Novo Nordisk pays China's Hengrui $300 million up front, and up to $2.6 billion, for a once-a-week weight-loss pill no person has taken yet

The rest of the day

13 more stories on this beat.

Each with its own sources. None of these is a link to the story above.

1. 02

   Huntington's gene therapy weakens at four years

   The Dutch company uniQure said on Tuesday that its one-time gene therapy AMT-130 slowed Huntington's disease by 44% after four years in 12 high-dose patients. [4][5] Huntington's is an inherited disease that destroys nerve cells. [5] The 44% was measured against a database of untreated patients, and it was not statistically significant, so chance could explain it. [5] At three years uniQure had reported 75%. [4] A second score, for daily tasks such as managing money, held at 61%, against 60% a year earlier. [4]

   Why it matters — The US drug regulator is reviewing uniQure's request for early approval, filed this month on the three-year data. [4] Shares fell about 40% on Tuesday, and uniQure says the database lost its fastest-declining patients, which makes its drug look weaker. [4][5]

2. 03

   Novo's records favour Ozempic over a switch

   Novo Nordisk said on Tuesday that health records from 636,525 adults with type 2 diabetes showed a 6% lower risk of serious heart and blood-vessel events in people who raised their Ozempic dose. [6] They were compared with people who switched to Mounjaro, the rival drug from Eli Lilly. [6] Within 720 days, 36.9% had raised their Ozempic dose to 2 mg and only 5.7% had switched. [6] The results were shown at the European diabetes meeting in Milan. [6]

   • Stayed on 1 mg57.4%
   • Raised the dose to 2 mg36.9%
   • Switched to Mounjaro5.7%

   Why it matters — This was a study of records, not a trial, so nobody decided by chance who switched and who stayed. Novo and Lilly are fighting over a US obesity market that analysts put at $100 billion to $150 billion by 2030. [6]

3. 04

   Lilly says its pill beats Novo's

   Eli Lilly said on Tuesday that its daily weight-loss pill Foundayo lowered weight and blood sugar more than a higher, 25 mg dose of Novo's Ozempic pill, in adults with type 2 diabetes. [7] The two pills were never tested in one trial. Lilly compared results from two earlier late-stage studies. [7] It put Foundayo ahead by 1.5% to 2.4% more weight lost, and by 0.3% to 0.6% more off HbA1c, a measure of average blood sugar. [7]

   Why it matters — Novo said people with type 2 diabetes lose less weight in studies than people with obesity alone. [7] Its 25 mg pill is not yet approved in the US for diabetes, and it expects a decision by the end of 2026. [7]

4. 05

   A no-fasting diabetes pill passes a final test

   A phase 3 trial, the large final stage of testing, at 46 sites in China tried safiglipron, a daily pill that copies the gut hormone GLP-1. [9] It can be taken with food, at any time of day. [9] The 284 adults had early type 2 diabetes, managed with diet and exercise alone. [9] After 32 weeks, HbA1c, a measure of average blood sugar, fell 1.20 to 1.45 points more than on a dummy pill. [9] Between 71.4% and 77.8% reached the usual target, against 25% on the dummy. [9]

   Best safiglipron dose77.8%

   Dummy pill25%

   Why it matters — The trial's main measure was blood sugar, and weight fell only 0.65% to 3.56% more than on the dummy. [9] Side effects made 6.9% stop on the highest dose, against none on the dummy. [9]

5. 06

   BioNTech closes three German sites

   BioNTech, the German company that first developed the COVID vaccine it sells with Pfizer, said on Monday it will close its sites in Idar-Oberstein, Marburg and Tuebingen. [10] It had tried since May to sell them, while moving its COVID vaccine production to Pfizer. [10] No buyer was found. [10] The wider plan, which also covered a site in Singapore, affects up to 1,860 jobs. [10]

   Why it matters — BioNTech says it has agreed redundancy terms with its works council, the body that speaks for staff. [10] Its two co-founders said in March that they will leave by the end of this year to start a new company. [10]

6. 07

   Roche pill tried on repair-fault cancers

   Roche's experimental pill RO7589831 was tested in 88 people with advanced cancers that carry MSI, a fault in how cells fix errors when they copy DNA. [11] The pill blocks WRN, a protein that these cancer cells need to survive. [11] Of 66 patients whose tumours could be judged, seven saw them shrink clearly, and in 74.2% growth stopped for a while. [11] Those responses lasted a median of 10.2 months so far, and some were still going when the results were published in Nature Medicine on 29 September. [11]

   7 of 66

   patients whose tumours shrank clearly

   Why it matters — This first test in people was about safety and dose, and nausea affected 54.5% of patients, mostly mildly. [11] MSI is found in 4% to 7% of advanced bowel cancers and 15% to 25% of advanced womb cancers, and immune drugs do not help all of those patients. [11]

7. 08

   A Lilly-paid comparison of Zepbound and Wegovy

   A study paid for by Eli Lilly, published on 28 September in the International Journal of Obesity, compared tirzepatide, which Lilly sells as Zepbound, with Novo's semaglutide, sold as Wegovy. [8] The two drugs were never given in one trial. The authors lined up two separate trials, each against its own dummy group. [8] On that basis, the two higher tirzepatide doses took off 5.1 kg and 6.5 kg more than semaglutide, in adults with obesity but no diabetes. [8]

   Why it matters — Six of the ten authors work for Lilly. [8] Lining up two trials cannot rule out that their patients differed, which is what a trial giving both drugs to one group, split by chance, is for.

8. 09

   Years of drugs after a late transplant

   Doctors at the MD Anderson Cancer Center in Texas gave 13 people with myeloma, a blood cancer that had come back, the antibody daratumumab after a stem-cell transplant late in their illness. [17] For six of them it was a second transplant, and 11 also took the pill pomalidomide. [17] The aim was to keep the cancer from coming back, a job called maintenance. [17] After a median of about four years of follow-up, half of them had gone 45.1 months before the cancer grew again. [17] Four years after the transplant, none had died. [17]

   Why it matters — The study had one group and no comparison, and it was funded by Janssen, part of Johnson & Johnson. [17] Newer immune-cell treatments have challenged the place of late transplants in myeloma care. [17]

9. 10

   A gene-therapy virus that waits to be unlocked

   Gene therapies often use AAV, a small virus, to carry a new gene into cells, and it can land in the wrong tissue. [12] Researchers reported on 29 September in Nature Materials a coat that stops the virus entering any cell. [12] Different coats come off only in the liver, under near-infrared light, or where chemicals from inflamed tissue build up. [12] In mice with a damaged heart, the virus delivered a repair gene to the heart muscle. [12]

   Why it matters — It has been tested only in mice, and most results in mice are never repeated in people. [12] The aim is to put a gene only where it is needed.

10. 11

    Fat bubbles built for longer gene messages

    Lipid nanoparticles, tiny fat bubbles, carry mRNA into cells, but they work less well as the message gets longer, a team reported on 28 September in Nature Biotechnology. [13] Gene-editing tools are long messages. [13] The team screened 384 fat-like molecules with a long test message and picked the best, LC-1. [13] In mice it reached editing rates of up to 79% in the liver, 48% in the brain and 27% in the lung, up to four times the standard bubbles. [13]

    Liver79%

    Brain48%

    Lung27%

    Why it matters — It also carried a base editor, which changes one DNA letter, to genes including PCSK9, which affects cholesterol, and a cystic fibrosis fault. [13] All of this was in mice. [13]

11. 12

    Where mRNA vaccine code makes cells slip

    COVID mRNA vaccines use a modified form of one RNA letter, called N1-methylpseudouridine, which stops the body attacking the mRNA itself and makes it last longer. [14] Earlier studies suggested it can make the cell's protein-building machine slip one letter along and make stray protein pieces. [14] A team reported on 28 September in Nature Communications that the slip happens at one short run of letters, UUUC. [14] Rewriting that run as UUCC or UUUU stopped it. [14]

    Why it matters — The slip was measured in cells and test tubes, not in vaccinated people. [14] The authors say careful choice of letters can remove the risk from future mRNA medicines. [14]

12. 13

    AI helps find a glue that destroys a protein

    Researchers at Baylor College of Medicine used AI and a large protein screen to find molecular glues that destroy VAV1, a signalling protein in immune cells. [15] A molecular glue sticks a target protein to the cell's own disposal system, so the cell breaks it down. [15] Their best compound, NGT-201-18, lowered VAV1 in human T cells and calmed their activity. [15] It also destroyed a second protein, LIMD1, that it was not aimed at. [15]

    Why it matters — VAV1 is linked to blood cancers and to autoimmune diseases, where the immune system attacks the body. [15] The compounds have not yet been tested for safety or in animals with disease. [15]

13. 14

    Thirty patients switched eye drugs

    Doctors in Varese, Italy, moved 30 patients with wet macular degeneration from faricimab to an 8 mg dose of aflibercept. [16] Wet macular degeneration is leaking blood vessels at the back of the eye, and a leading cause of lost sight in older people. [16] After three monthly injections the retina was less swollen, and eyesight stayed the same. [16] Patients then went longer between injections. [16]

    Why it matters — The study had no comparison group and was partly paid for by Bayer. [16] Longer gaps matter because each dose is a needle into the eye, and frequent injections can make patients give up treatment. [16]