This is The Guardian. Hi Ian here. We're on a break at Science Weekly, but we'll be back on Tuesday with a new episode. As we emerge from our Christmas food comas, you might know someone whose new year's resolution is to start tracking their glucose as a way to lose weight and feel healthier.
So today we're returning to an episode from July where I tried to get to the bottom of whether these trackers tell us anything useful. I found it a fascinating dive into an area of science that's still up for debate. I hope you enjoy listening. You might have noticed that everyone has recently become a bit obsessed with blood sugar or glucose.
There are blood sugar influencers like the glucose goddess and Justin Richard, aka Stop Spiking Sugar, who will tell you how to keep your blood glucose levels down to stay healthy. If you eat the elements of a meal in a specific order, you can cut the glucose spike by up to 75%. It's crazy. It's crazy. Let's see what beans and rice do to my blood sugar. And yeah, there's the spike right here. You can see that went up by 44 milligrams. That's quite high.
There are also wellness firms that claim to give you insight into how your body processes food, like NutriSense, Levels and Cygnos. Here in the UK, there's one that you've almost definitely heard of. Zoe.
You almost certainly would have seen the little video ad with Tim Spector who's got one of these blood sugar monitors on his arm and he points to it and says something like, this device has changed how I have breakfast forever. This tiny device has changed the way I eat breakfast every day. He's only telling you what's happened to him, but the implication is clear, right? That by wearing this device, you will learn things about your blood sugar response and that will change the way you eat.
The idea is that measuring your blood sugar levels can give you an insight into what's good for you, what foods work for your body, and which ones you might do best to avoid. And while plenty of us are excited by the promise of personalised nutrition, many researchers have begun to question the science behind blood sugar monitoring.
In healthy people, glucose is utterly negligible, I think. The marketing is 15 years at least ahead of the science. So today on Science Weekly, we're asking, what do we really know about blood glucose levels and our health? Is the science nailed down on personalized nutrition? Or are glucose monitors just helping each of us adopt the healthy eating habits we've been told about for years?
I'm the Guardian Science Editor Ian Sample and this is Science Weekly.
Julian Bajini, you're a philosopher, journalist, and author with an interest in food and philosophy. You recently wrote about the rise of personalized nutrition apps and particularly ones based on blood glucose monitoring, the most famous of which here in the UK is Zoe. And one of the things that led you into this is a concept called nutritionism. Tell me about that. The basic idea is that modern nutritional advice has become rather obsessed.
around, you know, breaking things down into particular macro-micro nutrients and specifying how many of each of these you should get. And that's kind of like taken away from the idea that good eating is a more holistic thing. And we've become obsessed with seeing certain biomarkers like blood pressure, cholesterol level, blood sugar.
as not just sort of the indicators for health because they're certainly useful as health indicators, but almost like things which are important in their own right. So for example, if you have a certain blood sugar score, then that's kind of good or bad irrespective of what else is going on in your body. So this was really interesting.
And this really gets to the heart of the personalised nutrition industry, which purports to use data like biomarkers, like your routine, your microbiome, and give you specific advice off the back of that. Although this is an industry that's really growing, it's really early days. How well evidence is it? The problem is that it is a new field, right? Now, if you're doing new stuff, basically a lot of this is kind of research. Some of those are remarketing.
is kind of explicitly telling you that this has got an ongoing study, that they're finding out new things. That's a real problem, because if you're in the process of doing research, right, that's too early to say what the results are. So what the personalized nutrition field, I think, is doing object can't do, is he's trying to do two things at the same time, which has been the cutting edge of research, but already giving people advice. By definition, that advice is bound to be running ahead of the evidence.
What they're looking into is very promising and very interesting. The fundamental premises that all our bodies are different, that our nutritional needs are different, and that we could in theory tailor our dietary advice more to our particularities. All of that is entirely sound, but this is too early in the process to be able to turn in that promise and that premise into actual useful advice.
The current interest in glucose isn't limited to ZOE or the US personalized nutrition company levels. There are endless videos online extolling the virtues of avoiding spikes and explaining how to avoid them. But what's the science behind it all? What do we actually know about how blood glucose impacts our health?
I have oats every morning, so I have a grated apple. Sometimes when blueberries or strawberries are available, I'll chop some up in. I have almonds, raisins, a scoop of protein powder, because I live weights a lot, and soy milk. And I pay zero attention to what my glucose may or may not be doing after I eat them.
My name is Nicola Guess. I am an academic dietician working at the University of Oxford, and my research and clinical expertise is primarily in the prevention and management of type 2 diabetes. Nicola, I think to really dig into this, I want to go back to basics. What actually is glucose?
It's basically an energy source. So it's energy that's in our bloodstream that can then go to the organs that need it. So if you think about our muscles, glucose needs to get into the muscles, the muscles burn it, we oxidize it. Glucose happens to be a very, very useful fuel, particularly for exercise. So it's probably, it's fair to say to the primary fuel that we have in our bodies. So we need glucose, but does it matter if our levels get high?
Well, if they get too high, yes, it does matter. Certainly at the levels of glucose you see in diabetes, yes, it is a problem because that glucose, when it's too high, can cause damage to our tissues. So it can cause damage to our eyes, for example, it can cause damage to our nerves. And it's really important to clarify here.
In diabetes or indeed pre-diabetes, glucose is elevated chronically. So diabetes isn't your glucose going up once to, you know, let's say 14 millimals per liter. Your glucose is elevated constantly. So during the night when you're fasting, it's high. It then goes even higher after you eat and it stays high all the time. And that is what right now we understand to be particularly harmful.
So we know that chronically elevated levels are what's causing damage. But what about in non-diabetic or pre-diabetic people? Because right now there are a lot of health claims going around online that we should be trying to avoid having peaks and troughs in our blood sugar levels. Is that in itself actually damaging? I mean, I like to have porridge for breakfast with loads of fruit in it. So if I get a spike afterwards, does that mean it's bad for me?
It tells us absolutely nothing about the healthiness or not of that meal. And it is perfectly normal based on everything we know right now for a person's glucose to go up quite often in proportion with the amount of carbohydrate they've just eaten. So if you've had a massive pizza, as I've done sometimes myself, yes, your glucose is going to go up really high, your body will release insulin to deal with it, and it will come back down again.
In fact, it is normal for your glucose to go up to 7.8 or 8 millimoles per liter. It probably isn't unusual for your glucose to go up to 11 or 11.1 and come back down again in perfectly healthy people. And can I just say one more thing about what it tells us about that meal or not?
If it's the case, and again, let's put numbers on this and let's say your glucose is normally at a normal concentration. But on what occasion it goes up to 12 or 13. Now we're getting into the realms of glucose where we don't really see it in healthy people, but we don't really know what it means. But it's still the case that if you have a large pizza or if you have some oats and your glucose would get 13.
It's getting to 13 because of damage underneath. So you've probably got insulin resistance and you've probably got a pancreas that isn't working so well. And what will cause insulin resistance and what causes your pancreas to not work properly?
are things that have nothing to do with glucose. It's often being sedentary. It might be not having enough fiber in the diet. Very often it's weight gain. And it's all of those things that we need to fix to fix the glucose. Glucose is merely a symptom of lots of underlying damage, which is why eating low carb doesn't prevent type 2 diabetes because you're not fixing the underlying damage.
It's not just Nicola and Julian who have concerns about using blood glucose levels as a way to offer nutritional advice.
A recent study by scientists at University College London and Birmingham Children's Hospital found that there was little evidence on how accurate continuous glucose monitors are in people not living with diabetes, nor sufficient evidence on what health benefits or utility such information would provide. The paper points out that even the best performing CGM's have a 20% accuracy margin
meaning they could display readings outside the normal range, even when that isn't the case. But companies like Zoe argue that glucose levels alongside a number of other measures can provide useful nutritional insights. We do not advocate constant glucose monitoring. We do a two-week test phase and then participants stop glucose monitoring.
It is just one feature of many, many, many different features that we at ZOE use as part of the dietary advice. We look at people's blood fat levels. We look at people's health history. We consider factors like what's your current diet like? What are your dietary preferences? What are your goals? And then we also look at people's microbiome. And we use all of the research that's published in our algorithms
in order to deliver advice based on all of these different parameters. That's Dr Sarah Berry, Zoey's chief scientist and an associate professor at King's College London. I asked her why they decided to measure blood glucose as part of their programme. So I think that there's value in measuring blood glucose as long as we're looking at it alongside many different
outcomes. I don't think that we should only measure blood glucose in healthy individuals. We know that an increase in circulating blood glucose after carbohydrate containing meal is a normal physiological response. I think that an area that we do need to be cautious about is pathologizing what is a normal response. But what we do know is if this is excessive, repeated, prolonged and fluctuates considerably throughout the day,
that there is now evidence in healthy people that this is associated with increased cardiovascular disease, increased risk of type 2 diabetes, we don't measure it just because we can measure it. Why do you think a lot of researchers are quite cautious when you say things like there being these health effects from glucose going up and down in healthy people? I mean, we've had
the NHS England National Diabetes Advisor Professor Parthakar saying there's no strong evidence that these kind of monitoring glucose in this way helps people without diabetes. We've had the chief medical officer at the American Diabetes Association saying it's not clear whether lowering glucose levels in healthy people
gives you any better health. Kevin Hall, almost legendary figure at the NIH in the US, has said the idea that avoiding glucose spikes can help with weight loss. He says that's controversial and probably wrong. There does seem to be some interest in whether
looking at glucose and looking for the kinds of abnormally high spikes of elevated resting glucose, that might help you spot people who are at high risk of developing diabetes, of becoming pre-diabetic. But there was a lot of caution in the scientific community about the value of these spikes and these measurements in people who are at the time healthy. Yes, I think as scientists,
We are a cautious bunch. I know that I'm always very cautious when I'm thinking about the relationship between diet, disease, how we measure it, what recommendations we make on the back of it. But it's very difficult to prove beyond or reasonable doubt without many, many, many years of research. And I think as scientists, we often need to take more of a pragmatic approach that when we believe the evidence is good enough that we do something about it.
And I, in my opinion, think that the evidence is good enough to suggest that excessive reputed increases in glucose may, in the long term, have unfavorable effects on some health outcomes, not all health outcomes. So for example, we've started looking at dips in glucose. So this is where your glucose levels dip below fasting levels. We know from our own research that's published in ancient capitalism that people that are big dippers
have a reduced feeling of energy, they have an increased feeling of hunger and they actually go on to eat a couple of hundred more calories over the day than those that don't have dips.
While the nature metabolism paper looked across a 24-hour period, it didn't measure physical activity or sleep, which are known to impact on blood glucose levels. There's a multitude of complex factors that affect what we eat, why we eat it, how we respond, and what that means for our overall health. This makes offering genuinely personalized nutrition advice a big challenge. There are so many variables.
Nicola, I'm curious about the personalized nature of using a continuous glucose monitor or CGM. I mean, if you and I ate the same food, would you expect to see the same glucose responses in our blood? So far as we know, people have more or less the same response to the same food. So when we give people like a slice of bread, for example,
On four separate occasions, maybe a weaker part under standardized conditions, the rise in blood glucose is pretty much the same between people. Now, what differs is that people's glucose is starting from a different concentration. So you might see different peaks. The challenge is that the glucose response to any given meal is influenced by things that have nothing to do with what was in that meal.
For example, if you did exercise three days ago, that can influence your glucose response to the breakfast three days afterwards. If you did a really intense bout of exercise the day before, that's going to influence your glucose response to the meal. If you ate lentils versus white rice the evening before, that is going to influence your glucose response to the breakfast or to the meal. So you cannot look at the glucose response to any given meal and assume it's because of what people ate at that meal.
And so that's what I think people are seeing when they are testing thousands of people and seeing different glucose responses. It's because the glucose response was influenced by the millions of different things each of those individuals was doing in the three days before they were tested. There's also a ton of noise in this data. People will be exercising differently, sleeping differently, eating differently,
Females will have their menstrual cycle, and this is what's been challenging is that CGM's are so new. Even in type 1 diabetes, where CGM's have been used for widely, for probably over a decade.
we are just at the point where we have data sets where we can kind of look like, okay, what do these peaks mean? Should we be managing the various different peaks and troughs in different ways? And even that's not certain yet. So we don't quite know what we should be aiming for on a CGM in type 1 diabetes. And I think that says a lot.
So given what you've said, Nicola, what do you make of these companies that offer personalized nutrition guidance off the back of these CGM, these continuous glucose measurements? I mean, normally the way things should work is we think, oh, blood pressure is a real problem. Wouldn't it be great if we could somehow measure 24 hour blood pressure easily? You know, for example, you have a problem and you find a solution for it. What's happened with CGM's?
is we've got this solution. And now we've, we're trying to find a problem. And I think that's why they're marketing these things to people without diabetes. So there's no reason whatsoever that we should be personalizing nutrition based on glucose in a person without diabetes. It's really overregging what glucose can tell us. I mean, to me it's just absurd. If we look at the things
that kill people, it's high cholesterol and it's high blood pressure, so it's heart attacks and strokes. These are the things that are killing people early and causing very poor health. And so this is the stuff that we should be focusing on. If you walked into my clinic and I could choose five things that would tell me about your health, it would be cholesterol and blood pressure, it would be your resting heart rate,
possibly even fasting insulin or insulin and a waste circumference. Those things are going to tell me infinitely more than your glucose if you don't have prediabetes and type 2 diabetes. After what I'd heard from Nicola about the variability of an individual's glucose levels, I was keen to ask Sarah about that and how they take it into account in offering personalized advice. A really important premise of personalized nutrition has been
that the variability there is between one person and another needs to be bigger than the variability there is day to day within a person. Now, we find, and again, this is published research in Aids and Medicine, that the variability between individuals in blood glucose responses is two to three times higher than the variability within a person in blood glucose responses.
And so what this enables us to do is say to someone, look, these are the kind of foods that are giving you a very high glucose response and potentially a dip, whilst also considering these other factors. But we embrace this variability that there is within a person day to day. So we embrace the fact that, yes, sleep is going to impact it.
And within the programme that we offer people, we give lots of advice on this. So it's not only glucose-centric, so that we can offer advice to individuals that we believe will therefore improve their long-term health.
So if you monitor someone for two weeks, say, when they sign up for Zoe, and they'll have these risers and dips and say they have a rise and dip after having their morning porridge, would you then say, look, don't have porridge for breakfast? I mean, is the two weeks of data, given all the other factors that are going to go into what is happening in their physiology, are those two weeks of data enough to
Say don't do this, do that instead. Are you extrapolating a bit too much from the two-week small amount of data you've got? I think that's a good question. It's something that we have spent a lot of time investing in using the ZOE membership, the two-week test phase, as well as the period that people are on the ZOE program in educating people as well.
So join the two week testing period. We have some glucose experiments where we'll ask people, for example, to have their typical breakfast, look at their glucose response, try the next day adding some extra protein or healthy fats or fiber to it and we'll give them personalized advice on that depending on what the breakfast is. That also adds additional healthy nutrients to their meal. Another important point to make though regarding that is we know from other published research that
particularly in the UK and the US, we eat from actually a very narrow group of meals. So actually in a two-week testing period, you do get a really good idea of people's responses to most of the types of meals. So is there any evidence then that this kind of personalised advice that you're giving is better for people's health than standard dietary advice?
So when we think about personalization, I think we need to think about it in two ways. So how effective is delivering advice that's personalized to people's biology in bringing about change in health measures? But also, and this is really important, how does the delivery of advice in a personalized way that people believe is personalized actually bring about better adherence to following that advice?
So we recently published a study where we looked at how our ZOE program compares to standard dietary advice. And what we found from this study is there was significant improvements in a number of health outcomes for people following the ZOE personalised programme compared to standard care advice.
But Sarah, this paper didn't isolate the effect of the personalized advice. I mean, one group followed the ZOE program with the tests, the CGM monitor, and they had to log their food intake. Whereas the other group
just got standard advice from a leaflet and a weekly follow-up email. So you don't know if it was just the way the advice was given that made the difference. What you could have done is blind the study, put both groups through your programme, but give one of them personalised advice and the other one the standard advice. So this is our first study that we wanted to conduct to firstly see, does the programme
which includes the biological personalization as well as personalized delivery of advice, does that improve health outcomes compared to standard care? Now that we've shown that it does, that's when we can now dive deeper. And we've actually got a large longitudinal study going on at the moment, looking at many different kind of features of personalization in order to disentangle which has the biggest effect.
So how would you respond to the criticism that your use of glucose monitors in giving personalized dietary advice is very far ahead of where the science actually is? So I think that is fair to say that we are slightly ahead of the curve within the timeframe of how we have traditionally actioned evidence. This is because of scientists
we want to have all of the evidence before we actually put anything into action. But I think that we have to be brave sometimes and take a call based on the body of evidence that there is here and now. If we wait until we are 100% certain on anything in science, we will never, ever change anything or deliver updated advice in my opinion.
Nikola, whatever the shortcomings of these personalized nutrition programs, some people will say, look, they still have health benefits because they make people think more about eating healthily. What do you make of that line of argument? Oh, I think that's a great point. And I think that's a really fair point. If a CGM can help you stick to a healthier diet, then that's a great thing. But I think it's really important to emphasize a CGM
isn't going to tell us anything we don't know about what a healthy diet is. And if you look at how we eat, if you look at the lives we live, we are largely sedentary. We eat a bunch of high-sugar, nutrient-poor foods. We don't eat many whole plants. But that is the stuff that I think most reasonable people think.
is 90 to 99% of the reason we are metabolically and in the health that we are. Could things like personalizing based on glucose or the microbiome explain a minuscule part of this? I would bet everything I own on it never being clinically useful. And you can come back to me in a few years if I've got that wrong.
Whether or not personalized nutritional advice makes a difference to our health, one thing's for certain, and that's that we love monitoring ourselves, but it's having all this data at our fingertips really good for us. It's something I put to Julian. I mean, this idea of the quantified self is a phrase that has been used in recent years, the idea that we're measuring
everything about ourselves, hours of sleep, steps taken and it seems how could it not be a good thing, it's all more information.
But I don't think it is obviously a good thing because the point is this, that we are hugely complex organisms. And the more I look into health, food and nutrition, the more I think we need a much, much more humility about it. People always go back to that simple principle of Michael Pollan, you know, just eat food, mainly plants.
Not too much you know that's kind of it you know and other advice is the same move around you do exercise be active you know to sort of like Think we should be more precise like that and specify the number of steps or how many grams of this or how many milliliters of that just I think it is unhealthy you become neurotic but also it also tends to lead to
obsessive behaviors. The other things I should say about this there as well is what does it mean to be personalized? I think this is a really important, almost philosophical issue. A good GP would already give you personalized advice, right? They look at your weight, your height, your blood test, your lifestyle, what they know about you. And they recommend things that they think that you can do in your life, which will help you. So it's not like current advice isn't personalized.
Generally speaking, the most healthy, happiest people are those who take care of themselves without worrying too much. I think we should all take a collective chill pill and be a bit more relaxed.
And that's it for today, this episode was produced by Madeleine Finley. It was sound designed by Joel Cox, and the executive producer is Eddie Burey.
Science Weekly Tea is planning an episode all about how to get healthy in 2025. And for that, we need you to send us your questions. It could be anything you're curious about, from whether intimate and fasting really helps us stay young, or time of day is best to hit the gym, in which exercises will help most. We'll choose our favourites and bring you answers backed up by science. Just email Science Weekly at TheGuardian.com. And we'll be back on Tuesday with a new episode. See you then!
