CGM data, Carnivore and Blood Glucose levels versus protein intake

(Michael) #1

Stats: Male, 168 cm, Weight is 142 lbs or 64.3 kg (started at 139 lbs/63 kg ended up at 64.3)

After doing a fast in August I went Beef only carnivore, at the forums suggestion that protein was irrelevant to blood glucose, even as a diabetic (despite my research from before I joined this forum stating otherwise), I decided to go ahead and eat very high protein loads and just keep tabs on my glucose readings. Using my finger prick blood glucose monitor, it seemed quite apparent that my glucose readings were very much affected by my protein intake. The problem with finger pricks, is they are unreliable. I cut back on the protein and my finger pricks returned to normal levels. How much of an effect does protein have on blood sugar levels though? In order to determine this, I planned and tested a 10 day CGM protein test. I also wanted to see my body’s reaction to the re-introduction of food/protein after a long fast.

Here is what I did to prove (to myself at least so I can plan my diet according to the facts and not the unsubstantiated dogma) that protein affects blood sugar levels.

  1. Fasted for 7 days straight, I put the CGM during the last day of the fast/refeed day
  2. Ate as much protein as I could handle for the first 4 days (I actually got nauseous and had to cut back on after day 4, I had stomach/gall bladder pains and diarrhea) and brought my glucose readings above my expected baseline (based on daily measuring for the past 4 months) - up to 5.4 mmol/L
  3. Cut back to two protein meals a day and let my glucose drop down to a steady state for 3 days (shows with around 650 g of meat a day, my average glucose is about 5.1 mmol/L).
  4. When my baseline was established and I physically felt better, I greatly increased my meat intake and watched as my glucose levels quickly rose by a full point up to 6 mmol/L in just 3 days of extra eating.

Conclusion: There is absolutely zero doubt that protein intake (or at least intake above the standard doses) greatly affects my blood glucose readings. I am a type 2 diabetic (partially healed now), so the results may differ for others. However, based on Shawn Baker having an average glucose of over 7, I suspect this is true of everyone, not just the metabolically challenged. It is probably more noticeable with type 2 diabetics. It seems to take a half day/day for the extra protein to equate to higher blood sugar levels as the extra protein is converted to glucose.

Day 1
broke fast with 500 mL bone broth, 12 oz striploin and 7.5 bacon (see CGM for times), 20.8 oz + 500 mL = 590 g + bit = 600 g or so

Day 2
ate 1 lb buffalo ribs, 11 oz striploin, 6 pieces of bacon (200 g pre cooked = 7 oz), total 34 oz = 964 g

Day 3
ate 14 oz buffalo ribs, 14 oz T-bone+6 bacon (200 g pre cooked) , 35 oz = 992 g

Day 4
today ate 10 oz striploin with 3.5 oz liver, 10 oz ribeye and 7 bacon (8.2 oz = 233 g pre cooked), 31.7 oz = 899 g (started to feel stomach upset)

Day 5 - Felt nauseous and could not eat a lot this day, reduced eating a bit this day
could not finish 11.5 oz buffalo steak, ate 8? Oz for breakfast + 10 oz ribeye + 7 bacon and still could not finish buffalo, 26.2 oz=743 g

Day 6 reduced to 2 real meals a day for 2 more days
ate 10 oz striploin, small amount of buffalo tongue (3.5 oz?) with bone marrow and 7.5 bacon (8.8oz = 250 g pre cooked), 22.3 oz = 632 g

Day 1- though 6 daily Statistics

Day 7 - established baseline at 5.1 mmol/L over past 3 days
ate 10 oz ribeye, ate 7 bacon (233 g pre cooked) with 2 oz buffalo tongue, 20.2 oz = 573 g

Day 8
ate 10 oz buffalo ribeye, 10 oz striploin and 14 oz T-bone today, 34 oz = 964 g

Day 9
meal 1) 14 oz ribeye + 3 oz liver 2) 11 oz ribeye 3) 12 oz buffalo ribeye, 40 oz = 1134 g

Day 10
meal 1) 14.5 oz ribeye 2) 11.75 oz ribeye 3) 10 oz ribeye, 36.25 oz = 1028 g

Day 7 - 10 daily statistics

Looking for any interesting observations, thoughts or concerns about the data. I will note that in theory, my last two days MAY have been as high as 4 g of protein by kg of bodyweight. Certainly more than needed.

(bulkbiker) #2

Did you cross check your CGM readings with finger pricks to assess accuracy?
You state in your opening that finger pricks are unreliable yet for many CGM’s are even more unreliable…
Which one did you use?

(Michael) #3

Hi Mark,

Finger pricks are unreliable because you get very few data points - not because the machine is faulty (yes, there are some variations even within these, but I find mine to be VERY consistent, whereby 4 measures in a row read exactly the same). I calibrated my CGM twice daily to match the finger pricks - you can see in the diagrams when I calibrated and what my calibration was as compared to the previous data point from the graph. In general CGM can be very accurate as well (my first checks matched perfectly), but over time they get thrown off, and need to be calibrated. Both times I used my Dexcom it was exactly the same as my finger picks for a few days, and then started to fluctuate more and need more calibrations to keep it on track as the days passed.

I also can upload/download all information in an excel file, but I do not think that is needed for these forums.


any food or drink ingested will change blood glucose on some meter.

fact point blank.

every ingested food product WILL effect glucose, simple as that.

maybe read about dawn effect and more and read this too and maybe some of this means something to you

and remember ALL the foods you are not eating that rot and effect your veins and BS issues also…for all you don’t eat, even if your BS is effected by anything you do eat, you are in the better for your changes ya know. It is never all about 1 number, it is about a miriade of things combined so think a bigger picture also in a way :slight_smile:

and research Dr Bikman he is a massive insulin person, thru insulin resistance and all about the insulin our bodies and how insulin response thru glucose can effect fat loss etc…so yea this guy is ALL about insulin and might be a source for you to check out :sunny:

(Bob M) #5

Are those times when you ate, where it says “0 grams” and has the fork and spoon?

To me, not much is happening. The main thing affecting your blood sugar appears to be exercise. There are some slight increases for some times when you eat, but I can’t tell how much or why.

To me, none of this is a big deal.

(Michael) #6

The fork and spoon with 0 grams signifies that I ate my first meal. I listed those in front of each diagram so you can see “what” I ate at those fork and spoon times. For example, if you look at day 9, wed sept 29, you see I ate at around 10 am for my first meal. It is listed as 14 oz ribeye steak with 3 oz liver. The glucose spike is typical for most days of my first meal (a rise for a few hours immediately following), and a general rise in the daily average which then needs to take into account the total daily protein intake.

I do notice that I forgot to add in my second meal time on day 10 in my dexcom, it was at 1 pm on day 10 (oops).

I am not worried about the spikes, since they are small, more the fact that my average glucose goes from 5.1 up to 6.0 in just 3 days of increased protein. I consider an average glucose of above 6 to be too high, or at least I am concerned it is too high if not convinced. Thankfully, I noticed through finger pricking before, that I can lower it back to 5.0 by removing the extra steak per day of intake in just a couple of days. While I no longer have a CGM on, I will stop taking measurements for a few days, and then verify that my average is closer to 5 in a few days from now.

On Oct 6 I will be getting a1c, insulin, glucagon and creatinine blood levels (oh yah, and standard lipid profile).

(Bacon is a many-splendoured thing) #7

What happens when you forget about quantities and simply eat to satisfy your hunger? Does your body go haywire, or do things tend to sort themselves out?

(Michael) #8

I can (very quickly) acclimate my body to eating 3000+ Kcal / day or half that. Once I acclimate, I am fine with the new value. So, for example, in August after my 10 day fast, I started eating 3 steaks a day. After about 2-3 days, I was hungry first thing in the morning, and then again every couple of hours after that. I was craving 3 steaks a day and sometimes was hungry for more. When I decided to reduce to TMAD to reduce daily protein (but stay carnivore with the food I had), I was hungry for a few days in the mornings, but after about 2-3 days, I was fine just eating two steaks a day and felt “less hungry” than when eating 3 steaks a day. Outside of pushing myself a bit quickly after this last fast, same kind of idea. I am “satiated” over a day on a wide range of calories once I give myself just a few days to acclimate.
However, if I had to choose for the rest of my life, I would “prefer” to eat more. Despite eating over 1kg a day recently, I have not eaten yet today, and I would rate my hunger at a “4” currently. Having said that, I could easily eat 1 pound of steak right now too. No idea if I am normal in this regard, but it helps for experimenting at least.

(Bob M) #9

So, 6x18 = 108, in US units.

I guess I’m having a hard time analyzing this data. For instance, I see an amount of meat, but no protein listed. While there’s some correlation between meat and protein, this can wildly vary. A pound of ribeye has 1240 calories and 80 grams protein, 73% calories from fat, while other means are much leaner (liver, for instance).

You seem to be comparing days with high calories with low, but I’m not sure it’s protein per se causing this.

And any day you exercise, you’ll have a higher overall blood sugar. Or at least I do.

If you really think it’s “protein”, you have to count that, and compare only similar days together. That is, compare only days you do not exercise.

Also, if you want to know whether it’s “protein”, try to eat the same calories for a few days (no exercising). Eat ribeye one day and lean meats (eg, top round) another day. Compare these (ribeye = higher fat, while top round = higher protein).

(Michael) #10

I would have preferred to have 30 identical steaks in my freezer for this experiment, but I did not at the start at least. However, you really only need to look at the last 4 days in the data, since that is after I initialized the baseline off which to compare (ate past the baseline after the fast so that it was no longer a concern AND kept a constant value of about 5.1 mmol/L for days eating a “standard for most” amount of protein in a day). I ate mostly ribeye’s during the final days (all from the same cut of beef as my delivery arrived late on day 6), and the differences in meals exists I am sure with some non ribeye choices/additions, but not in huge quantities of proteins overall. If you look at just those last 4 days alone, you can see the trend quite clearly (noting that if I eat a lot of protein one day, it shows up the next day, not immediately like carbohydrates).
How does calories affect blood sugar? I am confused by this reference. AFAIK, it is grams of macronutrients, having nothing to do with total calories. And therefore, while we might not have an exact protein count, if I eat 2 steaks one day and then 3 of the same steaks (essentially) the next day, the only thing that should change my blood sugar is the content of the macronutrients (essentially either protein or fat in the extra steak), and since fat does not really effect blood sugars as compared to protein, that would leave the greatest effect of the change due to more protein. Unless you feel the extra fat is also a significantly contributing factor to blood glucose levels? Am I missing something else with your calories reference? I do not get it.
Also worth noting, I exercised and had 5.2 mmol/L, the on Monday next day I did not exercise but went up to 5.7 mmol/L on Tuesday and finally I did exercise on the last day and went to 6.0 mmol/L. Exercise makes a difference, for a few hours, but not that much or for an average in a day, not even close (on a whole day it MIGHT make a 0.1 to 0.3 change for the days average). As noted though, you cannot get this from finger pricks, you need a CGM over days to note any such changes. I have been paying attention, both this CGM and the previous 10 day CGM.
I guess my response is you raise good points, all of which I think I have covered off (except calories which I did not consider important, simply the volume of the macronutrients which while related, is not the same thing - am I wrong here?). And you could argue that I do not have exact proteins (again, you are correct, I do not), but I cannot see how comparing 2 steaks to 3 steaks isn’t a clear cut case of more everything, of which only the protein is essentially relevant at this juncture.
Regardless, the data seems crystal clear to me (again, unless I am missing something about calories in addition to macro-nutrients), and I am sharing to help others, so hopefully it will (I am a numbers guy in general). If not, at least I know how protein affects my body and will re-formulate my diet accordingly (for now, up until my Insulin Resistance is greatly reduced again). I will be including dairy fat again, as I can stomach a lot of fat without issue or the inconvenience of having it packaged with protein.

(bulkbiker) #11

On a finger prick meter those readings are well within the permitted margin of error i.e. could all be exactly the same… in a CGM (which is even more inaccurate) they could too be exactly the same reading…
I think you are drawing too many conclusions from what is pretty much the same level.

(Michael) #12

But yet, when I repeat multiple times, I get the same results. I have redone the experiment again with finer details and have the exact same (essentially) type of results over the past 10 days. I used protein powder additions for many days to avoid any other confusions by people as well (data to follow in a week since I am too busy to put it together today - plus I should wait on my insulin results I just took after the experiment and before) There now is no possibility of error readings coincidentally giving thousands of results that match (if we consider taking 290 measurements a day). Anyway, I will post new data, and I also now have graphs of protein versus average glucose that I can use based on my HOMA-IR and QUICKI numbers. This last test buried any type of arguments to the data, outside of someone just deciding to ignore the data because it did not fit their narrative.

(Michael) #13

Hi Fangs,

You sent me that link before, so thank you again. I have found Ben Bikman’s stating the same thing in one of the first videos I watched by him. Here is the quote and the link for your records as well, since Bikman probably holds slightly more sway with the indoctrinated.

“some people will eat protein and will in fact have massive glucose spiking effect, and it could be that if someone becomes a type II diabetic, their glucagon producing alpha cells have become insulin resistant.”….“that might be part of why the very insulin resistant type 2 diabetic, … will in fact find as they adopt more carnivore or protein heavy version of a ketogenic diet, they might have some struggles with their glucose and they might have problems with their insulin,… especially (if they were) to the point of diagnosed hyperglycemia.”

Unfortunately I am now clearly one of those people specifically mentioned by Ben. Again, I have all the data and numbers which match over multiple tests now, so I will graph and put it out when I can. Quite interesting, and allows me to know my eAG based on calculations instead of guesswork.

(Jane) #14

Interesting - thanks for posting your data. Have you gotten fasting insulin and glucose data yet to calculate your HOMA-IR score?

I paid for the tests myself when I was 9 months out on keto and wished I had a baseline because I probably was IR before keto, but not diabetic.

My results:

Glucose: 83 mg/dL
Insulin: 4.6 uIU/mL
HOMA-IR: 0.9

I believe less than 1 is optimal so was happy with my results. I will run out of blood ketone strips long before glucose and have the older Keto Mojo so can’t get anymore strips for it. I was wondering what cool experiments I could do with just the glucose strips, such as check to see if stevia or Truvia raise my glucose. I could do one with protein now to see.

Not as comprehensive as your data but you need a prescription to get a CGM in the US and I doubt my doctor would write me a prescription with my numbers. I haven’t asked so could be wrong.

(Michael) #15

I did bloodwork on October 5th, in between my two sets of 10 day CGM data, and I JUST sent blood in for insulin and glucagon today, so will take 2 weeks for the I/G ratio update. In terms of my oct 5 results, my HOMA-IR was 2.6 on that day, and my QUICKI was 0.33. I put the reference ranges for IR as well so people can compare. With an HOMA-IR under one, you would see zero glucose fluctuations from protein (so I am truly jealous :slight_smile: )

QUICKI= 0.331423302 >0.3 = IR
HOMA-IR= 2.633650794 >2 = IR

(Jane) #16

Awww… I hope you get there also - maybe these latest results!

That was 3 years ago so I should probably test again. I live in the middle of nowhere so no labs within an hour’s drive to draw the blood. I was in Ft Worth for a week so scheduled an appoint with the local LabCorp. Next time I am in a big city I can do that again.

(bulkbiker) #17

Maybe look up the definition of “massive” cos looking at your CGM graphs you sure as heck aren’t experiencing that.


yea you must keep everything in true proportion ya know and ‘massive’ might not be a ‘wrong thing if off all plant matter’ and your body is healing so?

tons of shades of gray

but also, there are ‘unicorns’ truly in every aspect of life and how they do effect us when we make changes so? Again, unicorns exist for sure and if one then more research and learning thru what one needs is just that, go off and learn more :slight_smile: and change as that individual needs.

We can blanket statement alot of things on our eating plans cause it suits the majority all the time and we all know this, but if that one person is outside the box, then they have to find their individual way that is best…we know some don’t ever fit all blanket statements :slight_smile:

(Michael) #19

CGM experiment part 2

  1. Ate base meal of 7 bacon and 2 eggs twice a day for 7 days straight. This is about 86.5 g of protein from these two meals daily.
  2. Ate first meal in the afternoon (starting between 1 and 3 pm). Ate last meal around 9 pm each night, for the whole week regardless of food intake. Later in the week I added a third meal at around 5 pm.
  3. After 2 days when average glucose fell to 5.2 mmol/L (around what I consider my normal baseline), I then ate the base meal for 3 more days, but I added in 50g of protein through protein powder (mixed with water shakes) with both meals. (100 g or protein extra per day) for two of the next three days. Due to thanksgiving dinner on the Sunday, I ate an approximate amount of equivalent meat (also why the protein was slightly lower), so larger error bars on the protein intake this day).
  4. After 3 days of this data, I then had the base meal with two 50g protein shakes and the exact same size additions of one beef ribeye steak (330 g each day – 82.5 g of protein) with measured out 100g of liver (26g of protein) for an additional total of 108.5 g of protein eaten at around 5 pm.
  5. used two different glucose meters and took the average of those two readings to calibrate my CGM at least twice a day (did 3 times a couple times)

Results of second CGM data (only summary since the details seem to confuse people) from Friday Oct 8th until until Thursday Oct 14th. Started at 5.6 mmol/L the previous Thursday Oct 7 when also started eating the bacon and eggs meals twice a day.

note I did stop the CGM data on Thursday Oct 14 at 9 pm since I did not eat any food on Thursday at all, and normally I would have finished eating by 9 pm. Clearly using the previous days food intake is more appropriate since I would never expect an average glucose of 6 on a non eating day otherwise. This is also true due to my normal eating pattern of eating in the latter half of a day. Had I eaten mostly before noon, perhaps it would have been better to use same day data.

I also went back and did a rough calculation of protein intake for the first CGM experiment (which due to fasting data only had a few days of data to use for the protein experiment. Here is the data for both experiments, and then their combined data graph. It is interesting (although probably not related) that the QUICKI calculation is approximately the slope of the graph. The data from the second version of the experiment is probably more precise (averaged glucose over two meters as calibration) and was certainly measured and calculated more precisely in terms of both protein intake, consistency between days/meals and time of eating was also kept consistent to stop TRE effects on the data during the second run.

However, regardless, the data from both runs is entirely consistent and demonstrates a clear affect.
Here is the combined graph

P.S. Since he is probably interested in this information, I will be sending this to Ben Bikman. Perhaps he will officially run some experiments to confirm my data and publish. Clearly this is not confirmed data results in a clinical setting.

GNG, protein and blood glucose re-visited