Short sight is an eyeball that grew too long, and the thing that seems to slow the growing is bright daylight

The rest of the day

42 more stories on this beat.

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

1. 02

   Slower chewing cut 369 calories a day

   Forty-one adults ate two ultra-processed diets for two weeks each, in random order. The two menus were matched for taste, portion size served, total calories served and energy density. The only real difference was food texture, which sets how fast people eat. On the slower-textured menu they took in 369 fewer calories a day, and the gap held steady across all fourteen days [14]. A separate review lists faster eating as one of the routine features of ultra-processed food [19].

   Why it matters — Both arms of the trial were ultra-processed, so whatever the label is picking up, a large part of it here was chewing speed.

2. 03

   The unprocessed diet won on weight, not willpower

   In 2019 a US feeding trial found people ate about 500 calories a day more on an ultra-processed diet than an unprocessed one. Researchers have now gone back to the plate-by-plate records of those twenty adults. On the unprocessed diet people picked more low-calorie items, mostly fruit and vegetables. Their dinners came out 57% heavier and still held fewer calories, 719 against 830 [15]. Nobody had told them to choose that way.

   Why it matters — The difference sat in what people reached for when offered a choice, which is a different thing to change than how much food is put in front of them.

3. 04

   A food trial that found nothing, except in teenagers

   Twenty-seven people aged 18 to 25 spent two weeks on a diet that was 81% ultra-processed and two weeks on one with none, matched for protein, fat, carbohydrate, fibre, added sugar and energy density. Afterwards each was offered a free buffet. Across the whole group there was no difference in how much they ate. Split by age, the 18 to 21-year-olds ate more after the ultra-processed fortnight and the 22 to 25-year-olds did not [16].

   Why it matters — An effect that is real in late adolescents and absent in young adults vanishes entirely once the two are averaged into one number.

4. 05

   Nobody agrees what the category is

   There is no accepted definition of ultra-processed food, and a critical review argues the published definitions predict different health risks from one another. Most of the evidence tying processing to obesity is observational, and the review calls those links weak and inconsistent. It also notes that in community trials people often did not manage to cut their intake, so the trial tested the advice rather than the food [17]. A scoping review separately links higher intake to depression, anxiety and attention problems, rising with the amount eaten [18].

   Why it matters — Several governments are being asked to write dietary rules around a category whose boundary is still being argued over.

5. 06

   Creatine added a little to lifting

   Eight randomised trials, 482 older adults, all doing resistance training two or three times a week for between 8 and 104 weeks. Half also took creatine, a compound found in meat and made by the body, which helps muscle produce short bursts of power. Against training alone it added a small gain in leg strength and lean tissue, and nothing measurable in the arms. In the trials running 32 weeks or less the gains were larger [20].

   Why it matters — The extra is real and it is small, and the training is still the part carrying almost all of the result.

6. 07

   They looked for the dial and found none

   Fourteen trials of 528 older adults with sarcopenia, the age-related loss of muscle, were pooled to ask which training settings matter. The researchers pulled out sessions per week, load, number of weeks, sets, whether the programme got harder over time and whether people trained to failure. None of those explained why some programmes built muscle and others did not. The one thing that did was age: the trials that failed had recruited significantly older people [21].

   Why it matters — The field went looking for the setting to change and found that who was recruited mattered more than any setting.

7. 08

   Older women got stronger without getting bigger

   Twelve trials covering 518 older women with sarcopenia were pooled. Resistance training improved grip strength, walking speed, knee extension strength, a timed stand-and-walk test and a 30-second chair-stand count. It did not shift skeletal muscle mass index at all. Muscle mass is one of the measures the condition is defined by, and it is the one that did not move. Women who also carried obesity gained more knee strength than the rest [22].

   Why it matters — Getting out of a chair is what the training bought, and the scan of how much muscle is present reports nothing happened.

8. 09

   Amino acids on top, rated near the bottom

   Nine randomised trials, 496 people with sarcopenia, testing whether adding amino acids beats resistance training alone. Amino acids are the building blocks of protein; the trials used essential and branched-chain forms, leucine, and the derivative HMB. Grip strength, walking speed, a physical performance battery and a sit-to-stand test all improved with supplements added. Skeletal muscle mass did not. The reviewers graded the certainty of all of it as low to very low [23].

   Why it matters — A large supplement trade rests on this question, and the evidence under it is rated near the bottom of the scale researchers use.

9. 10

   In bed rest, only some kinds of lifting worked

   Long bed rest strips muscle fast, which is why countermeasures get tested there. A pooled analysis found that under the single heading of resistance training, plain resistance work produced no measurable gain in muscle volume. Flywheel training, which loads a muscle hard while it lengthens, and concurrent training, which mixes strength and endurance work, both produced large gains. Training three days a week worked; more than three days a week did not [24].

   Why it matters — One heading covered three different things, and two of them did the work while the third did nothing.

10. 11

    Two rankings, two different winners

    A Bayesian network meta-analysis pooled 35 randomised trials of 2,331 older adults with sarcopenia, diagnosed by the European and Asian working-group criteria. Exercise combined with nutritional supplements ranked first for grip strength, walking speed and limb muscle mass, adding 3.69 kilograms of grip and 0.11 metres per second of walking speed [25]. A separate network analysis of 17 trials in 1,107 pre-frail older adults put elastic-band exercise top for grip, and a progressive programme with tai chi top for balance [62].

    Why it matters — Both analyses are ranking whole programmes against each other, so neither can say which part of the winning programme won.

11. 12

    The largest gain is from starting at all

    Muscle loss begins around age 35 and runs at 1% to 2% a year, rising to about 3% a year after 60. An adult who does no strength training can expect to lose four to six pounds of muscle a decade, usually replaced by fat [28]. It makes stairs and standing up harder and raises the risk of falls and fractures [27]. Researchers say the biggest health gain comes from starting from nothing, not from training harder [26].

    Why it matters — The people with most to gain are the ones currently doing none, and they are the hardest group for any trial to recruit and keep.

12. 13

    Cut off and feeling lonely came apart

    Social isolation means having few contacts and can be counted. Loneliness is the feeling of not having enough, and has to be asked about. In 354,537 UK Biobank participants followed for a median 11.6 years, 38,103 developed cancer. After adjustment, isolation carried an 8% higher risk and loneliness carried none. The size of the isolation effect also differed strongly between men and women [29].

    Why it matters — The two are routinely reported together as one risk factor, and here only one of them tracked the disease.

13. 14

    Five million people, one heart figure

    Six cohort studies covering 5,253,128 people, followed between four and 11.3 years, were pooled for heart disease and stroke. Together, isolation or loneliness carried a 17% higher risk. Split apart, people with a history of social isolation carried 39% higher risk, more than those reporting loneliness. The studies disagreed with one another a great deal [30]. A separate study of 8,046 Chinese adults aged 45 and over found loneliness alone carried 42.9% higher risk over seven years [63].

    Why it matters — Two careful analyses put the bigger number on different halves of the same pair. Isolation looked worse in one of them and loneliness looked worse in the other.

14. 15

    For asthma it ran the other way

    In UK Biobank participants followed a median 11.6 years, 9,453 developed asthma as adults. Here it was loneliness that carried the risk, 27% higher, and social isolation showed nothing once other factors were adjusted for. The researchers also ran a Mendelian randomisation analysis, which uses inherited genetic differences to test whether a link is likely to be causal. It gave suggestive genetic support for feeling lonely and for taking part in fewer social activities [33].

    Why it matters — Cancer pointed at isolation and asthma points at loneliness, so whatever these two questionnaires capture, they are not measuring one condition.

15. 16

    A dose pattern in 219,086 people

    UK Biobank participants were matched into two equal groups, obese and not, and followed a median 12.4 years. Among the obese half, 18,280 developed a mental disorder recorded in primary care or hospital records. Compared with the least lonely, the moderately lonely carried 37% higher risk and the loneliest 73% higher, rising steadily across the scale. The same pattern appeared for social isolation [31].

    Why it matters — The risk climbed step by step as loneliness rose. That pattern makes a cause more plausible, and it still does not prove one.

16. 17

    Isolation, frailty and broken bones

    Isolation and loneliness were each linked to thinning bone and to fractures in a large cohort, with hazard ratios of about 1.12 and 1.19 for osteoporosis and 1.11 and 1.18 for breaks. Inherited risk did not change the pattern. The researchers then tested what the link travelled through, and found frailty and body-wide inflammation carried part of it, anywhere from under 1% to about 37% of the total, depending on which pairing was measured [32].

    Why it matters — A range running from 1% to 37% is the honest answer here, and it is wide enough to mean the route is not yet known.

17. 18

    What isolation does inside a cell

    A review pulls together what happens in animals kept alone: disrupted signalling between nerve cells, a stress-hormone axis running differently, and changes in how the body handles fuel. It connects long-term isolation to metabolic syndrome and heart disease through oxidative stress, inflammation, faulty mitochondria and poor clearing-out of damaged cell parts [34]. A review of the human evidence, published separately, says the field still cannot say whether loneliness is a cause of poor health, a companion of it, or a result [35].

    Why it matters — The animal work has a mechanism and the human work has the numbers, and neither one can supply what the other is missing.

18. 19

    A cold wrist cut severe hot flushes by 46%

    Twenty-seven people took part: ten with breast cancer, twelve with prostate cancer and five postmenopausal women, all having at least two moderate-to-severe hot flushes a day. Each wore a wrist device cooled to 47F for two weeks and an identical sham that only blew air for two weeks, without knowing which was which. Severe episodes fell 46% against baseline and all daily flushes fell 18%. Nobody dropped out and no side effects were reported [36].

    Why it matters — Hormone-blocking cancer treatment produces the same flushes as menopause, and the drug options for both carry risks a cold wristband does not.

19. 20

    Hot flushes and the traffic to the blood vessels

    Sixty-nine postmenopausal women were split by whether they had ever had hot flushes or night sweats. Researchers pushed a fine electrode into a leg nerve to count the bursts of signal running out to the blood vessels, which is one of the main things setting blood pressure. Resting blood pressure was the same in both groups. The flush group fired more bursts per hundred heartbeats, 53 against 44, with a slower resting heart rate and more variation between beats [37].

    Why it matters — Hot flushes are handled as a comfort problem, and this measures a difference in the nerve traffic that helps set blood pressure.

20. 21

    It was the evening flushes that wrecked sleep

    Ninety-nine women recorded their hot flushes and their sleep every day for a week, separating daytime, evening and overnight episodes. Two different things turned out to matter. Within a woman, more severe flushes in the evening predicted a longer time to fall asleep that night, while more frequent flushes during the night predicted worse sleep quality. Daytime episodes predicted neither [38].

    Why it matters — A study counting only night sweats would miss the evening half of this, and severity and frequency are turning out to do different jobs.

21. 22

    What actually produces a hot flush

    Menopause is the permanent end of periods, defined as twelve months without one, and most often arrives between ages 45 and 56 [40]. A flush is not simply the hormone drop. As the ovaries run out of follicles, oestradiol and progesterone fall, and the brain's control network rewires around three signalling systems called kisspeptin, neurokinin B and GABA. That changes the pulsing of the hormone driving the cycle, and produces the flushes [39]. Inflammation, oxidative stress and failing mitochondria run alongside it [39].

    Why it matters — Naming that machinery is what makes a non-hormonal drug possible, because it supplies a target that is not oestrogen.

22. 23

    Long-term itch is mostly not histamine

    Antihistamines fail most people with long-term itch, and the reason is that histamine is one messenger among many. Mast cells, immune cells sitting beside nerve endings in the skin, also release enzymes, signalling proteins, fat-based messengers and nerve growth factors that make sensory nerves fire more easily. Those nerves answer back with their own chemicals, which switch the mast cells on again. A review proposes treating the two together as one unit that feeds itself [41].

    Why it matters — It explains a very common experience, that the cream and the tablet do nothing, as a property of the wiring rather than a failure to use them properly.

23. 24

    The nerve keeps the itch after the rash goes

    In eczema, a signalling protein called IL-31 lands on sensory nerves in the skin and turns up two channels that make those nerves fire more readily. How much of its receiver sits on the nerves tracks how badly patients rate their own itch. In affected skin that receiver stays high after the visible inflammation has settled, which researchers describe as a memory held in the nerve rather than in the skin [42].

    Why it matters — It supplies a mechanism for the itch that outlasts the rash, and a reason that clearing the rash can leave someone still scratching.

24. 25

    The brain sorts itch and pain into two circuits

    Gut pain and itch both feel bad and both resist treatment, and it was unclear whether the brain separates them. In mice, researchers found two different groups of cells in the basolateral amygdala, an almond-sized region that handles threat. One responded to gut pain, one to itch, and they ran to different destinations, the hypothalamus and the striatum. Two different genes were switched on in the two groups, and silencing each one damped only its own sensation [43].

    Why it matters — Separate circuits mean a drug could in principle quiet itch without touching pain, which the drugs now in use cannot do.

25. 26

    In older skin a touch registers as an itch

    Old mice scratched no more than young ones when left alone, but scratched far more when lightly touched, a state called alloknesis. Their urine carried more serotonin, and their sensory nerve bundles and spinal cords carried more of one serotonin receiver, 5-HT7. Blocking that receiver reduced the touch-triggered scratching [44]. Long-term itch with no cause ever identified is common in older people, and a systematic review says its mechanisms are still poorly understood, with several competing explanations [45].

    Why it matters — It points at one specific target in a complaint usually written off as dry skin and age.

26. 27

    Long COVID immune activity still running at six months

    142 people were followed from 2020: 35 never infected, 54 acutely ill, 24 recovered and 28 with long COVID. More than 180 days after infection the long COVID group still showed switched-on inflammatory pathways and signs of exhausted T-cells, compared with people who had recovered from the same virus. A second group recruited in 2023 and 2024 showed the same pattern [46]. A review of studies to August 2025 reports the same persistent activation, plus antibodies aimed at the body's own tissue [50].

    Why it matters — Two separate groups of patients three years apart landing on the same signature is what turns a hypothesis into a candidate test.

27. 28

    One cell type proposed as the centre

    A review argues that a single cell type explains most of long COVID: the macrophage, a large immune cell that swallows debris and raises the alarm. The case is that these cells get locked into an inflamed state by leftover viral protein, by reservoirs in the gut, and by chemical marks laid on their DNA that keep the inflammatory genes readable. The authors present it as a framework for testing rather than as an established finding [47].

    Why it matters — Naming one cell gives drug developers something to aim at, which the long list of competing causes does not.

28. 29

    Another proposal points at the blood vessels

    A second review puts the cause in the endothelium, the single layer of cells lining every blood vessel. The argument is that infection pushes those cells into senescence: alive, no longer dividing, and leaking inflammatory and clot-promoting chemicals. At the barrier around the brain, in the gut and in muscle, that could account for the symptoms, and a poorly working immune system never clears the senescent cells [48]. Long COVID and ME/CFS share many of these features, with the balance between them unsettled [51].

    Why it matters — Three reviews in the same month name three different central causes, which is what an unsettled field looks like from outside it.

29. 30

    400 million people and a marker from the first illness

    A 2024 estimate put 400 million people worldwide living with long COVID, at an annual economic cost near one trillion US dollars. Researchers are now working backwards from who develops it, testing antivirals and immune drugs [49]. In 161 patients assessed six to nine months on, those who had run higher peak temperatures and lower blood oxygen during the acute infection carried more of one inflammatory assembly, the NLRP3 inflammasome, and more of the chemicals it releases [60].

    Why it matters — A measurement taken during the first illness that predicts the later one is the closest thing this field has to a handle.

30. 31

    Hearing may fail before the cells die

    Hair cells are the sensory cells of the inner ear, and their death has long been treated as the cause of hearing loss. A review argues the order is often the other way round. The connections between hair cells and the hearing nerve, and the fine bristles on top of the cells, are typically lost first, while the cells themselves are still there but working badly. On that reading, hair cell loss is frequently a later consequence [52].

    Why it matters — It moves the moment worth intervening earlier, to a point when the cells are still alive to be helped.

31. 32

    A protein named for the ear's amplifier

    Outer hair cells are the ear's amplifier: they physically push back on the sound wave, which is what gives hearing its enormous range. For that to work their tallest bristles must stay attached to a gel-like sheet above them, through a structure nobody had fully identified. Researchers now name a protein, TMEM145, as its core element. Mice bred without it lost two known partner proteins from the bristles, the bundle came free of the sheet, and the animals were profoundly deaf with no amplification at all [53].

    Why it matters — It names one protein whose failure would look, in a clinic, like ordinary hearing loss with no cause found.

32. 33

    WHO counted the vaccine studies, then weighed them

    The World Health Organization reviewed everything published between 2010 and August 2025 on vaccines, the mercury-based preservative thiomersal, and autism. On thiomersal the count was tied: nine papers reported no link, nine reported one. Graded for design quality, all nine reporting a link came from one group of authors in the United States and all nine were rated very low strength of evidence and high risk of bias. On vaccines generally, 17 of 19 studies found no link, covering nearly 800,000 children in six countries [54].

    Why it matters — Counting papers gives a tie and weighing them does not, and that distance is where most of the public argument has been sitting.

33. 34

    The sums check out, when anyone can get the data

    Researchers took 296 systematic reviews of health and education work and tried to redo each one's first meta-analysis. Of the 121 that did not already show their working, only 22 sets of authors supplied data or code when asked. Where the material could be had, 104 of 121 came out the same and none of the differences would have changed how a reader read the result [55]. Before any of it runs, somebody chooses whether the studies measure one true value or several, and that changes both the answer and how precise it looks [64][65].

    Why it matters — The arithmetic held up almost every time, and the thing that failed was researchers answering an email.

34. 35

    Hearing a thing twice makes it feel truer

    Repeating a statement makes people rate it as more likely to be true. A meta-analysis pooled 182 studies and 366 effect sizes published between 1977 and 2025, covering 31,184 people, and corrected for the tendency of small studies to overstate. The effect survived at a small size and varied a lot between studies. What mattered most was the first exposure: the instructions people were given then, whether a truth cue was present, and how long they saw the statement [56].

    Why it matters — Repetition is the cheapest thing to buy online, and this is a measurement of what it buys.

35. 36

    Teaching people to leave the page

    Sixty-four controlled experiments tested four ways of teaching people to judge whether a source is credible. Overall the teaching worked, at a moderate size. Lateral reading, which means leaving the page you are on to check who wrote it elsewhere, had the largest effect. Results varied enormously between studies, from slightly harmful to strongly helpful. Exercises run on the open internet worked better than closed ones [57].

    Why it matters — The technique that helped most is the one that involves not staying where the claim is.

36. 37

    Old mice hold on to water they do not need

    Vasopressin is the hormone telling the kidneys to hold water back. In old mice, one cluster of brain cells that make it, in the supraoptic nucleus, grew larger and fired more readily, and its gene was among the most switched-on with age. Turning those cells on in young mice reproduced the picture of ageing: low body temperature, less energy burned, less drinking. Turning the gene down in old mice restored water balance [58]. In humans the proposed trigger is low-grade inflammation acting on the brain's salt sensors [59].

    Why it matters — It gathers several separate complaints of old age, feeling cold, drinking little and a stress axis stuck on, into one hormone running too high.

37. 38

    Heat is becoming a kidney problem

    Heat injures kidneys directly, and a clinical perspective argues that the kidney specialists have not yet written rising temperatures into their risk models or guidelines. The most exposed are people working in hot conditions, indoors as well as out, particularly in low-income countries, and repeated bouts of heat-related acute injury can progress into long-term kidney disease [61]. Hard or prolonged exercise loads the same system, cutting blood flow through the kidney and raising the very markers used to diagnose injury [66].

    Why it matters — A diagnostic marker that also rises from running makes heat injury in outdoor workers harder to count than it looks.

38. 39

    The eye drops whose mechanism is still argued over

    Low-dose atropine eye drops are among the main treatments used to slow a child's eye from lengthening further. A review of the field says the basics remain open: how the drug works is not established, the right dose is not settled, who will respond cannot be predicted, and eyes often speed up again once the drops stop. The authors call for markers to guide individual treatment and for combinations with other methods [6].

    Why it matters — The most widely used drug in this field is being prescribed to children while its mechanism is still under discussion.

39. 40

    The target moves to children who do not have it yet

    The field's 2025 consensus digest puts its weight on pre-myopia: children who are not short-sighted yet but have already used up too much of the long-sightedness they are born with. That remaining reserve is being treated as a predictive marker, with different thresholds for different populations. The digest states that delaying onset by a single year may be worth more across a lifetime than several years of slowing progression later. It also reports that some young adults with strong short sight keep worsening [7].

    Why it matters — It shifts the work from treating children who already have it to identifying the ones about to get it.

40. 41

    Fourfold difference between two school systems

    Short sight is not evenly spread. Among 2,595 children aged 4 to 18 in one large Chinese municipality, 56.5% were short-sighted and 51% had eyes long enough to count as visually impaired by length. Girls were affected more than boys from age nine, and having a short-sighted parent raised the odds [9]. Among 3,564 children aged 6 to 15 across 44 schools in Beirut, Lebanon, the figure was 14.2%, higher in the older group and in children from lower-income families [10].

    Why it matters — A fourfold gap between two school-age populations is the strongest single argument that this is about how children live, not only what they inherit.

41. 42

    Green space, screens and a marker in the eye

    Dutch child-health researchers describe how the Netherlands handles rising short sight through routine child health checks rather than a separate screen-time campaign [12]. A European editorial adds that green space around schools matters, because it pulls children outside and gives them things to look at at many distances [11]. The same editorial reports a marker being tested to measure sun exposure objectively: a pattern of fluorescence in the white of the eye, which was larger in children with less short sight [11].

    Why it matters — Everyone here agrees children should be outside more, and the argument is over whether to ask them, measure them, or rebuild the schoolyard.

42. 43

    The same light signal reaches mood and eye growth

    A review proposes what it calls a light-eye-brain axis. Cells in the retina that sense brightness rather than shape feed both the body clock and the circuits handling mood, and the same light signalling influences how the eye grows. The argument runs both ways: children who are low or anxious go outside less and get less light, while strong short sight may raise anxiety through shared nerve and immune pathways. The authors offer it as a framework for future work, not a finding [13].

    Why it matters — It is one light measurement doing duty in two fields that rarely read each other's papers.