Drifting Cells and Dropped Patients

Jeremy Faust, an emergency physician who writes Inside Medicine, is giving a lecture at the American College of Emergency Physicians annual meeting. It covers the most important infectious-disease research published since last year's meeting, and this is his third year giving it. Before the talk, he posted a video version for general readers that walks through seven recent papers on sepsis and related problems such as pneumonia. He says he slows down in the video to explain background that his fellow doctors already know.

One caveat up front: the written post lists the questions the papers address. It gives no effect sizes, trial designs, sample sizes or conclusions. Those are only in the video.

The questions covered

• Real infection or contamination in children's blood cultures. When a child's blood culture grows bacteria, doctors have to decide whether it is a true bloodstream infection or a harmless contaminant. Faust notes that this call decides whether a child stays in hospital for days on antibiotics or goes home safely.
• Vasopressors or IV fluids for septic shock. He covers which works better: drugs that raise blood pressure or intravenous fluids. A second study compares types of fluids.
• Steroids in sepsis. Whether they improve outcomes, in both children and adults.

Eric Topol reviews two papers: one published in Nature by Vadim Gladyshev's Harvard group, and an earlier Cell paper from Juan Carlos Izpisua Belmonte's team at Altos Labs. Together they propose that aging is driven partly by cells losing their identity, alongside the older model of accumulated damage. How much each process contributes, and how the two interact, is unknown.

How identity is held, and how it slips

Every cell carries roughly the same DNA. What makes a cell a liver cell or a neuron is its 3D epigenetic structure: methylation, histones and chromatin. The paper describes three layers of control. A responds to acute stress within minutes to hours through transcription factors such as AP-1. An handles state changes over days to weeks, such as activating a cell for wound healing and then returning it to baseline. A made of chromatin architecture locks identity in place.

In 2016 the American Academy of Pediatrics (AAP) told doctors to offer medication for opioid use disorder to adolescents and young adults. A study published Friday in JAMA Network Open checks how that has gone nationally. More young people now start medication, but only slightly more. The share who stay on it for six months was already very low, and it may have fallen.

The care cascade

Scott Hadland of Mass General Brigham for Children and colleagues analyzed Medicaid enrollment and claims data. They looked at people aged 13 to 25 who were diagnosed with opioid use disorder between 2016 and 2023, nearly 230,000 in all. Each step lost more patients:

• About half started treatment by seeing a clinician.
• About a third had at least two visits in the first month.
• Of those engaged in care, about 1 in 6 received any medication.
• Just 1 in 32 stayed on medication for six months.

Black youth and young people from other minoritized groups dropped out more at each stage than white youth. Methadone was close to unavailable for minors. Fewer than 10 received it over the whole eight years, too few for the researchers to publish the exact figure. The article doesn't give the exact size of the rise in starts or the fall in retention.

The data cover Medicaid claims only. They show what was billed, not why patients left care. That means they can't separate patient choice from failures in the system.

Why minors fall through

Until late 2024, federal law required minors to have two documented "failed" attempts at recovery without medication before they could start methadone. The rule is gone, but many methadone clinics still won't take minors. Buprenorphine is easier to prescribe, yet few residential treatment programs for adolescents offer it. Psychiatrist Sivabalaji Kaliamurthy says that at 18, "a whole bigger field of people are open to helping you."

KFF Health News reporter Jordan Rau uses one family's lawsuit to show a problem that the statistics only hint at. Facilities can pressure families to consent to sedating drugs for relatives with dementia, and they can evict residents whose families say no.

The evidence on prescribing

More than 5 million Medicare beneficiaries have a dementia diagnosis. A JAMA study estimated that 1 in 4 are potentially inappropriately prescribed antipsychotics or other psychotropics. These drugs can cause delirium, falls and hospitalizations. The FDA requires a warning that antipsychotics raise the risk of death in older people with dementia.

In 2012, CMS launched its Partnership to Improve Dementia Care. It has reduced antipsychotic use in nursing homes, but in 2025 the drugs were still given to 1 in 7 residents who did not have schizophrenia. In March, an inspector general report flagged "alarming instances" of inappropriate use. Assisted living looks worse. States regulate it, and nearly half of its roughly 1 million residents have dementia. A 2023 JAMDA study found that residents with dementia were on an antipsychotic for 13% of the time they lived there. Researchers who tracked prescribing through 2017 found the federal campaign "did not appear to affect" prescribing in these facilities.

The aggression has a physical basis. Dementia damages the brain regions that control impulses and perception, so patients may lash out from fear or because they can't say they are in pain or hungry. University of Michigan geriatric psychiatrist Lauren Gerlach says non-drug approaches usually work better, such as checking for underlying illness and avoiding known triggers. Some severe cases still need medication, she says.

A review in Experimental Gerontology argues that GLP-1 receptor agonists, the drug class that includes semaglutide, should be studied as gerotherapeutics: drugs that slow aging itself, not just treat obesity and diabetes. Reason, who writes Fight Aging!, thinks this is mostly the drug industry doing what it always does with a blockbuster, which is look for new uses. He says the underlying evidence is weaker than the framing suggests.

What the review cites

The review's authors say GLP-1 drugs cut major cardiovascular events, all-cause mortality and systemic inflammation in randomized trials by more than their effects on blood sugar and weight alone would predict. The main numbers:

• SELECT: a 19% reduction in all-cause mortality (hazard ratio 0.81) in people with obesity but no diabetes.
• FLOW: a 24% reduction in the primary composite kidney endpoint (HR 0.76).
• Epigenetic clocks: a 2025 study was the first randomized evidence that a GLP-1 drug changes validated DNA-methylation clocks. Over 32 weeks of semaglutide, it reported significant slowing of DunedinPACE, PCGrimAge and PhenoAge. The summary doesn't give effect sizes.

The review proposes three overlapping mechanisms:

• the GLP-1 receptor in the hypothalamus
• AMPK/SIRT1 signaling, a cellular energy-sensing pathway
• more short-chain fatty acids produced by gut bacteria

The authors conclude that the class is a candidate geroprotector, not a proven one. Trials still need to: