(skip to content) jana e. beck

HFCT #2: The Plan

HFCT is the acronym I’m using for a high-frequency cholesterol testing study I’m conducting on myself (along with a group of QS-ers conducting similar studies) in the fall of 2017. This is my 2nd blog post about it. Start back at the first to catch up.

Over a period of about three months, I’ll be following three different diet protocols and collecting both lipid panel and blood ketone data alongside detailed food logs and approximate hydration logs, as well as lots of contextual data that I always collect anyway1. The CardioChek Plus machine I’ve been loaned measures total cholesterol, HDL cholesterol, and triglycerides via PTS Diagnostics’ three-analyte Lipid panel test strips. For blood ketones, I’m using my own Abbott Precision Xtra blood glucose and ketone meter.

The structure of the testing I’ll be doing is to take ten lipid and blood ketone readings every two hours on four days across my experiment: an initial baseline, my baseline while eating a very low carbohydrate diet, my baseline while eating a low carbohydrate diet, and my baseline while eating a high carbohydrate diet. On every other day that isn’t a “baseline” intensive testing day, I’ll be taking a lipid panel reading in the morning after fasting for at least twelve hours.

The three diet protocols I’m following are:

  1. a very low carbohydrate and ketogenic diet of <=20 grams of carbohydrate daily for two weeks
  2. a low carbohydrate diet of 20–50 grams of carbohydrate daily for 13 days; this may or may not turn out to be consistently ketogenic2
  3. a high carbohydrate diet of approximately 225 grams of carbohydrate daily for 16 days3

I picked these protocols for several reasons. First, it has been shown that, in a side-by-side comparison with a low fat diet, low-carbohydrate diets are correlated with a higher level of LDL-C4, the so-called “bad” cholesterol5. On an individual level, however, levels of LDL-C may rise, fall, or stay the same on a very low carbohydrate diet compared to the individual’s previous levels prior to making the dietary change. So I’m interested in seeing where I fall as an individual. When going from a very low carb ketogenic diet to a low(ish) carb diet to a high carbohydrate diet, how will my LDL-C change? I’m also very interested to see how my other lipid levels change or don’t change in ways that may correlate with greater or lesser risk of CHD, particularly my very high HDL.

Secondly, the motivation for my choice of these particular carbohydrate goals for each dietary protocol—<= 20g, 20-50g, and ~225g, respectively—stems entirely from my history with type 1 diabetes. In particular, the ~225g protocol is designed to mimic the dietary advice that I was given upon my diagnosis in 2003: I was instructed by the dietician I saw, who specialized in diabetes, to eat 75 grams of carbohydrate across three meals a day. This is an example of the kind of advice that is unfortunately still quite typically given to those with type 1 diabetes: eat a “normal” amount of carbohydrates6, just make sure to “cover” your carbs with enough insulin.

This is bad advice on a couple of fronts. First, the larger the dose of insulin, the more danger involved. A ten unit dose of insulin to cover a large meal could easily drop my blood glucose into the fatally dangerous zone if I time it wrong. (And timing is tough with insulin because there are lots of variables involved: the glycemic index of the foods you’re eating, the fat and protein content of the food you’re eating, where you inject the insulin, how warm your body is or gets7, any exercise or other activity you engage in, &c.) Second, the lower your carbohydrate intake, the easier your blood glucose is to control, even abstracting over the insulin dosing and timing issues. Low carbohydrate intake means smaller, slower spikes in blood glucose (because you’re typically consuming most or all of your carbohydrates from low-GI sources, like leafy greens and non-starchy vegetables). The smaller, slower curves—they can’t even properly be called “spikes”—give you a lot more time to watch changes and react to them. Simply put: low carb is not a rollercoaster; it’s only moderately more exciting than driving across Nebraska.

In the last year, I’ve been heartened to see a prominent voice in the diabetes community take on the “just cover your carbs with insulin” advice. Adam Brown, a columnist in the popular diabetes newsletter diaTribe, wrote this in one of his columns:

I’ll never forget the diabetes food advice I received from my doctor at diagnosis: “You can eat whatever you want, as long as you take insulin for it.” In my view, this advice is misleading, overly simplistic, and damaging. In fact, I’d nominate it for the “worst” diabetes food advice out there.8

All of this is to explain that my third dietary protocol is designed to mimic the advice I got at diagnosis as a means to proving—with numbers—exactly how awful this advice is. Despite the fact that I love carbs (🍞 🥐 🥖 🍌 🍐 🍒 🍓 🍅 🍿 🍱 🍝 🍨 🍩 🍪, and of course 🍫 among them), I’m almost dreading the third phase of my experiment. It’s not that I never eat 225 grams of carbohydrate in a single day anymore, but I don’t do it that often, and certainly not every day for sixteen days. The consistent high carbohydrate intake is what worries me the most, because I worry that it will make my blood glucose rollercoaster constant, and that’s a dangerous situation to be in. I know what to do if I get myself into the very rare situation of euglycemic diabetic ketoacidosis (DKA)9 while eating very low carb (eat some carbs and take some insulin; drinks lots of electrolyte-supplemented fluids; hie thee to a hospital if symptoms don’t improve). And let me emphasize again: euglycemic DKA is very, very rare. Thus, I’m far more worried about the risks of low blood glucose from the rollercoaster 🎢 I’ll be on during my third dietary protocol, not the very outside chance of euglycemic DKA on my first.


The next post gives more detail about the protocols I’m employing in my (self-)study.

  1. Namely: diabetes data including insulin dosages and blood glucose via the Tidepool uploader, menstrual cycle data in the Clue iOS app, activity and sleep data through my Jawbone fitness tracker and the Moves iOS app. 

  2. My experience in the past is that if I start with very strict low-carb, I can gradually introduce more carbs and remain in ketosis. But I’ve never tested this at the high resolution of blood ketone testing I’m going to be doing in this study. 

  3. The slightly uneven lengths of each dietary protocol are due to planned travel during the ‘Blood Testers’ group experimental period. Scheduling each dietary protocol and lining up each protocol’s “baseline” day of intensive lipid testing on a day when I’d be able to do intensive testing every two hours was a rather over-constrained problem and quite a challenge to overcome! 

  4. Volek, J.S., M.J. Sharman, and C.E. Forsythe, Modification of lipoproteins by very low-carbohydrate diets. J Nutr, 2005. 135(6): p. 1339-42. Cited by Phinney & Volek in their The Art and Science of Low Carbohydrate Living

  5. There is also a great deal of evidence in support of the idea that LDL-P (LDL particle size and number) is a much better predictor of CHD than LDL-C (LDL concentration). (Howard, B.V., et al., Low-fat dietary pattern and risk of cardiovascular disease: the Women’s Health Initiative Randomized Controlled Dietary Modification Trial. JAMA, 2006. 295(6): p. 655-66. Also cited in Phinney & Volek.) LDL-C is much easier and cheaper to estimate than LDL-P, which is one of the reasons there has been so much focus on it up until recent years. I may even add an nuclear magnetic resonance (NMR) lipoprofile to measure my LDL-P (as well as many other more advanced lipid measures) at the final stage of each dietary protocol in my experiment, just to get more useful information. 

  6. I would argue today that 225 grams of carbohydrate a day for my particular height and body type crosses beyond “normal” into “high.” For example, I gained weight above what my pre-diabetes weight had been on the 225g/day diet. 

  7. This is one not many people know! Increased circulation vastly speeds up the absorption of insulin injected subcutaneously. I’ve experienced this in all sorts of situations, from having my insulin pump infusion site on my thigh while wearing dark jeans on a sunny day to taking a hot shower or bath shortly after taking a dose of insulin. 

  8. Adam’s Corner, “The Best and Worst Diabetes Food Advice I’ve Seen”. Brown also discusses his low opinion of this advice in his book Bright Spots and Landmines: The Diabetes Guide I Wish Someone Had Handed Me

  9. Euglycemic Diabetic Ketoacidosis, a Misleading Presentation of Diabetic Ketoacidosis