Astegolimab’s Split Verdict: What ALIENTO’s Win and ARNASA’s Miss Reveal About COPD Trial Design

Pull up the two Lancet papers side by side. In ALIENTO, Roche’s Phase 2b trial of astegolimab in moderate-to-very-severe COPD, the anti-ST2 monoclonal antibody met its primary endpoint — a statistically significant reduction in annual exacerbation rate across 1,301 patients with frequent exacerbation history. Then open ARNASA, the confirmatory Phase 3 that enrolled 1,375 patients under a nearly identical protocol. Astegolimab delivered a 14.5% reduction in annualized exacerbation rate at 52 weeks. The FDA threshold was not crossed. The trial failed.

That is not a molecule problem. A 14.5% directional reduction in a disease where frequent exacerbators cycle through hospitalizations, emergency bronchodilators, and steroid bursts is a clinically plausible signal. What ARNASA exposed is a structural problem in how the trial was designed to confirm it.

The Eosinophil Stratification Trap

The assumption baked into both ALIENTO and ARNASA was that the ST2/IL-33 pathway would deliver benefit regardless of baseline blood eosinophil counts — a design choice that distinguishes astegolimab’s program from the approved dupilumab strategy. When the FDA approved dupilumab in September 2024 for COPD with an eosinophilic phenotype, it did so on the back of two Phase 3 trials — BOREAS and NOTUS — that explicitly enrolled patients with blood eosinophil counts ≥300 cells/µL. The regulatory signal was clear: enrichment works.

Roche’s bet was different. The ST2/IL-33 pathway drives innate immune activation and epithelial barrier dysfunction — mechanisms that operate upstream of eosinophilic inflammation and, in theory, benefit patients across phenotypes. That scientific rationale is credible. But credible mechanisms do not automatically translate into powered endpoints.

When you design a trial for a biomarker-unselected population, you accept dilution. Patients with low eosinophil counts may derive minimal benefit from an anti-type 2 pathway agent, while high-eosinophil patients may drive the signal. Average those two groups together across 1,375 patients and a true treatment effect in a subpopulation becomes a statistical noise problem in the primary analysis. ARNASA’s 14.5% reduction almost certainly masks a wider spread of responses across eosinophil strata — and the Lancet authors acknowledge as much, noting that subgroup analyses suggest a role for ST2/IL-33 targeting in patients with limited treatment options.

The old assumption: an eosinophil-agnostic mechanism permits an eosinophil-agnostic trial design. The new assumption, after ARNASA, must be that even pathway-level arguments require prospective stratification with pre-specified subgroup tests — or a narrower enrolled population from the outset.

ARNASA vs. the Competitive Field

The stakes here go beyond one molecule. AstraZeneca’s tozorakimab — also an anti-IL-33 inhibitor, targeting the cytokine rather than the ST2 receptor — has now scored a third Phase 3 win with its MIRANDA trial, announced April 20, 2026. Tozorakimab’s program has succeeded in patient populations overlapping with ARNASA’s. That parallel outcome compounds the design question: if the pathway works — and tozorakimab’s three Phase 3 wins suggest it does — why did ARNASA miss?

One answer may lie in receptor-level biology. Tozorakimab blocks IL-33 at the cytokine; astegolimab blocks ST2, the receptor. These are not interchangeable mechanisms. ST2 has both membrane-bound and soluble forms; the soluble form acts as a decoy receptor that sequesters circulating IL-33. An anti-ST2 antibody that blocks the membrane-bound receptor may paradoxically free more IL-33 to drive downstream inflammation in patients with high decoy-ST2 expression — a dynamic that would manifest as heterogeneous response across phenotypes. That hypothesis is speculative, but it is the kind of mechanistic question that a biomarker-stratified adaptive design could have answered prospectively rather than retrospectively in subgroup analyses.

Meanwhile, a narrative review by Kelsen et al., published March 2026, confirms that astegolimab was generally well tolerated in over 580 patients across Phase I and II studies, with no new safety signals emerging in the COPD program. The failure in ARNASA carries no safety asterisk — which keeps the molecule’s development path technically open, even as it narrows.

Sanofi and Regeneron’s itepekimab, another IL-33 pathway asset, has reported its own mixed results in COPD. The field now has a patchwork of Phase 3 outcomes from mechanistically adjacent molecules — some succeeding, some missing — and no clean framework for predicting which patient phenotype each agent will serve. That ambiguity is precisely what adaptive enrichment designs exist to resolve.

The Trial Design Signal for Ops Leaders

If you are designing or amending a Phase 3 protocol for an innate immunity or airway inflammation target in COPD, ARNASA just changed your powering calculus. A biomarker-unselected design that demonstrated a 14.5% AER reduction across 1,375 patients failed to reach significance — meaning your assumed treatment effect in a heterogeneous population must be sized more conservatively than Phase 2b results suggest. ALIENTO’s success in a smaller, similarly unselected population was not reproducible at Phase 3 scale, likely because Phase 2b sample sizes do not adequately penalize you for the dilution you will encounter at full enrollment.

The operational directive: pre-specify eosinophil strata as co-primary or key secondary endpoints in your SAP before locking the Phase 3 protocol. If your molecule has mechanistic reasons to work across phenotypes, design the trial to prove it stratified — not to assume it averaged out. An adaptive enrichment design with a pre-specified interim at 50% enrollment would have given Roche the option to narrow the enrolled population toward higher-responding eosinophil strata based on accumulating data, without compromising the blinded primary analysis. That option was not available post-hoc.

The COPD frequent exacerbator population — typically defined as two or more moderate-to-severe exacerbations per year — represents a high-cost, high-need segment with limited add-on options beyond dupilumab’s eosinophilic indication. The COPD and asthma therapeutics market reflects that unmet need in its trajectory. A molecule that delivers 14.5% exacerbation reduction in an unselected population almost certainly delivers more in a defined subpopulation — but the regulatory system rewards demonstrated significance in pre-specified analyses, not retrospective logic, however sound.

Genentech has indicated it is evaluating next steps for astegolimab following the ARNASA readout. The critical signal to watch now is whether Roche pursues a biomarker-enriched confirmatory trial — or whether the tozorakimab data, combined with ARNASA’s miss, effectively closes the ST2 receptor blocking strategy in COPD and redirects investment toward cytokine-level IL-33 inhibition. AstraZeneca’s MIRANDA results, expected to support a regulatory filing, will set the comparative standard against which any future astegolimab protocol will be evaluated. That filing timeline, and the FDA’s eventual label language around eosinophil count thresholds, will tell sponsors more about the agency’s biomarker enrichment expectations in COPD than any guidance document currently on the docket.

References

1. The Lancet — “Safety and efficacy of astegolimab for COPD with frequent exacerbations regardless of baseline blood eosinophil counts (ALIENTO and ARNASA)”
2. Genentech — “Genentech provides update on astegolimab (ALIENTO Phase 2b and ARNASA Phase 3 results)”
3. Pharmacy Times — “FDA Approves Dupilumab as Add-On Maintenance Therapy for Adults with COPD”
4. Fierce Biotech — “AstraZeneca’s faith in IL-33 inhibitor continues to pay with latest COPD Phase 3 win (MIRANDA)”
5. ResearchGate / Kelsen et al. — “Safety and tolerability of astegolimab, an anti-ST2 monoclonal antibody: a narrative review” (2026)
6. Grand View Research — “COPD and Asthma Therapeutics Market Report”