Search This Site!
Nutrition Past and Future

16 The Journalist Gary Taubes 16: A Case Against Fiber (Barely Strung Together) 

Gary Taubes invokes Francis Bacon again in this passage, this time as he attempts to undermine the healthy reputation of dietary fiber. Taubes says that the ideas put forth by Denis Burkitt have been shown to be merely “wishful science.”

Denis Burkitt was a British surgeon who observed that bowel disease was strikingly uncommon among Africans, who consumed much more dietary fiber than did people in developed nations. Their carbohydrates were not refined and stripped of their natural fiber. Burkitt believed that the replacement of whole food carbs with refined carbs was a likely cause of bowel disease in wealthy nations.

Taubes tells us that the biggest and best trials have all demonstrated that dietary fiber does not protect us from the many diseases and discomforts that Burkitt said it did. To support this claim, he mentions a number of studies, including a half-dozen randomized trials, that he says prove that the public has been misled about fiber.

Before I get into this any further, let’s first consider this systematic review of the literature on the topic of fiber and colorectal cancer from 2011.  A high intake of fiber, especially the fiber in whole grains, was found to be protective against colorectal cancer. This is a recent meta-analysis, so it should cause us to at least consider the possibility that Taubes is wrong and it may be he who has a problem with “wishful science.”

Here you can see his references for this section. This  is another list so I think you know what I’m up to. Let’s get started. The first is Giovannucci 1994.

And here you see Giovannucci 1994. This choice to me is classic Gary Taubes. The man with so much contempt for the epidemiological research that shows that red meat consumption is associated with colon cancer here has decided to use epidemiology to say fiber doesn’t prevent it. That’s inconsistent enough. But this very study he is citing says that “an elevated risk of colon cancer was associated with red meat intake.” Can you wrap your head around this? His reference against the claimed benefits of fiber, which simply failed to show an association between fiber and cancer one way or the other, did show a strong association between red meat and cancer. I guess this is the sort of reporting one has to do to win journalism awards and be called a Robert Wood Johnson independent investigator in health policy. It’s quite a career he’s made for himself with these tactics. This research was from the Health Professionals Follow-up Study.

Look at the tables in the paper and you’ll get an idea why they wound up with these results. If we compare the lowest to the highest consumers of red meat here, we see that the biggest red meat eaters were eating seven times more red meat than those who ate the least! We definitely had a meaningful range of red meat consumption. But if we do a similar comparison for fiber, we’ll see the biggest fiber-eaters ate only 2.3 times more than the least, and they really weren’t consuming all that much fiber. Apparently these health professionals were far more enthusiastic about red meat than they were about cereals, beans, fruit, and vegetables. Didn’t they learn about those in school? I mean elementary school?

The authors felt they had enough of a spread in fiber consumption to test its effects because the highest fiber consumers did reach the pathetically low recommended levels for fiber intake, and because they did show that fiber was protective against diverticular disease. Yes, Gary Taubes is arguing against fiber with a study that showed that fiber was protective against diverticular disease. You wouldn’t believe me if I just said that, would you? But you can read it here for yourself.

Think about this. Taubes calls the fiber hypothesis of Denis Burkitt  “wishful science.”

And Burkitt claimed that diverticular disease is caused by a deficiency of fiber in the diet.

And Taubes has just used as his first reference in challenging Burkitt a study which strongly supported Burkitt’s hypothesis! How can Taubes do this? Why is he not also telling us that this study says that red meat does associate with colon cancer? Would this meet your personal standards of honesty if you called yourself a journalist? I am not asking much, Gary Taubes.

TAUBES: Just, stay honest, Dean. That’s all I’m asking, ok?

Next on our list is Fuchs 1999.

Here it is, Fuchs 1999. This one focused just on fiber and indeed it found no protective effect with respect to colorectal cancer or polyps among women in the Nurses’ Health Study.

But again, you see we are talking about low levels of fiber consumption. The nurses topped out at 25 grams per day. This is not a real test of Burkitt’s hypothesis.

Now the authors did try to control for obvious considerations, such as long-term habits of fiber consumption.

But there may have been other shortcomings in their study for which they could not account. This is interesting. Their method of calculating dietary fiber was based on a method proposed in 1976. Do you see the name associated with those standards for reference 7? Southgate is the author. Remember the name David Southgate for a moment.

The definition of “fiber” has been somewhat of a moving target. This adds an element of imprecision to this whole enterprise and renders older studies less reliable than newer ones, at least with respect to their accuracy in the measurement of fiber. This paragraph comes from an editorial that accompanied this study. John Baron explains the problems inherent to the quantification of fiber intake in a trial like this. Reading this now may help you understand this issue better as we move along.

This study received some very good published responses. Dr. Neil Ravin made it clear that Burkitt’s concepts about fiber were in relation to truly high-fiber diets. Read the second paragraph on the left for his vivid description of Burkitt’s idea of a proper human stool. Another correspondent found fault with the use of those standards of quantifying fiber that were developed by David Southgate. Did you remember that name? The critic of the old Southgate methods used here was, as you can see, David Southgate. How funny is that?

I think the best comments were from Dr. Mary Ellen Camire. She touched on some of the most important considerations in the study of fiber. Modern food processing changes the nature of fiber in important ways. The benefits of fiber are associated with whole plant foods, not fiber supplements. Yes, fiber supplements are mostly what is at issue in these studies. But I am getting ahead of myself.

Many of the same investigators from that 1999 study that Taubes picked took part in this pooled analysis published in 2005. For example, you can see both Fuchs and Giovannucci at the bottom of that list of authors. Here they did find an inverse relationship between fiber consumption and colorectal cancer, but that relationship lost statistical significance once dietary folate was considered. Folate is generally believed to be protective against colon cancer. The primary sources of folate in the diet are fiber-rich plant foods, so you can see why that might add a confounding influence if we are trying to consider fiber in isolation.

If you look at the Wikipedia entry for Denis Burkitt, you will see that a news article about this study was used by someone to claim that most cancer researchers reject the fiber and cancer hypothesis. There is no mention on this Wiki page that their adjustment for folate probably explained this finding. Wikipedia may not be the best resource for health information out there.

This 2012 paper which came out of the huge and rigorous EPIC project states the folate issue succinctly here. Adjustments for factors like folate may be muddling our understanding of the effects of fiber.

This EPIC study did find a protective effect for fiber against colorectal cancer after following up with participants over a period of 11 years. The authors said this finding strengthens the evidence in support of high-fiber diets in efforts to prevent colorectal cancer. This is a major study that takes us up to the present day. Gary Taubes couldn’t have seen this EPIC study before he wrote his book.

But he could have seen this one. A 2003 paper from EPIC found a similar result. Fiber did seem to be protective against colorectal cancer in this one. Did the journalist Gary Taubes not hear about this one?

The researchers here sought to explain the lack of association in the study of nurses in the United States, our second reference used by Taubes. They surmised that that study failed to find an association because their participants consumed less fiber from a more limited variety of high-fiber foods. The latter point, along with that wrinkle added earlier about folate, leads us to an important concept regarding fiber.

Fiber isn’t the only thing in plant foods that might affect cancer risk. Plant foods are far more complex than that. They bring nutrients like folate along for the ride. For this reason, the authors said, “our results cannot be extrapolated to any potential benefit of dietary supplements or additives containing fiber alone.” They didn’t think their finding would necessarily relate to fiber supplements, a point that sets us up perfectly for Taubes’ other references.

Taubes claims that a half dozen randomized control trials have disproven any benefit from fiber against colorectal cancer. I’ll run through these quickly so you don’t need to simply take his word for it. The first is McKeown-Eyssen 1994.

And there it is. This is quite representative of the other ones Taubes picked. Patients who already had polyps removed were used in this study. They could have been as old as 84. They were given a supplementary wheat bran snack. Bran is just one part of the wheat kernel. A snack made of wheat bran should be considered a fiber supplement.

Look at the data and it tells the story in this one. The columns for each group are ND for Normal Diet, as in normal western diet, and LFHF, which means low-fat high-fiber. Look on the left and you can see that many of the participants had already encountered problems with their bowels. Now look at the right. Their normal fiber consumption before they started the trial had been pathetically low, only about 15 and 18 grams per day. So here is my interpretation of this: People selected for polyp removal probably got themselves into that situation by eating less than 18 grams of fiber per day. This is a bit like the diet-heart trials. People showed up for the trial with plaque built up over many decades, and a bad diet dripping with refined oils was supposed to make a noticeable difference for them somehow. It’s asking a bit much. We shouldn’t be surprised that they didn’t find much benefit from their bran snacks here. Recall for a moment that the lead author was Gail McKeown-Eyssen.

Here is her name on a much more recent study. This was from an analysis of three randomized trials. Dietary fat and red meat intake were associated with an increased risk of polyps. The author Taubes picked has recently found that fat and red meat appear to be linked to the early stage of colon cancer.

The next on his list is MacLennan 1995.

This one was more of the same. In order to be included, one must have had polyps removed. Again, they were getting a dietary intervention to people very late in the game. What was the intervention? Supplemental wheat bran was given to people on a reduced-fat diet. Yes, fiber supplements are the distraction here, too.

They noted that they were only dealing with a small number of patients. They also commented that they only really increased fiber intake by about 7 grams, and this came from a supplement. Moreover, patients were not told to avoid red meat. They really hobbled themselves in this trial.

As you can see, even under these difficult conditions, the intervention did seem to show some benefit. They believed they were able to prevent the development of larger polyps. Taubes didn’t mention that. I guess if you are going to have polyps you’d rather they were small.

After the MacLennan paper, it’s time to look at Alberts 2000.

Once again, we need only glance over the abstract to see the limits of this paper. The subjects had polyps in the recent past. The intervention was only 13.5 grams of refined wheat-bran fiber. Notice the conclusion. They clearly limited the wording in their findings. “As used in this study, a dietary supplement of wheat-bran fiber does not protect against recurrent colorectal adenomas.” Gary Taubes must have thought they said “all dietary fiber.” That’s because he doesn’t allow information into his mind that might cause him to rethink his biases. If Gary Taubes wanted to limit his claims the same way these authors limited theirs, I would have no objection. The only people who would even contemplate taking a wheat bran supplement would be people who were consuming inadequate quantities of whole plant foods. Now who would do that? Who would be so foolish as to believe fiber supplements could compensate for a deficient diet?

Robert Atkins.  This is from Dr. Atkins’ New Diet Revolution.  He asks, “How can you get the benefits of fiber without the carbs contained in these foods?” His answer is supplementation. He goes on to recommend psyllium husks and wheat bran or flax meal. He also advised a lot of water to avoid constipation. Low carbers should look at Taubes’ half-dozen trials, and then look at Robert Atkins’ diet ideas, and realize that they just cannot square this circle. They should reject the whole crazy low carb delusion and eat some real foods that allow their bowels to work normally. But that won’t happen anytime soon so I’ll just return to talking about his study.

When we look at the paper’s contents, we see that their previous adenomas, or polyps, had been removed only three months earlier or less. As in our previous trial, they had been consuming diets estimated to be quite low in fiber. We’re only up to 19 grams here, and they were probably from overly processed foods. The fact that they had developed adenomas eating like this seems to actually support Burkitt’s hypothesis, don’t you think?

When these nutritionally challenged individuals began consuming their refined fiber supplement, as a group they began to experience a range of difficulties, from constipation to cancer. This is just one more strike against the Atkins approach.

Even with those supplements, these individuals were only consuming 27.5 grams of fiber a day, which tells us that they were only consuming enough whole plant foods to reach 14 grams of fiber without the supplement. The doctors running this study tell us that it would have been impractical to have older individuals consume any more refined fiber. Perhaps it would have been a bit more practical to simply have them eat a reasonably healthy diet naturally full of fiber and skip the supplements altogether.

Next is Schatzkin 2000.

This one was a diet trial, but again, we are talking about recurrence, so we are confining ourselves to people who likely had poor diets for decades before this four-year period.

Had Taubes actually read this study, he might have seen why his other studies don’t tell us much about fiber. These authors spell out that those other trials used supplements, without real dietary change. We see that the subjects in this one were overweight – a BMI of 25 is overweight – which is another indication beside their polyps that they had been consuming unhealthy diets prior to the study. For these individuals, the typical intervention brought them to a total of 5 to 8 servings of fruits and vegetables. Beyond that, they did make real improvements in their overall diets.

However, as the authors noted, it all might have been too little too late. I think this about sums it up. Once again, this was not a study that had much chance of testing Burkitt’s hypothesis. These people already had polyps removed. “A change in diet later in adult life may be ineffective,” they said. It’s not reasonable to expect a four year improvement in diet to compensate for decades of failure.

Maybe Bonithon-Kopp 2000 will support Taubes a little better.

Or maybe not. We’re dealing with recurrence again, and our intervention is a measly 3.5 grams of a particular fiber supplement called ispaghula husk along with a calcium supplement. Really, Mr Taubes? 3.5 grams?

At least these authors gave us the information we need to figure out how useless their study was. Their fiber intervention group increased their fiber by a whopping one-half gram at the end of three years. I’m not kidding. Read the table. It was a half-gram difference. Denis Burkitt fans will be devastated to learn that this did not produce any effect!

The authors acknowledged that they cannot draw any conclusions about other types of fiber beside ispaghula husk from this one. Once again, Taubes has an issue with over-interpreting data. Here it comes!

TAUBES: You do not over-interpret your data. You do not say your studies say something they do not say. And to do that is bad science.

Yes, very bad, Mr Taubes. It’s almost unbelievable that someone would over-interpret a trial that made so little difference in the intervention group. It’s a bit more believable when we consider that you profit from this misinformation.

Ok, that study was super-lame, but Gary Taubes is an amazing researcher with the truth on his side. I’ll bet the next one will be a direct hit that will bring down Big Fiber forever.

In the nine years this one has been out it’s been cited only three times in the medical literature. As you see, it was a diet trial.

It studied the rate of cancer cell proliferation after four years, which is an intermediate marker of disease.

The authors tell us that by focusing only on this, they put themselves at a disadvantage in trying to assess the efficacy of their intervention. They did their best to try to compensate for this.

We see that their subjects were once again individuals who needed polyps removed recently, and we see that they were overweight. We are not told if they were low carbers.

The authors gave us some proper context for this study. A disease process had already been set in motion long ago in these patients. It may have been too much to expect a quick fix for their problems.

Now we’ve been through his six randomized control trials. I don’t know about you but I find this line of argument to be pathetic so far. The next reference he lists just says “fruits and vegetables”. What could this be about?

Here’s the text from his book from which we are working. Look a little past halfway down on your screen. Do you see that he says “and a half-dozen randomized trials concluded that fiber consumption is unrelated to the risk of colon cancer, as is, apparently, the consumption of fruits and vegetables”? How does he support that? Isn’t that a rather big claim to bury in the middle of that sentence?

Here’s the paper. The frequent consumption of fruits and vegetables conferred no protection against colorectal cancer. What is frequent?

Well, you can see that they didn’t want to factor in the effects of the most frequent fruit and vegetable consumption. If you ate more than ten servings per day of fruits and vegetables, you were considered an outlier in this study. You only got credit for your first ten fruits and vegetables. From my perspective, the outliers, had there been enough of them, might have consumed enough fruits and vegetables that the effects of their diets might have been more apparent, since at a certain point one might be consuming so much produce that there isn’t much room left for junk food and meat. But this isn’t the biggest problem with this study.

Remember, Gary Taubes is referencing this study in a discussion of fiber. Well, don’t fruits and vegetables have a lot of fiber?

Not if they are juice. That’s right, The Nurses’ Health Study, which Taubes considers one of the larger and more rigorous trials, equates a serving of apple cider with a serving of strawberries. The Health Professionals’ Follow-Up Study, which was included in this paper, had the same flaw. Now I know that Gary Taubes considers himself to he a very smart and important leader in the world of nutrition now, but he may not be aware of this fun fact:

It is possible to drink a fruit juice and consume absolutely no fiber. Moreover, what little fiber might be in juice is not in its original form and therefore will not regulate the digestion of the fruit sugars. And by the way, there will be a lot more of those sugars in a single serving of juice than in a single serving of fruit. Mr Taubes should watch my “Evolved Fuel System” videos to come up to speed on this very basic piece of common sense.

The results of her study got the lead researcher doing some serious pondering on the topic of measurement error. There is no mention here of the possibility of a faulty definition for “fruit” in their fruit and vegetable study. Sometimes very smart people overlook the obvious.

Despite that, a later paper looking just at the Nurses’ Health Study data did find that fruits appeared protective against polyps. Taubes didn’t tell us that because he is a low carber who likes to pick cherries.

At the bottom, you see he brings up the Dietary Modification Trial of the Women’s Health Initiative. Do you see that at the bottom of the slide he says this study confirmed that increasing fiber consumption had no beneficial effect on colon cancer, heart disease, or breast cancer? Viewers of my channel will know this is one of the less impressive studies out there. Let’s look at it again briefly just to make sure this is covered in this discussion.

As I’ve shown my viewers already, this was really just an effort at intervention more than an actual intervention. It didn’t focus on saturated fat at all and it didn’t really change the diets of the participants much.

Remember once more that Gary Taubes is putting this in his chapter on fiber. I’ve highlighted for you here changes over six years in fiber and fruit and vegetable consumption. Taubes can quote Francis Bacon or any other great thinker as much as he cares to if he really wants to be that much of a blowhard. It doesn’t change the fact that this great journalist and researcher never bothered to subtract 15.4 from 16.9. That’s pretty poor considering he has a degree from Stanford in aerospace engineering. Mr Taubes, 16.9 minus 15.4 is 1.5. That’s the increase in fiber these women achieved and that is a small number.

Here you see Taubes talking about fruits and phytochemicals on his blog. Check out the bottom of this slide. This award-winning researcher is actually supporting his views with a quote from Wikipedia. He is also missing the point about phytochemicals. His Wiki quote says the evidence is limited for the benefits of specific phytochemicals. Mr Taubes, fruits contain a whole mess of phytochemicals. You don’t just consume specific phytochemicals eating fruit. You consume specific phytochemicals in supplements, and that is the sort of research being referenced there.

Because the anti-fiber studies Taubes supplied were so weak, I decided to add another one to try to bolster his case. This made some news recently. Anne Perry and her co-authors made an absolute statement in their title: “A High-Fiber Diet Does Not Protect Against Asymptomatic Diverticulosis.”

This reminds me of the obviously overly broad statement in the conclusion of the big saturated fat meta-analysis that the low carbers love. There is no significant evidence of increased risk, they say. How ridiculous! You’d better believe I’m going to talk about this one again. That will be in my Andreas Eenfeldt videos.

So how did these brilliant researchers reach this definitive conclusion? First, they only did a cross-sectional study, so they are on the weak end of the research spectrum from the start. But look at who they studied. They only studied people undergoing colonoscopies. That means, as they spelled out, those people were getting colonoscopies because they had rectal bleeding or a family history of bowel cancer or polyps. I helpfully included a Wikipedia definition for sampling bias for them here. They started off with people who were very likely to have diverticulosis, and from that they drew a conclusion for everyone else. Good grief!

Once again, I find the remarkable feature of this study to be the fact that these people showing up for colonoscopies were eating so little fiber to being with. Notice that the researchers could not detect an association between fiber intake and bowel movement frequency. That should tell you something about the people and the data they were working with. I doubt most of you need a reference from me showing you that increased fiber intake is associated with increased bowel movement frequency. I’m guessing you kinda know that already.

Here it is anyway. “In general, fiber is found to increase the frequency of bowel movements and may prove to be of considerable benefit in treating constipation, a common childhood disorder.” I included a couple other good lines from this paper. Notice the bottom one in particular. Once again, the physical form of plant-foods is important.

Let’s look at what other studies have had to say about diet and diverticular disease. The Health Professionals Follow-up Study followed men prospectively for four years. They found that dietary fiber was inversely associated with diverticular disease, and this was primarily due to the fiber in fruits and vegetables. Read the right chunk of text carefully. Low-fat and high-fiber diets did well, and by comparison high-red-meat and low-fiber diets did terribly.

Here is the result with regard to fiber and low fat stated clearly. There was a strong association with reduced diverticular disease.

This study of vegetarians and non-vegetarians using the Oxford cohort in the EPIC project found that both a vegetarian diet and a higher intake of fiber were significantly associated with a lower risk of diverticular disease. This study was not plagued by obvious sampling bias.

This recent systematic review of the literature on inflammatory bowel disease found that high intakes of fat and meat were associated with an increased risk of Crohn’s disease and ulcerative colitis. High intakes of fiber and fruit were associated with lower Crohn’s risk, and high vegetable intake was associated with lower ulcerative colitis risk.

Here is another excerpt from this study for you to read.

Before I end this video, let’s see what Denis Burkitt himself had to say about fiber. He stated on the left that the fiber deficiency hypothesis was based on a lack of unrefined carbohydrate foods, not a lack of fiber supplements. He understood that fiber-containing foods didn’t just provide bulk to the stool. Cereals promote a healthy microbiome and supply needed phytochemicals. This was written in 1974. I find that impressive.

From here on out Gary Taubes will not be my exclusive focus but there is more debunking of his book yet to come. In the next video, he will have to take a back seat to Thomas Dayspring, a well known authority on lipids. He will be getting an intervention of his own.