Cortisol has been called the body's master stress hormone, regulating metabolism, blood pressure, immune response, and sleep, yet clinicians have never been able to track it continuously.
A single blood draw captures one moment. A 24-hour urine test averages an entire day. Neither reveals the dynamic rhythm that defines how cortisol actually works in the body.
Stanford University spin out Adaptyx Biosciences, a company that has invested 18 years into molecular biosensing research, recently presented the first continuous, multi-day measurement of free cortisol in humans at the American Diabetes Association's 86th Scientific Sessions.
Adaptyx’ wearable sensor is the first to measure cortisol from dermal interstitial fluid, the fluid just beneath the skin, in two first-in-human studies.
The Menlo Park, California-based company demonstrated a wearable sensor that tracks cortisol from dermal interstitial fluid, the fluid just beneath the skin, using programmable DNA-based molecular switches that reversibly bind to the hormone and generate a continuous electrical signal.
The breakthrough addresses an immense technical challenge: cortisol concentrations in the body are more than one million times lower than glucose, the target of today's CGMs.
Unlike CGMs, which rely on enzymes evolved over billions of years, Adaptyx's platform uses synthetic DNA structures designed to detect specific molecules at physiologically relevant concentrations, maintain accuracy over multiple days, and operate without sample preparation in real biological fluid.
In two first-in-human studies, Adaptyx's sensor tracked cortisol during a controlled oral hydrocortisone challenge—showing concordance with paired venous blood samples analyzed by liquid chromatography–tandem mass spectrometry (LC-MS/MS)—and captured the body's natural cortisol rhythm overnight, including the cortisol awakening response and overnight nadir that static tests routinely miss.
Adaptyx has collected more than 400 hours of in-body monitoring data through its IRB-approved human program and is pursuing FDA submission as a Class II cortisol monitor.
The company, which was also named winner of the 2026 American Diabetes Association Innovation Challenge, is initially targeting two clinical groupings: adrenal diseases such as Cushing's syndrome and adrenal insufficiency, where cortisol levels are the diagnostic focus; and cardiometabolic conditions including difficult-to-control type 2 diabetes, treatment-resistant hypertension, metabolic dysfunction-associated steatotic liver disease (MASLD), and polyendocrine metabolic ovarian syndrome (PMOS), where cortisol dysregulation shapes disease progression and treatment response.
Beyond these initial applications, continuous cortisol monitoring could provide objective biological context for psychiatric conditions including depression, anxiety, and PTSD; track stress physiology in critical care settings; accelerate pharmaceutical development by measuring hormone response within hours of dosing; and enable consumer applications for managing stress, sleep, metabolism, and athletic performance.
Founded on more than 18 years of continuous molecular biosensing research from the laboratory of Stanford professor and Chan Zuckerberg Biohub investigator H. Tom Soh, Ph.D., as well as technology from the laboratory of Stanford professor and National Medal of Technology and Innovation recipient Joseph M. DeSimone, Ph.D., Adaptyx holds an exclusive license to nine Stanford patents and has filed 18 additional patents since founding.
The company has raised $23 million in seed financing to date. Adaptyx co-founder and CEO Vijit Sabnis, Ph.D., connected with
MD+DI
to explain how the technology works, what the first human studies revealed, and how continuous cortisol monitoring could transform clinical care and consumer health.
Can you explain in depth how your sensor works? What are these "DNA-based molecular switches," and how do they detect cortisol continuously instead of just once like a traditional blood test?
Sabnis:
The technology comes out of a professor's lab at Stanford University. He is a cofounder of our company, and before I jump into the technology behind it, to put things into context—CGMs utilize an enzyme called glucose oxidase to convert the binding of glucose into a measurable electrical signal. It works beautifully well because glucose oxidase is an enzyme that has been evolved in nature for billions of years. The problem with enzymatic-based biosensing, like CGMs, is that it's entirely dependent on the enzyme. We do not have enzymes for every molecule that we want to measure continuously in the body. Companies like Abbott and Dexcom have continuous ketone monitors, which use an engineered enzyme system—and it took them a very long time to engineer that—and there are also continuous lactate monitors, but that is not a technology that allows measurement of a large range of other molecules. So that technology is limited in what it can do. A different approach is needed to measure other molecules continuously.
What Professor Tom Soh invented at Stanford is what we are calling programmable molecular switches, which are individual strands of DNA that are designed to bind to a particular aptamer molecule of interest. Aptamers have been around for a while, but Tom's breakthrough was recognizing that we could find an aptamer for a target of interest and then conjugate other DNA structures attached to that aptamer to create a switch that binds to a molecule of interest, allows us to read out that binding electrically or optically, but that also lets go. What is remarkable is that the switch is reversible.
Take a cortisol molecule, for example—one will float by, the switch will latch on to that cortisol, but it will then let go. Two things that Tom really pioneered are that he can make these switches operate across relevant physiological ranges in our bodies. So for cortisol, for example, concentrations are very small. Cortisol also changes at a certain rate in our body, so you have to make sure that that switch is letting go at a rate where we can track these changes with this reversible molecular switch. That is the real breakthrough here. We designed these aptamers for a target of interest, we then did molecular engineering on these aptamers, and we created this reversible switch that allows us to do continuous sensing.
This is very different from a blood test. With a blood test, you will bind to a target of interest, you will measure that binding, and then you are done and you have to throw it away because that binding is permanent. That change in the signal is a permanent, one-time change. So that is how we measure hormones or cholesterol or whatnot. There are technologies out there, but they are singular time-point measurements. It is the reversibility here that is unique. Also, oftentimes with molecular diagnostic assays, you have to prep the sample. For example, you will get blood, you will spin down that blood, you will then flow your sample over the assay and put some other agents in there to then prepare the assay to function, so that sample preparation is needed. In many, many different types of blood tests, for there to be a measurement, sample preparation is needed. What Tom did with this technology was create a way where no sample preparation is needed. You need that for continuous sensing. It has to work in that real, biological fluid. This is available with continuous glucose monitoring, but now with what we are doing, it's available on a platform that enables small molecules, electrolytes, and proteins. It is a very general technology.
Cortisol levels are incredibly tiny—a million times lower than glucose. What was the biggest challenge in building a sensor that could detect something so small and keep working accurately for days on someone's skin?
Sabnis:
It's a number of things. It goes back to the chemistry piece where you have to get that molecular switch to do three things. Does it work over the right concentration range? Does it switch on and off at a rate that is consistent with the changes inside your body? Can you make that switch very specific to cortisol? Oftentimes, these switches respond to other molecules that are present in the biological fluid, and particularly in hormone monitoring where many of the hormones and the metabolites of those hormones, are very structurally similar. One of the things we had to do in developing this switch was make sure, for example, that our cortisol switch does not also react with other molecules that are very structurally similar. That is extremely difficult to do, and it is one of the reasons why hormone monitoring is incredibly difficult. There are hormones, hormone pathways, metabolites of those hormones, and they often go back and forth between each other based on your body’s processes. Enzymes typically catalyze these conversions in different directions, and you have to make sure that you are measuring the right molecule. That was one of the things that was really difficult to do.
You tested the sensor in two ways: giving people a cortisol pill and comparing the sensor to blood tests, and then tracking people's natural cortisol overnight. What did you learn from each test, and how accurate was the sensor?
Sabnis:
The first test allowed us to do a lot of paired blood testing. What you will see is the blood level rises before you see the change in interstitial fluid concentration. It is very similar to how a glucose monitor works as well. We wanted to make sure, physiologically, that the time lag was consistent and that cortisol would be transferred from blood into interstitial fluid quickly, and at the same concentration we saw in blood. We were able to validate that, and we were able to measure blood and take individual blood points as a function of time. We found that the time lag was consistent across that entire duration of monitoring. When you time shift those curves, they lined up really well.
These sensors were getting about 19% MARD. Today's state-of-the-art CGMs are getting about 9% MARD. So we are actually within a good striking distance of making this commercially viable. We are working with the FDA to understand what MARD levels we need to actually validate, as we think about different commercial and clinical applications that we want to go into.
In the second set of experiments, which we are now doing pretty routinely, we are wearing our sensor overnight or over multiple days. We do some blood testing, but it is not easy to do blood testing on someone who is sleeping at home. The goal is not necessarily, initially, to do heavy validation with blood testing, but to really understand sensor drift, sensor longevity, and whether we are seeing the classic textbook cortisol curves.
Cortisol has a particular rhythm, rising in the morning and going up periodically throughout the day depending on meals, stress, and exercise. With things like blue lights and late nights, you won't see those cortisol levels go down as rapidly as they ought to. We are really looking at sensor longevity and whether we are able to maintain a good cortisol curve over multiple days. And we are looking at whether drift is occurring—where you would expect to see that sensor performance get worse over time. We wanted to test that in our sensor. So those were the main types of studies we were undertaking in that second set.
Now we are doing all sorts of really exciting things, for example, looking at our cortisol data in conjunction with conventional consumer wearables that are tracking things like sleep, heart rate, and HRV, and looking at the combination of that data together and what new insights we can gather. Today's consumer wearables are really measuring things related to the PPG signal, which is really far downstream of cortisol.
You mentioned in a press release that cortisol problems can make diabetes and high blood pressure harder to treat. How would tracking cortisol continuously help doctors figure out the right treatment for those patients?
Sabnis:
The way this will work initially is that cortisol as a regulatory hormone impacts every organ system in the body. The way we think this is going to evolve clinically is that cortisol dysregulation exacerbates a wide range of conditions. Continuous cortisol monitoring will enable us to identify patients who are not initially responding to conventional treatments.
For example, in diabetes management, there is a classification of diabetics that are called "difficult to control" diabetics. In those patients where they are adhering to their medications and following protocols—and their A1c is not improving—something else is going on. Physicians can then use our device and look at that cortisol curve and immediately understand what's happening. This is going to be 100x faster and better than conventional cortisol diagnostics. They can look at that curve and the interpretation we are providing to them and say, hey look, this person's cortisol levels are dysregulated. They also have diabetes that is difficult to control. Now we can take a look at normalizing those cortisol levels by prescribing a drug, pairing that drug with other lifestyle modifications, and improving their cortisol rhythms. Then you can start to understand the impact of those improved cortisol rhythms through lifestyle interventions and determine if their diabetes is getting better.
What's your timeline for FDA approval? When could doctors and patients actually start using this sensor?
Sabnis:
What I can share is that our device is actively under development. It is in human studies every single week and with another round of financing, we hope to accelerate that and push it towards FDA submission. We are in active discussions with FDA on initial use cases, and what those claims might be.
Companies like Dexcom and Abbott are leaders in continuous glucose monitoring. Do you see them as potential partners, or are you planning to go it alone?
Sabnis:
We would love to partner with large, successful companies. We are seeing a convergence between clinical and consumer care where there is a lot more appetite for data. On the consumer side, there is a lot more interest in longitudinal data collection from clinicians, and we do see Abbott and Dexcom as moving into consumer care and wanting to get into that data aggregation—not just clinical experiences, but consumer experiences. We are going to play a role in both clinical care as well as consumer care. When I think five and 10 years out, we don't know what the world is going to be, or what platforms are going to be available. So I think there is potential for us to partner with them.
We think there are going to be large platforms that are handling both consumer and clinical care. Abbott and Dexcom may be some of those companies, and we may want to partner with them to help give their patient populations additional continuous data. That data is really delivering more health context into the entire ecosystem. Ultimately, what each of these companies is going to build are AI agents that are ingesting a lot of data and trying to give each of us better clinical and consumer health advice. Those platforms are being built today. We have Hims, Hers, and Ro, and Abbott and Dexcom are playing a role in that. We see Adaptyx as growing into that ecosystem with not just cortisol, but eventually a wide range of continuous markers that are going to be available. That is going to be an additional data layer that we want to move into that ecosystem. I think there will be lots of partnership opportunities, with some focused more clinically and others more on the consumer side. There will also be a large number of companies playing in that crossover space.
Is there anything else you would like to expand on?
Sabnis:
One thing we haven't really covered is that this is a platform technology. Cortisol is just the first marker. We are aiming to bring other hormones into our sensors, so multi-analyte continuous sensing has been our mission from day one.
CGMs have been incredibly valuable for the world, and for diabetes management. That is a disease where looking at one marker was a game changer for managing that health condition. But the body doesn't run off of one marker—it runs off of thousands. Our mission with Adaptyx is to build a multi-analyte sensor that can do multi-analyte continuous sensing. Our R&D sensors are measuring 16 biomarkers continuously. Cortisol is the most mature biomarker, so we are commercially focused on that one initially, but we are constantly developing and testing a wide range of other molecules—other hormone markers, other small molecule markers—and the opportunity to look at multiple biomarkers continuously, look at their relationships with each other, and look at that continuous data in the greater context of blood biomarkers that you measure.
The context here is incredibly rich, and we want to be a data layer that makes the overall health context much richer and more powerful. It is not hyperbole to say that this type of data is way more valuable today, as it gets integrated into an AI-driven health management layer, compared to what it would have been a decade ago. That is why we started Adaptyx. We are building our entire roadmap around continuous chemistry, trying to be careful about how we build around cortisol—looking at cortisol first and then adding other hormones around that using adjacent opportunities. We will soon have testosterone on our patch in human studies as well. We can measure cortisol and understand how it impacts testosterone, and that is really valuable in men's health, and for women who have PCOS.
It also plays a role in perimenopause and the treatment of that. So you can start to see, as we add other markers, the applications we are enabling and the overall context we are delivering into the system will help our products become richer, and hopefully even more valuable to the patient and/or consumer. That is how we are thinking about building the company over time. We have lots of other hormones, other deep clinical applications to enable, and that is the world we want to build. We want to build the continuous data streams that are missing from healthcare today.
