No matter how often I write articles defending the place of red meat in a healthy diet, there are always people who get bent out of shape whenever a new study is published with a different twist on the same old “red meat is killing us” story.
Although I can’t say addressing these (often ridiculous) studies is the most exciting undertaking, I wanted to write about this one because it’s a prime example of the limitations of observational evidence, and what happens when you control for even the most simple confounding variables.
Does red meat really increase your risk of diabetes? Read this to find out!
The study in question is titled “Associations between red meat intake and biomarkers of inflammation and glucose metabolism in women,” and was conducted by Walter Willet and a group of other Harvard researchers. (1) Using data from the Nurse’s Health cohort, they identified a positive relationship between red meat intake and higher levels of plasma C-reactive protein (CRP), ferritin, fasting insulin, and Hb A1c, after adjusting for demographic, medical and lifestyle factors. CRP (and sometimes ferritin) are markers for inflammation, and elevated fasting insulin and hemoglobin A1c indicate impaired glucose metabolism.
Limitations of Observational Evidence
Right off the bat, there’s a glaring weakness in this study that has been largely glossed over by many media reports: all but one of these associations disappeared after adjusting for BMI. As the study authors mention, excess body fat is the biggest risk factor for type 2 diabetes, so it shouldn’t come as a surprise that people with higher BMIs have less-than-ideal biomarkers for glucose metabolism. Additionally, it’s common for people who are overweight or obese to have have underlying chronic inflammation, so it makes sense that people with higher BMIs would tend towards having higher levels of inflammatory biomarkers.
A more accurate title for this study would probably look something like this:
But that wouldn’t garner nearly as much press, would it?
This study is also a classic example of “if you look for something hard enough, you will probably find it.” The association between red meat and biomarkers for inflammation and glucose metabolism has already been studied on numerous occasions, and the results have been inconclusive. But from the researchers’ perspective, there must be an association between red meat and these unfavorable biomarkers, so they felt it was necessary to try yet again to establish this relationship. This type of confirmation bias can be a big problem in epidemiology; the China Study is a case in point.
I don’t want to imply that all observational evidence is worthless. That’s not true. Epidemiology can study trends and associations that are nearly impossible to evaluate with clinical trials, and these observational studies can be valuable for generating hypotheses. They only become a problem when we start treating these hypotheses like conclusions, and base public policy or nutritional advice on them. (For further discussion on this, I recommend checking out this recent paper on the limitations of observational evidence in nutrition.)
So, what of the hypothesis that red meat contributes to (or even causes) inflammation and impaired glucose metabolism?
Red Meat and Inflammation
I’ve already examined the relationship between red meat and inflammation at length, and although a few different mechanisms have been proposed by which red meat could cause inflammation, none of them really hold water. Additionally, a clinical trial that came out just a couple months after the Willet study was published actually found that regular red meat consumption decreased markers of inflammation. (2)
In this study, a group of older women living in retirement communities were given two 80g servings of red meat per day to consume six days per week for four months, in combination with resistance training. Compared with the control group, the red meat group actually ended up with significantly lower levels of interleukin-6 (an inflammatory marker) by the end of the study. If red meat caused inflammation, eating it twice a day for four months probably wouldn’t produce these results.
Red Meat, Iron, and Glucose Metabolism
One relationship that the Willet study brings up is the link between heme iron consumption, iron metabolism, and glucose metabolism. If you recall from earlier, only one association didn’t lose significance after adjusting for BMI, and this was the association between red meat intake and serum ferritin.
Ferritin is a marker for iron stores, so on the surface, this may seem like just a side effect of consuming more heme iron. But ferritin is also an acute-phase reactant (like C-reactive protein) that can be elevated in chronic inflammation even when iron levels are completely normal, so it’s not sound to assume that elevated ferritin is related to iron intake from meat. I frequently see patients in my practice with high ferritin and normal iron levels, so it’s necessary to test other iron markers like iron saturation, TIBC, UIBC and serum iron to determine what’s going on.
This is an important point, because the study authors imply that increased iron consumption from red meat could contribute to impaired glucose metabolism, causing the positive association between red meat intake and biomarkers such as fasting insulin and Hb A1c. As I’ve talked about pretty extensively in the past, iron overload is certainly harmful, and even high “normal” iron stores can lead to metabolic issues. However, making this connection assumes that the additional heme iron obtained from the increased red meat consumption is actually causing iron overload, and this is unlikely for a few reasons.
First of all, this study speaks mostly in terms of relative red meat consumption – the quartile of subjects who consumed “the most” red meat had higher levels of these biomarkers. But did you notice the absolute amount of red meat consumed? According to the paper, women in the 95th percentile for total red meat consumption only consumed 141g (or 0.31 pounds) of total red meat per day. A third of a pound of red meat per day is not a lot, and although this amount of red meat provides more iron than the same amount of chicken or fish, it’s unlikely to cause iron overload, especially for premenopausal women who have higher iron requirements.
Further, even if the amount of iron in a third of a pound of red meat were excessive, the body has mechanisms that regulate iron absorption so that if the body is replete in iron, additional iron isn’t absorbed. These regulatory mechanisms break down in disorders such as hereditary hemochromatosis, but for most people (and again, especially premenopausal women), iron intake from reasonable amounts of red meat would not cause iron overload. So it’s unlikely that the high serum ferritin observed in the highest consumers of red meat is due to the heme iron content of meat.
In summary, I still haven’t seen any convincing evidence that red meat contributes to inflammation, and there’s actually evidence that red meat decreases inflammation. And although iron overload can contribute to impaired glucose metabolism, red meat consumption does not cause iron overload in the absence of a disorder such as hereditary hemochromatosis. That said, if you do have iron overload, it’s a good idea to limit red meat consumption.
I hope you’ve enjoyed this latest installment of “don’t take observational red meat studies at face value”! Please feel free to share your thoughts in the comments.