Clinical Question
Among patients with idiopathic pulmonary fibrosis (IPF), does inhaled treprostinil slow the decline in lung function compared to placebo over 52 weeks?
TL;DR
Inhaled treprostinil significantly slowed FVC decline in patients with IPF — a median difference of about 95 mL compared to placebo. It also reduced clinical worsening events by roughly 30%. The drug didn't show a difference in acute exacerbations or mortality, though the trial wasn't powered to detect those effects. The main caveat: nearly half of patients on treprostinil developed cough, and one in three discontinued the drug before the trial ended. A promising signal for a disease with few options — but the tolerability question is real.
The Deets
IPF is a relentlessly progressive fibrotic lung disease. Median survival without treatment is 3 to 5 years. We have two approved antifibrotic agents — nintedanib and pirfenidone — that slow progression but don't stop it. The search for anything that adds meaningful benefit on top of those drugs has been largely disappointing.
Treprostinil is a prostacyclin analogue we know well from pulmonary arterial hypertension. What made investigators think it might work in IPF isn't just the pulmonary hypertension angle — it's that treprostinil targets multiple prostacyclin receptors that modulate fibrogenetic pathways in the lung. A post hoc analysis from the INCREASE trial (inhaled treprostinil in PH associated with ILD) showed a signal toward FVC preservation, which set the stage for the TETON program.
TETON-2 is one of three phase 3 trials in that program. This trial enrolled patients from the Asia-Pacific region, Europe, and South America. It randomized 593 patients with IPF — confirmed by ATS/ERS guideline criteria and central HRCT review — to inhaled treprostinil or placebo at 12 breaths (72 μg) four times daily over 52 weeks. Notably, 75% of patients were already on background antifibrotic therapy with nintedanib or pirfenidone. This wasn't a treatment-naive population.
The primary outcome was change in absolute FVC at week 52. The result was clear: median FVC decline was −49.9 mL in the treprostinil group versus −136.4 mL in placebo — a between-group difference of 95.6 mL (P<0.001). That's a meaningful preservation of lung function in a disease where every milliliter counts.
Clinical worsening — a composite of death from any cause, respiratory hospitalization, or a ≥10% relative decline in predicted FVC — occurred in 27.2% of the treprostinil group versus 39.0% of controls (HR 0.71, P=0.02). That's a 29% reduction in the hazard of a clinically meaningful event.
The trial hit a wall at the next prespecified endpoint in the hierarchical testing sequence. There was no significant difference in time to acute IPF exacerbation (3.0% vs 5.8%, HR 0.54, but CI crossing 1). Because the hierarchy stopped there, no formal statistical inference was drawn on mortality or patient-reported outcomes — though numerically, those trended in the right direction.
The safety data deserve attention. Cough occurred in 48.3% of the treprostinil group versus 24.1% of controls. That's a 24-point difference, and it wasn't trivial — cough was the most common reason patients stopped the drug (4.0% of the treprostinil arm). Overall, 33.6% of treprostinil-treated patients discontinued prematurely compared to 24.7% of controls. For a chronic therapy requiring four inhalation sessions per day, tolerability is a central practical question.
The subgroup data are worth noting. The benefit appeared consistent across patients on nintedanib, pirfenidone, and no background therapy, though the CIs for the no-background-therapy group were wide and crossed zero (Hodges-Lehmann estimate 43.9 mL, 95% CI −45.0 to 132.8). The clearest benefits were in patients already on antifibrotics — raising the possibility that treprostinil adds something on top of standard care rather than serving as a standalone alternative.
Design
Phase 3, double-blind, randomized, placebo-controlled trial
N=593
Inhaled treprostinil (n=298)
Placebo (n=295)
Follow-up: 52 weeks
107 centers across 16 countries (Asia-Pacific, Europe, South America)
Randomization stratified by background antifibrotic therapy (nintedanib vs. pirfenidone vs. none)
Population
Inclusion Criteria
Age ≥40 years
Diagnosis of IPF per 2018 ATS/ERS/JRS/ALAT guideline, confirmed by central HRCT review
FVC ≥45% of predicted
If on nintedanib or pirfenidone, stable dose for ≥30 days prior to baseline
Exclusion Criteria
FEV₁/FVC <0.70 (to exclude obstructive physiology)
Supplemental oxygen requirement >10 L/min at rest
Pregnancy or lactation
Baseline Characteristics
From the full cohort
Demographics: Mean age 71.7 years (range 41–91), 80.1% male
Lung function: Mean FVC 76.8% predicted, mean DLCO 48.5% predicted
Time since IPF diagnosis: mean 3.9 years
Supplemental oxygen use: 13.0%
Background therapy: 36.1% nintedanib, 39.3% pirfenidone, 24.6% none
Smoking status: 69.5% former smoker, 28.3% never smoker
Interventions and Controls
Patients were randomly assigned in a 1:1 ratio to one of two groups:
Inhaled treprostinil — 4.2% inhalation solution administered via TD-300 ultrasonic nebulizer. Started at 3 breaths (18 μg) four times daily during waking hours, titrated to a target of 12 breaths (72 μg) four times daily or maximum tolerated dose. Median target dose of 12 breaths reached by week 4 in both groups; 87.1% of treprostinil patients at target dose by week 16.
Placebo — Identical nebulizer device and session schedule, matched for appearance and administration.
Outcomes
Comparisons are treprostinil vs. placebo.
Primary Outcome
Change in absolute FVC from baseline to week 52
−49.9 mL (95% CI, −79.2 to −19.5) vs. −136.4 mL (95% CI, −172.5 to −104.0)
Hodges-Lehmann between-group difference: 95.6 mL (95% CI, 52.2 to 139.0; P<0.001)
Secondary Outcomes
Outcome | Treprostinil | Placebo |
|---|---|---|
Any clinical-worsening event | 81/298 (27.2%) | 115/295 (39.0%) |
Death from any cause (component) | 4 (1.3%) | 3 (1.0%) |
Respiratory hospitalization (component) | 30 (10.1%) | 38 (12.9%) |
≥10% relative FVC decline (component) | 47 (15.8%) | 74 (25.1%) |
Acute exacerbation of IPF | 9 (3.0%) | 17 (5.8%) |
Death by week 52 | 19 (6.4%) | 24 (8.1%) |
Change in % predicted FVC (percentage points) | −1.4 (95% CI, −2.2 to −0.5) | −3.8 (95% CI, −4.8 to −3.1) |
Change in K-BILD total score | −1.0 (95% CI, −1.9 to 0.0) | −2.6 (95% CI, −3.7 to −0.8) |
Change in % predicted DLCO (percentage points) | −2.8 (95% CI, −4.1 to −2.0) | −4.1 (95% CI, −5.5 to −3.2) |
HR for clinical worsening: 0.71 (95% CI, 0.53–0.95; P=0.02)
Note: No formal statistical inference on endpoints after acute exacerbation due to hierarchical testing stopping rule.
Subgroup Analysis
Benefit in FVC preservation appeared consistent across subgroups based on age, sex, geographic region, ethnic group, smoking status, and supplemental oxygen use. By background therapy:
No antifibrotic therapy: between-group difference 43.9 mL (95% CI, −45.0 to 132.8) — wide CI, not conclusive
Nintedanib: 97.6 mL (95% CI, 26.1 to 169.2)
Pirfenidone: 127.9 mL (95% CI, 58.8 to 196.9)
Adverse Events
Cough: 48.3% (treprostinil) vs. 24.1% (placebo) — most common AE; severe in only 0.7% of treprostinil patients
Headache: 19.8% vs. 11.9%
Diarrhea: 13.8% vs. 16.9%
Throat irritation: 7.0% vs. 2.4%
Serious adverse events: 25.2% vs. 23.1%
Premature discontinuation of drug: 33.6% vs. 24.7%
Discontinuation due to adverse events: 16.8% vs. 12.5%
Most common causes of death: respiratory failure and worsening IPF.
Criticisms
The trial was sponsored and run by United Therapeutics, the manufacturer of inhaled treprostinil. All data collection, management, and analysis were performed by the sponsor. That doesn't invalidate the results, but it's worth naming.
The high discontinuation rate in the treprostinil arm (33.6% vs. 24.7%) creates real concern about selection bias over time. Patients who couldn't tolerate the drug dropped out, and those who remained may represent a more resilient subgroup. The composite strategy used to handle missing data — assigning the 2.5th percentile value to patients who died — is a reasonable but imperfect approach.
Nearly half of patients on treprostinil developed cough. This is a chronic inhaled therapy requiring four sessions per day. In routine clinical practice, adherence will be a significant challenge. The clinical trial environment almost certainly overestimates real-world tolerability.
The hierarchical testing sequence stopped at acute exacerbations, which means the numerical mortality benefit (6.4% vs. 8.1%), quality-of-life trend, and DLCO difference cannot be interpreted as statistically significant. These may be real effects, but the trial wasn't designed — or didn't have the statistical power — to confirm them.
The no-background-therapy subgroup showed a wide CI that crossed zero (43.9 mL, 95% CI −45.0 to 132.8), raising the question of whether treprostinil's benefit is primarily additive to existing antifibrotics rather than meaningful as monotherapy. The trial wasn't powered for subgroup analysis, so this remains unresolved.
TETON-2 enrolled patients predominantly from Europe, Asia-Pacific, and South America. TETON-1, the companion trial in US/Canadian patients, is still ongoing. Whether the results replicate in a North American population — with different practice patterns, background therapy selection, and disease monitoring — remains to be seen.
The 52-week trial duration doesn't allow conclusions about long-term mortality. IPF survival curves tend to diverge over years, not months. Whether FVC preservation at one year translates into meaningful survival benefit will require longer follow-up.
Dashevsky's Dissection
For patients: IPF is a disease with no cure and limited options. This trial offers a cautious reason for optimism. Lung function declined significantly less in patients treated with treprostinil — and fewer patients experienced major events like respiratory hospitalizations or large drops in FVC. That said, the drug didn't reduce deaths over the one-year period studied, and nearly half of patients developed cough. For someone already managing a progressive lung disease with an inhaler four times a day, adding another inhaled therapy with that side effect profile is a real burden. The conversation between patients and their pulmonologist will need to weigh those tradeoffs carefully.
For pulmonary and critical care physicians: TETON-2 is a genuinely positive trial in a field that has seen a lot of disappointments. A 95 mL between-group difference in FVC is clinically meaningful — the annual rate of decline in untreated IPF is roughly 150–200 mL, so this represents a substantial slowing of progression. The 29% reduction in clinical worsening hazard is also a hard endpoint we can point to with confidence.
What I find most interesting is the subgroup data by background therapy. The benefit signal was weakest and statistically uncertain in patients on no antifibrotic therapy, but clear and robust in those on nintedanib or pirfenidone. That pattern could reflect disease severity differences between groups, or it could mean treprostinil's prostacyclin-mediated antifibrotic mechanism works synergistically with kinase inhibition or TGF-β suppression. We don't know yet — but it reframes how we might think about adding this drug: not as a replacement for standard antifibrotics, but as a third agent in a combination approach.
The tolerability issue is the elephant in the room. Nearly one in three patients stopped the drug before the trial ended. In a real-world pulmonology practice, you're asking patients who are already short of breath to inhale a prostacyclin analogue four times a day and accept 48% probability of cough as a tradeoff for slowing decline. That's a harder sell than it looks on paper. Patient selection will matter — motivated patients with preserved functional status and strong health literacy will do better with this regimen than those with significant comorbidities or limited support.
One more thing worth noting: the acute exacerbation endpoint (3.0% vs. 5.8%, HR 0.54) didn't reach statistical significance by the hierarchical testing rules, but numerically that's a near-halving of acute exacerbation events. Acute exacerbations of IPF carry 50–70% in-hospital mortality. If that signal holds up in TETON-1 and the open-label extension, it would be the most impactful finding of the whole program.
For the health system: Inhaled treprostinil (Tyvaso) already carries an indication for PH-ILD. An IPF indication would substantially expand its market and its cost burden to payers. Tyvaso is not a cheap drug — annual cost is in the range of $100,000+ — and the patient population with IPF is not small. Health economists and payers will scrutinize the cost-effectiveness data closely, especially given that the primary endpoint is FVC preservation, not mortality or hospitalization reduction.
The case for coverage will hinge on whether the clinical worsening benefit — and its component reductions in respiratory hospitalization and FVC decline — translates into downstream cost offsets. IPF hospitalizations are expensive and common. If treprostinil reduces those meaningfully, the cost argument becomes more defensible. If payers view FVC as a surrogate endpoint that doesn't justify the price, access will be uneven.
In summary, TETON-2 is the most meaningful positive trial in IPF since the approval of nintedanib and pirfenidone. The FVC preservation signal is real, and the clinical worsening benefit gives it practical teeth. But the tolerability burden and unresolved questions about long-term mortality mean we should adopt this thoughtfully, not reflexively. The next question isn't whether inhaled treprostinil works in IPF — it's which patients benefit most and whether they can actually stick with it.
Nathan SD, et al. "Inhaled Treprostinil for Idiopathic Pulmonary Fibrosis." New England Journal of Medicine. 2026. DOI: 10.1056/NEJMoa2512911.
