How Rare My Heart Scan Improvement?

In 2009, I had a heart scan — a three-dimensional X-ray. The scan was used to calculate an Agatston score, a measure of arterial plaque. Higher scores mean a higher risk of heart attack. A few months after that, I discovered that butter improves how fast I do arithmetic.

Because butter was good for my brain, I started eating half a stick of butter (66 g) every day. Surely the butter was improving overall brain function. The effect of butter on the rest of my body I didn’t know. However, I thought it was highly unlikely that a food that greatly improves brain function is going to damage the rest of the body. The food you eat, after digestion, goes to the whole body (leaving aside the blood-brain barrier). Every part of the body must have been optimized to work well with the same food.

As I have posted earlier, I had a second heart scan, producing a second Agatston score, about a year after the first one. Amazingly, the second score was better (lower) than the first score. The woman in charge of the testing center said this was very rare — about 1 time in 100. The usual annual increase is about 20 percent.

Now I have gotten more information about the annual rate of change in Agatston scores. The graph above (thanks to Harry Rood) shows data from 40 people who listed their scores at the Track Your Plaque site. It is based on pairs of consecutive scores: it plots change versus level (average?). Because some people provided more than two scores, the data allowed 77 points to be plotted. My two scores were 38 (log 38 = 1.58) and 29 (log 29 = 1.46). So the decrease in log units was 0.12. If you look at the graph, you can see what an outlier this is — as I was told, it really is about 1 in 100.

Here we have the conjunction of two unusual things: 1. Eating half a stick of butter per day. Almost no one eats so much butter. 2. An extremely rare drop in the Agatston score over the same period. A principle of reasoning called Reichenbach’s Common Cause Principle says if two rare events might reflect cause and effect, they probably do. You can think of it like this: Lighting doesn’t strike twice in one place for two different reasons. Indeed, there is other evidence that high levels of saturated fat cause heart-scan improvement (even though this contradicts everything you’ve been told). Mozzafarian et al. (2004) found that in postmenopausal women, “a greater saturated fat intake is associated with less progression of coronary atherosclerosis.” So it is quite plausible that my butter intake improved my Agatston score.

22 Responses to “How Rare My Heart Scan Improvement?”

  1. Chris Says:

    Do you need to eat half a stick of butter a day to maintain the cognitive benefits? Or do you think that once a deficiency has been corrected you can reduce the dose?

  2. Seth Roberts Says:

    You need to eat the butter regularly to maintain the benefits. When the butter goes away, the brain gets worse.

  3. Tuck Says:

    Fascinating result. Dr. Davis might reconsider his anti-butter stance…

    Of course Peter at Hyperlipid has already explained how Dr. Davis mis-read this… (I’m a big fan of Dr. Davis, but no-one is perfect.)

  4. Wilhelm Says:

    Do you eat the butter mostly as it is or cooked/fried/heated ?

  5. Alex Chernavsky Says:

    I know that Eri Gentry’s “Buttermind” study suggested that coconut oil doesn’t have the same positive effect as butter. Aside from flaxseed oil, is there any other plant-based food that may have the same (or similar) beneficial effect as butter?

  6. Paul Winter Says:

    K2, found in butter. Dr. Davies has a couple of posts on K2 causing a reduction of the arterial plaque.

  7. Aaron Ashmann Says:

    Seth do you consume a lot of choline? I’m worry that if I eat a primarly vegetarian diet that if I add a lot of butter, I might not be injesting enough choline to break down the fats. I don’t consider taking a supplement to be adequate way of getting the nutrient– any ideas here?

  8. Jonathan Shewchuk Says:

    While I think saturated fat is probably good for you, I don’t see why you’re focusing on it for heart scans. My guess for the most likely hypothesis is that the Vitamin K2 in the butter (and maybe the Vitamins D and A too) is regulating your calcium so it goes into your bones instead of into your arteries.

    Check out these posts on K2 and heart disease:

    Of course, the butter-calcium relation could be a many-factored effect, but the fat-soluble vitamins have a lot of evidence behind them. It’s not clear to me whether the fat itself is important or just a carrier for the vitamins.

  9. Alex Says:

    It’s “Agatston.” I only mention it because it’s a search keyword. I love the interesting insights from your blog.

  10. Seth Roberts Says:

    Thanks for the correction.

  11. Vic Says:

    Seth, is the yogurt you eat also full fat?

  12. Seth Roberts Says:

    yes, the yogurt I eat is full fat.

  13. Aaron Says:

    Dear Seth –

    What do you think about the (rather conflicting) research regarding flaxseed consumption and prostate cancer? I assume you are not worried?

  14. Seth Roberts Says:

    Aaron, I haven’t looked into it much. The benefits of omega-3 are overwhelming. I will eventually compare different sources of omega-3 but haven’t yet.

  15. JohnN Says:

    There are probably more causation than correlation between your consumption of butter and improved brain functions. Have you considered, however, other factors such as K2 (mentioned by other posters), added sodium as electrolyte and brain running on ketone (not uniquely limited to butter)?
    PS: is there any positive effect on teeth and dental health as well? The butter fat can break up dental plaque and neutralize the acidic environment created by the bacteria.

  16. The Ancestral Health Symposium | Perfect Health Diet Says:

    […] On Saturday I started with Seth Roberts and Tucker Max. Seth’s talk was excellent – the gist can be found in recent blog posts, such as the one on butter reducing his coronary calcium score. […]

  17. Seth Roberts Says:

    JohnN, re correlation or causation. The Genomera experiment, based on my findings, was an ABA design, where A = no butter and B = butter. It found a difference in outcome between A and B supporting causation.

    Re K2. I haven’t noticed a difference in effect between cheap and expensive butters, which I would expect to differ in K2 content.

    Re added sodium. I got the effect with unsalted butter. I haven’t noticed a difference between salted and unsalted butter.

    Haven’t noticed any dental changes. My gums are pink due to flaxseed oil.

  18. Chris Says:

    With K being a fat soluble vitamin, I would guess that it is not responsible for the cognitive benefits since, as Seth wrote, you need to eat the butter regularly. I would think that the K would accumulate in the body and the benefits would be maintained for some time after due to the body mobilising its stores.

    Does cheese provide a similar effect? If not, the differences between cheese and butter might hold clues to the mechanism.

  19. harry Says:

    Hi Seth,

    I’d be a little cautious about concluding that eating a half-stick of butter every day will result in a significant CAC score reduction. Not that I think that the butter is detrimental, I eat a quarter of a stick every day myself, and lots of other saturated fat, a total of about 80 grams of it per day. My caution about concluding that saturated fat reduces CAC score is due to 2 things. First, the fact that my CAC score isn’t going down, and second, the large uncertainty in each CAC score measurement, especially at the lower values of total score. Your Agatston score was initially 38 and it changed to 28.8 a year later, if I understand what you have said. That’s great, but it’s a relatively low level of calcium as these things go.

    If we look at studies of reproducibility of CAC score where they have taken into account the score uncertainty as a function of score, we see considerable uncertainty. For example, the study “Serial Electron Beam CT Measurements of Coronary Artery Calcium: Has Your Patient’s Calcium Score Actually Changed?”, by Sevrukov, Bland, and Kondos, 2005, suggests that if the measured CAC score is 40, the 95% repeatability limits on that value fall between 8.8 and 71 agatston units. Similarly, the 95% repeatability limits on a score of 30 is between 3 and 57. This is based on a large number of scores repeated serially within a short time. And the other major study of CAC reproducability is even worse. See “Repeatability Limits for Measurement of Coronary Artery Calcified Plaque with Cardiac CT in the Multi-Ethnic Study of Atherosclerosis, by Chung, McClelland, Katz, Carr and Budoff, 2008;

    I think we need a lot more tests of the hypothesis that butter, or saturated fat reduces calcium score before we can conclude that the hypothesis is valid. I’ve been eating a high-fat, high-saturated fat diet for several years, and my CAC score continues to increase (although it may be slowing down, percentagewise).

  20. Seth Roberts Says:

    Harry, the graph suggests that a reduction like mine is quite rare — about 1 in 100. That probability calculation includes error of measurement. But I certainly agree that more tests of the hypothesis are needed. Yet it is still meaningful that my results were so different than predicted by conventional wisdom. Conventional wisdom says that butter causes heart disease. My results, to the extent they say anything, point in the opposite direction.

  21. Kelly Says:

    Thanks Seth…this “improvement” story by eating a half stick of butter a day was truly hilarious.

    I needed the laugh.

  22. Seth Roberts Says:

    Kelly, thank you for your sneer.