Before I believed a single percentage in this week’s heart failure inflammation headlines, I wanted to know who had paid for the counting. I checked the three big recent studies behind those headlines, and all three were funded by Novo Nordisk. The same company is now running a late-stage trial of an anti-inflammatory antibody in exactly the patients being counted.
The latest count landed Monday at the Heart Failure Society of America meeting in Phoenix. There, researchers reported that cardiovascular inflammation was highly prevalent among U.S. heart failure patients, a registry finding MedPage Today said illustrated “the unmet need for good inflammation-targeting pharmacotherapy.” The talk was the U.S. arm of a study called POSEIDON. It sits on Novo Nordisk’s own HFSA 2026 congress page as a seven-minute short oral presentation, and one of its nine authors, Jeffrey Skaar, shows up again as a Novo employee on another of the studies below.
Three counts, one paymaster
POSEIDON’s global results appeared in the European Journal of Heart Failure in May. The study drew on 11,809 heart failure patients at 317 sites in 18 countries and flagged “high inflammatory risk” at a high-sensitivity C-reactive protein (hsCRP) level of 2 mg/L or above. It found nearly the same share in every type of heart failure: 38.8% in the preserved-ejection-fraction form and about 38% in the other two. Healio reported that Novo Nordisk funded it and that lead author Carolyn Lam of Singapore holds Novo research grants and advisory roles. Four of her co-authors list Novo Nordisk A/S as an affiliation.
The second count went looking inside Novo’s own drug trials. Barry Borlaug of the Mayo Clinic and colleagues pulled baseline blood work from the 5,468 participants with heart failure in three semaglutide outcome trials, SELECT, SOUL and FLOW. These were trial volunteers, not a random slice of clinic patients, and they ran hotter. Among those with preserved or mildly reduced ejection fraction, hsCRP was elevated in 52.5%. In FLOW, a trial built around kidney disease, the same subgroup hit 65.5%. Novo Nordisk funded the analysis, and three of its authors work for the company.
The third, published in Frontiers in Cardiovascular Medicine, mined a U.S. claims-and-lab database for 11,015 people with preserved-ejection-fraction heart failure and found systemic inflammation in 44.0% of them. It also has the strongest outcome signal of the three. Inflamed patients had 103.2 composite heart failure events per 1,000 person-years, compared with 83.3 for everyone else, an adjusted hazard ratio of 1.21. Death from any cause didn’t differ significantly (HR 1.14, P = 0.09). Novo Nordisk Inc. funded the work, and Skaar is among the authors listed as company employees.
So the share now in circulation runs from 38.8% in a global registry to more than half of semaglutide trial volunteers, depending on who’s being measured, and all three numbers came out of one company’s research budget. I don’t dispute that lots of heart failure patients carry elevated CRP; the studies agree on that much. What bothers me is the conclusion the field is being handed.
Lam spelled out the purpose to TCTMD: POSEIDON “provides a real-world estimate of the potentially eligible population for therapies directly targeting inflammation.” That’s market sizing, said out loud. The therapy in question is Novo’s. The design paper for the ATHENA and HERMES trials describes ziltivekimab, a monthly 15 mg injection that blocks interleukin-6, being tested in heart failure patients who meet the same bar POSEIDON used: hsCRP of 2 mg/L or higher. In September 2025, Novo cut the planned size of HERMES, the outcomes trial, from 5,600 patients to 4,900. By March 26, 2026, it had enrolled 4,769, or 97% of the target. Novo Nordisk is the sole funder, and the drug remains investigational.
Where the alarm starts
CRP is a protein your liver pours into the blood when it gets a chemical alarm, mostly interleukin-6, and the Borlaug paper notes that IL-6 sits upstream of CRP. Shut off the alarm and the liver stops responding. That’s why ziltivekimab cut hsCRP by 77% to 92% versus placebo in kidney patients, depending on the dose, according to a 2025 review of the anti-inflammatory trials.
So what sets the alarm off? I read POSEIDON’s predictor list twice, because it reads less like a heart disease profile than an inventory of waistlines, cigarettes and kidneys. The factors independently associated with high CRP were smoking, obesity, an eGFR under 45 (kidney filtration well below normal), abnormal blood lipids, a history of rheumatic, autoimmune or inflammatory disease, and worse functional class. Body mass index mattered most in the preserved-ejection-fraction form, the type most closely tied to obesity. In the semaglutide-trial analysis, greater adiposity was “the strongest determinant of hsCRP,” and CRP climbed about 20% with each added comorbidity.
Those are associations. They don’t prove cause. Even so, Lam herself told TCTMD that in preserved-ejection-fraction heart failure, inflammation “is primarily attributed to comorbidities.” Charalambos Antoniades of Oxford, who founded the inflammation-imaging company Caristo Diagnostics, made the sharper point that CRP picks up systemic inflammation but doesn’t identify where it comes from. A reading above 2 means the liver is responding to something, whether that’s inflamed fat, struggling kidneys, a smoker’s airways or the heart muscle. The number can’t say which, and it can’t say whether quieting the signal will keep anyone out of the hospital.
Lowering CRP has never been the same as helping hearts
Cardiology has tried to cool the inflamed, failing heart before, and the randomized record is humbling. RENEWAL tested the TNF blocker etanercept in 2,048 patients and was stopped early for lack of benefit. Infliximab did worse: in the 150-patient ATTACH trial, the 10 mg/kg dose raised the risk of death or heart failure hospitalization, with a hazard ratio of 2.84.
Colchicine, the cheap gout pill, shows how a lab marker can improve while the patient doesn’t. In 267 patients with chronic heart failure it lowered hsCRP by an average of 5.1 mg/L and did nothing for symptom class. In the 278-patient COLICA trial in acute heart failure, CRP fell 70.8% on colchicine versus 51.1% on placebo, with no difference in NT-proBNP or in death and worsening heart failure. The field’s best case, CANTOS, wasn’t a heart failure trial at all. It enrolled 10,061 heart attack survivors with hsCRP of 2 or higher, and at the 150 mg dose the IL-1 blocker canakinumab lowered hsCRP 37% versus placebo and cut major cardiovascular events 15%, at the price of more fatal infections.
Then, this summer, Novo’s own antibody faced its first big outcomes test. On July 31 the company announced that ZEUS, which randomized more than 6,300 people with atherosclerotic disease, chronic kidney disease and hsCRP of at least 2 mg/L, showed no reduction in heart attacks, strokes or cardiovascular deaths, with a hazard ratio of 0.99 and a 95% confidence interval of 0.88 to 1.11. Serious infections were more common on the drug. Chief scientific officer Martin Holst Lange said ziltivekimab “did not achieve the MACE benefit we had hoped for.” I read that hazard ratio three times. The drug was the same and so was the CRP entry bar, and the antibody and the placebo came out even.
| Result | Value (HR) | Lower bound | Upper bound |
|---|---|---|---|
| Ziltivekimab vs placebo | 0.99 | 0.88 | 1.11 |
HERMES tests a different population and could still succeed. It’s aiming for roughly 845 adjudicated events, and Novo says it is “anticipated to read out in the first half of 2027.” It deserves a fair reading when it does. For now, though, the only finished outcomes trial in this program is a null, and a CRP value on a lab report isn’t evidence that anyone with heart failure needs this injection.
If the antibody is the drug-cabinet answer, the alternative is to go after the items on the predictor list. Here the evidence is thinner than I’d like. Exercise training clearly helps heart failure patients function, but a pooled analysis of 106 patients found it lowered TNF-alpha by 1.9 pg/mL and left IL-6 essentially unchanged, even as peak oxygen uptake rose about 9%. Nobody should promise that a walking program will erase a high CRP.
Weight is where it gets awkward for Novo. In the 529-patient STEP-HFpEF trial in obese heart failure patients, CRP fell 43.5% on semaglutide versus 7.3% on placebo, alongside weight loss of 13.3% versus 2.6%. The trial can’t separate how much of the CRP drop came from lost fat and how much came from the drug’s other effects, but the result lines up neatly with POSEIDON’s predictor list. It was also a Novo drug in a Novo trial. Whether the eventual answer is shrinking the fat tied to the alarm or blocking the alarm itself, the same company gets paid.
| Group | Value (% reduction) |
|---|---|
| Semaglutide | 43.5 |
| Placebo | 7.3 |
None of this makes the prevalence numbers wrong. Until an outcomes trial says otherwise, though, they work as a sales argument, and doctors are learning to read a high CRP rate as an empty shelf in the drug cabinet from the company that owns the cabinet. The predictor tables point at smoking, excess fat, damaged kidneys and bad lipids, which medicine already knows how to treat.
If I had heart failure and my doctor showed me a CRP above 2, I’d put my effort into the cigarettes, the waistline, the kidneys and the lipids. I wouldn’t take a monthly anti-inflammatory injection until HERMES shows it keeps people like me out of the hospital.
Sources
- European Journal of Heart Failure: Lam et al., high inflammatory risk across heart failure subtypes in POSEIDON (2026)
- JACC: Heart Failure: Borlaug et al., systemic inflammation across the spectrum of ejection fraction
- Frontiers in Cardiovascular Medicine: systemic inflammation and outcomes in HFpEF (2026)
- European Journal of Heart Failure: rationale and design of the ziltivekimab ATHENA and HERMES trials
- Novo Nordisk: update on the ZEUS phase 3 trial of ziltivekimab (via BioSpace)
- Heart Failure Reviews: targeting inflammation in heart failure, evidence from randomized trials (2025)
- TCTMD: Inflammation common in HF, no matter the ejection fraction (POSEIDON)
- Healio: Up to 40% of people with heart failure may benefit from therapies targeting inflammation
- TCTMD: STEP-HFpEF, semaglutide cuts weight and symptoms
- Pooled analysis: exercise training, IL-6, TNF-alpha and peak oxygen uptake in heart failure
- MedPage Today: Rampant inflammation in heart failure, HFSA 2026
- Novo Nordisk Science Hub: HFSA 2026 congress presentations