Read the first chapter of real food for pregnancy for FREE.

CGM Experiment: What I learned as a non-diabetic from wearing a continuous glucose monitor

Yes, you read the title correctly. I do not have diabetes and I willingly chose to wear a CGM.

Overall, this was an eye-opening “CGM experiment” and this post covers what I learned as a non diabetic from wearing a continuous glucose monitor (aka CGM).

My CGM Experiment: What I Learned as a Non-Diabetic From Wearing a Continuous Glucose Monitor

I wasn’t sure I’d write much about it, but when I asked folks on my email list if they’d be interested in hearing about this CGM experiment (which I lovingly referred to as my semi-bionic arm), the response was HUGE.

I even polled my Instagram followers in my stories and 99% said “yes, tell me more about what you learned from wearing a continuous glucose monitor.” (Not sure why 1% aren’t interested, but I’m gonna go with majority vote here!)

I also received A LOT of specific questions about this:

  • Why did you choose to wear a CGM?
  • Which CGM did you choose?
  • What was your average blood sugar?
  • How did you get a CGM?
  • How much does a CGM cost?
  • How bad does it hurt to put in the CGM sensor?

And on and on and on. I received over 60 questions, so I obviously can’t answer them all here, but I’ll at least cover the biggest and most common questions.

Per usual, what I had envisioned as a short ‘n’ sweet write up turned into a bit of a research beast. I should know myself by now, ha!

Why did I wear a continuous glucose monitor (CGM)?

There are several reasons.

1. I’ve been curious about my blood sugar patterns ever since my interesting experience with taking the glucola (glucose tolerance test) during my pregnancy several years ago to screen for gestational diabetes—and my results from home blood sugar monitoring during the few weeks following that test. In short, I wanted to check on my non-pregnant blood sugar balance beyond standard labs.

2. Second, several mamas in my Real Food for Gestational Diabetes Course have used CGMs to better understand and manage their blood sugar levels. In one case, CGM readings revealed really interesting blood sugar patterns overnight and at fasting (first thing in the morning) that ultimately prevented misdosing medications that could have caused hypoglycemia overnight.

In another case, CGM data made a participant realize her blood sugar was spiking sooner and higher than she had anticipated after certain meals. With solely relying on fingersticks to check her blood sugar, she had been missing the “highest highs.” She was able to pinpoint what was causing her blood sugar spikes and better tailor her diet/lifestyle (which ultimately helped her stay diet controlled and have the out-of-hospital birth she had hoped for).

In other words, using a CGM was like having a window into their blood sugar patterns. It eased stress, improved their sense of control over what otherwise feels like a rollercoaster of a ride, and helped them better manage their gestational diabetes overall. As a certified diabetes educator, I’ve found CGMs cut out a huge amount of guesswork for clients and care providers alike.

3. Third, I had comprehensive blood work done a few months ago, which included a variety of metabolic markers. My HOMA-IR, a marker of insulin resistance, was almost too good (indicating favorable insulin sensitivity) and my fasting insulin and blood sugar were relatively low. While my results were overall very good, I was curious about what was going on with my blood sugar on a day-to-day basis.

4. Fourth, I have a family history of type 2 diabetes and reactive hypoglycemia. I know from all of my mindful eating observations that I feel best eating a moderately low carb diet, which probably is more kind to my pancreas (thanks to avoiding major blood sugar spikes). I know that heredity plays a role in your risk for disease, but also that lifestyle is a powerful way to moderate that risk. If we can catch any blood sugar issues very early on, we can take action to help prevent the progression to type 2 diabetes. If I inherited a crappy pancreas (no offense, pancreas… I don’t know for sure; I still love you man!), maybe I should take extra precautions?

If I can sum it up in one word CURIOSITY is what made me do this CGM experiment.

How do continuous glucose monitors work?

CGMs use a “minimally invasive electrochemical sensor” that’s inserted below the skin to measure blood sugar levels in interstitial fluid (that’s the fluid between your cells).

It emits a low frequency signal to communicate blood sugar data to a reader device. Different CGM systems vary in the specifics, but you essentially end up with a vast amount of data on your blood sugar patterns. Instead of relying on fingersticks and a regular glucometer, you can keep tabs on what’s happening 24/7 (without having to constantly prick yourself, hallelujah!).

How do you choose which CGM to get?

First off, this is not a sponsored post whatsoever. I had to pay out of pocket for everything (see below). I WISH the manufacturer would have donated one of these things to me since I’m about to give it a ton of free PR. The things I do for science!

I chose the Freestyle Libre for ease, functionality, and cost. Most CGMs require calibration with fingersticks using a glucometer (often with twice daily). Freestyle Libre is factory calibrated, so no need to poke your fingers all the time. This is both a benefit and a limitation of the system. It also tends to be less expensive than other CGMs.

It’s way beyond the scope of this post to discuss pros and cons of all CGM systems, so please take up that conversation with your care provider if you’re interested.

How do I get a CGM if I don’t have diabetes?

Continuous glucose monitors are by prescription only. Yes, they are typically used for people with diabetes. Even 10 years ago, it was pretty rare for anyone other than those with type 1 diabetes to get one, but new technology (that’s more accurate and affordable), provider awareness, and consumer demand has driven many non-diabetics to try out CGMs for themselves.

I simply talked it over with my doctor and he was happy to write a prescription for one. I have a feeling that my professional work as a dietitian and certified diabetes educator specializing in gestational diabetes (as well has our talk about some research studies on GD I’m consulting on) helped sway the decision in my favor.

If in doubt, it doesn’t hurt to ask! If your provider errs on the side of preventative medicine or functional medicine, I’d wager they’d be more likely to recommend a CGM.

How much does the Freestyle Libre cost?

This will vary by your location, your pharmacy, and whether insurance covers it or not. For me, insurance did not cover it. The only pharmacy near me that carried the Freestyle Libre was Walmart, which apparently has great pricing on prescriptions.

The Freestyle Libre has two components to function: a reader and a sensor.

Reader: $79.45 (you only buy this once)
Sensor: $43.66 (you have to replace this every 10-14 days)*

Another colleague who has one (but lives in a different part of the U.S.) paid $110 total for the reader and her first sensor; replacement sensors were ~$40. As I said, this will vary based on where you live.

The cool thing is the reader doubles as a regular glucometer, so if you don’t have a blood sugar meter already, you can use this for double duty. It uses the Freestyle Precision Neo test strips.

*At the time of writing the 10-day sensor is the current norm in the U.S. They also manufacture a 14-day reader that was recently approved by the FDA (which, for reasons beyond me, is already the standard in other countries).

Did it hurt to insert the Freestyle Libre sensor?

I’ll admit, the insertion of the sensor was something I feared most. The thought of having a sensor under my skin freaked me out, then thinking about how I would get it there freaked me out more.

Luckily, they’ve designed the Libre to be pretty foolproof. The sensor comes with a contraption that takes all the guesswork out of insertion. It’s essentially a plastic device with a spring-loaded needle—ok, that sounds scary when I write it out…. It’s a plastic thing that attaches to the sensor. You line it up on the right area of your arm and push down. Before you can blink, the spring-loaded needle thing has poked a hole, retracted, and in its place left the sensor attached to your arm.

I swear it was just as easy as a fingerprick. It did not hurt at the moment, but it does make a tiny wound where the sensor is implanted under your skin. For me, that was a little bit sore, like a very very very dull ache (pain scale 1/10) the next day, akin to the feeling after getting blood drawn.

Short answer: no, it doesn’t hurt.

CGM experiment non diabetic continuous glucose monitor freestyle libre
This is the Freestyle Libre sensor; a little larger than a U.S. quarter.

Was it uncomfortable to wear a CGM?

Not really. For the Freestyle Libre, the sensor is worn on the back of the upper arm (tricep area). The part of the sensor that’s external is a little larger than the size of a quarter. The internal sensor is ~ a quarter inch long and aside from the very mild ache I had the day after placement, it didn’t bother me.

I was worried the adhesive on the sensor would irritate my skin, as I’m very sensitive to the adhesive on bandages/tape, but this was surprisingly a non-issue for me.

After a few days, I became less aware of something being stuck on my arm. I’d accidentally bump it every once in a while or catch it on a shirt as I was getting dressed or notice it as I rolled over in bed, but it otherwise it wasn’t bothersome.

I will say, I’m extremely grateful that I don’t have to wear one at all times. I can only imagine what my friends/clients with type 1 diabetes have to put up with having a CGM and often an insulin pump constantly attached to their body. Just one more thing to always have on the back of your mind.

Was the CGM easy to use?

This shows how easy it is to take a real time blood sugar reading from the Freestyle Libre continuous glucose monitor. Press the button on the reader, bring it close to the sensor on your arm, and voila!

 

What are normal and abnormal blood sugar levels?

This is a surprisingly hard question to answer. If you go by American Diabetes Association standards, “normal” blood sugar is less than 100 mg/dl fasting (such as first thing in the morning before eating), and less than 140 mg/dl two hours after eating.

In general, they state that non-diabetic people have blood sugar in the range of 70-130 mg/dl.

For an official diagnosis of type 2 diabetes, fasting blood sugar would be 126 mg/dl or higher, and 200 mg/dl or higher 2 hours after a 75 gram glucose tolerance test (and/or A1c of 6.4% or greater).

For an official diagnosis of prediabetes, blood sugar levels are below the criteria for type 2 diabetes, but above 100 mg/dl fasting and 140 mg/dl 2 hours after a glucose tolerance test (and/or A1c of 5.7% of greater).

What is not specified in these guidelines is how high it’s “normal” to see your blood sugar spike after meals. Since these guidelines are focused on diagnosing and treating diabetes/prediabetes, they don’t give us much insight into truly normal blood sugar levels. I’ll explain more on this later in this post.

Lily, what did you learn about your blood sugar from wearing a CGM?

This is obviously the reason you’re here. This is the reason I did the whole CGM experiment in the first place.

In some ways, my results were surprising and in other ways, exactly as I would expect. My goal was to test my regular diet and see how my blood sugar fared with my usual low-ish carb, real food, mindful eating approach. There were a few unusual meals, including Thanksgiving dinner and the morning I intentionally ate oatmeal (more on that later!); otherwise, I was just eating as I always do.

My average blood sugar over the 10 days was on the low end: 79 mg/dl. No, I did not have symptomatic hypoglycemia. The lower average was due to my blood sugar running primarily in the 60’s and 70’s at night (“fasting”). During the day, my blood sugar mostly hung out in the 80’s-90’s with occasional peaks depending on the carbohydrate content of my meals.

My lower carb real food meals rarely spiked me above 100-110. I had a few higher spikes from meals that included a small to moderate amount of carbs, such as: a slice of real sourdough, rice, potatoes/sweet potato, or hidden sugar (psst – scroll down for my oatmeal experiment).

Average blood sugar 10 day CGM experiment in non-diabetic
This was my 10 day blood sugar average from the CGM.

For example, we had some Thai food from a local Thai restaurant and I’m pretty sure the curry and sauces have sugar in them (it’s pretty standard in Thai cooking to add sugar). I was surprised after eating a big bowl of curry with chicken and vegetables and the tiniest serving of white rice (we’re talking ⅓ cup, friends), to see my blood sugar spike to the 130’s.

Even though the rice was leftover (therefore had been cooled for >12 hours, which converts a portion of the starch into “resistant starch”, which is praised as healthful for our microbiome and less-impactful on blood sugar levels), it clearly still spiked my blood sugar. I can personally feel my blood sugar spiking at fairly low levels with symptoms like racing heart and jitteriness. It’s subtle, but noticeable, probably because my blood sugar tends to run stable in the lower range.

I repeated the same Thai food leftover meal the next day, just without the rice (so I’m sure there was still hidden sugar in the curry) and my highest peak was only 105. White rice and I are apparently not friends, even in portions that provide only 20 grams of carbs.

Sorry, everyone who’s #teamwhiterice. It doesn’t work for me.

Does my blood sugar run low, are the criteria for “normal blood sugar” wrong, or was my sensor off?

I’ll be honest, seeing my numbers average on the low end made me question a lot of things.

First off, is my blood sugar truly low? Maybe. Last time I had blood work drawn (meaning a venous blood sample), my fasting glucose was 68 mg/dl. This was after a 14 hour fast (not intentional; my appointment got pushed back that day).

This makes me think the Freestyle Libre readings were accurate. My fasting insulin levels were also on the low side, which is physiologically to be expected when fasting. The body is going to preferentially switch to burning fat/ketones and meanwhile both glucose and insulin will decrease during extended periods without food.

I spot tested with my glucometer a few times to check the accuracy of the CGM. It was always within 10 mg/dl of my fingersticks, but interestingly the “lower” values from the CGM tested 5-10 points higher on fingersticks, and the “higher” values on the CGM tested 5-10 points lower when double checked with the fingersticks (in other words, CGM made my lows look lower and my highs look higher than was indicated by my blood sugar meter).

My readings were still within an acceptable range of variation, but I did dig up a few articles in the medical literature that suggest that Freestyle Libre readings can be a little off, either from a factory calibration error or other factors. Even so, observing the overall patterns in my blood sugar was invaluable.

It’s also well-known that interstitial glucose (what you measure with a CGM) can have a lag time compared to capillary glucose (what you measure with a finger stick). That in and of itself could explain the slight discrepancy I observed between fingersticks and CGM readings.

If I assume that my blood sugar truly runs on the lower end of the normal range, maybe our current accepted “norms” for blood sugar aren’t perfect.

As with most lab values, the normal range is defined by a certain number of standard deviations from the average. I may be a person who naturally runs on the lower end of the range.

Nonetheless, the research thus far suggests that:

“In individuals with normal glucose tolerance, glycemia is maintained within a narrow range between 68.4 and 138.6 mg/dL.” (Diabetes & Metabolism Journal, 2015)

I tried to dig up some data on blood sugar patterns in hunter-gatherers (modern-day, obviously), to shed some light on the issue.

In the Hadza (people who are indigenous to Tanzania), researchers have yet to observe fasting glucose readings above 85 mg/dl. Similarly, fasting glucose levels from the Shuar (indigenous people of Ecuador and Peru), run on the lower side, with “fasted glucose levels among rural Shuar men (73.6 ± 13.2, n = 32) and women (82.1 ± 21.2 mg dL−1, n = 49).” (Obesity Reviews, 2018)

That means some indigenous people subsisting on their traditional diet have fasting glucose levels as low as 60 mg/dl.

Perhaps my blood sugar readings in the 60s and 70s at night aren’t so low after all?

Are you sure your blood sugar is running low? Enter: OATMEAL.

Another interesting finding was my oatmeal test. Now, for the duration of the experiment, I ate my normal breakfast, which is typically some variation of 2 eggs cooked in butter or lard, vegetables (usually rotating between kale, broccoli, mushrooms, onions, etc.), and possibly some breakfast meat, like sausage or bacon. Alongside this, I have black tea (unsweetened) with heavy cream. [Read: I eat a low carb, high fat, moderate protein breakfast.]

This style of breakfast was a dream for my blood sugar, essentially flatlining it in the 80s or 90s. It’s also excellent for my energy levels, satiety, and productivity (low carb + real food + mindful eating for the win!). I can easily go for 3-5 hours without getting hungry (depends on the day and how active I am), which is nothing short of a miracle for someone who used to be a huge snacker.

There’s a reason I tend to return to a variation of this day after day. Even when I added a small slice of sourdough one morning and ½ cup of leftover roasted potatoes another morning, my blood sugar didn’t exceed 100 (the magic of not eating “naked carbs”).

I started to wonder if I was just super insulin sensitive in the morning or maybe had more wiggle room for carbs.

So, the final day of my CGM sensor, I decided to eat a breakfast similar to the Academy of Nutrition and Dietetics’ sample meal plan in their pregnancy guidelines. Their meal plan is oatmeal, skim milk, and strawberries. In other words, all carbs.

(For those of you familiar with my book, Real Food for Pregnancy, this is the same meal plan that I use to make a comparison on the nutrient-density and macronutrient ratio compared to my real food meal plan. This excerpt of the book is included in the free chapter download; see the bottom of this post or this page to get it.)

I don’t have skim milk in the house (and I never will), but I do have whole, grass-fed milk for my son. I also didn’t have strawberries, but I have raspberries, which are nutritionally similar. So I whipped up measured portions of rolled oats (1/2 cup dry, which is 1 cup cooked) prepared with water, a little milk poured on afterwards (I’m really not a fan of straight up milk, so I only used a few Tbsp), ½ cup raspberries, and because it was in-edibly plain, I added 1 measured teaspoon of honey (not heaping). Total carb count was 45g.

All things considered, this was NOT a large bowl of oatmeal and it was essentially NOT sweet, despite adding a little honey. (I say this to point out that the average person adds A LOT of extra sweeteners to their oatmeal, either with sugar/honey or dried fruit. My version would be unpalatable to many people.)

At first, I thought my blood sugar was doing ok after the oatmeal, but I then watched with horror on my Freestyle Libre as the readings climbed. When you scan the sensor, the Freestyle Libre reader shows an arrow next to the numerical reading with an up, down, level, or slightly up/down error, indicating your real time blood sugar trends. This was the ONLY time during the entire 10 days that I saw the straight up arrow, indicating my blood sugar was rising FAST.

My blood sugar went from 74 to a peak of 178 in an hour. By two hours, I was down to the 120s and by three hours, finally back down to 100.

This perfectly illustrates why I preach “no naked carbs” (meaning carbs eaten without a source of fat or protein). The spike from oatmeal was 40 points higher than my Thanksgiving dinner, which included stuffing, sweet potatoes, cranberry sauce, pumpkin pie (in other words, the same or more carbohydrates than this unsatisfying, bland oatmeal breakfast). I’d wager that I didn’t spike as high from Thanksgiving because there was sufficient fat, protein, and fiber in the meal to slow digestion and absorption of the carbohydrates.

This image shows 4 days of blood sugar data. The first 3 days, I’m eating my regular diet. The last day, I did the oatmeal experiment. See the peak on day 4?

Continuous glucose monitor experiment in a non diabetic - oatmeal spike
3 days of happy blood sugar readings on my moderately low carb diet. Day 4: OATMEAL for breakfast.

What was interesting, though, is that I got ravenously hungry when my blood sugar started plummeting. Now, this is a mindful eating/blood sugar physiology teaching I’ve been using in practice for years (go back to my viral post, The Healthy Breakfast Mistake, for a hilarious run down).

It was really interesting to watch it pan out in real time on my CGM. I had always assumed that I wouldn’t be hungry until my blood sugar was back down to normal or in the hypoglycemic range, but this experiment showed me that the hunger trigger (for me, anyways) is in response to the impending crash.

I actually didn’t end up hypoglycemic in response to this meal, but I literally had to eat something with substance (fat/protein) around the two hour mark (~ 2 oz of leftover grass-fed beef burger patty) to avoid going hangry. That probably stopped me from going into reactive hypoglycemia.

After this oatmeal experiment, I started wondering if my glycemic response to oatmeal was exaggerated or unusual. It’s surprisingly hard to find data on blood sugar responses in non-diabetics, especially when trying to examine the “peak” blood sugar response.

As you can recall, as long as blood sugar is back down to 140 mg/dl by two hours after eating, then you’re supposedly “in the clear” by conventional guidelines for diabetes/prediabetes. If a study has people measure their blood sugar only at 2 hours, you’re likely to miss the peak glycemic response in many people. Moreover, different people peak at different times, so without a million finger pricks or CGM, the results aren’t going to be very meaningful.

Blood sugar patterns in non-diabetics

After spending entirely too much time on Google Scholar (this is my M.O. in life), I was able to find some studies on continuous glucose monitoring in non-diabetics (hooray!).

In one study from Stanford, 57 participants—comprised of people with and without diabetes—wore a CGM for 2-4 weeks. (PLoS Biology, 2018)

They stratified the glucose data into 3 subtypes of glycemic variability (essentially, how much a person’s blood sugar spiked or dropped throughout the day). The healthiest subjects tended to have the lowest glycemic variability. In other words, their blood sugar stayed within a tightly controlled range for most of the day.

These people tended to have a lower BMI, be younger, and have lower readings of glucose, A1c, fasting insulin, and triglycerides (this resembled my glycemic patterns).

A subset of the study group were given standardized meals to measure the glycemic response. In the test of cornflakes and milk (which is nutritionally similar to my oatmeal breakfast), fully 80% of people without diabetes experienced blood sugar spikes beyond 140 mg/dl (prediabetic levels). Furthermore, 23% of this group saw blood sugar levels exceeding 200 mg/dl (diabetic levels) after cornflakes and milk.

They write:

“It is interesting to note that although individuals respond differently to different foods, there are some foods that result in elevated glucose in the majority of adults. A standardized meal of cornflakes and milk caused glucose elevation in the prediabetic range (>140 mg/dl) in 80% of individuals in our study. It is plausible that these commonly eaten foods might be adverse for the health of the majority of adults in the world population.” (PLoS Biology, 2018)

In a press release on the study, one of the authors comments further on the cornflakes result:

“Make of that what you will, but my own personal belief is it’s probably not such a great thing for everyone to be eating.”

Ya think?

This is the whole argument for a lower carbohydrate diet as a treatment and/or preventative strategy for type 2 diabetes.

They also note that occasional high blood sugar readings (arguably, not ideal), are common, even among people without diabetes:

“Importantly, we found that even individuals considered normoglycemic by standard measures exhibit high glucose variability using CGM, with glucose levels reaching prediabetic and diabetic ranges 15% and 2% of the time, respectively. We thus show that glucose dysregulation, as characterized by CGM, is more prevalent and heterogeneous than previously thought and can affect individuals considered normoglycemic by standard measures, and specific patterns of glycemic responses reflect variable underlying physiology. The interindividual variability in glycemic responses to standardized meals also highlights the personal nature of glucose regulation.” (PLoS Biology, 2018)

This study brings into question the overall insistence on using glucose tolerance tests, A1c, and fasting blood sugar to define diabetes, as if it’s this black and white diagnosis. (Spoiler alert: it’s not.)

Imagine if you could go to the doctor, have a CGM sensor inserted on the spot, then return in 2 weeks to have your data analyzed and graphed using the criteria in this study. It would give us a much more nuanced look into your blood sugar regulation. It would also reveal what foods are great for your blood sugar and conversely, which ones are a glycemic disaster. This would be preventative medicine. Intervene before blood sugar levels are consistently in the diabetic range, before insulin resistance gets too severe, and before beta cell burnout.

Then again, this would be a royal pain for practitioners/insurance companies who like the black and white nature of single tests. I don’t think I’ll hold my breath waiting for this to become the standard of care.

In another CGM study (this one from Spain), glycemic patterns were tracked in 322 people (both prediabetic and non-diabetic) eating their usual diets, providing 1521 complete days of data. (Diabetes Research and Clinical Practice, 2018)

Participants had blood sugar in the non-diabetic range (<140 mg/dl) 97% of the time.

Interestingly though, 73% of participants experienced episodes of high blood sugar in the prediabetic range (and 5% in the diabetic range).

The average amount of time spent above the normoglycemic range (blood sugar over 140 mg/dl) was 32 min per day in the normoglycemic group and 53 min in the prediabetic group. The peak blood sugar reading for non-diabetics averaged 119 mg/dl (compared to 128 mg/dl in the prediabetic group).

Also interesting was that 44% of non-diabetic people in this study had some bouts of “hypoglycemia” defined as less than 70 mg/dl. (This again made me question if 70 mg/dl is really an appropriate cut off for hypoglycemia.)

I think this data highlights just how hard our bodies work to keep our blood sugar in a tightly controlled range 24/7. I don’t think a few high readings are necessarily a sign of something wrong, it just shows you which foods force your body to work especially hard to bring those “highs” back to the normal range as quickly as possible.

Blood sugar patterns in non-diabetic people who are morbidly obese

Another interesting dataset I came across was an analysis of CGM data from morbidly obese participants with and without prediabetes (literally these participants were applicants to The Biggest Loser; it’s in the study methods) found significantly higher glycemic variability and overall higher glucose levels when compared to CGM data from non-diabetic, non-overweight adults. (J Diabetes Sci & Tech, 2014)

Specifically, prediabetic Biggest Loser applicants had 17.6% of blood sugar readings above 140 mg/dl while non-diabetic Biggest Loser applicants had 11.8%. Meanwhile, non-diabetic adults who are not morbidly obese only experienced 0.3-4.1% of readings above 140 mg/dl.

This data highlights the role that excess body weight plays in blood sugar regulation, even if you “pass” diabetic/prediabetic screenings by conventional standards.

Should we be eating so many carbs?

All things considered, I think what these studies highlight is that modern, high-carb diets are a mismatch for our physiology. The research has a name of it: evolutionary discordance hypothesis.

In other words “departures from the nutrition and activity patterns of our hunter‐gatherer ancestors have contributed greatly and in specifically definable ways to the endemic chronic diseases of modern civilization” (Nutrition in Clinical Practice, 2010)

To quote Eric Sodicoff, MD, “CGM is useful to prove that you need to be eating stone age food.”

Is there harm—or even benefit—in occasional blood sugar excursions?

Some argue that elevated blood sugar after meals is normal, while others argue that you should aim to keep your blood sugar as “flatlined” as possible. Honestly, I don’t know which is true and I don’t think we have an easy answer. 

Where we need more research is in blood sugar patterns of metabolically healthy individuals. I’d be especially curious to see some CGM data from modern day hunter-gatherers.

We have an abundance of data on the harms of consistently elevated blood sugar (and certainly as a country are observing first-hand epidemic levels of beta cell burnout and insulin resistance). On the other hand, I can see from the perspective of maintaining metabolic flexibility (“testing” your pancreas just like you “test” your immune system during occasional illness, seasonal differences in food availability ancestrally) that it may be harmless to experience occasional blood sugar spikes.

From an ancestral perspective, carbohydrate intake varied seasonally. So It’s certainly plausible that having some occasional high blood sugar readings is not a concern, so long as there are spans of time where they can return to baseline (which was likely the case before the convenience of a globalized food supply, grocery stores, and take out).

The key here is that if occasional blood sugar spikes are ancestrally normal, it was also ancestrally normal to have periods of time WITHOUT blood sugar spikes. That is something that is missing from modern life for most of us.

Moreover, as I cover in Real Food for Pregnancy, the most comprehensive assessment of dietary patterns in modern-living hunter gatherers shows that they eat far fewer carbohydrates than we do with an average of 16-22% of calories from carbs. Compare that to our dietary guidelines, which push 45-65% of calories from carbs. It’s just A LOT of carbs for the human body to process.

Studies looking at glycemic variability, which refers to swings in blood glucose levels, tend to point towards minimizing the frequency of blood sugar spikes for heart health:

“Since 1997, more than 15 observational studies have been published showing that elevated [postprandial glucose], even in the high nondiabetic [impaired glucose tolerance] range, contributes to an approximately 3-fold increase in the risk of developing coronary heart disease or a cardiovascular event [13]. Moreover, the meta-analysis of the published data from 20 studies of 95,783 individuals found a progressive relationship between the glycemic variability (GV) and cardiovascular risk [49]. In summary, the accumulated data that GV seems to be associated with the development of microvascular complications appear to be impressive.” (Diabetes & Metabolism Journal , 2015)

Bigger implications for diabetes and prediabetes

At this time, 49-52% of the adult US population has either diabetes or prediabetes, many of whom remain undiagnosed. (JAMA, 2015) It’s estimated that upwards of 70% of people with prediabetes will go on to develop type 2 diabetes. (PLoS Biology, 2018)

A 2018 analysis of the most recent National Health and Nutrition Examination Survey data found that only 12% of Americans are “metabolically healthy.” TWELVE PERCENT. (Metabolic Syndrome & Related Disorders, 2018).

That means 88% of Americans are NOT metabolically healthy.

By the way, their definition of metabolic health included fasting blood sugar, A1c, blood pressure, triglycerides, HDL cholesterol (checking to see they had sufficient “good” cholesterol), and waist circumference. I was pleased to see they left total and LDL cholesterol as well as BMI out of their criteria, as these are less reliable markers of overall health.

I have no doubt that if I continued to eat high-carb meals for my lifetime (like oatmeal), pushing my blood sugar levels to prediabetic levels again and again and again… gradually resulting in a breakdown of my insulin sensitivity, I would become one of these statistics. Furthermore, eating this way leaves me feeling like garbage. It makes me crazy hungry for all the wrong foods as my poor body attempts to regulate blood sugar between spike after spike.

Fortunately, I don’t have to go down that path because this experiment has highlighted just how crucial a lower-carb, real food diet is to maintaining blood sugar balance, even for people like myself, who supposedly have normal glucose tolerance. If you watch the spectacular results of the Virta program or similar diabetes interventions using a lower carbohydrate approach, it’s just common sense. (Diabetes Therapy, 2018; JMIR Diabetes, 2018)

Even without blood sugar data, simply relying on mindful eating is what keeps me eating this way. If I feel good, if my energy is solid, if my mind is clear, if I’m sleeping well, if my menstrual cycles are normal, if my digestion is happy, if my thyroid is normal… obviously my body is communicating that this works for me. I can check all of those boxes (well, now that my toddler sleeps, I can check the sleep box… hooray!).

Furthermore, all my other lab markers of blood sugar and insulin regulation are good. Fasting insulin is low, HOMA-IR (a marker of insulin resistance) is good, leptin:adiponectin ratio is good (this is another surrogate marker of insulin resistance), A1c is good, all lipids are optimal, CRP is low, BMI is normal, etc.

The point is that even if you’re “metabolically healthy,” your blood sugar can still spike pretty high if you eat a bunch of naked carbs.

My oatmeal experiment made that crystal clear.

Why is carb tolerance so individual?

Everyone’s carb tolerance is different and that’s what is so cool about wearing a continuous glucose monitor. For me personally, total carbohydrate intake at meals was the biggest determining factor of my blood sugar response. Non-starchy vegetables, which contain small amounts of carbohydrates but a higher proportion of fiber (meaning low in net carbs), didn’t spike my blood sugar whatsoever, even at meals with BIG portions of non-starchy vegetables.

I was able to tolerate potatoes, sweet potatoes, winter squash, beans, real sourdough bread, and fruit (eaten whole) without a spike my blood sugar so long as I ate them in reasonable portions (about half a cup at a time) and alongside fat and protein (thus lowering the glycemic load) and lots of non-starchy vegetables (fiber & phytochemicals). I learned that rice, oatmeal, and hidden sugar did not work well for my body.

It could be that the latter sources of carbohydrates are “acellular,” which some research argues has a more severe glycemic and inflammatory response.

It could be that my microbiome has some preferences of its own (yes, the bacteria in your gut appear to play a role in your blood sugar).

It could be that I inherited a pancreas that doesn’t enjoy pumping out big boluses of insulin (could be from generations past or even my mom’s diet while I was in utero). Yay, #epigenetics.

Technically, it doesn’t matter what the mechanism is. What matters is that I can identify the foods that spike my blood sugar and make a choice as to if I eat them, in what portions, and how often.

Final thoughts on my CGM experiment

There’s a lot of “noise” in the nutrition world.

People telling you that you need to eat a certain macronutrient ratio. People telling you that going low carb is guaranteed a bad choice for women (there’s some nuance here, but I overall disagree so long as you’re eating sufficient quantities of nutrient-dense food, managing your stress, and not over-exercising). People telling you that you must go plant-based or carnivore.

Essentially we have a lot of people telling you that whatever you’re doing is wrong.

Wearing a CGM cuts through all the noise.

Eating for better blood sugar balance has carryover benefits to the rest of your health. I think the key with CGM is to try to observe other symptoms and how they relate to your glycemic response to a meal.

Note how your energy and hunger/fullness levels relate to your blood sugar.
Note how your sleeping patterns affect your blood sugar.
Note how eating to satiety vs. eating to discomfort affects your blood sugar.
Note how snacking or not snacking affects your blood sugar.
Note how eating an earlier or later dinner affects your blood sugar.
Note how exercise, stress, meditation, deep breathing, light exposure, nutritional supplements, etc. all affect your blood sugar.

Then play around with all of these factors to make adjustments to bring your blood sugar back to a healthy range.

I see CGM as a huge opportunity for preventative medicine. Here’s to hoping this technology becomes available more widely, for a lower cost, and without a prescription.

If you have questions or comments on CGM, please share them below and I’ll do my best to address them here or in a future blog post/interview.

Until next week,
Lily

PS – I know a big portion of my audience is interested in pregnancy. If you’re curious how blood sugar balance is affected by pregnancy, I encourage you to check out Ch 9 of Real Food for Pregnancy, where I sort through all the pros/cons of the different testing options for gestational diabetes and how blood sugar patterns naturally shift during pregnancy.

Also, see this post that explores 9 myths about gestational diabetes.

Research into using CGM as a replacement for glucose tolerance tests are underway, so we’ll hopefully have more insight into this as a screening option for gestational diabetes sometime in the future.

References
1. Suh, Sunghwan, and Jae Hyeon Kim. “Glycemic variability: how do we measure it and why is it important?.” Diabetes & metabolism journal 39.4 (2015): 273-282. 2. Pontzer, H., B. M. Wood, and D. A. Raichlen. “Hunter‐gatherers as models in public health.” Obesity Reviews 19 (2018): 24-35. 3. Hall, Heather, et al. “Glucotypes reveal new patterns of glucose dysregulation.” PLoS biology 16.7 (2018): e2005143. 4. Rodriguez-Segade, Santiago, et al. “Continuous glucose monitoring is more sensitive than HbA1c and fasting glucose in detecting dysglycaemia in a Spanish population without diabetes.” Diabetes research and clinical practice 142 (2018): 100-109. 5. Salkind, Sara J., et al. “Glycemic variability in nondiabetic morbidly obese persons: results of an observational study and review of the literature.” Journal of diabetes science and technology 8.5 (2014): 1042-1047. 6. Konner, Melvin, and S. Boyd Eaton. “Paleolithic nutrition: twenty‐five years later.” Nutrition in Clinical Practice 25.6 (2010): 594-602. 7. Menke, Andy, et al. “Prevalence of and trends in diabetes among adults in the United States, 1988-2012.” Jama 314.10 (2015): 1021-1029. 8. Hall, Heather, et al. “Glucotypes reveal new patterns of glucose dysregulation.” PLoS biology 16.7 (2018): e2005143. 9. Araújo, Joana, Jianwen Cai, and June Stevens. “Prevalence of Optimal Metabolic Health in American Adults: National Health and Nutrition Examination Survey 2009–2016.” Metabolic syndrome and related disorders (2018). 10. Saslow, Laura R., et al. “Outcomes of a Digitally Delivered Low-Carbohydrate Type 2 Diabetes Self-Management Program: 1-Year Results of a Single-Arm Longitudinal Study.” JMIR diabetes 3.3 (2018): e12. 11. Hallberg, Sarah J., et al. “Effectiveness and safety of a novel care model for the management of type 2 diabetes at 1 year: an open-label, non-randomized, controlled study.” Diabetes Therapy 9.2 (2018): 583-612.

Get a Sneak Peek of the Book!

Download your FREE chapter from

Real Food for Pregnancy.

You'll also receive Lily Nichols' weekly newsletter.
Unsubscribe at any time. Privacy Policy

Low Carb Korean Beef Bowls
Embracing Your Body & Mindful Eating in Pregnancy
Choline in Pregnancy Folate's Long Lost Cousin
Deep Dive into Real Food Prenatal Nutrition

Lily Nichols is a Registered Dietitian/Nutritionist, Certified Diabetes Educator, researcher, and author with a passion for evidence-based prenatal nutrition and exercise. Her work is known for being research-focused, thorough, and unapologetically critical of outdated dietary guidelines. She is the author of two bestselling books, Real Food for Pregnancy and Real Food for Gestational Diabetes.

91 Comments

Leave a comment
  1. Thank you for this excellent article Lily. Really interesting!
    Happy New Year, and keep up the great work!

    • Thank you! Glad you enjoyed it Afifah.

  2. Great piece. I had *exactly* the same experience. Non-diabetic, low-carber, wore a Libre for a month and tasted a wide range of foods. No scrip needed at the time (a year ago) in Canada.

    • No script? That would be amazing to see in the states!

  3. Wow, that was very insightful!! Thank you for doing the legwork to get this info. I read your book after being diagnosed with GD, and followed that (as well as a Trim Healthy Mama S meal approach) for the remainder of my pregnancy. I’m still doing random glucose checks (fasting and the odd post-prandial) to see how my body responds to carbs and sugar (we eat a mainly keto diet, with the odd splurge), and I find a CGM so fascinating. I might see if my doctor will prescribe one even though I’m not diabetic. Thank you for sharing this wealth of knowledge!

  4. That was a fascinating write up! You are by far the standout in this space regarding pregnancy nutrition and healthy diet. So scientifically sound, honest, reasonable and real. Thanks for this detailed article!

    • Glad you enjoyed it!

  5. Thanks so much for doing this experiment and writing it up. I can’t say I’m surprised by any of it as it confirms what we both expected, but it’s helpful to have this information out there for people to gain a better understanding of how “healthy” foods are really just contributing to metabolic dysfunction.

    I’m also not on #teamwhiterice despite being half-Persian. If I eat any more than a 1/2 cup of rice when eating out, my husband needs to drive us home because I fall asleep on the way home

  6. This is super interesting!! I want to monitor mine as well. I still finger prick often even after a great A1C. I am just so curious and cautious…and I find it so fascinating!

    The whole naked carb thing fascinates me too. I find if my son eats any sort of naked carbs he goes batshit crazy. It’s really interesting. But paired with fats/protein he has a complete different reaction. Not diabetes related but because of you I tried experimenting with this and it seriously makes a big difference in his behaviour and mood. Naked carbs are the devil.

    • Yeah, as I share in the article, my Thanksgiving meal had the same (or not more) carbohydrates than the oatmeal and my glycemic response was significantly lower, likely because it was a mixed meal with fat/protein.

  7. Very interesting! If you had to hypothesize, do you think that women who consume higher amounts of naked carbs during pregnancy might predispose their offspring to diabetes epigenetically?

    • No need to hypothesize. There’s research showing precisely this. See Chapter 2 of Real Food for Pregnancy in the section on carbohydrates.

  8. Thanks Lily, brilliant article and insight to your experience.

    Personally before I was pregnant I followed a very strict Paleo Ketogenic protocol, 15 weeks into my pregnancy and replaxing my carb intake ive gained 12 lbs, I’m very interested in running this experiment on myself to see what I can and can’t tolerate well. Do you think being pregnant would throw off the readings, or wouldn’t make a difference?

    I would be more than happy to share my results with you.

    Thanks again Estelle (estelleskitchen)

    • Pregnancy tends to result in slightly lower blood sugar readings (see this article) and the “goal” blood sugar readings are lower than non-pregnant.

      If you have my book, Real Food for Pregnancy, see Ch 9 for an in-depth discussion of blood sugar goals and different screening options for gestational diabetes.

  9. Thank you Lily for insight and experience taking this test.

    Personally I have been considering doing this myself, as for many years I followed a strict Paleo Ketogenic protocol and use to measure my BG and Ketones regularly.

    Fastforward and I’m 15 weeks pregnant, relaxed my carb intake and concerned about the rapid weight gain of 12lbs. Thinking allowed this might provide me the tool to see what carbs and I can can’t tolerate well (very much inline with Robb Wolf’s carb test). Do you think being pregnant would throw off the results somewhat? I would be happy to share my results with you for your readers.

    Thanks so much Estelle

  10. Thanks so much for this great article! I have a history of GDM and will still occasionally check my sugars, but have struggled trying to figure out what is a truly normal response to a higher carb load so your research is very interesting!

    I’ve read a little about “physiologic insulin resistance” in those who eat a lower carb diet (especially Keto) and I wonder how that plays into your spike from eating oatmeal? If you started eating more carbs daily would you spike less? For those that overall are eating a lower carb diet that is having good metabolic results what is the result/outcome of then the occasional higher carb meal (and potentially higher blood sugar spike)? Do we need to occasionally make our pancreas do a little work so it doesn’t forget how to do it’s job??

    • Yes, that’s the argument for carb loading prior to a glucose tolerance test, which essentially “primes” the pancreas to produce larger boluses of insulin. Technically, I had several instances of moderate carb loading — Thanksgiving, the Thai food with white rice — with significantly lower spikes than the oatmeal. Technically, my response to oatmeal would still not classify me as either prediabetic or diabetic because my blood sugar returned to “normal” within 2 hours.

      It’s very difficult to find data on what’s “normal” or not in terms of blood sugar responses. Hopefully more CGM studies, similar to those I mention in the article, will be completed in the coming years. For now, what keeps me feeling well and keeps my blood sugar in check is moderately low carb, nutrient-dense, real food.

  11. I guessed the spike! Although I did think I was pushing it. What’s my prize?

    • A bowl of oatmeal to celebrate? Haha.

  12. A real eye opener, thankyou for the great article.
    One of the doctors I work with asked my advice for a low carb/ real food dietitian for pregnancy and resources for a friend and I gave her your name. I hope she read your work!

    • Thanks for the referral!

  13. This is a great article. I’ve focused on eating protein, fat, fiber, and greens at meals (and keeping carbs moderate/non-starchy for the most part) for the last year… but now I’m traveling Europe and have noticed a STARK contrast when eating naked carbs (weight, acne, mood, energy). This article helped me to fully articulate the science of blood sugar balance to my husband. We’ll be using #nonakedcarbs as our reminder. 🙂 Thank you for your work!

    • It really struck me how much mindful eating observations were spot on to my blood sugar readings. Glad to hear #nonakedcarbs is working for you.

  14. Hi Lily

    Thanks for this fascinating insight, I was really interested to read it. I’m following your real food principles to gain control of my own gestational diabetes at the moment (doc’s say they think it’s actually type 2 but can’t be sure) and like many women was struggling with my first thing in the morning fasting readings. For a couple days now they’ve been normal. They were never in the normal range before, even with the slow acting insulin I was taking before bed, and even when this was dialled up slightly. I noticed it go down when I did two things: a) have a snack before bedtime and b) not take my prenatal multivitamin right before bed as usual. It’s probably the insulin kicking in and the snack, but do you know of any correlation between taking a prenatal multivitamin at a certain time and a spike in blood sugar?

  15. super interesting Lily! As a fellow RD and CDE, I think you’ve inspired me to do this experiment as well. I specialize in PCOS and am very interested in a low glycemic diet. A few questions/thoughts: what type of oatmeal? I wonder what would have happened it you had used steel cut oats, added chia or flaxseeds and an ounce of nuts? (and not added the honey or fruit) Second, what do you think about Day Two? I realize it’s not even close to the CGM, but have you had any experience with it?

    • It was rolled oats. Years ago, I ate steel cut oatmeal for breakfast every day (as I explain in this post from my archives). I added all the “good” things to up the fat/protein/fiber. I didn’t get hungry quite as quickly, but felt just as awful. My expectations is my blood sugar probably wouldn’t spike as high, but it would still spike from a high carbohydrate load. If I ever get another CGM sensor, I may experiment with this more.

      • I loved reading this about this whole experiment! I would also be interested if you ever did more experimenting with oatmeal. I’ve been a type 1 diabetic for over 30 years, and before I ate a low carb diet, I would occasionally have oatmeal. For some reason, oatmeal (rolled oats, soaked overnight) would really spike my blood sugar more than any carb (rice, white potatoes, etc.), even when I added coconut oil to it. I have another type 1 friend who says the same about oatmeal. I’d be interested to see a non-diabetic compare white rice, white potatoes, and oatmeal (with and without fat/protein), or something along those lines. I’ve often wondered if there’s something about oats…

        Thank you so much for doing this experiment and writing it up so thoroughly!

  16. Great article Lily and hope one day soon the Libre will be available wo a prescription like glucometers are now. The glucometer and CGM are such powerful behavior modification tools as you demonstrate. I need to get one and wear it for the 10-14 days to learn more.

    • I hope so too, Mark!

  17. Excellent article! Thank you for doing this experiment, all the added research and sharing it with us! I’m curious why the two week cut off? Perhaps that was all your doctor would write the script for? Anyway I was just wondering why you didn’t wear the CGM longer to be able to test more foods and food combinations.

    You’ve inspired me for sure! Keep up the great work; I love it!

    • The Libre sensors automatically stop working after the 10-14 day cycle. I can get another sensor to continue measuring glucose should I run another CGM experiment.

  18. Great article Lily, thank you! I was wondering if there was a specific brand of glucometer you use? I had a relion prime and it never seemed to read reliably. Thank!

    • I have the most reliable results with the Freestyle line of meters (no financial affiliation).

  19. I would be curious to see how a dessert (higher sugar than oatmeal) would affect your BG. Do you ever eat dessert/treats?

    Also curious about your usual physical activity level. It would be interesting to see how a day of just moving more (life chores/tasks) affects BG versus an hour of high intensity exercise followed by a full day of desk work (my typical day).

  20. Thanks for your great n=1 research work Lily. Very useful for me because I am about to embark on a similar test with a Freestyle Libre (I’m also non-diabetic) so you provide really helpful benchmarks. I particularly like your sub-head – ‘Wearing a CGM cuts through all the noise’ – that’s the biggest plus, for us all.
    My only concern is that this type of ‘individual’ approach gets used to reduce dietary generalizations and provide arguments for how different we all are, and how we need more ‘made-to-measure’ approaches to nutrition. While some individuals will be less responsive to oats or rice or potatoes, the essential principles advocating a lower-carb, real food diet seem to be the best basis from which to start.
    Many, many thanks for your patient and complete approach… Sammy

  21. Thank you for the excellent article. Very well written and SO interesting. Makes me want to do an CGM experiment on myself. Now if only my doctor would prescribe me one. 🙂

    • Hi Lily,
      Thanks for the great insights.
      As you are already low carb, it is not surprising to see spikes in blood sugar with even little bit of carbs (rice and oats) in your experiment. During low carb eating, glucokinase is down regulated and physiologic insulin resistance kicks in which means glucose takes lot longer to get disposed of as compared to someone healthy who is not on low carb. This is at least true for most low carbers including myself. An OGTT taken in this state looks scary. It would be very interesting if you ate 150g carbs per day for 3 days at least and then monitored your blood sugar response to carbs.

  22. Just excellent! Thank you for taking the time to do this experiment and write up afterwords!

  23. Hi Lily, I really enjoyed your post! Thank you! After reading it, it made me think of one of Chris Kresser’s post regarding low carb and pregnancy. I wondered if you had thoughts on the points he raised at the end when discussing the “Fetal Origins of Disease Hypothesis”.
    For example, he writes…

    The fetal origins of disease
    “Nutrition during pregnancy is important for more than just supporting the growth and development of the fetus, though. It also programs the metabolism of our children for the world they will be born into.

    This is more formally called the “fetal origins of disease” hypothesis, which has been well supported by the scientific literature. The most oft-cited example is children who were born to women during the Dutch Hunger Winter during World War II. Severe undernutrition of these women early in pregnancy produced children who were metabolically programmed for a world where food was scarce and where conserving energy and storing fat was advantageous. When the famine ended and food became plentiful again, these children were predisposed to obesity and metabolic syndrome in adulthood (4).

    There are only a few studies that have looked at carbohydrate restriction and how this affects the health of offspring later in life, but in general, they don’t favor low-carb diets.

    One study in humans found that the offspring of mothers who had consumed higher levels of protein and fat (likely resulting in lower carbohydrate intake) had significantly reduced insulin production in response to a glucose challenge 40 years later. For mothers consuming adequate protein, there was also an average 9.3 mm Hg increase in adult blood pressure for each 100 gram decrease in maternal carbohydrate intake (5).
    A similar study found that a high-protein, low-carbohydrate diet during pregnancy was associated with increased cortisol levels in the offspring 30 years later (6). (Unfortunately, these studies did not report micronutrient intakes.)
    Animal studies have suggested that a ketogenic diet during pregnancy may reduce the size of brain regions like the hippocampus in offspring, while increasing the size of others, such as the hypothalamus (7).”

    Here is the link to his article from March 27, 2018:
    https://chriskresser.com/do-low-carb-diets-during-pregnancy-increase-the-risk-of-birth-defects/

    No worries if you don’t have time for this! 🙂

    • I encourage you to read chapters 2, 7 and 9 of Real Food for Pregnancy for an in-depth look at macronutrient “requirements” in pregnancy and the implications of elevated blood sugar as well as chapter 11 of Real Food for Gestational Diabetes if you’d like to understand the science for/against low carb in pregnancy.

      • Thanks Lily! I have your Real Food for Pregnancy book and have read portions, but will check out those chapters. Epigenetics is such an interesting field in relation to pregnancy. Thank you!

  24. Super article. It reinforces my own experience as a T2 using a CGM. I am a 67 year old British man so not your usual reader.

    However, I will also add that for me, wheat and oats are a one way switch. My BG rises to as much as 20 (360 American) and does not go down until the grain is expelled. Rice spikes a bit less and goes down. Even straight sugar is not so potent, although I have never attempted to eat a similar amount. Using cream as a source of fat moderates the spikes and eliminates hunger on the downslope.

    Cereals have an enormous role in diabesity that is hardly discussed.

    My usual breakfast since my CGM opened my eyes is grapefruit or microwaved blueberries with cream followed by two strips of bacon with tomatoes/avacado/mushrooms/baked beans.

    Fasting until about 13:00 also avoids huge spikes as insulin resistance seems to decline over the day. A large roast dinner has almost no impact on my blood sugar.

    Great writing too.

    • Thanks for sharing your experience, Phillip.

  25. Interesting results. Could you elaborate on the type of oats you used? Was is the plain, large chunks (which you have to boil for ca. 15min) or the express type (done in 2-5min)?

    • Old fashioned rolled oats (not quick cooking; the kind that take 20+ minutes to cook), organic, certified gluten-free.

  26. Perhaps if one adds some whey protein and some fat ( eg nut butter) to the oatmeal the response might be better?

    • Likely yes, hence the #nonakedcarbs discussion. The meal I tested was specifically to mimic the sample meal plan from the Academy of Nutrition and Dietetics (oatmeal, milk, berries) and see what happened with a meal of primarily carbohydrate.

  27. Interesting!

    1. For the geeks in your audience, it would be nice if the references were more specific so people could read them. “JAMA, 2016” isn’t too useful.

    2. JS Christiansen of EASD in 2006 used CGMs on allegedly nondiabetic people after a high-carb breakfast and reported the results at a conference that used to be online at diabetes-symposium.org, but I think no longer is. They showed individual results, and some people didn’t go over 100 at the peak whereas others went quite high.

    3. I also get ravenously hungry when a peak starts coming down. The analogy I use is, if you had a toddler coming down a slide and there was broken glass at the bottom, you’d run to stop the kid *before* she reached the bottom rather than waiting until she was in the glass. That’s what the body is doing.
    Sciarc.

    • See the references tab at the bottom of the post for full citations.

  28. Lily,
    Very informative post thank you.
    I’d like to suggest 4 extra CGM tests:
    1. Measure the effect of 2 measured glasses of dry white wine, one before a meal and one with a meal. (100 ml is a standard serve, but 150 ml is a restaurant serve)
    2. Measure the effect of a sugared soda with a meal.
    3. Measure the effect of a diet soda consumed alone when fasted. (I expect there will be no blood sugar response, but may be wrong. But I would expect that there would be an insulin response, which a CGM would not pick up – I think.)
    4. Whether reduced sleep (say 6 hours versus 8) affected the response to a breakfast.
    Best wishes from Brisbane Australia

  29. Quote ‘This data highlights the role that excess body weight plays in blood sugar regulation, even if you “pass” diabetic/prediabetic screenings by conventional standards.’

    Or could this observation show raised blood glucose; spiked insulin leading to fat storage & weight gain ie being over weight was the symptom not the cause!

  30. Great article and experiment!

    I was wondering what it was like to remove the sensor after your experiment was done?

    Also, the earlier comment where the gentleman recommended having you drink diet soda would be really interesting! You might be able to experiment with other no calorie sweeteners like Splenda, Truvia, or monk fruit and see your body’s response to them.

  31. This is fantastic Lily! Well done! Great to see this in action!

    Happy new year and keep up the fab posts! 😉

    Andrew

  32. You go girl! THanks for putting yourself out there. I literally read this while my in laws were in town. I had just told my MIL that her blood sugars were in prediabetic range and we may want to rethink how she starts her day (homemade muffins and oatmeal). I was glad to share with her your results. It’s amazing how people don’t think they can get fiber or have normal bowel movements without oatmeal!

  33. This article is super validating. I can’t eat something like oatmeal or toast for breakfast because I’ll feel ravenous within a half hour. However, I can eat an egg with spinach or a piece of buckwheat toast with butter and be satiated for hours. Thanks for your experiment!

  34. I am wondering about alcohol consumption, which includes wine with dinner, or a glass of wine with hubby in the evening after a tough day, etc. Some think this is beneficial to health. Is it the combination with good food that makes the difference? What do you think?

  35. Yes! I Thank you for this post. have been wondering about all these things. Thank you for doing this experiment—we definitely need more data on glycemic response in healthy people. I am also very interested in how the fourth trimester affects glucose response in healthy individuals vs those with GD, particularly in the initial phases of lactogenesis. I have yet to find much data on that either, but I suspect that the 24-hour feeding cycle (ie interrupted sleep) plus all the hormonal changes happening during that time must impact insulin sensitivity. However, the typical medical advice seems to go only so far as to say that “usually glucose levels go back to normal after birth”. I would love to see more data on what is actually going on in the post-partum breastfeeding momma. Thanks again!

  36. Randy Woodward So, there is a VERY important lesson to be learned here, which I have confirmed myself. On the oatmeal experiment, recognizing how much her blood sugar level spiked is one thing, but the more important thing was what followed later…”What was interesting, though, is that I got ravenously hungry when my blood sugar started plummeting.” What this means, and I’ve confirmed in my own experiences, is that even if you are a full-time low-carb eater, which allows you to access your stored fat for energy, which is ultimately the goal here, THIS ABILITY WILL IMMEDIATELY CEASE WITH ONE SERVING OF “NAKED CARBS”, as she puts it. With a good low carb diet, I can ride my bike for hours on end. As a matter of fact, I did the Leadville Race Series MTB 100, that stands mountain biking 100 miles at a 9,300 to 12,500 altitude, in 11.25 hours. I woke that morning and had eggs, sausage, and bacon. During the race, I had ONE protein bar, more so out of boredom, than need. And of course lots of water. Think about that! I had to burn 8 to 10k calories, and I had maybe 250 calories during the race. By the way, Gary Taubes taught me this could be done in “Why We Get Fat”. Anyway, so one day I decided to experiment on a 50 miles ride, that would last about 4 hours. I ate my no-carb breakfast as I always do, and set out on the ride. At the 1 hour mark, I stopped and had a donut and some orange juice, and set off again. Within an hour I completely BONKED! Which is exactly what Lily described…your block sugar level tanks, leaving you know carb energy, and your access to stored fast has been SHUT DOWN! It happens that quickly and that easily. Period! I’ve tested this theory several other times and in different environments, and it happens every single time.

    It’s NOT what were “told” about the benefits of eating low-carb that make it real or truthful, it’s about what we actually do and feel to prove to ourselves that it’s real and truthful.

    Interestingly, Atkins referred to this without really knowing why, at the time. He made a comment in his book that when you’re going zero-carb for a period of time, even a mint or a stick of gum can throw the whole thing off.

  37. Thank you for doing this! I would be interested in reading more if you ever experiment more in this way. I’d also love to see some continuous monitoring before and during the 3 hour GD test. For me, my goal is to pass the test (which I have always done, but sometimes barely) and then self regulate with a gd diet without a diagnosis or finger pricking. ( I have used three different home monitors and tested them against my blood results. EVERY SINGLE TIME my home moniter reads my fasting and post meal readings 20-30 points too high compared with my blood that was sent off to the lab, taken at the same time. So if I get diagnosed with GD, my provider will look at my finger prick numbers and say that I can’t control with diet and need meds.) Anyway, my ‘hack’ for the last 2 pregnancies has been: low carb, then super high carb for 2 days, than a day of super low carb and a ton of exercise, take the glucose test, then go back to a low carb diet. I dont know if that’s all necessary, so I’d love to see it tested changing each of the variables leading up to the test! Even better, I’d love to see the test changed to a not naked carb intake!

  38. Hi Lily,
    Thank you for taking the time to write up this piece and make it available for us. I am a Type 1 of 16 years, and I have struggled to find much useful cgm info from non diabetics (specifically individual day charts). Your oatmeal experience, especially the notes about being very hungry during the falling glucose, regardless of the actual level is a terrible phenomenon that we Type 1s deal with often. Unfortunately we are taught that we can eat whatever we would like as long as we cover it with industrial doses of our fast acting (actually very slow) meal time insulin. This behavior creates a roller coaster effect of levels that induces a lot of binge eating because of that feeling you experienced. I have just recently begun to be a low carb eater and get my cgm data to look a lot more like your normal data. I have always wondered what that maximum spike a non diabetic might experience when eating heavy carb foods, and the oatmeal chart was exactly what I have been looking for.
    Thank you again,
    Cheers!

  39. Fascinating, enjoyed the breakdown of your experience. In regards to your spike in blood glucose after the standard american breakfast, I’d imagine that’s to do with your low carb adaptation. If you’ve eaten mostly low carb for the trailing ~14 or so days (or even the past 24-48 hours before the meal really), you wake up fasted, and as a low carber have physiological insulin resistance, and slightly higher circulating ketones. On eating oatmeal this would have driven your blood sugar to multiples of that of a normal person who wasn’t on a low carb diet. If you at moderate amounts of carbs for three days in a row, and repeated the experiment, I’d guess the spike would drop down significantly. (same reason why you have to eat a significant quota of carbohydrates in the days prior to the OGTT for it to be statistically valid.)

  40. Thank you Lily! As a fellow RD who also has a low carb philosophy, it is great to read blogs like yours that are helping educate folks and hopefully get them on a healthier path! Keep it up!

    • Glad you enjoyed reading about my continuous glucose monitor experiment, Sandra! Cheers to more real food and low carb friendly dietitians. 😉

  41. Great work Lily. Thank you for challenging the status quo and for your investigatory mind. YOU ARE AWESOME and changing dietetics. Hugs 🙂

    • Thanks Tara. You’re right there with me helping to move dietetics forward. 😉

  42. Excellent self research! I have had a similar experience with testing my own glucose levels with a glucometer, especially with rice. I have 2 deceased parents with type 2 diabetes and am prediabetic myself. I am an almost 70 senior female and would love to try CGM myself. My daughter in law is type 1 and wears one and a pump and it has made a huge difference in her monitoring. Thank you for this self reporting, we need to see more of this!

  43. I am non-diabetic who tests my post prandiel blood sugar after meals for fun, well initially in an effort to lose weight and because I am seriously obsessed in my own data. I ate low fat, high carb for years because my father died from a heart attack and that was the recommendation. I have the same experience with carbs meals. My worst meal ever is a bagel, and it doesn’t matter if I slather it with cream cheese or put egg and sausage on it. I was even able to eat a donut after having a bowl of cottage cheese with less of a spike than the bagel. BTW I lost 13 pounds of post baby weight that I couldn’t lose for 15 years by ditching added sugar foods and reducing grains to a small portion a few times a week. I think your advice is spot on.

  44. Hello Lily,
    Good article. I am a Health Architect from the Netherlands and have been testing the device five times. Next to see what food is doing to your blood glucose levels, you can als see the effects of physical actions such as workouts, running, etc.
    What also is surprising is what taking a sauna is doing for you blood sugar. Actually it shoots your BS up to 200+!! It also falls back to normal levels pretty quick.
    With regard to your Thai food experiment I have my own observations. If I order (Chinese, Thai types) food from restaurants I will see a huge rise in BS even if I don’t order rice or noodles. I first thought it was due to a lot of sugar in the sauce. However, I after investigating it was not the sugar but more the MSG (monosodium glutamate) used to flavour the foods. Just imagine what happens if you are diabetic and go out to eat this.
    What is your experience with this issue?

    • Interesting, thanks for sharing your thoughts. I know a lot of Chinese food uses corn starch as a thickener in the sauces, so even if it’s not sweet, it has hidden carbs. I’d be curious to see how pure MSG would affect blood sugar levels, but I’m not about to sign up to test that out!

  45. Whoa! Lily. What a shock on your results over the oats! I have been making oat pan bread in the morning with only turmeric and water added, for my dad. (Plus giving him a drizzle of coconut oil over top as well.) He just moved in with us and was previously constipated all the time until he began eating this. He’s clockwork regular ever since. I eat it now too because it’s so good. But NOW I am going to add egg to the mix. Why? I was really wondering why I was feeling weak and hungry all morning after that oat bread. RAVENOUS by lunchtime. Of course! Carbs, regardless that they are laminated (rolled) oats. I figured: Hey. It has fibre and will slow down the digestion. Thanks for this eye-opening post. Glo

  46. So informative. Totally want to get my hands on a CGM. I did Robb Wolf’s carb test on bread, potato, white rice, oatmeal, and banana. The only one that didn’t spike my blood sugar over 130 after an hour was the bananas. All the others pushed me to 180+ and just like you, starving! Explains so much of my childhood to college years and feeling so hungry all day.

  47. I loved this article ! So much information! I wanted to ask you, for a pregnant woman struggling with very severe constipation and wanting to increase fiber content, without resorting to oatmeal and grains and high carbohydrate fruit (too much of it), what are good options? Thank you in advance

  48. Thank you for the writeup. I’m not sure I would have had the nerve to try a CGM experiment without having read this post.

    My father has longstanding (20+ years) T2D, gone insulin-dependent after a couple of years, and T2D runs in his family. I’ve been low-carbing since the South Beach Diet became popular, and exercising regularly since my early 20s; having seen what Dad deals with, I don’t want to end up T2D at 60 years old like he did.

    Dad got a FreeStyle Libre 10-day last spring, and has loved it. It suits him a lot better than a smartphone-type CGM would–he doesn’t do computers, even little ones. He upgraded to the 14-day Libre last week, and still had one 10-day sensor left, so he gave me the 10-day reader and sensor.

    Three days into the sensor, I have a good idea of what my “normal” is, and I’m about to start experimenting more extensively.

    My husband is also T2D, on meds only, moderately low-carb, doesn’t exercise, and rarely checks his BG. He’s open to the idea of getting a sensor prescription from his doctor and giving it a try. I believe it would be an eye-opener for him. It already has been for me.

    • Glad to hear you got your hands on a Libre for your own CGM experiment! It really is an eye opening experience.

  49. Excellent write-up! I wish these things had been around fourteen years ago, I had to work out what was happening longhand with a glucometer.

    I discovered that my glucose would spike around 1 hour postprandial, be almost back to normal at two hours and then at three-four hours it would drop low pretty much as a reciprocal of the spike. Not just the low but the speed of the drop was what brought on my (reactive) hypo symptoms due to the release of counterregulatory hormones, not just glucagon but cortisol, epinephrine, norepinephrine etc. Like “hangry” on steroids.

    I found the same as you with those “healthy whole grains” – wheat was my kryptonite, the only thing that spiked my glucose worse than wheat was wheat mixed with other carbs – I suspect wheat germ agglutinin – and curiously when I gave up wheat I mostly stopped farting and my GERD resolved. Who knew? Well lots of people actually, just most of them aren’t doctors.

    The other important thing I discovered was the pattern of insulin resistance being highest in the morning. When I started I was limited to about 15g carbs at breakfast and 30g by evening. Once I got my insulin resistance under control my breakfast dropped to about 10g but by evening I could eat 50 – 80g and sometimes 100g without spiking. Not that I do but I CAN – mostly I have stuck to around 50g carbs mainly in the evening. This pattern is common in Type 2s but many Type 1s have the opposite pattern. Non-diabetics? Who knows, are there many left?

    Seriously though I think the CGM is a game changer, there is a strong financial case for giving every diabetic and prediabetic one on loan for a few weeks now doctors are finally permitted to believe diabetes is “reversible”.

    Already though I have seen clueless dieticians (other types of dietician exist, obviously) claiming that it is dangerous to use a CGM because you may give up “healthy foods” like cornflakes and fruit juice just because they spike your glucose. One such claimed that since he found glucose spikes after breakfast, that just meant that “nondiabetics like him” also had glucose spikes. Ermmm, you may not be diabetic yet but you are already fat, just maybe the CGM is telling you something . . .

    A recent JAMA paper

    Assessment of a Personalized Approach to Predicting Postprandial Glycemic Responses to Food Among Individuals Without Diabetes

    hopefully more such research will be done to back up all the “just anecdotes”

  50. What concerns me is that all this religious monitoring of glucose levels seems to begin with the assumptions that:
    – all glucose peaks are bad
    – all large glucose peaks are the fault of the carbs

    Regarding the first assumption, what is the evidence that momentary glucose peaks are bad for health? And to be clear, I am referring to people who are not overweight and already in a disease state.

    Regarding the second assumption, I see no where in your write-up where you are considering non-immediate effects of your diet on your instantaneous glucose readings. There have been studies going all the way back to the 20s that show pretty clearly that fat consumption can have an impact on glucose response and insulin resistance and yet you seem to be completely ignoring that. It seems to me that if we ignore the non-immediate impacts of diet we are in danger of doing the equivalent of determining that the pain in our leg is caused by running or walking instead of the sprain we suffered days previous.

    As a side-note, I don’t find your comments on your “feelings” after a certain type of meal very compelling. Those “observations”, particularly on N=1 studies like this, are ripe for confirmation bias. I think the controversy over the sugar/hyperactivity relationship in children and the fact that many people on a whole-food, plant-based diets DO NOT experience shakiness or roller-coaster energy levels is enough to at least question whether your “feelings” are legitimate for everyone or even a significant portion of the population. There is also again the possibility that your “feelings”, if relevant and un-biased, are reflective of your diet as a whole and not just what you ate at your last meal.

  51. Lily,
    Great article, second time to read it. I saw you presentation at Low Carb Denver and now ‘experimenting’ with my CGM (I am non prediabetes) to see how accurate and correlate with my LCHF diet.
    My CGM (FreeStyle Libre) was showing 25 point higher than my KetoMojo blood control, not sure if is a sensor issue (poor answers from Customer service).
    I am a Surgical Oncologist, super interested in Keto and Cancer. But we need the basics of DM control first, horrified with my patients poor control and lack of diet knowledge.

    • Reach out to Abbott. That sounds like a faulty sensor.

  52. Perhaps your oatmeal experiment had everything to do with the added HONEY. Rolled oats leave much of the germ behind, leaving mostly endosperm vs. steel cut oats. Rolled oats also have a higher glycemic load and less fiber than steel cut. Perhaps steel cut oats with only berries would not have had the same effect.

    • To blame a 100-point spike solely on only the 5 grams of carbohydrates coming from honey (while ignoring the 40 grams coming from oatmeal + milk) is a far stretch in my opinion. I agree that it *contributed* to the spike.

  53. Pingback: The Best Way To Test For Ketosis & Track Your Blood Glucose. – No Fat Living
  54. I really like reading a post that will make people think.
    Also, thanks for allowing for me to comment!

  55. Just an FYI, the Libre tends to read low, especially at night. They actually have a disclaimer about this. Also, not sure how low-carb you typically eat, but if you primarily eat low-carb, a sudden influx of carbs will cause a bg spike as the enzymes needed to handle a carb load are made on an as needed basis, but there is a delay. It takes about 72 hours to start making them again, which is why it is recommended to eat at least 150g for 3 days prior to an OGTT. A low-carber will fail an OGTT even if they are not diabetic.

  56. Thanks for sharing this. I’m a 42 years old man and was diagnosed with T2D with an A1c of 12.1% this xmas. I have been wearing a Libre CGM on and off and made significant changes to my diet (I still eat carbs but more moderately usually around 30-45% of total calories–and trying to focus on whole grains and vegetables) and added some exercise. I’m also a vegetarian. It’s super helpful to me to have data from non-diabetics to compare to as I’m trying to get my own diabetes into remission (my last A1c was 5.2% so I’m making good headway :-). So thank you so much for sharing your experiences and a little bit of non-diabetic raw data. A few comments on my experiences.

    1) Oatmeal causes a big spike for me as well, it’s almost exactly the same shape/profile although actually not usually as high. This was a big surprise to me as it was my breakfast of choice before wearing the CGM. I’ve switched mostly to Chia seed pudding now.

    2) I’ve only worn a few sensors so far, but have read obsessively in diabetes forums on other’s sensor results and so far my experiences have been similar.
    a) The Libre sensors can vary quite a bit.
    b) Most sensors tend to error low most of the time (compared to pin pricks), but some “hot” ones error high in the mid range (my current sensor often puts me 15-20 higher than pin pricks on average–my last one put me 10-15 lower on average). (I have 4 meters and they vary some amongst each other too–but my AccuCheck Guide and an ContourNext One are usually pretty close to one another and pretty consistent with multiple tests in the same time period)
    c) It seems many report it’s fairly common for the Libre sensors to report errors at the extremes in both directions (highs too high, lows too low). This is also consistent with my experience. That might be a deliberate calibration for safety?

    3) It’s getting harder as the 14-day sensors now encrypt the data and don’t work–but there’s fascinating 3rd party products (search BluCon or MiaoMiao) that continually scan and relay the data to phone from the older devices, and 3rd party apps like xDrip or Spike that can perform additional calibration with pin pricks to improve Libre accuracy. Fascinating how active the diabetes DIY community is. I haven’t been able to try these as I have a 14-day sensor/reader.

    I also appreciate Daniel’s comments above. I find it’s easy for me to start villanizing all glucose spikes (and that conclusion doesn’t seem entirely unreasonable), but I’m not sure there’s sufficient data to really support that. I do find that most large peaks for me do seem to center around carb consumption though (although other factors do also affect me, particularly sleep).

    Today I ate one of the largest carb meals I’ve eaten in a long time–a potato breakfast hash that I estimated at 121g of carb. Surprisingly rather than causing the massive spike, I’d anticipated/feared–I ended up with a prolonged high (120-135) for around 3 hours. This may support your idea of not eating “naked” carbs–as my hash included eggs, cheese, and oil as well.

    Cheers!

  57. Glad to have found your blog. My response to keto is similar with a CGM. A stable line with little variation around 65-70 in winter and 75mg/dl in sommer. It took me about 2 years to get there. Now I can tolerate some fruits without spikes in BG. Spikes up to 90mg/dl happen only with more exercise. Thankx for sharing!

  58. I have just been wearing a cgm for fun and was horrified to see my glucose go to 11.2 mmol after some oats ! I am on a low carb diet usually and worried this has ruined my carb tolerance but reading your blog has reassured me it’s the oats! That’s how toxic they are!!

Comment Policy