Just spent 4 weeks tracking my blood sugar with a continuous glucose monitor, and here’s what I learned!
Just did another 2-week round of blood sugar tracking (using the same setup), as my daily routine has changed since last year, plus there were a few things I didn’t get to try last time…
Here are my take-aways:
- My blood sugar stays stable while fasting for a full day. I don’t currently plan to fast regularly, but good to know?
- Foods with rice flour (incl rice noodles) definitely cause large blood sugar spikes, and eating large portions of fries (sweet potato, or regular) results in hours of elevated blood sugar (indigestion?).
- Taking digestive aids doesn’t seem to do anything (other than strip the enamel off my teeth?).
- Blood sugar appears to spike a bit after walking outside in the cold.
- The hot shower after drinking the post-workout milkshake may deserve some credit for the moderate blood sugar response.
- Watching movies while lying on a bean bag may be problematic.
- Track blood sugar on an overnight backpacking trip.
- Track blood sugar after swimming (incl sauna).
- See if drinking a coke with the fries helps.
Eric, really interested in your setup (must have missed the original discussion on the setup). Are you using a CGM? If so, how did you get one?
Also, I’m curious if you’ve done any of the carb tolerance testing that Robb Wolf talks about. It might help you understand more about the carbs that cause concerns.
The setup is described in the blog post linked above.
Doing a “carb tolerance test” could be interesting, but not all that useful, as you’d be taking the food out of the context you’re usually eating it, and if this experiment taught me anything, it’s that context (like time of day, activity levels before and after etc) matters!
Hi @ejain, thank you for sharing these results!
It continues to amaze me many things effect blood sugar to such a strong degree.
I would like to run an experiment to determine which is healthier: (1) eating a salad at my desk or (2) eating a delicious burrito from the food truck a block away, where I have to walk before and after the meal.
My question is: do you (or anyone) have resources where I can learn about the effect of blood sugar spikes on human health? I’m very curious about the research on how resting levels, spikes and drops effect health and how direct the relationship is.
Blood glucose resources anyone?
Thanks for this, looking forward to your next post!!
Thanks for posting this, @ejain!
I just inserted my Dexcom CGM sensor on Wednesday. I’m not diabetic–just tracking for interest. Maybe want to track with mood and energy levels and sleep. Maybe weight control. Maybe gym performance. So I now have about 48 hours worth of data.
I’m seriously impressed with your data collection and analysis–wow, thanks for the link to your Jupyter notebook!!
I have started thinking about how I want to connect the glucose data to the other things I want to compare it with. So this was a big help. Thank you.
Do a search for “diabetes complications” to get an idea of the long-term effects of elevated blood sugar (>140mg/dL). Surprisingly, low blood sugar doesn’t seem to cause any health issues (other than death).
Your “post-prandial” blood sugar may be a better predictor of problems than fasting blood sugar or A1C, but it’s less easy to measure, plus it’s still unclear if it’s an actual risk factor, or just a marker.
Please share, if you learn anything interesting
Looks like Dexcom gives you measurements at 5min rather than 15min intervals?
Had a “wellness exam” this week where my fasting blood sugar (from a finger prick) was in the pre-diabetic range (115mg/dL), while the A1C was fine (5.1%, which corresponds to an average of ~100mg/dL).
This surprised me much less than it did the doctor, since I now knew that after walking half an hour in the cold to get to the doctor’s office, my blood sugar would spike briefly!
Yep, 5 min intervals. It makes beautiful smooth curves
That’s so cool that you knew what caused the spike!
I will definitely share.
So far, I’ve been trying to isolate single things: like: a weightlifting workout seems to cause a very slight rise in BG, and by the time I’ve showered and walked to my office – there’s a pretty decent dip. At some point I’ll see if a protein shake fixes that, or if it needs carbs too. And how many
I have also noticed that a glass of wine doesn’t seem do much to BG for me. I was surprised.
I don’t quite know if I’m aiming for something: like the steadiest line? or no sharp spikes? or the lowest steady state… or if I’m just recording. Experimenting for now.
But I’m absolutely fascinated with it.
This is really nice, Eric!. Great experiment!
How did you find working with Altair over the more popular Matplotlib? The plots look great, but I’m curious how you found working with/customising it over any alternatives you’ve worked with in the past.
For starters, Altair doesn’t spike my blood sugar quite as much as matplotlib
Doing a third round of blood sugar tracking, summer edition!
- I’m now using the LibreLink app to scan and to manually record meals. This app hadn’t been available for Android in the US previously.
- Previously, my fasting values were around 80-90mg/dL, with spikes up to 170. Now, fasting appears to be around 60-70, with spikes up to no more than 150. Maybe I need less antifreeze in my blood in summer, or (more likely) the sensor is simply off by 20…
- The sensor survived 20 minutes of lap swimming and 5 minutes in a sauna.
Finally got to do a more extended hike while tracking blood sugar, this is what it looked like:
a. ate some yogurt with muesli for breakfast
b. started hiking, drinking a cup of plain water every half hour or so
c. shifted to burning fat after 3.5h?
d. reached the summit, ate a landjaeger and a piece of bread
e. sports drink (Vitalyte) containing 22g of sugar (and vitamin C, which could throw the sensor off)
f. back at the car
g. late dinner with lots of white rice, bad idea
Yeah, I had a lot of issues with freestyle sensors being low by this and more (which I was using when I had gestational diabetes last year). Do you not have some test strips to check against? I was doing that pretty regularly.
Not currently; I did some checking with test strips when I had the first sensor, but it left me none the wiser because I got results that differed by 20mg/dL using blood from the same finger prick… Also, my fingers don’t bleed easily
In your experience, was the sensor off by a consistent and constant value?
IIRC yes in the short run, but not so consistently in the long run. The freestyle always ran lower, sometimes wacky crazy low. The company was unsupportive with help since it’s not FDA approved for use during pregnancy.
I had trouble with drawing blood too. I always went for the side of my ring/pinkie finger, and I took the safety off the lancet device. The CGM was so much easier! mostly I used it to get a feel for the effect on glucose various foods had (based on spot checks I concluded the delta glucose was accurate, just the value was consistently off), it’s so much easier to learn from continuous data than with test strips.
I’d complain and ask for a replacement, but the sensor I’m using is slightly past it’s use-by date
Another factor I had forgotten to take into account: I started taking magnesium (240ug/day), which apparently can reduce your fasting blood sugar levels (though the effect is probably much larger in magnesium-deficient diabetics)…
Good news, my hypoglycemia has been cured… by switching to a new sensor
Readings are now around 10mg/dL higher on average, but still 10mg/dL lower than previously seen. So the question remains whether the lower blood sugar is a seasonal pattern, or if it’s due to magnesium supplementation, the use of a standing desk, or just more random sensor variation…
Doesn’t look like it does… Including a highly acidic drink (Diet Coke®) does not seem to alter the response to a high-fat meal (a cheeseburger with a generous serving of sweet potato fries) much:
Turns out saturated fat is known to temporarily increase blood sugar, so it’s no surprise that simply speeding up digestion won’t help.