Personalised nutrition from a poo sample

It's a bit shit.

Globally 2 billion adults are overweight. In Western countries, the prevalence of overweight or obesity reaches two-thirds of the population, and in many countries more than a third of people have blood glucose levels which are above the healthy range. People with obesity are discriminated against in the workplace, experience structural discrimination and weight stigma and unsurprisingly are at higher risk of depression.  At any given time, nearly half of people are trying to lose weight. Although within any trial or any weight loss intervention there are people who can and do lose weight and keep it off, for a large number of people, sustained weight loss is incredibly hard.

Little wonder that any new intervention comes along promising to help people finally “get control” of their weight, people are interested. And willing to pay lots of money for it. 

Photo by Diana Polekhina on Unsplash

As scientists and healthcare professionals we have a moral obligation and professional requirement to present the evidence for weight loss and type 2 diabetes interventions in an honest, and unbiased way. No healthcare professional code of conduct would permit you to sell a supplement or treatment without an evidence base behind it.  

This is why I have am very concerned with the new field of so-called personalised nutrition. To be clear, personalised nutrition has always been part of good clinical practice in the sense that any dietary advice should be personalised based on a person’s needs and preferences. 

However, a brand of personalised nutrition on the rise makes a difference claim. It’s based on the premise that individuals have inherently different biological responses to foods or nutrients. There is definitely a biological rationale for thinking that people may indeed respond differently. For example, people with type 2 diabetes appear to have a greater insulin secretory response to protein than people without type 2 diabetes. Similarly, riboflavin (vitamin B2) may only lower blood pressure in people with a specific genetic make-up.

So personalised nutrition based on biological differences has some legitimacy behind it and for good reason is a subject of active investigation across major national scientific funders. However, in my opinion, these companies are taking what is scientific plausible and prematurely presenting it as a panacea for weight loss or type 2 diabetes. And, ABSOLUTE YAWN, their claims of course involve the microbiome.

One company claims “By avoiding dietary inflammation and improving your gut health, our program can help to unlock energy, manage hunger, manage your weight”. Yet looking at their publications, the data is cross-sectional. There is no interventional data to answer the all important claim of whether a diet devised based on “avoiding inflammation” and “your individual microbiome” causes better weight loss than an otherwise healthy and varied diet. 

The website also reports results from a pilot study which found that 86% of users experienced increased energy, a reduction in hunger between meals, weight loss and improved bowel movements. Yet every scientist knows that any intervention will help people lose weight. Part of this is the long understood Hawthorne effect: essentially if people know they are being observed, they change their behaviour. Even asking people to track what they eat can lead to changes in dietary intake. So these claims are a marketing tool, not evidence. 

Other companies backed by academics have carried out trials but to date these are not convincing. Zeevi et al developed an algorithm to predict individual glycaemic responses to foods based on an array of data including post-prandial glycaemic responses, lifestyle factors (sleep, activity) and microbiome analysis collected in 800 people. The investigators describe the significant inter-individual differences they observed in the glucose response to the same foods. Indeed, this is the primary selling point of these personalised nutrition products and services: our unique biology means we react very differently to the same foods, and therefore the “best”diet for a person’s metabolism may differ from another persons. 

Yet, the data from their predictive algorithm tells a different story when they tested it in a pilot clinical trial. The algorithm recommended that all individuals consume more high-protein foods and intact grains. The algorithm also recommended that all individuals reduce high-fat and high-glycaemic carbohydrates.

The predictive algorithm also recommended a high-protein, reduced carbohydrate diet when it was tested in a subsequent trial in people with type 2 diabetes. 

In contrast to the “personalised based on your unique metabolism” approach promised on glossy websites, it sure looks like this algorithm is recommending a pretty specific diet to everyone. This makes sense when you base a diet entirely on the glucose response because dietary protein lower glucose, and a high-protein, reduced-carbohydrate diet lowers glucose even in people with type 2 diabetes. What clinical utility the gut microbiome plays in all this is unclear. 

A recently published randomised control trial by a separate research consortium concluded that “against a background of a mostly plant-based and generally healthy diet, biomarker driven personalization of the diet does not further improve body weight homeostasis, body composition, and cardiometabolic risk factors”.

In summary - the evidence behind this GIVE US YOUR POO AND OUR CUTTING-EDGE SCIENCE WILL DO THE REST is….[based on the data currently available] …pretty shit. 

My advice: save your money and spend it on fresh produce, a gym membership or a nice holiday!