A Very Different Path
If you haven’t already, be sure to first read Part 1 of this experiment by Siobhan Huggins.
Originally, Siobhan and I were going to simply do the same protocol at the same time, perhaps something similar to the one I did with my sister two years ago. But it occurred to me that it might be much more interesting if we tested both directions in lowering LDL that would lend further evidence to the Lipid Energy Model.
I began exploring the idea of going carb-centric in order to showcase this shift in energy metabolism having a likewise effect on my cholesterol. But while I’m at it, why not go big? Why not choose foods that would achieve the task yet from a category no one would associate with “healthy” or “whole food”? I settled on white bread and processed meat.
And heck, while I’m at it, I should exceed my prior levels of carb intake to drop below my previous fasting record LDL of 103 mg/dL, courtesy of the Capstone Experiment. In fact, I predicted this outcome and shot a video in advance of the experiment.
Whereas Siobhan was invoking the Inversion Pattern on her end of the experiment, I was straight up changing my metabolic pathway. As mentioned in the video above, I was moving from a fat-based metabolism to a glucose-based metabolism.
I had originally wanted to hold out on the “Peak Levels” until I saw that plateau I mention in the video above. But honestly, the massive levels of glucose throughout this experiment being indicated by my Continuous Glucose Monitor (CGM) was making me very uncomfortable. Thus, once below 100, I found myself ramping down soon (starting at 9/25).
For some perspective, here’s my CGM before this experiment (while keto):
You see that slight rise about 2/3rds of the way in? That’s when my glucose was just 105 mg/dL. Yes, that’s how steady this graph is when I’m fully keto — you can’t even tell when I’ve eaten.
You see that vertical line on the left side? That’s marking that peak at 183 mg/dL (shown in upper left). In other words, I’m clearly riding massive waves of high glucose throughout this experiment. While I expected this going in, it sure doesn’t help the ol’ psyche to see it in real time… although it does make for some good data.
The CardioChek Data
Let’s combine all three cholesterol markers into a single graph:
During the Ramp Up period where my carbs were slowly being increased, we don’t see a big drop in LDL, nor an increase in triglycerides (TG). Yet once entering into the Peak Levels portion, this changes and we observe LDL dropping dramatically while TG rises in kind.
The Lab Data
I took a total of three lab draws. One on the 17th right before the ramp up, one on the 24th which was midway in, and one on the 28th at the very end of the experiment. Ideally, I’d have had even one more at the end of the Peak Levels period, but I just didn’t know for certain when that would be and if I could get into the lab in time.
I actually got quite a bit of data, but I’m going to highlight the most notable ones here:
Certainly the two that most stood out to me were the Ferritin and Glucagon.
The Ferritin dropping to 11 seems like a legit lab error. For context, note that I’m quite experienced with Ferritin as it is the one marker I’ve most worried about. Before Keto it was in the 500s and remained as such into the diet. However, in recent times it was in the 200s (as shown above). I’ve never seen it below 200, much less lower than 20. Thus, lab error is my my top suspect. (But hey, if it wasn’t, that would be quite a find!)
I blame Ben Bikman for the money I’ve been putting out to test Glucagon <shaking fist>. That said, I think I may be getting the value of Glucagon a lot more this time around. Rather than share my own thoughts on this (for now), I’ll let commenters pontificate down below as to why my fasting Glucagon did go down correspondingly with the rise of fasting insulin and why that last score of 80 pg/mL was actually the highest number I’ve gotten to date.
Of all the experiments I’ve done to date, this one was the most mood-altering, hands down. While I’d like to think I’m usually optimistic, fun-loving, and easy-going in most situations, instead I was much more irritable, temperamental, and chronically anxious. I just flat out complained several fold more, even though a lot of it was repetitive. Even more frustrating was that I was self aware of it but couldn’t seem to help myself.
I actually have several anecdotes from that period that are a bit embarrassing to reflect on. But surely the worst was my wanting to post a complaint I had with a company on Twitter. I didn’t ultimately do this, but it’s quite weird that I ever even considered it in the first place. It’s so not me.
I should emphasize that I don’t think this was strictly a carb thing by any means. I’ve done other carb-based experiments that brought me some issues but didn’t always alter my mood appreciably. I don’t think even the most avid high carb low fat advocate would push for an all white bread and lean meat diet.
Yes, while I mainly focus on cholesterol and lipids, I’ve become very aware that certain combinations of diet in these experiments will have profound impacts on my mood in many ways: temperament, irritability, contentment, and even the axis of outlook on life.
(Not So) Final Thoughts
There’s actually a lot more to unpack with this experiment than I’ve laid out here. The key goal was achieved in making for a record change in my LDL cholesterol over seven days. It came at a cost, of course, and I’d never settle for the resulting HDL and triglyceride levels I was observing during the experiment.
Thanks for sharing your excellent work with us.
Happy to! Hope it helps further the discovery.
Thanks for your amazing research!!
“… but didn’t always alter my moon …”
I hate having my moon altered…
It sure can be a problem!
Thanks for sharing your experimental data, Dave. Really interesting. I, doing a 3-month experiment at present (3months3ways – a month on different diets, monitoring blood glucose 24/7, blood and lipid panels, ‘washout of 24 hr fast in between each month), and I’ve found my blood glucose monitor readings identical to yours for Keto. The only thing that creates a ‘blip’ during my Keto Diet month that is still well below the upper end of the reference range is doing infrared sauna. Of note on the carb and mental health response, I have founds similar thing – my mental healt has been more ‘stable’ and less irritated when in Keto than on a diet with more carbs, especially more refined carbohydrates. I’ve been tracking it using an app that asks me to subjectively select an emoticon that reflects how I feel at that moment, and taking that measurement first thing so I’m not too awake to ‘game it’. Keep up the awesome work!
Damn — I wish I had thought to track my mood via the app as you do above. I likely will again if I ever do this kind of experiment. Knowing me, I probably will.
1. When you give lab results, I assume they’re fasting. But it would be nice to specify that.
2. When you give BG printouts, it would be nice to have the Y axis labeled. You indicated value of one point, but one would like to know the range on the other graph.
3. You mention “moon” when you obviously mean “mood.”
4. These are all picky comments. A lot of work went into this, and I salute you for that.
1. Yes, always fasting around 12-14 hours unless explicitly stated otherwise.
2. That graph was a screen shot for the CGM. I have no control over the axis (I really, really wish I did). So I can only move that crosshair to one spot at a time.
3. Corrected — thanks!
Wow! Are the the levels of HDL, Triglycerides, LDL and glucose within “normal” ranges even as they elevate?
At the peak, they were not. Fasting glucose of 112, HDL 38, and TG of 221. But most of the time I was within range or just outside of it.
Why would you check the A1c 3 times in 2 weeks? Isn’t the erythrocyte life span 3 months? Also could you “bottom line” the message for me? Is the message “high carb means low LDL cholesterol”?
I check the A1C that frequently for that very reason. I’m interested in its variability and how much it can be influenced in the short term.
Sure, that bottom line sentence can be accurate (although oversimplified). Conversely, more fat means (potentially) more LDL cholesterol — all else being equal.
Great work, Dave!
So why is there so much TG? If you changed to carb-burner during the experiment, why TG went up? Ok, I get that you use less fat as energy, but where does it come from in the first place? The processed meat you ate? Your adipose tissues through lipolysis? Thanks!
My guess would be the major pathway for increased serum TG is from de novo lipid synthesis in the liver. All those excess carbons in glucose have to be used for something.
Short answer: underutilized VLDLs. You don’t need to have higher rates of VLDL secretion. You just need lower turnover of the existing ones.
Interesting experiment Dave, amazing to see how dynamic LDL-C and TC can be in a matter of days. I think it is interesting question whether the drop in LDL-C you observe is unique to individuals that have low adiposity and are metabolically healthy. In other words, perhaps the drop in LDL-C is transient. My prediction would be that a persistent carb-centric diet would cause LDL-C to eventually increase as you increased adiposity and became less metabolically flexible.
Did you measure LDL-P pre- or post-experiment?
So given we can all agree this carb-centric diet is unhealthy for many reasons, it does get back to the million dollar question you have been tackling, related to the risk at the other end of the spectrum. Does being protected from metabolic syndrome, but with high LDL-C and LDL-P, have an acceptable (low) risk of CVD? Since this is still open for debate, in so far as no large RCTs have been performed, let’s assume that a hyper-responder with high LDL-C is not comfortable with this unknown risk, but yet wants to benefit from the keto lifestyle. Have you considered or tested whether fat sources have differential effects on your LDL-C (or -P) levels? For example, would a keto diet of foods that have higher proportion of MUFA versus SFA induce lower LDL-C and LDL-P? Given what we know about the general population, my guess would be yes. If so, this would presumably provide some insurance against the unknown risk?
Todd, in this study they brought the LDL down with safflower oil (mostly Poly), https://www.fasebj.org/doi/abs/10.1096/fasebj.27.1_supplement.127.4 but did not see the anticipated ‘All Cause Mortality’ benefits, and they did not have the decency to publish the results. I will not loose sleep over ‘unknown risk’ popularised by industry when there are known benefits in (Lo Trig + Hi HDL) whatever the LDL.
Thanks Solomon for this paper, interesting. I would agree, the claimed benefits of the LDL-C lowering effects of PUFA from processed vegetable oils are not convincing. I would definitely choose a low risk of metabolic syndrome + low inflammation over vegetable oil-dependent lowering of LDL-C. For now, perhaps a good compromise for certain individuals would be to enrich their fat macro with MUFA from avocado and olive, or even perhaps algae oil.
>>> My prediction would be that a persistent carb-centric diet would cause LDL-C to eventually increase as you increased adiposity and became less metabolically flexible.
As is my prediction as well. Although it would take a while to unpack that, I suspect our positions would be quite similar with regard to downstream LDL-C resulting.
>>> Did you measure LDL-P pre- or post-experiment?
Yes. LDL-P was 3235 at the start of the experiment (9/17) Bear in mind I didn’t get LDL-P at the peak day drop day on 9/26 when my LDL-C was 83. But I got it on the 28th when my LDL-C was 126. My LDL-P on that day was 1290.
>>> So given we can all agree this carb-centric diet is unhealthy for many reasons,…
Actually, it may surprise you to know I’m agnostic in this area. I think it could be the case one could have a high carb, low fat diet — so long as their metabolism remains flexible and they can make all the endogenous sterols needed (if not, they might need more fat and/or cholesterol than currently ingested).
>>> Have you considered or tested whether fat sources have differential effects on your LDL-C (or -P) levels? For example, would a keto diet of foods that have higher proportion of MUFA versus SFA induce lower LDL-C and LDL-P? Given what we know about the general population, my guess would be yes. If so, this would presumably provide some insurance against the unknown risk?
I think for most, MUFAs will have some level of lower LDL-C and LDL-P, with PUFAs having an even more pronounced effect. This degree of difference seems very individualized.
The protective-ness of it is an entirely different story. Implicit in this assumption is that high LDL-P is a *net* increase in overall mortality. That even if we assume the lipid hypothesis is correct, that immunological, reparative, and other potentially beneficial roles of LDL-P are still at a loss when considered in balance. I don’t know that I think this question has been answered.
Great work Dr. Feldman. I have been studying a lot of your work since Low Carb Houston, this has opened up a whole new can of worms! Love the paradigm shifting work.
Thx! Although I’m not a doctor — I’m an engineer. (Some would say, a “Lipid Engineer”)
All your discoveries give me hope that I am not doomed! On that note, if you were a 55 yo post menopausal (on BHRT, Thyroid replacement) active and lean female who feels best eating mostly carnivore with Total chol Leslyn Keith, HDL Dan Glover, LDL Roger Johnson, Trigs Dave Lindberg, Homocysteine @12.5 (have started supplemented with methylated Bs) HS CRP @0.2 and is APOE 4/4. What would be your focus for optimizing health/longevity?
Without more information, I’d likely be doing the same as you (supplementing methylated Bs, if there are signs of deficiency/poor processing), and trying to find a way of eating that makes me feel my best and shows good results on my labs, exactly as you’re doing 🙂
Because of the apoE4 I’d be a bit more concerned about insulin resistance, and sources of oxidative damage, like from dietary sources, smoking, etc, so I’d likely try to find a diet that could minimize this (probably, at least, whole foods based at minimum, just personally).
One thing I do find interesting is your LDL. It’s a bit lower than I’d expect for a lean carnivore. Do you do a lot of exercise/resistance training, or anything like that?
I so appreciate your reply! Despite being LCHF/Keto for decades, going mostly carnivore has enabled to eat much less frequently. Although LCHF/Keto made a difference in my health, I would still be ‘chomping at the bit’ to eat every 2-3 hours. Now, I easily get by on 3 meals a day with an overnight fast of 13 hours or more. In regards to exercise, my days of long duration workouts are behind me–left me crushed hormonally. Instead I do ‘mini-exertions’, a 3-10 minute bout engaging my large muscle groups, I walk daily, and I avoid sitting during the day. What are your thoughts on an HDL of 125? Because of my high Alz risk and due to the impact on neuroinflammation, know that I partake of ‘Green Medicine’ (small keto edible with supper) and a keto cocktail most evenings.
Know that you and Dave are appreciated!
Thank you for appreciating us! I’m so glad we could be of some help 🙂
So it sounds as though you’re doing short sessions of HIIT? That would be likely to drive LDL down a little bit. I’ve also seen a study suggesting that alcohol can raise HDL a little bit while lowering LDL, so perhaps that’s contributing as well. If you ever forego the cocktails for a week or so I’d be curious to see if your labwork would change. That’s just curiosity though 🙂
I’m not sure what edibles do to lipid profiles, so I can’t comment there!
Over all it seems like a profile I wouldn’t mind having, but I’d be sure to not rely on lipid profile alone. You’ve already gotten hs-CRP which is one I always try to get, but fasting insulin can provide a lot of useful information as well, and can help flesh out the context of your results.
That all makes sense, thanks. Never had a fasting insulin done. During an internship at the Cooper Clinic (I was in my 20s), I let a fellow intern practice blood draws on me over a series of days (I had a crush on him). Everytime, my glucose level was flagged as out of range to the point of being too low for me to still be standing. I was told I was hyperinsulinemic. That was a good 3 decades ago, I imagine being LCHF/carnivore now, that it would be a given that my FI would be quite low. Is that generally the case and if so, what else would it ‘flesh out’?
Ah yes, you’re right, that going off of your other bloodwork your insulin is probably low. I just meant it’s a marker that I like to get every time because it’s such a good way to keep your finger on the pulse of how things are doing.
Insulin is something that can be high for a multitude of reasons (including infection, smoking, poor sleep, stress, poor diet, etc), and thus it’s a good “flag” that can catch problems before you’re even aware they’re happening (because it’s tied to things like cancer, and diabetes, and a multitude of other issues). Essentially it’s just very valuable information to have. I really wish it were included in annual bloodwork, and I myself get it on every blood test as a bit of a check engine light situation (if it’s low, I’m probably doing okay, if not I might have something I need to look into). I hope that makes sense!
It does! Love it–thanks
Try doing it again with carbs other than wheat.
NO don’t actually do this! But it might be intriguing to see how much of your mood problem relates to the glucose per se and how much to its form, white bread must be just about the most toxic way to eat wheat.
I discovered that my mood swings, along with drops in physical energy, were pretty closely related to my blood glucose which spikes badly about an hour after eating – lack of Phase 1 insulin – then drops precipitously after 2 – 3 hours as the Phase 2 insulin fails to shut down once it’s started (oh for a pocket insulin meter!) which causes a dump of cortisol and the other counterregulatory hormones – think hangry with a turbocharger (reactive hypoglycemia).
Most of my physical and mental symptoms improved or resolved on low carb but completely cutting out wheat (and omega 6 seed oils) was the icing on the cake. Far from uncommon.
Your work is fascinating. I’m not a hyperresponder, just a responder, I think if I’d done this my trigs would have spiked a lot more and my HDL would have tanked. According to my lipid panels when I used to bother with them, my LDL seems to be fairly constantly high but (I hope) not dangerously so with very low trigs and high HDL.
Sorry I haven’t (yet) managed to psych myself up to watch the Attia podcast, life’s too short! Some things I’d like the LDL as the One True Cause guys to explain – active weight loss often increases LDL, so should we not lose weight then? Hyperthyroid can drop LDL through the floor so should we all be taking thyroxine with our statins? I think not. And I’d love to know how LDL is responsible for these results correlating HbA1c with cardiovascular and other risk (you probably know these studies but for anyone who hasn’t seen them)
and a recent one
I definitely think I noticed a difference in the white vs wheat. How much of that was happenstance, I can’t be sure. But note I *did* do a wheat bread experiment for the first portion of my Capstone experiment last December. https://cholesterolcode.com/the-capstone-experiment-its-time-to-rethinkldl/
I don’t think the Attia podcast is required listening. It is, after all, 3.5 hours. But it was nice to get a solid talk in with Peter.
Is there an experiment just on High Carbs, Low Protein and Low Fat?
Example: Just Sweet Potatoes for a month?
Is there a video presentation of your tandem drop experiment with the turkey and the the wonder bread?
I believe it was discussed at Low Carb Houston, and I believe also mentioned in passing here.
were your total calories based on your BMR requirement calculation?
Dave have you ever done a low carb medium protein and a medium fat diet. I wonder what the blood work would be on that.
What is meant by medium fat and medium protein? This could be interpreted differently, percentage-wise, depending on the individual!