The Libre Duo. A platform prototype to move care from the hospital to the home?

Abbott has put glucose and ketones on the same wearable, and the launch coverage is all about DKA. Worth having if you’re on a pump, which is fine as far as it goes, but it misses the bit that makes this interesting. Because if you look at what Abbott has actually built across its other divisions, and what they’ve patented alongside it, the Libre Duo doesn’t really look like a ketone sensor. It looks like the first visible piece of an infrastructure designed to move chronic disease care out of outpatients and into the home. And the policy environment for that shift has just been locked in on both sides of the Atlantic for the next five years.

So what is it?

The Libre Duo is a dual Glucose and beta-hydroxybutyrate (ketones, to you, me and everyone else) sensor, which will be available in two forms:

A 15 day adult version and a 10 day version that can be used with kids aged 2 and over.

According to the press release, industry partners that are already lined up to use it include Tandem, Betabionics, CamDiab and Ypsomed.

It adheres to the familiar Libre design and application process, so allows existing users continuity if they choose to adopt it. The ketone alert thresholds Abbott has mentioned align with what those with type 1 diabetes should already be familiar with. Below 0.6 mmol/l is unremarkable, 1.5 and up warrants attention, while above 3.0mmol/l is considered serious. As with existing Libre products, data is delivered every minute. It does raise questions whether alarms and readings will raise concerns, as there tends to be a lack of understanding about ketone production in relation to starvation (e.g. overnight), higher fat diets, etc. There’s a question whether users are educated enough not to drown emergency units with calls about high ketones.

What’s perhaps more interesting is that Abbott appears to be following their own playbook on launching new sensors. Launch first in Europe, under the CE marking scheme, which we already know to be far less stringent than any FDA requirement (as I’ve discussed previously here).

Further to this, digging in to what Abbott has published recently on the patent front, the story appears to be much bigger than simply a ketone sensor.

How does it work?

Glucose sensing is the older, simpler half of the device that most people reading will be familiar with. Glucose oxidase on an electrode reacts with glucose, an electron transfer mediator carries the resulting electrons to the electrode, and you measure the current. Current scales with the glucose concentration. That’s the basic recipe behind every continuous glucose sensor since the original Medtronic and Abbott trials in the early 2000s and we have roughly 25 years of incremental refinement on the same idea.

Ketones use a related approach but with a more involved enzymatic cascade. Abbott’s published work, and the patent family covering the dual-analyte architecture (WO 2020/159956, granted in Europe as EP 3 917 396, with the membrane-architecture refinements published in 2024 as the divisional EP 4 509 046) describe it like this.

Beta-hydroxybutyrate dehydrogenase oxidises BHB and reduces NAD+ to NADH. A second enzyme, diaphorase, then reoxidises the NADH, with the electrons carried to the working electrode by an osmium-based mediator. The current at that electrode scales with BHB concentration.

Here we have two enzymes working in sequence on one electrode, both held in place by covalent bonding to a polymer matrix, with albumin in there as a stabiliser. That’s quite a bit more complicated than the basic glucose sensor.

What’s clear is that BHB sensing is much harder than glucose (and I’m sure SiBionics would agree, as they’re currently the only other provider with a continuous ketone monitoring system on the market).

This is primarily down to the lower concentrations and the ranges involved. If you consider glucose, it normally sits in a range that’s 4-8mmol/l (72-144mg/dl). BHB at its peak is around 5mmol/l (90mg/dl) and in normal circumstances, 1/10th of the normal levels.

In terms of technical difficulty, the range that the ketone sensor covers is typically wider than that of glucose sensors. Most glucose sensors run from ~40mg/dl to  400mg/dl, an approximately 10x measurement range. The Duo ketones run from 0-8 mmol/L, which is around 144x compared to the glucose at 10x, and the detectable concentration is lower, compounding the challenge.

The two sensing analytes live in adjacent spots on the same sensor filament, without interfering with one another, or with each other’s signals. That’s pretty smart engineering, given the size of the filament.

And of course, being a Libre, this is all factory calibrated.

So what?

Okay, we’ve described the detail of what it is and how it works, but really, so what?

Stepping away from the engineering, we can read the consensus paper on continuous ketone monitoring for a clinical perspective. To quote from it,

The increasing occurrence of euglycemic DKA, with glucose concentrations below established diagnostic thresholds, makes the availability and use of CKM technology an important addition to the diabetes management toolkit. CKM data could alert the user when the risk of acute DKA is high on sick days in addition to signalling that individuals might be predicted to be at greater overall risk of future DKA on the basis of the distribution and degree of ketone measures in daily life.

So the pitch here is straightforward. DKA is common, and when it does occur, expensive to treat and potentially deadly. A glucose-only CGM can’t see it coming reliably because glucose and ketones don’t always move together. If we can get an earlier warning, then lives can be made easier and healthcare costs reduced. US DKA costs run into the billions a year and DKA is a leading cause of death in under-24s with diabetes. These are big numbers and appear to present a real problem.

But why are these DKAs occuring? One argument (and the list of partners shown) is that device malfunction drives the need for sensing to avoid DKA when cannulas fail, pumps are occluded, etc.

There’s also a discussion about the use of SGLT2 inhibitors and whether the introduction of CKM makes these safer to use and would, perhaps, force manufacturers to relicense them for type 1, given the higher risk of euglycaemic DKA they generate.

That’s all well and good, but reviewing the literature, this doesn’t look like the main driver of DKA. What’s a far bigger player is socioeconomic status, especially when it comes to repeated DKA. And insulin omission appears to play a high part in that.

UK data shows recurrent DKA clustering in young people, those living with social deprivation, and those with very high HbA1c. The ADA consensus report has insulin omission behind more than two-thirds of admissions in some cohorts, with social circumstances and mental health as the strongest drivers. Inner-city studies add depression, alcohol and drug use, and homelessness to that picture.

So the sensor is brilliant for pump-user device-failure DKA, and that’s a real population worth protecting. It’s not very useful for socio-economic driven DKA that does most of the harm. It doesn’t change the game on DKA in aggregate. It changes it for the subset of users it reaches, which is roughly the opposite of the high-risk group.

That last bit is uncomfortable but worth saying. The natural market for a premium pump-linked CGM is the already-looping user cohort. The group already at lower DKA risk, although there is a subset of users within that who will benefit. Left to the market, will the sensor mostly land on people who probably aren’t at the highest risk?

That can change with how it’s funded and targeted. It won’t change on its own. The question is, how is Abbott seeking to address this with national funding bodies? Taking the NHS as an example, will it be bundled in place of the Libre3+ in the NHS HCL procurement model? Given the push to replace Libre2+ with Libre3+ in the UK, this seems likely.

You mentioned that the story might be bigger?

While we have a CE marked dual metabolite sensor ready to be distributed, the bigger story may be waiting in the wings.

Over the past few years, Abbott has issued a number of patents from the same people for multiple metabolites.

With that they have demonstrably developed a useable multi-metabolite sensor, with the manufacturing processes and factory calibration required that go with it.

Given this, the Libre Duo appears to be the open, public “first release” of their multi-metabolite platform rather than “just adding ketone sensing to CGM”. It’s worth taking a look at the other patents we’ve just mentioned to try and work out how that might play out.

So what’s actually in the patent portfolio? Same inventors, same multi-enzyme cascade architecture, same Abbott Diabetes Care assignee, all filed alongside the dual glucose-ketone work. They cover lactate, ethanol, creatinine, glutamate, and the amino acid pair asparagine + aspartate, plus a broader filing on multi-enzyme analyte sensors in general.

The first question to ask is what each of those would actually be used for clinically. The second, and probably more interesting, question is where they sit inside Abbott as a company. Because the more time you spend with this patent portfolio, the more it stops looks like something quite a bit bigger.

Where Abbott already plays

Abbott runs:

Abbott Diabetes Care: Libre, Lingo, AID partnerships with Tandem, Medtronic, Beta Bionics, Ypsomed and CamDiab.

Abbott Point of Care: the i-STAT business, which already sells lactate, creatinine and high-sensitivity troponin cartridges into emergency departments, ICUs and primary care.

Cardiovascular: Structural Heart (MitraClip, TriClip, Tendyne, Navitor TAVR), Electrophysiology (Volt PFA, AVEIR leadless pacemaker), and an implantable cardiac monitor business (Confirm Rx) that already does long-term remote monitoring with clinician dashboards.

Neuromodulation: deep brain stimulation for Parkinson’s, essential tremor and dystonia, spinal cord stimulation for chronic pain, and an active trial running DBS for treatment-resistant depression.

Cardiovascular at-home services: Acelis Connected Health, which provides at-home VAD monitoring and remote support today.

Cancer Diagnostics: newly created following the $21 billion acquisition of Exact Sciences in March 2026, which brought Cologuard, Oncotype DX, Oncodetect and Cancerguard under Abbott. Many of these are direct-to-home tests.

Software and integration: Libre Assist (launched January 2026), an in-app generative AI decision support tool, and a first-of-its-kind agreement with Epic to integrate Libre data directly into hospital electronic health records via Epic’s Aura software.

Once you have that list in front of you, the patent portfolio reads very differently.

Mapping the patents to the divisions

Acute and critical care. Lactate sits at the centre of sepsis management, cardiogenic shock and heart failure decompensation. Creatinine has a parallel story for acute kidney injury and renal monitoring after cardiac surgery. Abbott already sells both as i-STAT cartridges in hospital. A continuous wearable version of either slots directly into the same buyers, the same workflows, the same channel. The internal handoff isn’t something to engineer. It’s already there.

Chronic care and CKD. Creatinine for outpatient CKD progression tracking, which currently runs on eGFR from blood draws every few months. The i-STAT version is already in primary care. The continuous version is just the next step.

Cardiovascular. Two analytes — lactate and ketones — both have roles in heart failure that we don’t currently have wearable readouts for. Lactate for decompensation, ketones for cardiac fuel switching under stress. Abbott’s structural heart and electrophysiology businesses sell hardware to the same cardiologists who manage these patients, and the Confirm Rx implantable monitor business already does remote cardiac monitoring with home-to-clinician data flow. The clinical home is there.

Lifestyle and consumer. Ethanol fits Lingo. The dual glucose-ethanol framing is in the patent itself. Addiction medicine, liver disease and post-transplant compliance sit underneath, but the consumer wellness route is the simpler commercial path.

Neurology. Glutamate is harder to call. It’s the principal excitatory neurotransmitter and is implicated in epilepsy, stroke, TBI and some inborn errors of metabolism. Peripheral glutamate doesn’t directly read CNS glutamate, so the clinical case is less obvious. But Abbott runs a substantial neuromodulation business with DBS, SCS and the depression trial, so it has somewhere to land even if which clinical use comes first isn’t clear.

Oncology. The asparagine + aspartate filing (EP 4,271,825) lines up almost perfectly with L-asparaginase therapy monitoring in paediatric acute lymphoblastic leukaemia. For anyone unfamiliar: L-asparaginase is a backbone drug in paediatric ALL, working by depleting asparagine in the bloodstream, which leukaemia cells can’t synthesise themselves. Therapy currently requires intermittent blood draws to confirm depletion is happening, because the drug can be neutralised by anti-drug antibodies and you need to know whether to switch formulations. A continuous sensor would be a meaningful improvement on current practice. This was filed before Abbott had an oncology business and with their acquisition of Exact sciences, now sits inside one.

Every patent in the portfolio maps cleanly onto an Abbott division that already exists.

The bigger pattern

This is where it gets genuinely interesting, because once you list what Abbott is already doing in adjacent areas, the patent portfolio stops looking like a sensor roadmap and starts looking like one piece of something larger.

Look at what Abbott has built in the last couple of years. CGM data flowing automatically into Epic EHRs. A generative AI decision support tool in the Libre app. Long-term implantable cardiac monitors with home-to-clinician data flow. At-home VAD monitoring with clinical support behind it. Cologuard and Oncodetect doing direct-to-home cancer screening and recurrence monitoring. Patents on the comms layer that connects ketone sensors to AID systems. A patent on the decision support algorithm for what to do when the readings move.

Sensors. AI on the readings. Direct EHR integration. Existing at-home monitoring services. A salesforce in primary care that’s already the largest in the world for diabetes tech.

If you were going to build the infrastructure to move chronic disease management out of the outpatient clinic and into the home, with clinician alerting wired into the EHR they already use, this is essentially what you’d put together. The multi-metabolite sensor platform is the diagnostic layer it’s been missing. Diabetes Care has been quietly building the model. The Exact Sciences deal and the i-STAT and Acelis businesses suggest the other divisions have been running parallel.

Whether Abbott is explicitly framing it that way internally is, of course, unknown publicly. They haven’t said so. But the picture from the outside is reasonably clear. The sensors solve the bit that’s still missing, actual continuous chemistry data from outside the hospital, and slot into an infrastructure that’s mostly already built around them.

That’s a much more ambitious play than “we shipped a ketone sensor”. It’s a bet that the centre of gravity for chronic disease management is going to shift from in-person outpatient review to data flowing from home into the EHR, with clinicians intervening on signal rather than schedule. If they’re right about that direction, Libre Duo isn’t really a product launch. It’s the first time their full stack has been visible in one place.

What’s not in the portfolio?

It’s worth a quick note on the absences.

No cortisol, which is the obvious wellness analyte everyone in the continuous monitoring space talks about. Stress, burnout, athletic recovery, a clear consumer market, a small molecule that ought to fit the architecture, and Abbott already has the Lingo channel. Given they’ve gone after everything else, the absence is conspicuous. Either they can’t make it work, someone else has the IP locked up, or they don’t see the commercial case.

No troponin, BNP or the classical cardiac biomarkers on the wearable side. These are proteins rather than small molecules, and would need an antibody-based sensing architecture rather than enzymatic. The interesting wrinkle is that Abbott already sells the troponin assay as an i-STAT cartridge. They have it, they just can’t yet put it on a continuous filament. Cardiology is the one therapy area where the platform argument is currently partial, and closing that gap is a much harder engineering problem than another small-molecule cascade.

There are also no bilirubin and no specific drug-level monitoring patents (vancomycin or methotrexate would be the obvious ones, where therapeutic drug monitoring is currently a hassle). We assume these are either out of architectural reach, or not on the roadmap yet.

Which metabolites might be combined?

The interesting question is what gets paired with what on the same filament. The Duo gives us glucose and ketones because that pairing solves the DKA problem and fits the existing AID ecosystem. The filings suggest other combinations are clearly being considered.

Glucose + lactate would be the metabolic stress combination. Useful in heart failure, useful in sepsis, useful in critical care generally, and arguably the most clinically valuable pairing in the pipeline. Given Abbott Point of Care already sells the lactate cartridge and the cardiovascular business sells the structural heart hardware, this looks like the obvious next dual sensor.

Lactate + ketones would be a pure metabolic-fuel-monitoring play, mapping which fuel the body is burning at any given moment. Research market in metabolic physiology, sports performance, and probably heart failure where cardiac fuel switching is now active research.

Glucose + ethanol is explicitly named in the ethanol patent and is the obvious consumer/wellness pairing — adjacent to Lingo, with possible insurance and behavioural-change use cases.

Glucose + creatinine would be the diabetes-and-CKD pairing. Given diabetic kidney disease is one of the most common reasons people with diabetes end up in nephrology, this is the chronic-care combination with the biggest patient population behind it. It’s also the cleanest fit for the home-monitoring thesis — a single sensor covering two chronic conditions that currently each require separate outpatient pathways.

Past that, things get speculative. A critical care multi-analyte sensor with lactate, ketones and creatinine on the same filament would give an emergency department or virtual ward a real-time read on metabolic status, fuel use and renal function from a single stick-on. Whether the engineering scales to three or four analytes on a single filament is the constraint, and we don’t yet have public information on that.

What this actually tells us

What’s clear is that Abbott isn’t filing for vanity analytes. Every one of these has a real, sizeable clinical use case where intermittent blood draws are the current standard of care, continuous would clearly be better, and Abbott already has a division that sells to the relevant clinicians.

The strategic shape it suggests is that Libre Duo isn’t the destination. It’s the first published proof point of an infrastructure that spans most of what Abbott already sells, with a clear direction of travel: from outpatient care to home monitoring, with clinician alerting wired into the EHR they already use.

In the NHS, the government’s 10-year plan “Fit for the Future” sets a clinical hierarchy where care happens digitally by default, in the patient’s home where possible, and in a hospital only when absolutely necessary. Virtual ward capacity is being doubled to 40 beds per 100,000 GP-registered patients. There are already over 12,500 virtual ward beds in England, and NHS Shared Business Services is launching a standardised Virtual Wards and Hospital at Home procurement framework in 2026 to eliminate regional vendor fragmentation. NHS England has specifically targeted heart failure for virtual ward expansion.

In the US, the Acute Hospital Care at Home waiver was extended for five years in February 2026, through September 30, 2030, decoupling the programme from short-term continuing resolutions for the first time. 366 programmes across 139 health systems in 37 states are now approved to deliver acute hospital care at home under Medicare reimbursement.

Both of those policy moves were locked in within the last 18 months. Abbott’s patent filings on the multi-metabolite platform date back years. This isn’t reactive positioning. Abbott has been building toward this moment, and the funding architecture they need has just clicked into place around them.

So when we ask whether the Libre Duo matters, the honest answer is: a bit, on its own merits. The DKA argument is real for pump users but limited in aggregate. The bigger answer is that this is the visible edge of something considerably larger, and the ketone story is genuinely just the first one out of the door.

---

References

The various components of this article draw on the following references, if you require any further reading:

Gibb FW, Teoh WL, Graham J, Lockman KA. Risk of death following admission to a UK hospital with diabetic ketoacidosis: recurrent DKA concentrates in young, socially deprived adults with very high HbA1c. Diabetologia, 2016.

American Diabetes Association. Hyperglycemic Crises in Adults With Diabetes: A Consensus Report: social determinants, mental health, insulin omission and food insecurity as the strongest drivers of recurrent DKA. Diabetes Care, 2024.

Maldonado MR, et al. Recurrent diabetic ketoacidosis in inner-city minority patients: behavioural, socioeconomic and psychosocial factors: depression, substance use and homelessness in recurrent DKA, and DKA as a leading cause of death in under-24s. Diabetes Care, 2003.

Galindo RJ, et al. Global Burden of Diabetic Ketoacidosis: epidemiology, cost, and DKA triggered by device malfunction in pump users. Diabetes Technol Ther., 2025.

International consensus group. Continuous ketone monitoring for people with diabetes: international expert recommendations: no published natural history of ketone fluctuation, response algorithms still “proposed”. Lancet Diabetes Endocrinol., 2026.

Abbott. Abbott secures CE Mark for world’s first dual glucose-ketone sensing technology: Libre Duo specs, AID compatibility, European rollout. Press release, 27 May 2026.

Euglycaemic diabetic ketoacidosis with SGLT-2 inhibitors: mechanism, and the diagnostic blind spot for glucose-only monitoring. Clinical review, 2025.

Le Neveu F, Barlow Y, Harvey JN. Euglycaemic ketoacidosis in patients with and without diabetes: alcoholic, starvation, pregnancy and anorexia-related ketoacidosis, and the value of bedside beta-hydroxybutyrate. Practical Diabetes, 2013.

Ketone body induction: insights into metabolic disease management: physiological versus pathological ketosis thresholds, and GLP-1 receptor agonist–associated ketosis. 2025.

Soni S, Skow RJ, Foulkes S, Haykowsky MJ, Dyck JRB. Therapeutic potential of ketone bodies on exercise intolerance in heart failure: ketones as cardiac fuel and signalling molecule, and the ketone-ester trials. Cardiovasc Res., 2025.

JMIR Diabetes. Continuous ketone monitoring: data from a randomised controlled trial: apparent sensor drift in BHB readings over 14 days of wear. 2026.

Lawrence E. Abbott receives CE mark for dual glucose-ketone sensor: 15-day adult version and 10-day version for children aged 2+. MedTech Dive, 28 May 2026.

Abbott Diabetes Care. Continuous Dual Glucose-Ketone Sensing Technology: Factory Calibration for the Ketone Sensor Component: two adjacent working electrodes, precise dispensing for the separated chemistry matrices, 0–8 mM BHB range. Diabetes Technol Ther., 2025.

Feldman BJ, Kiaie N, Kumar A, Love MR, Sloan MK (Abbott Diabetes Care). Analyte Sensors and Sensing Methods Featuring Dual Detection of Glucose and Ketones: two adjacent working electrodes with separate ketones-responsive and glucose-responsive active areas, an enzyme system of at least two enzymes acting in concert on the ketone electrode (BHB dehydrogenase + diaphorase cascade, NAD+/NADH cycling, osmium electron transfer agent, polymer-bonded enzymes with albumin stabiliser), and a single membrane with compositionally differentiated portions overcoating each active area. WO 2020/159956 / US 20200237275; granted in Europe as EP 3 917 396 B1 (granted September 2024), with the membrane-differentiation architecture covered in greater detail in the 2024 divisional EP 4 509 046 A3 (filed September 2024, A3 search-report publication March 2025; notice of intention to grant despatched April 2026).

Bansal S, et al. Continuous ketone monitor patent landscape 2026: clinically relevant concentration ranges, baseline BHB ~0.1 mM. PatSnap review, April 2026.

Williams JN (Abbott Diabetes Care). Systems, Devices, and Methods for Applications for Communication with Ketone Sensors: the communications and applications layer for ketone sensors, relevant to AID interoperability. EP 4 401 633 A1, published 24 July 2024 (PCT/US2022/043466; WO 2023/043797).

Abbott Diabetes Care pipeline analyte sensor patents (same multi-enzyme cascade architecture as the dual glucose-ketone family): lactate, US 12,076,145 “Lactate sensors and associated methods” (2018); ethanol, EP 4 552 565 A1 “Analyte sensors and sensing methods for dual detection of glucose and ethanol” (2019); creatinine, EP 4 552 566 A1 “Analyte sensors and sensing methods for detecting creatinine” (2019); glutamate, EP 4 267 752 A1 “Analyte sensors for sensing glutamate” (2020); asparagine and aspartate, EP 4 271 825 A1 “Analyte sensors for detecting asparagine and aspartate” (2020); and a broader filing on multi-enzyme analyte sensors, EP 4 663 121 A1 “Analyte sensors employing multiple enzymes” (2019). Identified through the cited-families analysis of the EP 4 509 046 A3 patent family.