Does red meat kill you? Only in a vegetarian's dream!
Red meat is the favorite enemy of nutritionists nowadays. Their studies and publications are often (ab-)used by those evangelical vegetarian types who would love to impose their no-meat religion on the rest of us. Don't buy it. Now let me show you how you can profess your love for steak AND support it with the data from the same studies which the zealots use for their vegetarian crusades. Earlier this year Pan et al. published a study titled "Red meat consumption and mortality" [1]. They had pooled the data of two large prospective studies, the Nurses' Health Study and the Health Professionals' Follow-up Study. Collectively these studies had followed 121,000 people, who were free of cardiovascular diseases at baseline, for more than 20 years. Altogether, the participants accumulated close to 3 million person years for observation. During the observation period close to 24,000 deaths occurred of which 6,000 were of cardiovascular causes, that is heart attack, stroke, heart failure. The researchers discovered that for every increase of 1 serving of unprocessed red meat per day the hazard ratio of dying from any cause was 1.13 and the hazard ratio of dying from a cvd-cause was 1.2. That means for every increase of a serving of red meat per day the chances of dying from any cause and from a cvd-cause increased by 13% and 20% respectively. Those rates were a little higher for processed red meat. To put this into perspective the researchers also calculated that if all participants had eaten less than half a serving of red meat per day (42g/d), 9% of deaths in men and 7.6% of deaths in women could have been prevented. Wonderful. Sounds impressive, but it isn't for one simple reason: Unreliable data acquisition. Just ask one question: how did the researchers know how much red meat those people ate? This question cuts to the heart of many, if not most, studies on diet-disease associations. Data on food consumption are typically acquired through food frequency questionnaires (FFQ). These FFQs ask you about your consumption of food items over the past days, weeks or even months. And as you can imagine, such recall can be terribly unreliable. So much so, that other researchers wanted to quantify this effect. So they used FFQs and compared the results with objective quantitative measurement of energy intake and protein intake [2]. And lo and behold, they discovered that if relative risks (such as the hazard ratio mentioned above) were calculated from FFQs they overestimate the true diet-disease association very severely. In fact so severe, that a hazard ratio of, say, 2 would in reality be around 1.3. What does that mean for a hazard ratio which is, as in the study of Pan and colleagues, less than 1.3 to begin with? It means possibly nothing. You certainly can't conclude from these data that red meat kills you. That's what it means. And mind you, this inaccuracy of FFQs shows up with recall periods of a few weeks. Pan and colleagues had to rely on FFQs which were conducted YEARS apart. In fact, data acquisition based on FFQs is so flawed, that the question been raised "is it time to abandon the food frequency questionnaire?" [3]. And the authors state: "We should be very circumspect about analyses of current studies that have used FFQs for dietary assessment." That was 7 years ago. We still have those FFQs and you still have the media telling you how bad red meat is for you. And I'm going to have a real nice steak now. How about you?1. Pan, A., et al., Red Meat Consumption and Mortality: Results From 2 Prospective Cohort Studies. Archives of Internal Medicine, 2012: p. archinternmed.2011.2287. 2. Kipnis, V., et al., Structure of dietary measurement error: results of the OPEN biomarker study. American Journal of Epidemiology, 2003. 158(1): p. 14-21; discussion 22-6. 3. Kristal, A.R., U. Peters, and J.D. Potter, Is it time to abandon the food frequency questionnaire? Cancer Epidemiology, Biomarkers and Prevention, 2005. 14(12): p. 2826-8.
