Fibre – ask why, not how
* This week’s note is about fibre. A recently published paper assumed that there is an optimal fibre intake (25-30g daily) and it investigated how people can be encouraged to consume this.
* The evidence presented in the paper, to make the claims about optimal levels of fibre intake, came from population studies.
* Cereals were singled out as especially beneficial for health. The lead author of the paper had a number of conflicts with cereal companies.
* The authors reported that only 7.4% of US adults consume adequate fibre intake. The authors concluded that this was because people are ignorant. Their recommendations were patronising in consequence.
* There is no robust, quality, evidence to support the authors’ starting point – that 25-30g fibre daily is required. The epidemiological studies, upon which they rely, have fundamental flaws. There is a large, long randomised controlled trial on this research question, but the authors ignored it.
* When it comes to fibre, we should be asking why, not how.
This week’s note comes from my most prolific paper finder – Dr Peter Brukner from Australia. I don’t manage to do all the papers that Peter sends through, but they’re usually varied and interesting, so I try to cover as many as possible. This one is on a classic topic, which keeps coming up – fibre – or fiber, as it’s spelled in the US.
The paper was published in the BMJ in July 2022 and it was called “Fibre intake for optimal health: how can healthcare professionals support people to reach dietary recommendations?” written by McKeown et al (Ref 1). It was part of the BMJ SwissRe “Food for Thought” collaboration, which disappointed me, as I thought that this collaboration was focused on challenging the status quo, not indulging it (Ref 2). The title tells us the researchers’ beliefs – there’s an optimal intake of fibre; how can we ensure that people consume this?
I’ve been particularly outspoken about fibre. The main reasons for this are 1) the claims made about fibre are not evidence based and I don’t like claims being presented as evidence based when they’re not. 2) I think that fibre is being promoted as the only ‘justification’ left for pushing carbohydrates and 3) fibre health claims are made so often and so stridently that they need countering. I’m by no means the only person countering them. However, collectively, our counters are tiny relative to the relentless fibre public relations campaign.
The paper was a narrative review. It opened with a definition of fibre. It noted that the definition of fibre has evolved from “remnants of plant cells that are resistant to digestion by human enzymes” to “carbohydrates with three or more monomeric units.”
I have previously defined fibre as follows. The carbohydrates that we eat fall into three categories: monosaccharides (single sugars); disaccharides (two sugars) and polysaccharides (many sugars). The single sugars are glucose, fructose and galactose. The disaccharides are sucrose, lactose and maltose (which break down into single sugars). The polysaccharides include digestible forms of carbohydrate (glycogen and starch) and indigestible forms of carbohydrate (collectively called fibre). The indigestible forms of carbohydrate comprise insoluble fibre (which doesn’t dissolve in water e.g., whole-wheat flour, bran, and vegetables) and soluble fibre (which dissolves, or swells, in water e.g., oats and legumes).
The key fact to remember is that, during digestion, all carbohydrates need to be broken down into monosaccharides (simple sugars) to be absorbed. This means that all digestible carbohydrate either is, or breaks down into, sugar (Ref 3).
We thus agree on the fact that fibre is carbohydrate and indigestible.
The paper discussed soluble and insoluble fibre. It presented the benefit of soluble fibre as delaying gastric emptying and the benefit of insoluble fibre as promoting laxation and regularity. You can’t have it both ways! It was also claimed that soluble fibre improved glycaemic control, which is an often-made claim for fibre. It’s a disingenuous claim. It means that, relative to refined carbohydrates, fibrous carbohydrates have less of a detrimental effect on glycaemic control. Compared to meat, fish and eggs, which don’t impair glycaemic control, fibre has a harmful effect.
The paper noted that current guidelines differ around the world and by age groups, but 25-30g or more daily is widely recommended for adults. “In Europe and North America, grain-based foods are the predominant contributors to dietary fibre, followed by vegetables, potatoes, and fruits, with little contribution from legumes, nuts, and seeds” (Ref 4). The researchers said that “strikingly”, few people meet those recommendations. A study published in 2021, which related to US adults (with and without diabetes), estimated that approximately 7.4% of US adults met adequate intakes for dietary fibre (Ref 5). The rest of the McKeown et al paper extolled the virtues of fibre and suggested ways in which to increase fibre consumption from current low levels.
The benefits of fibre
The paper referenced 11 papers as evidence for the benefit of fibre (their references 10-13 inclusive, and 21-27 inclusive). They were all population (epidemiological) studies and/or papers that pooled together population studies. As we frequently note, population studies can only establish associations. Populations studies are supposed to precede randomised controlled trials. An observed association is supposed to be tested in a clinical trial. This doesn’t happen nowadays and has not happened for years. Epidemiology is taken as evidence in itself, which is a major reason for nutritional research being so poor.
The praise was mostly for fibre generally. However, there were three incidents of cereals being singled out as particularly beneficial. The first claimed that “observational evidence indicates that higher intakes of fibre from cereal sources (that is, primarily whole grains) seem to provide a greater degree of protection against the development of type 2 diabetes.” The second claimed “Based on a recent meta-analysis of data from seven cohorts, a higher intake of whole grains, a rich source of cereal fibre has been linked to a 13% reduction in the relative risk of colorectal cancer” (Ref 6). And “a more recently published prospective cohort study [found that] only fibre from cereal foods was inversely associated with colorectal cancer incidence” (Ref 7).
The lead author, Nicola McKeown, has received funding from General Mills Bell Institute of Health and Nutrition, Cereal Partners Worldwide, Procter & Gamble (to support dietary fibre research), and is an unpaid scientific adviser on the Oldways Whole Grains Council.
How to increase fibre intake
One of the recommendations in the paper was that people should eat their five-a-day. That’s another non-evidence-based health message (Ref 8). The paper went further and suggested that people should eat at least five portions of fruit and vegetables each day. The authors opined that people should eat more whole grains, legumes, nuts and seeds, and fruits and vegetables at mealtimes and snack times to reach a target of 30g or more of fibre a day. They suggested, “Adding one tablespoon of flax or chia seeds to salads or oatmeal will provide an extra 3g or 5g of fibre, respectively.” Finally, they advised, “Check nutrition labels to identify fortified high fibre cereals – some wheat or bran based products might have up to 11 g of dietary fibre per serving.” May I suggest that healthy eating means not eating foods with labels?
The paper also contained policy suggestions to help populations, not just individuals, to increase fibre intake. The first suggestion was “Fibre is generally listed on processed foods, giving consumers that ability to choose foods rich in fibre.” May I suggest that healthy eating means not eating processed foods? But no, apparently we should eat processed foods because they report grams of fibre on the packet.
I was shocked by the authors’ disregard of income and circumstances. The authors questioned why dietary fibre guidelines are not being met. The suggestions given were related to consumer ignorance – people don’t understand the type or quantity of individual foods or combinations needed to achieve intakes. People don’t know how to identify whole grains from refined grains. Some consumers are (stupidly, was the implication) following low-carb diets etc. The word “afford” did not appear in the paper. The idea that chia and flax seeds (and Jerusalem artichokes, no doubt) are not on regular, weekly shopping lists did not seem to occur to the authors.
I’ve just looked at the largest UK grocery store online (Tesco) and chia or flax seeds cost £1 per 100g. Apples are 50-75 pence each. Meanwhile, I can buy entire 400g meals (the Hearty Food Company range) for 75p. There are many options including cheese & tomato pasta, lasagna, mac and cheese, sweet & sour chicken with rice, chicken curry with rice, and more. All the pasta is white and all the rice is white. Low fibre, but convenient and cheap.
Rebutting the fibre claims
The simplest rebuttal to the claim that we must eat 30 grams of fibre is this: Fibre is only found in carbohydrate and there are no essential carbohydrates. De facto, fibre is not essential. We have no requirement for it. Full stop.
The question then becomes, we don’t need it, but is it beneficial? To answer this, we need to turn to evidence:
1) The top-level evidence would be a meta-analysis (pooling together) of evidence from randomised controlled trials (RCTs). This is not possible. There has only been one long, large randomised controlled trial (RCT) where (cereal) fibre was a specific intervention and actual health outcomes were measured. There have been a number of small, relatively short, trials where surrogate end points, such as blood pressure (Ref 9) and cholesterol levels (Ref 10), have been measured. Those are just a couple of examples of such trials.
The one large, long RCT was the 1989 Diet and Reinfarction Trial (DART) (Ref 11). DART was a 2-year trial designed to examine the effects of 3 different dietary interventions in 2,033 men who had already had a heart attack. The men were randomly allocated to receive, or not to receive, advice on each of three dietary factors: fish; fat and/or fibre. (This study was not generalisable beyond UK men who had had a heart attack).
The fibre advice was to increase the intake of cereal fibre to 18g daily. 1,017 men were given fibre advice; 1,016 were given no-fibre advice. There were 123 deaths (12.1%) in the fibre advice group vs. 101 (9.9%) in the no-fibre advice group. There were 109 deaths (10.7%) from heart disease in the fibre advice group and 85 deaths (8.4%) from heart disease in the no-fibre advice group. Those results were not statistically significant, but they clearly provided no support whatsoever for cereal fibre.
The Women’s Health Initiative trial was indirectly a large enough, long enough, test of fibre intake (Ref 12). The main aim of this study was to test the hypothesis that a diet low in fat and high in vegetables, fruits and grains would reduce breast and colorectal cancer in postmenopausal women. A secondary aim was to test if the same diet could reduce cardiovascular risk. The study involved 48,835 post-menopausal women, aged 50-79, of diverse backgrounds and ethnicities. They were followed up for 8.1 years on average. There was no significant difference between the diet and control group for deaths from any cause or deaths from cardiovascular disease. The high fibre intervention made no difference. (This study was not generalisable beyond American postmenopausal women).
2) ‘Evidence’ for fibre claims, as the McKeown et al paper confirmed, is epidemiological. This suffers from three flaws and starts from a base of another flaw, which is the unreliability of food frequency questionnaires. The three flaws are i) the studies can only find association, not causation. Observed associations should be tested in RCTs. In the case of fibre, the one large, long, RCT preceded epidemiological studies, but has been ignored. ii) relative risks are claimed, absolute risks are tiny. Differences of approximately 10% (relative risk) are trumpeted as significant findings when absolute risk differences may be 1 in 1,000 vs 1.1 in 1,000. iii) The healthy person confounder. This is huge in fibre studies.
Small associations are found in population studies showing that people who eat fibre have fewer incidents of disease (whether cardiovascular disease, cancer, or type 2 diabetes etc). However, that’s because of this: people who eat fibre (quinoa and cashew nuts) are completely different people to those who consume fibre-free items (burgers and fizzy drinks).
The issue with all nutritional epidemiology was best summed up at the Denver conference (2019) by Gary Taubes. Nutritional epidemiology compares two entirely different people and assumes that if only person A ate the same as person B, they would be as healthy as Person B.
We have used these images many times, but that’s because they are so powerful. Nutritional epidemiology assumes that, if only the couple in the fast food outlet ate like the family having breakfast together, they would be as healthy as the family. This is complete nonsense.
I presented another way of looking at the healthy person confounder in this post (Ref 13). In it, I explained that epidemiological studies, which compare high fibre diets with low fibre diets, are guilty of the following…
I defined nine food groups as: 1) meat; 2) fish; 3) eggs; 4) dairy; 5) vegetables; 6) fruit; 7) nuts & seeds; 8) legumes (beans, pulses, etc); and 9) grains (Ref 14). I designed a diagram to put these food groups into a table, as follows.
Box A contains food groups high in fibre and low in carbohydrate.
Box B contains food groups high in fibre and high in carbohydrate.
Box C contains food groups low in fibre and low in carbohydrate (with the exception of some carbohydrate in dairy, these foods are essentially zero fibre and zero carbohydrate).
Box D contains food groups low in fibre and high in carbohydrate. Box D contains foods that don’t fit into natural food groups, because they aren’t natural foods – all junk is in Box D.
Every epidemiological study and every meta-analysis of these studies is comparing people who eat things in Box B with people who eat things in Box D. They do not compare people who base their diet on Box B with people who base their diet on Box C. Box C contains the most nutrient dense foods. The foods richest in what we need to consume – essential fats, complete protein, vitamins and minerals (and in the form in which the body wants these nutrients) – are naturally zero/low carb and zero/low fibre. That should immediately tell us something about the ‘value’ of carbohydrate/fibre.
Figure 2 in the McKeown et al paper reported reduced risk ratios for high fibre diets vs the control diet in each study. Such population studies compare the highest fibre intake group of people with the lowest fibre intake group of people. The low fibre diet is Big Mac and fries not a real food/Paleo/ancestral diet. These studies compare a bad diet with a diet that includes fruit, vegetables, nuts, legumes and other whole foods. They compare Box B with Box D and that has nothing to do with fibre. Every subsequent claim for fibre will continue to do this. It needs to stop.
Until next time
All the best – Zoë
Ref 1: McKeown et al. Fibre intake for optimal health: how can healthcare professionals support people to reach dietary recommendations? BMJ. July 2022. https://www.bmj.com/content/378/bmj-2020-054370
Ref 2: https://www.zoeharcombe.com/2018/06/swiss-re-bmj-nutrition-conference/
Ref 3: https://www.zoeharcombe.com/2022/05/why-eat-carbohydrate/
Ref 4: Stephen et al. Dietary fibre in Europe: current state of knowledge on definitions, sources, recommendations, intakes and relationships to health. Nutr Res Rev. 2017.
Ref 5: Miketinas et al. Usual dietary fiber intake in US adults with diabetes: NHANES 2013-2018. Curr Dev Nutr. 2021.
Ref 6: Reynolds et al. Carbohydrate quality and human health: a series of systematic reviews and meta-analyses. Lancet. 2019.
Ref 7: Hullings et al. Whole grain and dietary fiber intake and risk of colorectal cancer in the NIH-AARP Diet and Health Study cohort. Am J Clin Nutr. 2020.
Ref 8: https://www.zoeharcombe.com/2012/03/five-a-day-the-truth/
Ref 9: Streppel et al. Dietary Fiber and Blood Pressure A Meta-analysis of Randomized Placebo-Controlled Trials. JAMA Internal Medicine. 2005.https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/486375
Ref 10: Brown et al. Cholesterol-lowering effects of dietary fiber: a meta-analysis. The American Journal of Clinical Nutrition. 1999.https://academic.oup.com/ajcn/article/69/1/30/4694117?
Ref 11: Burr et al. Effects of changes in fat, fish, and fibre intakes on death and myocardial reinfarction: diet and reinfarction trial (DART). The Lancet. 1989.
Ref 12: Howard et al. Low-fat dietary pattern and risk of cardiovascular disease: the Women’s Health Initiative Randomized Controlled Dietary Modification Trial. Jama. 2006.
Ref 13: https://www.zoeharcombe.com/2019/01/why-do-studies-conclude-that-fiber-is-associated-with-better-health/
Ref 14: https://www.zoeharcombe.com/2015/05/food-groups/