Podcast

The future of hardware for continuous glucose monitoring

Episode introduction

The future of hardware is really exciting. There’s a lot going on in academia as it pertains to things like monitoring different molecules, glucose being one of them that hopefully, one day, we’ll be able to monitor continuously. When trying to understand the way that this technology works, there’s no one better to talk to then Josh Clemente, one of the cofounders of Levels, and Josh, being as nerdy as he is, understands hardware at a very deep level, not just the way it currently works, but exploring what is possible with the future of hardware and that’s what we got into in this episode:

  • How Levels is adding value by layering technology
  • Adapting therapeutic devices for everyday life
  • The potential for measuring multiple factors of metabolic health
  • Whether you should measure interstitial or arterial glucose
  • Advances in metabolic monitoring technology

Key Takeaways

Adding value by layering technology

Levels is focused on using existing medical hardware and becoming the operating system that allows people to use them in everyday life.

What Levels is doing is taking existing hardware technology, the continuous glucose monitoring system that is its own category of medical device with a few manufacturers here in the US and a few others abroad. And we’re building the insightful, actionable, accessible user experience on top of it. So, again, you can think of this as the mainframe versus personal computer. The companies have taken the lab assay to measure glucose and turned it into a personal device. And then you have windows, the software that lays on to it and makes it really accessible and useful. That sort of Levels is the software the operating system for those devices.

Adapting a therapeutic device to measure glucose in everyday life

The Levels wearable adds a Performance Cover to make a therapeutic device easier to use and less likely to be disrupted during day to day activity.

that was developed for therapeutic purposes to be able to measure glucose, primarily during diabetes. But we obviously see the potential for continuous molecular monitoring in the body to give insight into the reactions that follow the actions we’re taking every day. So, that’s adapting it to a different use case and a different population. And we want to solve all of the user experience problems, not just software, but also the daily life, physical interferences with the sensor and things like that. And so, that’s what that Performance Cover does, that fabric patch. It just sits over the sensor, provides a larger adhesive surface area to adhere it to the skin and also smooths the profile. So, you have this thing stuck to your arm. You want to make sure that when you’re taking your shirt on and off, or toweling off after a shower, you don’t accidentally brush that off because it is so in conspicuous you can forget that it’s on your arm, and accidentally peel it off and then lose that data. So, that fabric Performance Cover was designed to help provide you know a little extra water resistance, a lot of extra retention for snagging and other things you might experience in your daily life. And if you ask me a little bit of additional good looks.

The importance of a device that can handle movement

By using a thin filament instead of a rigid needle measurements can be taken comfortably in real-time during activity.

it’s about a quarter millimeter thick and about half millimeter wide. And then it extends into the cutaneous space about five millimeters for the Libre and about 11 millimeters for the Dexcom, which goes in at a slight angle. The Dexcom filament is slightly narrower. It’s round instead of that flat profile. But ultimately, these dimensions are, it’s hard to imagine just how small these devices are and how flexible they are.

And the reason for that flexibility is certainly for comfort. Having that filament in the top layer of the skin, if it were rigid or something like a needle, which is nothing like that, you would be able to feel that rigidity. But because it’s a slender filament, that can really shape and move along with you, it’s a very comfortable feel. And it’s also quite resilient to activity when you’re flexing muscles, when you’re moving around. And it’s in that way very conspicuous to someone when they’re wearing it.

The potential for measuring multiple factors of metabolic health

In a perfect world glucose measuring will be a small part of a more comprehensive array of metabolic factors that are trackable in real-time.

when you talk about CGMs, they’re only able to measure glucose. And that’s it, which is great. It’s giving you real-time feedback on what your glucose levels are. But there are so many other things related to metabolic health and or advisors specifically, Dr. Robert Lustig, he’s notorious for saying glucose is a very small part of the entire profile. When you talk about things like lactate, you talk about things like uric acid, you talk about things like insulin and all these other important factors, which go into truly understanding metabolic health.

So, there are a number of factors that have to be taken into account and to use your words, glucose is not a panacea. It’s a very important aspect of metabolic health, but it is one of many factors.

Should you measure interstitial or arterial glucose

The Levels wearable tracks interstitial glucose which concentrates in the tissues, muscles and brain in comparison to arterial blood glucose.

what’s interesting is outside of diabetes monitoring, where you want to maintain blood glucose levels of a certain range, for the general wellness population, I believe that the interstitial glucose is what’s most important. It’s what you care about the most. Those are the values that glucose concentrations that your tissues, your muscle, your brain is bathed in. That is the concentration that the energy consuming tissues the vast majority of them are experiencing. And so, it is possible that in the near term, we’ll start to see products that are specifically developed not to measure blood glucose, but actually to measure interstitial glucose and represent it as such. Right now, we’re measuring interstitial glucose and attempting through software to adjust it to represent blood glucose. That’s what the manufacturers do today. And I think in the future, you’ll see interstitial glucose for its own sake.

The rise of Aptamers

Aptamers represent a potential leap forward in monitoring technology by allowing non-invasive monitoring using specifically designed single strands of DNA or RNA.

There’s another really interesting technology that’s coming on to the … Well, it’s been around for a long time in academia, it hasn’t really hit the big time I think in in lab work or in biowearables. But that’s aptamers. And aptamers are a single strand of DNA or RNA. And the way the aptamer works is it binds to the target molecule in a unique way. So, it has a unique shape. And the shape of the molecule if you were to be able to zoom in on it, you’d see all these different interesting geometries and patterns and charges based on the electron densities of various regions.And so, you have all of this interesting attraction or repellent potential. And these certain aptamers that are shaped in certain ways will bind to these molecules. And it is because of the way that that binding occurs and the way that you can design an aptamer so that it has a recording molecule or some other tag on the end of it, you can create a sensor where when that aptamer binds to a molecule, it creates a different electrical signal than when it’s unbound. That’s essentially what it boils down to.

The benefits of Aptamers over existing monitoring technology

Aptamers are highly specific and allow monitoring of hormones and other molecules in extremely small concentrations.

there’s a fascinating potential for these aptamers to then unbind from that molecule they just bound to and then revert their signal back to where it was, and then repeat the process with a new molecule. So, if you think about it, it’s this potential for a very high specificity bind. And when I say specificity, I mean, it’s specific to one molecule. So, if you introduce something very similar to for example, glucose, a molecule that’s very close to glucose, but slightly different, it won’t bind. That’s a specific interaction. So, you can have this highly specific binding occur, similar to for example, antibodies, but you could theoretically unbind and continuously measure and release measure and release. And that is really fascinating. Because although it’s only been shown in the lab, it has this potential to transform the future of biowearables. So, if you’re not just breaking things down at higher concentrations the way enzymes work and you’re not just binding and holding on like a permanent lock and key the way antibodies do, then you now have this combination effect where you can selectively bind and release to a very specific molecule. And once you get something like that, that can operate at low concentrations on the range of picomolar, which is many orders of magnitude lower concentration than blood is in the body, which is millimolar. So, when you finally get something that’s reversible and specific and operates at those very low concentrations where hormones are, that’s what unlocks the holy grail of metabolic and just general health monitoring technology.

The downsides of any invasive monitoring device

All invasive monitoring technologies stimulate the bodies immune response when the skin is broken which can reduce the accuracy of readings over time as the body isolates the area.

The interesting thing is that all invasive technologies will have that foreign body response. It’s an immune defense mechanism. So, anything that is not the biological material of the host of the person, even other biological material, for example, during organ transplants, human tissue is placed inside the human body, it’s just someone else’s human tissue. And the body recognizes that that’s not its own. And it sets up an immune response to it, because it’s trying to protect itself. And so, oftentimes, for organ transplants, you have to take these immune suppressants in order to make sure that you don’t have a rejection of that organ. That process occurs at a very, very small scale, whether it’s a splinter that you get from picking up a picnic bench, all the way through to having a small sensor of filament inserted into the top layer of your skin. It’s a very, very muted response. It’s not something that’s traumatic. It’s not something that can cause any serious side effects whatsoever, because again, it’s such a small, minimally invasive element, but it occurs. And what that does is it builds that protein layer up, as you said, it’s an encapsulation of the filament. And that actually interferes with the lifespan and accuracy of the enzymatic sensor. That is likely to also apply to aptamers. However, the way in which the filament is designed, the surface area, the geometry, and the material selection can all be modified relative to enzymatic sensors, such that you can, again, further minimize the invasiveness of that process.

Building a platform on a noninvasive sensor is kind of like unlimited nuclear fusion

Lots of companies have teased noninvasive sensors for metabolic monitoring but in reality it may not provide reliable enough data to build a platform on top of.

At the end of the day, if we don’t know that it’s measuring what we think it’s measuring and we’re just trying to, in the words of an expert that I read about in this great article, review of noninvasive glucose, but in search of the best hammer to nail in the screw, that’s the way that the multivariate analysis has been described. It’s like just try and kind of throwing the whole thing and the kitchen sink at it to see if we can maybe make it work. And these problems don’t get easier as you try and build an extensible platform on a noninvasive sensor. So, I think it’s very tantalizing thinking about a noninvasive tool that you can just put on a wristwatch and it can measure all the molecules in your skin. Yeah, it sounds fantastic. But so does unlimited nuclear fusion energy. And it’s still a massive, massive Manhattan scale project to make that happen. I hope that it occurs. I hope that both noninvasive molecular monitoring happens and nuclear fusion, but it’s not something that’s around the corner.

Episode Transcript

Josh Clemente: So, if you want to think about biosensors, you can kind of break them down into two categories, one talks about the physical, the invasive versus noninvasive sort of categorization. The other is the functional, how do they actually work. Invasive sensors are chemical. They use a chemical reaction of some kind to measure a molecule. Noninvasive use the rules of physics, typically electromagnetic radiation, electromagnetic interrogation to measure at a distance something about the molecule. So, chemical versus physics, chemistry, physics. That’s kind of how it breaks down.

Ben Grynol: I’m Ben Grynol, part of the early startup team here at Levels. We’re building tech that helps people to understand their metabolic health. And this is your front row seat to everything we do. This is A Whole New Level. The future of hardware, it’s really exciting. There’s a lot going on in research in academia as it pertains to things like monitoring different molecules, glucose being one of them.

Ben Grynol: But they’re all these different molecules that people can monitor. And hopefully, one day, we’ll be able to monitor continuously. That being things like insulin, that being ketones, that being lactate, the list goes on. There are all these reasons why a person might want insight into their own biometric data on more levels than just glucose.

Ben Grynol: True metabolic health is an amalgamation of all these different things that are happening in our body simultaneously. And so, when you start to think about hardware, what it currently does, how it currently works, it’s kind of this unknown thing unless you really deconstruct it. We know if you’re using Levels currently, we know that you’ve got this little device, a piece of hardware, and that sits on your skin, the little disc, and it gets covered up by this patch, the Performance Cover. And from there, you get these readings to your phone. And it’s sort of this magic that happens.

Ben Grynol: You exercise, you see a response. You eat something, you see a response. When that happens, that’s actually the software, the Levels technology, the software that we develop as a team on top of this hardware. The hardware is made by other companies. There are a couple of companies in the space of being Abbott and Dexcom, and they make hardware that gives you the ability to read glucose in real time.

Ben Grynol: And when trying to understand the way that this technology works, there’s no one better to talk to then Josh Clemente, one of the cofounders of Levels. So, Josh is a mechanical engineer by training. And if you’ve listened to him on other podcasts, you know that he spent a lot of time at SpaceX putting rockets into the air. And so, Josh, being as nerdy as he is, understands hardware at a very deep level, not just the way it currently works, but exploring what is possible with the future of hardware.

Ben Grynol: And what the media reports on is things like the Apple Watch, right? The Apple Watch one day maybe it will be able to give you insight into your own glucose levels. And that is research that is being worked on. There are many companies working on this type of noninvasive technology. So, a durable good, like an Apple Watch Could possibly give you glucose readings. But there are all these challenges into the process of doing it and doing it well, doing it accurately.

Ben Grynol: So, Josh is always reading up on things and understanding what’s going on in the space. When we get questions from other Levels members or the general public about what’s possible, is it possible that one day Apple will be able to read glucose readings?

Ben Grynol: Well, we have to understand whether or not it is. And so, Josh and I sat down and we dug into everything around the way current CGM technology works and what the future of hardware might look like. What are some of the benefits? What are some of the challenges? What are some of the limitations of the technology in its current state and the future state technology that’s being developed, being researched at the moment? Here’s where we dug in.

Ben Grynol: I think we got to start at the foundation, which is engineering, manufacturing, and then we’ll get into what the hardware is and how it works. So, let’s think through like what is engineering and it might sound elementary, but I think if somebody doesn’t have a lens on engineering, the word “engineer” can be hard to break down, right? So, it’s analogous to a doctor. There’s a million different types of doctors, being hyperbolic in every which way.

Ben Grynol: But there’s all these different types of engineers and classifications. And you need a bunch of different engineers to make the hardware. So, why don’t we go through a bunch of different types of engineers and just talk about that, and then we’ll hop into manufacturing.

Josh Clemente: Yeah, I mean, engineers are responsible for taking … I think they’re responsible for taking science and turning it into solutions for everyday life. So, it’s kind of hard to put a firm description on it. But that’s kind of how I think about engineering is using the guidelines, the guide book that’s developed by hard scientists, and then figuring out how to solve everyday real world problems like building bridges or creating medical devices that can measure glucose in the blood.

Josh Clemente: And as you said, there’s a ton of different categorizations and specialties in engineering. The ones that are most relevant here would be probably biomedical, biochemical, mechanical, and electrical, I would say. So, these are engineers who focus on for biomedical, the interface between products and biology. So, this would be device development, prostheses development, anything that’s going to be interfacing with the human body and changing its function or assisting as function.

Josh Clemente: Biochemical or just chemical engineers are focused on the chemistry and interactions and reactions that occur whether with molecules, that could be drugs, or with materials in devices, and how they create responses and reactions in the body or in tissues or even outside the body and in solutions on lab bench and things like that.

Josh Clemente: And then mechanical engineers, that’s really broad. Mechanical engineers can work on essentially anything that is a physical product that typically has some sort of structural or functional requirement mechanism, a motor, all the way up through cars, airplanes, spacecraft, bridges. Anything that’s structural and physical and interacted with physically would be in the domain of mechanical engineering, obviously, with specializations inside that.

Josh Clemente: And then electrical engineers will be primarily looking at designing electrical hardware and implementing the hardware-software interface, firmware, software functionality and making sure that essentially, the mechanical requirements can be carried out through electrical signals, which most mechanical engineers, they’ll have some competency in understanding basic software and electrical design.

Josh Clemente: But they’re not going to be the ones creating a printed circuit board, for example, that then send signals to the actuators and motors to drive these fine medical device components that’ll typically be in the realm of an electrical engineer.

Josh Clemente: And then I skimmed over it, but manufacturing engineering and process control engineering and quality engineering, once you get through the initial concept development and design of any product, you then need to build it. And in order to build it, there’s this whole other domain, which is focused on optimization of the construction of that product, whatever that is. It can be anything from a drug to, again, an airplane.

Josh Clemente: And that requires its own design entirely of the each step that occurs. You can kind of think of it as describe to someone how to make a peanut butter sandwich. And you can say, “Well, take the peanut butter out and put it on the bread and you’re done.” But actually, it’s not quite so easy, because that you can interpret that 100 times. You can take the lid off and dump the whole jar on the bread. You can waffle the bread over the peanut butter.

Josh Clemente: There’s a million ways to misinterpret that. And so, you actually need to break that down into every micro step that occurs so that you get the optimal peanut butter sandwich at the end of that process no matter who is in the loop.

Josh Clemente: And that’s what manufacturing process control engineers do. And the quality engineers are the ones that help them iterate and maintain points of quality control. So, testing various steps along the way to ensure what the minimum number of steps you get the maximum output. That’s kind of how the domain plays out.

Ben Grynol: Yeah, and there’s a software engineers as well. So, when working on a hardware and software product, you’re going to have the software side of things. It’s funny to have the conversation because if you’re close to engineering or if you yourself are an engineer, then it seems funny to be like, a civil engineer is an architectural engineer, a structural engineer. We can just extrapolate this forever.

Ben Grynol: But I think sometimes when having conversations, it’s easy to forget where, let’s see with friends, if it’s like, “Oh, we’re bringing on lots of engineers,” just using that word in a circle of people like let’s say in just within Levels, we understand an engineer is a software engineer. That is our heuristic for taking the shortcut of engineer means software engineer But to other people, they’ll be like, what kind of engineer. Because they don’t know necessarily what we’re referring to.

Ben Grynol: So, when we talk about having engineers, our entire team is software engineer. So, other than you, we don’t have any mechanical engineers because we’re not a hardware company. We’re a software company that’s building a layer on top of existing hardware from other companies. So, it’s important to sort of classify everything, understand it, and it paints a good picture and provides a foundation for everything else.

Ben Grynol: But on the hardware side of things, so if we go into manufacturing, and again, this is just a nerdy subject that if you’re deep into manufacturing, it’s fun to talk about. But there’s all these different levels within manufacturing. You can have invention. So, Dyson is a great example of a company that invents parts and processes to make entirely new goods. They’re not just creating a derivative, let’s use a vacuum. They’re not creating a derivative vacuum. They actually invented the part and the process to be able to change the way that a vacuum sucks up dust and dirt. So, that’s one part of manufacturing.

Ben Grynol: Then you’ve got design manufacturing, which would be reskinning an item. So, reskinning meaning, “Hey, we’re taking an existing good that has off the shelf parts. And we’re just putting some industrial design around it, meaning like some form and factor.” And some companies just do that. They take off the shelf parts. That’s what Apple did with the first iPod. They’re notorious for shipping that in, gosh, I can’t remember if it was like six or nine months that Tony Fadell and the team got that to market him.

Ben Grynol: But that was off the shelf parts that Jony Ives, who was head of design at Apple at the time, basically skinned it with his team and said, “Cool, here’s the casing in a little wheel. This is how it’s going to work.”

Ben Grynol: And then, the last part of it is assembly. So, you can be a manufacturer that only assembles off the shelf parts that are already designed. Like you take, let’s say, furnitures, you might buy prefab legs and prefab, whatever it is, arms and cushions, and there you go, and you just sort of like put it to this thing and get it to market. And so, that even within manufacturing, there are all these different levels of manufacturing that it takes to get a good to market.

Ben Grynol: And so, it’s important to understand what type of manufacturing is required for creating a new product, whether it is a rocket, like SpaceX, or whether it’s a car like Tesla, or a piece of hardware, like Abbott would do very different in every which way.

Josh Clemente: Absolutely.

Ben Grynol: So, on the note of manufacturing, we should break into hardware. So, let’s talk about CGMs to begin with, because I think there’s a misconception. So, if people have used Levels, they understand that the CGM is not part of the Performance Cover. You apply it, you understand, “Hey, these are two distinct things.” The cover is just an adhesive piece of fabric, like a giant, round, circular Band-Aid, if you want to call it that, a stylized Band-Aid. But that goes over top of a little coin-sized device, that’s, I guess, what, an eighth of an inch in depth about that by-

Josh Clemente: Thicker than that, but approximately.

Ben Grynol: One and a half across, something like that. And so, that is what adheres to your arm and you put the cover over it. So, it is entirely different. But why don’t you dive into what exactly is that hardware? How does it work?

Josh Clemente: What Levels is doing is taking existing hardware technology, the continuous glucose monitoring system that is its own category of medical device with a few manufacturers here in the US and a few others abroad. And we’re building the insightful, actionable, accessible user experience on top of it.

Josh Clemente: So, again, you can think of this as the mainframe versus personal computer. The companies have taken the lab assay to measure glucose and turned it into a personal device. And then you have windows, the software that lays on to it and makes it really accessible and useful. That sort of Levels is the software the operating system for those devices.

Josh Clemente: So, when you use the Levels program, you get access to a device and there are a couple versions that we have out there. But it’s a combined set of functions in a single product, one of which is interfacing with the skin of the body, so being able to actually measure molecules in the skin. And that is done via a conductive electrode, a little filament, this little flexible fiber-like thing that sits in between the skin cells and actually react with glucose molecules in what’s called the interstitial fluid. So, there’s that little filament.

Josh Clemente: That filament conducts electrons up through the skin to the sensor module. And that sensor module has a little chip in it so that that disc that sits on the skin contains a little chip with a microprocessor. And it turns that electrical current into a signal, and that signal represents glucose levels. And then there’s a wireless transfer chip. And this could be near field communication, which is an RFID tag, if you know about that technology, or it could be a Bluetooth chip, if you know about that technology, and there’s a little battery.

Josh Clemente: And so combined, you have that little filament and then you have a disc housing. And inside that is essentially a few electrical components and a wireless transfer chip. And that then sends the signal that represents glucose to the user’s device. That little sensor has been developed over the past few years. As we said, there are a few manufacturers. But in general, the functionality of it is the same across manufacturers that uses an enzyme that breaks down what’s called a substrate. In this case, the substrate is glucose.

Josh Clemente: So, it breaks that down in the presence of oxygen. And it creates some byproducts. One of those byproducts touches the little filament in the interstitial space. And electrons are, are broken off and run up the electrode. That’s essentially the core principle.

Josh Clemente: And so, we’re taking that and that was developed for therapeutic purposes to be able to measure glucose, primarily during diabetes. But we obviously see the potential for continuous molecular monitoring in the body to give insight into the reactions that follow the actions we’re taking every day. So, that’s adapting it to a different use case and a different population. And we want to solve all of the user experience problems, not just software, but also the daily life, physical interferences with the sensor and things like that.

Josh Clemente: And so, that’s what that Performance Cover does, that fabric patch. It just sits over the sensor, provides a larger adhesive surface area to adhere it to the skin and also smooths the profile. So, you have this thing stuck to your arm.

Josh Clemente: You want to make sure that when you’re taking your shirt on and off, or toweling off after a shower, you don’t accidentally brush that off because it is so in conspicuous you can forget that it’s on your arm, and accidentally peel it off and then lose that data. So, that fabric Performance Cover was designed to help provide you know a little extra water resistance, a lot of extra retention for snagging and other things you might experience in your daily life. And if you ask me a little bit of additional good looks.

Ben Grynol: I think you’re notorious for saying you don’t realize how close you come to door casings and start wearing a CGM, right, because that’s a very realistic scenario where we hear it from members and even our own team where they say, “Oops, knocked it off, I got a little too close to the door casing.” But it happens when you’re wearing it. And that’s the point of having something like the Performance Cover where it just gives it a little bit more protection so that it’s not a rigid profile, saying, “Hey, here’s a disc that is easier to knock off.”

Ben Grynol: One of the things that you said, which is maybe we’ll break this down even further is, so the little filament measures the interstitial fluid. That might not be familiar to people who might not have a background in it. But I think there’s a misconception that they think cool, blood glucose. This thing measures my blood and it doesn’t. You can measure glucose levels through the blood. But that would be taking a panel, blood tests. That is entirely different than what a CGM does. So, it’s not actually taking readings from a person’s blood.

Ben Grynol: And part of that, part of the reason is that the filament is so small, so if you haven’t seen it in person, it looks a little bit thicker than a hair. I can’t remember the profile, but I think it’s under a millimeter or something like that for Abbott or maybe just over for Dexcom, it’s like 1.1, 1.1 for Dexcom or something.

Josh Clemente: Yeah, it’s about a quarter millimeter thick and about half millimeter wide. And then it extends into the cutaneous space about five millimeters for the Libre and about 11 millimeters for the Dexcom, which goes in at a slight angle. The Dexcom filament is slightly narrower. It’s round instead of that flat profile. But ultimately, these dimensions are, it’s hard to imagine just how small these devices are and how flexible they are.

Josh Clemente: And the reason for that flexibility is certainly for comfort. Having that filament in the top layer of the skin, if it were rigid or something like a needle, which is nothing like that, you would be able to feel that rigidity. But because it’s a slender filament, that can really shape and move along with you, it’s a very comfortable feel. And it’s also quite resilient to activity when you’re flexing muscles, when you’re moving around. And it’s in that way very conspicuous to someone when they’re wearing it.

Ben Grynol: Yeah, and that’s something we’ll get into later when we talk about the difference between invasive and noninvasive technology. Because they’re totally different in application. And I think the media reports a ton on noninvasive, which is, when you hear things like, “Oh, the Apple Watch is going to be able to read glucose levels.” We’ll dive into that in a bit.

Ben Grynol: But we’ll just use Abbott for now, this little piece of hardware, this little disc that’s giving you readings. And sometimes we hear from people, “Cool, can the device measure ketones?” Let’s just use ketones and glucose. And right now, with technology in the world, it’s just not possible to measure multiple molecules at one time through these pieces of hardware.

Ben Grynol: So, when you talk about CGMs, they’re only able to measure glucose. And that’s it, which is great. It’s giving you real-time feedback on what your glucose levels are. But there are so many other things related to metabolic health and or advisors specifically, Dr. Robert Lustig, he’s notorious for saying glucose is a very small part of the entire profile. When you talk about things like lactate, you talk about things like uric acid, you talk about things like insulin and all these other important factors, which go into truly understanding metabolic health.

Ben Grynol: So, there are a number of factors that have to be taken into account and to use your words, glucose is not a panacea. It’s a very important aspect of metabolic health, but it is one of many factors. So, we’ve talked invasive hardware. There’s a couple other things to be nerdy about it. So, we would call a current CGM would be like a transdermal filament-based enzymatic sensor. Right?

Josh Clemente: Yeah. And actually, just to fall back on the blood to interstitial fluid discussion that you brought up, that sensor class that you just touched on that, as we mentioned, is measuring glucose levels in the peripheral tissues, out near the skin, near the muscle. And that is different than what’s called blood glucose, which when you have a needle stuck in your arm, and you pull blood for a lab draw, and go and process that. That’s coming from essentially your arterial blood flow. And it’s quite different.

Josh Clemente: So, as the heart is pumping and moving blood through those big arteries, and then it diffuses out into the tissues out all the way to the skin surface, the concentrations of a whole bunch of different analytes are changing throughout that process. And so, we’re measuring in the interstitial space, which, historically for diabetes management, the goal is to try to represent the blood levels, which these two things, interstitial glucose levels and blood glucose levels track the same. So, the patterns and trends are the same. They correlate very, very well. But there’s typically a time difference.

Josh Clemente: So, the peaks and valleys will be shifted a little bit in time. So, if you were wearing a CGM and you go and check your blood glucose, which is a little bit challenging, you can actually measure capillary glucose. So, not to get too pedantic here. But capillary glucose is again, it’s not representative exactly of what’s in your arteries, because it’s at your fingertip. So, you can use a little fingertip device. You measure capillary glucose, you might notice that the CGM to capillary glucose values are different.

Josh Clemente: And if you track those trends, they’ll follow each other, but there’ll be a few minutes separated. And that’s because again of this time-dependent process of the molecules flowing, the concentration is changing as it’s moving through the body, through the thickness of the body out into the surfaces. Now, what’s interesting is outside of diabetes monitoring, where you want to maintain blood glucose levels of a certain range, for the general wellness population, I believe that the interstitial glucose is what’s most important. It’s what you care about the most.

Josh Clemente: Those are the values that glucose concentrations that your tissues, your muscle, your brain is bathed in. That is the concentration that the energy consuming tissues the vast majority of them are experiencing.

Josh Clemente: And so, it is possible that in the near term, we’ll start to see products that are specifically developed not to measure blood glucose, but actually to measure interstitial glucose and represent it as such. Right now, we’re measuring interstitial glucose and attempting through software to adjust it to represent blood glucose. That’s what the manufacturers do today. And I think in the future, you’ll see interstitial glucose for its own sake.

Ben Grynol: Yeah, and that gets into the conversation around going deeper into invasive versus noninvasive and why invasive, so invasive anything that adheres and sits just below the skin, why that is necessary. So, one of the things we touched on was so transdermal is the current sensors, they’re enzymatic, so they are reading in a different way. The filament is reading in a different way than aptamer-based tech. So, maybe walk through the difference between enzymatic tech and then aptamer-based tech and why there are challenges and things like accuracy.

Josh Clemente: So, right now, the sensor modality or transducer, there’s a bunch of different ways to try and describe this. But what you have is you have again, that flexible filament that is what’s interfacing with the skin, and then on that filament is some sort of chemical reactant. You need some mechanism that interacts with very specific molecules like glucose and doesn’t interact with other ones.

Josh Clemente: And that’s the real magic of biosensor development is you’re trying to only find the molecules that you care about and only create a reaction with those molecules and not allow interferences from very similar molecules, or altogether different molecules that might attack your sensor and create an unspecific inaccurate sensor signal. So, today, the sensors use primarily enzymes.

Josh Clemente: And enzymes are, you just think of them as molecules that are very … They’re designed to break down another very specific molecule. And so, glucose oxidase is the name of the enzyme that is currently coated on the CGM filaments that we use. It oxidizes or in the presence of oxygen breaks down glucose. And that’s a similar process. There are enzymes that exist that can break down many different molecules. So, that is an enzymatic sensor. That’s what the state of the art is today.

Josh Clemente: And then there are a whole bunch of other specific and sensitive chemical sensor designs, and a few examples could be antibodies. So, antibodies are typically associated with the immune system. But they’re another specific molecule that binds to a site on another very specific molecule of interest, the target were anolyte. And antibodies, they bind with high specificity. When they bind, they don’t come apart. And so, they’re very common in lab-based tests. You take a blood sample, and you want to know the concentration of a whole host of different molecules or test targets, you’ll use antibodies typically.

Josh Clemente: And then there’s another really interesting technology that’s coming on to the … Well, it’s been around for a long time in academia, it hasn’t really hit the big time I think in in lab work or in biowearables. But that’s aptamers. And aptamers are a single strand of DNA or RNA. And the way the aptamer works is it binds to the target molecule in a unique way. So, it has a unique shape. And the shape of the molecule if you were to be able to zoom in on it, you’d see all these different interesting geometries and patterns and charges based on the electron densities of various regions.

Josh Clemente: And so, you have all of this interesting attraction or repellent potential. And these certain aptamers that are shaped in certain ways will bind to these molecules. And it is because of the way that that binding occurs and the way that you can design an aptamer so that it has a recording molecule or some other tag on the end of it, you can create a sensor where when that aptamer binds to a molecule, it creates a different electrical signal than when it’s unbound. That’s essentially what it boils down to.

Josh Clemente: And there’s a fascinating potential for these aptamers to then unbind from that molecule they just bound to and then revert their signal back to where it was, and then repeat the process with a new molecule. So, if you think about it, it’s this potential for a very high specificity bind. And when I say specificity, I mean, it’s specific to one molecule. So, if you introduce something very similar to for example, glucose, a molecule that’s very close to glucose, but slightly different, it won’t bind. That’s a specific interaction.

Josh Clemente: So, you can have this highly specific binding occur, similar to for example, antibodies, but you could theoretically unbind and continuously measure and release measure and release. And that is really fascinating. Because although it’s only been shown in the lab, it has this potential to transform the future of biowearables. So, if you’re not just breaking things down at higher concentrations the way enzymes work and you’re not just binding and holding on like a permanent lock and key the way antibodies do, then you now have this combination effect where you can selectively bind and release to a very specific molecule.

Josh Clemente: And once you get something like that, that can operate at low concentrations on the range of picomolar, which is many orders of magnitude lower concentration than blood is in the body, which is millimolar. So, when you finally get something that’s reversible and specific and operates at those very low concentrations where hormones are, that’s what unlocks the holy grail of metabolic and just general health monitoring technology.

Ben Grynol: Yeah, one of the other advantages is accuracy. So, one of the challenges with enzymes, so we know with the filaments when they’re coated in glucose oxidase, that there is a shelf life. There is a shelf life to that enzyme being on the end of that filament. There’s also a challenge around biofouling, which is, when you adhere the sensor to your arm, so you put it on and there’s, we’ll call it to keep it pretty colloquial, it’s like a spring-loaded applicator that has a needle, that is, gosh, how long is the needle, maybe 10 mils, something like that.

Josh Clemente: It’s a little bit long. It’s longer than the filament. But the interesting thing about the way it’s designed is it only introduces, it’s more of an introducer than anything else. It introduces the filament, which is very flexible to the surface of the skin and the needle itself, it doesn’t stay in the body, it just breaks the surface and allows the filament to thread its way in.

Ben Grynol: The delivery mechanism to say, cool, here’s the tiny portal into the body for the filament to sit in there. But anytime you break the skin, no matter how uninvasive, if you want to call that it is, it’s still breaking the skin. And so, your body goes, cool, to use your words, we’re going to gate off this area. We’re going to quarantine this area of the body or the tissue around it to make sure that the intruder isn’t going to cause harm, even though it’s still on a very, very small scale.

Ben Grynol: And so, over time, protein can form around that. And that can over the span of the life of the sensor, if it’s a 14-day sensor that can lower the accuracy over time. So, with aptamer base tech, it’s changing the way in which you can get accurate readings.

Josh Clemente: Yeah, I mean, the interesting thing is that all invasive technologies will have that foreign body response. It’s an immune defense mechanism. So, anything that is not the biological material of the host of the person, even other biological material, for example, during organ transplants, human tissue is placed inside the human body, it’s just someone else’s human tissue. And the body recognizes that that’s not its own. And it sets up an immune response to it, because it’s trying to protect itself.

Josh Clemente: And so, oftentimes, for organ transplants, you have to take these immune suppressants in order to make sure that you don’t have a rejection of that organ. That process occurs at a very, very small scale, whether it’s a splinter that you get from picking up a picnic bench, all the way through to having a small sensor of filament inserted into the top layer of your skin. It’s a very, very muted response. It’s not something that’s traumatic. It’s not something that can cause any serious side effects whatsoever, because again, it’s such a small, minimally invasive element, but it occurs.

Josh Clemente: And what that does is it builds that protein layer up, as you said, it’s an encapsulation of the filament. And that actually interferes with the lifespan and accuracy of the enzymatic sensor. That is likely to also apply to aptamers. However, the way in which the filament is designed, the surface area, the geometry, and the material selection can all be modified relative to enzymatic sensors, such that you can, again, further minimize the invasiveness of that process.

Josh Clemente: And the, let’s see how to say this, the degree of disruption to the tissues. That’s essentially how this immune response works is a significant response will occur when there’s a significant “trauma.” If you have a large wound that occurs from falling off your bike or something, you’re going to have a larger immune response to that then than having a very minimal skin break.

Josh Clemente: And so, continuing design that process and all of the selections of the various elements they’re in helps to drive that foreign body response down. And I think eventually, there will be a possibility for a much longer lifespan sensor on the order of months as opposed to currently days.

Josh Clemente: But going back to it, the reality is that fully eliminating the foreign body response won’t be possible without the intervention of some sort of suppressants like localized tissue immunity, or inflammation suppressant, which I don’t necessarily know is the right move. But again, anytime you’ve got something that’s in the body, the body is really, really good at identifying that.

Josh Clemente: And so, that’s where one of the philosophical or theoretical upsides of noninvasive monitoring comes in is that the whole concept behind it is that you can measure molecules in the scan in the tissue in regions of the body that are biologically relevant and active and measure those molecules without entering the skin.

Josh Clemente: So, if you want to think about biosensors, you can kind of break them down into two categories. One talks about the physical, the invasive versus noninvasive sort of categorization. The other is the functional, how do they actually work. Invasive sensors are chemical. So, they use a chemical reaction of some kind to measure a molecule. Noninvasive use the rules of physics, typically electromagnetic radiation, electromagnetic interrogation to measure at a distance something about the molecule. So, chemical versus physics, chemistry, physics. That’s kind of how it breaks down.

Josh Clemente: And with the physics-based sensors, you’re usually using, like I said, electromagnetic radiation, which sounds scary, but it’s actually anything from any waves, electromagnetic waves, which includes visible light, it includes infrared light, and it includes higher energy ultraviolet and stuff. But that typically won’t be used because it can cause damage. So, it’s usually visible light into the lower energy spectrum, radio waves and things like that.

Josh Clemente: And right now, noninvasive electromagnetic sensors are used for things like heart rate measurement. Anyone that’s got a smartwatch, you look at the little LEDs underneath it flashing, that’s a form of physics-based electromagnet magnetic interrogation at a distance. It’s measuring your pulse through your skin using light. And you can extrapolate that technology and a few others to all of which are kind of based on a similar concept where you send out a pulse, and then measure the response and use the signal processing tools that we’ve developed to determine what was in that signal that you got back.

Josh Clemente: And that basis of the technology is really broad. It’s also very complicated. There’s so much interference. There’s so much in that signal that can change. And you can’t quite know that it’s specific. And that’s what the challenge is. Now, we can get deeper into that. But basically, that’s the promise. The promise is that without breaking the skin, without a foreign body response, without any of the complexity of chemical sensors, you can measure the same molecule.

Josh Clemente: But the tricky thing is, now it’s a bulk measurement, it’s not specific, and you have a whole host of environmental interference that you don’t have with the chemical sensors. So, it’s a trade. And as we know, currently, the chemical sensors are winning.

Ben Grynol: So, one of the things about noninvasive is it also creates durable goods, which is way different. Because right now with invasive, they’re disposable. And so, there’s a lot of waste and we can go down that rabbit hole. But there’s a big difference when you talk about an Apple Watch is a durable good. It lasts until basically the battery runs out or you drop it or break it in some which way. Whereas any invasive tech is going to be disposable.

Ben Grynol: One of the other things with invasive tech and I think there’s these other classifications, it’s invasive we’ll call. Invasive sounds like such a brash word, but it’s just the way that it’s described. Because really, CGMs aren’t overly invasive. It’s not like this really large foreign body sitting in your arm. It’s just if you haven’t seen it, it really is that filament is like a hair.

Josh Clemente: Yeah. I like to think of it as it’s not noninvasive. And it’s the single step above noninvasive. Yeah, kind of have to draw a line somewhere. It’s either invasive or it’s not. It crosses the skin barrier. Thus, it is minimally invasive or just barely, noninvasive.

Ben Grynol: Exactly. And so, then one of the other classifications is minimally invasive, which there actually is, and that’s this technology that’s being researched and developed around microneedles. And so, microneedles are on the scale of a filament, they are, I don’t know, what would you say, what portion of a filament? I guess they’re 2% of the length. I mean, it’s very small.

Josh Clemente: It really varies. There are many teams working on microneedles and they all have sort of different geometry approaches. But there’s no hard cutoff. There’s no firm definition that I’m aware of, of what constitutes a microneedle. But some of these are on the order of 500 microns. So, it’s a few tens of 1000th of an inch long or fractions of a millimeter. It’s really almost imperceptibly small. If you were to look at a grid of microneedles, you could see the texture, but it would be hard to tell what it looks like or that it really has much of a protrusion at all.

Ben Grynol: Yeah, and the reason why these are still important is because they are using the technology of enzymes versus anything that’s physics related or light related, as you’re saying with durable goods or with noninvasive tech. And so, the idea is, hey, if we can get something that is even less invasive and it can give you interstitial fluid or readings, then hey, this might be good tech.

Ben Grynol: But some of the challenges around that are that the microneedles, as they’re referred to, are sitting so close to the surface that things like movement and things like sweat, or any properties that might give inaccurate readings can cause challenges as far as making the device useful.

Ben Grynol: And so, then you start to look at it, let’s remove microneedles, given the state they’re at right now in the world. So, the state being that they would not be accurate enough in the research that exists, they would not be accurate enough to provide glucose readings, let’s just use glucose, provide readings that would give people enough insight or accuracy to say, cool, this is benchmarked against what I would see through a traditional enzymatic sensor or a CGM.

Ben Grynol: So, those are still very much in development. And there are a lot of challenges with making that tech work. Will it work one day? Possibly, but given the state right now, it’s just not going to be. It’s not going to provide the readings and the data that is necessary to make them effective. And so you get back to this world of sweet, I’ll be in to the CGM thing when Samsung does it or Apple does it or Facebook or name a company that’s working on using this tech of noninvasive, which would be like PPG, so you referenced Garmin I think and Whoop uses. So, PPG is photoplethysmography.

Ben Grynol: And that is basically using light or spectroscopy is one form of reading. But maybe go into the challenges around, why this just can’t work and things around the hardware not being as close to your skin as possible. Things like sweat, all of the things that cause issues with accuracy, where I think there’s some nice anecdotes of people who are researching the tech right now and working on it from an academic level, like they say, “Hey, this is a very, very challenging tech to make.”

Ben Grynol: And I think your grandkids’ grandkids will be working on it. So, maybe dive into sort of what is PPG and why is the world of noninvasive, although the media loves to report on it, why are we so far away from it?

Josh Clemente: Well, it’s tantalizing. PPG is actually one of the few noninvasive technologies that I think is, it’s not really even in the ballpark. It’s too nonspecific the way that it personally works. It’s basically measuring a volume change. Plethysmography is a volume change measurement. So, PPG, it’s great for measuring heart rate, heart rate, even variability measuring the distance between pulses. And it’s great for potentially blood pressure and things where it’s a bulk measurement of tissue volume change or blood volume change. But it’s not specific to molecules.

Josh Clemente: Now, spectroscopy and other light-based specific measurements, more specific measurements like Raman spectroscopy, these are tools spectroscopy, by the way is just spectral measurement, spectral analysis. You’ll hear a lot of spectroscopy, when you’re talking about a spectrum of light or a radio waves, essentially, whenever you get a range of signal responses, like a frequency if you ever see a frequency plot that looks at you know, you’ve got frequency on the x-axis and amplitude on the y-axis or time and an amplitude. There’s a bunch of different ways to display it.

Josh Clemente: But looking at the spectrum of a signal, that’s spectroscopy. So, Raman scattering, infrared spectroscopy, these are the types of technologies that certainly on the lab bench show a lot of promise. You can measure concentrations of molecules of interest with these tools. But when you take that from the lab and then you then apply it to the body, it’s a very, very different world.

Josh Clemente: First of all, you’re trying to measure through the skin a specific distance and understand the concentrations that are changing of a colorless, water-soluble molecule glucose in concentrations of millimolar. So, several millimoles per liter, so 1000th of a mole, which is a chemistry term for the amount of a certain substance in each liter of blood and there’s not that much blood flowing through. So, it’s a very, very, very weak signal.

Josh Clemente: And you’re attempting to measure that and be able to determine small fluctuations in it. So, for those people who are looking to monitor the effects of stress, enough sleep, enough nutrition, being able to measure 10% fluctuations in that very, very, very weak signal is already extremely challenging, now introduce the fact that there are other constituents that interact with light and these spectral ways that they don’t corrupt the signal, they just contribute to the signal in a way that you don’t know.

Josh Clemente: So, you’re using all of these light-based or EMR, electromagnetic radiation-based methods, you’re measuring a bulk signal. You’re sending out a pulse of light or radio waves, and you’re receiving back what comes back. And you don’t know that light or that signal interacted with everything in its path, the skin, the sweat, the water from rain, the vibrations of the passing train, everything that occurred is captured in the signal that you get back. And you have no real knowledge because this is not a chemistry-based sensor. You have no way of confirming that certain parts of the signal are only impacted by certain molecules.

Josh Clemente: Now, there are tools like I mentioned Raman spectroscopy, which are more specific, but again, not something like sweat, which contains glucose. But many papers have been published showing that even though sweat contains glucose, it does not correlate with blood glucose, so it’s actually totally decoupled. So, it doesn’t track with blood glucose. And so, you can have these corruptions that occur where the signal is broken by the fact that there’s the presence of the molecule you think you’re measuring, but it doesn’t track to the concentration in the body that you care about.

Josh Clemente: So, it’s very complex in that way, where using an interstitial filament when you are in the tissue that you care about and you have a direct enzyme that only breaks down glucose, you know what you’re measuring, and you have high confidence and specificity of them measurement.

Josh Clemente: When you’re doing noninvasive, again, it’s bulk. You don’t know what you’re measuring. And it’s currently relegated to software processing, using essentially throwing processing power at the problem to try to figure out, can we use machine learning? Can we use pattern recognition? Can we use big labeled data sets of some kind to figure out how to fit the square peg in the round hole.

Josh Clemente: And this multivariate analysis, it might work at some point. But we know that it’s so complex technically that even if we can get to the point where it’s marginally useful for measuring glucose ranges, high, medium and low, we know that we will face orders of magnitude more challenge to then go and measure the next molecules that are interesting for metabolic health.

Ben Grynol: One of the other things is like lab setting versus human environment.

Josh Clemente: Yeah. There is no average person. And so, you can calibrate these things to the … Well, this is what’s happening is with noninvasive techniques, you can measure the signal with a known concentration on the lab bench. You mix glucose with blood, measure the signal spectroscopically, and then correlate the two. And so now you know what signal corresponds to that glucose concentration.

Josh Clemente: And then start to fluctuate or rather just take that and put on the human body and watch how the signal changes. And initially, things might track. But then invariably, all of the studies have shown this, they start to diverged. And the reason is that you were measuring what’s called spurious correlations. So, just because those two things, the concentration that you knew and the spectral signal that you were measuring, just because they could be fit to each other using quick mathematical technique, doesn’t mean that you actually are measuring the glucose signal like you may want to believe.

Josh Clemente: And so, as things change like blood pressure and other metabolites are circulating, and you’re starting to get a little sheen of glucose on your skin or a person to person variations where skin color changes, just a few shades can interact with the way that the signal is broadcast through the tissue. And so, all of these different things person to person and within person start to create a divergence between the signal you’re measuring and what you thought you were measuring.

Josh Clemente: And so, this is shown over and over, again, spurious correlations are common in the literature, but they don’t hold up over time. And that might require one solution is to calibrate, use a finger stick or something like that, to continually calibrate. But at the end of the day, if we don’t know that it’s measuring what we think it’s measuring and we’re just trying to, in the words of an expert that I read about in this great article, review of noninvasive glucose, but in search of the best hammer to nail in the screw, that’s the way that the multivariate analysis has been described.

Josh Clemente: It’s like just try and kind of throwing the whole thing and the kitchen sink at it to see if we can maybe make it work. And these problems don’t get easier as you try and build an extensible platform on a noninvasive sensor. So, I think it’s very tantalizing thinking about a noninvasive tool that you can just put on a wristwatch and it can measure all the molecules in your skin. Yeah, it sounds fantastic. But so does unlimited nuclear fusion energy. And it’s still a massive, massive Manhattan scale project to make that happen. I hope that it occurs. I hope that both noninvasive molecular monitoring happens and nuclear fusion, but it’s not something that’s around the corner.

Josh Clemente: And the news media kind of like, because it’s so exciting and it has so much potential it sounds very futuristic and sci-fi, it gets broadcast quite often. But as we know the Apple Watch 7 was promised to have CGM in it. And there was no word on the announcement the other day and I think that’ll be the case for several more years as those technologies are still stuck in research mode. But yeah, again, I don’t want to be really negative because there are a lot of amazing teams working on those types of things and microneedles, and any focus on biowearables, any focus on multimolecule monitoring is good.

Josh Clemente: It’s like we’re going to learn things. We’re going to find out where the dead ends are. And I hope that eventually this will all converge into your sensor that can measure everything we need with minimal and/or noninvasive ness in the long term.

Ben Grynol: What Greek food is going to spike you when you hit the Santorini for your honeymoon.

Josh Clemente: I’m sure it’s some sort of pastry drenched in honey and some sort of goat cheese and balsamic glaze, I hope.

Ben Grynol: I think it’s going to be lemon potatoes, those will crush you. If you eat a pita, go pork, double meat, extra tzatziki because all the pitas over there come with fries in them so you can’t not get the fries. You have to get the fries, but that’s going to be your best mechanism and then head for a hilly run Santorini.

Josh Clemente: That’s a good tip. My mouth is watering.