For more details on this graphic and the story behind it, see the post following.
“This is the very first public presentation of my data,” I began, “and I’m excited because I get to start it off with a bang.”
I was standing before the Ketogains Seminar audience on October 9th, 2016 in Las Vegas, and I planned to make that day just a little extra special. The first slide I showed included a picture of me smiling as my blood was taken, with the heading, “October 7th, 2016 – Two days ago”. I then explained I expect my cholesterol on that day will be very high.
I follow with a second slide. It has the same picture but I’m a silhouette with a question mark inside. The heading reads, “October 10th, 2016 – Tomorrow”
Finally, I present a third slide that includes both images and my predictions together, the second reading plainly, “Cholesterol Plummets in 3 Days”.
“How am I doing this?” I asked. “A super powerful statin? High dose of Niacin? Some new, special exercise?… No… just food.”
I’d be lying if I didn’t confess at this point that my heart was pounding. I’m no stranger to public speaking, with one of my very first jobs being the M.C. of a dive show and working my way through college as a D.J. No, my nervousness stemmed from the fact that I was trusting all my research to date and taking this to the next level of publicly predicting an outcome. And not just any outcome… the outcome a thirty billion dollar prescription industry tries to provide.
So what happened? Quite a lot actually. Let me break it down into small pieces…
Sure enough, my Total Cholesterol dropped 66 points. But the real star of the show is LDL-C, the so called “Bad Cholesterol” which dropped a whopping 73 points in just three days — that’s nearly a 30% drop!!!
If I could put these results in a drug, I’d be a multibillionaire right now. Bayer and Pfizer would be pounding down my door offering a slice of the world to package this magic in a bottle.
But what about the actual LDL particles which show on the NMR labs as LDL-P? Or for that matter, the “Small” LDL-P that gets the credit as being the most dangerous of all?
LDL-P is nearly cut in half! This was even more impact than I had predicted. While it is true my LDL-P continued to drop into the 5th day (to be covered in an upcoming post), the shift in just 72 hours was really impressive.
Even more strikingly, Small LDL-P drops from 442 into the unreadably small “<90” range. Like LDL-P above, this runs very counter to the endless articles I’ve read about clearance with LDL receptors, which has been one of the chief suspects as to why LDL gets “backed up” in the bloodstream. Over and over, my data keeps providing a very different story.
HDL-C and Triglycerides both shifted in different directions, each one considered more favorable.
HDL-C bumped up 18 to 68, which is likewise consistent with my data.
Triglycerides are still a little more mysterious to me as they have the lowest correlation of the bunch over my research span, but are still negatively correlated with dietary fat. In this case, I was excited to see it dropped to nearly half of its earlier reading.
Food. Just Food
Unless you’re just now joining us, you already know how I pulled this off. More fat.
On one hand, if this is the first blog post of mine you’ve ever read, you might think I mistyped the above sentence. I most certainly didn’t. As with the last eleven months of my life, I carefully logged my food in MyFitnessPal to track it closely. I likewise took pictures of everything I ate. Everything.
On the five days prior to the Oct 7th blood test that gave me the high cholesterol, I had brought my total diet down to an average of 748 calories per day. This wasn’t pleasant given I’m a 6’3 male who gets semi-regular exercise. (Not to be confused with “athletic”)
Moreover, I didn’t want to stay too low on the calories for very long given I’m also very lean and would actually prefer to weigh a little more. I know this has me on the other side of most people in a low carb lifestyle, but actually aligned well with many at Ketogains.
By the time Friday was rolling around, I was excited to finally move to the next phase and ramp up the total food to full throttle. But unlike my Intentional Outlier experiment that I describe in Part I, I planned to exceed everything and shoot for the gold: 5,000 calories per day for five days.
The first day was pretty easy, most likely due to having been so low cal for the previous five. Each meal I had to double or even triple my portions. But eventually I figured out it was just better to spread it out over the day. By the last day I could barely stand it. It’s incredibly difficult to eat very high calories on such a satiating diet. I kept telling my wife, “ugh, I can’t ever do this again,” to which she rolled her eyes knowing my variety of experiments before it and said, “yeah, we’ll see.”
Thus, many at the Ketogains Seminar noticed I was constantly pulling food out of my backpack in various forms: Adapt bars, peanut butter, and hot dogs I kept in a small cooler to name a few. This was especially annoying given the conference was catered with several amazing Keto-friendly buffet delights. But I couldn’t partake, given I needed to be certain of the exact macros and general ingredients I was consuming.
One particularly funny moment happened near the end of Sunday where I was sitting with other presenters at a Q&A and I realized I was still 1,200 calories short for my 5k goal which I had to meet by 7PM in order to have the proper 14 hour gap for the blood test at 9PM the following morning. I left my chair to go get a block of cheese and munched on it for the remaining 15 minutes of discussion.
Having said all that, let’s revisit our final version of the chart where we’ll include the food as well:
I’ll concede I’m still pretty awestruck. Even as I write these words and post this data, I keep rereading the labs and checking back over my logs. But there’s no doubting it now – the experiment delivered. The Dietary Inversion is very much real and yet almost no one knows about it.
Maybe this will move that needle just a little more.
This Part I is meant to be more layman, less geeky. In Part II I’ll be expanding on the data including the test before Oct 7 and after Oct 10th and how all of it compares to my current correlation timeline.
I’ve had quite a few things on my plate over the last six weeks. In fact, four and a half of those weeks were travel alone. Obviously this has quieted the experimentation and blogging somewhat, but that’s about to change.
I have a presentation coming up for the Ketogains conference October 8-9 in Las Vegas. I’ll share a summary of my N=1 data to date, including a recent new surprise that will be meaningful for fat-adapted athletes. I’ll also be giving my assessment of risk for high cholesterol on a low carb diet and why everyone should make an informed decision on this important subject.
For the short version with pictures, see below. For the long version, read on after…
A couple months ago I started talking to my sister about taking my data to the next level. But to do so, I’d need someone else’s help — to which she immediately volunteered. (Side note: my sister is uberawesome!)
In fact, my sister was perfect for what I wanted to test specifically. While her cholesterol numbers went up after going low carb last year, they didn’t rise nearly as much as mine. In fact, both her LDL-C and LDL-P were generally half of mine. Thus, we would have different starting points on our cholesterol when we ate, which is the goal.
I then started planning all our meals to be similar to my prior March week-long solo test. There would be the same day-by-day blood tests during one week. Only this time we’d add one test for the Friday before the week, and the trailing Monday that followed it, seven blood draws in all, each was an the advanced cholesterol test, NMR (Nuclear Magnetic Resonance).
Once started, we had to eat the same food at exactly the same time, no exceptions. Both of us had to take pictures of everything we ate, along with weighing them when possible. Half of the time she was in her state, then flew to me in my state where I’d then on prep, weigh, and cook our duel meals personally.
Generally, I tried to keep our food mostly home-prepared and stay away from processed or fast food. But both before the meal plan and following we ate out a little more. Also, my sister likes Zipfizzes, so we agreed to have one a day through the planned days as well.
Overall, the plan worked beautifully! My sister stayed religious to the diet and timing and we didn’t have any sudden surprises that thew us off the rails. I have a few fun stories that I’ll save for in person talks later.
The Comparative Data
Beyond the Total Cholesterol hook above, it’s worth looking closely at the other markers as well.
Our LDL-C was an impressive 88.9% correlative with each other! Did I put only one exclamation mark there? I meant three — 88.9% correlative!!!
And here’s a relative comparison to really see the match up:
This was especially relevant to me given my general theory encompasses energy trafficking as being the primary driver of these LDL cholesterol payloads. If I’m losing you here a little, don’t worry, I’ll cover this in a future post.
Like my own data before this, HDL doesn’t often move too much, but typically tracks with three day dietary fat in a positive correlation. More fat, more HDL. Between the two of us, we correlated a solid 71%.
This next piece of data is extremely relevant to me (which I’ll get into in the theory post). It also tends to have a high standard deviation relative to the other markers from my past tests. However — in this case it was remarkably close in comparison to each other’s at a 77%. Incredible!
So here is where things get interesting. On both LDL-P and Small LDL-P, Darla and I track very closely with the exception of the very last data point (7/18). In fact, the metric is so off course as to be suspicious to me. Up to that test, we had been eating everything identically as with the others, so what happened?
I’m loathe to suggest a lab error, especially since the non-P metrics appear to line up correctly. But unfortunately, there’s no easy way to find out as I have no direct contact with the lab (as it should be). For now, I’ll list both the complete results and what the correlation is without it and you can judge for yourself.
It’s hard to quantify in words how happy I am that we captured this data and confirmed the previous patterns I’ve observed to this point. Our next steps will a new N, Nicole Recine, a Ketogenic Practitioner who has graciously accepted being our #3. We’re currently in the planning phase and hope to set up the next capture in the coming weeks.
I couldn’t end this article without given a very sizable thanks to my sister, Darla, and her contribution to this science.
I just attended Low Carb USA at San Diego where I shared much of the data below. And while I was interested in a possible divergence that seemed to appear at the end of May in Part III, it turned out to be more of a one-off, probably due to a higher percentage of protein and a lower percentage of fat than my usual ratio.
In upcoming Part V, I’ll be revealing some new data on a “second N” to my study. I should have that up within the week.
For now, note that the new 21 to 28 data points include a 9 day period where I once again did a total of 7 days of blood draws. Thus, we again can see this mechanism in nearly real time.
I’ll let the graphs speak for themselves…
Three Day Average of Dietary Fat vs the LDL-C in the resulting blood test. The LDL-C still tracks inversely with total fat. (-81%)
Same blood tests, same dietary fat, but for HDL-C — which clearly tracks positively higher total fat. (65%)
Same blood tests, same dietary fat — but with a 2 day gap in between (Days -5, -4, and -3), but for LDL-P — which tracks inversely with higher total fat. (-82%)
And finally, same blood tests, same dietary fat — but with a 2 day gap in between (Days -5, -4, and -3), but for small LDL-P — which tracks inversely with higher total fat. (-72%)
If by this point you don’t see this is a highly regulated, highly responsive network in the lipid system (at least for my N=1), then you think I’m some kind of X-Men mutant. (In which case, I dib the name, Captain Cholesterol)
I now have very high confidence that this regulatory pattern is likely present with virtually everyone who is fat adapted (getting the majority of their energy via fat) without an underlying metabolic condition.
It’s also quite possible this applies to those who are not fat adapted yet still with no underlying metabolic condition. For that, we’d need more study.