Earlier this month, Nature Magazine published the results of a study that found that a simple blood test can spot diabetes a decade before even the first symptoms appear.1 Currently, doctors have no way to accurately predict the onset of diabetes other than generally observing weight and rising blood sugar levels. Obviously, ten years of early warning would represent a huge advance in dealing with diabetes and could help prevent complications like heart disease, stroke, kidney failure, blindness, and amputation.
And that is all well and good, but the real treasure of the study may lie just below its “press release” results. Dig a little deeper, as it were, than any of the media reporting on it, or the researchers conducting the study for that matter, and you will find some revelatory information concerning diet and diabetes.
Before we look at the fun stuff, let’s look at the study itself.
The premise of the study was actually quite interesting. It is understood that metabolic diseases such as Type 2 diabetes are often present for years before becoming clinically apparent. For instance, by the time insulin deficiency manifests and a diagnosis of diabetes is made, pancreatic beta cells have been failing for years. Unfortunately, current modes of testing aren’t able to reveal it until the disease is fully manifest. Yes, testing for body mass index and fasting glucose levels can be predictive, but in truth, they more often than not merely reveal a disease that already exists, even if only at an early stage. But now, thanks to emerging technologies, it is possible to test quickly and inexpensively for a wide range of metabolites in the body — to see if the presence of any one, or several of them, is predictive of the later onset of disease.
With this in mind, the researchers evaluated 2,422 people who had been followed for 12 years in the early 1990s as part of the ongoing Framingham Heart Study.2 Of the people they followed, 201 went on to develop diabetes. After analyzing the data for 189 of those individuals and comparing it to 189 individuals from the non-diabetic control group, the researchers discovered some fascinating consistencies among the diabetics. Specifically, they found that:
- The fasting levels of 5 amino acids found in blood plasma — isoleucine, leucine, valine, tyrosine and phenylalanine — were strongly linked with a diagnosis of diabetes later on in life.
- The participants with the highest blood levels of three of these metabolites in particular (isoleucine, phenylalanine, and tyrosine) were 5 to 7 times as likely to develop diabetes.
To verify their results, the researchers additionally screened participants from a long term Swedish study and found among those individuals too they were able to accurately identify those who went on to develop diabetes.
Obviously, being able to identify people with a high likelihood of developing diabetes ten years before they actually come down with the disease holds out the possibility of early intervention in the form of lifestyle changes and/or drugs. Then again, we already have strong predictors in obesity and high-glycemic diets, albeit at a later stage in the development of the disease. Nevertheless, those predictors do not seem to work well in motivating intervention. This is especially disheartening considering that the disease is far more imminent, in fact most likely already present, when these predictors rear their ugly heads. How much less motivating would a prediction ten years out be perceived? On the other hand, the study was able to identify prospective diabetics who were neither overweight nor consumers of bad diets. These were people who were obviously trying to live healthy lives and therefore would likely be more amenable to altering their lifestyles if they knew what to do. In other words, there is no doubt that at least some people would benefit from a ten year heads-up on becoming diabetic.
However, before getting too excited, it should be mentioned that the researchers acknowledged that they didn’t actually know if the high levels of amino acids were cause or effect. Or as study coauthor Robert Gerszten of Mass General Hospital said, “We now need to figure out what is the cart and what is the horse. In other words, are these early markers of the disease or do they participate in the causal pathway that leads to the condition? We are currently investigating this.”
Then again, the answer to that question may already be contained in the data they have on hand — if they looked at it from a slightly different angle.
Amino acids and diabetes
Let’s quickly repeat the facts on hand:
- Fasting blood plasma levels of three amino acids in particular (isoleucine, phenylalanine, and tyrosine) reflected a five to sevenfold increase in the risk of developing diabetes.
- The researchers don’t know whether the higher levels of the cited amino acids actually cause the increased risk, or are merely reflective of some other cause.
What I would ask at this point is: are there foods that are especially high in these particular amino acids that have already given indication that their consumption might lead to an increased risk of diabetes? Answering this question would go a long way to identifying whether high levels of isoleucine, phenylalanine, and tyrosine are cause or effect.
So with that in mind, let’s take a look.
Foods ranked high in the targeted metabolites
When looking for foods that are high in amino acids, you’re looking at protein foods. Here is a comparison of several major sources of protein in the modern diet. I included human breast milk and my own Nutribody® rice/pea protein blend to serve as reference points. I grabbed the data for egg whites, milk, soy protein isolate, and ground beef from the USDA National Nutrient Database — converting everything to a 25 g serving of protein so we could compare beef patties to smoothies.3
Certainly nothing leaps out as distinctly dangerous in “every” category — but then it couldn’t, could it since each serving totals 25 g. If one amino acid is higher, another has to be lower to compensate. However, some smaller differences do stand out, especially when you consider how many servings people might have of that food in any given week, month, year, etc. For example: egg whites are notably higher than breast milk in isoleucine, phenylalanine, and valine — three of the five targeted metabolites. And then when you consider that the difference will likely be exacerbated if someone is regularly having egg white omelets in pursuit of health. Another example is milk, which is notably higher in leucine and valine. And then there’s whey, which is notably higher in isoleucine and “way” higher in leucine. Soy protein isolate and rice/pea protein each have one amino acid a bit out of line, but not multiples. And hamburger, surprisingly, stacks up quite well in its amino acid profile. If it weren’t for the other issues associated with meat, you might have something there. Keep in mind, there is a huge difference between organic, grass-fed, meat and dairy and the hormone-injected, grain-fattened variety. But that said, let’s look at the two protein sources on the list that have several amino acids that deviate substantially from human breast milk — eggs and dairy.
Eggs and diabetes
So are eggs good for diabetics? Well, if you search on the internet, you’ll find most sites say that eggs are great for diabetics. They’re extremely low on the glycemic scale; they’re a great source of protein; and new studies have shown that contrary to popular belief, dietary cholesterol from foods such as eggs has little effect on the body’s blood cholesterol levels. And that’s absolutely true, as far is it goes. In fact, I’ve reported on those studies myself. Yes, if your focus is on heart disease, eggs get a clean bill of health. But it appears eggs have a detrimental effect on diabetes that has nothing to do with cholesterol.
According to a study out of Brigham and Women’s Hospital and Harvard published in Diabetes Care in 2009, men with the highest level of egg consumption — at seven or more per week — were 58% more likely to develop Type 2 diabetes than those who did not eat eggs, and women were 77% more likely to become diabetic if they ate at least an egg a day.4 Perhaps all of the focus on cholesterol and diabetes misses the point. Perhaps the question is protein, and specifically the amino acid profile of that protein. If so, then the high isoleucine, phenylalanine, and valine levels in eggs and egg whites might be a problem. And if that’s true, isn’t that one heck of surprise for all those people eating egg white omelets thinking they’re being extra healthy. (By the way, the amino acid profiles for egg whites and whole eggs are very similar.)
Milk and diabetes
If our theory holds true that high levels of certain amino acids as found in specific foods could lead to diabetes, then we might expect to see a higher level of diabetes in those who drink large amounts of milk as noted in our chart. And in fact, the data supports that.
In 2005, the American Journal of Epidemiology published the results of a study linking milk and dairy consumption with an increased risk of being diagnosed with Type-2 diabetes. The study, conducted over 37 years out of the National Public Health Institute in Helsinki, Finland found that diets high in fruits and vegetables were associated with a reduced risk of diabetes, whereas diets associated with high consumption of milk, cheese, butter, and potatoes were associated with a notably higher risk.5 In addition, a second study out of Denmark, published the same year, found that a high intake of milk increased insulin resistance in young boys .6 Especially interesting is the fact that this study found that high consumption of meat had no such effect — only dairy. But based on the chart we looked at earlier, meat’s amino acid profile is friendlier than dairy’s, at least according to the study findings.
It’s not a protein food, but I’ve got to talk about aspartame because it does relate to the study and our amino acid chart. The artificial sweetener aspartame is a methyl ester of a dipeptide composed of two amino acids — aspartic acid and phenylalanine. Once ingested, it tends to break down into its constituents. This is significant for two reasons. First, phenylalanine was identified in the study as one of the three primary metabolites associated with diabetes. Considering that the study data was gathered during the early 1990s when aspartame was the artificial sweetener of choice, especially in diet sodas, the likelihood that aspartame itself may have contributed to the high incidence of diabetes we have seen in the two decades since cannot be discounted. How ironic if the very thing that people were using to avoid weight gain and diabetes may actually have contributed to both.7 And in fact, a study presented at the Endocrine Society in 2009 found that the use of artificial sweeteners (particularly aspartame) was indeed linked to a twofold increase in diabetes.8 Most definitely ironic!
What it all means
As the researchers pointed out, cause and effect has not been established between the cited amino acids and diabetes — merely a relationship. Further, it has not been clearly established that dietary intake of said metabolites leads to an increase in their plasma fasting levels, although as we have seen, there certainly are indications. And finally, and perhaps most important, even if there is a cause and effect relationship, the reason for it is unknown at this time. The bottom line is that definitive conclusions are not possible — but inferred recommendations certainly are.
As we have seen, there is some evidence to indicate a relationship between high consumption of certain protein sources such as eggs and dairy, which are high in several of the cited metabolites, versus other protein sources for which no such relationship has yet been established. Then again, perhaps that is because no one has conducted such a study to establish a connection between soy, rice, and pea proteins and diabetes. In fact, the only studies indicate just the opposite. For example, there have been studies that indicate that soy reduces lipid production and prevents hyperinsulinemia (the presence of excess insulin in the blood). But as we saw earlier when examining eggs, proving that a food positively affects one marker of diabetes is not necessarily the same as saying that it ultimately does not contribute to the overall incidence of the disease. On the other hand, since each of the vegetarian protein sources were only high in one cited metabolite, not multiple ones, that’s probably a good sign.
Then again, is the problem that a particular protein is high in certain amino acids, or is it a question that high protein foods are simply high in amino acids in general — and if you eat enough of them, you raise all amino acid markers? Yes, that’s possibly true, but there is no indication in the study that those individuals who became diabetic ate larger amounts of protein than the non-diabetics — only that their blood levels of certain amino acids were higher. Nor for that matter was there any indication as to what type of protein they ate.
There is another study that might shed some light on the issue. In this 1997 study, Type 1 diabetics who reported protein consumption less than 20% of total calories had average albumin excretion rates (AERs) below 20 mg/min. But in those whose protein consumption was greater than 20%, the average AERs increased and were in the “microalbuminuric range” indicative of incipient diabetic neuropathy. The important point in this study is that trends reached statistical significance only for animal protein, not for vegetable protein.9 In other words, the type of protein consumed matters — not just in terms of fat, growth hormones, cholesterol, etc. — but for how the body processes the protein itself. This is a very key point and focuses right back, like a laser beam, to our original study’s findings concerning specific amino acids. Different protein sources produce different effects in the body.
Before we jump on animal protein in general, though, it should be mentioned that according to research presented at the American Heart Association’s Scientific Sessions in 2008, retired National Football League players had a significantly lower prevalence of diabetes and metabolic syndrome compared to the male population as a whole. This is significant in that football players are generally large consumers of animal protein (including dairy) as part of their training regimens. Then again, NFL retirees had a higher prevalence of elevated cholesterol and impaired fasting glucose, which might be indicative of future diabetes.10 Not to mention the fact that NFL players (active and retired) tend to exercise more than the population as a whole — and exercise changes outcomes.
The bottom line comes down to moderation in all things. The body hates imbalances and extremes even when it comes to amino acids. Or as a variation on what I’ve said before, “Too much of a good thing is bad. Too much of any single amino acid is bad. No matter how healthy an amino acid is, if you overindulge in it, disease will result, not health.”
Until we know more as to whether the five amino acids in the study are cause or effect, moderation is the recommendation.
The truth is that most people in the developed world eat more protein than they need. Food consumption surveys show an average protein intake of approximately 100 grams per day (regardless of body type, sex, or exercise levels), with about 70% of that from animal sources. Most people could do with significantly less protein per day (about 45-55 grams per day is adequate for most). Then again, people with special needs such as performance athletes, active adults, senior citizens, and people recovering from illness or injury may have requirements that run up to as much as 70-100 grams a day. Beyond that, we’re talking about bodybuilders looking to put on as much muscle mass as is humanly possible. And when it comes to that, each bodybuilder knows his/her own requirements.
As to what type of protein to eat, I would recommend backing off from egg and dairy. You don’t necessarily have to stop eating them if that’s your thing; you just might want to eat less of them, though — and not just because of the diabetes issue. They are both highly allergenic, as is soy for that matter. That means that high consumption of these foods is liable to cause some degree of systemic inflammation — a condition associated with many diseases, including heart disease and diabetes. My preference is for hypoallergenic proteins. These include organic grass fed beef, low-mercury, non-toxic (and dare I say non-radioactive) fish, and rice, pea, and hemp protein concentrates.
Oh yes, and stay away from artificial sweeteners — especially aspartame.
For more on protein, check out Protein, Parts 1, 2, and 3.
For more on diabetes, check out Diabetes — the Echo Effect.
1 Thomas J Wang, Martin G Larson, Ramachandran S Vasan, et al. “Metabolite profiles and the risk of developing diabetes.” Nature Medicine 17,448–453(2011). <http://www.nature.com/nm/journal/v17/n4/full/nm.2307.html>
2 Daniel Levy, Philip A. Wolf, Larry D. Atwood, et al. “Framingham Heart Study.” National Heart, Lung and Blood Institute and Boston University. 1948-ongoing. <http://www.framinghamheartstudy.org/>
3 “USDA National Nutrient Database for Standard Reference.” US Department of Agriculture. 24 April 2011. <http://www.nal.usda.gov/fnic/foodcomp/search/>
4 Luc Djoussé, MD, DSC, J. Michael Gaziano, MD, Julie E. Buring, SCD and I-Min Lee, MBBS, SCD. “Egg Consumption and Risk of Type 2 Diabetes in Men and Women.” Diabetes Care. February 2009 vol. 32 no. 2. <http://care.diabetesjournals.org/content/32/2/295.full?sid=95a8baab-acbf-49a4-81a0-5bb4849c21f9>
5 Jukka Montonen, Paul Knekt, Tommi Härkänen,et al. “Dietary Patterns and the Incidence of Type 2 Diabetes.” Am. J. Epidemiol. (2005) 161 (3): 219-227. <http://aje.oxfordjournals.org/content/161/3/219.full>
6 C Hoppe, C Mølgaard, A Vaag, V Barkholt, and K F Michaelsen. “High intakes of milk, but not meat, increase s-insulin and insulin resistance in 8-year-old boys.” European Journal of Clinical Nutrition (2005) 59, 393–398. doi:10.1038/sj.ejcn.1602086 Published online 17 November 2004. <http://www.nature.com/ejcn/journal/v59/n3/full/1602086a.html>
7 Pamela L. Lutsey, MPH; Lyn M. Steffen, PhD, MPH, RD; June Stevens, PhD, MS, RD. “Dietary Intake and the Development of the Metabolic Syndrome.” Circulation. 2008;117:754-761. <http://circ.ahajournals.org/cgi/content/full/117/6/754>
8 Kristofer S. Gravenstein. “Use of Artificial Sweeteners Linked to 2-Fold Increase in Diabetes.” 29 June 2009. Medscape Diabetes & Endocrinology. 24 April 2011. <http://diabetes.com.sg/index.php?option=com_content&view=article&id=271:use-of-artificial-sweeteners-linked-to-2-fold-increase-in-diabetes&catid=72:medical-findings&Itemid=50>
9 Toeller M, Buyken A, Heitkamp G, Bramswig S, Mann J, Milne R, Gries FA, Keen H, and the EURODIAB IDDM Complications Study Group: Protein intake and urinary albumin excretion rates in the EURODIAB IDDM Complications Study. Diabetologia 40:1219-26, 1997.
10 Science News. “Football Players: Staying Active May Lower Health Risks For Large, Retired Athletes” 12 Nov 2008. Science Daily. 25 April 2011. <http://www.sciencedaily.com/releases/2008/11/081111102757.htm>