My previous experiment where all of my calories for 4 weeks came via Keto Chow brought out some interesting data in my 3rd week. That week I changed my fat source from heavy cream over to avocado oil, this resulted in some remarkable blood tests. A question arises as to whether my elevated Triglycerides, lowered LDL, Lowered HDL, and more are a result of the reduced carbohydrates (under 2g net a day average), or the composition of the fats I was eating: Saturated vs. Monounsaturated, vs. Polyunsaturated. This experiment intends to answer that question by isolating the fatty acids as the only variable.

At the same time, I put another popular theory to the test: CICO or “Calories In = Calories Out”. The theory is that a calorie, is a calorie, is a calorie; and that the kind of calories makes no difference, only the quantity of calories. People get overweight because they eat more calories than they burn, which is the same as saying people get wealthy because they spend less money than they earn – both are technically true but don’t tell you the why, just the what. Even die-hard CICO believers will balk at giving kids cookies for dinner (which is funny, I thought a calorie was a calorie!). Regardless, I was able to definitively isolate the number of calories as a fixed variable with the only change being the type of calories: either from different kinds of fats, or from different forms of glucose and fructose. If the type of food is irrelevant, I should have seen no change in weight loss, blood sugar, energy, or other results. My hypothesis was that during the week I’m replaced fat with carbohydrates I would:

  • Gain weight, initially from glycogen (water weight) but it will likely go beyond that.
  • Feel like crap with little energy, despite an overabundance of glucose in my system.
  • My fasting insulin would go through the roof.
  • My A1c would go up.
  • My Blood sugar will also go way up.

Experiment Methodology

Each week I will be testing a different fat/calorie source. Each will be tuned for 2000 calories a day total caloric intake.

  • Week 0: Heavy cream for my “wash-out” period
  • Week 1: Oil high in PUFAs (Grape Seed Oil)
  • Week 2: Heavy Cream – high in Saturated Fatty Acids but does not contain Medium Chain Triglycerides which could confound the results.
  • Week 3: Oil high in MUFAs (Macadamia Nut Oil)
  • Week 4: Oil High in Saturated Fatty Acids (Liquid Coconut Oil – this is also high in Medium Chain Triglycerides which are metabolized directly into ketones)
  • Week 5: Carbohydrates – via dextrose, soda, and candy

I will be taking fish pills all 6 weeks, including week 5 where I will be only consuming 28g of fats a day. This will be the embodiment of conventional recommendations of high carb low fat. I expect to feel like crap and to gain weight using the same calorie amount as the previous weeks.

Test Plan:

  • After week 0, I had blood tests taken, these were repeated every week (Friday) at the end of that 7 day period.
  • At the beginning of week 5 (carbs) I would do a test to measure the effect of drinking Keto Chow with dextrose and candy on my postprandial glucose so I could compare this result to an earlier test of just drinking Keto Chow with heavy cream.
  • Every day I would test my blood ketones – I used a “KetoMojo” blood tester for this due to the lower cost per ketone test. I would also check my blood glucose during week 5 (carbs)
  • My weight would be tracked using a Withings (Nokia) scale.
  • There were a total of 3 DEXA scans: before week 0, before week 5, and after week 5

[visualizer id=”23810″]

[visualizer id=”23808″]

[visualizer id=”23814″]

[visualizer id=”24583″]

[visualizer id=”23815″]

[visualizer id=”23803″]

Food Logs and Nutrient Analytics

Fatty Acid Composition Effects on Biomarkers and CICO on Trial


One of the main criticisms of Morgan Spurlock’s “Supersize Me” is that he absolutely refuses to release his food logs. You can’t take an experiment, even an N=1 experiment, seriously unless you can look at the data. Here are my own food logs for the entire experiment. The data was collected using Cron-O-Meter.