Diseases, Food, science

We don’t know what to eat. Our blood glucose will show us.

Your blood glucose levels can show you what to eat to avoid type 2 diabetes, obesity and heart disease (among others).

Here is why, and how.

Untitled_Artwork 3

As part of your general checkup with your doctor, you’re probably getting your fasting blood glucose measured every 3 years1, if at all. Based on this number, you’re told whether your blood glucose levels are within a healthy or unhealthy (pre-diabetic/diabetic) range.

  • If your fasting blood glucose result is abnormal, keep reading to understand how you can improve your blood glucose levels.
  • If your fasting blood glucose result is normal, you may think you don’t need to make any dietary or lifestyle changes. However, new research shows that you may be in fact unknowingly inching towards diabetes2.

Here’s the problem: when you get your result back, you’re none the wiser about how to eat to prevent getting an abnormal result next time, or how to improve your glucose levels in the case of a (pre-) diabetes diagnosis.

Fortunately, there is now a a-cool-technology-that-i’ll-talk-more-about-in-a-sec that provides more information on what is going on in between those triennial checkups. Thanks to it, I discovered what I should eat to keep my levels in the good zone forever (spoiler alert: it includes burritos)

But first, a tiny recap on the importance of blood glucose.

Screenshot 2019-03-07 at 20.56.21

Chronic high levels of blood glucose are bad. They put you at increased risk of type 2 diabetes3, heart disease4, fatty liver disease5, and are associated with obesity6, some cancers7, and Alzheimer’s8. To add insult to injury, 1 in 4 people in the US have clinically high levels of blood glucose (pre-diabetes) and don’t know it9.

The main reason your glucose levels become too high is that you eat things that don’t work for you. And “what works” is different for everyone10, as you’ll see below.

Recap over.

Now here’s how to get out of this mess.

Three friends and I put on $45 Continuous Glucose Monitors that we got over-the-counter at a pharmacy in Europe (I’m french, bonjour!). These wonderful things are only available via prescription in the U.S.

Screenshot 2019-03-07 at 21.02.47

It’s painless to put on, you can’t feel it, you can shower and do all the normal human things while you have it on. These devices last for two weeks. While we had them on, we recorded everything we ate and when we exercised.

The monitors measure your glucose levels11 every 15 minutes and plot them for you. This is an example:


The green range represents the “good” range for non-fasting, non-diabetics, as defined by the NIH (< 6.9 mmol/L, or < 125 mg/dL). It’s critical to keep our glucose in this range as much of the time as possible to reduce risk for all the diseases above. 12,13 

And now, for the results.

All in all, our glucose was much spikier than expected – even though we were all healthy, not overweight, 20- something (and one 50-something) year-olds. Here is a daily glucose graph and the corresponding meals and workout.


We were surprised how often we were in the red zone, usually indicative of having pre-diabetes.

And since every spike into the red inches you more towards issues, we really wanted to avoid them. Studies show that problems arise with any increase in average blood glucose16, which means, any time spent in the red zone.

So I built a tool to measure the glucose spikes of individual meals. It uses a score – the bigger the number, the worse spike that meal created, and the worse that meal was for our blood glucose. (The score is called “food line” on the graphs).

The coffee shop decision

We got a clear winner for what to pick at the coffee shop: unsweetened flat white (just espresso and whole milk) beats an oat milk matcha latte hands down.

Quick note: science has shown that in a single person, the response to a food will be pretty much the same over time. So a matcha latte on Tuesday will do the same to your glucose as on Friday10.

Different person, different effect

You are unique. Your genetics, your age, physical activity, the types of bacteria in your gut, your levels of various biomarkers, all contribute to differentiating you from your neighbor. These factors also influence what your glucose response to a meal will be10. Researchers are finding that there is no blood-glucose stabilizing diet that works for everyone.  A food that spikes my blood glucose, won’t spike yours, and vice versa.

Case in point: cookies!

4 cookies had a terrible impact on my blood glucose, but did virtually nothing to Jeff. We ate them at the same time.

Untitled_Artwork 9Untitled_Artwork 8
I was pretty shocked at the sight of this result (ok, maybe “sad” is a better word). A blood-glucose-stabilizing diet for me would not include cookies (sob, sob), but they could be included in bob’s diet.

There is more and more scientific evidence supporting vastly different personalized response to an identical food.10

The easy choice

I compared –on different days– two mexican foods that I usually order interchangeably (a plate of nachos with pork vs. a large pork burrito). I don’t have a strong personal preference for one over the other, they have the same amount of meat, and similar amounts of carbs.15,16

In terms of glucose spike, though, the difference was startling:


The burrito was a much better choice. I dug a little deeper and thought, well maybe this is due to burritos having more fat in them, which slows the glucose response down? But nope! Nachos have more fat.15,16 Fiber is slightly higher in burritos, which may explain a little bit of the difference…

But ultimately,  I really just have a different response to these two very similar foods.

I order burritos from now on. 😊

In fact, the burrito did so well for me that it even beat my go-to low-carb cauliflower and chicken salad. That was remarquable.


When you love a food, hack, hack again

Some foods are non-negotiables. We need them in our lives. Cookies, I can do without, but pasta and parmesan, no way Giuseppe.

Unfortunately, sometimes things you love don’t love you back.

On march 3rd, at 6pm, I ate a big bowl of spaghetti with 3 boiled eggs and a LOT of parmesan. Sad news, the response was bad. Food line = 18.

*eggs_at_1145_pasta_1230 3

Determined to hack my way to eating pasta without spiking my glucose, i tried a few things. I’m happy to report that something worked, and I can now eat pasta without the spiking: I ate a 3 egg omelet with olive oil 20 mins before the bowl of pasta. Same amount of pasta though.

Success! Much smaller spike, food line = 1.

*eggs_at_1145_pasta_1230 2
I have a hypothesis about why this happened, but don’t know for sure: maybe the omelet being digested first slowed down the digestion of the subsequent pasta… There is much I do not know. But I’m happy with the result.

The confusing meals

Meals we thought were healthy were creating spikes in the red zone, like these ones:*stew

Other meals that we thought less healthy, weren’t that bad in terms of spike.

*sandwich and tater tots

Spikes and hunger

Feeling hungry, tired or cranky often correlated to blood glucose levels that were low or dropping rapidly right after a high spike. Here I noted with the blue lines a feeling of hunger.


Intense bouts of hunger seemed to correlate to our blood glucose dropping right after a big spike. The more stable our glucose was, the less intensely hungry we felt.

So what now?

First of all, I want to repeat my Diabetes at 36,000 feet experiment hooked up to a CGM.

Second, I dream for everyone who is interested to have access to a CGM to discover their own unique response to food. Just imagine how many metabolic problems and disease risk factors we could avoid with this information.

Food is personal, your body is unique, and information is the solution.

You can email me (full details here) if you have CGM data that you want me to run through my Food Line program. And stay tuned for individual case studies!


Caveat to this CGM approach for figuring out what to eat.

I think a glucose spike result should be used to choose between two nutritionally equivalent foods, and that you should run many experiments on yourself to find the perfect match between a nutritious diet and stable blood glucose.

For example, it’s not because cookies have a food line of 2 and salad a food line of 15 that you should replace salads with cookies in your life.

A way to think about it is: all else being equal, like vitamins, fiber, carbs, fat, how processed the food is, what’s the better choice for me?

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Thanks to Thomas! 

And mega props to https://apps.automeris.io/wpd/


(1) NIH Health screenings per sex and age. 

(2) Diabetic-level glucose spikes seen in healthy people

(3) American Diabetes Association Prevention or delay of type 2 diabetes. Diabetes Care. 2015; 38: S31-S32

(4) Gallwitz B. Implications of postprandial glucose and weight control in people with type 2 diabetes: understanding and implementing the International Diabetes Federation guidelines. Diabetes Care. 2009; 32: S322-S325

(5) Nishida T. Tsuji S. Tsujii M. Arimitsu S. Haruna Y. Imano E. Suzuki M. Kanda T. Kawano S. Hiramatsu N. et al. Oral glucose tolerance test predicts prognosis of patients with liver cirrhosis. Am. J. Gastroenterol. 2006; 101: 70-75

(6) Blaak E.E. et al. Impact of postprandial glycaemia on health and prevention of disease. Obes. Rev. 2012; 13: 923-984

(7) Lamkin D.M. et al. Glucose as a prognostic factor in ovarian carcinoma. Cancer. 2009; 115: 1021-1027

(8) De la Monte, Suzanne M., and Jack R. Wands. “Alzheimer’s disease is type 3 diabetes—evidence reviewed.” Journal of diabetes science and technology 2.6 (2008): 1101-1113.

(9) The CDC, Diabetes Prevention (note: the “1 out of 4” number comes from a rounding of 90% of 1 out of 3.)

(10) Zeevi, David, et al. “Personalized nutrition by prediction of glycemic responses.” Cell 163.5 (2015): 1079-1094.

(11) They actually measure interstitial glucose levels. More information here.

(12) Riccardi G. Rivellese A.A. Dietary treatment of the metabolic syndrome–the optimal diet. Br. J. Nutr. 2000; 83: S143-S148

(13) Livesey  G, Taylor  R, Hulshof T, Howlett  J. Glycemic response and health: a systematic review and meta-analysis: relations between dietary glycemic properties and health outcomes.  Am J Clin Nutr. 2008;87(1)(suppl):258S-268S.

(14) National Nutrient Database for Standard Reference Legacy Release: nachos

(15) National Nutrient Database for Standard Reference Legacy Release: burritos

(16) Association of Hemoglobin A1c with Cardiovascular Disease and Mortality in Adults: The European Prospective Investigation into Cancer in Norfolk


5 thoughts on “We don’t know what to eat. Our blood glucose will show us.

  1. Robert Dyson says:

    Extremely valuable study. So many results were unexpected, including large differences between people for same food. I am going to try a CGM device. This is inspirational.


  2. I’ve found that you might have to run some experiments several times to isolate the cause and effect. For example, the burrito vs. salad comparison seems very counterintuitive and could be because of some other variable. There are many that affect your levels at a given point. Great post!


  3. Andrea says:

    Great to see this. Useful links, useful information. Could not find the CGM in Mexico but it is available in Canada without a Rx.
    Thanks for the new post!


  4. Pingback: The anatomy of glucose curve: cupcake edition | HEALTH HACKING

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s