Jan 16 2014

Blood Type Diet – Disproved

Skeptics often confront various challenging situations. One is sophisticated spurious knowledge (or SSK – OK, I just made that up). These are elaborate systems that purport to describe some body of knowledge but which are ultimately based on nothing – they are completely disconnected from reality.

Astrology is a great example of this. There are several astrological systems that are very complex, include charts and diagrams, and follow a certain internal paradigm or logic and underlying philosophy. None of this prevents astrology from being 100% fantasy. There is no relationship between the positions of the planets and stars and the personalities or fates of humans, and no mechanism by which such a relationship might exist.

The problems with systems of SSK is that they can give the powerful illusion of real knowledge. Further, people can become highly invested (in many ways) in such systems making it difficult to abandon them. They can even become institutionalized.

Another type of claim that presents a challenge are those for which there is simply no evidence one way or the other. As promoters of science, we like to appeal to high quality scientific evidence to support our positions. When there simply isn’t any science, then we have to resort to arguments of plausibility. These are perfectly legitimate, just not as satisfying.

We can also appeal to the very legitimate position that the person making the claim has the burden of proof, especially if they are selling something or making medical claims. But still, saying there is no evidence is not as rhetorically effective as being able to point to evidence that disproves the claim.

The Blood Type Diet

The blood type diet was invented by naturopath Peter D’Adamo, following what appears to me to be the naturopathic tradition of just making stuff up. He created an elaborate system of hand-waving explanations, with lots of sciencey references to genetics and metabolism.

Ultimately, however, the blood type diet was based on nothing. A 2013 systematic review concluded: “No studies that showed the health effects of ABO blood type diets were identified.”

This did not stop D’Adamo from writing a best selling book (selling 7 million copies according to his website). He has built a career out of the blood type diet, without actually bothering to publish any clinical evidence that his diet works.

The appalling lack of evidence has also not stopped Dr. Oz from promoting the diet. (Apparently eating right for his type gives him “vital energy.”)

The blood type diet, in essence, is the astrology of nutrition.

Fortunately, researchers have finally decided to put the popular but unfounded diet to a rigorous clinical test, and the results are just published in PLOS One – ABO Genotype, ‘Blood-Type’ Diet and Cardiometabolic Risk Factors.

Researchers looked at 1,455 subjects on each of the four blood type diets, and followed several outcomes like cholesterol, BMI, and waist circumference. Not surprisingly, for the A, AB, and O type diets subjects had some improvement in some of the parameters. Actually I found it most surprising that the B diet showed no improvement in any parameter.

Improvement is not surprising because every diet will tend to work, in that people following a diet will tend to eat less overall, or avoid the worst indulgences, and this will reduce cholesterol or weight.

The real outcome in this study was to look at the relationship between the blood type of the subjects and their response to the various diets. Not surprisingly (to skeptics, anyway) there was absolutely no relationship. It mattered not one bit what blood type the subjects were.

Senior author, El-Sohemy, is quoted:

We can now be confident in saying that the blood type diet hypothesis is false.

So what happens now? If this were a medical intervention, and a clinical study of this rigor and statistical power were published, it would result in abandoning the intervention (maybe not as quickly as we would like, but that is what generally happens). Proponents might still do further research, hoping for a different result, but it would be unethical to continue to use an intervention when the only evidence we have shows that it does not work.

I suspect that D’Adamo, who has never bothered to test his multi-million dollar scheme, will shrug off the results without batting an eye. Perhaps he may look up long enough to offer some ad-hoc rationalization for dismissing the study. You can always find some flaw or weakness to point to, and pretend that it’s a fatal flaw, if you don’t like the results. Or he may point to the fact that most subjects had improved parameters on the diets (except the B diet) and proclaim that his diets work.

I wonder if Dr. Oz will now publicly pull his support from the blood type diet and inform his audience that the best evidence we have (the only evidence) shows that it does not work.

In any case, I can now say to those who ask (and this is a frequent question) that the blood type diet has been studied and it simply does not work.

27 responses so far

27 thoughts on “Blood Type Diet – Disproved”

  1. jt512 says:

    Steve, were you writing for dramatic effect with the title? “Disproved!” That’s a bit strong for a non-randomized, observational, cross-sectional study, don’t you think?

  2. There is no objective threshold for demonstrating a negative. But given the lack of plausibility, this large study is sufficient to reject the hypothesis. If there were any clinically significant effect we probably would have seen it in this study.

  3. jt512 says:

    First of all, I’m not convinced that the study is “large” given the aim of the study: to detect an interaction between blood group and compliance with diet. No indication is given in the paper that the investigators computed the power to detect such an interaction, so I don’t know whether the study was adequately powered or not.

    Secondly, this study has immense methodological limitations. First of all, it’s cross-sectional, so cause-and-effect can not be inferred; for all we know, compliers could have self-selected in response to previous health problems that improved as a result of their changing to a more blood-type-compliant diet (personally, I doubt it, but it is for reasons just like this that cross-sectional studies are considered poor evidence in epidemiology). Outcomes were just anthropometrics and biomarkers, as necessitated by the cross-sectional design, rather than hard endpoints that can only be assessed in a longitudinal study.

    Next, compliance was really uneven across diets. According to Figure 1, almost no one in the study had a positive compliance score for the type-A diet, and no one at all had a positive compliance score for the type-O. It’s kinda’ shaky to make claims about a diet effect for a diet that no one in the study is even on! The authors claim, then, is that if there were a type-O diet effect, it should scale with the degree of non-compliance with the diet, even though no one is actually on the diet. Should it? Who knows? On the other hand, almost the entire study population was in high compliance with the AB diet. So this study is trying to find a relation between blood type and blood-type diet in a study population that is largely on just one of the diets (AB). If d’Adamo pays any attention to this study at all, he will almost certainly object to this, and rightly so.

    Contrast this to what a definitive RCT would look like, with subjects randomized in even numbers to designed test and control diets based on their blood types, and changes from baseline compared between each test and control group.

  4. jt512 says:

    Quick follow-up comment: This study is typical of “first look” epidemiologic studies. The authors had a cross-sectional data set and decided to explore it for a relationship of interest. On one hand, when such a relationship is found, it can suggest that the hypothesis is worth following-up in a more rigorous study. On the other hand, because the study is weak, we cannot expect that such a relationship would be detected; so a negative result isn’t very informative about the hypothesis.

  5. I think this kind of study is useful to see if a hypothesis is worth pursuing. They tend to be more worthwhile in the negative than the positive. Any association found would be tentative, and would not establish cause and effect. But if the results are dead negative, then there probably is no signal there. Again – you can’t prove a negative, but this type of data, given the weak hypothesis, is probably sufficient to conclude that it’s not worth pursuing.

    Also, when the only data available is negative, it definitely puts the proponents on the defensive. They already had the burden of proof. Now, it is all the more unethical to promote the intervention without first providing some clinical evidence that it works. If they don’t like the study, then they should do a more rigorous one.

  6. norrisL says:


    In the US it seems you call it “just making stuff up”. Here in Australia we call it “just makin’ shit up”. Still means the same thing of course.

  7. It’s the same here – I’m just trying to keep it clean.

  8. jt512 says:

    @Steve: I agree. “If they don’t like the study, they should do a better one.” Well put.

  9. Bruce says:

    My argument is why waste money chasing negative studies? If we assumed every negative study was a false negative we would spend a lot of time chasing ghosts. I would argue the claim for blood type diets is quite extra-ordinary, and as far as I can see the mechanics of it don’t look plausible, so for me I would like to see multiple positive studies before I would accept it.

    Is one study enough to disprove the hypothesis? Perhaps not in some eyes, but it is within reason to say so in my opinion in this case. As a scientist/statistician we would say “most likely has no effect” but that would not be strong enough to stop people wasting money on it. I think what is said privately and what is said publicly needs to be different to get the right message across.

  10. RickK says:

    @Bruce “If we assumed every negative study was a false negative we would spend a lot of time chasing ghosts.”

    I think your analogy is more apt than you think.

    The question is: Is there ANY evidence that would convince the proponents of blood type diets that they are ineffective?

    We just had an incident at work – someone in our company arranged to offer IgG food allergy screening done for any employees that wanted it. This is basically a blood test that tells the patient what specific IgG antibodies for specific foods are present in their blood stream, the pitch being that you are very likely to be intolerant of any of those foods. Several companies make a good profit selling this testing. And since this is covered by the national health, people assumed it worked and signed up in droves. Fortunately, I managed to get it cancelled by pointing out that such testing is at best controversial. For evidence, I provided several statements like this one from the Canadian Society of Allergy and Clinical Immunology (CSACI): “The CSACI does not support the decision of licensed physicians and our pharmacist colleagues to offer such testing, given the overwhelming consensus against the validity of such tests.”

    But of course one of the employees told me that the testing does work – he was absolutely certain – because it is widely employed by naturopaths. There was no set of tests, official statements or any other evidence I could provide that would convince him that it doesn’t work. Only if the naturopathic community rejects the testing would he then change his mind.

    As we know, once an alt-med modality is invented (and sells), it is never abandoned.

  11. Bronze Dog says:

    I remember checking out a thick astrology book from my high school library a couple times because I just found it fascinating. Had fantasy illustrations of ‘avatars’ of the signs to convey a feel of the alleged personality types. I knew it was nonsense, but it was elaborate nonsense. Kinda got interested in tarot for similar reasons, but I didn’t find any books on it. These interests got started in part because of video games that referenced them. Final Fantasy Tactics had combat modifiers based on the characters’ astrology signs and their ‘compatibility’. Ogre Battle uses Tarot’s major arcana as magic items, often summoning a relevant figure to do an elemental attack.

  12. Bruce says:


    “The question is: Is there ANY evidence that would convince the proponents of blood type diets that they are ineffective?”

    Exactly, it is a question most true believers will give no satisfactory answer to. And even if they do most of them will move the goalposts when you do meet their criteria or claim conspiracy theories etc. It is a thankless battle and in all honesty I have no idea where Steve and some other commenters here find the patience.

  13. So these academics have nothing better to do with their time and research funding than to ‘debunk’ moronic diet fads – irrespective of how temporarily popular any one might be at present? They can’t think of any useful contributions they can make to expand our understanding of nutritional science?

    What next? A study to debunk the claims of the banana and coconut diet? Let’s do a study on the South Beach Diet. They can study everyone’s eating habits in South Beach compared to the rest of the planet…

    The only slight saving grace to all this is that the funding was apparently privately derived. A sad state of affairs. The authors of this study should be tarred and feathered, metaphorically speaking.

  14. Karin says:

    There’s a whole world of blood type dating in Asian countries.

  15. Bruce says:

    Apparently Will thinks no one should study anything unless it fits his world view of what is a useful contribution to expand our knowledge.

    Steve, can we get a veto system in place so all scientists pass their ideas by Will Nitschke before they go ahead with a study?

  16. Ahmed El-Sohemy says:

    Will Nitschke:

    I am the senior corresponding author of this study published in PLoS One, and one of the academics that you refer to as “having nothing better to do with their time and research funding than to ‘debunk’ moronic diet fads – irrespective of how temporarily popular any one might be at present?”

    This ‘fad’ diet has been around for almost 20 years now, with >7 million copies of the book sold and millions of dollars are probably being spent on related testing, cookbooks, shakes, supplements, and even apps! (www.4yourtype.com). All with the belief that an individual’s blood type affects what they should eat. These products are sold by many naturopaths and chiropractors (who sign up as Sales Affiliates on the website above), to their patients who trust these healthcare professionals to provide them with remedies and recommendations that are based on scientific evidence, not just a popular book. Our study is the first to ever test the validity of the blood type diet hypothesis.

    Why did we bother to do this? I’ve been a researcher at the University of Toronto for about 15 years and I’ve held a Canada Research Chair in Nutrigenomics for the past 10 years. My research program in nutrigenomics aims to understand why some people respond differently from others to the same nutrients or foods they consume (think of lactose intolerance; celiac disease; or PKU). I’ve given over 100 invited talks around the world and almost every time I have someone in the audience ask me “So, is this like the blood type diet?”. My response has always been “No. I’m not aware of any study that has shown that difference in the ABO gene (which determines whether you are A,B,AB or O) makes people respond differently to different diets”. Rather than appearing to be a passing fad, this diet seems to be increasing in popularity. Because some studies have shown that individuals with different blood types have different risk for certain diseases, like type 2 diabetes, it seemed possible (albeit remotely) that they respond differently to various foods (given that perturbed metabolism is a hallmark of diabetes). Our findings showed that adherence to either the A,B or O type diets was associated with some beneficial effects. However, these effects had nothing to do with an individual’s blood type. So, if a person followed the type A diet, they would see similar health effects regardless of whether they had A,B,AB or O type blood. This means the diets are generally prudent, healthy diets (eg A type diet is mostly a vegetarian diet, which has long been known to promote health – most people just don’t stick to it). This explains why so many people swear that it works. But it would work equally well on someone who had an ‘incompatible’ blood type. So, the bottom line is the benefits (which many people will see if they can actually stick to any of the A,AB or O type diets – no matter what their actual blood type is) have absolutely nothing to do with a person’s blood type. This could be very useful information for the millions of people who have purchased this book and are being sold pricey supplements, shakes and other products that supposedly provides them with some unique benefit, when the only scientific evidence to date (http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0084749) shows there is no unique benefit. So, you may argue that we wasted time and money to test the validity of a ‘fad’ diet, but consider all the time and money that millions of others may now be saving.

  17. Ahmed – thanks for dropping by and taking the time to leave a thoughtful comment. Don’t worry about Will – he is a recent troll here who manages to find fault with everything but so far as I can recall has yet to make a coherent argument.

    I too am frequently asked about the blood-type diet, and I was happy to find an actual study to shed light on it.

  18. y meyshar says:

    Please read Dr. Peter D’Adamo’s comments:

    Thank you.

  19. Bruce says:

    Thanks y meyshar,

    I found Ahmed El-Sohemy’s response to that article in the comments of that article very interesting and I would suggest you read them too.

  20. @Steve

    I’m not faulting ‘everything’. If I don’t comment on a post I either agree with it or I find nothing interesting about it, or nothing interesting to add.

    @Ahmed El-Sohemy

    I understand the type of argument you might mount to defend a study like this but I don’t agree with it for the reasons I set out. If there is no evidence to support claim X and nutritionists and medical practitioners inform the public of this, there is nothing more than can or should be done about it. People who adopt silly diets because they want to believe in them are going to take anecdotal information more seriously than a medical study. It worked for my sister-in-law, or there must be something to it if 7 million readers bought the book! Clearly your medical study is flawed…

    What I’m saying is, if you’re dealing with members of the public who are prepared to act without evidence, then counter evidence is hardly going to make much of a difference to them.

    Amateur skepticism has its good and bad points. Medical researchers shouldn’t be replicating the sorts of things amateur skeptics obsess about. (I appreciate that Steven Novella may be deeply offended by that perspective and call me names, but so be it.) What medical researchers should be doing is adding to our body of scientific knowledge concerning health, well being, and the treatment and prevention of disease. Your study even if well intentioned, doesn’t do this.

  21. y meyshar says:

    @ Bruce

    Accuracy, truthfulness, integrity are good enough for me.

  22. y meyshar says:

    Pleae read more comments by Dr. Peter D’Adamo:


    Thank you.

  23. BillyJoe7 says:

    Dr. Peter D’Adamo: “That the BTD theory is currently unproven by rigorous scientific study is not argued”

    Convicted by his own mouth.
    Nothing more needs to be said really.

  24. Bruce says:

    I challenge Dr. Peter D’Adamo to prove, scientifically, that it works if he has so many issues with this study.

  25. BillyJoe7 says:

    At the above link, there is an interesting exchange on this question:

    Armed El-Sohemy:

    “You state “That the BTD theory is currently unproven by rigorous scientific study is not argued.”
    So, what happened to that 10 year trial on blood type diet and cancer remission you described in your book?:

    Even now, as I write this, I am beginning the eighth year of a ten-year trial on reproductive cancers, using the Blood Type Diets. My results are encouraging. So far, the women in my trial have double the survival rate published by the American Cancer Society. By the time I release the results in another two years, I expect to make it scientifically demonstrable that the Blood Type Diet plays a role in cancer remission.” – Peter D’Adamo (p 307, Eat Right 4 Your Type, 1996).

    Where are the results of this study you claim to have conducted?

    Peter D’Adamo:

    “Sorry, but I am a not going to play your little game here. What I wrote in 1996 has very little to do with what you wrote in 2014. Perhaps if I felt your question insinuated less, I would provide an answer, but since I do, I won’t. You can ask around, or troll my website (as pastime you seem to enjoy) and find the answer there. In the meantime I’lll just eat my mac and cheese and lima beans.”

  26. Thomas Oellrich says:


    You find that interesting? What I find most interesting is that this blog proclaims that the Blood Type diet has been disproved based on a study where the participants weren’t even following the Blood Type Diet at all. I mean Hot Dogs rated as beneficial for Type 0? Last time I looked, they consisted out of sausages made out of pork (not to mention E-Numbers), buns made out of wheat flour often garnished with ketchup. Pork, wheat and ketchup are all rated as avoids for Type 0 in the Blood Type Diet book. I personally wouldn’t touch a hot dog with a ten feet pole. Not to mention beverages that aren’t even accounted for. A study subject that is Type 0 could drink as much coffee, liquor or black tea (all avoids for Type 0) as he wanted without them having a negative score on the overall outcome. What’s funny is that you guys have so much egg on your face and you don’t event realize it.

  27. Northcroft says:

    How much time should I spend on this? Yawn!

    I’ve been looking at the Blood Type Diet for about 15 years. 15 years ago I bought between 50 and 100 blood test kits for people I met that did not know their blood type.

    Rather recently, in the last 2 or 3 years, a whole lot of new stuff has been coming out about the bacteria in the gut. 10x the number of cells in the body. 5000 different kinds. It was 50 kinds 20 years ago, 500 kinds 5 years ago. Who knows how many different kinds there are – very recently scientists have discovered that they have only been able to see 1% of the bacteria in the sea! A couple of studies – one American, one European – have shown that people do not have any old combination of the 5000 types. Rather they fall into one of 3 groups. The group they fall into is not connected to sex, race, diet, or age. I suggest that the group they fall into depends on blood type – because regardless of what you eat, the bacteria eat you, and your spit, which has your blood type antigens in it. Whatever they are!

    Other recent research has shown that the presence of particular bacteria in the gut can cause immune system responses that cause arthritis, asthma and thickening of the arteries. Eating exactly the same foods will not result in the same bacteria in the gut of a blood type A person, a blood type B person, and a blood type O person. But eating exactly the same foods IS more likely to result in the same bacteria in people with the same blood type. And the same problems.

    Between the age of 20 and 40 most people can eat anything. As one gets older every bit of the body begins to function less and less well. Logically this must include the functions of the gut, and the very construction of the intestines.

    ANYWAY. So what? In my own case I can binge on Dadamo’s beneficial foods without seriously negative consequences. But if I binge on the foods that he suggest I avoid I get all sorts of problems. Interestingly different foods give me different problems! Potatoes give me bleeding gums. Fried potatoes give me heartburn. Orange juice gives me headaches. Milk makes me tired and makes it hard for me to think. White bread gives me gas and arthritis. Pork makes my skin itchy.

    If I stick to the diet – which I do not unless I am feeling bad – a number of other things improve. My dry eyes get much better. My feet become hot instead of cold. I become ill very much less – maybe once a year – I used to get colds quite often. My back feels a lot better. My tinnitus is much reduced. I am a bit thinner, maybe a half stone lighter.

    I had none of these problems between the ages of 20 and 45. I am now 62. And if I keep to a mild version of the Dadamo diet, I can pretend that I am still 25!!

    I do not think much of Ahmed El-Sohemy’s study. I have read a critical analysis of it, and it is hopeless. It would be completely idiotic, for instance, to use students for such a study. If one were interested in its healing potential, I would choose people aged between 50 and 60. And the people’s diets would need to be accurately monitored. I have been watching people these past 15 years, and people do not eat what they say they eat! Both fat and thin people! Ill people and well people.

    It interesting to note that ayurvedic medicine puts people into one of three groups, and suggests that they eat slightly different diets. If you have ever thought about the Indian caste system, you may know that some castes have stayed genetically pure for the past 3000 years, with no marriage outside the caste. I am speculating here, but it is possible that IF there were a caste system for doctors, that India will have been the only place on the planet where medical knowledge was passed down from father to son for 100 generations – maybe long enough to get a handle on the foods that suit or do not suit the three groups of people. Dadamo’s blood type food categories have a lot in common with the Ayurvedic food groups. And modern science is discovering that quite a bit of the ayurvedic system is effective. And some of it is hocus pocus … ! Quite a bit.

    The reasons that Dadamo gave 15 years ago for his system working were also rather hocus pocus. But that does not matter. Real science is gradually finding out why it works. And it does.

    The proposition that “eating healthily” is obvious and an easy thing to do is nonsense for many people. Just think for one second. If tomatoes, or oranges, were bad for you. Would “organic” tomatoes or oranges be better, or worse, for you? The answer of course is worse, because organic food is more full of the essence of whatever the plant or animal is. Because the food has taken longer to mature, and because it has had to develop its own immune system fully. So, take extra care in avoiding organic food if it is an avoid on the blood type diet!

Leave a Reply