Groundbreaking Results from 29 Continuous Glucose Monitor Experiments

As engineers turned health mentors, we’ve spent the last decade reading the research and self-experimenting with our health, and reporting our results in our newsletter.

Continuous glucose monitors are a new technology in the last 3 years, but only recently were they were approved for use without a prescription for those of us health optimizers who are glucose curious.

Alex and I were extremely excited to test these, learn about how our bodies respond to meals, stress, and exercise in real-time, and decide if they’re a helpful tool for people trying to optimize their nutrition and metabolic health.

Many health experts are raving about how elevated glucose is the cause of all chronic diseases and how using a CGM can help you “flatten the curve,” or eliminate foods that spike your glucose (and, therefore, lose weight!)

Sounds pretty straightforward, right? Well… it’s more complicated than that. Do we agree with this approach, recommend using a GCM, or think this latest health trend is a healthy one? We’ll share all our data, results, and takeaways from 29 experiments in this article.

What We Learned (Good and Bad) From Our CGM Data

We were also excited to test out many of the glucose hacks we see on Instagram, like how a shot of vinegar before a carb-filled meal could cut glucose spikes in half. There are tons of others like this too.

As engineers, we love collecting data on our health and running experiments, which we’ve been doing since 2017! Over the past several years, health tech has made huge leaps. Our hope for this experiment was to add another game-changing tool to our personalized health arsenal as we chase longevity and living our healthiest life.

Historically, data collection in the health realm has been quite basic and imprecise. study on the Lumen metabolism tracker. We’ve used them since 2023 and they’re eye-opening, indispensable, and surprisingly affordable.

However, even with the Lumen, you have to blow into the device to get results. These Continuous Glucose Monitors (CGMs) have sensor embedded in your arm and it’s continously (well… every 15 minutes at least) reading your glucose levels. That means nearly instant feedback on how that food, movement, or stress affected your glucose levels.

It’s pretty cool, and in this one experiment, we’ve basically collected more data that we have in all the experiments we’ve run for the past decade! Which, was honestly a bit overwhleming. More on that later.

Let’s dive into this ocean of data and see what we can learn together. We mean that… this is still new tech so please share questions, ideas for future experiments, and any personal experience you have in the comments.

Our 5 Conclusive Results and Exciting Findings

We’re going to jump right into our most conclusive results. And I specify “most conclusive” because, frankly, not all of the results synced up with what the experts say, or were just downright conflicting.

To give you a quick visual, below is a table of all the experiments we ran with color coded results.

Green –  Conclusive as expected (5)
Orange – Inconclusive or negligible (5)
Red – Conclusive but not expected (4)

We’ll talk about our inconclusive results in the next section. And if you’re interested in the finer details like the experimental design and controls, we’ll have a section for that too.

Finally, I guess it’s as good a time as any to remind you that every single person is different! The more we research, the more we genuinely believe that there’s no one-size-fits-all human diet, but that we each react to food individually.

Still, we hope our results will help you decide if this is something you want to try yourself. No matter what, you will learn something if using a CGM.

Here are a few results from the 29 experiments we ran as two healthy, non-diabetic people in their late 30s optimizing their health.

1. Food Order: A Dramatic Impact on Glucose Spikes

One of our most eye-opening findings came from experimenting with the order in which we ate our meals. We had a salmon bowl with rice, avocado, and veggies, and we tested eating the veggies first, followed by protein, fat, and carbs versus reversing the order.

What we found was striking: when we ate carbs first, followed by fat, protein, and veggies, we both experienced a more than double glucose spike.

The below graph is showing the change in glucose levels after a meal, so the difference between our highest peak and our level before the experiment. As the graph shows, this was a very conclusive and dramatic result.

Interestingly, this pattern aligns with the eating habits of many of the healthiest cultures in the world. For example, many traditional Chinese meals begin with vegetables, followed by protein and fat, and only have rice at the end of the meal if they’re still hungry.

After controlling the cooking method, meal time, weight of each food, and the amount of time taken to consume them, our results support the idea that the incredibly simple act of starting a meal with vegetables and saving the carbs for last can lower the glucose spike by an average of 67%.

Total mind blow! Of course, we don’t have a ton of data here so we’re looking forward to more experiments in the future.

Here’s a cheat sheet, in food order to eat: vegetables, protein, fat, and carbs.

2. Post-Meal Activities: The Power of Muscle and Movement

We also tested how different post-meal activities impacted our glucose levels. For this experiment, we had the same breakfast—pre-weighed oatmeal—and followed it with four different activities: nothing, a leg workout, a brisk walk, and a guided meditation.

Here is peek at the graphical results before we breakdown our findings below.

The leg workout had the most significant effect, leading to a noticeable reduction (an average of 39%) in glucose spikes. This was expected since strength training tends to speed up glucose utilization in the body.

A brisk walk post-meal also showed a small (an average of 11%) reduction in glucose spikes. Interestingly, for Alex there was no difference between walking and doing nothing. Maybe these results help explain why she had her most transformative fitness results from a bodybuilding program?

The meditation experiment, on the other hand, was a bit of a surprise. Despite our best efforts, the 20-minute guided meditation outside in the morning sun caused unexpectedly and unexplainably high glucose levels.

This finding was perplexing, as meditation is generally seen as a stress-reliever and often touted as beneficial for lowering blood sugar. Meditation is even part of some diabetes management protocols (to a small extent, but still) since gratitude regulates the physiological response to stress and stress directly impacts metabolism and blood sugar levels.

This is because, when stressed, our bodies release hormones like cortisol and adrenaline, making it harder for insulin to do its job.

However, in our case, it appeared to be worse than doing nothing after a meal. We’re still unsure why this happened, and it’s an area we plan to revisit in future experiments.

3. Eating Speed: Does It Really Matter?

Another hypothesis we wanted to test was the impact of eating speed on glucose spikes. We compared eating a large bowl of pasta within 5 minutes versus eating the same bowl over 30 minutes. Unfortunately, Alex’s sensor had an error in the first experiment so we repeated it with a gyro. 

For me, eating quickly led to a higher glucose spike, which made sense as rapid eating and less chewing means a more substantial digestive load and insulin response. However, with eating pasta, a food notorious for spiking insulin, I was expecting a far more drastic difference than 19%.

By the way, it’s really hard eating 300 grams of pasta in 5 minutes!

We should have just repeated the pasta experiment, but it’s not a very enjoyable or typical meal for us. So we repeated the test with a gyro and the results were the opposite of what we expected. Both of us had an equivalent or higher glucose spike during the 30-minute conversational meal.

A retest was in order and we ate another gyro in 30-minutes, but this time we were really distracted because we were running late to an event and getting ready. Surprisingly, this added stress and distraction significantly increased the glucose spike by over 100% for Alex and 25% for me! 

This result highlights the importance of eating mindfully. Like the Blue Zones says: never eat standing up.

Based on our data, if you’re distracted or rushed during meals, you will experience a higher glucose response. Also, and we’ll test this more in the future, but the rate at which you eat might not be as important you think. This was a relief for me because I tend to eat quickly. 

Lastly, when I say ‘based on our data,’ I understand that our data is not perfect. We learned pretty quickly that even though these CGM sensors are accurate, the body is far too complex to run a controlled experiment. Therefore, we had to be very careful about what results we trusted, and what results we didn’t. More on all of this later.  

4. Fake vs. Real Foods: Our Least Favorite Results

As health advocates who’ve been researching nutrition and traveling the world to better understand cultural links to health, we’ve come to one main conclusion: Ultra-processed foods (UPFs) are the devil.

After living outside the US for a decade and witnessing how simple a healthy lifestyle can be in Asia, Europe, and South America, we see it clear as day every time we return to the States. The cause of the health epidemic in the US is the food. Period. Plain and simple. 

So, we were particularly interested in comparing the glucose response to real foods, designed by mother nature, and fake foods full of man-made ingredients.

In the first experiment, we compared pizza from a local artisanal restaurant (using high-quality, fresh ingredients mostly imported from Italy) and the other from Dominos (full of highly processed and artificial ingredients).

Surprisingly, the artisanal, handmade pizza caused significantly higher glucose spikes in both of us, despite being made with what we considered far healthier ingredients.

Similarly, when we compared ultra-processed white bread (Bimbo) to freshly baked sourdough bread from a local bakery, we found that the glucose response was an average of 45% lower with the UPF bread.

This is a bit disheartening because a lot of CGMs are marketed like, “Flatten the curve to lose weight!” and clearly, you can’t always get caught up in the curve. Of course, the artisanal counterparts offered far better nutrition, despite the glucose spike.

This experiment wasn’t perfect, but the data acts as a reminder that CGMs are just a helpful tool, not a total solution to eating the perfectly healthy diet.

The good news is that we ran  more experiments after this one which gave us the results we were hoping to see.

The first, was that a simple oatmeal breakfast produced a lower spike than the equivalent weight of Lucky Charms. That’s right, for the first time in probably 30 years, we ate Lucky Charms.

They weren’t magically delicious, did nothing to satiate our appetite, caused bloating, and then made us crash very, very hard. With the oatmeal, we felt perfect.

Fortunately, “liquid obesity,” as our colleagues in China called it, met the same destiny. 

We tested fresh squeezed orange juice and Coke. While both are sugary drinks, the impact of processed sugar on blood glucose was more intense than that of naturally occurring sugar in fruit juice.

The difference was especially dramatic for Alex, and she couldn’t even finish the 12 ounces of coke because it was too gross. Really, if you drink that stuff, you need to quit and find an alternative.

5. Stress: Doomscrolling Spikes Glucose

We saw firsthand how stress can dramatically affect glucose levels. On Day One, Alex’s sensor was not working yet. I was stressed because I had spent the last couple of work days planning these experiments. 

As you can see below, the experiment calendar was jam packed, with only a couple of days at the end to repeat experiments. And we were already behind schedule on Day One.

Okay, definitely not the end of the world, but when I started scrolling the web I read that sometimes sensors don’t ever work, and often they fail early. I was pouting on the couch, catastrophizing about how the whole experiment would fail, when I got a notification on my phone. 

My glucose was at 142 mg/dL and steadily rising. Whoa! I didn’t just chug a chai latte, this had to be stress induced. So, I grabbed my guitar and started singing songs to our 3-month-old. It was! Over the next 30 minutes, my glucose levels turned around and dropped to 119 mg/dL.

This underscores the importance of managing stress to optimize health. After job burnout in my 20s gave me the testosterone of a 70-year old man and a stress-induced mineral imbalance made me chronically fatigued in my 30s, I understand the impact of stress on our physical health.

If CGMs can show users a physical response to the stress they’re currently experiencing, people will stop doomsday scrolling and start prioritizing self care.

Our Lessons Learned Using a Continuous Glucose Monitor

After wearing our continuous glucose monitors 24/7 for 15 days, and over a dozen experiments later, here are our takeaways.

We’ll get into the pros and cons, but one of the biggest one is that things aren’t that simple. The curves aren’t always easy to read, and so much can influence the results like stress.

We tested what we thought would be home runs, according to the glucose gurus we follow on social media… and so many didn’t work out as we thought.

Is it because we’re not pre-diabetic? Is it because the science is still too new, or that everyone is so uniquely individual? Maybe. Maybe not.

To be fair, there are millions of people using glucose monitors out of necessity and with absolute success. I don’t want to put any doubt on that. But does it work for healthy optimizers like us?

We did learn that our bodies respond differently to the same foods during the same experiments, so perhaps glucose monitors help get very personalized information, even if it’s mind-boggling.

Now let’s get into specifics.

Lesson 1: The Personalized Results are Key

As engineers and lifelong health optimizers, we were already aware that individual responses to food and activities can vary widely. This study confirmed that idea. For example, Alex and I responded very differently to the exact same foods and activities.

This personalized feedback is one of the main benefits of using a CGM. Instead of relying on broad, generalized health advice, you can observe how your body specifically reacts and tailor your diet and lifestyle accordingly.

For example, the following graph shows how we both responded to the same breakfast, lunch, and dinner. 

And here are Alex’s results on the same day. You can see she if far less glucose sensitive than Ryan, never exceeding the recommended 140 mg/dL limit.

Lesson 2: More Data Could Mean More Confusion

While the CGM provided a wealth of data, we found that constant glucose monitoring without understanding the full context of each data point can be overwhelming, confusing, and in some of our experiments, misleading.

It’s easy to get lost in trying to understand every fluctuation. But, we absolutely do not recommend that. Honestly, the majority of the data that we collected was not useful because we couldn’t replicate it or it didn’t make sense to begin with.

In a perfect world, we’d like to see a CGM that pairs with an app where you can journal your physical feelings, mental state, meals and daily activities to provide richer context. Then, AI can scrub all that information to find replicated trends, and prevent the user from drawing too many incorrect conclusions.

We’ve heard that this is possible, and that some AI-based software can even cross-reference your gut microbiome test to put together a full picture of what your body reacts to. 

While Ryan’s average glucose seemed higher than it actually was (according to blood testing at least), the data and glucose changes are accurate. Still, we think the body is way to complex to always draw helpful and accurate conclusions. At least it was for us.

We know this, because we ran very controlled experiments: weighing food so the meals were identical, eating them at the same rate and time of day, and controlling our post-meal activities. Still, our data was mostly unexpected and not replicable.

Maybe that’s because we’re not necessarily the target market for these sensors as we’re only glucose curious. As you can see below, over the 2 weeks we wore the CGMs, I was within range 94% of the time and Alex was 99% of the time. That means we’re doing pretty well, and far from being pre-diabetic.

This is a good time to share some examples of why you can’t always draw conclusions based on your CGM data. Some experiments that we thought were guaranteed to be successful, failed or yeilded the opposite of what we expected.

The first is our hydration experiment. According to this data (I swear I controlled everything I possibly could) being hydrated is bad for glucose spikes, which is not what the science says.

Then, there was our fiber experiment, where we used a juicer to remove the fiber and compared it to the same fruits and veggies blended in the Vitamix (with all the fiber). The difference was negligible, which is surprising. Fiber is definitely better for glucose regulation, per the science.

And finally, our intermittent fasting experiment, where a 16 hour fast (compared to a 10 hour fast) produced higher spikes and a much longer glucose plateau. Again, against what the science says.

Lesson 3: If It Sounds Too Easy to Be True, It Is

There are a lot of tips out there for reducing the glucose spike. For example, taking a shot of vinegar before meals, meditating after a meal, refrigerating rice and pasta before eating, etc.

We ran these experiments, and the difference was negligible. To us, that’s just a reminder to focus on the big picture rather than getting distracted with little tips you see on Instagram.

Lesson 4: Flattening the Curve Isn’t the Holy Grail

Many CGM proponents promote the idea of “flattening the curve” as a way to optimize health and prevent diseases. While reducing glucose spikes is important, our findings suggest that it’s not the only factor in overall wellness.

Intermittent fasting, for example, produced higher glucose spikes in our experiments. But, in the health world these days, fasting is considered, almost infallibly, a healthy practice. And from our intermittent and extended fasting experiments, we know that extended fasting and intermittent fasting are beneficial for our metabolism.

The focus should be on a holistic approach to health, not just minimizing glucose spikes.

Plus, if everyone starts wearing these things, how easy is it going to be for the big food companies to engineer the food to produce a lower, less dramatic spike? Piece of cake. Honestly, they could easily find a corn or soy-derived ingredient to make even a piece of cake appear flat.

About the Experiment

We used the Stelo CGM by Dexcom. We don’t think it’s the best CGM (the Dexcom G7 seems to be the best) but at the time it’s what was readily available without a prescription.

Applying the sensor was straight forward, but for Ryan, his arm was sore for 24 hours after and sporadically was sharply painful throughout the two weeks. Based on our research, this is quite rare, and probably just because he has a very low body fat percentage. Alex had no problems with hers and couldn’t even feel it.

To get the best data possible, there were several things that we had to control.

• Food weight: Obviously, every single food item had to be weighed.
• Preparation method: Since some people say that cooling and reheating pasta or rice can lower the glucose spike, all of our all of our food was cooked and eaten hot.
• Meal duration: We controlled how long the meal was so that we didn’t eat too fast.
• Water intake: Hydration was avoided 30 minutes before, after and during meals.
• Meal time: In each experiment, we had breakfast, lunch, or dinner at the same time of day.

Of course, there’s a lot of uncontrollable variable, especially with an infant in the house. So, how did we do? Share your thoughts and questions in the comments. We’re excited to do more testing!

Closing thoughts (for Now) On Continuous Glucose Monitors

Using a CGM and planning 29 meal experiments over two weeks was one of the more intense things we’ve done. Honestly, it took a lot of effort and more time than you’d think.

Often, we’d be waiting hours and hours after lunch for our glucose levels to flatten out to move onto the next experiment. This, in particular, was stressful for Alex who was breastfeeding our baby at the time and needed to be eating a lot of calories throughout the day to keep up.

So, it’s definitely not like those perfect curves you see presented by glucose gurus on social media, with the after-meal spike and dip soon after. It’s far more intermixed than that.

Still, we enjoyed using them a lot, and it was fairly easy and pretty affordable to get massive amounts of data in 15 days. Our recommendation overall is that it was interesting, and yes, we would do it again.

For us, it confirmed that our blood glucose levels were healthy and acting as they should, with no alarm bells. That already gives invaluable piece of mind right there (although you can figure that out from a simple blood test, too).

But when it came to better understanding which foods could “trigger” us and how to optimize our food and meals to be easier on our bodies, we honestly only became more confused. We didn’t walk away from this experiment declaring, “Ryan should avoid pizza but ice cream is okay!” like we were hoping. Or, “Ryan CAN eat the pizza as long as he drinks a bottle of water first!”

It wasn’t like that.

Perhaps, though, that would be oversimplifying things and missing the point. And the point is that, thank goodness, we discovered that we’re fortunate enough to eat just about everything without throwing our hormones into overdrive.

So, what will you find out?

Let us know in the comments what brought you to this article and if you’ve tried a continuous glucose monitor before! A good place to keep tabs on our journey and future experiments is by signing up for our newsletter, and better yet, to consider joining our membership where we have dedicated discussions and calls about topics like these. Good luck!

Written by Ryan Gleason
CEO and Co-Founder at Ryan and Alex Duo Life