Ketogenic Diets & Fatty Liver

Non-Alcoholic Fatty Liver Disease:  NAFLD.  Along with diabetes, a disease on the rise in the US and around the globe where the obesity epidemic is rearing it's ugly head.  Near as I can tell, Peter/Hyperlipid subscribes to a "damaged liver" hypothesis of obesity and diabetes.  That the liver gets damaged, throws your insulin and glucose outta whack, and then you get fat.  There can be no doubt that damage to the liver is an integral part of the metabolic syndrome(s) and the insulin resistance that underlies it.  So the question is, do we take it seriously in all contexts?  Or do we try to explain away inconvenient results when it is one's favored advocated diet that might possibly be a culprit?

Sadly, Peter seems bound and determined to spin ever more far-fetched mechanisms with which to explain away inconvenient results.  Having nothing to do with fatty liver, Peter found himself unable to backtrack on his definitive statement that it was fasting insulin levels that determined fat loss.  So bad was the evidence to counter this that he had to resort to manipulating study data to make his point.    It seems where high fat rodent studies are concerned, Peter was hell bent on blaming transfatty acids.  Now I'm no fan of the transfat, and they are indeed evil doers in any body, but they can't be blamed for everything. Indeed one just needs to look around to find high fat rodent diets that are not high in transfats.  Enter the other bad guy by which we can dismiss inconvenient results ... veggie oil PUFA.   True, many high fat rodent diets are very high PUFA ... but there are many that are based on lard.  You know, the stuff us moist eyed female types should wallow in to enter the glories of Slimville via the Ketogenic Highway?!


Well, now lard is not a good fat source to use for rodent studies, because .... drum roll ... it's full of too much PUFA!!  As Peter tells his readers:

The paper (and the F3666 pdf) specifies 16% of PUFA in the fat, which seems rather low considering that USA produced lard (the main ingredient) appears to have 32% PUFA (almost all omega 6) but, even at only 16%, these PUFA are 16% of 95% of your total calories. That's an awful lot of PUFA...

Firstly, the 16% PUFA comes from the manufacturer of the rodent diet used.  While there may be some slight differences depending on the source of the lard, to even suggest that the PUFA content of USA lard (and thus the chow) is twice that stated in the study is outrageously misleading.  It seems that Peter got his "facts" on all USA produced lard from a communique from a rodent chow manufacturer to Chris Masterjohn.   Apparently this one particular obesity-inducing high fat diet contained roughly twice the stated PUFA.  But I find it hard to believe that nutrition data bases are so far off.  For example:  Nutritiondata.com puts PUFA at 11% in lard,   10% PUFA from this source.  I suppose we can count the Canadians in on the pork-PUFA ruse, as using the fattiest products to minimize the contribution of rounding error, bacon comes in at 12.5% PUFA, and back ribs at  13.2%.  Surely Peter is familiar with the marketing of pork as "the other white meat" here in the US, and since PUFA are considered more beneficial by our misguided government one would think that "pro pig" organizations would exaggerate rather than downplay the PUFA content ... right?  According to the US Pork Center, pig fat exceeding 15% PUFA becomes soft fat, and "soft fat and pork bellies cause carcass handling and fabrication difficulties; reduced bacon yields; oily, almost opaque-appearing, unattractive products; reduced shelf-life; and, more importantly, discrimination by domestic consumers and export partners."  

Whatever the reason that the fatty acid content in Research Diets D12492 is so high in PUFA, however, simply doesn't translate to the diet in the study where the mice developed fatty livers.  Some links:

• The study:  Hepatic steatosis, inflammation, and ER stress in mice maintained long term on a very low-carbohydrate ketogenic diet  (email me for the full text if you're interested)
• My blog posts referring to the study:  Bloggo Science ~ Firko-ized and Mighty Metabolism Mouse?
• The diet:  Bio Serv Product# F3666 - Ketogenic Diet, AIN-76A Modified, High Fat, Paste

According to Peter:

The paper (and the F3666 pdf) specifies 16% of PUFA in the fat, which seems rather low considering that USA produced lard (the main ingredient) appears to have 32% PUFA (almost all omega 6) but, even at only 16%, these PUFA are 16% of 95% of your total calories. That's an awful lot of PUFA...

The ingredients of the diet are:  Lard, Butter, Corn Oil, Casein, Cellulose, Mineral Mix, Vitamin Mix, Dextrose.  The paper lists the diet as predominantly lard and butter.  Now c'mon here.  The order of listing of ingredients only means there's more lard than butter than corn oil -- for all we know there's 50% lard, 49% butter and 1% corn oil in this diet.  But here's the key folks.  Rather than rank speculation about what might be, why not just report what IS.  On the product page, we find that the product is "Nutritionally assayed" ... in the PDF they state content may vary.  Bottom line, feel free to call the 800 number if you're concerned over the PUFA content in this study, but given that low PUFA butter is part of the diet, I'll go with the 16% PUFA reported unless and until Bio Serv reports differently.  Given that the product is assayed (in other words directly analyzed for nutritional content, not just calculated from ingredients), I doubt highly that it differs to the tune of a two-fold error.

Moving on ... So what if PUFA is high?  Peter chooses this study to implicate PUFA in the development of fatty liver disease.  The study is in rats who were housed 4-to-a-cage and given either 7% fat standard chow, or free access to a high-fat chow OR standard chow.  The high fat diets consumed were either a coconut-fat diet or a butterfat diet that worked out to be 45 and 42% fat respectively -- by caloric percent.  All rats ate the standard chow for 3 weeks followed by 14 weeks on their diet.  Now a fourth diet was tested, a methionine choline deficient diet, MCD.  For whatever reason, this was given to 9 week old rats for 6 weeks duration.  While butter produced twice the liver triglycerides vs. the standard or coconut diet, the MCD diet produced 22X the liver triglycerides!   The two high fat diets (well within SAD range) produced mild macrovacuolar steatosis (5% of hepatocytes) vs. the MCD that produced severe steatosis affecting 60% of hepatocytes.   What was the MCD diet composed of?   I can't find the diet specs, but the discussion in this diet summarizes various other studies of NAFLD.  Here's that table: 

Pretty definitive that it's the PUFA causing fatty liver?  I'd say that mild steatosis on 42% and even 59% PUFA of total fat -- on 45% total fat diets (by weight - e.g. higher percent by calories) is hardly compelling.  Indeed in this study, the 45% butter diet was producing mild steatosis in 14 weeks.   This tells me that if I'm going to eat a high fat diet, there may be some lessons to be learned as to the sourcing of those fats, or perhaps there might just be lessons to be learned as to the consumption of a high fat diet at all.

I'm making a mental note to come back to this study as there are some other interesting things to be discussed.  For example the CO group ate a lot more calories than the butter group and both ate more than the standard chow group.  Body mass at 14 weeks didn't differ much, but the high fat rats became ... well ... highly fat rats, especially the buttery ones (one fat pad was 93% larger for the butterballs than the controls).

Moving on, I was looking for more information on ketogenic diets and fatty acid compositions of such diets, as I feel they are more applicable.  None of the diets in the previous study were low carb or anywhere near ketogenic.  One study I found:  Lipid and Fatty Acid Profiles in Rats Consuming Different High-Fat Ketogenic Diets.  (I have the full text of this one for anyone interested).

In this study, they were specifically looking at differing ketogenic diets where ketones were assessed along with seizure prevention efficacy.   They used very high fat diets -- 80% by weight fat -- of different compositions.   The fatty acid compositions of these diets are shown at right. 

Care to guess what happened?  Well:

"Liver triglycerides in the BUTTER and MIXTURE groups were about 10 times higher than in the other groups."

Note the following caveat:

Rats are an important research tool for modeling seizure mechanisms and treatments, but it may be difficult to extrapolate between rats and humans. This caution applies not only to understanding the changes in cholesterol metabolism but also to the significant rise in liver triglycerides in the BUTTER and MIXTURE groups, changes that were not clearly obvious in plasma lipids (Table 3). If children on the ketogenic diet have such an elevation in liver triglycerides over 2–3 yr, compromised liver function would become a concern.

I've not found any studies on children on keto diets vis a vis fatty liver.  Perhaps this is because the diet is usually relatively short term and tends to be calorie controlled as well.  Even if fatty liver were to develop it may not be a major concern in this context.  But so ... here we have a butter diet causing huge increases in hepatic triglyceride content.  Unfortunately we have no histology, but this increase in triglyceride content would usually be associated with steatosis.     I have to make another mental note to return to this study as it has some interesting other findings as well with respect to ketone levels, tissue arachodonic acid concentrations and body fat fatty acid compositions on the various diets.  More for another day ...

So Peter leaves his readers with the following advice:

So there is a message: If you are going to eat a very high fat diet in the long term, you should pay serious attention to PUFA creep. Having cirrhosis is not likely to help you make old bones in good shape. Perhaps go easy on the commercial mayonnaise and just spread some butter on your cheese... Actually I tend to add butter to my beef if it's too lean (only happens when eating out, no way we would choose lean beef to cook at home!)...

He seems to acknowledge that fatty liver on ketogenic diets is a pretty real possibility in humans after all.  Certainly it's worthy of consideration, which is all I'm really saying on this subject as well.  But perhaps it's the butter in the diet, not the lard, that caused the steatosis in our mighty metabolism mice.  How many low carbers eat tons of butter and heavy whipping cream and cheese -- all similar dairy fats lacking in PUFA.  Throw in that unreliable PUFA-ridden American pork and your low carb diet is a veritable fatty liver minefield!  The fact of the matter is, that if you're eating a truly ketogenic diet -- especially at maintenance -- you're eating a lot of PUFA -- if that's your concern -- and likely a lot of butterfat.  Heck, Mark Sisson hits the 16% PUFA range on one version of his sample menu!

So that's all I'm saying with these posts ... it is a valid concern that can't just be explained away by exaggerated speculation about the fat sources in rodent chows.  When I said "Do I think most humans on VLC diets have to worry?  Probably not ..." I meant it.  Because most humans eating a low carb diet for weight loss (a) aren't eating more calories and fat than they were before, and (b) aren't eating anywhere near zero carb 5% protein KD diets.  And yet still, when I read more and more on rodent studies of fatty liver, my beloved butter isn't smelling like such the rose any more....  Certainly the last study shows that butter is very good at packing fat away in the liver, while the high PUFA flaxeed oil?  Not so much.