Four drug programs died this week. One winner took every headline.

Biotech & Longevity 4 min 80 sources

While Eli Lilly's obesity pill hit 22.6% weight loss, four other drug programs were quietly killed off — a phase 3 gene therapy, a sickle cell drug, a cancer radioligand, and an eczema treatment bought for $1.1 billion. Plus a child's death in a China gene-editing trial, safer CAR-T signals, and a US bill to cut animal testing.

Key takeaways

Four drug programs were quietly killed this week — including a phase 3 gene therapy that failed 531 patients — while Eli Lilly's 22.6% weight-loss result took the spotlight.
Roughly 9 of every 10 drugs that reach human testing never make it to approval, but only the survivors get remembered.
A six-year-old's death in a China gene-editing trial and a US bill to cut animal testing show a field pushing hard at both its frontier and its rules.

The week’s quiet burials

The biotech headline of the week was easy to spot: Eli Lilly’s triple-hormone drug retatrutide cut body weight by 22.6% in a large trial [11]. What the headlines mostly skipped was the graveyard filling up beside it.

In the same week, four separate drug programs were killed. Kolon TissueGene’s cell-based gene therapy for knee osteoarthritis failed its phase 3 trial — the last and biggest test before approval [4]. One year after treatment, the 531 patients did statistically no better than those given a placebo [4]. Agios abandoned its sickle cell drug tebapivat after weak phase 2 data, axing an indication for the second time in two months [22]. Novartis discontinued all studies of a phase 2 cancer radioligand. Early data “did not support advancing it,” the company said — not a safety problem, just a drug that didn’t work well enough [30]. And Sanofi dropped plans to seek approval for an eczema drug it had acquired in a $1.1 billion buyout [2].

Four deaths, one jackpot. By next month, most people will only remember the jackpot. Today’s lesson is about that gap.

The obesity gold rush keeps its shine

Retatrutide is a “triple-G” drug — it acts on three appetite-and-metabolism hormone receptors at once (GIP, GLP-1, and glucagon) [11]. Lilly tested it against a placebo in two large trials. TRIUMPH-2 enrolled 1,152 patients with type 2 diabetes and obesity; TRIUMPH-3 enrolled 1,949 with severe obesity and existing heart disease [11]. It hit its main weight-loss target. Its effect on actually cutting heart attacks and strokes, though, was less clear [11].

The search for a gentler version is on. Stanford researchers used AI to find a natural molecule, nicknamed BRP, that curbed appetite in animals without the nausea, constipation, or muscle loss that current shots can cause [44]. The heavy caveat: this was an animal study, and most things that work in animals never work in people [44]. Separately, an FDA advisory panel recommended loosening US rules so pharmacies can compound three more weight-loss-related peptides — widening access, but to versions the agency hasn’t formally reviewed [25].

Gene editing’s hardest reminder

The week also carried a sobering story. Science reported that a six-year-old girl in China died after receiving an experimental, brain-directed CRISPR therapy, raising serious questions about how such trials are overseen [20]. CRISPR is a tool for editing DNA inside living cells; aimed at the brain, in a child, it sits at the frontier of what anyone has tried.

The lab news ran the other way. Researchers reported a method to insert new DNA precisely without cutting both strands of the double helix [26]. That cut is the very step that makes today’s gene editing risky and hard to control. Promise and peril, in the same field, in the same week.

Signals in cancer

Two quieter cancer results are worth holding. Yescarta is a CAR-T therapy — it reprograms a patient’s own immune cells to attack cancer. A study of patients from its trials found genetic markers that flag who is most likely to suffer severe side effects [28]. That could let future CAR-Ts be engineered to be safer [28]. Cevostamab is an antibody designed to grab a cancer cell and a T cell at the same time. It showed activity in a phase 1 trial of 324 people with multiple myeloma [3]. These were heavily pretreated patients, a median of six prior therapies, so any response counts. But phase 1 only tests early safety and signals — not proof it works [3].

A busy week for the brain

Alzheimer’s research produced a cluster of findings. One team reported that a gene variant called APOE2 seems to protect the brain from both Alzheimer’s damage and ordinary aging [19]. Another found a protein that shields brain cells from the damage seen in the disease [12]. And a third group grew tiny “mini-brains” from human cells that may help predict which Alzheimer’s treatments will actually work before they reach patients [38]. All are early-stage lab science, not treatments — but each chips at a disease that has broken more drugs than almost any other.

The under-covered story: rethinking the animal test

The US House passed the FDA Modernization Act 3.0 [5]. It directs the FDA to update its rules so drugmakers can lean less on animal testing and more on newer tools, like lab-grown tissue and computer models [5]. The bill is bipartisan and now goes to the Senate [5]. It matters for a quiet reason. Animal tests are themselves a filter — they let some drugs through and kill others. And most drugs that pass in animals still fail in humans. Changing that filter changes which drugs ever reach the graveyard the rest of this briefing is about.

02 · Lesson · why it matters

Why you only ever meet the survivors

We judge how well medicine works from the drugs that reached the shelf — and never from the far larger number that died on the way.

Four funerals and a headline

This week gave you a natural experiment in what you notice.

One drug won. Eli Lilly’s retatrutide cut body weight by 22.6%, and the number travelled everywhere. Four other drugs lost. A gene therapy failed its final trial after a year and 531 patients. A sickle cell drug was dropped for weak data. A cancer radioligand was quietly shelved. An eczema drug bought for over a billion dollars was abandoned before it even reached the regulator.

The winner got a press release, ads, a stock bump, and this briefing. The four losers got a line each in a corporate filing. In a month, almost no one will remember them. And that forgetting is not an accident. It is the exact shape of a trap called survivorship bias.

The graveyard you are never shown

Survivorship bias is what happens when you judge from the things that made it through, and never see the things that dropped out. The survivors are visible. The failures are gone, so they leave your sight before you even start looking.

The cleanest example comes from the Second World War. Analysts studied bombers returning from missions to decide where to add armour. The planes came back peppered with bullet holes on the wings and tail, and almost none on the engines. The obvious move was to armour the wings, where the damage was.

A statistician named Abraham Wald saw it backwards. The planes in front of him were the ones that came home. A plane hit in the engine did not come home to be counted. The undamaged engines were not proof engines never got hit — they were proof that a hit there was fatal. Armour the engines. The survivors were pointing at exactly the wrong lesson, and the missing planes held the real one.

Medicine is the purest case there is

No field runs this experiment as brutally as drug development. Of every ten drugs that reach human testing, only about one ever gets approved. The other nine die somewhere along the way, and each phase is a gate that most do not pass.

Retatrutide is not proof the field is winning. It is the rare plane that came home. Kolon’s gene therapy is the more ordinary story. It survived its early trials, reached the last and biggest gate, and died there — in front of 531 people who had volunteered. Phase 3 is the final test before approval, hundreds of patients, the drug against a placebo. Surviving to phase 3 means clearing years of gates the other nine never did. And it still was not enough.

When you read only the winners, you are reading a shelf that has already deleted the graveyard. You quietly upgrade your sense of how reliably medicine delivers, because the failures were removed before the story reached you.

Who fills the graveyard, and never gets thanked

The cost of the missing planes is not abstract. It has names.

The 531 patients in Kolon’s trial and the 59 in the sickle cell study carried the risk, gave their time, and got back a single flat no. Their contribution is real and it is invisible, because it produced no product to point at. Investors priced the winners and ate the losses. And you — reading “22.6%” without reading the four obituaries beside it — walk away with a rosier sense of the odds than the odds deserve.

That matters the next time a trial result lands near something you care about. A diagnosis, a parent, a hope pinned on a headline. If you only ever see the winners, you will expect the next drug to work far more often than it will.

Why the graveyard stays quiet

There is a machine that keeps the failures out of view, and it is not a conspiracy. Winning is worth announcing; losing is not. A good result buys a press release and a rising share price. A dead program gets a sentence in a regulatory document that no one outside the company reads.

So the record you inherit is pre-filtered toward good news, one press cycle at a time. The point is not that anyone lied. It is that the survivors shout and the fallen are silent, and if you only listen you will only ever hear from the survivors.

The whole is bigger than the shelf

The humble move is small and hard to remember: when you see a triumph, ask what you are not being shown. The graveyard is almost always larger than the winners’ club, and it is the part no one puts on a shelf.

You are inside this too. Every “studies show” you have ever read had its failures quietly removed before it reached you. Seeing the survivors is effortless — they are the only ones still standing. Counting the graveyard is the whole discipline, and none of us, not the reader, not the company, not the regulator, can ever see all of it.

03 · Lab · your turn

The Survivors' Shelf

Bet how many of 20 drugs reach approval, then watch the graveyard fill and feel how hiding the failures makes the survivors look like the whole story.

04 · Hope · carry this

Every one of this week's failures was also an honest answer - a door closed so the next team does not waste years walking through it. That quiet, unglamorous no is exactly how, over time, medicine keeps finding its yes.