Novo Nordisk has stopped two more Phase 3 trials of ziltivekimab, its anti-inflammatory antibody, after an independent data monitoring committee concluded neither was likely to succeed. Investigators were informed on September 4, 2026 that HERMES and ATHENA would end early.
The decision follows the failure of ZEUS, the program’s flagship cardiovascular outcomes trial — and the way ZEUS failed is what makes this more than a routine pipeline write-down.
What ZEUS showed
ZEUS (NCT05021835) was a double-blind, placebo-controlled, event-driven trial in more than 6,300 people with atherosclerotic cardiovascular disease, chronic kidney disease and inflammation, the last defined as hsCRP of 2 mg/L or higher. Participants received ziltivekimab 15 mg once monthly by subcutaneous injection or placebo, on top of standard care.
The result on major adverse cardiovascular events was flatly null: HR 0.99 (95% CI, 0.88–1.11). Not a near miss, not an underpowered signal — no effect.
And yet the drug worked, in the biochemical sense. Free IL-6 and hsCRP fell as expected. Target engagement was clean. The inflammation came down and the heart attacks did not. That dissociation between biomarker and outcome is the finding cardiologists will be arguing about.
Why anyone believed the inflammation hypothesis
The idea that residual inflammatory risk drives cardiovascular events — independently of cholesterol — is decades old, and it had real support.
CANTOS, published in 2017, tested canakinumab, an antibody against IL-1β, in post-heart-attack patients with elevated hsCRP. It reduced major cardiovascular events, the first randomized proof that targeting inflammation alone could improve cardiovascular outcomes. It also produced an increase in fatal infections, and the FDA declined the cardiovascular indication. Separately, trials of colchicine, a cheap generic anti-inflammatory, showed benefit in coronary disease.
IL-6 sits downstream of IL-1β in the same inflammatory cascade, and hsCRP is produced in response to IL-6. Blocking IL-6 directly looked like the cleaner, more precise version of the CANTOS experiment — hitting the pathway closer to the readout. ZEUS was supposed to be its vindication.
The two trials just stopped
HERMES (NCT05636176) tested ziltivekimab against placebo in patients with heart failure and inflammation, measuring time to first heart failure event — cardiovascular death, heart-failure hospitalization or urgent visit. ATHENA (NCT06200207) examined quality of life and heart-failure-specific measures in a similar population.
The monitoring committee reviewed the totality of the data and judged that neither was likely to produce a different answer from ZEUS. Patients in both trials will complete their three-month follow-up visits as planned.
What survives
One Phase 3 continues. ARTEMIS is testing ziltivekimab in the acute post-heart-attack setting — a genuinely different proposition. ZEUS, HERMES and ATHENA all addressed chronic, low-grade inflammation in stable disease. The inflammatory response immediately after a myocardial infarction is a different biological event: sharper, larger, and plausibly more responsive to a short intervention at the right moment. Data are expected in the first half of 2027.
That is a narrower thesis than the one Novo started with, but it is not obviously refuted by ZEUS.
What it costs Novo
Novo acquired ziltivekimab through its $725 million purchase of Corvidia Therapeutics in 2020, and had positioned it as a significant non-obesity asset — useful diversification for a company whose valuation has become tightly coupled to GLP-1 franchises facing intensifying competition. Losing three of four Phase 3 readouts removes most of that optionality.
What it means for the rest of the field
Ziltivekimab was not the only bet on IL-6. The pathway has been pursued in chronic kidney disease and in rheumatology, where IL-6 blockade is long established — tocilizumab has been used in rheumatoid arthritis for years, and its cardiovascular effects have been studied with interest precisely because of the inflammation hypothesis. A null cardiovascular outcomes result in a 6,300-patient trial does not invalidate those uses, but it removes the most attractive expansion story attached to the mechanism.
It also raises an uncomfortable question about patient selection. ZEUS enrolled people with both atherosclerotic disease and chronic kidney disease, a population with substantial inflammatory burden but also with competing causes of cardiovascular events — vascular calcification, uremic toxicity, volume overload — that an anti-inflammatory antibody would not touch. Whether the trial tested the hypothesis in the population most likely to benefit, or in a population where too much of the risk was inflammation-independent, will be debated when the full results are presented.
The wider lesson
The most durable consequence may be methodological. ZEUS is a clean demonstration that lowering an inflammatory biomarker is not the same as preventing events, at least via IL-6 in patients with ASCVD and chronic kidney disease. hsCRP has been used for years as a marker of residual inflammatory risk and, implicitly, as a surrogate worth treating. This trial suggests it can be moved substantially without moving what actually matters.
Whether that indicts IL-6 as a target, this particular population, or the surrogate itself is not yet settled. Clinical and business news, not medical or investment advice.