This is a great writeup. Only thing I'd add is that the CGM lets you cycle through a lot more of your n=1 experiments, assuming you buy the premise that glucose spikes matter. You should be able to preregister a bunch of hypotheses and then test which foods you can add to ex150 that (idiosyncratically for you) are a free lunch. My own findings are here: https://braff.co/advice/f/i-lost-4-pounds-in-4-weeks-using-data-and-analytics-ama
I just read your article. Largely concur with your assessment of the CGM. Mine's currently +35mg/dL off because I'm taking a huge dose of Vitamin C supplement (compared to finger prick blood test). The spikes are still exactly the same, it's just everything is moved up by 35 points.
After your experiment, would you say that the hypothesis ("small blood glucose swings => fat loss") is accurate? I kind of don't think so, but I will say that I do have pretty stable blood glucose on ex150 most of the day.
But e.g. I recently had 4 weeks where the spikes were almost exactly the same yet I didn't lose any weight.
I think the CGM can be helpful with weight loss if you're one of those people who accidentally eats "surprise" carbs, like your example with the brown rice. If you've never done low-carb, you go by the "healthy food pyramid," and so on, you might learn that the "organic honey/agave syrup" in your Whole Foods health food is just sugar, or that brown rice converts to sugar just the same as white rice.
For anyone who's already carb-aware, it'll just confirm your intuition, as you mention in your post.
Yeah, I agree with you and Justin that the whole enterprise of "CGM for fat loss" is precariously balanced on top of the premises I mentioned up top, especially the one about treating spike height (and area under the spiky curve) as the do-or-die metrics to manage down.
For me, there are still plenty of corner cases (foods that idiosyncratically spike me, or not) that surprise me, and I really benefit from the discipline of having an objective device pointing out when I stray from the low-carb path. But the fact that you haven't observed much of a connection between the spike metrics and weight loss is concerning, since you seem like the kind of person who would notice a real connection.
Yea, on ex150 any spikes are really obvious because with 88% calories from fat, the default is no glucose from dietary intake. And the little sugar I do consume is in relatively liquid form w/o much fiber, so it spikes and comes down super rapidly.
I did notice when I was off-diet and just testing random stuff with the CGM how the signal would get super muddled. If you eat a normal-fiber diet with 3 meals + snacks, your glucose never resets except overnight. So you can't tell what one food did from the other without waiting at least 4h or so, which for most people is already a dietary intervention close to intermittent fasting.
For what it's worth, I've been back on the CGM for 3 months now, on the sort of low-carb diet (with no special fasting rules and I guess a modest amount of fiber? I don't track that) implied by my blog post. I can see pretty clearly when the addition of a bad or even a marginally bad food causes a spike. You don't need your glucose to reset all the way back to the morning baseline to see a spike relative to "average glucose level in the 15 minutes before you log the meal," which is the baseline that Levels seems to use when measuring spikes. So I don't find the signal to be super muddled, just moderately muddled.
I've found that when I eat a lot of mixed foods including a bunch of fiber, my glucose goes up (more moderately) and stays up for hours at a time. So it can be very hard to tell if this small change is caused by the piece of dark chocolate I just ate, or is just the remainder of the big lunch 3 hours ago.
Is there any good evidence that glucose spikes matter? By this, I mean the height of the spikes doesn't seem to matter, at least not directly.
From what I've read, the important signal in glucose readings seems to be the time to peak glucose and the time it takes to go back to baseline glucose. These are thought to be related to how well the pancreas and the signalling hormones are all working to keep your glucose in homeostasis. Going back to baseline quickly is also related to sensitivity (which from the evidence we have from bariatric surgery and incretin hormones/semaglutide-like drugs seems to not be caused by too large or too many insulin spikes). It seems unfounded to be worried about the height of the spikes (or even the area under the curve) so long as the body is bringing them back down quickly.
It has been a few years since I've really looked into this though, so maybe new research has come out?
In terms of weight loss, I don't think it matters all that much. At least the spikes don't. What I did observe that when I ate ad-libitum of my regular (very meat-heavy but also otherwise not exactly clean) diet, my AVERAGE blood glucose would be way elevated. I'd basically baseline at 90 or so vs. 75-85 (on ex150)
So I'm thinking maybe constantly elevated glucose can be a sign something is off in terms of fat loss, but if you're below that, the spikes don't seem to matter much.
Two weeks of chocolate is nothing to worry about. Unless you have it atop spinach and almonds... Oxalates danger awareness is low even in keto community. Sally K. Norton is the #1 expert on oxalates. https://sallyknorton.com/oxalate-science/
This is a great writeup. Only thing I'd add is that the CGM lets you cycle through a lot more of your n=1 experiments, assuming you buy the premise that glucose spikes matter. You should be able to preregister a bunch of hypotheses and then test which foods you can add to ex150 that (idiosyncratically for you) are a free lunch. My own findings are here: https://braff.co/advice/f/i-lost-4-pounds-in-4-weeks-using-data-and-analytics-ama
I just read your article. Largely concur with your assessment of the CGM. Mine's currently +35mg/dL off because I'm taking a huge dose of Vitamin C supplement (compared to finger prick blood test). The spikes are still exactly the same, it's just everything is moved up by 35 points.
After your experiment, would you say that the hypothesis ("small blood glucose swings => fat loss") is accurate? I kind of don't think so, but I will say that I do have pretty stable blood glucose on ex150 most of the day.
But e.g. I recently had 4 weeks where the spikes were almost exactly the same yet I didn't lose any weight.
I think the CGM can be helpful with weight loss if you're one of those people who accidentally eats "surprise" carbs, like your example with the brown rice. If you've never done low-carb, you go by the "healthy food pyramid," and so on, you might learn that the "organic honey/agave syrup" in your Whole Foods health food is just sugar, or that brown rice converts to sugar just the same as white rice.
For anyone who's already carb-aware, it'll just confirm your intuition, as you mention in your post.
Yeah, I agree with you and Justin that the whole enterprise of "CGM for fat loss" is precariously balanced on top of the premises I mentioned up top, especially the one about treating spike height (and area under the spiky curve) as the do-or-die metrics to manage down.
I believe the reason why people subscribe to apps like Levels, as opposed to just buying their devices straight from Abbott/Dexcom, is that you get access to all their writeups on these topics, e.g., https://www.levelshealth.com/blog/8-reasons-glucose-is-worth-measuring , which links to https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3884108/ .
For me, there are still plenty of corner cases (foods that idiosyncratically spike me, or not) that surprise me, and I really benefit from the discipline of having an objective device pointing out when I stray from the low-carb path. But the fact that you haven't observed much of a connection between the spike metrics and weight loss is concerning, since you seem like the kind of person who would notice a real connection.
Yea, on ex150 any spikes are really obvious because with 88% calories from fat, the default is no glucose from dietary intake. And the little sugar I do consume is in relatively liquid form w/o much fiber, so it spikes and comes down super rapidly.
I did notice when I was off-diet and just testing random stuff with the CGM how the signal would get super muddled. If you eat a normal-fiber diet with 3 meals + snacks, your glucose never resets except overnight. So you can't tell what one food did from the other without waiting at least 4h or so, which for most people is already a dietary intervention close to intermittent fasting.
For what it's worth, I've been back on the CGM for 3 months now, on the sort of low-carb diet (with no special fasting rules and I guess a modest amount of fiber? I don't track that) implied by my blog post. I can see pretty clearly when the addition of a bad or even a marginally bad food causes a spike. You don't need your glucose to reset all the way back to the morning baseline to see a spike relative to "average glucose level in the 15 minutes before you log the meal," which is the baseline that Levels seems to use when measuring spikes. So I don't find the signal to be super muddled, just moderately muddled.
I've found that when I eat a lot of mixed foods including a bunch of fiber, my glucose goes up (more moderately) and stays up for hours at a time. So it can be very hard to tell if this small change is caused by the piece of dark chocolate I just ate, or is just the remainder of the big lunch 3 hours ago.
Super interesting. YMMV (your muddling may vary).
Is there any good evidence that glucose spikes matter? By this, I mean the height of the spikes doesn't seem to matter, at least not directly.
From what I've read, the important signal in glucose readings seems to be the time to peak glucose and the time it takes to go back to baseline glucose. These are thought to be related to how well the pancreas and the signalling hormones are all working to keep your glucose in homeostasis. Going back to baseline quickly is also related to sensitivity (which from the evidence we have from bariatric surgery and incretin hormones/semaglutide-like drugs seems to not be caused by too large or too many insulin spikes). It seems unfounded to be worried about the height of the spikes (or even the area under the curve) so long as the body is bringing them back down quickly.
It has been a few years since I've really looked into this though, so maybe new research has come out?
In terms of weight loss, I don't think it matters all that much. At least the spikes don't. What I did observe that when I ate ad-libitum of my regular (very meat-heavy but also otherwise not exactly clean) diet, my AVERAGE blood glucose would be way elevated. I'd basically baseline at 90 or so vs. 75-85 (on ex150)
So I'm thinking maybe constantly elevated glucose can be a sign something is off in terms of fat loss, but if you're below that, the spikes don't seem to matter much.
Spinach and turnip greens are two different "beasts" when it comes to oxalate content. Spinach is 20 times worse. Here's a handy reference I just found, from a rather unexpected source https://lafeber.com/vet/oxalic-acid-content-of-selected-foods/
Oh I assumed they were similar due to similar appearance. Thanks! I'll stick with turnip greens then.
100g dark chocolate bar every day? Watch out for oxalates. https://www.sciencedirect.com/science/article/abs/pii/S0889157511000822
Interesting. I only did it for 2 weeks, so not super worried, but maybe something to watch out for if you did it longer.
Two weeks of chocolate is nothing to worry about. Unless you have it atop spinach and almonds... Oxalates danger awareness is low even in keto community. Sally K. Norton is the #1 expert on oxalates. https://sallyknorton.com/oxalate-science/
I do eat greens like spinach, but, on e150, only about 60g per day. Today I had turnip greens.