Continuous Glucose Monitor

[Scott 2]

I've been wearing a Dexcom G7 continuous glucose monitor for the past 3 days. My doctor was able to get me a free 10 day sample from Dexcom. I'm not diabetic, so the numbers are relatively stable. Some observations:


1. The biological system is more resilient than I thought. After 3 hours on a bike, I purposefully waited a couple hours to eat. My blood sugar stabilized in the high 80's anyways. I feel emboldened to ignore the sense I'm "so hungry" after working out. I'll be fine.

2. Sometimes, my blood sugar will dip as low as 70. But it also recovers if I wait.

3. When I ate a meal with 100g of carbs, the blood sugar spike was just a blip. When I ate a meal with 200g, there was more of a long hill, that took several hours to decline. I suspect that's less desirable. I often eat only twice a day, because it's easier. I might get back to eating 3x a day.

4. There's some lag in how quickly the biological system responds to inputs. Maybe 15 minutes or so. It makes sense, but a lot of things I thought were changing my blood sugar, were in my head. That frees me from worrying about things that don't matter.

5. Walking around almost immediately takes 10-20 points off the reading. It makes more sense to me why fitness trackers prompt for hourly step quotas. I tried to see what sort of impact an immediate light walk has on that big meal hill. Does the hill rebound as soon as I stop? Or does it stay depressed? So far - it looks like the former. So a post dinner 10 minute walk blunts the peak, but doesn't truly flatten the curve.


6. It's interesting to see nutritional advice play out in real time, in my body. I had steel cut oats with pecans and blueberries for lunch. Also some lemon pepper roasted broccoli, olives and tomatoes. My blood sugar spiked to 141.

Dinner was a massive bowl of rice and beans. About 50% more food. But also eggs and cheese. Sugar only got to 124. However - the curve took longer to flatten out back to baseline. I hear "always have protein with your meal." Now, I have evidence. It's cool.

I think consciously blunting that curve can offer a positive cumulative effect. Seeing it in my body is a powerful teacher. Rather than believing certain things on faith, I know now them to be true. It feels different.


Overall - For a healthy person, there's nothing here that couldn't be accessed via sound nutritional advice. The value stands out for someone with blood sugar issues. It becomes trivial to see if any patterns throw you out of range. Or to learn if certain behavior modifications are not doing anything. And maybe it helps teach what high and low blood sugar feel like. One G7 patch yields a reading every 5 minutes, for 10 days. That's a lot of potential learning.


The tech has been annoying. My phone isn't officially supported, and I didn't want to buy a receiver. I had to download an "unofficial" version of the app, found via reddit, and then side load it into my phone. The device also uses a separate app for creating reports too, which feels dumb.

Keeping connected to the phone means staying within 20 feet of it. Data still accumulates when a signal is lost and uploads on the next sync. But I find myself fiddling with signal loss issues more than I expected. I am constantly wanting to check it. As someone who likes to disconnect, this is a major downside. There's also some effort to track meals / exercise within the app, if you really want to watch for patterns. I don't mind that, but some may.

Application felt a little nervous. There's a device that sticks it to you via button press. But before you do the deed, you can see the roughly 1/2" needle it uses. I had my wife do it, haha. The needle doesn't stay in your skin, just a filament. It burned for maybe 15-30 minutes after.

The filament used by the patch is super thin. I can't feel that. But the adhesive is heavy duty. I do feel that when moving around. It's like having two stacked quarters under a large bandaid. I put all my weight on it in bed last night and didn't enjoy it. I'm mostly nervous I might dislodge it and ruin my fun experiment. Though from what I've read, Dexcom support often replaces the device when an issue happens.

I can't see myself using a CGM continuously. With my body self stabilizing, there's not much action to take. Given the hassle and expense, I have better returns available to me. But I can see spot checking my patterns every year or two. With GoodRx coupon, cash pay for a 3 pack of 10 day patches is about 170. That's cheap for the data.


If I was concerned about potential pre-diabetes or diabetes, I would start here. Screw the fasting blood work. This is cheaper and far more informative. Yes it's a hassle, but totally worth it in that context.

[mountainFrugal]

Jokingly my XC skiing friend and I talked about getting one of these and setting a spiked level baseline by eating an entire box of Cinnamon, Toast, Crunch in Chocolate Milk.

More seriously it would be really interesting to see how it works after different levels of exercise and consuming different "recovery" foods that I normally eat. In the same way that you can start to get a feel for different heart rate zones, I wonder if you can get a feel for glucose levels? It is likely only on the extremes that you would feel something though initially.

[Scott 2]

More seriously it would be really interesting to see how it works after different levels of exercise and consuming different "recovery" foods that I normally eat.

I have been bounds testing. Yesterday was 90 minutes of weights and mobility. I got home and immediately ate a 1200 calorie pint of ice cream. Sugar spiked to 131, then slowly climbed down. I was surprised how muted the response was.

But, that's been one of my key lessons. If your body is insulin sensitive, it absorbs the metabolic assault. However, you may also be bathing in massive quantities of insulin. Over time, that's not great.


I spent a few hours learning how the tool is used for health optimization. I now see more value in the CGM for a healthy person. A disadvantage of the stock app, is it is setup for diabetics. It lacks the integrated health optimization features found via something like Nutrisense or Levels. I've gained new perspective:


1. The health optimizations ranges are much tighter and lower. I'm barely within the high end of normal. Levels publishes their guideline. Note that CGM readings do not perfectly mirror finger prick readings:

https://www.levelshealth.com/blog/what- ... mate-guide




2. The most immediately actionable optimization, is blunting the post meal glucose peak. I've got work to do:

2.1 Activity timing is the biggest lever. The worst thing you can do is eat a big meal, then sit. That is my pattern. Even something as simple as standing or doing the dishes will dramatically blunt the curve. So save the chores for after the meal. Don't "earn" a post meal nap, by doing all your work first.

I put this into action today. It's fun to pick-up donuts on the way home from a long bike ride. Today, instead of carrying them home, I ate at the donut shop. 3 donuts and 16oz of whole milk. Then I biked the last 4 miles, showered, etc.The highest my blood sugar touched was 120. It was back down to 99 an hour later. That's after well over 100 grams of high sugar carbs.

I'm taking this change. I'll make a point to have breakfast before morning training, since the activity will blunt that peak. And save my chores for after my meal. I don't really care when I eat. It's nearly free health gains.


2.2 Don't eat naked carbs. Add fat and protein to the meal. I've already encountered this. The highest sugar peak I've seen, has been from steel cut oats. Not ice cream or donuts. Another easy change I'll keep.

2.3 Ideally, eat fat and protein first. Create a buffer to slow digestion of those carbs. Rolls as an appetizer is terrible, from this perspective. I'm not sure how well I'll commit to this one. I like my greek yogurt and berries as dessert, for instance.

2.4 Carb dosing. The more you have, the harder it is to absorb. This is an area where I think the CGM can help find bounds. Doing it without seems very difficult. I might get away from 1 or 2 giant meals per day, because of this. That's a significant lifestyle change for me. I dunno if I'll adjust per meal carbs based on upcoming activity. That might be more ideal, but also a hassle.

2.5 Identify your high glycemic foods. The claim is our individual response to specific foods varies. So I might be sensitive to oatmeal but barely respond to corn chex. I might as well eat corn chex. The CGM seems necessary to puzzle this one out, maybe for an extended period of time. I'm not sure I'll chase it.

2.6 Don't drink a lot of liquids with your meal. The idea is thinning out your food with water makes it digest faster, which causes a bigger blood sugar spike. It's common for me to have 16-40 ounces of water and tea when I eat. I dunno if I'm willing to change that. I may test having my tea before breakfast, instead of after. I don't know if that's enough separation to matter.

What I don't know, is if training optimization would look different. Are there times where athletes want the post-meal blood sugar spike? I have no idea. It is certainly not what the health optimizers are looking for.


3. Another big challenge, is bringing down fasting glucose levels. This one's a little tougher to puzzle out. From what I can tell:

3.1 Keto does it. Because you're also running on ketones and need less blood glucose. I'm not eating keto though.

3.2 Maintaining general healthy patterns helps. Exercise. Sound diet. Low stress. Sleep. Etc. For the most part, I'm at my tolerable limits with these interventions though. So I feel a little stuck here.

3.3 You can mess with supplements and medicine (ie metformin). I'm not prepared to play that game.

3.4 I wonder if regularly blunting the post-meal peak will gradually make me more insulin sensitive, ultimately bringing down fasting numbers. I'm not sure. That's the best idea I have for action here.


4. There are expected blood sugar responses to exercise, based upon intensity. I'm not prepared to change my exercise intensity for blood sugar games, but I can see that would make extensive use of the CGM. Some sources say high intensity exercise will make the body more insulin sensitive, which could help bring down fasting glucose levels. But again, I already have other priorities for my activity.


5. My last take away, is that the CGM data informs sleep quality. Poor sleep wrecks the blood sugar curve and decreases insulin sensitivity the day after. There's a desirable curve over the night's sleep. Ideally it holds relatively steady and any changes are gradual. During REM sleep, the curve gets a little lower. If one sees spikes and hills during their sleep, that's a sign something is going wrong and needs to be addressed.

I'm not sure if a 10 day CGM cycle is enough to pick out bad patterns here. It's certainly not enough to resolve them. I have had some potentially questionable nights. Addressing that looks hard. I've been mulling over the potential value of a sleep study. CGM data could inform the decision.


I'm happy with my little experiment so far. I'm surprised to find so many potential action items. And how many assumptions I held that were simply not supported by the data. I'm aware there are limits to the device, especially around precision at the low end. I'm also conscious of the potential for developing disordered patterns. Imagine using glucose tabs and a standing desk, to achieve a perfectly level curve, as an example.

There's a school of though that this is frivolous use of valuable medical equipment. I see and acknowledge that perspective too. Consider the value for a parent with diabetic child, who's trying to keep tabs on things. Now they can have real time updates, on their phone, from anywhere in the world. Game changer. Instead, I'm using that patch to measure my over-consumption of donuts and ice cream. Yikes.

[Lemur]

There is a lot of logic to avoiding naked carbohydrates and eating them with protein/fat. The latter slows gastric emptying, in other words, slows the process of digestion down. Its this same reason that eating whole fruit would likely cause less spike then eating fruit juice with the former having the sugar bound to fiber and slowing down digestion. Sometimes people may be surprised to find that ice cream can create a smaller spike then something like grapes for that reason.

I think its neat that the CGM can validate this stuff outside of a lab. Individual differences are also fascinating. I'm betting in a group of 100, if you have all of them eat a bowl of white rice, 75% or so will see the spike and the other 25% won't. Do the non-responders have genetic advantages? Or do we also need to control for health measures and age groups? Because a metabolically healthy person, and likely younger person, could probably handle that white rice better being more insulin sensitive overall.

[DutchGirl]

Thanks for this, Scott, it makes an interesting read, even if n is just 1.

I think this was valuable for you to create a more healthy lifestyle for yourself, which in turn could lead to much lower health costs down the road. So an investment of 10 days and $100-$150 or so now could lead to savings of ... who knows how much.

On the other hand, please don't overdo it either. Don't forget to enjoy the long life that you're setting yourself up for this way.

By the way, I've seen some people who feel like they have spontaneous hypoglycemias even though they can't be objectified with the normal medical tests for that. For them it could also be nice to carry this type of device for a week or so and see if it can show them a relationship between glucose levels and how they feel.

[Scott 2]

Don't forget to enjoy the long life that you're setting yourself up for this way.

By the way, I've seen some people who feel like they have spontaneous hypoglycemias

Agreed on not overdoing. So far, I think my time return will be a net positive. Simply finding which long held assumptions are true/false, in my body, eliminates a good bit of mental overhead. For all the reading I've done on the importance of blood sugar, this small experiment has been far more informative. I see this as exactly the Medicine 3.0 strategy Attia talks about in Outlive. Also it's fun to me. I like playing with this sort of data.

My hope is to adopt some new patterns that mesh with my lifestyle, then run them mindlessly.

For what it's worth, I still cannot feel the difference between my blood sugar at 70 vs 140. Maybe I'm starting to feel when it trends up and down? I feel relatively in tune with my body, but 'm really not sure. That makes me much more skeptical of people's blood sugar claims.


Contextual perspective has become much more evident as well. IE:

1. In tools for a diabetic, my numbers are so far removed, they don't even register for action. They're an impossible consideration.
2. For someone managing pre-diabetes, my numbers are the goal. Achieving them would be a success story.
3. For the health optimizer, my numbers are trouble! I'm nearly pre-diabetic. Emergency!

It's very important to consider which audience education is targeted at. A big portion of the consumer grade CGM push, is really focused on catching undiagnosed diabetics and pre-diabetics. Extremely valuable, but not the standard I am playing against. I can't compare to someone guzzling Coke by the gallon, then claim I'm crushing it.

[ertyu]

I don't think this is relevant to you @Scott 2 but adding this for completeness: it's a good idea to keep in mind that blood sugar impact isn't the only dimension to the "healthiness" of something. E.g. ice cream might have a lower blood sugar impact than grapes, but it also contains emulsified fats, conservants, and other industrial chemistry that wreaks havoc on gut health: damages the villi and causes leaky gut, etcetera. Eat food, not too much, mostly plants is still the gold standard.

I do find the timing experiments very interesting. For instance, I seem to remember that in the "lean gains" diet, the recommendation was to do omad, take branched chain amino acids before your work-out, workout, then eat *after*. An acquaintance that was into bodybuilding once lectured me on the importance of eating a banana immediately post-workout, too. It's very interesting to see that when it comes to blood sugar, the optimal effect is to eat before exercising. Jives with the common practice in south east asia, which is to have dinner at around sunset and then go for a leisurely walk at say 8-9 pm. A lot of this is because it's simply too hot during other parts of the day so there is a lot more activity in the streets during late evening than you'd see in the west, but it's interesting to see that it dovetails with optimal meal/movement timing for blood sugar regulation.

[Smashter]

+1 to the activity timing and its relation to blood sugar being interesting.

For years I have been doing 20 pushups immediately after finishing meals I eat at home. This was based on the idea that doing a physical activity post-meal would shunt blood sugar into the muscles and reduce insulin spikes. I have no idea how true this is, but my fasting glucose was 76 last time I tested, which is in the the optimal range according to the Levels chart posted above. There are lots of confounding factors in my case, as I exercise a lot in general and walk about 15k steps a day.

I am tempted to run some Dexcom tests, but for now am happy to live vicariously through @Scott2. Thanks for the detailed writeup!

[DutchGirl]

I have one, or maybe two, questions.

The dubious one is: what does chocolate do? A good portion of high quality good Belgian milk chocolate.

The other one is that people are often told to fast for 8 or 10 or 12 hours before having blood taken. I'm curious what your glucose levels do when fasting, whether levels stabilize at some point and whether that's after 8, or 10, or maybe a surprising different number of hours.

[Scott 2]

@Smashter - That's a much better fasting glucose number than I've been seeing. It's a little surprising, to be honest. I thought between all my exercise and eating a vegetarian diet, that my numbers would be excellent. Nope.

When I did my blood work last, a1C was 4.9% - with an estimated average glucose of 94. My fasting glucose was 100. But I was training for a 10k then, with loads of cardio. The CGM gives an average glucose in the low 100's so far, which is just below pre-diabetic. Also of note - the fasting glucose test I had with my blood work, considers 109 at pre-diabetic. Which rule set do I trust? Why is the doctor's office using more forgiving metric?

It doesn't help that I keep eating terrible things, to see what happens. My first 5 days were highly exploratory. I'm now making some improvement seeking changes, over the next 5. I hit the grocery store this morning for test foods. For the sake of convenience, I've been sucking down a lot of breakfast cereal and sourdough bread lately. My thinking - It's low sugar, I just need calories. Not quite.

So I'm going to do the extra work to get a veg with my breakfast, replace one serving of grain with nut butter, have berries instead of a banana. More hassle, but small changes in the grand scheme. We'll see if they have much impact. I'm curious to see what a typical smoothie does. A whole wheat couscous bowl. A quinoa bowl. Etc.

The dubious one is: what does chocolate do? A good portion of high quality good Belgian milk chocolate.

In short - it depends just how much chocolate, how much sugar vs. fat, and what you do after eating it. Mid bike ride, even 3 donuts didn't hit that hard.

One of the foods I want to test, is a bar of dark chocolate from Aldi. 160 calories, 8 grams of sugar. My go to "healthy" treat. I don't expect it to do much. There's fat to slow down digestion. I also tend to savor my chocolate. I think eating a piece of sourdough bread would do more.

The other one is that people are often told to fast for 8 or 10 or 12 hours before having blood taken. I'm curious what your glucose levels do when fasting, whether levels stabilize at some point and whether that's after 8, or 10, or maybe a surprising different number of hours.

Time for the system to stabilize is highly individual, based on the person and their actions.

The system behaves like a taught rubber band. A fasting state is one with minimal vibrations. Events (eating or exercising) introduce energy into the system, causing fluctuations. Peaks and troughs in blood sugar result. They can resonate with one another. The more and stronger the events, the longer the fluctuations last.

Fasting recommendations need to consider the worst case scenario. Let's say you ate a pint of ice cream at 3pm, then lifted weights at 5pm - like I did last night. Chaos ensues. Stability might not occur until 2 in the morning. What if I was also strongly pre-diabetic? I might need a full 12 hours to be fasted.

Alternatively, I had some cereal and protein shake at 6am today, then did yoga at 8am. By 11am, my blood sugar looked stable.

You can see both patterns on the 24h graph:



Some screwy stuff happened after my yoga class at 8am too. I have not totally figured that out. Why the climb up to 126 by 9:30? The drop happened while grocery shopping. But 30 points? Why so fast? My head started to hurt a little. What the heck?

Was it a symptom of yesterday's chaos? Did yoga speed digestion of my breakfast? Is that simply what it looks like to move with moderate intensity and then sit? I don't know. I think understanding could require months with the tool, vs. my 10 day window.


Before the monitor, I incorrectly visualized the system mechanically - like the fuel tank in a car. Eat sugar, levels go up. Exercise, levels go down. I assumed a linear response. It's not that simple. Which makes sense, because it is a biological system. They exhibit elasticity, like a spring.

[Stahlmann]

https://optimisingnutrition.com/data-dr ... art-guide/

[DutchGirl]

I wonder if you can create an anticipatory rise in insulin and thus a drop in glucose by imagining you're going to eat sweets. Maybe that was why the drop happened in the grocery store: your body was getting all kinds of "Ooh, we're going to get some good stuff, soon!".

Anticipatory hormonal responses are discussed here:
https://www.sciencedirect.com/science/a ... 3421001470

[Dream of Freedom]

It could be cortisol. A rise in stress between yoga and a busy road and shopping experience would release glycogen from the liver.

[Scott 2]

@Stahlmann - That's an interesting program. Have you tried it? If so, any findings?


Turns out the spike after yoga is a standard response to exercise. Insulin goes down, allowing glucose to rise. The muscles use it, keeping levels relatively steady. The harder you are going, the stronger the effect. Once you stop, the body continues releasing glucose for a bit. After returning to baseline, you are more insulin sensitive for the rest of the day.

https://www.signos.com/blog/blood-sugar-spikes-exercise


Some observations over the last 24h:



1. My biggest spike so far, 50mg in 30 minutes. The source? A bowl of whole wheat couscous, with spinach, mushrooms and 3 eggs. Grapes for dessert. I suspected that might happen, but wanted to test. It has been a common meal for me. Quite the glucose excursion.

Breaking it down - I had 3 servings of tiny, low protein grains. Relatively little protein and fat, just the eggs. And more sugar from the grapes. 125g of carbs in total. I also camped at the computer, because it was late. Probably not great for my sleep.


2. I planned a 750 calorie breakfast to minimize blood sugar impact. Despite 90g of carbs, barely a blip. And I sat for almost 90 minutes after eating. Whey protein, peanut butter, kale, balsamic vinegar, 2oz of corn chex, a cup of milk, 1/2 cup of blue berries, tea. This is a strategy I used successfully yesterday too. Front load the meal with a veg and fat. Go heavy on the protein. Keep grains near 2 servings. Use a lower sugar fruit. I even had my tea last, with no impact.


3. Great visual of the exercise induced glucose spike and subsequent insulin sensitivity. I ran, breaking into Zone 3 towards the end. The harder I tried, the higher my blood sugar went. However - the 2nd bigger peak didn't come until I was already resting inside.


4. I had a smoothie around noon - blended fruit, whey protein, milk, almond butter. Fairly balanced nutritionally, but fast digesting. I wasn't sure how it would hit. The blip was small. Sometimes I throw 1/2c of ice cream in there. I bet that would substantially increase the impact.


My lessons from the experiment are consolidating:

1. Switch from eating 2x a day to 3x. The meals for 2x a day are simply too big. I cannot both respect my blood sugar and eat enough.

2. My regular meals can be easily adapted, to minimize the glucose excursion. The crux of this is described above. Start with a veg, include 1-2 fats, 1-2 proteins, limit grains to 1-2 servings, favor lower sugar fruits. It will cost more and takes more effort.

3. Dinner timing and contents deserve special consideration. Blood sugar bouncing around hurts my sleep. As much as I like to eat a big dinner, I think small (or early) is the better strategy. Something like today's smoothie might be ideal. That requires front loading breakfast and lunch.

4. When a meal will cause a large glucose excursion, move afterwards. Less necessary, if the meal has been planned well.

5. There's a deeper tier of optimization available. Plan glucose excursions around movement intensity and recovery. The current strategy - minimize glucose mean and variance - is crude. One can do better, but I'm simply not ready. Once my new behaviors are normalized, a subsequent round of CGM could be warranted.

[Scott 2]

Another small take away. I read that alcohol doesn't have much impact on blood sugar, unless one is binge drinking.

I haven't drank in the last 6 months, but I was curious. Check out the steady curve - 2 hard ciders at 5% abv. Each is 23g of carbs, 16g sugar:

[Scott 2]

The 10 day trail finished. I firmly believe my data indicates insulin resistance. Despite a healthy a1c of 4.9%. I requested another month of patches and will continue with changes described above. After a month of practice, I'm going to run another 10 day CGM. Repeat until findings stop.

The hardest part is designing meals that limit glucose excursions. That's the primary benefit I expect from subsequent patches. However, one thing is undeniable - I simply don't know what the numbers will do. I have to test them.

This may be the most valuable medical data I've ever had. Low cost. Immediately actionable. Extremely short feedback cycles. I am 100% sold.

[ertyu]

What about the data indicates insulin resistance? I'm not asking in the spirit of arguing, rather, I'm missing the step between the pattern of the data and how it leads to the conclusion. What would data from a non-insulin resistant person look like in comparison?

[Scott 2]

What about the data indicates insulin resistance?

Average glucose at 100. I'd like it around 80.

Post meal glucose spikes of 50+ points, for basic healthy foods. One example - Cheese, crackers, olives and grapes. I want that limited to 30, minimal as possible, really.

A cheat not directly related to the monitor. I have fasting insulin and fasting blood sugar numbers from recent blood work. Individually, they look fine. Together, they can be used to calculate a HOMA-IR metric. Mine is 3.05, indicating possible insulin resistance.


Keep in mind, this is all in the context of 10+ hours of weekly exercise, plus a relatively low calorie vegetation diet. I thought I was invincible! Not quite.

[DutchGirl]

Might some of this be caused by genetics? You might be genetically more susceptible to diabetes type 2 than other people? (And thus suffer from mild insulin resistance right now even with a decent lifestyle - life's not fair).

Weight, fat percentage and muscle percentage, are also important. Exercise is definitely good, so keep doing that.

I always learned that having more muscles could help. I can't remember whether you mentioned anything about strength training, but if you're not doing much of that you might consider upping it. Apparently during the exercise it can increase glucose levels, but the longterm effect is beneficial. Strength training is also good for general health of course.

[Scott 2]

@DutchGirl - could be. All I can do is make the most of it. Agreed on the other factors. I do strength train.