Intermittent fasting does not work the same way for everyone
The essentials at a glance
No one disputes that people respond differently to the same dietary pattern. The more interesting question is whether this can be predicted.
The largest study on this question found no significant interaction between a genotype pattern and diet type among 609 participants. We report this result even though it contradicts expectations of genetic tests.
You will read the cited source verbatim, three documented facts, three sentences in the fact check, a precise distinction of what the study examines and what it does not, and what a genetic analysis actually evaluates.
What to expect in this article
1. The claim and what it asserts
2. The documented source
3. What this study examined
4. Three documented facts
5. What the result means and what it does not
6. Three sentences in the fact check
7. What has specifically been established about intermittent fasting
8. The key point in one sentence
9. Why a null finding is valuable
10. What a genetic analysis says and what it does not
11. Why the source category matters here
12. Who can benefit from a genetic analysis
13. The DNA metabolic analysis in detail
14. Limitations: what remains open
15. What matters in this topic
Frequently asked questions
Sources
The claim and what it asserts
The headline of this article contains a claim that breaks down into two very different statements.
The first is: People respond differently. This is undisputed and visible in every study with many participants because their results vary.
The second is: One can predict who will respond in what way. This claim is the truly interesting one, and it is the one that must be examined.
This article deals exclusively with the second. We cover the question of energy balance and why someone does not lose weight with intermittent fasting in a separate, detailed article, which we will link to at the end.
And let us say upfront where the review leads: The result is not what one would expect from a provider of genetic analyses. We are writing it down anyway, because anything else would violate our own rules.
The documented source
The sentence comes from a randomized twelve-month study published in the Journal of the American Medical Association.
Documented source
“There was no significant diet-genotype pattern interaction (p = 0.20) or diet-insulin secretion (INS-30) interaction (p = 0.47) with 12-month weight loss.”
Gardner CD et al., DIETFITS Randomized Clinical Trial
JAMA 2018;319(7):667–679 · Translation by the mybody®x editorial team
Author's own translation: There was no significant interaction between diet and genotype pattern, nor between diet and insulin secretion, with regard to weight loss after twelve months.
The two numbers in parentheses are so-called p-values. Put simply, a p-value of 0.20 means that a difference of this size would occur by chance alone in about one out of every five studies without a real association. The usual threshold at which a result is considered statistically significant is 0.05.
So the value is not just slightly off the mark; it is far from it. That is important when interpreting the result.
What this study tested
To read the result correctly, you need to know exactly what was tested. The study design is unusually clear.
609 adults were randomly assigned to one of two dietary patterns—low-fat or low-carbohydrate—and followed for twelve months. 481 of them completed the study.
Before the study began, a genotype pattern was determined for everyone. The question was: Do people with a particular pattern lose more weight on one dietary pattern than on the other?
That is precisely the question behind the idea that a genetic test could tell someone which dietary pattern suits them. It was posed in advance, not searched for retrospectively in the data.
The answer was: no significant interaction. Neither the genotype pattern nor the measured insulin level predicted which of the two approaches would work better for a given person.
Three evidence-based points
Three pieces of information from the sources reviewed. They provide context and are not a diagnosis.
Evidence-based information
609
Adults were followed for twelve months; 481 completed the study (Gardner et al., JAMA 2018)
p = 0.20
the p-value for the interaction between diet type and genotype pattern was; a value below 0.05 is generally considered statistically significant (Gardner et al., JAMA 2018)
no recommendation
The DGE does not issue a recommendation on how often healthy adults should eat each day to lose or maintain weight (2012)
The third piece of information comes from a different context and still fits here. Two independent sources arrive at the same pattern for two related questions: The evidence does not support a general rule.
What the result means—and what it does not
A result of this kind is overstated in both directions, and both overstatements are wrong.
It does not mean that genes play no role in metabolism. The study tested a specific genotype pattern for a specific prediction, not the importance of genes overall.
It also does not mean that genetic analyses say nothing. An analysis can provide information about the metabolic pathways of individual substances, and that is a different kind of statement from a prediction about the success of a dietary pattern.
What this means is narrowly defined yet still significant: For the question of which of the two dietary patterns studied performs better for whom, the genotype pattern tested was not a useful predictor.
Anyone seeking to derive a dietary recommendation from a genetic analysis must measure it against this study. This applies to every provider, including us.
Three statements in a fact check
These three statements come up whenever genes and nutrition are mentioned together.
Checked against the evidence
Common claim
“A genetic test tells me which dietary pattern is right for me.”
Established
In 609 participants over twelve months, the DIETFITS study found no significant interaction between diet type and genotype pattern with regard to weight loss (Gardner et al., JAMA 2018).
Common claim
“So genes tell us nothing at all about metabolism?”
Established
The study does not say that. It tested a specific pattern for a specific prediction. It makes no statement about other analyses, such as those concerning the metabolic pathways of individual substances.
Common claim
“There are specific genetic markers for intermittent fasting.”
Established
We found no source in the class we require for claims about genetic predictions specifically for intermittent fasting. We therefore cite none.
What has specifically been established about intermittent fasting
The DIETFITS study compared a low-fat diet with a low-carbohydrate diet, not different eating windows. For intermittent fasting itself, it is therefore only an inference by analogy.
We therefore specifically searched for sources on genetic predictions in intermittent fasting and found none that met our requirements for source quality and currency.
What we found is the position of the German Nutrition Society on meal frequency. It states that, based on the available evidence, no reliable recommendations can be made about how often healthy people should eat each day to lose or maintain body weight (2012).
Together, these two findings paint a picture. There is no reliable rule either for the question of dietary pattern or for the question of frequency—and certainly not for the combination of genetic analysis and time windows.
We write it down and do not fill the gap with a conjecture. That is the uncomfortable option and the only one consistent with our own rules.
The core in one sentence
This is where the topic can be brought together.
It is undisputed that people respond differently. The largest study on the subject did not confirm that this could be predicted from a genotype pattern.
The entire difference between an observation and a prediction lies between these two sentences.
An observation only requires that something happens. A prediction requires that it be possible to predict in advance who it will happen to. The second step is the difficult one.
Why a null finding is valuable
A result that finds no association initially seems disappointing. For practical purposes, it is often more valuable than a positive result.
The reason lies in the design. This study posed the question in advance, randomly assigned the participants, and ran for twelve months. This is the design that would find an association if one existed.
When a study designed this way finds nothing, that is robust information. It is worth considerably more than a small analysis that searches the data retrospectively and happens to find something.
For someone considering whether to use a genetic analysis to make a dietary decision, this information is useful. It tells them what they should not expect and protects them from disappointment.
Why such results appear less often
There is a reason null findings are read less often in everyday life than positive findings, and it has to do with how the system operates.
A study that finds an association is easier to turn into a headline than one that finds none. It is cited more often, shared more often, and more often incorporated into product copy.
Anyone who wants to assess the strength of evidence should be aware of this. What people often encounter online is not the same as what the body of research as a whole shows.
For this article, that means: We cite the study that posed this question most rigorously, not the one whose results would fit best. That is the entire difference.
What a genetic analysis does and does not say
The table distinguishes which statements are addressed by the study and which are not. It provides context and does not constitute a diagnosis.
| Type of statement | Example | Addressed by DIETFITS? | Classification |
|---|---|---|---|
| Prediction of a dietary pattern | “A low-carbohydrate diet suits you better” | Yes, directly | No significant interaction found, p = 0.20 (Gardner et al., JAMA 2018) |
| Prediction of a time window | “Intermittent fasting works particularly well for you” | Not directly; the study did not examine time windows | We found no source of sufficient quality for this and therefore do not provide one |
| Information about a metabolic pathway | Categories for caffeine, alcohol, or lactose metabolism | No, the study did not examine that | A question of your own; the evidential value is described in the respective analysis report |
| Statement about the current state | “Your ferritin level is X” | No, no genetic analysis is responsible for that | That is answered by a blood test, not a genetic test |
The first line is the one that matters. Anyone who buys a genetic analysis expecting it to predict the right diet for them should have read this line.
The last line describes the most common confusion of all. A genetic variation never changes, while a blood value changes constantly. They therefore answer fundamentally different questions.
There is also a difference in the type of statement between the second and third lines. A statement about a metabolic pathway describes how the body handles a substance. A prediction says what will work better in the future. The latter is far more difficult to substantiate than the former.
Why the source category matters here
There are many publications in this field and very few that can be relied upon. So here is a word about our approach.
We primarily rely on German-language institutions such as IQWiG, the German Nutrition Society, and the Federal Joint Committee because they have already assessed the available evidence.
These institutions are silent on genetic prediction in this area. That is why we rely here on the primary literature and explicitly identify it as such, as we did in an earlier article on vitamin D and magnesium.
The selection criteria are therefore: randomized, a question formulated in advance, a sufficient number of participants, a long duration, and a reputable academic journal. The DIETFITS study meets all five criteria.
We do not cite studies that fail to meet these criteria for questions of this kind, even if their results would fit a product better.
One criterion deserves particular attention: the question formulated in advance. Anyone who searches a large dataset for long enough will find correlations, purely by chance. Only a question established before the analysis rules that out.
That is precisely where the significance of the study cited here lies. It posed the genotype question in advance and randomly assigned the participants, instead of looking for patterns afterward.
Who can benefit from genetic analysis
The comparison addresses the question of whether genetic analysis offers any benefit in your situation.
It makes sense for you if …
You are interested in the specific analysis categories, such as caffeine or lactose metabolism.
You want a one-time analysis that, unlike blood values, does not need to be repeated.
You understand that the current evidence does not allow for predicting a suitable diet, and you do not have that expectation.
You can allow for a processing time of 15 to 25 business days.
Probably not, if …
You expect it to predict the dietary pattern that suits you. The available evidence does not support this.
You want to know whether intermittent fasting will work for you. We found no reliable source for this.
You want to see your current values. That is what a blood test is for, not a genetic test.
You have symptoms. In that case, you need an evaluation at a medical practice.
The DNA metabolic analysis in detail
A genetic analysis does not provide a diagnosis or replace medical evaluation. It describes genetic variations that do not change over the course of your life.
The following information comes from the product page of mybody®x (MYBODY Lab GmbH), accessed on 19 August 2026. Based on everything stated in this article, we present it as a description, not a promise.
Genetic analysis from saliva
DNA metabolic analysis
According to the product page, evaluates more than 80 genetic variations across categories covering nutrition and diet; requirements for vitamins B6, B9, B12, D, and E, as well as iron, sodium, and potassium; and metabolic function, including alcohol, caffeine, and lactose metabolism, as well as gluten tolerance. What the analysis does not do: It does not predict whether intermittent fasting will work for you, measure current blood values, or provide a diagnosis.
Product-page information, accessed on 19 August 2026
Product-page information, accessed on 19 August 2026
Chapter at a glance
The analysis evaluates genetic variations related to diet, nutrient requirements, and metabolic function. The DIETFITS study found no robust association for predicting which dietary pattern works better for whom. The analysis does not measure current blood values, provide a diagnosis, or replace medical evaluation.
Limitations: what remains open
The first limitation is the transferability. The DIETFITS study compared a low-fat diet with a low-carbohydrate diet, not eating windows. For intermittent fasting, it is an inference by analogy, not direct evidence.
The second limitation is the scope of the finding. It concerns predicting a dietary pattern, not the overall significance of genes.
The third limitation is our source situation. We found no sufficiently high-quality source on genetic predictions specifically for intermittent fasting, so we do not make any.
The fourth limitation is the age of the study. It dates from 2018. If a newer study with a comparable design reaches a different conclusion, we will correct this article.
The fifth boundary is diagnosis. All assessments in this article provide context and do not constitute a diagnosis. A single laboratory value is always just one component of an overall diagnosis (IQWiG, as of April 2, 2025).
What matters about this topic
If you take away just one thing from this article, let it be this: Distinguish between the observation that people respond differently and the claim that this can be predicted.
For prediction, there is a large, well-designed study with a clear result: There was no significant interaction between diet and genotype pattern with regard to weight loss after twelve months (Gardner et al., JAMA 2018).
We could have written this article differently. The fact that we did not is the real answer: A text that takes numbers and sources seriously also reports results that do not fit the narrative.
Frequently asked questions
Does a genetic analysis predict which diet is right for me?
For this question, the DIETFITS study found no significant interaction between diet type and genotype pattern with regard to weight loss over twelve months among 609 adults (Gardner et al., JAMA 2018;319(7):667–679). Anyone purchasing a genetic analysis with this expectation should know this result.
Does that mean genes play no role in metabolism?
No. The study examined a specific genotype pattern for a specific prediction—namely, which of two diets would work better for whom. It makes no claims about other analyses, such as the processing pathways of individual substances like caffeine or lactose.
Are there studies specifically on genes and intermittent fasting?
We did not find a source that meets our requirements for source category and recency, so we are not citing one. The DIETFITS study compared diets, not eating windows; for intermittent fasting, it is an analogy.
What does the p-value of 0.20 mean?
Put simply: A difference of this size would occur by chance alone in about one in five cases in a study with no real association. The usual threshold for a statistically significant result is 0.05. So the value is clearly above that threshold, not just barely over it.
Why am’t I losing weight despite intermittent fasting?
This article deliberately does not address this question because it is a different one. We have a separate, detailed article on this topic that examines the energy balance using sources from the DGE, IQWiG, and WHO. It is linked under “Read more.”
Next step
What a genetic analysis actually evaluates
The DNA Metabolism Analysis evaluates more than 80 genetic variants, including categories related to alcohol, caffeine, and lactose metabolism. It does not predict whether intermittent fasting will work for you, measure current blood values, or provide a diagnosis.
Go to the DNA Metabolism AnalysisRead more
You might also be interested in this
The energy balance calculated with sources.
What the timing can demonstrably tell us.
Sources
- 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 – jamanetwork.com
- German Nutrition Society (DGE): Meal frequency and weight regulation in adults, technical information, July 2012 – dge.de
- Institute for Quality and Efficiency in Health Care (IQWiG): Understanding laboratory values correctly, as of April 2, 2025 – gesundheitsinformation.de
The verbatim English quotation, the p-values, the participant numbers of 609 and 481, the study duration of twelve months, and the study design come from source [1]; the German rendering of the quotation is a translation by the mybody®x editorial team. The statement about meal frequency comes from [2]. The sentence about being one component of an overall diagnosis comes from [3]. The statement that we found no sufficiently high-quality source for genetic predictions specifically relating to intermittent fasting is based on research conducted on August 19, 2026. The evaluation categories, the number of genetic variants, and information on price, sample type, processing time, and certification come from the mybody®x product page, accessed on August 19, 2026. All sources were accessed and reviewed on August 19, 2026.
mybody®x editorial & expert team
Laboratory diagnostics Blood analysis interpretation Nutritional science Nutrigenetics
This article was created by the mybody®x editorial and expert team. The team combines laboratory diagnostics, nutritional science, and the interpretation of blood analyses. Everyone involved can be found on the authors page.
Published on August 19, 2026 · Last updated on August 19, 2026
The content is for general information and does not replace medical advice, diagnosis, or treatment. Reference ranges depend on the laboratory, method, and age—the information on your report is always authoritative.
Share now:
The timing of your meals changes how your body responds
Why do DNA tests cost different amounts? The pricing tiers explained