Personalized nutrition: why there is no one right diet for everyone
Two people, the same meal, two entirely different blood sugar curves. What sensor data, gut bacteria, genes and life stage can reveal about what suits you.
More articles from the nutrition cluster
Imagine two people sitting side by side at the kitchen table. Both are wearing a small glucose sensor on their upper arm. Both eat the same ripe banana in the morning.
For one person, the curve shoots steeply upward. A little over an hour later, it falls again, even below the starting level. Shortly afterwards, hunger kicks in and concentration fades.
For the other person, the curve barely moves. A gentle hill, then calm. This person does not eat again until lunch and feels alert.
The same banana. Two bodies. Two stories. Which of them ate the right way? Both. And that is exactly what this article is about.
This is an illustrative example, not a real patient story; I cannot infer causality from it.
In my practice, I often meet people who have been driven to despair by the search for the right diet. They followed plans that were supposed to be the same for everyone. Maybe the problem was not the person, but the idea that one plan fits everyone.
Personalized nutrition sounds like high tech and apps. At its core, it is an old medical question: What does this particular person need? What is new is that today we can measure how differently bodies respond to the same food.
In this article, we look at what large sensor studies show. We examine the role that gut bacteria, genes, sleep and life stage play. And we are honest about where the research is still thin.
What to expect
- Why the same meal can produce two curves
- What 46,898 measured meals show
- How your gut microbiome eats with you
- What genes explain and what they don't
- Why sleep and menopause can shift blood sugar
- What personalized programs achieve in studies
- KPNI: systems thinking instead of diet dogma
- Three levers for your everyday life
The same meal, two curves
Does this sound familiar? You eat a muesli that is considered healthy. An hour later, you are tired and already hungry again. The person next to you eats the same thing and stays full until lunch.
Maybe you concluded that you lack discipline. Research from the past ten years suggests a kinder explanation: Bodies respond differently to the same food.
A team led by Segal and Elinav followed 800 people for one week with a glucose sensor, 46,898 meals in total. Participants responded very differently to identical meals. A model built from blood values, eating habits, body measurements, physical activity and gut bacteria predicted the curves, including those of 100 additional people. What this means for you: A general list of good and bad foods may only partly fit you.
Zeevi D, Korem T, Zmora N et al. Cell. 2015;163(5):1079-1094. DOI: 10.1016/j.cell.2015.11.001The same paper included a small, blinded, randomized trial. Meal plans tailored to the predicted curves were associated with lower values after eating. The researchers cautiously conclude that general recommendations might be of only limited use.
The PREDICT 1 study examined 1,002 twins and unrelated adults from the UK using identical test meals. The variation between people was 103 percent for blood fats, 68 percent for blood sugar and 59 percent for insulin. A prediction model passed the test in 100 people in the US. What this means for you: How your body responds to a meal is to a large extent personal.
Berry SE, Valdes AM, Drew DA et al. Nat Med. 2020;26(6):964-973. DOI: 10.1038/s41591-020-0934-0What is interesting is what explained the differences in PREDICT 1. The answer depends on the measure.
The nutrients in the meal explained 15.4 percent of the differences. Personal factors such as the gut microbiome explained 6.0 percent.
Here it was the other way around. Personal factors such as the microbiome explained 7.1 percent, the nutrients in the meal 3.6 percent.
A second look matters. For blood sugar, what is on your plate still counts for more than the microbiome. And no single factor explains the whole curve. It is an interplay of many small contributions.
Researchers at the Mayo Clinic followed 327 people without diabetes from Minnesota and Florida for six days with a sensor. A personalized prediction model matched the curves better than counting carbohydrates or calories, with a correlation of 0.62 versus 0.40 and 0.34. The company DayTwo, which offers such models, was involved. What this means for you: Counting carbohydrates shows you only part of the picture.
Mendes-Soares H, Raveh-Sadka T, Azulay S et al. JAMA Netw Open. 2019;2(2):e188102. DOI: 10.1001/jamanetworkopen.2018.8102The team led by Snyder at Stanford University gave participants standardized meals and took measurements with a sensor. Even people with unremarkable standard values spent 15 percent of the time in the prediabetes range and 2 percent in the diabetes range. Behind this were different patterns of insulin resistance and insulin secretion. What this means for you: A normal fasting value does not tell the whole story of your day.
Hall H, Perelman D, Breschi A et al. PLoS Biol. 2018;16(7):e2005143. DOI: 10.1371/journal.pbio.2005143At Tufts University, 63 healthy adults aged 18 to 85 repeatedly ate the same white bread. The glycemic index averaged 62, but varied by 20 percent within the same person and by 25 percent between people. The researchers therefore do not consider the index a good guide for choosing foods. What this means for you: A GI table gives you a rough direction, but not your personal curve.
Matthan NR, Ausman LM, Meng H et al. Am J Clin Nutr. 2016;104(4):1004-1013. DOI: 10.3945/ajcn.116.137208From the perspective of Clinical Psychoneuroimmunology, KPNI for short, I look at metabolism first here. A steep curve usually means that a lot of sugar enters the blood in a short time. The body responds with insulin, and the cells take up the sugar. If the curve then falls below the starting level, this is called a dip.
A single rise is normal and no reason to worry. It becomes interesting when a pattern repeats every day and you regularly feel tired or hungry afterwards.
What can you do with this? Observe first, instead of judging. Two to three hours after a meal, pay attention to how you feel: alert and full, or tired and restless. That is your first, free measuring instrument.
Supported by large sensor studies such as those from the Weizmann Institute and PREDICT 1: People respond very differently to identical meals. Personalized models predict blood sugar curves better than carbohydrates or calories alone.
Mechanistically plausible, human studies still thin: Whether every single spike causes long-term harm in people without diabetes is not well established so far. Most of the data are observations over days to weeks.
In my clinical practice, I observe that many people relate to food in a more relaxed way once they know their own patterns. This is an observation from my practice, not a study result.
And now you know why your muesli can affect you differently than it affects the person next to you.
Your gut eats with you
Maybe you know this from your own family. Whole grain bread agrees with you, but it sits heavy in your sister's stomach. Legumes keep you full for a long time, while others feel bloated afterwards.
Part of the explanation lives in your gut. Trillions of bacteria help process what you eat. And this community is composed differently in every person.
In random order, participants ate industrial white bread and sourdough whole grain bread for one week each. On average, clinical values did not differ between the two breads, and the microbiome remained personal and stable. Looked at individually, however, people responded differently, and which bread produced the lower curve in whom could be predicted from the microbiome data alone. What this means for you: An average of zero can hide two opposing groups.
Korem T, Zeevi D, Zmora N et al. Cell Metab. 2017;25(6):1243-1253.e5. DOI: 10.1016/j.cmet.2017.05.002A team led by Bäckhed had adults eat bread made from barley kernels for three days. In some, glucose metabolism improved, in others hardly at all. Those who benefited had more Prevotella, especially Prevotella copri, and a greater potential to break down complex fiber. Germ-free mice given the gut flora of these people also showed more favorable glucose metabolism and more glycogen in the liver.
What this means for you: Whether fiber has a favorable effect on your blood sugar may depend on your gut bacteria.
Kovatcheva-Datchary P, Nilsson A, Akrami R et al. Cell Metab. 2015;22(6):971-982. DOI: 10.1016/j.cmet.2015.10.001Researchers led by Segata analyzed 1,203 microbiomes from 1,098 people in PREDICT 1. The strongest associations appeared with healthy and plant-based foods. Prevotella copri and Blastocystis were associated with more favorable blood sugar levels after meals, and the composition partly predicted blood sugar, blood fat and inflammatory markers. What this means for you: Eating habits, gut flora and curves are linked, even if the direction is not yet entirely clear.
Asnicar F, Berry SE, Valdes AM et al. Nat Med. 2021;27(2):321-332. DOI: 10.1038/s41591-020-01183-8When gut bacteria ferment fiber, short-chain fatty acids such as butyrate, propionate and acetate are produced. These substances can nourish the gut lining and send signals to the liver, pancreas and brain.
Which ones are produced, and how much, depends on which bacteria are present. This could explain why the same portion of fiber plays out differently in two bellies.
The immune system also has a seat at the table. A large share of your immune cells lives in and beneath the gut lining. In PREDICT 1, the composition of the gut flora was also linked to inflammatory markers in the blood. So a diverse gut flora might matter for more than just your blood sugar.
What does this mean for your everyday life? Feed diversity. Many different plants each week, meaning vegetables, legumes, nuts, seeds, whole grains and herbs, provide food for different bacteria. If your gut is sensitive, increase fiber slowly. You can find more on this in the article on 30 grams of fiber a day.
Supported by the Weizmann Institute bread study and by PREDICT 1: The microbiome is linked to personal blood sugar responses and can partly predict them.
Mechanistically plausible, human studies still thin: The idea that individual bacteria such as Prevotella copri directly improve blood sugar in humans rests so far mainly on animal models and small studies. That is why a microbiome test cannot give you a reliable meal plan yet.
In my clinical practice, I observe that people with a sensitive gut often tolerate fiber better when they increase the amount slowly over weeks. This is an observation from my practice, not a study result.
And now you know why the same whole grain bread can write two different stories in two bellies.
Genes: less destiny than you might think
Maybe you have already considered taking a genetic test for your diet. The ads promise a diet that matches your DNA. The idea is tempting: a clear answer, in black and white.
The evidence is sobering, and that is good news. In PREDICT 1, hereditary factors contributed only a little to the prediction: 9.5 percent for blood sugar after meals, 0.8 percent for blood fats, 0.2 percent for C-peptide. So your genes play along, but they do not call the tune.
I describe in detail how ancestry and genes can shape metabolism in the article on genes, ancestry and metabolism.
609 adults aged 18 to 50 with a BMI of 28 to 40 followed a low-fat or a low-carbohydrate diet for twelve months. One group lost 5.3 kg on average, the other 6.0 kg, a difference of 0.7 kg without statistical certainty. Neither a gene pattern nor insulin secretion predicted who lost more weight on which diet. What this means for you: Both approaches can work, and a genetic test does not reliably pick the right one for you.
Gardner CD, Trepanowski JF, Del Gobbo LC et al. JAMA. 2018;319(7):667-679. DOI: 10.1001/jama.2018.0245In seven European countries, participants received nutrition advice online, either general or personally tailored. After six months, those who received personalized advice ate somewhat less red meat, salt and saturated fat, and their healthy eating index rose slightly. Additional blood values or genetic data brought no further benefit. What this means for you: Personal guidance can count, but the genetic data did not make the difference.
Celis-Morales C, Livingstone KM, Marsaux CF et al. Int J Epidemiol. 2017;46(2):578-588. DOI: 10.1093/ije/dyw186Researchers led by Marteau analyzed 18 studies in which people learned about their genetic disease risk. For eating behavior, there was at most a very small, uncertain effect, and none for smoking or physical activity. Motivation did not increase, but neither did anxiety or low mood. What this means for you: A genetic test alone hardly gets anyone to eat differently.
Hollands GJ, French DP, Griffin SJ et al. BMJ. 2016;352:i1102. DOI: 10.1136/bmj.i1102There are exceptions, and they matter. For certain inherited conditions, such as very high cholesterol that runs in families, genetic testing makes medical sense. But that is something different from a nutrition DNA test from the internet.
If you are considering such a test, ask yourself first: Which decision would I make differently based on the result? If the answer is unclear, your money is often better spent on good guidance or on two weeks with a sensor and a clear question.
Supported by DIETFITS, Food4Me and a meta-analysis from Cambridge: Gene-based nutrition recommendations brought no measurable added benefit over general or personal advice in these studies.
Mechanistically plausible, human studies still thin: Individual gene variants can influence how you process lactose, caffeine or fats. Whether this leads to better meal plans is still open.
In my clinical practice, I observe that results from nutrition genetic tests tend to unsettle some people rather than guide them. This is an observation from my practice, not a study result.
And now you know why a genetic test cannot give you a ready-made diet today.
Sleep, menopause and the dip afterwards
Do you know those days when the same breakfast does not carry you through? Often they are the days after a short or restless night. That is not your imagination. Your curve depends not only on who you are, but also on how you are doing right now.
Researchers led by Franks, Walker and Spector compared sleep data from 953 people, 41 percent of them twins, with more than 8,000 meals tracked by sensor. Poorer sleep efficiency and a later bedtime were associated with higher blood sugar levels after breakfast the next morning. Deviations from a person's own usual sleep pattern were also linked to less favorable control. What this means for you: Your blood sugar in the morning also tells you something about your night.
Tsereteli N, Vallat R, Fernandez-Tajes J et al. Diabetologia. 2022;65(2):356-365. DOI: 10.1007/s00125-021-05608-yThe team led by Bermingham compared women before, during and after menopause. After menopause, fasting blood sugar was 6 percent higher, HbA1c 5 percent higher and the inflammatory marker GlycA 4 percent higher than before. The curves after meals were less favorable, sleep was poorer, and the women ate 12 percent more sugar. What this means for you: What worked well at 35 may look different at 55.
Bermingham KM, Linenberg I, Hall WL et al. EBioMedicine. 2022;85:104303. DOI: 10.1016/j.ebiom.2022.104303Two details from this paper are important. Even compared with women of the same age, the blood sugar peak after menopause remained 4 percent higher. And women on hormone therapy had more favorable values for belly fat and blood fats. This is an association, not proof that the hormones had an effect. Some of the differences were mediated by diet and gut bacteria.
The team led by Wyatt and Berry analyzed 8,624 standardized and 71,715 freely chosen meals from 1,070 people. When blood sugar fell below the starting level two to three hours after eating, hunger showed up more often, the next meal came sooner, and more was eaten over 24 hours. The associations were weak, with correlations between 0.14 and 0.27. What this means for you: Hunger after a dip is a body signal and not a sign of weakness.
Wyatt P, Berry SE, Finlayson G et al. Nat Metab. 2021;3(4):523-529. DOI: 10.1038/s42255-021-00383-xThis is where two more systems come into play. The hormonal system: Estrogen can influence how sensitively cells respond to insulin. When it drops during menopause, the curves can shift. You can read more about this in the article on nutrition, hormones and blood sugar.
The nervous system: After a bad night, the stress system is often more active. Cortisol and the sympathetic nervous system can raise blood sugar further. Your internal clock keeps time as well, as I explain in the article on morning light and your internal clock.
What can you do? Take your sleep seriously as a nutrition topic. Regular bedtimes, light in the morning and a calm evening are part of it. If you often get cravings two to three hours after eating, try a meal with more protein, vegetables and healthy fats. Notice whether your afternoon changes.
Supported by analyses of the PREDICT studies on sleep, menopause and dips: Sleep, life stage and the pattern after meals are linked to blood sugar curves and hunger.
Mechanistically plausible, human studies still thin: All three papers are observational data. They do not show whether better sleep or hormone therapy directly improves the curves.
In my clinical practice, I observe that afternoon cravings often change when breakfast and lunch contain more protein. This is an observation from my practice, not a study result.
And now you know why the same meal can play out differently on two days or in two stages of life.
What personalized programs show in studies
If people respond so differently, a personal plan ought to win clearly. That sounds logical. The randomized trials paint a more mixed picture, and that is exactly what you should know before you spend money.
225 people with prediabetes ate according to a personalized algorithm or a Mediterranean diet for six months, followed by six months of follow-up. Time above 140 mg/dL fell by 1.3 hours per day versus 0.3 hours, and HbA1c by 0.16 versus 0.08 percentage points. The difference persisted after twelve months, while the glucose tolerance test showed none. What this means for you: A measurable but small advantage.
Ben-Yacov O, Godneva A, Rein M et al. Diabetes Care. 2021;44(9):1980-1991. DOI: 10.2337/dc21-0162In a blinded design, 23 people with newly diagnosed type 2 diabetes followed a personalized or a Mediterranean diet for two weeks each. On the personalized plan, blood sugar spent 2.42 fewer hours per day above 140 mg/dL, and the mean was 7.8 mg/dL lower. In a subsequent phase without a control group, HbA1c fell by 0.39 percentage points in 16 participants after six months. What this means for you: An interesting signal, but a small pilot trial is not proof.
Rein M, Ben-Yacov O, Godneva A et al. BMC Med. 2022;20(1):56. DOI: 10.1186/s12916-022-02254-y204 adults with obesity and impaired glucose metabolism followed a personalized or a low-fat diet for six months. The low-fat group lost 4.31 percent of their body weight, the personalized group 3.26 percent, a difference without statistical certainty. There was also no difference in body composition or in the adaptation of energy expenditure. What this means for you: For weight loss, the personalized plan was not superior in this study.
Popp CJ, Hu L, Kharmats AY et al. JAMA Netw Open. 2022;5(9):e2233760. DOI: 10.1001/jamanetworkopen.2022.33760347 people aged 41 to 70 followed a personalized nutrition program or the general US dietary guidelines for 18 weeks. Triglycerides fell by 0.13 mmol/L more in the program, while there was no reliable difference in LDL cholesterol. Weight, waist circumference, HbA1c, diet quality and gut flora developed more favorably, especially with high adherence, whereas blood pressure, insulin and glucose did not. What this means for you: Small advantages, especially for people who stick with it.
Bermingham KM, Linenberg I, Polidori L et al. Nat Med. 2024;30(7):1888-1897. DOI: 10.1038/s41591-024-02951-6| Study | Who | Comparison | Result |
|---|---|---|---|
| Ben-Yacov 2021 | 225 people with prediabetes | personalized or Mediterranean, 6 months | less time above 140 mg/dL, HbA1c lowered slightly more |
| Rein 2022 | 23 people with type 2 diabetes | personalized or Mediterranean, 2 weeks each | less time above 140 mg/dL, pilot trial only |
| Popp 2022 | 204 people with obesity | personalized or low-fat, 6 months | no reliable difference in weight |
| Bermingham 2024 | 347 people, aged 41 to 70 | program or US guidelines, 18 weeks | triglycerides slightly lowered, no difference in LDL |
| DIETFITS 2018 | 609 people with overweight | low-fat or low-carbohydrate, 12 months | similar weight loss, genes predicted nothing |
Ordovas, Ferguson, Tai and Mathers summarized the state of personalized nutrition. Their conclusion: Tailored approaches are promising, but more research is needed. What this means for you: Personalization is a tool in development, not a finished product.
Ordovas JM, Ferguson LR, Tai ES, Mathers JC. BMJ. 2018;361:k2173. DOI: 10.1136/bmj.k2173A word on competing interests. Several PREDICT analyses and the 2024 program study were carried out with the involvement of the company ZOE, which offers personalized nutrition programs. The Mayo Clinic study was carried out with DayTwo. That does not make the data wrong, but you should know it when you read advertising claims.
Supported by several randomized trials: Personalized plans can lower certain blood sugar and blood fat values slightly more than general recommendations. For weight loss, no clear advantage has emerged so far.
Mechanistically plausible, human studies still thin: No study has shown so far whether personalized nutrition prevents diabetes, heart attacks or strokes in the long term. The available studies are small or short.
In my clinical practice, I observe that the biggest gain often lies not in the algorithm, but in people paying attention to their bodies again. This is an observation from my practice, not a study result.
And now you know why personalized nutrition is a tool and not a magic bullet.
KPNI: systems thinking instead of diet dogma
Maybe you have already tried a few diets. Low carb, low fat, intermittent fasting, vegan, paleo. For each one, there were people who swore that this was the way. For you, it sometimes worked and sometimes did not.
Clinical Psychoneuroimmunology asks a different question. Not: Which diet is right? But: How are the major systems working together in this person right now? A meal is then not an item on a list, but a signal to all four.
Four systems, one meal
- Metabolism. How steeply does your blood sugar rise, and how does it fall again? How full are you after three hours?
- Immune system. How does your gut respond? Do you get bloating, abdominal pain or skin reactions after certain foods?
- Hormonal system. Which stage of life are you in? Your cycle, pregnancy, menopause and thyroid can shift the response.
- Nervous system. How did you sleep, how stressed are you, and do you eat in peace or while doing something else at a screen?
Freedom here does not mean finding the perfect diet. It means asking your body questions again and taking its answers seriously.
The old question is: Which diet is the right one? The new question is: Which meal carries me, in this stage of life, with this sleep, with this gut?
The answer is allowed to change over the course of your life. That is not failure. It is biology.
General rules are not worthless in all this. In DIETFITS, both groups lost weight. In the prediabetes study, the Mediterranean diet also lowered HbA1c slightly. I describe why the Mediterranean diet can be a good foundation in the article on the Mediterranean diet in studies and everyday life.
There are good basic rules for everyone and fine-tuning for you. Vegetables, legumes, whole grains, nuts and little ultra-processed food are a solid foundation. You find the fine-tuning by observing.
How do you start? With one system, not all of them at once. If you sleep poorly, sleep is your first lever. If your gut is rebelling, you start with the gut. If your curves are steep, you look at the composition and order of your meals.
Supported by the randomized trials in this article: Healthy general eating patterns can bring measurable benefits, and personal differences in the response to food are well documented.
Mechanistically plausible, human studies still thin: Whether a systemic view, as KPNI proposes, leads to better results than a standard diet has not yet been tested directly in studies.
In my clinical practice, I observe that people are more likely to stick with a plan when it fits their life, not the other way around. This is an observation from my practice, not a study result.
And now you know why I would rather ask your body questions than prescribe you a diet.
Three levers you can start with
You do not need an expensive program to get started. These three levers cost little and still give you a lot of information about yourself.
Test instead of believing
For two weeks, notice how you feel two to three hours after eating. A sensor can add to this if you have a clear question.
Feed diversity
Vegetables, legumes, nuts, seeds, whole grains and herbs in a changing selection provide food for many different gut bacteria.
Protect your sleep
Regular bedtimes and a calm evening are part of your nutrition, even if they are not on your plate.
Before you buy a test or sensor
- A clear question. What do you want to know? For example: How do I respond to my usual breakfast? Without a question, a sensor mostly delivers numbers.
- Guidance. Curves alone can be unsettling. Discuss your data with someone who can put it into medical context.
- Be skeptical of genetic tests. For choosing a diet, genetic data has so far brought no added benefit in studies.
- Check for competing interests. Who sells the product, and who funded the studies used in its advertising?
- Diabetes or medication. If you take insulin or other medication that lowers blood sugar, have any major change in diet medically supervised from the start, because low blood sugar episodes are possible.
- Your relationship with food. If food is tied to control or anxiety for you, constant measuring can be a burden. In that case, a sensor is probably not the right place to start.
I describe what two weeks with a sensor can look like in everyday life in the article on 14 days with a glucose sensor. And you can read how the order on your plate can influence the curve in keeping blood sugar stable.
And now you know why the best start is often a good question and not a new plan.
Frequently asked questions about personalized nutrition
What is personalized nutrition?
Personalized nutrition does not tailor recommendations to an average person, but to you. It takes into account how your blood sugar responds to meals, which bacteria live in your gut, how you sleep and which stage of life you are in. The goal is a way of eating that fits your body and your everyday life.
Is there one right diet for everyone?
The studies argue against it. In PREDICT 1, with 1,002 people, participants responded very differently to identical test meals, and the variation in blood sugar was 68 percent. There are, however, good basic rules that may make sense for almost everyone: plenty of vegetables, legumes, whole grains, nuts and little ultra-processed food. The fine-tuning is personal.
Why does my blood sugar respond differently from other people's?
Many factors interact: the composition of the meal, your gut bacteria, your sleep, your physical activity and your stage of life. At the Weizmann Institute, a model built from blood values, eating habits, body measurements, physical activity and gut bacteria predicted the curves of 800 people. In PREDICT 1, genes contributed only 9.5 percent to the prediction of blood sugar after meals.
Do I need a glucose sensor?
Not necessarily. You can already learn a lot by noticing how you feel two to three hours after eating. A sensor can make sense if you have a clear question and discuss the data with someone who can put it into medical context. If you take insulin or other medication that lowers blood sugar, have any major change medically supervised from the start.
Are genetic tests for nutrition useful?
So far, hardly at all when it comes to choosing a diet. In DIETFITS, with 609 people, neither a gene pattern nor insulin secretion predicted who lost more weight on a low-fat or a low-carbohydrate diet. In Food4Me, genetic data brought no additional benefit over personal advice. It is different for certain inherited conditions, such as very high cholesterol that runs in families, where genetic testing can make medical sense.
How important is the microbiome?
It is an important player, but not the only one. In the Weizmann Institute's bread study, microbiome data alone could predict whether white bread or sourdough whole grain bread produced the lower curve in a given person. In PREDICT 1, personal factors such as the microbiome explained 7.1 percent of the differences in blood fats and 6.0 percent in blood sugar. A microbiome test, however, cannot yet give you a reliable meal plan.
Is the glycemic index useless?
Not useless, but less precise than many people think. At Tufts University, 63 adults repeatedly ate the same white bread. The index averaged 62, but varied by 20 percent within the same person and by 25 percent between people. A GI table can give you a rough direction, but it does not show you your personal curve.
Why do I get cravings two to three hours after eating?
One possible explanation is a dip: After rising, blood sugar falls below its starting level. In PREDICT 1, with 1,070 people, hunger showed up more often afterwards, the next meal came sooner, and more was eaten over 24 hours. The associations were weak, though. A meal with more protein, vegetables and healthy fats may be worth a try.
Does my metabolism change during menopause?
It can. In PREDICT 1, fasting blood sugar was 6 percent higher, HbA1c 5 percent higher and the inflammatory marker GlycA 4 percent higher in women after menopause than in women before it. Even compared with women of the same age, the blood sugar peak after menopause remained 4 percent higher. These are associations from observational data, not proof of a single cause.
What role does sleep play in blood sugar?
A bigger one than many people think. In a PREDICT analysis with 953 people and more than 8,000 meals, poorer sleep efficiency and a later bedtime were associated with higher blood sugar levels after breakfast the next morning. These are observational data. Regular bedtimes and a calm evening can still be a good place to start.
Is personalized nutrition better for weight loss?
A clear advantage has not been shown so far. In a study from New York with 204 people, the low-fat group lost 4.31 percent of their body weight and the personalized group 3.26 percent, with no statistically reliable difference. In a program study with 347 people, weight developed somewhat more favorably in the program, especially with high adherence. What often matters is whether you can stick with a plan in the long run.
How do I get started?
With a question instead of a new plan. For two weeks, notice how you feel two to three hours after eating. Feed diversity with vegetables, legumes, nuts, seeds, whole grains and herbs, and take your sleep seriously. A sensor can come later, once you have a clear question and someone to help you interpret your data.
How this article connects to others
Personalized nutrition
Why there is no one right diet for everyone and what you can observe instead.
this article14 days with a glucose sensor
How a sensor in everyday life can show which meals make your curve steep.
Genetics and ancestry
What genes and ancestry can say about your metabolism and where the limits lie.
The order you eat in
Why the same meal eaten in a different order can show a flatter curve.
Hormones and blood sugar
How estrogen and cortisol can help determine how you respond to food.
30 grams of fiber
How you can feed your gut bacteria with diversity on your plate.
Satiety and protein
How you can learn to read your body's hunger and fullness signals again.
A calorie is not just a calorie
Why equal amounts of calories can be processed differently in the body.
Scientific sources
- Zeevi D, Korem T, Zmora N et al. Personalized Nutrition by Prediction of Glycemic Responses. Cell. 2015;163(5):1079-1094. DOI: 10.1016/j.cell.2015.11.001 · PMID: 26590418 [Cohort with randomized intervention, n=800, validation n=100]
- Berry SE, Valdes AM, Drew DA et al. Human postprandial responses to food and potential for precision nutrition. Nat Med. 2020;26(6):964-973. DOI: 10.1038/s41591-020-0934-0 · PMID: 32528151 [Cohort, PREDICT 1, n=1,002, validation n=100]
- Mendes-Soares H, Raveh-Sadka T, Azulay S et al. Assessment of a Personalized Approach to Predicting Postprandial Glycemic Responses to Food Among Individuals Without Diabetes. JAMA Netw Open. 2019;2(2):e188102. DOI: 10.1001/jamanetworkopen.2018.8102 · PMID: 30735238 [Cohort, n=327]
- Hall H, Perelman D, Breschi A, Limcaoco P, Kellogg R, McLaughlin T, Snyder M. Glucotypes reveal new patterns of glucose dysregulation. PLoS Biol. 2018;16(7):e2005143. DOI: 10.1371/journal.pbio.2005143 · PMID: 30040822 [Cohort, CGM with standardized meals]
- Matthan NR, Ausman LM, Meng H, Tighiouart H, Lichtenstein AH. Estimating the reliability of glycemic index values and potential sources of methodological and biological variability. Am J Clin Nutr. 2016;104(4):1004-1013. DOI: 10.3945/ajcn.116.137208 · PMID: 27604773 [Controlled test meal study, n=63]
- Korem T, Zeevi D, Zmora N et al. Bread Affects Clinical Parameters and Induces Gut Microbiome-Associated Personal Glycemic Responses. Cell Metab. 2017;25(6):1243-1253.e5. DOI: 10.1016/j.cmet.2017.05.002 · PMID: 28591632 [RCT, randomized crossover study, 1 week each]
- Kovatcheva-Datchary P, Nilsson A, Akrami R, Lee YS, De Vadder F, Arora T, Hallen A, Martens E, Björck I, Bäckhed F. Dietary Fiber-Induced Improvement in Glucose Metabolism Is Associated with Increased Abundance of Prevotella. Cell Metab. 2015;22(6):971-982. DOI: 10.1016/j.cmet.2015.10.001 · PMID: 26552345 [In vivo and human study, crossover design]
- Asnicar F, Berry SE, Valdes AM et al. Microbiome connections with host metabolism and habitual diet from 1,098 deeply phenotyped individuals. Nat Med. 2021;27(2):321-332. DOI: 10.1038/s41591-020-01183-8 · PMID: 33432175 [Cohort, PREDICT 1, n=1,098]
- Gardner CD, Trepanowski JF, Del Gobbo LC et al. Effect of Low-Fat vs Low-Carbohydrate Diet on 12-Month Weight Loss in Overweight Adults and the Association With Genotype Pattern or Insulin Secretion: The DIETFITS Randomized Clinical Trial. JAMA. 2018;319(7):667-679. DOI: 10.1001/jama.2018.0245 · PMID: 29466592 [RCT, DIETFITS, n=609]
- Celis-Morales C, Livingstone KM, Marsaux CF et al. Effect of personalized nutrition on health-related behaviour change: evidence from the Food4Me European randomized controlled trial. Int J Epidemiol. 2017;46(2):578-588. DOI: 10.1093/ije/dyw186 · PMID: 27524815 [RCT, Food4Me, 7 countries, n=1,269 completers]
- Hollands GJ, French DP, Griffin SJ, Prevost AT, Sutton S, King S, Marteau TM. The impact of communicating genetic risks of disease on risk-reducing health behaviour: systematic review with meta-analysis. BMJ. 2016;352:i1102. DOI: 10.1136/bmj.i1102 · PMID: 26979548 [Meta-analysis, 18 studies]
- Tsereteli N, Vallat R, Fernandez-Tajes J et al. Impact of insufficient sleep on dysregulated blood glucose control under standardised meal conditions. Diabetologia. 2022;65(2):356-365. DOI: 10.1007/s00125-021-05608-y · PMID: 34845532 [Cohort, PREDICT, n=953]
- Bermingham KM, Linenberg I, Hall WL et al. Menopause is associated with postprandial metabolism, metabolic health and lifestyle: The ZOE PREDICT study. EBioMedicine. 2022;85:104303. DOI: 10.1016/j.ebiom.2022.104303 · PMID: 36270905 [Cohort, PREDICT 1, 627 women]
- Wyatt P, Berry SE, Finlayson G et al. Postprandial glycaemic dips predict appetite and energy intake in healthy individuals. Nat Metab. 2021;3(4):523-529. DOI: 10.1038/s42255-021-00383-x · PMID: 33846643 [Cohort, n=1,070]
- Ben-Yacov O, Godneva A, Rein M et al. Personalized Postprandial Glucose Response-Targeting Diet Versus Mediterranean Diet for Glycemic Control in Prediabetes. Diabetes Care. 2021;44(9):1980-1991. DOI: 10.2337/dc21-0162 · PMID: 34301736 [RCT, n=225]
- Rein M, Ben-Yacov O, Godneva A et al. Effects of personalized diets by prediction of glycemic responses on glycemic control and metabolic health in newly diagnosed T2DM: a randomized dietary intervention pilot trial. BMC Med. 2022;20(1):56. DOI: 10.1186/s12916-022-02254-y · PMID: 35135549 [RCT, crossover pilot trial, n=23]
- Popp CJ, Hu L, Kharmats AY et al. Effect of a Personalized Diet to Reduce Postprandial Glycemic Response vs a Low-fat Diet on Weight Loss in Adults With Abnormal Glucose Metabolism and Obesity: A Randomized Clinical Trial. JAMA Netw Open. 2022;5(9):e2233760. DOI: 10.1001/jamanetworkopen.2022.33760 · PMID: 36169954 [RCT, n=204]
- Bermingham KM, Linenberg I, Polidori L et al. Effects of a personalized nutrition program on cardiometabolic health: a randomized controlled trial. Nat Med. 2024;30(7):1888-1897. DOI: 10.1038/s41591-024-02951-6 · PMID: 38714898 [RCT, n=347]
- Ordovas JM, Ferguson LR, Tai ES, Mathers JC. Personalised nutrition and health. BMJ. 2018;361:k2173. DOI: 10.1136/bmj.k2173 · PMID: 29898881 [Review]
I cross-checked all 19 sources via PubMed. The figures in the text come from the study abstracts.
- Possible conflicts of interest: Several PREDICT analyses (Berry 2020, Asnicar 2021, Wyatt 2021, Tsereteli 2022, Bermingham 2022) and the 2024 program study were carried out with the involvement of the company ZOE, which offers personalized nutrition programs. The company DayTwo was involved in the Mayo Clinic study.
- A large part of the data comes from observational studies. They show associations, not causes.
- The randomized trials in this article lasted twelve months at most, some only days or weeks. They measure intermediate values such as HbA1c, blood fats or weight, not hard endpoints such as heart attack or new-onset diabetes.
- The banana story at the beginning is an illustration, not a case report and not a study result.
- Observations from my practice are marked as such in the text. They do not replace studies.