From Meat to Microbes to Main street: is it time to trade in your George Foreman Grill?

If scientists keep publishing the results of their work in journals, we’re going to run out of stuff to eat. The latest nutritional no-no literally has meat-eaters on the ropes following a startling article published in Nature Medicine that draws a possible link between a nutrient (carnitine) in red meat and cardiovascular disease. But what makes this interesting and so different from past debates – like how saturated fat and meat was once thought to hasten your arrival at the pearly gates – is the role of gut microbes as an intermediary in this tangled web of cardiac arrest.

cows2The story goes something like this: Your run-of-the-mill 8-ounce steak contains ~180mg of the nutrient carnitine (click here to learn how to pronounce). In and of itself, carnitine is not a bad thing as your body produces it naturally and it’s used everyday in the transport of fatty acids into your mitochondria. However, dietary carnitine from a slab of red meat can be gobbled up (metabolized) by your resident gut bacteria and converted to something called TMA (trimethylamine). So far so good until it diffuses and makes it’s way from your gut into serum (blood). Since TMA is a gas at room and body temperature, translocating across the gut barrier is not a problem. Once in your blood, TMA makes its way to your liver where it is oxidized into TMAO (trimethylamine-N-oxide). From this point, the researchers argue that elevated levels of TMAO muddy up some mechanisms including cholesterol transport, which accelerate atherosclerosis. In short, not good news for red meat lovers – or so it might seem.

This is a complicated paper with a lot of moving parts that are worth (re)considering (explored below). But the most interesting finding is the role of gut bacteria in this process and the data presented in the paper suggests that vegans and vegetarians have a decreased capacity to generate TMA from dietary carnitine than do omnivores – due to the fact they harbor different gut bacteria. But the most striking finding in this bacterial connection – at least for me and the one that almost caused me to fall out of my chair when I read – was the role of one group of bacteria known as Prevotella. The researchers discovered that regardless whether you were a vegan/vegetarian or an omnivore, higher levels of Prevotella in your gut correlated with higher levels of TMAO in your blood. Using gut microbial data from three individuals recently analyzed as part of the American Gut project – one of which is myself – I will touch upon the irony of the argument that carnivorous habits may lead to accelerated atherosclerosis in the context of Prevotella. But first, lets look at the Nature Medicine paper.

To demonstrate that gut bacteria are necessary to convert carnitine from red meat to TMA in the gut, which if you remember, is then absorbed into the blood and makes its way to the liver where it is further converted into TMAO, researchers gathered together five omnivores and fed them each an 8-ounce sirloin steak (which contains ~180mg of carnitine) and a tablet of an additional 250mg of carnitine (d3-carnitine) tagged with a heavy isotope which allowed the researchers to follow this particular dose through the body. So, we have regular (dietary) carnitine from the red meat and an isotope labeled d3-carnitine (note the d3 denotes isotope labeled). The researchers then drew blood after the meal over a period of 24 hours and as expected, the levels of d3-TMAO went up over time, indicating that the gut bacteria did their thing and metabolized the isotopic d3-carnitine into d3-TMA that was absorbed and converted to d3-TMAO by the liver. Interestingly, while the d3-carnintine from the tablet ended up as d3-TMAO in serum, the native or natural carnitine and TMAO from the steak only showed “modest” concentrations in serum according to the researchers. In other words, the impact of the steak was semi-uneventful.

To further demonstrate the contribution of gut bacteria to the carnitine>TMA>TMAO development, the researchers put the same five omnivores who wolfed down the sirloin steak with the 250mg carnitine chaser on broad-spectrum antibiotics for a week (metronidazole 500 mg and ciprofloxacin 500 mg – twice daily) to suppress their gut bacteria – as broad-spectrum antibiotics have a tendency to do.  After taking antibiotics twice a day for a week, the 5 omnivores were called back in and given another 8-ounce sirloin steak and 250mg d3-carnitine tablet. However, unlike during the first visit, the week of antibiotics caused a “near complete suppression” of TMAO in blood or urine. On top of that, no d3-TMAO was noted either. In other words, no detectable levels of TMAO or d3-TMAO were formed suggesting that whatever bacteria were present before that were endowed with the genes to metabolize carnitine from meat or the tablet into TMA, were gone.

After this second meal at the end of the week of antibiotics, the researchers discontinued the antibiotics and sent the omnivores home. After a few weeks had passed and enough time had presumably elapsed for their gut microbiota to recover from the week-of-antibiotic-hell, the omnivores were called back in for another steak and 250 mg tablet. This time, the gut bacteria that were capable of metabolizing carnitine into TMA in the gut had reappeared as TMAO and d3-TMAO came back with a vengeance in blood and urine. Taking together, all three steak dinners and tablet chasers demonstrate the “TMAO production from dietary carnitine in humans is dependent on intestinal microbiota.”

At this point you’re probably wondering which bacteria were depleted during the week of antibiotics: who were the nasty culprits responsible for creating TMA, but were metabolically shut down during a week of broad-spectrum antibiotics only to re-bloom after the fog of antibiotic war had lifted? Would seem simple enough – they were there during the first meal, wiped out during the week of antibiotics, then reappeared after several weeks of recovery. As amazing as it might sound, it appears the researchers DID NOT collect stool samples during this phase of the project so have no idea how the microbial composition had shifted before, during, and after antibiotics. Oh my – how unfortunate. Lets move on.

At this point, the researchers noted that TMAO (native and d3-labeled) varied among individuals following ingestion of carnitine. More to the point, a post-hoc nutritional survey revealed that red meat consumption might enhance ones ability to generate TMAO from carnitine. So it was time to break out the George Forman Grill again, get some tablets of carnitine ready and invite a vegan to dinner.

Kudos to the researchers as they were able to get a long-term (>5 years) vegan to scarf down an 8-ounce sirloin steak and the additional 250mg of d3-carnitine. The figure below plots what happened next. Over the course of 24 hours they took several blood samples from the vegan. As you can see, no appreciable amounts of TMAO showed up in the blood as a result of ingesting the steak (left panel) or the isotope-laden d3-TMAO (right panel) for that matter. Keep in mind this is a sample size of exactly ONE person. For comparison, the researchers also plot a “single representative omnivore” that self-reported eating meat on a daily basis. Though the researchers do not specify, the assumption is that the data plotted below for this single “representative omnivore” is from one of the five omnivores that participated in the previous five omnivore study discussed above.

Screen Shot 2013-04-12 at 4.48.21 PMThat said, among the five omnivores to choose from, it’s not clear which omnivore they selected to plot against the vegan: was it the one with the largest concentrations of TMAO, the one with the least amount, or was it one from the middle of the pack? If it were me, I would probably select the one with the greatest concentrations of TMAO to plot against the vegan – for dramatic effect! You see some of this dramatic effect in the plot on the right (d3-TMAO). You will notice that the panel on the left (TMAO) has vertical axis ticks set at 0, 4, and 6 – clearly showing a spike in the level of TMAO for the omnivore. However, the panel on the right (d3-TMAO) has the ticks on the vertical axis set at 0, 0.125, and 0.250 – to create a pretty dramatic effect for one that really didn’t exist. I have taken the liberty and re-plotted (red circle and line) the concentrations of d3-TMAO for the omnivore it the researchers had used the same vertical scale as the panel on left. This, of course, is much less dramatic and clearly shows that in a sample size of exactly ONE for each group (omnivore vs vegan), the omnivore showed elevated levels of native TMAO but very nominal increase of d3-TMAO from the isotope-laden pill. In other words, the gut microbiota from both individuals failed to metabolize the 250 mg of d3-carnitine into meaningful levels of d3-TMAO in blood.

As a side note, it’s also not clear if the TMAO levels for the “represented omnivore” is from baseline (pre antibiotics) or TMAO levels measured during the third visit – the one following the several weeks of wash out to allow the gut microbiota of the five individuals to recover. I’m not being nit picky here, but this does matter (a lot). It matters as the researchers note in Supplemental Data that a “representative omnivore” significantly greater levels of TMAO in their blood after the weeks of washout from the antibiotics over baseline (we don’t know about the others). The assumption is that gut microbiota of all five individuals had returned to pre-antibiotic diversity and abundance. But we don’t know for sure, as previous studies on the antibiotic impact on the gut microbiome have shown that it can take weeks or even months for the gut microbiome to recover and it often settles back in a different composition all together (altered state). All we know is that our five omnivores – based on a representative omnivore – had bounced back with a greater metabolic ability to yield TMA from dietary carnitine.

In either case, would be nice to know how representative the omnivore really is and whether or not we are seeing the TMAO levels from a highly perturbed gut microbiota following antibiotic treatment. Increased gut permeability might be contributing to elevated TMAO levels as well. This, taken together with the sample size of ONE for each and the fact that no appreciable amount of d3-TMAO was noted among either (only TMAO for the omnivore), I think the jury is still out on this one. Fascinating nonetheless, but more data is needed. Sample of ONE does not allow for any statistical comparisons. Again, unfortunate – lets move on.

Now the study gets really interesting. This time they gathered together 26 vegans and vegetarians and 51 omnivores (note in the text of the article the researchers say 23 vegans and vegetarians but show 26 in the plot below – I will go with the 26). After fasting overnight for 12 hours, the researchers found that baseline levels of TMAO in the morning was significantly lower in the vegans and vegetarians compared to the omnivores (see figure left panel). However, there is considerable overlap in the box plots – that is, a considerable range. It appears there are vegans and vegetarians that have just as high or higher levels of TMAO than some omnivores. Given there are less vegans and vegetarians in the sample (26 vs 51), would increasing the sample of vegans and vegetarians to the same as the omnivores (n=51) change this plot in anyway – possibly capturing more vegan/vegetarian variability?

From the 26 vegans and vegetarians and 51 omnivores the researchers picked five from each group and gave them each the 250mg d3-carnitine tablet (no steak this time) and measured the d3-TMAO in the blood at several time points over 24 hours (see plot). As can be seen, the d3-TMAO levels went up in the omnivores but not the vegans and vegetarians. However, it’s striking to see the variability in the omnivores – ranging from near 0 to almost 30. Said differently, a SINGLE omnivore among the five can explain much of the elevated levels show in the plot. Also, since this was a subset of the larger group of 26 vegans and vegetarians and 51 omnivores, we do not know if the researchers selected omnivores that showed higher levels of baseline (fasting) TMAO. Since they had to pick someone, would be nice to know the selection criteria: did they pick omnivores with low, medium or higher levels of baseline TMAO levels? In either case, following the d3-carnitine challenge omnivores showed an increase as a group, but with some less so than others.

Screen Shot 2013-04-15 at 12.19.12 PM

To explore the possibility that plasma TMAO levels might be associated with specific gut microbial taxa, the researchers collected stool samples from 23 of the 26 vegans and vegetarians and 30 of the 51 omnivores (at baseline). Sequencing the gene encoding for bacterial 16S rRNA of the stool samples, the researchers found some interesting patterns. As diet can significantly shape the composition of your gut microbiota – and differences between vegetarians have been noted in previous studies – the resulting microbial differences were not completely unexpected. However, one particular group of bacteria stood out – Prevotella. When the researchers grouped all 53 people analyzed into so-called enterotypes – regardless of diet – they found that individuals that had gut microbiota enriched with Prevotella, had higher levels of TMAO in their blood. Conversely, those with gut microbiota enriched with Bacteroides showed much lower levels of TMAO. While the concept of enterotypes is a matter of debate, Prevotella-enrichment is still evident for those individuals with higher Screen Shot 2013-04-15 at 12.20.21 PMlevels of TMAO. Interestingly, previous studies have suggested that enriched levels of Prevotella are associated with a higher carbohydrate diet – that is, less meat. However, three of the four individuals (see plot) in the Prevotella-enriched group were meat eaters, and the fourth was a vegan/vegetarian.  On the flipside, the 49 that clustered together – more Bacteroides, less Prevotella – was a mix of omnivores and vegans and vegetarians. These findings suggest that previous dietary habits can drive microbial composition in the gut that in turn effect the ability of your microbial ability to synthesize TMA/TMAO from dietary carnitine.

The potential role of Prevotella as a key player in this saga is made more interesting when you consider the results of feeding d3-carnintine to mice. In a series of experiments, the researchers spiked the chow of a set of mice and compared the gut microbiota to mice receiving the same chow but without the d3-carnitine. You guessed it, the mice receiving the d3-carnitine had significantly higher levels of Prevotella and unclassified Prevotellaceae. Since the chow was the same between the two groups of mice, the carnitine seems to have the unique ability to nudge along the growth of Prevotella as well. Or more specifically, Prevotella spp. appear to harbor the genes necessary to synthesize carnitine. Though the microbiota between mice and humans are not directly comparable, it is interesting.

Quick side note: In attempt to demonstrate that carnitine is linked to heart disease, the researchers fed two groups of mice normal chose but spiked the water of one group with carnitine, which resulted in a doubling of atherosclerotic plaque and “doubling of disease burden” in the carnitine group at the end of 10 weeks. This part of the paper got a lot of attention from the popular press and was held up – along with other pieces in the Nature Medicine paper – as proof that dietary carnitine from red meat can lead to accelerated atherosclerosis. However, most failed to mention the mice used in the study (Apoe -/- mice) are genetically engineered to be prone to atherosclerosis. On top of that, the 1.3% of carnitine in the drinking water is the equivalent to a human eating  ~1,000 steaks a day.[i] You can decide for yourself if you think mice that are genetically engineered to be inclined to atherosclerosis that are fed the human equivalent of an entire cow every three days for 10 weeks is proof that carnitine from red meat accelerates atherosclerosis in humans.

The researchers also investigated the possible relationship of carnitine and cardiovascular disease in a cohort of 2,595 men and women undergoing elective cardiac evaluation. In short, they found an association between TMAO concentrations in serum and “cardiovascular event risks.” They found the TMAO-cardiovascular risk the highest in the fourth quartile – or, 650 of the 2,595 individuals. Average age of this group/quartile was 61, 80% were male, were overweight to obese on average, one out of three had diabetes, 75% were hypertensive, a whopping 77% were smokers, numerous medications like statins and beta-blockers were the norm, and every biomarker of a crappy diet was elevated, as well as biomarkers of inflammation. In short, these folks were a mess.

It’s important to note the TMAO association was just that, an association – among many. At this point it might be useful to mention that even though the researchers are focused on red meat, an amazing assortment of foods can generate levels of TMAO higher than that reported for red meat. To name a few: potatoes, peas, peanuts, eggs, mushrooms, bread, squid, prawn, crab, halibut, cod, herring, tuna and an assortment of other fishy things. So, who’s to say that red meat is responsible for the higher concentrations of TMAO in the sickest quartile of these 2,595 people? If a TMAO association is the issue, then we might want to start issuing warnings to the general public that omega-rich fish are now off the menu, as well as carrots and peas. Remember, correlation does not equal causation.

The research in the Nature Medicine paper is important, as it clearly shows that gut bacteria are needed in the conversion of carnitine to TMA/TMAO. This was demonstrated when antibiotic treated humans and mice were unable to produce TMAO after antibiotics were taken for a week but when they were discontinued the ability was restored – even enhanced (note I did not discuss the antibiotic treatment in mice as it showed more or less what was seen in the human experiment). Unfortunately, the researchers did not collect stool samples so were unable to determine which bacteria were suppressed during the antibiotic treatment. But, the role of bacteria in TMAO creation and TMAO as a risk factor or cause of atherosclerosis is a separate issue. If higher concentrations of TMAO (from unknown sources) are associated with an elevated risk of heart disease in a terribly sick group of predominately male adults, and mice genetically at risk for heart disease fed the human equivalent of a cow every three days is proof that TMAO causes atherosclerosis in the rest of us, then we might have a smoking gun. But I’m not sure. I will leave that up to the statisticians and biochemistry propeller heads to sort out.

If we assume for the moment that TMAO can accelerate atherosclerosis, then the observation that enriched levels of Prevotella are associated with higher levels in of TMAO become more interesting – especially as enrichment of this group of bacteria seems to occur in vegans and vegetarians as well as omnivores. So what causes enrichment of Prevotella? In a paper published in 2010, Italian researchers found that rural African kids in Burkina Faso had super high levels of Prevotella, while a similar aged group of kids in suburban Italy had none.

The striking dietary difference between the kids in Burkina Faso and the Italian kids was whole grain consumption. As with many rural African communities, grains – in this case millet and sorghum – are processed manually by stone grinding. Minus any winnowing afterwards, these whole grains made up >50% of the daily calories in this village. In addition, they did eat some legumes (black-eyed peas), some mango fruit, some butter and very limited amount of vegetables. All in all, this is a very monotonous diet dominated by whole grains and very little diversity of anything else. In contrast, the Italian kids got ~25% of their daily calories from highly processed bread, biscuits, pasta, and rice. The Italian kids also consumed legumes, a range of vegetables, milk, chicken, beef, fish, code, sole, eggs, extra-virgin olive oil, butter, yogurt, cheese, snacks, and a diversity of fruits. All in all, a much more diverse diet, though apparently devoid of whole grains but high in processed foods.

As mentioned, the Italian kids had no traceable amounts of Prevotella in their stool samples. Below is a figure showing seven of the kids from Burkina Faso plotted against me and two other individuals (Mr. A & B) recently sequenced as part of the American Gut project. While the sequencing data between American Gut and the African study are not directly comparable (different lab protocols etc), they are informative for this comparison.

While all of the kids from Burkina Faso showed very high levels of Prevotella (~>50%), the seven in the graph are the highest. In fact, the Prevotella levels among these Burkina Faso kids are among some of the highest reported for a cohort anywhere in the world. On left you see my Prevotella levels are 1.5%, Mr. A is at 0% and Mr. B at 32%. Interestingly, neither myself or Mr. A eating much bread, pasta or any other grain-based foods. In my house, we get an artisan loaf of bread every few weeks to make sandwiches, occasionally eat bread when dining out, but can’t think of the last time I had pasta at home. Likewise, after chatting with Mr. A about his diet, he avoids all processed foods and eats a lot of fish – occasionally eating beef. In short, he follows a Paleo diet. I, too, follow more or less a Paleo diet but not for ancestral reasons per se – but more due to the fact I have a type 1 diabetic daughter so I am hyper aware to the insulin spiking effects of processed foods. But I do love the taste and smell of freshly baked bread!

Screen Shot 2013-04-15 at 9.19.46 AM

As for Mr. B, an email exchange reveals a very diverse diet of plants and animals, but a taste for whole wheat and whole grain foods in general – including whole oats for breakfast. Mr. B cooks often with whole wheat and whole grain ingredients. All three of us are average BMI and physically active.

In the Burkina Faso paper, the researchers attributed the high levels of Prevotella to grain-based carbohydrates and specifically dietary fiber intake. However, while I personally do not eat grains on a regular basis, I do eat an extraordinary diversity of plants and thus have high fiber intake – yet low to no Prevotella. I would say some weeks I average 40-80g a day of dietary fiber from a range of foods. I go out of my way to try and consume a diversity of dietary fiber sources. Mr. A also eats lots of dietary fiber, but not from grain sources. That said, it doesn’t appear that that dietary fiber in general is driving enrichment of Prevotella. If it were, both myself and Mr. A would have higher levels of Prevotella. So, it appears some how specific to the starch in grains or the dietary fiber in grains.

Prevotella, a gram-negative bacteria, are a common member of rumen and also present in high numbers in pigs, poultry and evidently some humans. A recent genome analysis of two common Prevotella (P. ruminicola and P. bryantii) reveals they posses an extensive repertoire of genes targeted towards the degradation of non-cellulosic polysaccharides such as hemicelluose and pectin – which is present in the cell walls of grasses and cereals (a.k.a. grains). While they can also utilize starches and simple sugars, they have carved out a specific ecological niche when it comes to chomping down on the cell wall (think bran) of grains.

Interestingly, in a study of 51 obese Finnish adults fed either a fiber-rich rye bread (whole grain) or a highly refined wheat bread for 12 weeks, those consuming the highly processed bread (no whole grain) showed a 37% decrease in the phylum Bacteroidetes, which includes genus Prevotella. Suggesting further that something particular to the whole grain – presumably the bran in the intact grain – is potentially modulating levels of Prevotella. However, another study by the folks at the University of Nebraska didn’t show much movement (up or down) in the abundance of Prevotella in 28 healthy individuals eating either whole-grain barley or brown rice. This same research group also fed resistant starch (RS2 and RS4) in the form of crackers to ten people and saw nothing exciting with the abundance of Prevotella – suggesting at least in this small study, resistant starch (like that found in grains) effects levels of Prevotella.

It’s also worth noting that we see higher levels (15-30% of bacteria) of Prevotella among San Bushmen in Namibia who have shifted from hunting and gathering to more westernized foods, such as porridge made from goats milk and whole-grain maize meal (unpublished data from our work in Africa). Colleagues working among Amerindians in South America have noted elevated levels of Prevotella among more traditional groups eating lots of cassava. The contribution of other agricultural products is less clear, but overall, suggests some interesting patterns with this widely used root food.

I reached out to the researchers who published the Nature Medicine paper and asked if they might share the taxonomic summaries of the stool samples from the individuals in the study. They were kind enough to do so. Also known as OTU table (Operational Taxonomic Units), the table breaks down the relative abundance of bacteria identified to the genus level. The OTU table reveals an abundance of Prevotella for the Prevotella-enriched group (enterotype discussed above) that included a vegan and three omnivores ranged from 16-54%. The highest abundance recorded (54%) was for the vegan. At 54%, that puts this particular individual in ballpark for the kids in Burkina Faso, and as for the rest, at least within the range of our whole-grain-pasta-bread eating Mr. B.

So taken together, enriched levels of Prevotella appear to be associated with intake of whole-grain products, but not necessarily associated with high fiber intake from no-grain plant sources (otherwise Paleo-like eaters like myself and Mr. A would have higher levels). Interestingly, the OUT table from the researchers revealed that the vast majority of vegans and vegetarians in the study had low to no levels of Prevotella. I’m going to go out on a limb here and suggest that if vegans and vegetarians are anything, they are starch eaters. If starch that escapes digestion in the upper GI (resistant starch from amylose and amylopectin in grains) were driving Prevotella enrichment, then we would expect higher levels of Prevotella across this entire veggie cohort. But we don’t see this. This leaves the possible interpretation that the vegans and vegetarians that didn’t show elevated levels of Prevotella are not heavy whole-grain eaters.

One last thing of interest with the kids from Burkina Faso. In addition to high levels of Prevotella, they also exhibited high levels of a genus known as Xylanibacter (which was absent in our biscuit and pasta eating Italian kids). As the xylan in Xylanibacter implies, its especially adept at degrading xylan found in plant wall cells – like those found in whole grains. Below is the levels of Prevotella (bottom axis) plotted against the levels of Xylanibacter in the subjects in the Nature Medicine paper along with some addition data from another study. While the overall abundance of Xylanibacter is low compared to the African kids, the plant cell wall metabolizer Xylanibacter goes up as the abundance of Prevotella goes up. Taken together with the whole-grain-based diet of the Burkina Faso kids, the co-occurrence of these genera might have something to do with whole-grain consumption.

Screen Shot 2013-04-15 at 11.16.32 AM

To wrap it up, I’m not sure what to make of the carnitine-induced atherosclerosis in the heart attack prone mice fed the equivalent of mountains of red meat. Also not sure what to make of the sample size of ONE, comparing the newly meat-eating vegan to a “representative omnivore.” Hopefully future studies will shed some light on this. It’s also important to note the researchers never tested meat alone, it was always red meat and d3-carnitine. Not sure if it matters but it might. I do like the clear demonstration that gut bacteria play a role in TMAO formation, as antibiotic treatment for a week shifted the gut microbiota just enough to blunt down the bacteria and the genes capable of metabolizing carnitine. It’s just unfortunate that the researchers didn’t collect stool samples during this intervention (or maybe they did and are holding for a future paper – lets hope). But I do like the Prevotella connection, as it doesn’t discriminate between vegans/vegetarians from omnivores. Though the Prevotella group sample size was small, it’s informative.

For what its worth, our Mr. B above – the American Gut participant who had levels of Prevotella around 32%, also took a week of broad-spectrum antibiotics. And for giggles, he submitted another stool sample at the end of the antibiotic course. Interestingly, his levels of Prevotella went from 32% down to 13%. Does this mean his potential circulating levels of TMAO would have gone down if they had been measured? Who knows – and maybe the Prevotella went down in the antibiotic group above as well – i.e., the group that was unable to produce TMAO after antibiotics.

If enriched levels of Prevotella are in fact associated with higher levels of circulating TMAO – and you buy into the current thesis that TMAO accelerates atherosclerosis – then it’s worth paying attention to. And, if elevated levels of Prevotella might be associated with whole-grain based diets (but not highly processed carbs), then the genus Prevotella has come of age since the advent of agriculture (last 0-10,000 years more or less – depending on where you ancestors hail). A potentially interesting question then might be, is there any deleterious effects associated with the blooming of this genus in our recent evolutionary past? If humans (and hominids in general) have lived for millions of years with low to no levels of Prevotella, one might suspect that if all of sudden they constituted 30, 50, or even 90% of the total gut bacteria at any given time in some individuals, then there might be some downside. But nobody knows for sure – as nobody has asked the question. Though some have suggested they are reservoirs for antibiotic resistant genes – but not all antibiotics genes it appears.

As hard as it may seem by this point in the blog post, I don’t have a horse in this race – I’m not anti-meat nor anti-grain. I think if you want to eat meat, knock yourself out. Same goes for breads and pasta – I like those too – just don’t eat that often. I’m more interested in the microbes and what they might be trying to tell us. In the not so distant future, which can’t come soon enough, I predict we will all become proficient Microbial Whisperers when science hands us the tools and the basics of ecology become core modules in our educational systems – top to bottom. Betwixt the twain as they say lies some true understanding of human ecology and thus optimal human health.

Humans have been eating meat and thus carnitine for a long time. The question is, are the genes necessary for metabolizing carnitine a recent phenomenon, or have they always been with us? If there is some specific association between Prevotella and the genes capable of metabolizing dietary carnitine, then it appears – possibly – to have been ushered in with the agricultural revolution and has no precedent throughout much of our evolutionary past. How ironic would it be if the microbial ability to metabolize dietary carnitine from red meat is linked to whole-grain consumption! As for my hardcore Paleo eating brothers and sisters, this is the place in the story where you smile. Even if Prevotella is currently serving as a reservoir today, what’s to stop other bacteria from taking up the genetic call of duty to metabolize carnitine if Prevotella were to get depressed at a population level?

As for the title of the Nature Medicine article – “Intestinal microbiota metabolism of L-carnitine, a nutrient in red meat, promotes atherosclerosis” – a more appropriate title may have been “Researchers get vegan to eat steak – sun rises next day


***If you are curious about your personal Prevotella levels, join American Gut and find out.


[1] The 1,000 steaks a day was calculated by Dr. Chris Masterjohn in a very clever blog post. Highly recommend.


69 Comments Add yours

  1. valeriefrances says:

    Hi, wondering if there has been any thought to look at pastured animals versus those raised on “factory farms”. Not surprisingly, diet and exercise in animals has a huge impact on their biology of their fats…. and perhaps their protein as well. Just a thought? 🙂

    1. Jeff Leach says:

      Good Q – not sure. However, I doubt any appreciable differences. But….

      1. Inky says:

        Well the antibiotic consumption of factory animals if nothing else should make a difference.

      2. Valerie says:

        Perhaps the differences in fatty acid composition could have an impact on our intestinal flora. I would not rule out the potential for differences or their impact. I worked as a nutrition lab tech in the early 80’s while getting my nutrition bachelors. The professor asked me to review and synthesize a slew of articles on diet components and human blood cholesterol impacts. He had me throw out an “outlier” article from Africa that correlated milk consumption with improvements in blood cholesterol. I would bet money now that milk was produced by cows managed on pasture and not confined, and rather than an anomaly to be discarded, was actually pointing the way.

        1. Jeff Leach says:

          at the end of the day, it’s not the amount of fatty acids or carbs etc that one consumes, but how much of its escapes digestion and becomes available for the microbiota. so yes, fatty acids do matter. however, in the case of prevotella, its about undigested plant substances – which is what much of the genetic material is geared towards. they are also mucin degraders as well.

    2. Food Guy says:

      Unfortunately diet and exercise has very little effect on ruminant gut metabolism. The fat profiles of meat and milk are very similar whether they are pasture fed or fed on grain and silage.

      1. orionantares says:

        That is proven untrue. The fat profile does change significantly from grass and forage compared to grain and soy feed.

  2. Denise says:

    I thoroughly enjoyed this article – especially watching your thinking process. What amazing food for thought.

  3. Jeff, do you know if there have been any studies done on the gut microbiome of the preserved corpses which have been found? I’m thinking particularly of “cavemen” found preserved in things like peat bogs or ice, but Egyptian mummies and others would also be interesting.

    Sort of off-topic, I know, but your mention of the “paleo” diet made me curious. I also figured that if anyone knows this, I’m sure it would be you!

    1. Jeff Leach says:

      Hey Michelle – Great Q! Here is one cool study a talk about –
      I’m sure others are coming soon.

  4. skadley says:

    Might be interesting to know how much carnitine is in pasture-raised beef as opposed to grain-fed or even ‘grain-finished’ beef, since that is what the Prevotella was working on.

    1. valeriefrances says:

      yes, agreed!

  5. john pelley says:

    Hi Jeff, the only significance to this study is that it likely contributed to a grant renewal. The entire premise is based on the action of TMAO as an up-regulator of scavenger receptors on macrophages. Although this is among the first steps in the formation of atherosclerotic plaque, the scavenger receptor has little meaning if there is no “oxidized” LDL available to scavenge. A Paleo diet would have enough fat soluble antioxidants to protect LDL and this concept has been supported by experiments with vit E dosing. As you point out, this study is replete with oversights not only on a changing and/or undefined microbiome but, I would add that it is myopic with respect to the holistic nature of our metabolism. BTW, my wife and I are the two pins on the demographic map for the American Gut Project at Lubbock.

    1. Jeff Leach says:

      great insights john – sounds like TMAO etc is in your wheelhouse. and nice to see you and the missus in american gut. it’s pretty amazing to hold your own results in your hand as i have recently done – and correlate what you have been doing with the composition. eager to see how my gut microbiota changes after spending a few months with the hadza in tanzania this summer/fall.

  6. jon says:

    Great article. Thanks for taking the time to put it together.

  7. toofin says:

    I really had a question, do microbes change during our lifetime, or when we go abroad and eat something there, is there an addition of new microbes when we travel and how does our gut cope with the new microbes introduced when we travel?

    1. Jeff Leach says:

      yes, microbes change from birth, when solid foods are experienced, and so on. as for travel, yes, you can pick up news microbes. sometimes you cope, sometimes you don’t (ie you get diarrhea).

    2. john pelley says:

      You can influence what population of bacteria dominates in order to establish a favorable flora. An easy way is freshly fermented vegetables like sauerkraut. Many recipes for this on the web and it is extremely easy. Only requires consuming about 2 tbsp/day. Also, your genes influence the microbes, e.g. gut microbiome for identical twins is much closer match than with other siblings.

      You are what you eat, but you are also what you poop.

  8. Excellent post. This study had echoes of Metchnikoff’s 100-year old idea about the toxins produced by gut bacteria (“intestinal putrefaction”) being responsible for the diseases of aging. Just for the old guy’s sake it would be nice if they had been right. Maybe next toxin.

  9. A. Lynn Dolson says:

    Any comments on pH. It is very confusing that with systemic acidosis (usually low plant, high meat & processed diet, plus stress) we see disease. Yet we know that the GI tract is happier being acidic. In an alkaline state, the colon hosts an unfriendly microbiome (gaseous and otherwise unhelpful). Any comments on this and the interrelations between the GI tract and the body’s connective tissue would be greatly appreciated!

    1. Jeff Leach says:

      Yup. Fermentation equals lower pH =’s good. As you probably already know, lower pH improves gut barrier, improves colonic health, improves pathogen resistance, lowers numbers of Proteobacteria )e.g., some e coli, etc), and this list goes on. From your comment are you suggesting low plant diet lowers pH? Quite the opposite in fact.

      1. A. Lynn Dolson says:

        Thank you. Also doubting the attempts to alkalinize the body by consuming alkaline products that travel through the gastrointestinal tract (plant-based or medicinal).

        1. adalca says:

          Hi, please don’t confuse two separate things here: 1) GI tract pH; 2) Blood pH, or “body” PH, or Potential Renal Acid Load (PRAL) of food!
          Sure the lower pH the better, at least in our stomach, since pancreatic and epatic bile do neutralize such acids right at intestine start (ileum).
          But “systemic acidosis” is a WHOLE different story, which tells us of the stress the foods we eat pose on our kidneys, especially the sulfur-based acids coming out after metabolism (digestion) of animal proteins (the richest in methionine and other sulfur-containing amino-acids).
          Our kidneys filter our entire blood stock every 2 hours or the like, so those metabolic acids have the time to damage our tissues (starting from red blood cells, whose will then need vitamin B-12 to be replaced). Upon arrival to the kidneys, those acids will require calcium to be drawn from the blood in order for them to be neutralized, since urine pH can’t be lower than 4.
          So “systemic acidosis” means calcium depletion (osteoporosis), damage to blood red cells and tissues, etc.

          1. Jeff Leach says:

            thx for the comment. note i’m not confusing – understand the diff. my interest is in the pH of the colonic environment – >acidity = a good thing 🙂

    2. Give Peas a Chance says:

      Ph seems to be very important, particularly for inflammatory bowel disease and risk of colorectal cancer. The pH factor may be one reason mammal meats increase risk of both IBD and colorectal cancer.

      I do not think we can assume the gut biomes our ancient ancestors had. Depending on location and food choice I am sure it varied widely and provotella may have been present in the island climates where taro root (starch), and other plant fiber foods (coconut, fruits) were the staples along with some fish, but very little on anything resembling red meat.

  10. Rebecca says:

    “….and every biomarker of a crappy diet was elevated….” Hahaha. Thank you for a dose of humor with your BRILLIANT commentary on the latest hype to be splashed around the news. Highly informative and thought-compelling.

  11. Mark Wlodarkiewicz says:

    There has been a similar study that has just been published in the New England Journal of Medicine that found a similar TMAO situation with the lecithin in egg yokes.

    What do you think about that? It seems to me that these are really complex systems and though being able to identify one aspect of it (TMAO production in relationship to certain gut bacteria) there has to be other aspects that are in play. The notion that whole grains may play a part in this mechanism seems pretty important. I mean it could be sugars or broken down starches in relation to the carnitine or lecithin in relation to the bacteria. I am interested in this – especially since my wife wants me to stop eating eggs and red meat…

  12. Jim McGinn says:

    It’s funny, but in reading the comments here I don’t think most of the respondents absorbed the full impact of what you were saying here–or, maybe, I’m the one not getting it. What I thought was most amazing about your analysis was the fact that you made a distinction between the consumption of red meat, eggs, etc., which are contributing factors and the prevalence of provotella in the gut, which is (presumably) the cause of the production of TMA (which, as you pointed out is then converted to TMAO by the liver). And then the big slap in the face was the role played by WHOLE GRAINS.

    Here is the reason I even looked into this. I am now 55 years old. When I was I was 26 (about 1983) I came down with chronic stomach, constipation, diarrhea (IBS). It’s been something that I’ve been trying to figure out for 30 years now. I’ve long suspected some kind of gut microbe as being a contributing factor. Recently I’ve theorized (after reading an article on the internet and doing some research) that, possibly, my gut problems are a result of my gall bladder being filled up with little balls of cholesterol which reduced the capacity of the gall bladder to deliver bile in sufficient amounts and since bile is instrumental in controlling gut flora (again, I read this on the internet, so it must be true) this (theoretically) is a contributing factor to my continuing stomach issues.

    So, in the last two months I’ve done three gall-bladder, liver flushes. Each time the amount of cholesterol balls (about the size of a pea) that came out was pretty dramatic. And it does seem I’m feeling somewhat better. (But this could very well be placebo effect.)

    I’d often wondered why I came down with this condition. Unlike a lot of my friends and associates I was fairly health conscious even in my late teens and early 20s. I was especially concerned about refined foods. My solution, you guessed it, WHOLE GRAINS.

    After I send this message I’m going into my kitchen and grab the loaf of whole wheat bread and the other whole grain products (are quick oats a whole grain?) and put them in the trash. I’m also going to continue to stay clear of red meats, ham and egg yolks and I’m going to start to look into doing whatever I can do to detect and control prevotella. Maybe this is just another red herring. But, I tell you, being chronically ill for thirty years is no joke. At least now I have some hope.

    Thank you for your wonderful analysis!

    1. Michael says:

      Jim hit the nail on the head. Why aren’t other people here commenting on the whole grain issue? Whole grains appear in this entire page to be completely bad for us. AVOIDING whole grains appears to shift the gut away from provotella, and thus shift our bodies away from heart disease. So the obvious answer is to eat starches that have had their bran removed, instead of whole grains (and of course eat the other parts of a healthy diet too, like vegetables, protein, healthy fats). Easy enough. White pasta, white rice, potatoes, to fulfill starch part of our diets.
      Are we making things more complicated than they need to be in these long discussions?

  13. Henry says:

    Suppose elevated TMAO causes heart disease. TMAO may be elevated if the gut has a lot of Prevotella bacteria, AND those bacteria are fed plenty of carnitine or choline. High-carb people are likely to have high Prevotella populations, while low-carb types are likely to have low Prevotella populations (

    Then maybe the seemingly conflicting evidence makes more sense. Then vegans and low-carb people *both* would have low TMAO and lower rates of heart disease. The vegans don’t provide their abundant Prevotella with TMAO precursors, and the paleos simply have low Prevotella populations despite consuming lots of carnitine and choline.

    Maybe the real problem here is a “balanced” diet? (I’m looking at you, medical establishment and US RDAs)

  14. tatertot says:

    I just sent off for a test kit. I have been eating paleo for about 2 years, and have been on a resistant starch kick for the last 6 months. I eat about 30-40g/day of RS, probably more than any American. I get all this RS from paleo foods, raw potato starch and plantain flour. You can read about my journey here: I’m ‘tatertot’ in the article and comments.

    If I figure correctly, my gut microbes will show an extremely high percentage of bifido strains.

    Any advice for diet tweaks before my test kit arrives?

  15. Christine B. says:

    I wonder about cultures, like Ethiopia, who eat only fermented bread with their meat stews…

  16. In my last post I mentioned I had given up wheat bread. It had a significant effect. Not only am I feeling better in general but an oral inflammation problem (very slight) that I’d had for about two years (since a root canal surgery) cleared up. Nevertheless, the issues I’ve had with IBS (flatulence, diarrhea, and a persistent [and somewhat debilitating] sensation of nausea) lingered on.
    Then, in reading the comment section of this page (or, maybe it was from a page I linked to off this page) I read that prevotella feeds off anything plant based. So, I thought, what would happen if I just eliminated ALL plant based foods from my diet? It took me a while to work out a routine but about Tuesday of last week I did it and kept it up for about five day. My IBS symptoms disappeared! No joke. I was symptom free from Wednesday to Sunday. Then, on Monday, after seeing something on TV about the importance of fruit and vegetables, I thought, “This is crazy. I can’t eliminate vegetables and fruits. I’ll get scurvy or something.” That evening I chopped up and cooked (sauteed) an onion which I ate along with some chicken, I woke up Tuesday with my IBS symptoms once again. I remained vegetable and fruit free on Tuesday and felt good all day Wednesday when that evening (yesterday) I thought, “Okay, I can’t eat onions. But a simple green-leaf salad with one tomato couldn’t hurt me. Could it?” Wrong, I woke up this morning and once again my old friend IBS was back.
    Now I’m taking the next step. Now I’ve got a pretty good diet routine worked out. I’ll only eat white meats, egg whites, cheeses, yoghurt and such (I am going to allow myself some corn syrup based sodas since that didn’t seem to bother me over the 5 day period I mentioned above). Other than this I’ll just take vitamins (you know, to avoid scurvy 🙂 ). And I’ll report back to this forum and let people know how it goes.
    To me, the possibility of living a life free of IBS is dream I never thought could come true. I’m very excited.
    Thanks again for the insight provided by this website.

    1. Barb says:

      Hi Jeff, you might want to check out the FODMAP diet for IBS. You need some veggies and fruit for your gut flora (only source of fiber!), but when your digestion is impaired, some kinds of sugars make it down to the colon and the bacteria kick up heck. Here is a link:

      1. Jeff Leach says:

        Thx, very familiar with FODMAP

    2. Barb says:

      Oops, I meant the note below to be for Jim McGinn, not Jeff. Sorry Jim and Jeff!

  17. Patrick says:

    My American Gut profile came back at 54% Prevotella. I have had severe recurrent nausea for years now and my doctor put me on Omeprazole (prilosec), a proton pump inhibitor, and in a week my nausea was gone. It took 3 days without treatment to notice symptoms again. I wonder, does proton pump inhibition prevent TMAO production from Prevotella as well? Great read.

    1. Kraut says:

      Prevotella digests fiber and carbohydrates, causes fermentation in gut and gas production, this gas can override the pylorus sphincter and push the stomach contents through the les causing Gerd, Lpr, if ppi helped then this is the likely cause! follow Dr. robillards fast tract diet to decrease fermentable carbs and solve the root of your problem.

  18. julianne says:

    Prevotella copri has been found in high amounts in recently diagnosed rheumatoid arthritis patients. Does this bacteria also respond to a high grain diet? Do you think one of the results of paleo is to decrease this bacteria leading to improvements in RA?

    1. Jeff Leach says:

      Good Q – hope we can find out soon.

  19. This is great Jeff, many thanks for this detailed post.

    I’m wondering if the Prevotella differences (and maybe the bacterial variety in general) were due to other non-dietary factors? Let’s not forget that Burkina Faso has wildly different levels of hygiene and sanitation compared to a sterile laboratory in the west. Did you factor this in? There are other routes to the stomach, such as bacteria riding on the food itself.

    Raw veganism seems to be all the rage lately, and you have to admit, many of them are showing health improvements (though who knows what they were eating before switching over). Could the presumed benefits be due to simply more bacterial intake (in a vegan’s case, soil-based bacterial too) due to non-heating of food?

    1. Sorry, my sentence “There are other routes to the stomach” was unclear. I meant there may be other ways to populate the colon with bacteria, such as through hygiene, cooking methods, etc.

  20. priya says:

    recently, researchers have found a link between provetella copri microbiome and newly onset Rhematoid arthritis. Do you jave any idea what is going on with that project? I’ve been recently diagnosed with RA and hit upon your article. Thanks

  21. AM says:

    Some studies that are showing that high levels of Prevotella copri is linked to different autoimmune conditions, such as ciliac disease, gut leak, rhumatoid arthritis, rosasea, seasonal allergies, and hyper sensitivity to certain chemicals.

    I developed some of these conditions in my early 20s, off and on periodically. That correlated to a time when I started to eat more whole grains, rice, and other such materials as opposed to the mainly white bread and pasta I often ate growing up.

    To combat a form of rosacea specifically and an almost hive-like reaction at times, I have been systematically eliminating certain foods. Most of my berries and fruits have to be organic, as the pesticides cause immediate reaction, and it seems that coffee sometimes aggrivates the condition (which is a shame because I LOVE my morning coffee). Eating more yogurt seems to help, especially the stuff we make at home from organic milk and a nice starter (probiotics at work).

    However, I never thought to move AWAY from whole grains. I eat oatmeal at least once a week and all of our breads are the “squirrely” variety as my fiance calls them. Only on occasion do I eat white bread or pasta anymore. Perhaps it’s time to make the switch back and see how it goes?

  22. Carol Willis says:

    There could be a connection between prevotella and dopamine levels. Carnitine raises dopamine (and so often lowers serotonin — the dopamine-serotonin see-saw). Recent research showed that Parkinson’s patients, who are low on dopamine, are 77% lower on prevotella. Carnitine (plus alpha lipoic acid) improves dopaminergic neurons [Zaitone, et al, 2012]. Wonder if prevotella are enhanced also by the carnitine, and if implanting provotella would by itself raise dopamine?

  23. Snowy62 says:

    There is a study out now showing significantly reduced Prevotellaceae in the gut flora of Parkinson’s patients. My husband suffers from PD and we have been Paleo for a few years now with no improvement to his condition. I’m wondering what you think of trying to increase his Prevotellaceae make up?

  24. Fidel says:

    Many vegans avoid grains which could indeed partly account for the mostly low levels of prevotella for the vegans in that research group. The few vegans who had high levels of prevotella and TMAO might have been grain eaters who also liked eating a lot of tempeh which has almost as much carnitine as bacon, according to .
    Since carnitine itself feeds prevotella growth, the people here who are so quick to want to avoid whole grains should be just as quick to avoid meat, especially beef.

    1. Jeff Leach says:

      Interesting. Do you have a published reference to the “carnitine itself feeds prevotella growth.” Thanks

      1. Tim Steele says:

        Jeff – I just did an interesting experiment. I sent a stool sample for 16SrRNA analysis after eating nothing except potatoes for 7 days.

        Several things stand out in this report:
        – Prevotella went from .001% to 3.9%
        – Akkermansia from .1% to 2%
        – Blautia from 1.5% to 7%.

        The only “big guy” that was reduced was Bifido…7.3% to 1.3%.

        The diversity seen from this mono-food diet was off-the-charts high. Nearly double that of my pre-potato-intervention diet, which was very high in plant fiber.

        Could this be the paradox we are looking for? Too much diversity in foods of the Western diet leading to lack of diversity in gut flora?

        It’s discussed in the literature, but so counter-intuitive that no one took notice, ie:

        I’d love to see variations of this dietary intervention tried by others. I did it to the extreme to test the hypothesis, but there are surely variations that would be easier with just as great results. I always think back to people that had one staple food they rely on (stale maize porridge, meat and potatoes, rice, etc…). Massive variation is a modern invention.


        1. Jeff Leach says:

          Good lord! What was the dominant genus in your results?

          1. Tim Steele says:

            Top 10 genera:

            Bacteroides – 25%
            Eubacterium – 12%
            Fecalibacterium – 10%
            Blautia – 7%
            Roseburia – 7%
            Ruminococcus – 4.9%
            Clostridium – 3.9%
            Prevotella – 3.9%
            Lachnospirea – 2.9%
            Akkermansia – 1.7%

        2. Jeff Leach says:

          That is, what’s your excel spreadsheet say (a link can be found online in your account at

        3. Timar says:

          That is an interesting hypothesis you have derived from your n=1 experiment. I like counterintuitive food for thought. However, most current research into the microbiome – at least on the intracultural scale – seems to suggest that in general the intuitive “more dietary diversity (in healthy foods) equals more microbial diversity” holds true. See this excellent presentation by Dr. O’Toole, particularly the section beginning at 37:40.

  25. Pati says:

    How do you test for protovella in your stomach my son has chronic recurrent multigocal osteomyelitis a recent study showed that kids with that rare disease have protovella in their stomach I am trying to get him tested as right now they use bone scans and MRI but that Only shows already inflamed areas so not preventative. We of course already eliminating grains from our diet by just wondering if I can go to a lab and request testing of that bacteria?

Leave a Reply to Mark Wlodarkiewicz Cancel reply