends not to behave like an inflammatory disease at all. Whether an unnamed, IPAF-labeled process with a UIP-dominant scan follows IPF's rules or a treatable connective-tissue disease's rules is the actual question nobody in the room can yet answer with real data.
Nine months of watching, and an answer
Start nintedanib today, not mycophenolate. Her scan is UIP-predominant — the same pattern that, in idiopathic pulmonary fibrosis, taught us the hard way not to reach for immunosuppression. PANTHER-IPF stopped its prednisone-azathioprine-N-acetylcysteine arm early because it raised mortality and hospitalization in patients with this exact pattern, not because the trial ran out of time. She's already lost seven points of predicted FVC watching; I'm not willing to spend more lung function testing whether that lesson doesn't apply to her too.
You're right that PANTHER-IPF is the strongest reason to be careful here, and I'm not dismissing it. But her scan isn't pure UIP — there's a real NSIP overlay on the same CT report you're citing, and interstitial pneumonia with autoimmune features exists as a category precisely because some of these patients have a genuinely treatable, antibody-associated process underneath. Reaching straight for nintedanib because the dominant pattern resembles IPF discards the half of her scan that doesn't.
“Same pattern” is doing more work in that argument than the report itself supports — it’s a predominant pattern with a documented second component, not an unambiguous one.
I'd rather settle this on the guideline's own terms than on which pattern reads as more familiar. The 2022 criteria for progressive pulmonary fibrosis require decline despite appropriate management of the underlying disease — they don't require a specific treatment trial to have failed first. Nine months of real monitoring, with no clear indication a different management would have changed her trajectory, is appropriate management for a disease this ambiguous. That gets us to nintedanib without either pretending we skipped a step she was owed or pretending the NSIP component doesn't exist — it just isn't the deciding fact here.
Agreed: start nintedanib, with pulmonary function testing repeated at three months rather than the usual six, given how fast she's already declining.
Not agreed: whether a limited trial of mycophenolate should still run in parallel for the NSIP-overlay component specifically — the rheumatologist would add it now; the pulmonologist wants one antifibrotic-only interval first, worried that layering both drugs at once would make a future decline uninterpretable.