Ebola vaccines teach the body the shape of the virus's outer protein, and the Ebola spreading in Congo has a different one

The rest of the day

34 more stories on this beat.

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

1. 02

   Two treatments tested against Bundibugyo

   A trial called PARTNERS began enrolling Ebola patients in Congo on 2 July. [21] It tests two drugs against the Bundibugyo virus, alone and together: MBP134, a lab-made antibody, and remdesivir, an antiviral drug. [21] Treatments exist for Zaire Ebola, but none is approved for Bundibugyo. [21] No treatment has been shown to work across every type of Ebola virus. [21] Patients who join are followed for at least 28 days. [21]

   Why it matters — Until a drug is proven, the WHO says people who reach a treatment centre early have a better chance of surviving. Four nurses in Bunia recovered that way in late May. [4]

2. 03

   Congo's Ebola response misses its targets

   A report from the US Centers for Disease Control and Prevention (CDC), published on 10 September, measured Congo's Ebola response against five targets up to 21 August. [2] Only 15 to 20% of new cases had been known contacts of earlier cases, against a target above 90%. [2] 59% of Ebola deaths happened outside a treatment unit. [2] Fewer than half of the affected health zones had a team for safe burials. [2]

   Why it matters — When most new cases come from people nobody was watching, the virus is spreading along chains the response cannot see. Bodies of people who die at home can pass Ebola on at funerals, which is why burial teams are one of the five targets. [2]

3. 04

   A vaccine made for Bundibugyo itself

   Oxford scientists built a vaccine aimed at Bundibugyo, called ChAdOx1 Ebola BDBV, with the same method as the Oxford-AstraZeneca Covid vaccine. [22] Its first trial in people began in Oxford in July, and University Hospital Southampton joined on 2 September. [22] It will give 50 healthy adults aged 18 to 55 one or two doses or a dummy injection, and track safety and immune response for a year. [22] In June, CEPI gave about $62m to three Bundibugyo vaccine projects. [23]

   Why it matters — This first trial checks safety, not whether the vaccine stops Ebola. The WHO says vaccines made for Bundibugyo may work better than Ervebo, which was made for another species. [6]

4. 05

   The outbreak spread for weeks unseen

   Congo declared the outbreak on 15 May, when 246 suspected cases and 80 deaths had already been counted. [9] The first alert, which reached the WHO on 5 May, described an unknown illness in Mongbwalu, in Ituri, that had killed four health workers within four days. [9] A CDC computer model published in June suggested the virus jumped from an animal to a person in mid to late February. [19] An infectious disease doctor at Emory University said the size of the first count suggested it had been going on for weeks. [24]

   Why it matters — Every week an outbreak goes unrecognised is a week with no isolation or contact tracing. Bundibugyo's first symptoms, such as fever and tiredness, are hard to tell from malaria without a laboratory test. [1]

5. 06

   The first approved test for Bundibugyo

   On 2 July the WHO added the first test for the Bundibugyo virus to its emergency list, which lets UN agencies and governments buy it knowing it meets quality standards. [5] The test finds the virus's genetic material in a blood sample. [5] The WHO had asked test makers to apply less than two weeks after declaring the emergency. [5] Congo's testing had relied on a few laboratories, mainly in Kinshasa and Goma. [5]

   Why it matters — Without a test, Ebola is easily mistaken for malaria or typhoid, and cases are missed. By late August the CDC found only 72% of checked alerts about possible cases had been tested. [2]

6. 07

   US overdose deaths fell for a third year

   About 70,000 people in the United States died of drug overdoses in 2025, around 14% fewer than in 2024, according to early government figures released in May. [25] It was the third yearly fall in a row, the longest decline in decades. [25] Deaths had peaked at nearly 110,000 in 2022. [26] Seven states went up, including Arizona, Colorado and New Mexico, each by 10% or more. [25]

   2022110,000 deaths

   202480,000 deaths

   202570,000 deaths

   Why it matters — Roughly 11,300 fewer people died than the year before. [27] One researcher suggested the rises in the West may come from more people using fentanyl and methamphetamine together. [25]

7. 08

   Nobody can name one reason deaths fell

   A Commonwealth Fund report on 3 September asked addiction officials in at least two dozen US states and other countries why overdose deaths fell. [28] Every one of them said no single thing did it. [28] They named naloxone, a spray that reverses opioid overdoses, wider use of addiction medicines such as methadone, and treatment in prisons. [28] Virginia saw deaths fall 40% from 2023 to 2024 after a plan to put naloxone everywhere. [28]

   Why it matters — Deaths also fell in Scotland and Canada in 2024, but Scotland's may have risen again in 2025. Experts named cuts to Medicaid, the US health insurance for people on low incomes, as the biggest threat. [28]

8. 09

   New street drugs naloxone often cannot reverse

   US street drugs now hold a fast-changing mix of industrial chemicals that chemists call a synthetic soup, NPR reported in April. [10] They include medetomidine, a powerful sedative that can damage the heart, and new opioids such as cychlorphine and nitazenes, often stronger than fentanyl. [10] These mixes often do not respond to naloxone, the standard drug for reversing an opioid overdose. [10] Quitting medetomidine suddenly can be life-threatening. [10]

   Why it matters — A US government chemist said a substance never seen before turns up every month or two. [10] Nobody buying street drugs can know what a batch holds, and naloxone is made to reverse opioids. [10]

9. 10

   Cocaine and meth deaths without opioids rose

   US overdose death rates fell 27.9% from their 2022 peak to 2024, mostly because fewer people died from fentanyl. [11] But deaths from cocaine or methamphetamine taken without any opioid rose slightly from 2023 to 2024, by 1.7% and 2.4%. [11] The biggest rises were in the northeastern US. [11] Methamphetamine deaths rose among Black, Native American and over-55 groups. [11] University of Minnesota researchers published the study after it was accepted on 23 August. [11]

   Why it matters — The tools behind the overall fall were mostly aimed at opioids. The authors say prevention now has to reach the harms of stimulants too. [11]

10. 11

    A fentanyl shortage may explain part of the fall

    Overdose deaths fell 27% in 2024 to around 80,000, the largest one-year drop ever recorded. [26] A paper in the journal Science by University of Maryland researchers pointed to the drug supply. [26] It argued that rule changes in China cut the chemicals used to make fentanyl, so the drug got harder to make and weaker. [26] US drug agents had reported fentanyl strength rising early in the pandemic and falling after 2022. [26]

    Why it matters — If weaker fentanyl did much of the work, deaths could climb again when the supply changes. Researchers named a shift in the drug supply as one thing that could reverse the fall. [25]

11. 12

    A second weight-loss pill approved

    The US Food and Drug Administration (FDA) approved Eli Lilly's daily weight-loss pill orforglipron, sold as Foundayo, on 1 April. [29] It acts on the same target as the weekly injections Wegovy and Zepbound, the body's receptor for a hormone called GLP-1. [30] In a 72-week trial in people with type 2 diabetes, the highest dose took off almost 10% of body weight. [31] Novo Nordisk's Wegovy pill, approved in December 2025, came first. [31]

    Foundayo pill10%

    Wegovy pill17%

    Tirzepatide injection21%

    Why it matters — Pills are easier to ship than injections and need no cold storage. [29] In separate trials the top dose of the Wegovy pill took off about 17% and the tirzepatide injection about 21%. [31]

12. 13

    Orforglipron beat oral semaglutide in diabetes

    A trial of 1,698 people with type 2 diabetes compared Lilly's orforglipron with Novo's oral semaglutide, the pill Rybelsus, for 52 weeks. [30] A standard blood sugar marker fell by nearly 2 points on the top orforglipron dose, against nearly 1.5 on semaglutide. [30] People also lost more weight on orforglipron, but fewer of them stayed on it. [30] The results were published in the Lancet in February. [30]

    Why it matters — It is a direct comparison of two pills rather than two separate trials. Doctors hope pills like these could let some people with diabetes rely less on insulin, which can cause weight gain. [30]

13. 14

    The weight-loss pill tested in people over 65

    A new analysis looked at 616 people aged 65 or over in two orforglipron trials, published in September in the journal Obesity Pillars. [32] After 72 weeks those on the top dose lost 12 to 13% of their body weight, against about 2% on a dummy pill. [32] Results and side effects matched younger people's. [32] Orforglipron is a small molecule, not a peptide, the protein-like chain the Wegovy pill is made of. [32][29]

    Why it matters — The Wegovy pill has to be taken 30 minutes before any food, and orforglipron has no such rule. [29] Few studies have tested these drugs in older adults, the authors note. [32]

14. 15

    262 obesity drug trials compared

    A review in The BMJ this year pooled 262 trials with 99,791 people, testing 19 weight-loss drugs. [33] After a year, tirzepatide took off about 14.9% of body weight compared with diet and exercise alone, and orforglipron about 9.9%. [33] Injected semaglutide was the only drug linked to fewer deaths from any cause. [33] Orforglipron was among the drugs people most often stopped because of side effects. [33]

    Why it matters — Tirzepatide cut the most fat, 25.7%, but also the most muscle and other lean tissue, 8.3%. [33] No drug clearly improved quality of life by more than the smallest change patients notice. [33]

15. 16

    A third company's weight-loss pill

    Structure Therapeutics, a US biotech company, reported results for its daily weight-loss pill on 16 March. [34] After 44 weeks, people lost about 16% of their body weight compared with a dummy pill. [34] The result came from a phase 2 trial, the middle stage of testing. [34] The company is racing Eli Lilly and Novo Nordisk to sell a pill version of the GLP-1 injections. [34]

    Why it matters — Lilly's orforglipron produced about 11% in its own 72-week trial, reported last year. [34] The two figures come from different trials of different lengths.

16. 17

    An unapproved drug given to one man

    Eli Lilly and the FDA let one 79-year-old man receive retatrutide, an experimental weight-loss drug, STAT reported on 23 June. [18] He got it through compassionate use, a route meant for patients with serious and immediately life-threatening illness. [18] Three sources said the request, made in April, drew interest from top health officials, suggesting the man was well connected. [18]

    Why it matters — Retatrutide has produced weight loss on the scale of stomach surgery in trials, and millions of Americans are waiting for it. [18] Some are already getting it through unofficial sellers. [18]

17. 18

    43 studies, no Tylenol-autism link

    A review in the Lancet in January found that paracetamol in pregnancy, sold as Tylenol in the US, did not raise the risk of autism, ADHD or intellectual disability. [35] Researchers across Europe narrowed thousands of papers to the 43 studies least likely to be biased. [35] Studies comparing brothers and sisters, one exposed in the womb and one not, found the exposed child was no more likely to have autism. [35] In September 2025 US officials had advised using it only if needed. [35]

    Why it matters — Doctors use paracetamol in pregnancy because fever raises the risk of birth defects, and ibuprofen and opioids carry their own risks to the baby. [35] The US health department said the study did not refute other research. [35]

18. 19

    1.5 million Danish children, no link

    A Danish study published on 13 April tracked more than 1.5 million children born between 1997 and 2022, including 31,098 exposed to paracetamol in the womb. [36] Autism was diagnosed in 1.8% of the exposed children and 3% of the others. [36] A similar Swedish study in 2024 found a small link that vanished when siblings were compared. [36]

    12

    0%2%4%

    1. 1Exposed to paracetamol in the womb1.8%
    2. 2Not exposed3%

    Why it matters — Comparing siblings matters because brothers and sisters share much of their genes and home life. The Swedish result suggests autism is strongly genetic, which other studies had already shown. [36]

19. 20

    A Taiwan result nobody can explain

    A Taiwanese study of more than 2 million children born from 2004 to 2015, published in JAMA Pediatrics in March, found a link between paracetamol in pregnancy and autism or ADHD. [16] The link disappeared when siblings were compared, as in Sweden and Japan. [16] But when only the older sibling had been exposed, that child had more risk. [16] When only the younger one had, its risk fell. [16] The authors could not explain it. [16]

    Why it matters — The authors used the oddity to question sibling studies in general, including their own. A Swedish researcher called it a local quirk that does not overturn the other results. [16]

20. 21

    Tylenol use in emergency rooms dropped

    After US officials linked Tylenol to autism in September 2025, emergency-room orders of it for pregnant women fell 16%, then settled 10% lower over 12 weeks. [12] There was no significant drop for women who were not pregnant. [12] Prescriptions of leucovorin, an old cancer drug officials promoted for autism, rose 71% over the same weeks. [12] The study was published in the Lancet on 5 March. [12]

    Why it matters — Paracetamol treats fever in pregnancy, and an untreated fever carries its own risks to the pregnant woman and to the baby's developing brain. [12]

21. 22

    Plastic found in living brain tissue

    Researchers analysed 156 samples from 113 people having brain tumours removed, and 35 samples from five healthy people who had died. [37] Tiny plastic particles turned up in 99.4% of the tumour patients' samples and in every healthy one. [37] Tissue around the tumours held more than healthy brain, perhaps because tumours damage the barrier that protects the brain. [37][38] The study appeared this year in Nature Health. [37]

    Why it matters — More plastic went with faster-growing tumours. The authors say this shows the particles are present, and that more research is needed to learn whether they cause disease. [37]

22. 23

    Doubts about plastic in the body

    Scientists told the Guardian in January that many reports of tiny plastics in the brain, arteries, testes and blood may come from contamination or false readings. [15] The Guardian counted seven studies challenged in their own journals. [15] One common method heats tissue into gas, and fat can give the same signal as polyethylene, a common plastic. [15] The brain is about 60% fat. [15]

    Why it matters — Nobody doubts that plastic pollutes food, water and air. What is in doubt is how much is inside people, and researchers warned that wrong numbers could lead to wrong rules. [15]

23. 24

    Sports betting and binge drinking

    A study reported in March found that US men aged 35 or under who already binge drank did so 10% more often after online sports betting became legal in their state. [39] Binge drinking meant five or more drinks in one sitting. [39] Betting apps are always open, and people often watch sport in bars. [39] The figures come from the men's own reports of their drinking. [39]

    Why it matters — The senior author, an economist, says the study shows a cause but does not test the reason behind it. [39] A 2024 survey found one in four US men aged 30 or under bets on sports online. [39]

24. 25

    Betting on anything, from a phone

    Prediction markets such as Kalshi and Polymarket let people bet on almost any event, and they are regulated as financial markets, not as gambling. [40] Anyone aged 18 or over can join. [40] In Pennsylvania, requests to block oneself from gambling apps rose from 2,200 in 2023 to 3,500 in 2024. [40] During this year's Super Bowl, Kalshi handled more than $1bn in trades. [41]

    Why it matters — Gambling disorder is the only behavioural addiction in the main US psychiatric manual, and it shares brain features with drug addiction. [42] Fewer than one in ten people with gambling problems seeks treatment. [40]

25. 26

    Early strong treatment for Crohn's

    A UK trial followed more than 350 people with Crohn's disease, a lifelong inflammation of the gut, for about five years after diagnosis. [43] Some started strong drugs such as infliximab straight away, instead of waiting until the disease got worse. [43] They were half as likely to be admitted to hospital, and five times less likely to need abdominal surgery. [43] The results were published on 11 September. [43]

    Why it matters — Current guidance usually saves these drugs for people whose disease is already more severe. About one person in 350 in the UK has Crohn's. [43]

26. 27

    Approved, and harder to get

    The FDA approved Rasonque, a pancreatic cancer drug made by Revolution Medicines, in August. [17] Approval shut the company's early-access programme at once, but insurers take weeks or months to start paying for a new drug. [17] Families already in line, such as that of nurse Kelsey Solano, whose husband has advanced pancreatic cancer, were left waiting. [17] One Medicare patient was refused cover. [17]

    Why it matters — Rasonque has been shown to nearly double how long patients live on average, and pancreatic cancer is one of the most aggressive cancers. [17]

27. 28

    A many-cancer blood test missed its goal

    The NHS-Galleri trial gave 143,000 people in England aged 50 to 77 either usual care or a yearly blood test for many cancers at once. [44] After three rounds it did not cut the combined number of stage 3 and 4 cancers, its main goal. [44] Stage 4 cancers fell, but more stage 3 cancers were found. [44] Of people who tested positive, 52% really had cancer. [44]

    Why it matters — Stage 3 cancer is often still treatable with surgery or radiation, so an oncologist writing in STAT argued the shift from stage 4 matters. [44]

28. 29

    The morning cancer trial was retracted

    In February a trial in Nature Medicine reported that lung cancer patients given immunotherapy in the morning did far better than those treated in the afternoon. [14] Other scientists soon pointed to inconsistencies in its data. [14] The journal retracted the paper on 24 June. [14]

    Why it matters — The result had raised the hope that simply moving infusions to the morning could make cancer drugs work better. [14] That hope has now lost its main trial.

29. 30

    Mpox reaches new countries

    The WHO's mpox report of 23 January counted 52,845 confirmed cases and 215 deaths in 98 countries during 2025. [45] Clade Ib, one form of the virus, was reported for the first time in Czechia, Israel, Madagascar, Nepal and France's island of Mayotte. [45] Madagascar had an outbreak spreading inside the country from December. [45] France, Italy and Spain reported it passing between people locally. [45]

    Why it matters — The global emergency over mpox ended in September 2025, but the virus keeps reaching new countries. [45]

30. 31

    Ethiopia's first Marburg outbreak ended

    Ethiopia declared its first ever outbreak of Marburg virus, which causes an illness like Ebola, on 14 November 2025 in the southern town of Jinka. [46][23] There were 14 confirmed cases, nine of whom died, plus five probable cases, all of whom died. [46] Health workers followed 857 contacts for 21 days each. [46] It was declared over on 26 January, after 42 days with no new case. [46]

    Why it matters — There is no approved vaccine or treatment for Marburg either. It was stopped by following contacts, testing and safe burials. [46]

31. 32

    The cruise ship hantavirus outbreak ended

    The WHO declared the hantavirus outbreak linked to the cruise ship MV Hondius over on 2 July. [47] It caused 13 cases, three of them fatal. [47] Health authorities followed more than 650 contacts in 33 countries and territories. [47] Spain helped passengers and crew leave the ship safely in Tenerife. [47]

    Why it matters — The WHO is now running a study in 21 countries on how the disease develops, to help build tests, drugs and vaccines for future outbreaks. [47]

32. 33

    US addiction grants cut, then restored

    On 13 January, SAMHSA, the US federal agency for mental health and addiction, cancelled hundreds of grants worth about $2 billion. [13] After protests from both parties, the money was restored the next evening. [13] The grants paid for mental health and drug treatment services around the country. [13]

    Why it matters — Medicaid, the US government's health insurance for people on low incomes, pays for a quarter of US spending on mental health and addiction care. The 2025 budget law cuts its federal funding by $1 trillion over 10 years. [13]

33. 34

    Britain's new clinical trial rules

    The UK's biggest change to its rules for medicine trials in more than 20 years came into force on 28 April. [48] The Health Research Authority and the Medicines and Healthcare products Regulatory Agency, the UK drug regulator, wrote them together. [48] They line up how the regulator and ethics committees review applications, and aim to speed approvals while protecting volunteers. [48]

    Why it matters — Every new medicine reaches patients only after trials, so these rules set how fast UK trials can start and how volunteers are protected. The law was signed in April 2025 and given a year to take effect. [48]

34. 35

    Few medical AI tools tested properly

    A review published this year in npj Digital Medicine sorted 4,667 studies of medical AI. [49] 88.2% were early lab-stage work, and only 2.4%, or 113 studies, were randomised trials in patients. [49] Two thirds of those trials ran in a single hospital. [49] 82.3% reported good results, but many had weak methods. [49]

    Why it matters — The research came mostly from a few countries: ten produced three quarters of it. [49] Only a trial shows whether a tool helps patients, and fewer than one study in 40 was one. [49]