Jan 17 2014
A study purporting to ‘debunk’ the blood type diet theory has recently been published. [1,2] However, a closer look at the study’s experimental design raises serious questions about its conclusions, including whether in fact the participants were actually following the blood type diet at all, and given its other parameters, would it have even been possible for the study to have any other outcome.
In response, I’ve written about some of my concerns with the design of the study and consequently the strength of what conclusions can be drawn from it. I’ve also asked a few colleagues who are versed in research evaluation to read the article, look at the data, and weigh-in on the conclusions. Let’s take a look at some of the more serious flaws in the study:
1. None of the subjects actually followed the Blood Type Diet.
The study was done on 1,455 participants of the Toronto Nutrigenomics and Health study. The study’s subjects simply kept diet diaries and a ‘diet score’ was calculated to determine their adherence to the respective diet for their blood type. This is how the researchers calculated the subject’s adherence to the blood type diet:
“Based on the food items listed in the ‘Blood-Type’ diets, subjects received one positive point for consuming one serving of each recommended food item and one negative point for consuming one serving of an item on the list of foods to avoid. Foods that are listed as ‘Neutral’ were not included in the equation and do not contribute to the final score.”
From the start it should be obvious that this method is a gross simplification. For example, a type A subject eating 12 ounces of high-fat hamburger three times per week would have a ‘0’ rating if they garnished the hamburger with onions. Simply giving +1 and -1 values for following or not following the food choices will most likely result in the subject’s results simply marching backwards and forwards and often just canceling out.
Dr. Ryan Partovi: “My problem with this methodology is that two people —one who eats mostly neutrals and a few beneficials could end up with the same score as someone who eats heaps of avoids and then just covers them up with a slightly greater number of beneficials. That’s not the way I’ve found the BTD to work. You can’t offset avoids with beneficials.”
Dr. Natalie Colicci: “I’m fairly certain you do not recommend white bread and potato chips to any patients”.
Dr. Ryan Partovi: “I think one of the biggest issues that I see is one of food quality. The difference between a grass-fed, grass-finished beef steak and a corn-fed beef steak is as large as the difference between a corn-fed beef steak and a piece of salmon. The average reader/researcher is clueless to those distinctions as well as their impact on what they seem most interested in measuring (lipids, weight, and blood sugar regulation).
The researchers themselves point to the limitations of this model, admitting that “Since the scoring system in the present study only assessed relative adherence to each of the four ‘Blood-Type’ diets, we could not determine the absolute number of people who strictly followed any of the diets.” 
Dr. Joseph Veltman: “Whenever I see a researcher using a food frequency questionnaire to evaluate someone’s nutrition, the results and conclusions are meaningless. The American Dietetic Association has said as much too. Research subjects who use this instrument bias their intake on what they think the researchers want them to eat.”
The study misclassifies numerous foods or lumps them into impossible-to-categorize groups. For example, ‘Mac and Cheese’ is listed as neutral for types A, B and AB. However, ‘mac and cheese’ is almost universally made with processed American cheese, which is listed as an avoid for those types. ‘Other grains’ (whatever they are) is listed as neutral for types A, B and AB, which effectively removes the majority of lectin specificities.
Dr. Ryan Partovi: “Many of the food values they used are also flat out wrong. Generic hotdogs (almost always pork) beneficial for O’s? Fried potato chips neutral for B’s and AB’s? Fried corn chips neutral for A’s? Sandwiches beneficial for O’s?! Did they look at the typical ingredients of any of these foods?! Fish clumped up into fish cakes, Dark meat fish, Other fish, and assigned random values? You can’t conclude anything from this nonsense.”
If this wasn’t bad enough, the very design of the study resulted in almost the entire study population being high compliance with the AB diet. So in essence, this study is trying to find a relation between blood type and the blood type diet in a study population that is largely on the AB diet.
2. The study used healthy young adults.
The study was performed on young adults aged 20 to 29. Using a population comprised sole of heathy young adults is virtually guaranteed to produce very negligible differences between the blood types, especially since each of the blood type diets is fundamentally healthy whole-foods diet. A much more useful study would have looked at a more health-compromised population, for example individuals with digestive disorders. Most chronic illness is accompanied by changes in cell glycosylation and host micro-biome, and it is these changes that sharply define the differences in how each blood type might benefit from a specific diet.
Dr. Natalie Colicci: “These people are healthy and it would seem unlikely that you’d see much variation due to genetic differences in as short a time as one month.”
Dr. Todd LePine: “Is the Blood Type Diet the cure of all ailments? No, but food affects the immune system directly and indirectly via the interaction with lectins on the gut mucosa and how the food shifts bacterial populations which in turn affects host metabolism.”
3. The study was conducted over a very short period of time.
The study analyzed what the subject reported eating over a one-month period. Even under the best of conditions, this is a very short period of time to observe any difference as subtle as variation between individuals. Our observations have consistently shown that a minimum three-month period is required for even the earliest demonstrable differences to be discerned.
Dr. Natalie Colicci: “I don’t think 30 days is enough to cause a sustained improvement. I’ve always said that superficially the blood type diet is a whole foods diet. That can’t be argued with. So of course taking any one blood type and mismatching it to one of the diets would show improvement in 30 days to some extent just because you are ‘cleaning’ up someone’s diet, especially if their current diet was contributing to elevated cardio risk factors.”
Dr. Todd LePine: “Measuring a few ‘biomarkers’ in a month’s span that are associated with cardiovascular disease and which take years to develop is like observing just one mole on the skin not change in a month’s time and saying you can predict the probability of melanoma.”
4. The diets actually benefited many of the subjects.
Given even its very short length the study’s authors themselves admit that the diets produced positive effects: “However, the observed results showed that even relatively high adherence to Type-A, Type-AB and Type-O diets were associated with favorable levels of cardio-metabolic disease risk factors, albeit in an ABO-independent manner.” 
Dr. Ryan Partovi: “I did find it interesting that the Type O diet lowered triglycerides (much more of a problem for Type O’s anyway) much better than the Type A diet, which was better at lowering cholesterol (much more of a Type A problem anyway).”
5. The study’s scope was overly simplistic.
As previously discussed, nobody in the study actually followed the blood type diet. In addition, if the researchers were truly interested in whether there were differences between individuals of different blood types (in such a short time frame) were truly significant, they should have also included the subject’s secretor status in their workup, as I have repeatedly stressed in my subsequent writings.  Since ABO type was determined genomically, adding this additional data (and adjusting food values to include variations based on secretor status) would have minimally added to the cost and complexity of the study but might have produced observable differences in the short (30 day) time frame.
Dr. Todd LePine: What is the Blood Type Diet? As I’ve always understood it, it was based on both ABO status and secretor status. I don’t see that they measured secretor status in the paper.
Dr. Mitchell Stargrove: “On first pass it seems to have a narrow set of evaluation factors, and a shallow comprehension of what actually constitutes adherence; let alone missing out on deeper approaches such as secretor status.”
6. When you go looking for something (or not), you very often find it.
Dr. Todd LePine: “In the study the ‘drug’ was the food, for which they did not fully control.”
Dr. Ryan Partovi: “This is a retrospective study. When you know the data set ahead of time, it’s fairly easy to structure a study that ends up with the result that you’d like to find. Because there is no control group, and this is only a retrospective study based on general trends in a person’s eating, there’s no way to really say for certain that the people who were following a diet of avoids for their blood type weren’t doing much worse than the average. Put another way, this study looked at the wrong thing. I’d say that the fact that they didn’t involve you, the author of the blood type diet books, in the study design is evidence enough that they weren’t serious about figuring out what the diets can do for cardiovascular disease.”
7. Possible conflicts of interest.
According to the disclosed ‘competing interests’ section of the article, one of the study’s principle investigators, Ahmed El-Sohemy, holds shares in Nutrigenomix Inc., a genetic testing company for personalized nutrition. Nutrigenomix markets a variety of nutrigenomic test kits, genotyping and customized reports to dietitians.  Although the article states that Dr. El-Sohemy holds shares in Nutrigenomix, in reality he is actually the founder of the company. It should not be too difficult to imagine the possible benefits to invalidating the blood type diets under these circumstances.
This study had the opportunity to shed new light on a complex topic in nutrigenomics. However, its definition of what actually constitutes following the Blood Type Diet was simplistic to the point of uselessness. In addition, the length of the study was too short and the significance of its negative results over-extrapolated. Far from proving the ineffectiveness of the blood type diet, it simply demonstrates the previously known fact that, in the short term, there are a multitude of approaches to eating healthy –if you are an already healthy young person.
In addition, pejorative elements in the article’s syntax, such as the continued use of the phase “Blood Type Diets” in quotation throughout the article; the simultaneous publication of a press release hailing the ‘debunking’ of the blood type diet, and the direct links between the principal investigator and other nutrigenomic business interests call into question, at least in my mind, if the study’s outcome was an already pre-ordained forgone conclusion.
Dr. Todd LePine: “From this study you can’t conclude anything, except that the Blood Type Diet was beneficial and not harmful in all cases. But in this limited time-frame, relatively unscientific study we don’t have the power or time-frame to conclude much of anything. The thousands of real patients, who have followed the diet for a myriad of aliments, especially those with inflammatory and autoimmune issues, benefit in the real world of do I feel better, think better and move more easily.”
- ABO Genotype, ‘Blood-Type’ Diet and Cardiometabolic Risk Factors
- ‘Popular blood type diet debunked’
- Metabolic and immunologic consequences of ABH secretor and Lewis subtype status.
- Nutrigenomix Scientific Advisory Board