Read the Endpoint, Check the Sponsor

Mandrola uses the Danish-led POET II trial as a lesson in critical appraisal. His point: a trial's primary endpoint should be its most important outcome. In POET II, he argues, the outcome that matters most was listed as a secondary endpoint.

The trial

Bacterial endocarditis is an infection of the heart valves. It can destroy a valve, or keep seeding the bloodstream with bacteria from a mass (vegetation) on the valve until the patient develops sepsis. The earlier POET I trial showed that patients could switch from IV to oral antibiotics once they met "POET criteria," signs that the infection was resolving, such as falling white count and inflammatory markers and no fever.

POET II enrolled about 500 patients with left-sided endocarditis. All had already received the standard 2 to 4 weeks of antibiotics and met POET criteria for stabilization. Half were randomized to stop antibiotics ("response-tailored" therapy) and half continued for the standard duration.

• Primary endpoint (superiority): days alive without antibiotics over 6 months. Result: 183 vs 169 days, a statistically significant gain of about two weeks for the tailored arm.
• Primary safety endpoint (non-inferiority): death, unplanned heart surgery or embolic events. Result: 8.2% vs 10.7% (risk difference −2.4 points; 95% CI −7.7 to 2.7; P<0.001 for non-inferiority).

This is a newsletter roundup, not a full report. The cancer-vaccine story in the headline is a short teaser that links out to the full article. Below is what the newsletter itself actually says.

Nobel Prize for light-switched neurons

Karl Deisseroth (HHMI/Stanford), Peter Hegemann (Humboldt University) and Georg Nagel (University of Würzburg) won the 2026 Nobel Prize in Physiology or Medicine "for their discoveries concerning light-gated ion channels and optogenetics." The method genetically modifies neurons so they carry ion channels that open in response to light. Researchers can then switch individual cells on or off in a living brain with light. That lets them test how specific nerve cells shape memory, emotion and behavior. The technique is also being studied as a disease treatment.

Cancer vaccines: one win, one stop

In August, two companies reported that an mRNA vaccine combined with an immunotherapy drug substantially cut the risk of melanoma coming back. Days later, a German drugmaker ended its trial of a colorectal cancer vaccine because early data showed it wasn't working. The newsletter says experts are trying to work out why one succeeded and the other failed, and what that suggests about which cancers vaccines can treat. It gives no effect sizes, company names or proposed explanations. Those are in the linked story.

Other items

• Stem cells and the FDA. STAT's Margaret Manto reports that stem-cell advocates see the FDA's "tentative embrace of peptides" under Robert F. Kennedy Jr. as an opening to move stem cells out of the grey market. STAT notes that the data don't support claims that stem cells are a revolutionary treatment.

This is a short recap of a radio segment, not a write-up of a study. Paula Span writes "The New Old Age" column for The New York Times and KFF Health News. On Sept. 30 she appeared on WAMU's Health Hub to discuss polypharmacy, the build-up of prescriptions as people age, and deprescribing, which means reviewing drugs and stopping the ones a patient no longer needs.

The claim and the number

The main figure comes from K. Eric De Jonge, director of geriatrics at MedStar Washington Hospital Center: "Once you get to six or seven prescription medications, you get a nearly 100% chance of some sort of side effect or intolerance."

The article doesn't name a study behind that figure. It also doesn't say what counts as a "side effect or intolerance" or which patients it applies to. Read it as a doctor's rule of thumb, not a measured risk.

The mechanism the article describes is simple. A drug that made sense when it was prescribed stays on the list. Later prescriptions are added, and the drugs interact to cause side effects.

Why drugs linger

Span argues that research moves faster than prescribing habits. "Doctors prescribe medicine based on the evidence available at the time. But science doesn't stop," she said. "Your doctor might not be up on the latest research. And patients get into a routine. They may not question if they still need the drugs they've been taking."

Her examples come from her Sept. 17 column. It names three types of drugs she says some older adults overuse: benzodiazepines, antibiotics and daily aspirin. According to the column's summary, not all older people need them, and taking them "can carry increased risks."

The Karolinska Institutet awarded the 2026 Nobel Prize in Physiology or Medicine to Karl Deisseroth (55, Stanford and Howard Hughes Medical Institute), Peter Hegemann (72, Humboldt University Berlin) and Georg Nagel (73, University of Würzburg) "for their discoveries concerning light-gated ion channels and optogenetics." The three will share 12 million Swedish kronor, about $1.2 million.

How it works

The work started in the early 1990s at the Max Planck Institute. Hegemann wanted to know how the single-celled green alga Chlamydomonas swims toward light. With Nagel, he found the answer: channelrhodopsin, an ion channel in the cell membrane that opens quickly when light hits it. Ions then flow into the cell. The pair showed that the protein also worked when they put it into other cell types.

Deisseroth then put the channelrhodopsin gene into rat nerve cells. In those genetically modified neurons, a pulse of blue light opens the channel and makes the neuron fire. Because only the engineered cells respond, researchers can turn specific cells on or off in a living brain. That lets them work out which cells drive which functions and how circuits connect. Deisseroth says light in medicine has mostly been used to observe things. Here it is used "to cause things to happen," with "millisecond precision and the cellular resolution that is essential to brain function." The article says the technique is being studied as a disease treatment but gives no clinical details or results.

Who was left out

Nobel rules allow at most three laureates per prize, and the choice drew criticism. On social media, many people argued that should have been included. He was first author of Deisseroth's landmark 2005 paper and a PhD student in his lab at the time, and is now at MIT. Boyden did not respond to STAT.