Novartis lowered the blood particle that genes tie to heart attacks. In 8,323 people, the heart attacks did not fall.

The rest of the day

32 more stories on this beat.

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

1. 02

   First ever drug for Alexander disease

   The US drug regulator approved Zanvastro, also called zilganersen, on 3 September for Alexander disease in children and adults. [2][5] Alexander disease is caused by a faulty gene for a protein called GFAP. [2] The protein builds up in the brain's support cells and slowly damages the nervous system, causing seizures, difficulty walking, and the loss of skills a child had already learned. [2] It affects fewer than one person in a million and until now there was nothing to offer but supportive care. [2] The drug is injected into the spinal canal every three months and reduces how much of the faulty protein gets made. [2] In the trial, patients aged five and over walked significantly faster at 61 weeks than untreated patients; children aged two to four improved at standing, walking, running and jumping while untreated children declined. [2] Serious side effects were more common in the untreated group than among those on the drug. [3]

   Why it matters — It is the first approved medicine for this disease and the first to attack its cause rather than its symptoms, and it is the first independent launch from Ionis Pharmaceuticals' brain-disease pipeline. [3][4] Ionis says the drug will reach patients in the United States within weeks. [4] It is the thirty-seventh brand-new drug the regulator has approved in 2026. [5]

2. 03

   Immune cells rebuilt inside the body ease MS

   Sixteen people with multiple sclerosis or other autoimmune diseases got better in a small trial published this week in the New England Journal of Medicine. [6] Multiple sclerosis is a disease in which the immune system strips the coating off nerves. The usual version of this treatment, called CAR-T, means taking a patient's immune cells out, rewriting them in a laboratory over weeks, and putting them back. [6] Here doctors used a modified virus to deliver the rewriting instructions inside the body instead, so the patient's own T cells were reprogrammed where they were. [6] The rewritten cells hunt the B cells that make the antibodies attacking healthy tissue. [6]

   Why it matters — Making the cells in the body rather than in a laboratory is far cheaper and far faster, which decides how many people a treatment can ever reach. David Simon, a doctor and researcher at the Charite university hospital in Germany, called it a proof-of-concept study, and the researchers say it has to be tested in many more people. [6]

3. 04

   AbbVie myeloma drug beats standard care

   AbbVie reported that its drug etentamig produced a response in 74% of patients in a final-stage trial, against 45.7% for the treatments doctors would otherwise have chosen. [7] Multiple myeloma is a blood cancer that starts in plasma cells in the bone marrow and causes bone pain, fractures and infections; more than 130,000 people are diagnosed with it each year worldwide. [7] The trial included 393 patients whose cancer had come back after a median of three earlier treatments. [7] Etentamig is a bispecific antibody: one end grabs the cancer cell, the other grabs a T cell, and it holds them together so the T cell kills it. It is given as a monthly drip. [7]

   Why it matters — Myeloma is still counted as incurable and almost everyone eventually relapses, so the fight is over how long each successive treatment holds. [7] Follow-up is short at a median of 11.4 months, which is not long enough to say anything about survival.

4. 05

   GSK flu vaccine beats the standard shot

   GSK said its mRNA flu vaccine produced stronger immune responses than existing flu shots in a trial of 971 adults, and will start a final-stage trial this month. [8] Flu vaccine effectiveness has swung between 20% and 60% over the past fifteen years, which is the opening every new flu shot is aiming at. [8] Licensed vaccines mostly target one protein on the virus, called hemagglutinin; GSK's targets that and a second one, neuraminidase, on the theory that hitting both protects better. [8] The company's press release carries no numbers behind the claim; a researcher presented details at a flu conference on Monday. [8]

   Why it matters — GSK's flu vaccine business shrank by about 25% last year, so a better shot is the way back into a multi-billion-dollar market. [8] A stronger immune response in a blood test is not the same as fewer people getting flu, and only the final-stage trial can measure that.

5. 06

   A cancer pill at $480,000 a year

   Revolution Medicines set the price of Rasonque, its new pancreatic cancer pill, at about $480,000 a year, or $663 a tablet. [9] The US regulator approved it on 26 August for pancreatic cancer that has already spread, six and a half months before its scheduled decision date. [10] In the trial, patients lived a median of 13.2 months against 6.7 months on standard chemotherapy. [10] Nine cancer drugs launched in 2024 alone at sticker prices above $400,000 a year, according to the Institute for Clinical and Economic Review. [9] Two decades ago a $100,000-a-year cancer drug caused an outcry. [9]

   Why it matters — The New York Times reports there is little constraint on these prices beyond what a company believes the market will bear, and the bill lands on government health programmes, employers, patients and taxpayers. [9] A company spokesman said the price reflects the drug's benefit to patients. [9]

6. 07

   A leukaemia drug shrinks the HIV reservoir in monkeys

   Researchers at Emory University in the United States reported in Nature Microbiology on a short course of venetoclax, a cancer drug already approved for some blood cancers. [11] It shrank the hidden virus reservoir in rhesus macaques infected with SIV, the monkey relative of HIV. [11] Antiviral drugs keep HIV suppressed but cannot clear the small pool of infected cells that survives quietly and restarts the infection if treatment stops. [11] Venetoclax blocks a protein called BCL-2 that helps cells resist dying, which is what lets those infected cells persist. [11] The drug was given at the start of antiviral treatment. [11]

   Why it matters — Clearing that reservoir is the single obstacle between long-term HIV control and an actual cure, and the senior author, Mirko Paiardini, says nothing has managed it so far. [11] This was done in monkeys, and most things that work in animals do not work in people.

7. 08

   A rare immune-storm drug files to go public

   Electra Therapeutics filed to list on the Nasdaq stock exchange in New York, to fund a late-stage trial of its antibody ipsoprubart. [16] It is aimed at secondary HLH, a rare condition in which the immune system overreacts so violently that it attacks the body's own tissue. [16] Cancer, infection, autoimmune disease or immunotherapy can all set it off, and it kills quickly without immediate treatment. [16] The antibody selectively strips out the immune cells that have gone wrong, rather than suppressing the whole immune system the way current care does. [16] In an earlier study of twelve patients whose HLH was driven by cancer, all twelve were alive at eight weeks. [16]

   Why it matters — Both the US and European regulators have given the drug priority status, which Electra says makes it the first for this condition to hold both. [16] Twelve patients with nothing to compare them against is the weakest kind of evidence, and the late-stage trial is not expected to finish enrolling until 2027. [16]

8. 09

   Roche backs a weight drug that spares muscle

   Roche said it is excited about an experimental obesity drug licensed from the South Korean company Hanmi, after seeing results in animals. [17] The drug copies a hormone called UCN2 and acts on a receptor called CRF2, which is a different route from the gut-hormone drugs that dominate weight loss now. [17] In animals, switching that receptor on protected against muscle wasting. [17] Manu Chakravarthy, who leads Roche's heart, kidney and metabolism development, said the animal data suggest weight loss with lean mass preserved and muscle function improved. [17] The first human results will come from a phase 1 study Hanmi is already running. [17]

   Why it matters — Losing muscle along with fat is the main complaint about the current weight-loss drugs, so a drug that avoids it would compete on something other than pounds shed. [17] Chakravarthy said himself that not everything seen in animals carries over into people. [17]

9. 10

   Most RSV vaccine studies rated unreliable

   A review published in JAMA for the coming winter graded the evidence behind RSV shots and found most of it weak. [12] Of 33 observational studies that compared vaccinated with unvaccinated people, only two were low risk of bias; nine were serious risk and nineteen were critical risk, all because of confounding. [12] All sixteen studies with no comparison group at all were also rated critical risk. [12] The randomised trials fared better: five were low risk. [12] RSV is a common winter virus that is dangerous mainly for babies and the very old; in the 2024-25 season US surveillance recorded about 1,117 hospital admissions per 100,000 infants under one year old. [12]

   Why it matters — Confounding here means the people who get a vaccine differ from the people who do not before the vaccine does anything, so the difference measured afterwards is partly just the difference between the two groups. It is the same problem that sits under this edition's lead story.

10. 11

    Two leukaemia drugs turn out identical

    A study of 652 people newly diagnosed with a hard-to-treat form of acute myeloid leukaemia found no difference between the two standard starting drugs. [13] Both azacitidine and decitabine work by loosening chemical tags that switch genes off in cancer cells. Earlier evidence had suggested decitabine cleared the faulty TP53 gene more deeply and produced better responses. [13] Among the 321 patients who received one of them, and after the researchers matched patients to make the groups comparable, survival was the same either way, with p-values of 0.92 for both measures. [13] Median overall survival ranged from 5.7 to 9.2 months across the four regimens. [13]

    Why it matters — TP53-mutated leukaemia is among the worst cancers to be diagnosed with, and doctors have been choosing between these two drugs on a difference that appears not to exist. This was a look back at records rather than a trial, which is why the matching was needed at all. [13]

11. 12

    COVID antibodies fall to a fifth in a year

    A Japanese study of 25,800 adults measured how antibody levels change over two years across every combination of vaccination and infection. [14] Compared with levels thirty days after a third dose, antibodies fell to about 0.18 of that after one year and about 0.10 after two. [14] People who had three, four or five doses ended up with higher levels than people with two, but the rate at which they faded was much the same after the third dose. [14] The data came from four nationwide blood surveys run between December 2021 and March 2023. [14]

    Why it matters — How fast protection fades is what decides whether a booster is needed every year or every few years, and this is one of the few datasets that follows it past the twelve-month mark. [14]

12. 13

    Decades of money worry linked to brain ageing

    Researchers at University College London followed 2,759 people from the British birth cohort of 1946 and found that persistent money trouble in early and middle adulthood went with worse scores on thinking tests by age 53. [15] Among those who later had brain scans at ages 69 to 71, persistently low income also went with poorer brain health, including more shrinkage. [15] The link held after the researchers accounted for childhood thinking ability, education level and childhood disadvantage. [15] The work was published in the journal Innovation in Aging. [15]

    Why it matters — It measures a long exposure across a whole life, which is exactly what a trial cannot do. It also cannot say what happens if the hardship stops, because nobody randomly assigned anyone to poverty. [15]

13. 14

    Radiotherapy drugs may be dosed too cautiously

    Louise Emmett, a leader in radiopharmaceuticals who works with AstraZeneca and Novartis, told STAT that the drug industry is too frightened of these treatments' power. [18] Radiopharmaceuticals are drugs that carry a radioactive atom to a tumour and irradiate it from inside. She and others believe they can safely be given more often than they are. [18] Over the past five years developers have found some of these treatments damage patients' kidneys, liver or bone marrow, and those findings have caused holds and delays. [18]

    Why it matters — The dose a drug is given at is chosen by the organ it hurts second, not by the tumour it is aimed at, so the same medicine can look weak simply because nobody dared give enough of it. [18]

14. 15

    UK pharma pull-back a year on

    A year after several large drugmakers froze research investment in the United Kingdom, Fierce Biotech looked at what changed. [19] The dispute began when the UK government moved to raise the share of new branded medicine sales that companies must pay back to the National Health Service from 15.5% to 31.3%. [19] That rate has since been set at 14.5% of sales for 2026, down from 22.9% in 2025. [19] In April the UK also agreed with the Trump administration that prescription drugs sent to the United States would be exempt from tariffs for three years, in exchange for the UK paying 25% more for new medicines. [19] AstraZeneca then unfroze a 300 million pound investment programme. [19]

    Why it matters — In the United Kingdom the health service buys nearly all the medicines, so whatever rate it sets is what a drugmaker earns there. When the UK government raised the rate, the companies moved research money out; when it cut the rate, AstraZeneca put 300 million pounds back. [19]

15. 16

    China licensing deals pass one hundred

    More than 100 licensing deals between Chinese biotech companies and drugmakers in the United States and Europe have been signed since the start of 2025, with the peak in the first three months of 2026. [20] AstraZeneca, AbbVie and Roche each paid at least $550 million up front for rights to new cancer and obesity drugs. [20] GSK and Roche each signed another this week, with Hutchmed and Simcere Pharmaceutical. [20] The pace has not slowed even as Western governments and investors ask whether the West is losing its lead in drug discovery. [20]

    Why it matters — Buying a Chinese molecule is now a routine way for a large drugmaker to fill a gap in its pipeline, which changes where the early science gets done and who is paid for it. [20]

16. 17

    The bill for shutting a country down

    A modelling study published in a Nature journal estimated the total cost of future respiratory pandemics under four different school and business closure strategies, adding up health, economic and educational losses. [21] For low and lower-middle income countries the mean losses ranged from 2.9% to 439% of a year's national output, depending on how severe the disease was; for high-income countries the range was 1.9% to 290%. [21] Strict closures came out best for the most severe diseases and for richer countries. [21] Reactive closures, triggered by conditions rather than fixed in advance, usually beat long-term school closures. [21]

    Why it matters — The numbers come from a model of six invented diseases, not from any outbreak that happened. It ranks one closure strategy against another; nobody is forecasting a real pandemic here. [21]

17. 18

    A new drug wakes a man from brain inflammation

    Doctors described treating a man in a coma with herpes simplex encephalitis, inflammation of the brain caused by the cold-sore virus, using a drug called pritelivir. [22] He had received a bone marrow transplant from a donor, developed the infection, was treated with acyclovir, and relapsed a month and a half later despite continuing preventive treatment. [22] Two further antiviral drugs gave only a temporary response while his brain swelled and tissue died. [22] After pritelivir was added he became more alert, moved more, and was well enough to be discharged to rehabilitation; the virus became undetectable in his spinal fluid. [22]

    Why it matters — This form of brain inflammation kills about 30% of patients even when treated properly and leaves permanent damage in up to 70% of survivors. [22] It is a single case report, which is the weakest kind of evidence there is.

18. 19

    Chemotherapy makes leukaemia visible again

    Researchers reported in Nature Communications that chemotherapy pushes some leukaemia cells into a state called senescence, in which they stop dividing but do not die. [23] Those stalled cells then start displaying far more identifying markers on their surface, which lets the patient's own T cells recognise and attack them. [23] In laboratory samples and in mice carrying human leukaemia, this made the cancer sensitive again to immune checkpoint drugs, which release the brakes on T cells. [23] The team traced the change to reduced activity of a protein complex called PRC2. [23]

    Why it matters — Resistance and relapse are what actually kill people with acute myeloid leukaemia, and this suggests the chemotherapy that fails to kill a cell may still be setting it up for something else. [23] It was done in cells and mice.

19. 20

    A map of myeloma from 341 patients

    Scientists published a single-cell atlas of multiple myeloma built from 341 people at different stages of the disease and different points in treatment. [24] Reading cells one at a time, they found five recurring patterns of malignant cell behaviour plus a separate growth programme, each linked to genetic features, resistance to treatment and how patients actually fared. [24] They then used the map to hunt for a target found on myeloma cells and almost nowhere else, and picked out a surface protein called FCRL2. [24] Cell therapies aimed at FCRL2 killed the right cells in the laboratory and extended survival in animals. [24]

    Why it matters — Myeloma varies so much between patients that a single classification has never worked well, and a target restricted to one cell type is what makes a cell therapy safe enough to give. [24]

20. 21

    US regulator warns on copied weight-loss drugs

    The US drug regulator issued a warning about unapproved versions of GLP-1 weight-loss drugs, the family that includes semaglutide and tirzepatide. [25] These copies are made outside the approval system, so nobody has checked them for safety, effectiveness or quality before they were sold. [25] The agency published a list of warning signs for consumers buying through telehealth companies, and separate guidance for doctors and pharmacists. [25]

    Why it matters — Demand for these drugs has outrun the approved supply for years, and the gap has been filled by compounded and imported versions whose contents nobody has verified. [25]

21. 22

    Brain-damaging immune cells get orders elsewhere

    Researchers at Washington University in St Louis found that the immune cells which pile into the brain in Alzheimer's-like disease are given their instructions in lymph nodes outside it. [26][27] Patients with Alzheimer's and related conditions have far more T cells in the brain than healthy people, and the more tau protein has built up, the more T cells there are. [26][27] The team showed that a particular kind of immune cell, in tissue outside the brain, is what activates those T cells; mice engineered to lack it were largely protected from nerve damage. [27] The work was published in Nature Neuroscience. [26]

    Why it matters — Getting a drug past the blood-brain barrier is one of the hardest problems in neurology, and if the instructions are issued outside the brain, a drug may not have to get in. [26] This was done in mice.

22. 23

    How bacteria hide from a last-resort antibiotic

    A study in Science Advances explained how bacteria develop heteroresistance to colistin, an antibiotic kept in reserve for infections nothing else touches. [28] Heteroresistance means a small minority of cells in a population survives the drug while the majority is killed, so a standard laboratory test reports the infection as treatable and the patient does not get better. [28] The researchers traced it to several gene-regulating systems acting in combination rather than to a single mutation. [28]

    Why it matters — A resistance mechanism that ordinary hospital testing cannot see is worse than one it can, because the wrong answer arrives with confidence. [28]

23. 24

    Blood test to track a cancer drug's effect

    Natera agreed to work with Angiex, a small Massachusetts company, using Natera's Signatera test in Angiex's first human trial of a cancer drug. [29] Signatera looks for fragments of tumour DNA circulating in the blood, which rise and fall with how much cancer is present. [29] Angiex's drug, AGX101, carries a cell-killing payload to a protein found on tumour cells and on the blood vessels feeding them. [29] The trial is enrolling patients with advanced solid tumours that cannot be removed by surgery. [29] Financial terms were not disclosed. [29]

    Why it matters — Scans often show that a tumour has changed shape without showing whether any living cancer is left, and a blood measure taken repeatedly is meant to fill that gap. [29]

24. 25

    A language model trained on RNA

    Researchers described NucleicBERT, a model that learns the structure and function of RNA from raw sequences alone, without needing comparisons to related species. [30] Most of the human genome does not code for proteins and much of it works directly as RNA, but there is very little three-dimensional structural data on RNA to train on. [30] The team applied the same masked-word training used for language models, hiding parts of a sequence and having the model predict them. [30] It matched or beat existing RNA prediction tools while needing only single sequences. [30]

    Why it matters — Custom RNA drugs are now a real category, and designing one starts with predicting what a sequence will fold into and do. [30]

25. 26

    Where fatty liver treatment now stands

    A review in Nature set out the state of treatment for MASH, the damaging form of fatty liver disease, which raises the risk of liver scarring, liver cancer, heart disease and kidney disease. [31] Two drugs now have accelerated approval for people with MASH and scarring who do not yet have cirrhosis: resmetirom, which acts on a thyroid hormone receptor in the liver, and semaglutide, the weight-loss and diabetes drug. [31] The authors argue MASH is not one disease but several overlapping ones, some driven by fast liver scarring and some by insulin resistance and heart risk. [31]

    Why it matters — If the disease is several diseases, the same drug will keep working brilliantly in some trial populations and failing in others, and combinations chosen by patient type are the way out. [31]

26. 27

    Spit test tells tooth decay from gum trouble

    A pilot study measured nine immune signalling molecules in the saliva of 65 people using an ultra-sensitive test. [32] Two of them together, TNF-alpha and IL-17, separated people with tooth decay from healthy people well, with an accuracy score of 0.901 where 1.0 is perfect and 0.5 is a coin toss. [32] Adding more molecules did not help. [32] Telling tooth decay apart from ordinary gum inflammation was much harder, and a different pair did it only modestly. [32]

    Why it matters — Tooth decay is one of the most common chronic diseases in the world and is still diagnosed by looking, so a saliva measure would be a genuinely new instrument. [32] Sixty-five people is a pilot.

27. 28

    Antibodies from a monkey Shigella outbreak

    Researchers reported in Science Translational Medicine on the antibodies produced during an outbreak of Shigella in a primate colony. [33] Shigella is a bacterium that causes severe dysentery and is a major killer of small children in poorer countries. [33] Resistance to antibiotics is rising among Shigella strains, which is why several groups are working on a vaccine and why knowing which antibodies actually protect matters. [33] The team found antibody responses that work through more than one mechanism at once. [33]

    Why it matters — A vaccine target is chosen from what protected somebody, and outbreaks are where that becomes visible. [33]

28. 29

    Sending drugs straight to the lung

    A review in the journal Signal Transduction and Targeted Therapy surveyed nanoparticle delivery for lung disease. [34] The lung is directly exposed to the outside world, which makes it vulnerable to infection, pollution and inherited disease, and also means a drug can in principle be breathed straight in rather than sent through the bloodstream. [34] Nanoparticles are meant to solve the problems that hold gene and drug therapies back: unstable ingredients, poor uptake by cells, and no way to aim at one tissue. [34] The authors compare breathing a drug in against injecting it, and set out where each wins. [34]

    Why it matters — Getting a drug to the right organ, and only that organ, is what separates a therapy from a side effect. [34]

29. 30

    Alzheimer's blood tests checked in Nigeria

    Researchers measured Alzheimer's blood markers in 967 older adults in Nigeria, part of a study called VALIANT. [35] The markers, including p-tau217 and GFAP, rose step by step from people with no thinking problems through to people with dementia, and the pattern held on two different measuring platforms. [35] Men had higher levels than women. [35] Almost all validation of these blood tests has been done in wealthy countries, and African populations are barely represented in the research at all. [35]

    Why it matters — A blood test is only as good as the population it was calibrated in, and Alzheimer's is a growing burden in low and middle income countries where the tests were never checked. [35]

30. 31

    Pooled heart failure trials answer a fear

    Researchers pooled individual patient records from nine randomised trials, 38,753 people in total, to work out what happens to blood pressure, kidney function and blood potassium when heart failure drugs are combined. [36] Fear of exactly those three things is among the commonest reasons doctors do not start patients on the full recommended combination. [36] The team built a model to estimate the short-term effect, over two to twelve weeks, for a given patient on a given combination. [36] The pool covered both main types of heart failure, with 16,877 patients of one kind and 21,876 of the other. [36]

    Why it matters — It answers a practical worry with pooled trial data rather than with impressions, and the model gives a number for an individual instead of an average. [36]

31. 32

    How to fit a chemotherapy port

    A multicentre study compared two ways of implanting the small device under the skin that lets patients receive chemotherapy without repeated needle sticks: guiding the insertion by feel and anatomical landmarks, or with imaging technology. [37] The researchers found complications too rare to separate the methods statistically. [37] They also noted that most of the procedures were planned in advance and done by experienced operators at high-volume centres, which would push the complication rate down. [37] Because the study looked back at records rather than assigning patients, a proper matching analysis would have shrunk the sample too far to be useful. [37]

    Why it matters — The authors say plainly what they cannot conclude, which is rarer than it should be. [37]

32. 33

    The largest Bundibugyo outbreak on record

    The outbreak of Bundibugyo virus in the Democratic Republic of the Congo is now the largest ever recorded for this member of the Ebola family, and there is no licensed vaccine against it. [39] A modelling study in The Lancet Infectious Diseases compared ring vaccination, which means vaccinating the contacts of each case, against vaccinating whole communities, assuming a vaccine that is 45% effective and also protects people already exposed. [38] A separate paper studied antibodies in survivors in Guinea that cross-react with this virus, which is how a vaccine for a related Ebola strain might be shown to help against this one. [39] A third argues the mental health of outbreak responders is left out of the response, though they face infection risk, exhaustion and the fear of infecting their own families. [40]

    Why it matters — A vaccine that exists for one Ebola strain does not automatically work on another, and every choice about who to vaccinate has to be made before anyone knows how well it works. [38][39]