Cholesterol Code

[Note, much of this data is discussed in two videos from last week. One where I first announced the data, and another where I go over Q&A with Siobhan.]

The last week and a half has brought me some powerful new data and insights on triglycerides.

While most of what I’m about to show you is data I’ve collected directly with myself (AKA, an N=1), I think you’ll agree it is quite compelling. This is certainly one moment where my prior body of experiment data was extremely useful to compare to newly collected evidence as you’ll see below.

A Surprise Morning Spike

As many of you know, I began a “carnivore diet” experiment on December 17th. It was very restrictive in that I’d be eating only beef at first. Ten days later on the 27th I pulled out my CardioChek and did a lipid test right after I awoke. I was stunned to see:

My triglycerides (TG) were exceptionally high. I know this to be unusual because I’ve done many just-woke-up tests from experiments before. In fact, the only time I can recall having TG above 100 after having just awoken, was for three days following the second intervention of the Resistance Training Experiment — even then, that was almost 100 mg/dL lower than this number!

[Side note: If you’re new to this site, the LDL number might have surprised you too, but this was entirely expected. And if you haven’t already, consider reading this before you continue…]

Naturally, I confirmed this number for a second time a few minutes later. Then a few hours passed and I was about to have breakfast with my family. Just before, I took two more tests to confirm what I suspected — the triglycerides will have dropped. And they did… quite a bit, actually.

A longer version of this played out the next day (28th) where I woke up with even higher TG than then the day before. Only this time, I went ahead and took my lipids every hour on the hour until they were to return to below 100 mg/dL. (Shown left)

Triglyceride Carryover Effect is Observed

By this point, I couldn’t help be feel this looked too much like a postprandial (following a meal) sized marker. So I started working my way backwards toward the last time I ate.

What followed were several days where I captured both the morning lipids and those from the night before. Indeed, it appeared I was going to sleep with high TG and thus waking up with it as well.

A Carnivore Connection?

So what’s going on here?

If I’m metabolically healthy, I shouldn’t be waking up with triglycerides this high, right? Again, I wasn’t seeing anything like this before, so why now? Was there something special about the carnivore diet in particular?

Actually, the phenomenon of unexpectedly high triglycerides with some carnivores is not new. The first dramatic case I observed was with Brenda Zorn who collaborates with the 2 Keto Dudes and now works for IDM. For an experiment, she ate steaks for one month and was shocked to see her TG shoot up to 500. Gradually it dropped back down after she left the diet, but it was no less a mystery to us at the time.

Since then, there seemed to be a notably higher proportion of these reports with carnivores relative to low carbers, even if still a small subset. A fun fact to know and tell is that the carnivore diet was originally called the “Zero Carb” diet. And there in might be the biggest clue.

The Magic Twelfth Hour

While I go through more graphs and slides in the accompanying video. I want to put extra focus on one particular scatter plot for this blog post:

This is all 35 tests together on the plot and something immediately jumps out if you look closely.

Let’s divide the graph into quadrants with the vertical line at the 12 hour mark and horizontal at triglycerides …

• In the top left area (1) we see most of the dots on the graph, all above 100 TG (above green line) and taken less than 12 hours from when I last ate (left of the red line).
• Conversely, we see all the plots on the right side of the 12 hour line at below the 100 TG line in bottom right area (4).
• Or to put it another way, wherever a test was taken at more than 12 hours, I always showed a TG of less than 100 mg/dl regardless of where it started at in the morning. There are no plots in the upper right area (2)

As it happens, we here at CC always recommended water-only fasting for at least 12-14 hours before a blood test. But I’ll concede I assumed 12 hours to be conservative. When someone said they had fasted 9 or 10 hours, I assumed it probably didn’t have that much impact. Now I’m starting to rethink that, particularly if this data is further reproduced by others.

Low Insulin Hypothesis

So what if this all comes down to operating at lower insulin levels?

Insulin is important for fat metabolism, although not in the way most people tend to think of it. In fact, one could argue one of insulin’s main roles is to inhibit lipolysis (the release of fatty acids from your fat cells). Thus, longer residence time of fat in the blood makes a lot more sense in this context.

Alas, to really unpack this subject will take a bit of writing and referencing with regard to technical technical things like hepatocytes, VLDL-TG, and NEFA. But I’ll save that for a later post.

Gastric Clearance Hypothesis

A couple of people hypothesized that perhaps this may just be lower overall gastric clearance from the GI tract. Perhaps my food is taking a bit longer to digest on net as I have a higher protein to fat ratio on this diet (around 35% and 64%, respectively). Thus, entry of TG into the bloodstream via Chylomicrons would take longer overall. I think this is also a very interesting possibility and may have to think on how I could test that.

Retrospective Lab Analysis

Having taken pictures of all my food for over three years is about to pay off big for this situation. I can see on my spreadsheet 18 out of the 107 blood draws where I had TG above 100 which I could match to the last time I ate from the evening before as it will be on a picture that is times stamped. Scientific OCD for the win!

Of the 18 draws, there were 11 that can be explained by my having eaten below maintenance levels. Long time readers already know this is reflected in the Inversion Pattern, which includes TG levels (see right). Generally, I need to eat at least 2,000 calories on keto or I’ll lose weight.

And another five are explained by the carbs I was taking for an existing experiment these were taken from.

This left just two blood draws that were both high enough calorie, yet also low carb.

These draws were also in the 12-14hr window, so that wouldn’t explain the increase. But of course, these increases aren’t very dramatic anyway (102 and 129).

Final Thoughts

For over three years now I’ve been very interested in how fat-based energy is distributed in the body, particularly when carried in the blood by lipoproteins. Yet, I’ve probably put a little too much focus on LDL-C and LDL-P as these markers are much more stable and predictable when compared to triglycerides for a blood test. Triglycerides are “noisy” after all.

Or… maybe they aren’t after a long enough fasting time.

Nearly every test taken at 12 hours or longer was under 100 mg/dL for triglycerides, so long as I wasn’t (1) eating low calorie, low carb (invoking the Inversion Pattern) or (2) eating moderate to high carbs (typically due to an experiment)

Thus, this data is extremely powerful for two big reasons:

1. It shows just how easily one could be misdiagnosed with persistently high triglyceride levels, otherwise known as hypertriglyceridemia . Had I gotten a blood test at 10 hours fasted from almost any of the tests for last week, I’d have gotten that diagnosis. Just two hours later and I’d have instead gotten the optimal 100 mg/dL or less. (Think about that for a second — just a two hour window!)
2. It provides further weight to the lipid energy model. I certainly suspect we are seeing high redistribution of TG to both adipose and non-adipose tissue alike during the rapid fall in total levels. Fasting triglycerides in the hundreds under a Standard American Diet (SAD) is rightly associated with disease because it is usually continuously at these levels, potentially due to poor metabolic health. And yet, in this case, we see mine move from the levels of hypertriglyceridemia to very ideal levels over the course of the morning every single time.

So if you’re looking for the “tl;dr” — just look at the graphic below…

Note From Siobhan

After a discussion with Dave I decided to likewise check back through the data I’ve collected and see if I had any unexplained triglyceride readings at or above 100 mg/dL. I certainly found it interesting that, even though I’d been carnivore for over a year, I had never – as far as I remembered – seen oddly high triglycerides.

This could be for a few reasons though:

1. Up until recently, I’ve been drinking diet soda – some speculate that this could cause insulin to rise, which could possibly counteract the proposed carryover effect
2. I try to always follow my own advice and get blood draws around 13 hours fasted, so even if it was high at the 9-10 hour mark I’d likely never know
3. I may just not have enough data points to have had the opportunity to spot it yet.

Currently, I have 23 data points, where triglycerides were measured. Of those, there are 6 in which triglycerides are at, or above, 100 mg/dL:

• One is from 2015 where I was on a Standard American Diet and was likely due to my high carb/western diet.
• Two were taken while I was intentionally multi-day fasting (e.g. outside of the 12-14 hour window, reflecting the inversion pattern).
• Two were intentionally taken while I was sick, and thus are expected to be from an immune response from the illness.
• There is one where I wasn’t explicitly sick, with triglycerides at 102 mg/dL, however I was sneezing and felt a little “off” at the time. I retested a few days later and was in fact sick with a cold – so again, this could be explained by an immune system reaction that was currently under way.

This means that I have no unexplained high triglyceride results, perhaps owing to the fact of my consistent fasting time prior to the blood draw – although of course I’ll now have to see if I can replicate Dave’s data to confirm that.