Hey readers! Here's a twist most of us didn't see coming: the wearable that tells you your glucose is running high may soon also tell you why it matters by tracking ketones in real time. That shift, from checking a urine strip in a panic to watching a continuous trend, is the story we're leading with this week, along with a sweat-sensing ring, a hard look at CGM alarms, and a wave of research into what's happening inside your islet cells.
🩸 The dual-analyte moment arrives
5 questions on Abbott's glucose-ketone sensors walks through what changes when a sensor tracks glucose and ketones at the same time, and it's the clearest picture yet of where continuous ketone monitoring is headed. Abbott's Libre Duo already earned Europe's CE mark earlier in 2026 and is under FDA review in the U.S., with the company expecting approval before the end of 2026; Europe is planning adult and pediatric versions aimed at people at higher DKA risk. - MedTech Dive
The promise is real: diabetic ketoacidosis is, in Abbott's framing, the most common hyperglycemic emergency for people living with diabetes, and catching rising ketones early could keep people out of the hospital. But the endocrinologists in the piece are careful not to oversell it. Ketones are highly nonspecific, so a reading is closer to a check-engine light than a diagnosis.
"It's not the ketones; it's the insulin deficient state, which is really what we need to understand for set failure," Lal said.
That nuance is why education matters as much as the hardware. Breakthrough T1D helped publish an international Lancet consensus that includes a 3 mmol/L "urgent high" threshold, though the article notes the evidence for using ketones alone to diagnose DKA is low. The open questions, who should wear it, how often to update readings, and how to avoid alert fatigue, are exactly the ones the community needs settled before launch. If you're at higher DKA risk, this is the device to watch this fall.
💍 A ring that reads your sweat
Engineers create smart ring that tracks biomarkers in finger sweat is the research-side companion to the Libre Duo story, and it points at the same destination from a different angle. UC San Diego researchers built a wearable ring that reads six markers from finger sweat, including glucose and ketones, using an osmotic hydrogel to draw sweat in for an electrochemical sensor array. - Medical Design & Outsourcing
In tests with both type 1 and nondiabetic participants, the ring's results closely matched commercial CGMs and blood ketone meters, and its subject-specific calibration held stable for about two months. The reason to care is captured well by the lab's lead:
"For example, the ring's ability to track both glucose and ketone continuously and simultaneously would greatly benefit optimal insulin dosing for the management of diabetes," said Joseph Wang.
It's early, non-invasive, and years from your finger, but it signals that continuous glucose-plus-ketone sensing is a direction multiple teams are chasing, not a one-off product.
🔔 Do CGM alarms actually help in the hospital?
The ketone-sensor debate about alert design has a useful data point this week. Inpatient Continuous Glucose Monitor Alerts Show Mixed Results in Trial reports on the SIGNAL trial, a three-arm randomized study of 533 hospitalized adults with type 1 or type 2 diabetes in China. Threshold alerts nudged time-in-range up by 3.9% versus no alerts, clearing a commonly used 3% bar for a meaningful change; predictive alerts, layered on top, added nothing significant. - MedPage Today
"Unlike pharmacologic interventions, the effectiveness of alarms depends not only on the technology itself but also on the clinical systems through which alarms are delivered, interpreted, and acted upon," Faulds noted.
The suggested reasons, predictive algorithms not accounting for injected insulin, plus workflow constraints and alarm burden, are exactly the design traps the Libre Duo experts flagged. Simpler, actionable alerts beat clever-sounding ones. (Full text is also archived here.)
🧪 Reading the sensors we build
One more monitoring caveat worth your attention:
FairGlucose benchmark exposes hidden CGM accuracy gaps - Across 33 glucose-forecasting models, type 1 patients had 6 mg/dL higher two-hour prediction error than type 2 patients (p<0.001), a gap the authors say is baked into the prediction task itself, not any one model. - The Neural Feed
They call for subgroup-disaggregated reporting to become the default standard in clinical AI validation, a move that could reshape how CGM-based tools are evaluated before deployment.
If you have T1D and lean on predictive CGM features, this is a reminder that the algorithms tend to have a harder time with your data. Worth a healthy dose of skepticism toward glucose forecasts.
🔬 What's happening inside your islets
A cluster of research this week is trying to measure and protect beta-cell function better than we can today.
NIH funds new blood biomarkers to track islet dysfunction in type 1 diabetes covers a USD 1.92 million first-year NIH R01 grant to Battelle Pacific Northwest Laboratories, running through August 2028, to develop blood markers of islet dysfunction. The idea is that C-peptide only reflects insulin secretion, while intermediate prohormone-processing peptides from alpha and beta cells could give cell-type-specific signals and predict progression. - AllSci
Greenstone Biosciences Wins Breakthrough T1D Grant to Model Diabetes Using Patient-Derived iPSCs is a complementary bet: building 3D islet organoids from people with T1D to study how beta cells respond to inflammation and stress, paired with multi-omics to find therapeutic targets. - BioInformant
"Human pancreatic tissue is scarce, and animal models don't fully capture T1D's autoimmune process," said Alex Bashore, Scientist at Breakthrough T1D. "This islet model could become a key resource for T1D researchers."
A New Platform for the Development of Targeted Treatments for Type 1 Diabetes - ULB and Johns Hopkins engineered lipid nanoparticles that deliver genetic material to beta cells more than 148 times more efficiently than a Moderna formulation in the lab, and shifted STAT2 and PD-L1 signaling in human cells under inflammatory stress. - myScience
None of these are treatments you can use today, but together they aim at the thing continuous ketone data can only hint at: the underlying insulin-deficient state.
💊 Worth a quick look
Type 1 diabetes: Patient receives genetically modified insulin cells without immunosuppressants - Uppsala University Hospital reports a man in his 40s who received gene-edited donor islet cells engineered to evade immune detection, followed for 14 months and described in NEJM, with no immunosuppressive drugs needed. - Sweden Herald
GLP-1 RAs in Type 1 Diabetes: New Data Shows No Increased Risk of DKA or Pancreatitis - A real-world AACE 2026 analysis of over 7,000 adults found no DKA or pancreatitis signal; among 255 T1D GLP-1 users, zero admissions for either over a year, though the single-center, small-cohort design limits generalizability. - HCPLive
Australian PBAC declines teplizumab funding after TN-10 uncertainty - PBAC accepted that teplizumab delays stage 2 to stage 3 progression but declined PBS listing over screening-equity and long-term economic uncertainty. - AbangeLabs Editorial
Diamyd Medical Q1 25/26 report - The pivotal Phase 3 DIAGNODE-3-a trial of retogatein has completed patient screening, with an interim efficacy readout expected at the end of March 2026 after FDA alignment shortened primary follow-up from 24 to 15 months. - Inderes
A lot of this week points the same direction: better sensing (glucose plus ketones), smarter alerts, and better measures of what beta cells are actually doing. See you next week.