Mar 14

Thoughts on Shawn Baker’s Labs

This week, Robb Wolf had Dr. Shawn Baker on his podcast, The Paleo Solution.

Baker has been a controversial figure for his embracing a meat-only diet (no plants). Naturally, this has made him a favorite target of vegans, particularly youtubers. What makes him especially interesting is that he’s not just an athlete, but he’s actually setting a number of world records for his age class (over 50!) in athletic benchmarks.

Full disclosure: I now know a number of “carnivore” dieters like Amber O’Hearn who I consider friends. But I likewise have a number of vegan friends as well, many of whom I have worked with to improve their lipid numbers.

Throughout his time on this diet, Baker didn’t take any blood tests until just a couple months ago. This first round of tests were then revealed on the podcast and proved to be quite interesting. While there are a number of them that would be worth commenting on, those I want to focus on were about the energy metabolism (no surprise).

The Glucose and Insulin Labs

  • HbA1c: 6.3
  • Fasting Glucose: 126 mg/dL
  • Fasting Insulin: 2.6 mIU/L

The A1c and fasting glucose are certainly consistent with someone who is prediabetic and likely on their way to full-blown Type 2. Well — that is — until you look at the fasting insulin. In what circumstance would you expect a high fasting glucose with a low fasting insulin?

I see them all the time: low carb athletes.

Ironically, I just got done talking about this on Ketoconnect’s podcast (airing in a couple months). In fact, Lean Mass Hyper-responders (LMHR) are actually the most likely to have a fasting glucose in the 90s or even over 100s while also sporting a fasting insulin below 3. Moreover, this profile typically has lower blood ketones (BHB) when testing compared to their more sedentary cohorts. (I’ll have a blog post on this soon)

I myself tend to be on the borderline of a LMHR or just past the line when fully keto. In that context, I often have fasting glucose in the 90s or lower 100s and an A1c of between 5.5 and 5.7. And that’s bad, right? Creeping back toward risk of Type 2 diabetes? I certainly don’t think so given my average fasting insulin of 3 or less.

But wait — Baker is much, much higher, right? Almost 20% higher in both fasting glucose and A1c. Surely he’ll be piling up the insulin resistance as we don’t see these numbers in any other low carbers, yes?

Energy Demand Leading to Higher Glucose Sparing?

This is where I part ways with so many people in and out of LCHF. What got me into cholesterol in the first place was seeing how it was really just “ridesharing” in a larger energy metabolism. So energy delivery and homeostasis on a greater network-level scale is what I continue to find endlessly fascinating. I’ll let the Benjamin Bikmans and Michael Eades of this world tackle the finer details of the mitochondria, I want to know how the Human OS manages to traffic that energy to our cells so effectively in the first place.

Which brings us back to Baker. This isn’t any ordinary guy. He isn’t just working out for recreation, he’s an athlete’s athlete. He’s training to break these various world records in addition to a variety of other workout regimes. So if you observe (as I do) a degree of adaptive glucose sparing as being more common with low carb athletes, what do you suppose it would be like for this human cyclone?

It’s not enough to just think of where his cells are getting their energy, you have to think of how timely it needs to be at a systemic level. Is Baker more of a slow jog in the morning kinda guy? No. Is he more of an exploding HIIT adventurer that is probably keeping high muscle confusion? Yes. The latter suggests more need for glycogen stores in the muscle with a strong rotation of glucose via the liver.

This is why I asked Baker and many other low carb athletes to please do more testing for us if they (heaven forbid) get injured or for some reason can’t exercise intensively for a period of time. I suspect in such cases if all other things were equal, their fasting glucose would drop. This is what happened to me when comparing my morning glucose in and out of marathon season.

The Cholesterol Labs

Baker had the following lipids:

Total Cholesterol: 192

LDL-C: 141

HDL-C: 40

Triglycerides: 54

For these I had the following notes:

  • LDL-C and LDL-P are both on the lower end for a low carber who is as lean and fit as Dr Baker is. I suspect this is due to his emphasis on resistance training, which can reduce LDL scores due to a higher rate of use for muscle repair.
  • Small LDL-P is low at 283, and clearly very Pattern A.
  • HDL-C — 40 mg/dL is low for a zero carber. However, as he discussed with me, he apparently has low HDL running in the family as well as a prior history of it running closer to 30. At some point, he may what to check into his SNPs and see if there’s some pushdown there.
  • HDL-P was highlighted as being low (out of range) at 28. However, I see this frequently with low carbers across the board.
  • Triglycerides — 54 is certainly very correlative with a very athletic, insulin sensitive metabolism.
  • Lipoprotein (a) — 2 nmol/L is one of the lowest scores I’ve ever seen (maybe THE lowest).
  • The score I most care about is Remnant Cholesterol which is calculated by subtracting both HDL-C and LDL-C from Total Cholesterol. His score of just 11 mg/dL is extremely low risk and suggests he has a very efficient fat metabolism. (You can use our reporting tool here if you’d like to calculate your own)

Again, I’m not surprised by the lower LDL-C and LDL-P as I’ve covered this before with my own data. As the theory goes, on one end he’s being powered by fat, necessitating more VLDL secretion. On the other, he’s engulfing LDL-P at a rapid rate for tissue repair, particularly for all that resistance training.

Final Thoughts

Baker, O’Hearn, and the many other “carnivores” are certainly conducting a very interesting, real-time experiment. I’m excited to see how this new data will manifest and what positives and/or negatives will be revealed.

At some point, I plan to do an experiment of my own going meat-only for a month, just as I plan to do likewise with a plant-based diet. I’ll be very curious as to what happens to my blood markers when comparing to my many, many labs before it.

Mar 06

#LowCarbBreck ’18

Credit: Ted Eytan, MD @tedeytan

This last weekend I got to present at easily one of my favorite conferences to date, Low Carb Breckenridge 2018. Moreover, I had been working extra hard on this particular deck as it was meant to illustrate the energy model of lipids as I see them for a layperson audience.

It was also easily the busiest conference for me as there were quite a lot of contacts and connections with many key people that have led to many new opportunities moving forward. But while it’s fun to connect with A-listers, I really got to hear some remarkable personal stories from many of the attendees and their own journeys on this road.

It’s hard to emphasize enough just how happy I feel in helping others to understand this challenging subject. Cholesterol is so difficult to work through given the many dense texts and literature that surround it. Which is why my favorite compliments aren’t from doctors with 14 letters after their name, it’s from people like Stacy from Ohio. (Hi Stacy!) As these statements were made to me in private, I won’t repeat them here. But I’ll just say they were all very powerful and moving to me. Thank you!


Feb 19

Sugar and Cholesterol Experiment – Findings

Okay, before we get started, you should know I’m not going to have enough time to cover everything that came out of this experiment. Needless to say, there’s lots and lots of interesting stuff in the numbers. But the good news is that I’m making the raw data available to everyone to find these things out for themselves. (See downloadable spreadsheet below)

First, let’s the stages and dates as I’ll be referencing them quite a bit:

Baseline Food

Baseline food eaten every single day from January 30th to February 14th are the following:

  • 10 eggs
  • 8 oz cheddar cheese
  • 60 ml Heavy Whipping Cream
  • 2 sausage patties
  • 43 almonds
  • 15 grams of butter
  • Supplements: Multivitamin, Magnesium Malate, K2 MK-7, D-3, Sports Salts

These were divided into three meals eaten at roughly 10 am, 3 pm, and 8 pm — with one floating “snack” between 3 pm and 8 pm of almonds (usually close to 5:30 pm).

Intervention Food

  • In Stage 3 and 5 I had four packets of Skittles a day. One with each meal and one with the almond snack.
  • In Stage 3 the Skittles were eaten following each meal/snack.
  • In Stage 5 the Skittles were eaten before each meal/snack

The Biggest Revelations

Below are the biggest findings from this experiment that I’ve noticed (so far). There are many more than this which you can discover for yourself in the downloadable Excel file at the bottom of this post.

Does the order in which I eat the Skittles matter for cholesterol? Apparently so!

As with the rest of the experiment, Stages 3 and 5 were the interventions where I ate exactly the same food over the course of the day in three time slots (again, 10 am, 3 pm, and 8 pm) but with the addition of the Skittles. The only real difference between the two stages was the order in which I ate the Skittles.

Here is a comparison of the two stage groupings of 2-3-4 and 4-5-6 so that you can see the interventions side by side. First, the Total Cholesterol:

(Note, all measurements in this second were taken with the PTS Diagnostics CardioChek Plus)

Now, unfortunately, as I had a number or readings where triglycerides (TG) fell below 50 mg/dl (more on that below), I wasn’t able to get a step-by-step accounting for LDL-C. So we’re going to do a comparison of its close cousin, Non-HDL:

All I can say is — wow! These are not inconsequential differences. Same food, same intervention, same times eaten — but that one little change in order had a fairly profound impact.

Naturally, there are already some studies on this very thing and physiologically speaking the order in which the liver will be impacted by the sugar will have measurable impacts on glucose and lipids postprandial. But again, these are effects rippling into the fasting lipid numbers of the next morning by a whole lot.

I actually have quite a bit more to say on this, but that could be a blog post by itself… and I’m a bit pressed for time right now in preparation for Low Carb Breckenridge. So on with the show…

What Blood Markers Changed the Most?

I had a total of four blood draws:

  • 2/2 – Last morning of first Washout Stage
  • 2/3 – Morning after 24 hours into the first Intervention Stage
  • 2/5 – Morning after 72 hours into the first Intervention Stage
  • 2/12 – Morning after 72 hours into the second Intervention Stage

Keep these four fresh in your mind as you look at the charts below…

Note: I had actually meant to get a lab done on 2/9 as well to precede the second Intervention Stage, but there was a problem with the lab order that I won’t get into here. Regardless, we have 2/2 serving as our “clean” washout sample along with a lot of historic lab tests I’ve done overall to compare with.

Glucose and Insulin:


So naturally, we aren’t surprised I’d have higher fasting glucose days after bringing in the sugar. This is especially true if you are fat-adapted, by the way. This is why it is generally advised you leave the diet and reintroduce carbs for a few days before taking an Oral Glucose Tolerance test (OGTT) since you will likely fail it if you don’t.

No, the really big news is how different 2/5 is from 2/12. The first is three days of Skittles after the meal, the second is three days before — and boy oh boy is that fasting insulin remarkably different! In fact, this is the highest reading I’ve ever had by far! (My average insulin is usually under 4 uiu/ml when keto)


I had only had my Cortisol checked in two tests, the before and after mornings of the first three days Skittles. Certainly, this marker isn’t one I normally watch very closely, but then, it’s usually very consistently around 10-13ish. Very first time I saw it drop this low.

WBC (White Blood Count):

Another interesting surprise. While I’ve had my WBC dip below range before, my overall average is 4.2, this is certainly unusual.

I also got NMRs in my blood draws for advanced cholesterol testing:

Okay, this one I’m super excited about. A small group of you uber-geeks are familiar with my theories know why I’m calling out the change in HDL-C and triglycerides. But for fun, I’m just going to pass the ball back to you in the comments below. Why, pray tell, would you think HDL-C would go up and triglycerides would come down in the wake of all that added sugar? (Some of you already know my answer)

Does the order matter in personal experience? Absolutely!

Once I decided I was going to add the extra intervention stage, I figured it would pretty much feel like the first one. Nope! That one tiny difference of order made the entire experience very different.

In the first intervention, I ate the fatty meal, then ate the skittles for around 20-30 minutes following. I didn’t feel especially hungry for the Skittles, but I didn’t mind eating them either. It simply felt like it was a simple dessert.

On the second intervention stage, I tried to time eating the skittles for about the same amount of time (20-30 minutes), but this time before the meal. And each time I just wanted to stop right afterward. I looked at my prepared meal with dread as I had no appetite at all for it.

It’s hard to explain just how radically different both stages were in spite of everything else that was intentionally the same.

Here’s the kicker — I had notably more gut pain in the first stage relative to the second one (again, I have Fructose Malabsorbtion, so this was somewhat anticipated during the experiment). This difference could just be my having acclimated to the diet better by the second sprint, or perhaps the order did really matter in this case. I can’t say for sure.

Lots More!

I’m stopping here and just posting the raw data for everyone. Note I don’t have all the numbers entered yet, but I hope to add these at a later point.

Data Spreadsheet for Added Sugar Experiment

Feb 16

Cheap Blood Labs Coming Soon

In the near future, I hope to have a means of ordering cheap, private blood labs through this website for those who live near a Labcorp. Moreover, I want to incentivize everyone to consider sharing their labs back to our data pool anonymously.

Here’s how it would work

  1. We’d have an Order Labs section on this website where you’d add them much like a shopping cart.
  2. If you further agree, we’d give a special discount for additionally answering a few questions along with submission of your labs to the data pool anonymously. These questions would be similar to “what diet do you follow?” and “how long have you been on (specified diet)?”, etc. This anonymous pool of data would be updated in batches and made available to the community at no cost.

Our incentives

  1. Obviously, my greatest incentive to get involved with this is the potential for extraordinary data growth! This could go a long way to helping us get very strong numbers for analysis in the community.
  2. There will be a margin built into the price, even if very small — and this could be beneficial to doctor and/or myself.
    1. This margin is to be first maintained in case of problems or unexpected costs.
    2. The ordering doctor will be obligated to follow up on unusual labs and thus will be compensated out of this margin for time required.
    3. Any remaining margin will go to offsetting the cost of my own labs.
    4. If this margin grows to the point of actually exceeding the costs of (a), (b), and (c) — I’ll disclose that directly in a future post. (But I actually doubt it given how small the margin is and how large my total lab costs are.)

It’s worth emphasizing that the money collected for these labs goes directly to the ordering doctor, not to me or a company entity I’m in any way connected to. So effectively, I’m just letting this site set up the lab options and offer a discount if one agrees to provide more data, then handing it off to the Doctor’s PayPal to fulfill the order.

(Again, I do not accept any money from any business or affiliate of any kind and will let everyone know if that changes.)

Our DISincentives

Let me state plainly that I only want to do this if it doesn’t become a big bandwidth drain. The added data will be great, but if it ultimately results in a lot of admin and footwork time due to how small our margins are, we may have to make adjustments to accommodate. And if that happens and eventually the pricing looks like it will be close to what is offered online, then I’d probably just pass on it altogether. I only want to do this if (1) our labs are substantially cheaper and (2) we get great community data from this. (If I wanted to start another business primarily for money, I’d just do so with my software devs and make far better returns for time spent.)

When is it coming?

Soon! I can make no commitments of timing. I’ll be doing the beta with my Patreon at first, but only after thorough testing will we open the doors here to ensure we can handle the volume!

Feb 09

Sugar and Cholesterol Experiment – Midmortem


So I started out this experiment with a different design. But along the way, I made a few different adjustments which I’ll list below and why.

  1. Changed baseline food. I wasn’t feeling very well when I started out and this could have been for a number of reasons I detail in the second post of this series. As such, I decided to switch baseline food from the keto shake to a combination of eggs, cheese, heavy whipping cream, and some almonds. I further supplemented as I always do with magnesium, K2, D3, and a multivitamin.
  2. I effectively made this three day period of food reset the new Washout Stage, pushing the Intervention Stage with the Skittles back to the weekend.
  3. Given the new timeline, I decided to leave open the possibility of cutting the Aftermath Stage short if my numbers appeared to “snap” back before Friday. This would leave open the unlikely but possible scenario that I introduce a second intervention while still having this level of dietary control. (More on that in a moment…)

Technically, the experiment completed as of today given I’d speculate my numbers have indeed returned to baseline. Here’s the raw data:

Date 1/28/18 1/29/18 1/30/18 1/31/18 2/1/18 2/2/18 2/3/18 2/4/18 2/5/18 2/6/18 2/7/18 2/8/18
Stage Washout Washout Washout Washout Washout Washout Intervention Intervention Intervention Aftermath Aftermath Aftermath
Shake Shake Shake Baseline Food Baseline Food Baseline Food Baseline Food Baseline Food Baseline Food Baseline Food Baseline Food Baseline Food
 + Skittles  + Skittles  + Skittles
CC TC 263 270 261 246 259 232 247 245 245 245 198 286
CC HDL 70 73 68 64 64 62 63 70 70 73 75 66
CC TG 55 52 62 76 67 70 <50 58 <50 51 51 59
CC Glu 95 97 108 90 102 96 117 116 111 111 98 105
CC LDL 182 186 181 167 182 156 163 163 113 208
CC Non-HDL 193 197 193 182 195 170 184 175 175 173 124 220


Okay, so technically, we can’t judge the three day average of both the Washout and Aftermath stages because TG got <50 for two of the Intervention days. Meh! Thus, we can’t effectively see if I met the original hypothesis that the LDL of the Washout Stage would be higher than that of the Aftermath Stage.

We could compare Non-HDL, and those numbers look pretty interesting. But before going any further, I need to bring something else to your attention.

The Adjustment

There are two days where I didn’t feel I got the right amount of blood onto the CardioChek device strip, Feb 2nd and 7th.

On the 2nd, I was recording from all my devices as normal – Keto Mojo glucose, then ketones, eGlu, then lipid strip. But when releasing all the blood from the capillary tube for the CarioChek, a tiny bit sucked back up instead of emptying completely. I scrambled to squeeze the tube harder to release it, but there was no use and the device already started its analysis. I figured at the time it was probably so small it wouldn’t matter, but couldn’t be sure.

One other problem with the 2nd is that I had just enough stips and tubes to carry me to the next delivery date, so I couldn’t have retested anyway without risking a blind spot in my continuum.

On the 7th, the same thing happened, only this time the amount reabsorbed back into the tube was far more substantial. But this time around I had enough supplies to do a second test. The second test yielded numbers that I felt were more likely.

(Side note: please don’t ask why I don’t just do multiple CardioChek tests every time I check. This test is $11 a pop and I’m already out of pocket for very expensive lab tests ($100-450) throughout this experiment that together runs higher than my mortgage. I’m saying this preemptively because some assume I’m somehow funded by research grant or company. No, I allow only funding from the awesome people who support my Patreon or contribute directly. Thanks!)

So given this, I’m going to provide an adjusted table that uses the second CardioChek test here:

Date 1/28/18 1/29/18 1/30/18 1/31/18 2/1/18 2/2/18 2/3/18 2/4/18 2/5/18 2/6/18 2/7/18 2/8/18
Stage Washout Washout Washout Washout Washout Washout Intervention Intervention Intervention Aftermath Aftermath Aftermath
Shake Shake Shake Baseline Food Baseline Food Baseline Food Baseline Food Baseline Food Baseline Food Baseline Food Baseline Food Baseline Food
 + Skittles  + Skittles  + Skittles
CC TC 263 270 261 246 259 232 247 245 245 245 257 286
CC HDL 70 73 68 64 64 62 63 70 70 73 68 66
CC TG 55 52 62 76 67 70 <50 58 <50 51 52 59
CC Glu 95 97 108 90 102 96 117 116 111 111 95 105
CC LDL 182 186 181 167 182 156 163 163 178 208
CC Non-HDL 193 197 193 182 195 170 184 175 175 173 189 220

So yes, I do think had I gotten the second test on the 2nd, it would have been higher and more in line with the days before it.

And yes, I’m advocating to use the second test on the 7th, even though (ironically) it pushes the trendline away from my original hypothesis. But I’m not doing this to appear right, I’m doing this to get to the truth.

Current Thoughts

Funny as this may sound, this data was exciting precisely because it didn’t change that much! I know you, dear reader, are used to seeing my dramatic shifts in cholesterol with the carb swap experiments of Phase II in my research. But again, those were removing of fat and replacing with carbs. As I said at the beginning of this experiment, this is the first really controlled experiment where I added carbs.

Now that said, I did introduce several new things at once, which is a little uncharacteristic of me.

  1. This was the first time I experimented with high fructose corn syrup. And I had hoped it would be the last. Naturally, many will point out fructose is metabolized differently via the liver of course, and there are many studies stratifying how this plays into lipogenesis, adipose storage, NAFLD, etc. But I won’t get too deep in that rabbit hole here.
  2. Obviously, there are other ingredients in Skittles providing more potential confounding, such as hydrogenated palm kernel oil. (Which I’m sure Tucker will tell me is A-Ok)
  3. And finally, while I didn’t think of it until halfway though my first Intervention day, I was eating the Skittles at the end of the fatty meal, whereas I usually ate my carbs at the beginning in prior experiments.

On this last point, I should emphasize this was entirely by accident. I’m used to thinking of sweets as naturally following a meal — you know, dessert! Once I realized I was doing that, I decided not to change gears and instead keep the pattern the same.

Now, this order change may be indeed relevant to my resulting blood glucose following the meal; this wouldn’t surprise me. But would it actually make a difference to the lipid scores of the following day? That question has been haunting me the last couple days.

Experiment Extention!

Alas, this is where I reveal I’ll be prolonging my misery and doing a second round intervention followed by a second round washout!

Why? Because if there’s even a tiny chance the order in which I ingest the sugar component of the meal could have a longer-term impact in changing lipids for the morning of the next day, we need to know!!!

So without further adu, here is the new, extended schedule:

Thus, this is not a postmortem, it’s a “midmortem”.

I won’t lie — these experiments are very taxing not just for the money spent, but because I likewise eat to the schedule (as shown in the first post) of 10 AM, 3 PM, and 8 PM to keep consistent with all prior experiments and equalize against other possible confounders. So I don’t relish this extra seven days added, especially given right now I’d kill for a fat, juicy steak.

Yet there’s no getting around it, I’m now very uniquely well placed in controlled, personal lipid data now and I might as well take this opportunity to find this out for all of us as well. The things I do for science!

Older posts «

» Newer posts