This past spring I adopted a diet resembling Dave Asprey’s. After about 5 weeks on it, I took a fancy blood test (trying to optimize things such as my vitamin D levels, Omega-6/Omega-3 ratio, etc). When I finally got the results back, I was shocked to find that the most important results involved my cholesterol levels.
Six years earlier my cholesterol had been good enough that I hadn’t had any reason to pay attention to it. But this May it was 369, with an LDL of 208, and an amazingly good HDL of 108.
If I’d been unlucky enough to get those results from Kaiser, I’d have felt helpless for quite a while. But the WellnessFX report also mentioned that my ApoE Genotype is 3/4. When I got around to researching the implications of that, I figured out that the ApoE4 gene can contribute to some unusual reactions to fat, and that a few people with a copy of it had reported skyrocketing cholesterol in response to a diet of paleo foods plus much coconut and grass-fed butter.
So I cut back on my fat intake, and figured out how to test how my diet affects my cholesterol. I’ve been using CardioChek about once a week. It’s unreliable (about 50 mg/dL too low, judging by my two comparisons with Kaiser). And I’ve only been getting total cholesterol numbers. I could get better info by also measuring HDL with it and subtracting that from the total to get the number that I really want to reduce. But that would require getting a second blood sample, and I barely have enough patience to take one good sample (I often end up with too little blood and need to start over).
With all the uncertainty about the accuracy of those numbers and the time delay between altering my fat intake and when my cholesterol levels showed my response, I wasn’t too optimistic about getting clear results. But the first ways I quantified the evidence gave correlations of 0.540 or 0.418 using 12 or 13 data points (the first of those excludes an implausibly low reading the first time I used it). That’s using an exponential moving average of my saturated fat intake that roughly corresponds to the past week. Other averages of my saturated fat intake have produced somewhat similar results.
My saturated fat consumption ranged 23.5 to 53 grams per day for that data, and the cholesterol levels at the high end of the range are something like 30-50 mg/dL higher than the low end.
So I see a clear and moderately strong connection.
My limited attempts to analyze subclasses of saturated fat have been inconclusive.
Now, how good or bad is the combination of very healthy HDL and disturbingly high LDL?
There are a variety of algorithms to estimate my risk based on my cholesterol levels. Some weight LDL heavily and say I ought to be concerned about the higher readings. Others weight HDL more, and say I’m unusually healthy. The TC/HDL criterion says I have improved with each of the tests I’ve taken [excluding the home tests that only measured TC].
Without strong evidence about which algorithm is more reliable, I’m disposed to keep my cholesterol somewhat close to normal levels.
What about the evidence that societies with high coconut consumption are healthy? That could mean that high cholesterol with ApoE4 is a healthy sign – the body sees lots of rich food, and decides it devote more resources than normal to reducing inflammation. Or it might mean that ApoE4 genes got selected out of those populations.
It looks like coconuts have been eaten for the longest time in islands of the west Pacific, with the Philippines currently having the highest consumption for a sizable country, and high consumption extending west to India.
According to this study the ApoE4 prevalence in countries from the Philippines to India (including China, which presumably hasn’t had much dietary coconut in the north) ranges from 6 to 10 percent, as opposed to 17 percent in the US (which is about average). (This paper(paywalled) has some different-looking numbers on a not very readable map.)
So it looks like there has been enough selection against ApoE4 in those areas that I shouldn’t feel safe eating lots of coconut fat. But I suspect that there are other relevant genes which make it hard to generalize to everyone with an ApoE4 gene.