The Marathon Experiment



The word wasn’t well enunciated due to the lancer I was holding between my teeth. It was a very cold, very early morning on a road just outside of Disney World. I was in a sea of fellow runners about to start the first Disney marathon of the year. But what set me apart was the pricking of my fingers for blood to apply to my Precision Xtra strips for both glucose and ketones. And unfortunately, the strips were failing.

As our corral moved slowly to the start line, waiting for our turn to be released, my head was tilted down trying desperately to get a solid reading. After a couple of failed strips, I managed to get a successful test from each.

Before I reveal that reading, let’s share what they were when I woke up two and a half hours earlier:

3:31 AM -- Glucose: 98, Ketones (BHB): 0.8

And right then, before the race began:

5:56 AM -- Glucose: 79, Ketones (BHB): 1.1

The shift in the numbers was as expected. I have an opinion of the “Dawn Phenomenon” you may have read about, but more on that later.

The plan was simple… on paper. At about every 5 kilometers, I was going to slow down and take both glucose and BHB. While I was excited for the new data, I was woefully undertrained, having my longest practice run before this day at only 6.5 miles. Ack!

To my surprise, I was able to run the first 30k with only slow downs to take my numbers.

Sorry for the blur, but I was actually moving while doing the test.

About 5k : 6:45 AM -- Glucose: 71, Ketones (BHB): 0.6

About 10k : 7:17 AM -- Glucose: 101, Ketones (BHB): 0.6

About 15k : 7:53 AM -- Glucose: 70, Ketones (BHB): 1.1

About 20k : 8:39 AM -- Glucose: 68, Ketones (BHB): 1.3

About 25k : 9:06 AM -- Glucose: 53, Ketones (BHB): 1.8

Drop Off

In college one of my professors once said, “Successful experiments make for great data. Failed experiments make great stories.” In this post, I have a little from Column A and a little from Column B.

At the 30k mark I whipped out my iPhone which by this point was at just 30% power. I then reached for the lancer to my Precision Xtra only to find it missing. As I was running, the lancer had fallen out of my pocket.

A wave of frustration and disappointment hit me all at once. I was so careful to use the zipper pocket specifically to prevent something like this from occurring. Yet somewhere between the 25 and 30k markers the zipper had widened on its own and the lancer snuck its way out while I was probably listening to Jump Around by House of Pain.

After checking and rechecking my other pockets several times, I finally mustered a restart to my pace. But my rhythm was different now. I couldn’t keep from stewing as I made my way forward. If you’ve ever ran a half or full marathon, you know how in many ways it’s much more a mental exercise than a physical one. And unfortunately, this was clearly a moment when my spirit left me.

About 30k-40k : ?

Finishline: ?

Over the next several miles, a few other key things went wrong which I won’t go into here. But regardless, I wasn’t able to resume my blood testing until after my wife joined me at the finish line and we returned to the hotel.

Back On Track

1:25 PM -- Glucose: 73, Ketones (BHB): 4.7

I was pretty hungry, so I had two bowls of AdaptMeal (4 portions total) while my wife was icing her legs. I planned to keep testing my blood every hour for the next several hours.

2:26 PM -- Glucose: 84, Ketones (BHB): 4.0

We then both went off to TGI Fridays where I had a salad and a steak with lots of butter. (Just an aside, running full marathons makes steaks very, very good!)


In the middle of the meal my alarm went off so I took my numbers again.

3:29 PM -- Glucose: 110, Ketones (BHB): 3.6

We finished up and went back to the hotel.

4:34 PM -- Glucose: 76, Ketones (BHB): 3.5

Both my wife and I laid down to watch a movie and found ourselves falling asleep (shocker!). I set an alarm for one hour, but my sleep self hit cancel when it went off and I instead woke up after two hours.

6:28 PM -- Glucose: 95, Ketones (BHB): 3.2

7:27 PM -- Glucose: 100, Ketones (BHB): 3.8

At this point my wife and I settled in to watch the Golden Globes and I decided not to take my numbers again until I was heading to bed.

(Bedtime) 12:27 AM -- Glucose: 92, Ketones: 2.1

I went to sleep, then woke up on my own at 6:25 am. Since I was planning to get my blood drawn for a full lab workup as part of this experiment, I decided to start my day.

6:29 AM -- Glucose: 95, Ketones 2.3

After a very long wait, I finally get my blood drawn and happily closed the chapter on this experiment.


What was I looking to find?

This was the final experiment in a series related to exercise that actually goes back to August of last year. If you’ve been following my blog for a while, you’ll note I’ve been adamant about trying to control for variables I assumed would affect my cholesterol numbers. The biggest of these variables would likely be energy demand from exercise.

One of the coolest things I’ve come to understand is that there is an anticipatory management effect with metabolism from regular, ongoing exercise. In layperson’s terms, the body is constantly figuring out what you keep doing with it and trying to front run those energy needs.

There’s clearly a global “priming” of cells with energy by the body. Yet there’s an equally strong counter regulation to preserve the energy you already have for survival. Use only as much as you need and save everything else.

So how much exercise and with what frequency drives this regulatory anticipation? How much of that is glucose vs fatty acids vs ketones? How much does this depend on sleep? Timing of the day? Warm ups, pre-exercise routines, or snacking?

Obviously there’s a lot to unwind and I can’t reveal the patterns I have discovered yet until I’ve gotten more of them reproduced and verified.

Next Steps

I’ve now done two very long stretches of exercise testing. The period from January to July of 2016 was low to moderate amounts of exercise. Then I started my training for the half and full marathons from August to now, keeping me at medium to high endurance exercise. (Note I was intentionally trying not to do any extra strength or resistance training as I believe this will affect many markers differently.)

  1. The next few months I’ll be… sedentary! Probably from now until about mid March.
  2. After that I’ll be shifting gears heavily into intense resistance training.

Extra credit if you know why I consider these the obvious next phases.

A Simple Guide to Cholesterol on Low Carb – Part I

If you’re on a Low Carb High Fat diet (LCHF), there are a few things you should know about cholesterol and how it is related to this lifestyle. In this guide I plan to cover the topic in very simple terms. It is by no means complete and is intentionally simplified to make it easier to read and understand for the layperson.

Our Energy on a High Fat Diet

Before talking about cholesterol, we have to talk about the energy you get from fat when on LCHF. Of course the main reason for food is to supply your body with energy. But how does that energy get to everywhere it needs to go in your body?

Like every other living thing, your body is made up of cells. Your heart is made of cells. Your brain is made of cells. So are your fingers, knees and toes. Almost all these cells need energy. And the vast majority of these cells ultimately get their energy from the blood circulating around your body.


The most commonly talked about source of energy is carbohydrates (carbs). Your body turns carbs into glucose to put in the bloodstream. From there, the glucose can circulate throughout the body, allowing hungry cells to grab some for themselves, and this is done with the help of insulin.


The other major energy source for your cells is fat, and by that I mean fatty acids. And like glucose, your cells also get their fatty acids from the bloodstream. Only there’s an important catch: glucose can swim in the bloodstream easily, but fatty acids cannot.

Fatty acids and the bloodstream are like oil and water, they don’t mix well. To fix this, your body cleverly does two things:

  1. It packages three fatty acids into a combo pack molecule called a “triglyceride”.
  2. And it makes a kind of boat for these triglycerides to travel in called a “lipoprotein”.

In fact, the kind of lipoprotein that delivers all these fatty acids is known as a very low density lipoprotein — or VLDL. After it delivers its energy, it remodels to a low density lipoprotein — but you probably know it already by its abbreviation, LDL.

Trigs vs Glucose

Common Confusion with LDL

Odds are you have most likely heard of LDL being used to describe cholesterol on a blood test. “Your LDL is high…” for example. So what gives?

Here’s the thing about cholesterol, like its triglyceride cousin, it also doesn’t swim well in the bloodstream. In the medical world, these molecules are commonly referred to as lipids. And lipids are actually repelled by water, so they are commonly called hydrophobic (hydro = water, phobic = repel). So when someone tells you they love to go to the beach but hate the water, mention they must be hydrophobic like cholesterol!

Yet what if the body has reason to have cholesterol available in the bloodstream as well? (We’ll get into that more in Part II) And while we’re on the subject, there are a few other things the body wants available to cells that are also hydrophobic, such as fat soluble vitamins (like Vitamin E).

So should it make a separate lipoprotein container for each of these molecules? No! It effectively packages all of them into the same boat: the lipoprotein.


That’s the genius of the human body. It has a kind of FedEx for all the hydrophobic elements needed by the cells. And most of whatever isn’t used gets recycled by the liver for many other possibilities, such as hormone or bile salt production.

Common Confusion with Triglycerides

The odds are likewise that you’ve heard “triglycerides go down on a low carb diet”. Indeed, blood tests for those going LCHF are almost universally lower in triglycerides. But a measurement of anything in the bloodstream is counting what is traveling around in that moment and not yet in use.

For example, Type 2 Diabetes has a common symptom of having very high glucose in the blood. This is because these diabetics are insulin resistant and have trouble getting the glucose out of the blood and into their cells. They may eat the same quantity of food as someone who isn’t diabetic, but glucose in the blood will spike higher and last longer by comparison.

If you have reduced your carbs and now get your energy mainly from fat, without question your cells are absorbing more of it from the bloodstream now. So even if you’ve increased the total amount of triglycerides going into the blood due to the diet, it is still brought down by the amount getting taken back out and used by the cells.

Common Confusion With Ketones

Another common assumption with LCHF is that you “get the majority of your energy from ketones” since you are in a state of ketosis. It’s certainly true your body makes many more ketone bodies from breaking down fatty acids, which will likewise feed your cells. This is especially important for proper brain function as ketones have special access that lipoproteins do not.

Yet while ketones are both produced and used much more on LCHF, they are still a secondary source of energy. The primary source of energy is still fatty acids brought to cells in LDL particles.


More Cholesterol is Trafficked on a Low Carb High Fat Diet

Now that you understand your body has need to move around more triglycerides to fuel your cells while getting the majority of your energy from fat, you may have already connected the dots.

  1. Your cells need energy
  2. On a high fat diet, their primary source of energy is triglycerides
  3. To get the triglycerides to your cells, your body sends them in very low density lipoproteins (VLDLs), which eventually remodel to low density lipoproteins (LDLs).
  4. All very low density lipoproteins (VLDLs) are made containing both triglycerides and cholesterol (but mostly triglycerides)

What are the Risks?

If you read the above and are struck with fear, I don’t blame you. It has been well drilled into our heads that more cholesterol in the blood = higher risk of cardiovascular disease and stroke.

But if you’re early in your research on this topic, let me help you skip ahead with one very crucial point (which I alluded to above). Making something available is not the same as using it.

Here’s a simple analogy – life rafts in the water and being used are typically a sign of trouble. Yet all ships sail with them on board. This is a good idea in case of an emergency, of course. But if you were only counting life rafts whether in use or not, then you’d assume a lot of ships entering view was by itself a sign of trouble.

Cholesterol is like the life rafts on the LDL ships. Even if it travels with your triglycerides, it is a much smaller passenger (in quantity) and mostly recycled back at the liver. You don’t actually care how much cholesterol is in your blood — you care how much cholesterol leaves the bloodstream and causes a build up of plaque in your arteries (atherosclerosis). And this is at the core of the inflammation debate with cholesterol. Is it a life raft for damage to the blood vessels? Or is the sheer presence of it risk alone? (You can probably guess where I fall on this one.)

In Part II we cover part of the journey of cholesterol in more detail through a very visual comic form. (Part III is coming soon…)

How To Do The Cholesterol Drop, Step by Step

After several have approached me both on and offline about it, I’ve decided to make a page that lays out my experiment and how you can do it to.

And once again, PLEASE keep close track of your data and share your results, whatever they are. The more people who do this, the more data we have.


Dropping My Cholesterol at Record Speeds – Part II

In Part I, we focused on the three day gap where I arranged a massive shift around the first public presentation of my data. There’s quite a bit more to the story, though, which we’ll dig into here…

Knowing I’d be presenting on October 9th made it immediately clear I needed to test on the Friday before (7th) and the Monday following (10th). This should reflect the shift in LDL-C and HDL-C, given they both follow that three day window immediately before a blood test.

But what about LDL-P, and for that matter, small LDL-P (smLDL-P)? Since these follow a three day window, but with a two day gap in between, I needed to be sure I captured all five days before Oct 7th and the five days following. So in short, I’d be on the “low calorie diet” on October 2nd, 3rd, 4th, 5th, 6th until the morning of the 7th (Sunday, Monday, Tuesday, Wednesday, Thursday). I’d then take the blood test and ramp up the food intake to a “high calorie diet” for the rest of the 7th, then the 8th, 9th, 10th, and 11th (Friday, Saturday, Sunday, Monday, Tuesday) up to the morning of the 12th.

Below, in each graph, I mark the blood test associated with these four tests over the ten days in red, which will always be at the right of the graph.


So let’s start by looking at our old friend, LDL-C.


For those of you who regularly read my blog, this is a familiar graph. In this case the blue solid line represents the LDL cholesterol that resulted from the blood test. The dashed orange line represents the dietary fat I ate in the three days just before the blood test was taken.

At the far right end you’ll see a solid red line replacing the blue. Those four data points took place October 5th, 7th, 10th, and 12th. You can also see the dashed orange lines showing my extreme low, followed by the extreme high in dietary fat.

Let’s invert the axis on the left side of the graph (the Three Day Dietary Fat) so that we can compare these trends visually.


Sure enough — I “broke” my own correlation given how quickly I shifted from one extreme to the other. Given our correlation on the left side of the graph, you’d expect my extreme low on dietary fat of 63g would theoretically push my LDL-C up to around 348.

However, I did manage to bring the LDL-C down to 155 which did match the expected position relative to the diet. Yet this completed on the 12th, not the 10th. Again, this was probably due to the degree of shift, but we’d need to do more testing to know.

How about LDL-P?


Same thing. Purple is pre-conference, red on the right is during. Let’s invert again…


Like LDL-C above, the massive shifts seem to surpass my metabolism’s attempt to catch up.

Now let’s look at HDL-C.


As you can see from the green line on the right side, we have a long established coupling in the positive between dietary fat and HDL-C. But this too is broken by the experiment as shown on the right.

Other Interesting Markers

I had a number of things change, but there are two specific ones I wanted to call attention to in this post.

Glucose vs Insulin

Now I don’t want to blow your mind too much, so prepare yourself… here comes fasting glucose vs insulin…



Whoa — really?!? As my fasting glucose went up, so too did my insulin? No way!

Okay, kidding aside, it’s difficult to tease out the impact of each macronutrient on my fasting glucose to see what the primary drivers of the increased insulin were. Since all of my macros went up, this could’ve been from higher carbs or protein — or possibly just overall caloric load.

What is interesting is that even after 5 days at 5000 calories, my peak insulin was still at the bottom third of the reference range, at 8.3. Incredible!

Calories vs Weight

Easily one of the most hotly debated subjects in low carb communities. I’m firmly in the camp that says Hormone Balance Matters Most, But Calories Are Still Energy. Is that a side? If so, should we start a website for

I’m 6’3″ (190.5 cm) and thus feel a bit underweight at my running average of around 176 lbs (79.8 kg). My typical average calories right now are around 3,100. So what happens when I drop it to 750-ish for five days, then ramp up to 5,000+ for five days?


calories_weight_tableI’m missing 10/2 and 10/3 due to not being around a scale (was on a camping trip), but presumably I had a similar trend as 10/4-7 given I hover around 176 lbs as mentioned above.

Clearly the massive drop, then rise, in calories had a very acute effect on my weight. Over this span of time, I was careful to keep my water intake roughly the same for both low and high periods, although my non-water beverages were a little higher on the last half.

Final Thoughts

I’m now at the point of being so sure about the dynamism of the lipid system that this is becoming almost comical. Getting to this point took a lot of money, time, and blood (literally). I was fortunate to have this experiment wrapped into my first presentation. Doing the demonstration was a no-brainer. Why only show the data if you can display its predictive power as well?

There’s actually a lot more data I couldn’t get to due to the balancing act I’m now performing between this, contract work, interviews, and helping others to gather N=1 data as well (upcoming).

Personally, I wish I could work on this research 100% of the time, but I do have to pay the bills…

Presentations, Data, and Exercise

The egghead in me is dying to complete the geekier Part II of Dropping My Cholesterol at Record Speeds. But alas, there are a number of high priority developments rolling in.

  • I had to post a draft of my Food Logging Protocol given there’s some discussion of putting together a study based on my research. It’s been an effective tool for me to keep accurate, “audit-able” track of my diet. I still have some more to put in the page regarding the text portion of logging, but the visual language is the most relevant.
  • Since my last posting, I was approached on doing another presentation which I’ll be putting on tomorrow in a private meeting. Unfortunately, I can’t take a video of it, but I might do a separate one to post online later.
  • New data coming in from two new participants, more on that in future posts. My hope is that I’ll have something up at the end of this month.
  • How will endurance exercise impact my trend lines? This too we’re finding out right now. Over the last month I’ve now done a half dozen more blood tests both before and after running for long periods.