While recording a podcast with Ben Greenfield, I was discussing how the French, Tibetan yak herders and early American farmers all had a diet combining saturated fat (butter) with starch yet appeared to stay lean and healthy. Ben then suggested that there were studies showing that combining saturated fat with starch increases LPS. LPS is a bacterial toxin that can get into your bloodstream during digestion. It is generally seen as being “bad” as it can lead to inflammation1 and is typically higher in those who are diabetic.2
The answer that I gave Ben was that based on my knowledge of traditional cuisines, that doesn’t sound right but I don’t really know enough about LPS to refute it.
After the show I quickly discovered that a long term (12 week) dietary intervention trial had been conducted to determine whether or not this was true!3
Burying The Lead
Researchers randomized 75 subjects into four groups. One group (HSFA) ate a “high fat” diet (40% of calories), mostly from butter. Another group (HMUFA) ate the same amount of fat as olive oil. A third group (LFHCC – the carbohydrates were complex, they can be so complicated. The “complex” carb used was white bread and it was the same in all three groups) only ate 28% of calories as fat using both butter and olive oil. The last group (LFHCC n-3) ate the same as the third but supplemented with fish oil.
The four groups ate these diets for 12 weeks then were tested for fasting LPS. They ate a test meal that reflected the test diet they’d been eating.
Maybe Ben was right! Right there in the abstract it says:
They present this graph as proof. Clearly the butter group increased post-prandial LPS the most, although I’d argue the olive oil rise looks significant as well.
But what were the fasting values for the groups after 12 weeks on the test diets?
The high fat butter group has HALF the fasting LPS of the other groups. Even if there is a brief post-prandial spike, the butter group has by far the lowest overall exposure to LPS. Kind of a curious thing to leave out of the discussion, don’t you think? They ran some statistical analysis that is never explained, came up with a high p-value and therefore concluded that all of the groups are the same.
If you add the incremental post-prandial increase from the previous graph to this table of fasting values, you can see that after the meal – when butter is supposedly having it’s pro-inflammatory effect – LPS in the butter group would be 0.5, while the olive oil group would’ve been at 0.67. They never present total values, though, only baseline and the amount of change. It almost seems like they don’t want us to notice.
The only way their conclusions could be justified is if there was some massive error in the randomization of the groups. If the ~19 people in the butter group all just happen by random chance to have started out dramatically lower than the other groups. This is unlikely in a group of that size, but of course a standard study design would have been to randomize the LPS levels of the groups before the dietary trial. Yet they don’t report doing that and they don’t report the pre-diet intervention LPS levels of the groups. This is a curious omission. Overall, the study seems well done, but they didn’t randomize the groups as to the very thing they were studying? Either they are not very good or they don’t want to tell us.
You have to read the whole paper and not just the abstract!
I’m still no expert on LPS, but after reading this paper, I am going to stand my ground that the traditional French/American/Tibetan diet of starch and butter is probably healthful in regards to LPS exposure until proven otherwise.
- 1.Cani PD, Amar J, Iglesias MA, et al. Metabolic Endotoxemia Initiates Obesity and Insulin Resistance. Diabetes. Published online April 24, 2007:1761-1772. doi:10.2337/db06-1491
- 2.Pussinen PJ, Havulinna AS, Lehto M, Sundvall J, Salomaa V. Endotoxemia Is Associated With an Increased Risk of Incident Diabetes. Diabetes Care. Published online January 26, 2011:392-397. doi:10.2337/dc10-1676
- 3.López-Moreno J, García-Carpintero S, Jimenez-Lucena R, et al. Effect of Dietary Lipids on Endotoxemia Influences Postprandial Inflammatory Response. J Agric Food Chem. Published online August 22, 2017:7756-7763. doi:10.1021/acs.jafc.7b01909