How Gut Bacteria Affects Blood Sugar - with Dr. Momo and Dr. Casey | Empowering Neurologist EP. 139

When we know how food is impacting our glucose levels and which foods and ingredients specifically trigger these high glucose spikes, we can start to modulate that a little bit. We can start to learn which foods actually keep us more stable and flat, which is what we want. We want to be more stable and flat in terms of our glucose levels throughout our lifetime.

Hello everyone, I'm Dr. David Perlmutter. Welcome again to The Empowered Neurologist. You know, there is probably nothing more important to know in terms of your biometrics than what your blood sugar does after you eat, your postprandial glucose measurement. You know, as that number starts to increase and we become more and more resistant to insulin, it really does set the stage for just about every bad bad thing to happen to you that you don't want to happen, like increasing your risk for coronary artery disease, cancer, diabetes, obesity, and even Alzheimer's. So we really need to know what is happening in terms of our metabolism of the foods that we are eating, especially as it relates to our blood sugar. And we would like to know with respect to our insulin levels as well. So we are going to be spending some time looking at 2 really different technologies that ultimately, I think, arrive at the same goal, and that is to develop a personalized approach to understanding how foods affect ultimately our blood sugar and therefore what the ramifications are of the food choices that we, we make. We'll be talking to Dr. Casey Means, and Dr. Means has been on the program before. You'll remember her from Levels. That is a way of looking at our data that we obtain from wearing a continuous glucose monitor that really gives us, as she calls it, a closed circle feedback in terms of the events related to our blood sugar, the foods that we eat, certainly, but also other activities like when we exercise, how much we exercise, what time of day, how our sleep is, how much sleep we get, what is the quality of that sleep, how much stress we are experiencing. How do all of these variables over which we have control, how do they affect our blood sugar? And we know this moment to moment when we're wearing a continuous glucose monitor and we can input this data into the Levels app and then really gain a much better understanding of how we are controlling our blood sugar. We're also going to be looking at a very new and exciting technology that is brought to you by a company called Viome, V-I-O-M-E. And what Viome is doing in their research is trying to gain an understanding about personalized nutrition, originally looking at the gut bacteria, but as we're going to learn today, looking specifically at the metabolites, the actions of our gut bacteria, and how that is reflected in terms of our blood sugar moment to moment. What is the effect on blood sugar of the metabolites, what the bacteria within us are actually actually making in terms of how those metabolites are influenced by bacterial gene expression. Now, that sounds a bit complicated, but we're going to unpack that for you today on the program. Let me tell you a little bit more about our guests. Dr. Momo Vujicic is a co-founder and chief science officer at Viome. He's an entrepreneur scientist whose passion is to build a healthier future in which chronic diseases and cancers are covered, as he In the history books, not TV commercials. He's used his extensive scientific expertise and business acumen to lead the development of the core Viome technologies. He obtained a PhD in chemistry from the University of Utah and a BS in microbiology from the University of Texas at El Paso. He is also an adjunct professor at the New Mexico Technology University. Dr. Casey Means, you'll recall, has been on the program before. She's a Stanford-trained physician and the chief medical officer and co-founder of the metabolic health company Levels. She's also the associate editor of the International Journal of Disease Reversal and Prevention, and her mission is to maximize human potential and reverse the epidemic of preventable chronic diseases by empowering individuals, by giving individuals tech-enabled tools that can inform smart, personalized, and we're going to talk about today, and sustainable dietary and other lifestyle choices. She's an award-winning biomedical researcher with past research positions at the NIH, National Institutes of Health, Stanford School of Medicine, and NYU. Dr. Means' perspective has been recently featured in the New York Times, Men's Health, Forbes, and many, many more. And I had the opportunity of writing, co-writing a recent op-ed in MedPage Today, which was actually a letter addressed to President Biden about the importance of the messaging of around sugar in the American diet. So we are going to have a really incredible program today. I can't wait to jump right in. Let's get to it. Well, welcome, friends. This is really great, and I'm so glad that both of you could make this. I want to start off because I think central to the work that you're both doing, although Momo, you know, I think yours is certainly more widespread beyond just glycemic response. Casey, you're really focused on glycemic response and certainly other things moving forward. But I'd like to ask you both, starting Casey with you, why should we care what our blood sugar does after we eat a meal?

Yeah, it's such an important question and now more than ever because we are dealing with truly epidemic levels of metabolic dysfunction in our country. As you know, as you talk about so often, 88% of American adults with at least one biomarker of metabolic dysfunction. And the reality is, is that our post-meal glucose levels have a huge impact on whether we're going to develop metabolic problems down the road. When we're spiking our glucose levels, and food is a key determinant in whether we do spike our glucose levels, when we do that to a high magnitude and frequently throughout the day and throughout the course of years and decades, We're really setting our body on this glucose roller coaster that has so many immediate effects on our health, whether that's triggering inflammation, oxidative stress, glycation, but then also the long-term ramifications associated with developing insulin resistance, which we know is related to really most of the main killers in the United States. So many of the leading causes of death, like we talked about when I was on your Yeah. In the podcast earlier, Dr. Perlmutter, we talked about how 9 of the 10 leading causes of death in the US are related to blood sugar, high blood sugar, and insulin resistance. So this is something we just really need to care about. And what's so interesting is that we have so much control over this. When we know how food is impacting our glucose levels and which foods and ingredients specifically trigger these high glucose spikes, we can start to modulate that a little bit. We can start to learn Which foods actually keep us more stable and flat, which is what we want. We want to be more stable and flat in terms of our glucose levels throughout our lifetime. And, you know, we know that there are so many things that affect that— meal composition, whether we move after a meal, how much sleep we're getting, and many other factors that I'm sure we'll get into today. So really, the big picture is preventing downstream metabolic disease by controlling our— Blood sugar. glucose levels day to day and over time.

You know, there's so many things that come to mind. This is the week where a new so-called Alzheimer's drug was received FDA approval. And again, what you're talking about is looking at the front end, is looking at how do we remain healthy. And yet there seems to be, you know, this mentality that we're going to live our lives come what may, and then we're going to have an Alzheimer's drug and or a a drug for our diabetes, coronary artery disease, whatever the downstream effect of this persistent elevation of blood sugar is. But before we leave the topic, because it's gonna segue very nicely into the paper that we're gonna be talking about today that Momo authored, the beauty of this CGM, or continuous glucose monitoring, is rather than us attaching ourselves to broad stroke recommendations for individuals, though we can, in terms of what they should or shouldn't be eating. This is a technology that allows the individual to understand his or her particular glucose postprandial response and therefore to make better decisions moving forward.

Yeah, that's exactly right. This technology, continuous glucose monitoring, which I'm wearing right now, It's the first time ever, ever that we've had closed-loop biofeedback on nutrition. For forever, we've basically put stuff into our mouth and then just sort of had this no real feedback on what it was doing to our body immediately. And I would say that throughout human history, that was probably an okay way to go because we were primarily eating whole foods, eating close to the earth. We were eating food in the way it was supposed to be Mm-hmm. Intended to be eaten. Now, over the past 50, 75 years, we're not eating really anything that looks like food. The vast majority of our calories are coming from ultra-processed foods, food-like substances made in factories that our physiology and our metabolic machinery has no idea what to do with. I mean, average American eating 152 pounds of refined sugar per year, our mitochondria just have no idea what to do with that. It's orders of magnitude more than what we were designed to process. And so, Because we are now facing this new and really clever, monumental sort of threat of essentially the modern food system and what it's doing to our physiology, having tools like a continuous glucose monitor just gives us so much more empowerment and agency in terms of understanding what these substances are doing to our bodies and being able to make informed choices. And I think what a lot of us find who wear them is that we actually move away from the processed food and actually towards whole foods with fiber, with healthy whole food fats and protein, you know, with carbohydrates that are surrounded by the actual whole food in this sort of protective way. We find that our glucose on these, on the CGM monitors actually looks a lot better. So it really reinforces some of what we, I think, know intuitively, which is that eating Whole foods, natural foods can be a lot better for our overall glucose responses, our metabolic health. But yeah, we're facing kind of this really interesting new cultural norm of being surrounded by packaged food that is just wreaking havoc on our blood sugar levels. It's designed essentially to be absorbed quickly, and that those carbohydrates are so refined that they can just get in and cause that spike really quickly. So having this knowledge and this closed-loop biofeedback is is just really knowledge is power in this situation. And then knowing how specific foods are affecting our specific body and having that real insight for the first time into biochemical individuality, it's just, it's really a game changer. And when we've got 72% of Americans who are overweight or obese and 128 million Americans with prediabetes or type 2 diabetes, having some tools on our side can really be a good thing.

And yet the editorial in the Journal of the American Medical Association that was published several days ago, I emailed you about that and you had mentioned that they actually wanted to, or they did interview you, really I think made the point that, well, we should only be checking these in diabetics. And the point is that this is how you don't become a diabetic in the first place, 'cause you get some understanding as to what is leading to blood sugar elevation. So that leads me, Momo, to you. And you just completed, and you're gonna tell us about another study that I didn't see yet where I can't wait to hear what you have to say. But a really interesting study that as Dr. Means was just talking about, that there is an individual, we learn about our individuality in terms of our blood sugar response to a given meal, given type of meal. You studied 550 individuals, adults, and a total of something like 30,000 meals. And you were able to specifically categorize those meals. And you learned that it wasn't— you didn't study the individual's genetics. So it wasn't like you were looking at how an individual may be genetically programmed to have a particular response, nor did you look specifically at the bacteria, but you looked at the metabolites of what our gut bacteria, those gut bacteria of those individuals, were producing. And you really came upon some really, I think, empowering findings. So maybe just to start off, you can walk us through the nuts and bolts of the design of that study. Then we'll talk about what you found and then we'll move on to why it's important.

Yeah. So I agree with everything that Dr. Amin said, but I would like to add that we are actually now at Viome, This was our focus about 2 to 3 years ago and our customers have been benefiting from that. But really, I would like to highlight that we firmly believe, and there's plenty of literature evidence, that really the type 2 diabetes or prediabetes are really sort of the last stages of a disease, a metabolic disease that starts actually years or decades earlier. So what we are focused on at Viome right now, and that's one of the focus areas of the preprint that we just published is that the microbiome and the nutrition, the interaction between the microbiome and the nutrition actually create a condition where you first have fasting insulin go up. The next stage will be fasting glucose goes up and then the HbA1c goes up. And so really that's sort of the third stage. And so what we are focused on is we want to prevent that fasting insulin from going up in the first place, which is probably the very first event that happens. And so, so what we just published is a paper where we actually developed a machine-learned model for HbA1c based on 50,000 study participants. So it's a very powerful machine-learned model, state-of-the-art. And then what we did is we actually enrolled 2,500 participants in a longitudinal, so prospective trial, where basically we observed them for close to a year and we computed this type 2 diabetes risk based on these 50,000 data points from a training set. And then we applied that risk to these 2,500 people, and half of those people did not follow the Viome diet well as self-reported, and half of them did follow it well. Those that followed it well, actually a year later had a 30% reduced risk. So it was sort of a really— we're getting to the point where we're generating data that have to do with causation on a very large population and a molecular risk that has not yet been reflected in an elevated HbA1c, not yet. And so we are very, very excited about this. But to just back up to the CGMs and so on. So So this started at Viome about 4 years ago, and my wife and I were experimenting with CGMs. And since we ate the same meals at that time, we didn't have a precision Viome diet. We wore these CGMs and then we looked at the graphs. And one of the first things that really struck us was sweet potatoes and potatoes. We had completely the opposite response, right? Potato— they're called potatoes, so you would think they're bad. And if you look at the glycemic response or glycemic load of those 2, They're basically— glycemic index is what it's called. They're basically the same for the average human. However, with us, she can eat all the potatoes she wants and she basically— her glucose does not reach 100 and she can eat like a pound of mashed potatoes. And if I eat that, mine goes up to 180. And so that was crazy. And for sweet potatoes, mine only goes up to about 110 or 120 and hers spikes to like 165. And we did this multiple times and it was very reproducible. And that was really striking how the same food can have such a profoundly different effect on 2 different people. And so then we launched this study at Viome. And so yes, you summarized that study nicely. So we enrolled 550 people, we collected their stool samples prior to, just prior to the study, and then 2/3 of their meals over the next 2 weeks we provided to them. We actually had 10 people shop at Whole Foods and we spent $350,000 at Whole Foods buying food. Wow. and handing them the food. And so 2/3 of their meals were given to them and 1/3 they prepared, but they took pictures of them and they recorded exactly what they put in them. So we developed a machine-learned model based on 30,000 meals, and all of these people wore CGMs for the entire 2 weeks, right? So we had very accurate glucose responses. And so we built a machine-learned model that says multiple things. First of all, microbiome plays a very important role in determining how you're going to respond to, to the same food that other people respond differently. Second, very important thing is that it's, it's not just individual food. You— so you can't have— rice doesn't have a glycemic index. Rice has a different glycemic index depending on what you eat rice with. So depending on how much fat you're consuming with rice and how much protein you're consuming with rice and what types of fats and proteins, That same exact amount of rice can spike your blood sugar very differently. So you can't look at— no one eats rice for a meal, so it's sort of pointless to look at individual foods. And this is really where it's really— it gets really challenging in using CGMs. So one of the earliest things we thought was, hey, well, let's provide our customers with a CGM for 2 weeks, we'll record the data, and we'll tell them what to eat. It turns out context, the entire meal, matters so much that that We just, we just, we just, we couldn't do it ourselves. And so we developed this machine learning model that now not only identifies the glycemic index for each food for each person, but does that in the context of the other foods. And so now when a Viome customer gets their recommendations on their app, it computes that entire thing and it tells you what to eat and what not to eat. So it tells you to avoid, let's say, quinoa because it's gonna spike your blood sugar. But it also tells you to avoid foods that that could spike your blood sugar with other foods that you're okay to eat. So if you could, let's say you let's say the computer will say you can go, you can eat potatoes, but it it takes out other foods that can make the potatoes spike your blood sugar even more. So it's a very complex algorithm, and it takes all these features into account. And you don't have to, you don't have to do the you know a normal person doesn't have to do this machine learning and algorithm interpretation. So that's really where we are now. But like I said, we're now shifting backwards in time to really focusing on fasting insulin to try to see— because here's, here's a simple fact, right? You can have a person who has a normal insulin response drink a liter of soda every day for 3 months, right? And their response will be the same after 3 months, meaning Yes, it's going to spike, but it's going to go back to normal and it's fine, right? But then you have a person who is— who's developed insulin resistance, and their blood sugar is going to spike way more than that, right? And so if you have a healthy response, you know, we evolved to eat a lot of fruit, right? And so we evolved to eat a lot of sweet fruit, but that's natural to consume sugar, and that's why we have the insulin response to put it away. and not affect the rest of your physiology. It's now when we develop insulin resistance that that sugar stays in the bloodstream for longer and spikes higher and has these negative effects that Dr. Means mentioned that are really disastrous. And so if we can maintain a proper healthy insulin response, we can then consume more sugar, not excessive amounts like most Americans do, but I'm saying, you know, now they're saying don't drink any juice, don't eat any fruit, you know, like avoid all sugar. And that's, I think, bad advice because we need fruits.

Well, I think there's a lot of data out there that indicates that the big issue as we talk about fruit would be the sugar and the concern, of course, being the fructose. And I think with both fruits and vegetables, it doesn't really seem like that is, A, a risk factor for diabetes, and B, ultimately going to have an effect even more proximal to that in terms of insulin resistance, and even more importantly, as far as what I'm looking at now, doesn't seem to spike uric acid levels. And therefore, I think the notion of fruit with its attendant fiber, vitamin C, hopefully some quercetin or other types of phenols that can regulate or can have an impact on the genesis of uric acid by helping to inhibit the important enzymes, xanthine oxidase, that it can allow lower levels of uric acid and therefore make— obviously, it's what your mom told you probably these days seems to make the most sense. Let me say one thing. Actually, I wanna say a couple things, if I may. I'm glad I've introduced the two of you because this looks like a match made in heaven in terms of this information. And second, I have to go back to the striking news, at least for me as a neurologist this week, the approval of this so-called Alzheimer's drug because it fits in very nicely with what we are talking about. We're talking about being preemptive. We're talking about being proactive, not reactive. We're talking about what happens way before your blood sugar starts to go up and how we can keep that from happening. And as a matter of fact, it matters a whole heck of a lot for me as a neurologist because we know that becoming a type 2 diabetic may as much as quadruple your chances for becoming an Alzheimer's patient, and I will go on record even now as saying there is no pharmaceutical treatment for that disease. The reality is that the drug was, on its good day, downstream with a tailwind, showed that it slowed the decline. Well, we're not interested so much in that. We'd like to reverse Alzheimer's. And now we know that a more personalized approach Dr. Means, I introduced you just last week to Dr. Dale Bredesen, and he's reversing Alzheimer's. I mean, it's a wonderful time to be alive to see our paradigms being so challenged by this personalized approach that really, I think, defines what both of you are doing. It is really about, and the empowerment part, giving people data then that they can make changes. to be so preemptive and preventative in terms of these devastating conditions. I get back to you, Moma, just for a moment, and that is this, the idea that you looked at metatranscriptomic data. And I wonder if you could just sort of unpack that a little bit. Yeah,

let's do that.

In comparison to, I think a lot of people are, do have some fundamental understanding of the RNA categorization of their gut bacteria, but you were not as involved with that in the study as you were looking at the product. So why don't you walk us through that?

Yeah, let's talk about that. That's a really important piece. So yeah, so about 10 years ago when I was a scientist and I was looking around at studying microbiome and studying, using technologies to study the microbiome, I have a PhD in chemistry and I understand that our physiology is driven by chemicals and chemical reactions. It's not driven by names. In other words, let's take a very simple concept. Butyrate is a very well-known metabolite produced by the gut microbiome. One of the cool things that most people don't know is that humans have evolved to depend on butyrate as a food source for colonocytes. These are the cells that are lining the intestines. Those cells are actually consuming a byproduct of microbial fermentation as a food source. We're using that as a carbon source. That's an amazing phenomenon that we have— we are so in such symbiosis with these microbes over the course of evolution that we've basically— we've become one organism. So we need that butyrate. Now, butyrate is also recognized by our immune system as, oh, everything in the gut is healthy. There is no infection. Butyrate is being produced by the friendlies, and therefore I need to calm down and take a chill pill, right? I need to reduce the inflammation. There is no emergency. There's no alarms going off, right? So these are chemical signals. My immune system and my colonocytes know nothing, nothing about the name Faecalibacterium prausnitzii. Okay, the fact that humans have named a microbe Faecalibacterium prausnitzii and that it codes genes and codes genes for producing butyrate, that means zero to my physiology. The only thing that matters to my physiology is Is butyrate produced? Okay, and it doesn't in fact have to come from Faecalibacterium prausnitzii. It can come from Clostridium difficile. It can come from any number of other bacteria. That is the that is really we need to shift our focus from studying the microbiome and giving things names. So if you walk into somewhere and your your teeth hurt and you need a dentist, and you say I need a dentist, and someone says. Well, we have a Joe and a Samantha and Momo here. You're like, I don't care what their names are. I need a dentist. Who is the dentist here? Well, we don't know because we don't study their functions. We just study their names. That doesn't help me at all because I need the function. And so about 11, 12 years ago in my lab, we started using these DNA-based tools for studying the gut microbiome, and I realized early on that I'm not gonna get any answers that I'm seeking. And now, fast forward, my team of brilliant scientists spent 6 years developing the technology before we started the company. So we didn't just look around and say, hey, let's put some pieces together. But here's the main point. The main point is that if you do the fanciest DNA analysis today called metagenomic analysis, all you're doing is sequencing the DNA. What you're telling someone is, You have a microbiome that has hundreds of species of bacteria. They encode 2 million genes with every function known to man, right? And they— and you say, wow, you have butyrate producers, you have bacteria that are associated with good serotonin, you have all these things. And then the simple question of, are they actually doing those functions for me? Well, we don't know because the data doesn't tell me that. All you can measure is the potential that DNA is a physical object that is there in space and it doesn't actually have any function. It codes for a function. But our data tell us, aha, of those 2 million genes that your microbiome encodes, only 50,000 are actually expressed. So meaning they're active, they're functional, they're doing something. And because we're quantifying them, we know exactly the level of expression. So now we can say, well, you have no expression of butyrate-producing genes, so butyrate is absolutely not being produced. Or we say you have mild or moderate or really strong expression of butyrate genes, so your butyrate is being produced. And we can do that for hundreds or thousands of bacterial or microbial metabolites. And that's really what we need to shift our focus from, calling names to identifying the functions. That's sort of the core technology that's important. But there are many layers on top of that. And I'll just list them, and we can delve into details if you're interested. Another extremely important piece of the puzzle is when we took— when others take depressed people versus healthy people and do a study, they say, you know what, these microorganisms are a little bit more enriched in depressed people than these. Well, simple question is, do depressed people have a different lifestyle and different diet, and could that be the reason, right? Is that really related to depression, right? Because when we did a study on depression, we didn't study tens of people or hundreds of people. We enrolled 11,000 participants in our clinical research study because that's how complex depression is. And 9,000 of those did not have depression, and 2,000 did have depression. And then we developed a machine-learned algorithm that not only was developed on such a large number, it was independently validated on an independent cohort. And so this is really phenomenal. I can't wait to publish this. If we take a randomly picked stool, someone freshly pooped somewhere, and we pick that stool and analyze it, we know nothing about the person who provided that stool sample, nothing. We don't know who they are, where they live, what their age is. nothing. We have 80% accuracy in determining whether they're depressed or not from our functional microbiome test. That's remarkable that 80% of the determinants for depression are actually found in stool. That's amazing. But—

No, I'm just gonna say it really is amazing. And to get back to your point about looking at the specific bacteria, I mean, you mentioned Faecalibacterium prausnitzii and for the past decade or so, we've seen lower levels compared to controls in patients, for example, with inflammatory disease of the gut, autoimmune conditions, including multiple sclerosis. And so the thought was, gee, maybe we can give them some Faecalibacterium prausnitzii and then all will be well. But I think you've refined that whole notion to be the actual gene expression that we should be thinking about.

Exactly. Plus what people also need to study is not isolated microorganism that Faecalibacterium prausnitzii actually lives in a complex community and it actually cross-feeds. And so there was an elegant paper that just came out showing that for it to produce butyrate, a specific species of Bacteroides has to be present and be fed a certain diet to be producing acetate. And then acetate as a byproduct of this Bacteroides is used by Faecalibacterium prausnitzii to make butyrate, right? So it's not just one organism. We need to get away from that infectious disease hypothesis that we've been driven by for 100 years— one organism, one disease, right? That doesn't exist with chronic diseases and with the microbiome. It's a complex community with tens of thousands of functions. And food is not food the way we see it with our eyes. Broccoli is a sack of hundreds of different molecular ingredients, and so is every other piece of food. And when we chew it, We basically combine all of those ingredients into one mixture, and that mixture of 10,000 different molecules hits this microbiome that has 50,000 different functions. And now all these complex chemistries happen— direct feeding, cross-feeding, some absorption into our body, control of our own physiology by the gut microbiome, processing of these nutrients into secondary metabolites or these byproducts by the gut microbiome, and then flow of those across the intestinal barrier into our bloodstream. So it's a very complex thing. We can't be thinking, oh, let me just give you one probiotic and you'll be healthy. That's just such a simplistic view and it doesn't happen, unfortunately. I wish it did.

Well, it does, I think, get to your point earlier in that the probiotics in and of themselves may not have an overwhelming biological activity, but do influence functionality of other organisms. In fact, we know that in tagged studies, there really isn't much recoverability of those probiotics that an individual may take, but we know that they do influence gene expression, for example, in other residents— resident organisms. Yes. So it— and then getting to another point, this notion of eating the broccoli. I mean, it really affirms this notion of food as information. And that when we eat a particular food, it changes the genetic expression of our resident bacteria, which then influence our gene expression as well. So it really does tend to close the loop. I'd like to go back to Kasey, and if I could, what do you think of this information and what might you think about in the future in terms of what you could do with it?

Yeah, I mean, it's really, paradigm shifting. It's incredible work that you're doing, and it's exciting to hear about even these pre-publication results that you've been finding. I think something, you know, really interesting that you mentioned was this idea that insulin levels are going to become dysfunctional long before we see rises in fasting glucose. And I think a paper I saw from The Lancet suggested that even up to 13 years before we saw sort of clinical changes in fasting glucose levels, which is, of course, what we're picking up in the doctor's office, we might be seeing signs of insulin resistance and hyperinsulinemia. And so such an important thing. You know, I was— one thing I was curious about in your paper that I thought was fascinating was that increasing BMI and increasing waist-to-hip ratio seemed to be associated with a decreasing postprandial response. So as BMI— I mean, as expected, I think we saw that as AUC— let's see, what was it? I don't have the figure in front of me, but as hemoglobin A1C went up, area under the curve went up. But it wasn't the same with BMI and waist-to-hip ratio. I was curious if there might be something there related to insulin levels. Like, as people go from— we know that insulin levels tend to go up as BMI goes up and waist-to-hip ratio goes up. So if we're kind of catching people in that hyperinsulinemic phase there, where they might be a little bit hyperinsulinemic, that's actually potentially driving glucose down a little bit in that very early stage of insulin resistance. But I was curious what your thoughts were on that finding.

Yeah. Let me tell you the actual truth. The actual truth is that we have many, many programs like this one on depression and IBS and anxiety and many other diseases. And unfortunately, we're not an academic institution to where every interesting finding we can follow and dive deeply into to try to tease that apart. So we have now built machine-learned algorithms models for about 30 different chronic conditions. And each one of them has given us hundreds of different molecular features in the microbiome. And also, I don't know if you guys know that we do a blood test because we look at the human body as an ecosystem, also on the human gene expression side. And so we now have literally 10,000 different molecular features in the microbiome and the human genome data, gene expression data. That are of interest. And there's no possible way for us to have enough time and money to follow up on all those. So what we are really, really focused on is converting that knowledge into something that can help improve people's lives. That is our number one objective. And at some point in time, we may establish scientific collaborations with like academics, and they can perform all these studies in detail to tease apart the mechanisms. mechanisms, right? So unfortunately there's just not enough time. So we have shifted, like I said, to now understanding the mechanisms that drive insulin resistance and increase in fasting insulin. And we're trying— and we're building diets, we've already built diets that can modulate the gut microbiome function to prevent that. And so this year we're excited to launch 2 different clinical trials where we can hopefully show We're gonna test and see what happens, whether we can prevent the rise of fasting insulin, right? And so, so these are the kinds of fundamental things that that we're wanting to do. And then, you know, we we also focused on depression and IBS and anxiety. I don't know if you guys realize how bad depression is. I mean, it is to me. It is it is the worst of all diseases. Not only is it the most prevalent and such devastating disease to others, to so many people, but it has so many downstream. problems. You know, it disables people from having a life, social life, professional life, and so on. It just completely occupies the entire society. And so we really need to address it. So we now— so this preprint that I'm really excited about that just got posted today on Research Square, it's under review in Nutrition Journal. It's basically we're showing the efficacy of our diet, precision diet and supplements, on the symptoms of depression, anxiety, IBS, and this type 2 diabetes risk. And we show a reduction in clinically validated scores by 35 to 40% for severely affected individuals with nothing but what you can buy at the grocery store. So there were no drugs, there were no mushrooms from China, there were no, you know, sort of a basement kind of a thing. Just literally you can walk into Whole Foods or Trader Joe's and buy everything that we recommend to people, and if they do that, their symptoms are reduced significantly. And that's really the paradigm shift that I would like everyone to understand is that food is medicine. It's, it's really that simple. You just have to figure out what foods are gonna feed your microbiome such that the byproducts of that interaction are gonna be good for your physiology and maintain your healthy homeostasis and avoid foods foods that if you consume and give to your microbiome, and I emphasize your because everyone else will be different, your microbiome is going to produce now metabolites or byproducts that are going to tell your body something is wrong. I need to increase inflammation. I need to eat more. I need to, you know, I need to have more anxiety. Something is wrong. So if we can have people avoid those foods, that their microbiome is going to convert into pro-disease metabolites, then we should be able to maintain a healthy homeostasis forever.

Momo, I want to really be very clear about what it is that you're saying. When our viewers hear that there are specific foods that can, that may be effective, and the numbers you quoted, I think were pretty astounding in comparison to SSRI medications, which we recognize have maybe minimal effectiveness at best, and the scope of depression, anxiety, and mood disorders in general in our modern world. When you're talking about specific foods, going to Whole Foods, Trader Joe's, whatever it is, and getting those foods, that can— and it can be helpful. I'd like you to really just spend a little bit more time talking about how that is not a generalized food recommendation. These are not the anti-inflammatory anti-inflammatory foods that you could write a book about, and then everybody goes and eats them and feels better, that this is personalized nutrition, a branch of personalized medicine. Please spend a little more time on that.

Yeah, so the way that our— so everything at Viome is driven by chemistry and mathematics, right? Literally everything, 100%, not 99.9%, 100.00% is chemistry and math. So a stool and a blood sample come from a customer, They're analyzed in a clinical lab, and we measure the chemistry of the microbiome and the chemistry of the human genome, meaning the gene expression levels. That's measured chemistry. Those chemistries are now used to apply mathematical equations. Here's the stepwise process. First step is to quantify the activities of the microbiome for every single type of activity, production of this, production of that, lack of production of this. And then the computer takes our food ontology database. It's a database of all foods that we normally see at the grocery store translated into 20,000 different molecular ingredients. And so the computer uses those ingredients and asks the question for this particular person. So it does this for every single person. For you, I've just analyzed your microbiome. I know that if you consume these ingredients, your microbiome is going to convert them into molecules that our machine-learned algorithms tell me from our big studies on IBS and depression and so on, that they're going to induce some kind of a disease process in your body. So I'm going to put those on the avoid list. These other molecular ingredients— remember, I haven't mentioned the word broccoli yet. We're not working in the broccoli space. This is molecular ingredients. the types of polysaccharides, the types of fats, the types of all kinds of molecules. Now, the computer says, okay, but if I give you these foods, these molecular ingredients— sorry, not foods, molecular ingredients, then your microbiome is going to convert them into beneficial, healthy molecules for your physiology. That's the first order. If we could one day, maybe in 100 years, our computer will be able to see send that recipe to a company and the company will be able to produce a shake for you that only contains the 10,000 molecular ingredients that are perfect for you and avoid every single one of the other ingredients. We don't have that and that'll probably taste like crap, so probably not a good idea anyways. But then the computer does the next thing, which is, okay, now I'm going to map these ingredients to the foods that a person buys at the grocery store and I'm going to do the best. It's called a global optimization algorithm. It basically says, okay, I want to find the foods that are going to give me the biggest benefit and the least harm. And so it finds broccoli and it says, aha, broccoli only contains beneficial chemicals, ingredients for you, so broccoli is superfood for you. Then it goes to spinach and says, oh, spinach has something good like vitamin K2, but it also has oxalates. That's not good for you. So now I'm conflicted. What do I do? Okay, well, let me find— let me put spinach on your avoid list and let me find other foods that can replace K2, right? And if I can do that, great. If I cannot, I'll tell you to take K2 as a supplement, right? And so the computer does this iteration but does it thousands of times because every time there's a new conflict, it has to resolve prior conflicts. So it's a supercomputer that runs in the cloud, and it basically makes that final calculation, which is it categorizes all foods into 4 categories from avoid to superfoods, and then it gives you a list of all supplements that you need to be taking and the amounts that you need to be taking. And that's— so that's all 100% objectivized. There was no human opinion anywhere in here. There was no anecdote. There was no experience. There's not— it's literally chemistry, and math, right? So that's really cool and it's really making an impact already, huge impact. We're extraordinarily proud. But what's really even cooler is that if you take a Quest test, let's say, or any other test, they're not learning anything from those tests, meaning the millionth customer is going to have the exact same information as the very first customer. For us, every single day, seven days a week, the AI wakes up in a supercomputer and says, "Oh, I have new data because we have customers from seventy countries." Right? I have new data, and it starts to it starts to learn. And every single day, it improves these algorithms and it adds new ones. And we have algorithms in our data that we don't understand. We we don't understand mechanistically. There's no biological explanation as to why that should be that way. But the machine-learned models tell us this really works. And they do these— I can give a 3-hour talk about this, but basically—

It's leaving us behind. That's amazing.

Yeah, so it's continuously improving. So every single new customer benefits from the vast prior knowledge. It's a spinwheel. You keep adding speed to it and knowledge. It's never linear where, sorry, that's all we know. That's really the power of our technology. We've created this platform that's self-learning and self-improving continuously, and we're very proud of it. Going back to your personalized diets, we have 250,000 customers in 70 countries. There are no 2 customers that have been given the same diet. There are 250,000 diets because our computers don't know anything about the Mediterranean diet or the ketogenic diet. That they are not aware of that. That doesn't exist in their algorithms. The only thing that exists is here's the chemistry I measured, here's the chemistry that I need to support, and here's the chemistry I need to deactivate because it's bad for you. And I'm going to deactivate that chemistry by withdrawing the molecular ingredients and foods that your microbiome needs in order to activate that chemistry. And I'm going to provide you with the chem— with the foods that have the molecular ingredients that are gonna activate the beneficial chemistry for you. That's how our computers think. How cool is that?

I'm speechless. It takes a lot to get me to that point. I'll tell you, Casey, the feedback we get via a CGM though also allows the— it really exploits personalized input, but it's really dependent upon that interaction with that individual, isn't it? But having said that, how is their data used for the collective much in comparison to what we just heard?

Yeah. So to give just the broad brushstrokes of what we're doing at Levels. So like I mentioned, this is giving real-time closed-loop biofeedback about how foods are affecting our glucose levels. And glucose is a powerful biomarker to Yeah. Like, you know, Momo was saying, it's interesting. Fasting glucose isn't going to rise until probably many years after you start having— developing insulin resistance. But something that we're learning from our members is that this paradigm we have of like your fasting glucose is, you know, you're 88 and that's who you are, that's actually really flawed. It changes day to day drastically. And having that type of information about yourself that I could have a fasting glucose of 75 on Monday and 88 on Tuesday based on the behaviors and the choices I've made each day, that's really powerful because what we start to unlock there is this realization that how I'm living and the hundreds of micro choices I'm making each day are actually very much changing sort of this number. That could be what foods you're eating, what time you're eating, whether you exercised, what type of exercise, how much sleep you got, what your sleep quality was, how you responded to stress during the day. You know, there's many, many factors. It's a very multivariate model that determines where you're at every day. But what you can sort of safely assume is that if you're making those choices that are putting you at those higher levels each day, whether it's your fasting glucose or your postprandial, your post-meal responses, if you're doing that day after day, you're essentially leading your body down a course where it's going to, unfortunately, adapt in a negative way and become— lead you down that road of insulin resistance. And one of the key markers other than just post-meal responses, area under the curve after a meal, or fasting glucose that we focus on is glycemic variability, which is a metric that we really have had very little insight into prior to the advent of CGMs, continuous glucose monitors, because we've been just using these snapshot measurements, fasting glucose, or averages like hemoglobin A1c, which of course looks at about, you know, 90 days average of glucose levels. But that doesn't tell us how much we're going up and down day to day, and that is a really important biomarker for risk of future disease and development of obesity and type 2 diabetes. So Yeah. That is something we can track and can be actually a really kind of early biomarker of whether we're going down a path that we don't want. Great research out of Stanford, Michael Shiner's lab, a couple years ago, a paper about what he called glucotypes, which was showing that as you increase your variability, meaning the ups and downs, these peaks and dips that you're doing each day, it correlates with your levels of fasting insulin. So even though we may not have a continuous insulin test yet, which it has not been technically feasible, looking at things like glycemic variability and how much of those ups and downswings you're having can be, in a sense, kind of a window into where we might be at. So putting all of that together with levels, glycemic variability, post-meal glucose, peaks and area under the curve after a meal, what your fasting glucose is, you know, on average, and how much it's varying day to day. When you take all of these things into account, you can just get a much more kind of granular picture of where you're at. And ideally, in the future, it may be that we can actually predict your fasting insulin levels based on all of these sort of complex metrics about glucose. Certainly not there yet, but there's a lot of research really sort of around that. So how does continuous glucose data streams basically correlate with fasting insulin, which we know is such an important biomarker? And fasting insulin is something that we are gonna be testing in our members actually as well to kind of give a better insight into the overall metabolic picture. So Yeah. Really, so much of it comes down to this personalization and just really thinking about the changes day-to-day and where we actually have control over that. So I'm very excited about learning about what you're doing, Momo. And I think that there's just— so the sky's the limit in terms of sort of the intersection of this research, like, for instance, how glycemic Variability relates to our microbiome metabolites and whatnot. So very hopeful.

Well, I am so glad that I got to introduce the two of you. This is very exciting, especially our viewers are right here when it happened. Let me say thank you to both of you and thank you for being here. But, and I think more importantly, thank you for all that you're doing, all the work that you're doing. And I know I can extend that from all of our viewers. And we will certainly have available to our viewers links to your respective companies so that they can learn about what you're doing and participate. So again, thanks, you guys, for being with me today.

Thank you, Dr. Perlmutter. Thank you, Dr.

Perlmutter.

Thank you, Dr. Means. Thank you. Good day.

Well, again, thank you for our guests today. And I think this is, uh, having 2, 2 guests like that to meet for the first time on our program, very exciting. I feel honored to have made that introduction. Look, uh, how empowering that, uh, that conversation, uh, was. And, uh, who knows where this is going to go. Uh, that said, we learned about levels and about the, the value of continuous glucose monitoring, and the website should be appearing on your screen right now. We also learned about this company Viome, which allows people like yourselves to learn about specific foods that you should be consuming, specific foods that you should be avoiding based upon information not just from you that you provide in terms of your genetic information and your gut bacterial information, but how that information is manipulated in terms of thousands and thousands of other people in terms of their response to the diets that have been implemented with them. And then that inform— the data is crunched through artificial intelligence, then you benefit from that. So what an interesting program we've had today. I'm Dr. David Perlmutter. Thanks for joining me, and we will be back soon. Bye for now.