What's wrong with the Banting argument (part 1)

Anyone living in South Africa with access to the media can’t help but have heard of the Banting diet (also called the Low Carb High Fat or LCHF diet) and its biggest proponent Professor Tim Noakes. Prof Noakes is undoubtedly popular but he has suffered a great deal of criticism from within the academic circles that he espouses. Several high profile academics from his own University of Cape Town Department of Medicine have publicly attacked his views and Professor Jacques Rousseau, also from the University, has had a long running exchange on his blog. This may well confuse the outsider; why is Noakes so popular outside academic circles and so unpopular within them? Whether you are already a LCHF follower or a sceptic this article aims to explain the apparent contradiction.

I generally side with those who criticise Prof Noakes but I am not anti-LCHF. Instead I am pro-Evidence Based Medicine (EBM) and this is an important distinction. EBM is the science of using the best possible medical evidence to provide the best possible healthcare and at the moment it is largely at odds with LCHF. In short, it’s not about the diet, it’s about the use and misuse of evidence by LCHF followers. There is actually some good quality evidence for some aspects of LCHF, such as short term weight loss, and I am happy to endorse it for that purpose. However, there is an enormous gap between the majority of the claims about LCHF and evidence to back them up.  

Evidence for medical interventions comes mostly from studies published in the peer-reviewed scientific literature and there is a standard hierarchy for publications. The lowest quality and therefore the weakest evidence is merely someone’s opinion based on their experiences, slightly higher but still very weak evidence comes from anecdotes about single patients, the hierarchy continues to descriptions of multiple patients, then formal trials such as case-controlled studies and cohort studies. Higher still are randomised controlled trials (RCTs) and even higher than that are meta-analyses which group together all the evidence from multiple RCTs. Below this is displayed visually as the Evidence Based Medicine pyramid. It is wider at the bottom because the evidence is much easier to create and is therefore more plentiful, as you move up the pyramid the evidence is of a higher quality but there is less of it. 

The search for evidence always begins with a question. The more specific the question the better and while you might find good answers to one question there might be limited evidence for a different but related question. In general you move up the pyramid looking for the highest quality evidence that answers your question. Even if there is contradictory evidence lower down it is typical to make decisions based on the highest level of evidence available. Another generalisation is that you need evidence at the level of either a very good RCT or a meta-analysis for an intervention to be considered as proven. Although exceptions exist, evidence below that level is generally a way of generating hypotheses to test in trials. 

Let’s look at some examples of claims made by proponents of LCHF to see what evidence is available. Many claim it helps people to lose weight so the question would be ‘what is the highest quality evidence that eating a LCHF diet compared to a control diet helps people lose weight?’ Fortunately this question has been tested often and there is quite a lot of high quality evidence. In fact there are a number of meta-analyses, the highest form of evidence, on the subject. The answer is that over a time period of about 1 year LCHF diets are superior to High Carb Low Fat (HCLF) diets when it comes to weight loss. No EBMer could argue with that and I am in total agreement with Professor Noakes on this point. It is important however to appreciate the narrowness of the question. The control diets are HCLF and the duration is fairly short. The evidence doesn’t say much at all about keeping the weight off for a prolonged period, nor does it say much about comparison diets other than HCLF, which itself has been largely discredited.

While the LCHF and EBM movements can agree on this point we don’t have to delve much deeper to find disagreement. At times it can be difficult to pin down LCHF proponents but there seems to be a general claim that LCHF is the optimal diet for a long and healthy life regardless of age or background. Prof. Noakes once even tweeted that a mother should wean her baby on to a LCHF diet for example. When we go to the medical literature we find there are relatively few RCTs to help determine the healthiest diet. There are very many observational studies where scientists observe what people eat and see how healthy they are but this is weak evidence that is often contradicted by RCT data because healthy people tend to choose different food from unhealthy people. Despite a general lack of large well conducted trials there are a couple of notable exceptions to guide us. Unfortunately these trials have rarely included LCHF diets. There is much better evidence for the comparison between HCLF diets and what is known as the Mediterranean diet (MD). This isn’t the diet all current Mediterranean’s eat nowadays so the name is perhaps a little unhelpful. It is actually based on historic diets of people living in that geographical area and includes lots of olive oil, lots of oily fish and vegetables, some carbohydrates mostly from non-refined sources, a little alcohol and some red meat. The evidence is very clear that, compared to HCLF, the Mediterranean diet protects high risk patients from cardiovascular disease. So right now, based upon what we know, the Mediterranean diet is probably the healthiest. Despite what devotees might suggest there is simply a paucity of evidence to suggest even medium term health benefits of LCHF.

If you are a scientist who thinks that the LCHF diet is superior to MD then the obvious thing would be to do an RCT between the two to test the hypothesis. A good EBMer would keep quiet about their hypothesis until the data was in. If this is done and the LCHF diet is shown to be superior to MD I will happily change my view. One of the great things about EBM is that when new evidence is made available you are duty bound to change your view. There is no shred of embarrassment or regret, you simply accept the new evidence and move on. Until that time there can be no reconciliation between the LCHF and EBM movements unless the former change their own views to fit with the current evidence.

Future posts in this series will discuss the claims that LCHF proponents actually do make and the types of evidence they use to back up the claims

^{Selected references}

Nordmann AJ et al. Effects of low-carbohydrate vs low-fat diets on weight loss and cardiovascular risk factors: a meta-analysis of randomized controlled trials.Arch Intern Med. 2006 Feb 13;166(3):285-93.

Estruch R. Primary prevention of cardiovascular disease with a Mediterranean diet.N Engl J Med. 2013 Apr 4;368(14):1279-90. doi: 10.1056/NEJMoa1200303. Epub 2013 Feb 25.

Conflicts of interest-

I attended the LCHF summit in Cape Town in 2015 at the kind invitation of Prof Noakes

I make a tiny income from an article on Kindle called "We need to talk about statins"