This week I kept circling back to a small, uncomfortable thought: how much of our trust in medicine is really just an assumption that someone, somewhere, is checking. We walk past a shelf and take it on faith that a grown-up signed off on what’s there.
That faith took a couple of hits for me lately. The one that stuck was the toddler food aisle. I have watched people I love read a label in that section with the careful attention of someone who believes it matters, believes the words on the front mean a rule was met. It turns out the aisle is mostly guarded by our own good intentions and a company’s marketing team. Nobody is standing at the door.
What unsettles me is not that a product is imperfect. It is the quiet gap between how safe a thing feels and how safe it actually is, and how often that gap is where the real story lives. A shelf that looks vetted. A penalty that looks like justice until you learn its size. A study that looks like an answer until you count who was actually in it.
I read the week’s work looking for that gap. It was never hard to find.
Autoimmune & Inflammation
A 30-Minute Sauna Floods the Blood With Immune Cells, and the Heat Isn’t What Triggers Them
Finnish researchers put 51 people in a 73-degree sauna and watched every class of white blood cell surge, then found the surge had almost nothing to do with how hot anyone actually got.
Histamine Was Always the Wrong Target for the Eczema Itch
Eczema itch runs on cytokines, not mainly histamine, which is why the pills in the allergy aisle keep failing the trials, and a network meta-analysis just said so one more time.
Investigations
A Gene-Editing Trial Killed a Six-Year-Old. The Only Official Penalty Was a $3,500 Fine.
Xinhua Hospital paid about 24,000 yuan in September 2025. The death it was fined over stayed out of the medical literature for another ten months, and out of the Nature paper entirely.
The Phone-Radiation Study Everyone Called a Debunking Was Too Small to Find the Tumor
Japan and Korea spent two years ‘validating’ the NIH finding that cell-phone radiation gave rats heart tumors. They used one dose, one sex, and too few animals to detect it. Korea’s rats grew the tumor anyway.
Drugs & FDA
Rice Chemists Taught an Enzyme to Recycle Drug Synthesis’s Most Expensive Ingredient
The chemistry that builds drugs like Spinraza just got cheaper. The $125,000 price tag never had much to do with the chemistry.
The FDA Approved the First Cholesterol Pill of Its Kind Before Its Trial Could Show It Prevents Heart Attacks or Deaths
Lipfendra lowers LDL cholesterol by up to 60 percent. The 14,550-patient trial that would show whether that prevents heart attacks or deaths does not finish until November 2029, and Merck is charging roughly $3,800 a year in the meantime.
Cancer
A Week Instead of a Month: The Faster Prostate Radiation Was Only “As Safe” on the Coarser Ruler
PACE-C put five-fraction radiation against the four-week course in 1,208 men. On the old grading scale they matched. On the finer one the short course logged 17 percent bowel toxicity against 10 percent, and the trial still can’t say whether the shorter beam controls cancer as well.
The Same Cancer Mutation, Opposite Effects: What 581 Mouse Tumors Say About Who Gets Sick
Cambridge and Yale gave four strains of mice one identical dose of a carcinogen. The tumors converged on the same pathway but took four different genetic routes to get there, and today’s screening ignores the reason.
Brain & Aging
Same Diet, Same 15 Pounds: Only the Women Who Ate in an 8-Hour Window Got Sharper
Two groups of older women cut the same calories and lost the same 15 pounds over six months. The ones who packed eating into an 8-hour window still scored better on problem-solving, from a 47-person pilot that hasn’t been peer-reviewed.
Nutrition & Food
A Microwave Can Fry a Potato That Soaks Up Less Oil: That Isn’t the Same as a Healthier Fry
University of Illinois engineers made French fries that hold less oil by frying them from the inside. Nobody ran a taste test, nobody measured the acrylamide, and the pitch was aimed at food manufacturers.
No One Guards the Toddler Food Aisle: 81 Percent of Its Foods Are Ultra-Processed
Parents assume the baby aisle is vetted. It isn’t. No US rule makes toddler food meet a nutrient standard, and an Austin audit of 2,783 products marketed for 6-to-36-month-olds found 81 percent ultra-processed and nearly half failing a WHO benchmark.