Two of the most genetically bulletproof cardiovascular drugs ever put into patients just failed — and they did not fail by a little. Novartis tested a drug to lower Lp(a), a genetic risk factor for heart attacks. Novo Nordisk tested one aimed at IL-6, an inflammatory marker widely believed to drive heart disease. Both hit their targets. Neither produced any cardiovascular or survival benefit in trials that ran for years and cost hundreds of millions of dollars each.

“One researcher said to me, this is kind of what you’d expect if you gave people water,” STAT senior medical writer Matthew Herper said on The Readout Loud, reflecting on the fallout from the American Heart Association meeting.

That image — a decade of genetic evidence, thousands of patients, and a result indistinguishable from water — is the story of heart drug research right now. And the lesson, Herper argues, is bigger than two trials. It is that the industry keeps mistaking a validated target for a validated drug.

The null result nobody can explain away

The puzzle is not that the drugs missed their endpoints. It is that there was no hint of efficacy at all. People who carry Lp(a) low for life appear to have lower heart attack risk. Inflammation is so central to heart disease that companies have chased it for decades. Both premises survived the trials; the drugs did not.

ProgramSponsorTargetResultLp(a)-lowering drugNovartisLp(a), a genetic risk factorNo cardiovascular or survival benefitIL-6 drugNovo NordiskIL-6 inflammatory markerNo cardiovascular or survival benefit

Part of the reason the field gets stuck is structural, Herper explains. Cardiovascular trials have to count heart attacks and strokes, and those take a long time to happen — longer now that statins, blood pressure drugs, and blood thinners are everywhere. “If you want to study cardiovascular disease, you have to count heart attacks and strokes, and they take a long time to happen,” he said. “It’s much faster to develop drugs for things like cancer where you can see, do you shrink the tumor at the beginning of the disease?” That is “a real problem for a disease that is the thing we’re most likely to die of.”

Genetic validation: necessary, no longer sufficient

For a decade, the industry’s working theory was that genetic evidence could de-risk a target — if human genetics says this pathway matters, a drug that moves it should work. The AHA results blow a hole in that shortcut.

“Just because you have a whole bunch of genetic data showing that people with lower Lp(a) have lower risk of heart attacks, it doesn’t mean giving somebody a drug that lowers Lp(a) is going to have that same effect,” Herper said.

He does not expect genetics to fall out of use — it remains one of the main ways targets get found. What he expects is that every company marks down its probability of success on similar programs. And his warning extends well past genetics, straight into the AI hype cycle now gripping drug discovery.

“There’s often this idea in the drug business that somebody decides they figured it out,” he said. “And my observation having covered drug development for a long time is that usually that gets you in trouble. It’s going to be more complicated than you think.”

His summary of where the science actually stands is blunt: “Biology is maddeningly complex. It doesn’t make sense. And maybe someday we’ll have some set of tools that will involve genetics and AI and all sorts of other things where we won’t have one out of every 10 drugs fail, but that success rate hasn’t meaningfully changed.”

There is a counterweight in the record. Some of the best cardiovascular wins of the past decade came from drugs nobody expected to help the heart — the GLP-1 diabetes drugs among them. Herper’s read is that companies overestimated how easy genetics would make validation, not that the genetics was wrong.

The retreat already happened — and it is getting worse

If you are waiting for the AHA failures to trigger a pullback from heart drug research, Herper has news: it started years ago. Cardiovascular coverage has already shifted from mass-market prevention of heart attacks and sudden death toward niche medicines.

The mechanism is a financing mismatch. Biotech innovation often runs on raising something like $20 million to reach a milestone a large pharma will buy, and that model has always been hard in cardiovascular — and is getting harder. Herper’s stated worry is losing “that innovation engine that we get in other parts of biotech.”

And yet the commercial prize stays enormous. Cholesterol drugs were the last medicines of GLP-1 scale financially, and Lipitor is still among the most prescribed drugs in the country. That is exactly why companies will keep coming back. “With inflammation, it’s even more complicated because there’s every reason to believe that inflammation plays a role in heart disease,” Herper said. “And if you find the right target, you’re going to have a huge drug.”

Vertex’s kidney drug: approval is likely, the label is the question

The episode’s second strand is a concrete catalyst. The FDA is expected to rule by November 30, 2026 on povetacicept, Vertex’s treatment for IgA nephropathy — a chronic autoimmune kidney disease affecting roughly 200,000 people in the U.S. Host Adam Feuerstein expects approval “almost certainly,” which would make it the third drug in the class to reach market, behind Japan’s Otsuka and Vera Therapeutics.

The U.S. market is worth about $20 billion, so three competitors will be slicing a large pie under intense pressure. The controversy is safety: a risk of serious infections that has emerged recently and could make the drug less attractive than its rivals.

Vertex executives have been adamant that povetacicept does not raise patients’ risk of serious infections, pointing to pivotal-trial data showing infection rates comparable to placebo. That has not settled it. An anonymous citizen’s petition filed with the FDA asks the agency to require monthly immune monitoring for patients who might be prescribed the drug. Vertex says it does not expect monitoring to be required — but the prescription label, unseen until approval, will decide.

Novartis’s gene therapy: nine tumors, and a disclosure question

Novartis confirmed nine total reports of tumors in children treated with Zolgensma, its spinal muscular atrophy gene therapy, after a European advocacy group publicized seven previously unknown cases. Of the nine, two so far show evidence the tumor may be linked to the treatment itself.

Scale matters: the tumor risk is already in the Zolgensma label, and more than 6,000 children worldwide have been treated. But the FDA is investigating, and because many other gene therapies are built similarly, the finding raises class-wide questions. SMA is a progressive neuromuscular disease; untreated babies with the most serious form are likely to die, and Zolgensma has helped children who would likely have died. The episode frames this as one item in a difficult stretch for Novartis, which has also disclosed a few deaths in its CAR-T trials in autoimmune conditions, a high-profile cardiovascular failure, and a neuromuscular trial failure.

Alnylam’s $40 billion lesson: pills beat injections

Then there is the cautionary tale that has nothing to do with biology. Alnylam has fallen to roughly a $20 billion company from about $60 billion when Feuerstein wrote a column arguing it was on its way to $100 billion — a call he conceded on air was as bad as his preseason football take.

The driver is disappointing sales of Amvuttra, a once-every-three-months injectable for ATTR-CM, the progressive heart condition. Amvuttra is a blockbuster, just not blockbustering enough. It competes with pills from Pfizer and BridgeBio, and Alnylam is struggling to grow prescriptions. The company acknowledged lower demand and said Medicare reimbursement cuts starting in January 2027 have clouded its ability to forecast sales.

Feuerstein attributes the stock collapse mainly to that lost forward guidance rather than demand alone — and says the consensus view is that Alnylam needs two or three consecutive quarters of sales at or above expectations before sentiment turns.

What to watch

The through-line across the episode is that validated biology and validated targets keep failing to convert into revenue or clinical benefit on schedule — a sharp counterpoint to the AI optimism now coursing through biotech. The near-term catalysts are concrete: the povetacicept label on November 30, 2026 will show whether the FDA imposes immune monitoring; the January 2027 Medicare reimbursement cuts will test Alnylam’s forecasting credibility; and the ongoing Lp(a) and inflammation programs will reveal whether the AHA nulls were about dose, patient population, or the targets themselves.

For investors, the practical implication is a repricing of certainty. Genetic validation and AI-driven target discovery are still valuable tools, but they no longer justify the near-certain probabilities the market once assigned to them — Vertex’s own chairman has said his company tested roughly 50 AI firms without a single viable drug emerging from the exercise. Herper’s warning is the one worth carrying: the industry’s success rate has not improved, and each new toolkit — genetics, now AI — tends to arrive with more confidence than the underlying biology supports.