Impact of a low-carbohydrate, high-fat diet on gut microbiota.

NPR recently reported on a study where the participants ate either a meat-based, fiber-free ketogenic diet or a vegetarian diet and had their gut microflora analyzed.  The low carb diet was much higher in fat, and as such, increased the prevalence of a microbe involved in fat digestion.  “Bilophila.”  The article focused on this one and cited a 2012 study where Bilophila was associated with intestinal inflammation… however, the ketogenic diet increased the levels of Bacteroides and decreased Firmicutes.  These are the two that brought the whole gut microbe-obesity connection into the spotlight.  The microbiome in obese mice is characterized by low Bacteriodetes and high Firmicutes. Fecal transplants from obese mice to lean mice causes them to gain weight.  Little is known about Bilophila relative to Bacteriodetes & Firmicutes, and I suspect the focus was on Bilophila because the authors wanted something negative to say about a meat-based, fiber-free ketogenic diet, and that 2012 mouse study suggested Bilophila could be their answer.





Obesity alters gut microbial ecology (Ley et al., 2005)
This study shows increased Firmicutes & decreased Bacteriodetes in genetically obese mice but not their lean siblings.  This is important because microbes are usually inherited from Mom and are common among littermates.  Apparently, genetic obesity overrides both.


An obesity-associated gut microbiome with increased capacity for energy harvest (Turnbaugh et al., 2006)
This one shows decreased Bacteriodetes & increased Firmicutes in obesity.  This is also the study that shows microbial transplantation from obese to lean mice causes weight gain (with no change in food intake).  Oh yeah, and by “microbial transplantation,” they mean wiping the poop of a fat mouse all over a skinny one.  Yes, that’s how they do it.


Human intestinal microbiota composition is associated with local and systemic inflammation in obesity (Verdam et al., 2013)
This study shows decreased Bacteriodetes & increased Firmicutes is associated with obesity and inflammation in humans.


Firmicutes associated with human obesity here, too.  Sorry, everyone’s favorite Lactobacillus acidophilus is a Firmicute.  Bifidobacteria, on the other hand, is not. #Bifidobacteria.



Diet rapidly and reproducibly alters the human gut microbiome (David et al., 2013)
The current study.  The animal-based, high fat, very low carb, fiber-free ketogenic diet induced a microbial shift that correlates with leanness in humans, and causes leanness in mice.  This is what should’ve made headlines, not Bilophilia.


Dietary information was buried in the supplemental files, but great data presentation:


Calories and macro’s show that the animal-based diet was indeed, ketogenic:


And if you still think 126 grams of protein is too many for ketosis, ffs:



The diet only lasted 4 days; that’s not a disadvantage because diet-induced microbial changes occur within 1-2 days and remain stable.




This study is interesting because it tested two widely differing diets.  Critics will say we can’t know what caused the changes because there were too many variables.  To them, I say this: those studies will come; that doesn’t make this one useless or uninteresting.  Also, I don’t care much for Firmicutes or Bacteriodetes; they’re obviously not the answer to obesity… the tone of the news reports and article itself just didn’t sit well with me.


calories proper


Become a Patron!




Be Sociable, Share!
  • Jim

    Interesting. I am all over the map eating high fat low carb but also having things like kale and bitter melon (raw) in an effort to feed the “good guys”. Would that be the best of both worlds or am I just confusing things in my tummy? Also, I wonder what long term use of Terbinafine will have on the microbiome? Would it be overall positive or negative? Thanks for your interesting work.

    • Hi Jim, thanks!

      Fermented foods like sauerkraut & kefir seem like a good (safe) way to support gut health, and they’re compatible with your low carbohydrate diet. Foods rich in insoluble fibres aren’t as good as selectively promoting the “good guys” as fermentable and soluble fibres… (but that definitely doesn’t mean they’re bad for you.)

      Given it’s side effect profile, Terb will impact your microbiome …and if it leads to GI problems, might want to consider a prebiotic like galactooligosaccharides (?)… just a thought.

      Thanks again 🙂

      • Dana

        Personally, if there were some way I could just get soluble fiber (preferably from food sources) and skip insoluble entirely, I would totally do it. Eating a lot of insoluble has roughly the effect on your intestines of eating Brillo pads. At least with soluble along for the ride it gets slowed down somewhat.

        And in turn I find soluble useful for feeding gut bacteria (hopefully the good kind) and also for slowing down gut transit on LCHF since animal fat seems to have the effect of speeding things up, which can be great if you were constipated previously to adopting said diet but not so hot if you weren’t! The soluble fiber and the fat seem to balance one another out pretty well in my experience. Pumpkin is a particularly favorite source for me.

        • I expect each soluble fibre will have a different impact on the microbiota, primarily because studies comparing purified prebiotics always seem to show very different effects.

          That’s an interesting point about animal fat & soluble fibre…Stefannson mentioned they would get constipated if they didn’t get enough animal fat.

          love pumpkin 🙂

          • Thomas Hemming Larsen

            I didn’t recall Stefanssono mentioning it but it would definitely fit my own experience. When I was eating the least, I would only go number 2 once each week.

          • I hear that a lot from people losing weight or not eating a lot of food. It also happened to the contestants on the TV show “The Biggest Loser” (I’ve never seen an episode, but spoke with one of their on-site MD’s about health problems they encountered).

          • johnnyv

            Eating less food makes you poop less? what wizardry is this?!

            Is this really a health problem?
            If you are not suffering from diarrhea or passing bricks and you do not have bloating or cramps you should be fine with a reduced frequency.

          • Not a health problem… if you exclude all the reasons why it could be a health problem 🙂 Reduced frequency is oftentimes accompanied by those symptoms – bloating & cramps – which can be a pretty big impediment to people’s daily lives. It’s the number 2 cause of missed work days (no pun intended!).

          • Thomas Hemming Larsen

            Haha, I would never think that bad about you Bill 🙂

        • Aeryn

          Acacia fiber is wonderful. All soluble, as far as I’m aware, makes stools soft, and does feed the gut as I have to moderate my intake to keep gas from getting too powerful or stools too soft.

      • Jack Kruse Nuts are a better choice for many who want to try the RS experience.

        • love almonds! Raw (and not Californian) as I suspect heating does bad things to the nutritional quality… probably more RS, too.

          I think buying them in their shells is the best way to guarantee they’ll be raw.

          • Bill

            Interestingly, wild almonds (along with most wild nuts) are extremely poisonous — full of cyanide. Thank goodness for domestication.

          • thanks, good to know!

            (actually, if it wasn’t for the Google image search I just did, I probably wouldn’t have even recognized a wild almond tree.)

          • Bill

            You can’t use Almonds as a substitute for RS. Almonds have only minuscule amounts of RS in them. You would have to eat an enormous amount of almonds just to get even close to the RS in Potato Starch.

            There’s nothing to fear with Potato Starch. Potato Starch is essentially pure RS. It’s 80% RS and 20% water — the water is trapped inside the RS granules and explode, like popcorn, when exposed to heat, which turns then turns it into a real carb.

            While nuts certainly have some benefits to them, I don’t think they could have possibly been a major part of the Paleolithic-era diet. The overwhelming majority of wild nuts are very poisonous. Some archaeologists have found evidence of very early nut cracking, but for all we know they were using the nuts to create poisons 🙂

          • Bill, I don’t use almonds as a substitute for RS. I just like them, regardless of their Paleo status 🙂

            The study Phil cited shows that cold raw potato starch is 54% RS & 27% digestible starch (I’m not sure where the 80% RS & 0% digestible starch came from).

        • I’ve written about this: “The day almonds became interesting.”

          • Thomas Hemming Larsen

            Great post! I definitely have to ‘try’ this. Since finely ground almonds have such a positive effect would almond flour also be viable? Are there studies on other types of nuts?
            As you mention it must have something to do with the fatty acid composition, I’m just thinking that other nuts would not be that different from almonds in that respect which would also give them a very positive effect on ‘gut health’.

          • Thanks! it’s one of my favorites 🙂

            There haven’t been any stories like this about other nuts (that I know of), but I wouldn’t be terribly surprised if there were.

          • Thomas Hemming Larsen

            I guess the same would go for different kinds of seeds. As you mention in a previous comment, the different types of fibre will impact the gut bacteria in different ways.

    • Anthony DiSante

      Richard Nikoley and crew over at Free The Animal have been doing lots of experimentation with resistant starch and its effect on the different types of gut bacteria, with some pretty amazing results:

  • Melissa

    I thought the NPR article was pretty nuanced for a mainstream publication and I’m sure other outlets will report it in a far more distorted way. Buried in that paper is something else interesting- the survival of bacteria from fermented meat into the gut. I wonder what the sources of their fermented meats was, but I suspect it might be an authentic enough source optimistically just because they bothered to use the names of the fermented meats rather than just “sausage.”. I wonder if that hasn’t been detected before because of differences between a truly high quality fermented meat product and cheap stuff.

    • yeah, that part of the study was pretty cool… the researchers really went the whole nine yards (aiming for a Nature paper?).

      Good point about the sources of fermented meats, they only said: “prosciutto & salami.” They also detected a few cheese microbes in the feces of the animal-diet group. And a weird spinach-associated pathogen in the plant-diet group.

      I think these haven’t “been detected before” because no one looked! those data are a nightmare to analyze.

    • Bill

      Some indigenous tribes have been known to bury their meat in the ground and return to it days later. The burial process ferments and pickles the meat. I’ll bet that’s been happening for millennia.

      The Japanese have also been known to pickle meat in a nuka-bed.

  • Wenchypoo

    After not having any fruit in my diet for over a year, I tried a slice of cantaloupe the other day, and O…M…G!! I could’ve added to the atmosphere of Mars from right here!! It took 3 days and handfuls of probiotics to get my old digestive system back.

    Poop pills? No thanks.

    Needless to say, I’m never eating fruit AGAIN no matter WHO screams at me about fiber!

  • rs711

    [slide 2/22] i love this: chimps ingest loads of ‘sugars with fibres’ (aka fruit) ===> assumption would be that they’re on a (relatively) high-carb diet…turns out, they’re digestive system is so different that it allows them to ferment the carbs/fibres into fat (butyrate mainly I’m assuming)….it really seems nature tries to extract as much as fat as possible from its environment from a purely ‘efficiency’ perspective.

    This – in contrast with the RS debate – makes me wonder why we’re so quick to frame any health issue in terms of either a low or high-carb perspective…maybe it says more about our general ignorance?

    Anyways, great blog post as always – keep ’em coming!

    • johnnyv

      Chimps have a decent sized small intestine so they should absorb simple carbohydrates easily, also protein and fat as they are omnivores that will eat what ever they can catch.
      They do indeed have a large large intestine compared to a human and so can derive a significantly larger proportion of scfa from bacterial digestion of fibre.

      Despite the SAD allowing humans to achieve big pot bellies our ribcage does not flare outwards like a chimp or gorilla to accommodate them by design.

      • I once read that a captive gorilla caught a bird and ate it. Omnivores, when they can get it, and make a lot of SCFAs with the rest.

        and… Ha! I hadn’t thought about their ribcages before. Humans just get bloated and uncomfortable when they eat too much fibre 🙁

    • Thanks!

      Yeah, I’ve heard that… and humans, too! (convert a lot of prebiotic oligosaccharides into short-chain fatty acids).

      • rs711

        I briefly described what I ate to Richard Nikoley yesterday – which was quite representative of my general diet – as I was surprised by his affirmation that “ketogenic diets are fucking moronic”. This is what he answered:



        Could you maybe enlighten me as to why I could possibly be delusional about my diet being LC/keto (or cyclic ketogenic)?

        • J. Stanton

          There is no positive correlation between the definitiveness of a person’s statements and the correctness or utility of those statements.

          On some days I’m tempted to assume the correlation is negative.


          • rs711

            J. Stanton – what surprise! Big fan of yours here 😀 …I loved the Gnoll Credo and funnily enough it’s the only non-science based reading I’ve done over the past year (I know, silly silly me!). Also, your approach to explaining satiety, satiation, hunger, fasting etc., is BY FAR the most grounded, clear and well supported [some might even go as far as calling it ‘good science’!]. Ok – I’ll get back to the subject at hand.

            I will start by posting RN’s reply.


          • Michael

            Interesting effect of RS:

            “… this study showed an improvement in insulin secretion following HAM-RS2, it did not demonstrate an improvement in insulin sensitivity…”


          • Higher insulin secretion with no improvement in insulin sensitivity?

            I think by definition that means insulin sensitivity declined. Perhaps I’m misinterpreting, but that is what it appears to suggest… rather directly.

          • Bill

            And yet, this n=1 suggests a dramatic increase in insulin sensitivity.


          • that’s fantastic!

          • Rs711

            I went through that study and was ‘confused’ by that? Feels like im committing an elemental mistake…am i? Surely if the insulin secretion pattern doesn’t change…?

          • the Bodinham study is an interesting one, that’s for sure. Supports the idea that RS improves glucose control by enhancing the incretin response, which improves insulin secretion. More about it here:

          • Isn’t that how George McGovern got the whole “low fat diet” thing rolling in the 70’s?

        • Your diet looks rather healthy. And low carb despite the inclusion of sweet potatoes & raisins. +1

  • “Oh yeah, and by “microbial transplantation,” they mean wiping the poop of a fat mouse all over a skinny one. Yes, that’s how they do it.”

    I laughed out loud so much, Bill explaining the scientific process to wooo at home.

    Yes again, thx for pointing out protein is not antiketogenic, and I suspect problems dietary controlled epileptic pts have with protein has less to do with ketone concentration and more toward excitatory properties of protein or insulin increases independent of ketone level. I use keto for mood and I observe over, and over, and over again that eating lean meat will keep my pee pink and ketones high but I experience symptoms as if I were not in ketosis. There is a certain state of calmness and quietness I feel when in a proper low protein keto; the gold standard is an overnight fast with low calorie intake the night before. I suspect that with epilepsy, seizure from higher protein isn’t about ketones as much as it might be that amino acids reduce inhibitory tone of brain independent of ketones, just like millions of other things such as sleep deprivation or stimulants, which are notorious seizure risks.

    Obesity is NOT a seizure disorder, or neurological condition.

    BTW, the only ketosis that matters for obesity, is fasting ketosis. Measuring ketones after a meal = whocurr. People on low cal low fat diets might even be in a trace ketosis after a fast. It’s not about dietary fat, or postprandial ketones…weight loss is about oxidizing body fat and insulin levels over 24hr period.

    • Ha! 😉

      Truth –> 45% carbs (mainly sugar), 35% protein, 20% fat = ketosis… because 525 kcal/d.

      P.S. yes, agree, epileptic children need to restrict protein on a ketogenic diet. Perhaps I should’ve included that as a “disclaimer” in my protein/ketosis blog post.

  • Rs711

    I will try a RS experiment as i now own a glucose/ketone meter, it’s likely ‘safe’ and i am without-Grokette at the moment so temporary flatulence is no dealer breaker.
    Research wise – i will start at the beginning: where does RS appear within our evolutionary history? How much? Was it significant? What current science supports/or doesn’t support supplemental RS? What plausible mechanisms are there etc.? Etc.

    The only thing I’m certain of is that these questions are nowhere close to fully answered and thus imply the bold claims are definitely mismatched.

    Die biting the throat!

  • Danny J Albers

    Hi Bill,

    Tonnes of confusion on the protein vs keto issue.

    There is a fair margin between simply producing ketones, and having your skeletal muscle use those ketones preferentially especially for things like distance running or cycling. There seems to be a certain blood level range required in terms of ketone concentration.

    During my zero carb period I certainly found it important to up fat, and reduce protein slightly on heavy training or distance running days.

    For weight loss though, just to lose weight, I am not sure the level of ketones matters or even their presence. Nothing requires you to be ketogenic or at least on the keto diet to lose weight, people lose weight a lot of ways.

  • WAHG

    Can you clarify the Table “Calories and macro’s show that the animal-based diet was indeed, ketogenic” please?

    Animal-based diet:
    Fat 137 g = 69% (% kcal)
    Protein 126 g = 30% (% kcal)

    Isn’t it 52% and 47% resp.? Or am I not understanding something about grams and % kcal?