When we went to Low Carb USA (San Diego) earlier this summer, over a meal I mentioned to Dave Feldman that one of our sets of twins had one doing keto (hereafter referred to as “KT”) for a year and the other not doing keto (hereafter referred to as “NKT”). Dave got really excited (that’s a SEVERE understatement) and wanted to know if we could convince them to get some blood tests to see how keto was affecting their cholesterol, insulin, glucagon, and other factors. Dave wanted to make sure they had similar activity levels. Both have Fitbits Alta HRs, you can see the raw data from those here (KT) and here (NKT). They are relatively close in activity 169,928 steps for KT vs 158,278 for NKT over a 2 week period before the blood tests. Sleep patterns are noticeably different: the Fitbit on KT recorded 5524 minutes of sleep, the Fitbit on NKT recorded 6388 for the same 2 week period; that’s 92 vs 106.5 hours.

The same 2 weeks leading up to the blood draw, each was taking Dave Feldman style photos of everything they ate, which proved to be the biggest annoyance for both during the experiment and they were looking forward to concluding so they could stop with that. This was to ensure that we had a record of what they were eating and that both stayed pretty much consistent with their food during the 2 weeks before the blood draw. The day of the blood tests came, KT was worried as she has had issues with blood and fainting – strangely KT had absolutely no problems at all with the blood draw and was done in about 3 minutes. NKT, on the other hand, went pale and ended up needing to lie down before the lab could finish the blood draw. We were very happy that she came out OK, though I doubt she’ll be donating blood anytime soon. =)

Before we get to the results I should explain that both parents in our home are quite strict with keto for ourselves. Two of our daughters are also strict about staying keto, though both are more “lazy keto” and don’t track carbs or such – frankly, it’s not really necessary. As young and healthy as our children are, simply the act of cutting out sugar and flour was likely enough to prevent problems in the future. Our other children are not doing keto and will eat candy, pizza, fruit, and such if given the opportunity – though we, as parents, do not buy anything non-keto anymore so it’s usually at school or when visiting.

Now for the data! Dave wanted to get a recording of his first read-through of the blood tests:

Most of the analysis was done by Dave during the video so refer there for commentary. If you prefer to look at the data yourself, you can either read it in tabular format on the “Twins” tab of this spreadsheet (the other tabs are data from previous experiments I’ve done) which also has % difference and absolute value difference columns for comparison, or you can grab the original blood tests in PDF format: KT, NKT. I’ve made some handy graphs (also on that spreadsheet in the “Twin Charts” tab) to illustrate some of the takeaways.

To start, we have a comparison of their NMR Lipid Panels. Nothing here is really a surprise at all, KT has higher numbers for all of the measurements except LDL size which is the same for both, this follows Dave’s hypothesis regarding the lipid system as an energy distribution aparatus. Both are low on HDL but have fantastic Triglyceride levels. Both also have a remnant cholesterol of 8. Dave has explained this better but this number represents energy parked in the blood in VLDL particles and you want this as low as possible since it tends to be a legit indicator of risk; an 8 is fantastic.

Both have really good fasting glucose.

Which should make their nearly identical Hemoglobin A1c not a surprise at all. This number represents the percentage of their red blood cell hemoglobin that has become “glycated” or exposed to glucose. It serves as a way to measure your average glucose over several months, usually 6 months. HbA1c is the test used to diagnose Type 2 Diabetes Mellitus, although elevated HbA1c is a symptom of T2DM, not the cause. Most experts agree that the underlying cause is hyperinsulinemia or chronically elevated insulin levels.

Boom.

KT’s low fasting insulin doesn’t indicate how close she is to having hyperinsulinemia or her risk of T2DM or Alzheimers – that would require a Kraft test and measuring insulin response to glucose over time instead of once while fasting. NKT’s elevated fasting insulin does indicate that even without eating for 12 hours, her insulin was still elevated. For now, she’s likely to be able to handle that insulin load; but Dave feels that he’d like to see insulin below 10 in the future. This higher number for NKT was one of the few surprises that were in the test, most of the other markers fell in-line with what Dave expected to see based on the other results he has seen from both keto and non-keto subjects.

The other surprise was the low glucagon levels in both. A typical level is 50, and one would expect KT to have higher than normal levels. While hers is higher than NKT, both are abnormally low. Glucagon is something that Dr. Benjamin Bikman presented about at the 2018 Low Carb Breckenridge conference, interesting stuff.

So, in summary: It’s interesting to note the difference in the LDL numbers of the two. Many professionals are of the opinion that LDL is causal in atherosclerosis and would insist that KT has a higher risk of heart disease in the future. The cool thing with this data is that genetically they’re identical, with the same environment, same living conditions, same lack of cholesterol-lowering medication. They actually have almost the same triglyceride numbers too! NKT and KT both are handling the glucose load really well (that’s rather expected, given all the machinery is still new and fully functional). It is important to note the massive difference in their respective fasting insulin. So many of the chronic diseases in modern society are a direct result of too much insulin: chronic hyperinsulinemia. NKT is in danger of hyperinsulinemia causing problems in the future which is something to be truly concerned about.