In a noteworthy first, a large-scale study in Italy has found that following the so-called Mediterranean diet reduced heart disease risk only for those in top of the educational and income cohorts. Even more disconcerting, the researchers pushed the data around and didn’t find strong explanations for what might have caused the disparity in results.
We’ll discuss the study, High adherence to the Mediterranean diet is associated with cardiovascular protection in higher but not in lower socioeconomic groups: prospective findings from the Moli-sani study in more detail below. I encourage the medically and statistically minded to read the study in full and add their views.
We pointed out in 2007 that socioeconomic status played a big role in health outcomes, and even more so in countries with high degrees of economic disparity. So one has to wonder to what degree the stress of being poor, and of increased precariousness for older low and middle income individuals (the people selected for this study were over 35 and were in less than perfect health) offset the benefits of a good diet, meaning even though the Mediterranean diet may not have led to better results, they would have been worse off had they not followed it.
The study was large scale, starting out with over 24,000 participants and winding up with nearly 19,000 after various screens were applied, like missing data and implausible inputs. Subjects provided extensive information on health indicators, and in many ways, the diet reporting was detailed (for instance, it scored the variety of fruit and vegetable consumption as well as “healthier” versus less healthy veggie and protein cooking methods, as well as the level of organic food consumption). The findings were controlled for age, smoking level, body mass index, and amount of physical activity.
They also scored the participants’ diets on “food antioxidant content.” But there were some odd gaps, like looking only at whether the participants ate whole grain bread as the measure of whole grain consumption. Anyone who had made “whole grain” bread will tell you it still consists primarily of white flour, plus looking only at bread would omit those who ate oatmeal or other cooked cereals (where does polenta fit in the “healthy grain” spectrum?) Nevertheless, the effort to score the diets on a more granular basis looked sound.
The researchers divided participants into three socioeconomic groups based on income and educational levels.
The social stratification impact was strong enough that the authors flagged it as a “key message:”
Cardiovascular advantages associated with the Mediterranean diet are confined to high socioeconomic status individuals.
The International Journal of Epidemiology article also highlighted that there were “inequities” in “diet-related behaviours and nutrient intake”. However, if you read the piece, you can infer that the authors are scratching their heads a bit, in that the differences in eating behaviors that they observe by socioeconomic group don’t seem to be sufficient to add up to the difference in results.
However, the biggest disparity, and this could potentially be significant, is that the best-off ate more fish and less meat. However, the meat they did eat was prepared more often in the methods the researchers deemed to be less healthy, namely frying or grilling as opposed to stewing or boiling. The highest status groups ate more organic food, but since many studies have found that organic foods don’t contain more nutrients, they seemed unwilling to attribute much significance to that.1 Conversely, less well off participants used the “less healthy” cooking methods more often on veggies. Higher income people ate more whole grain bread; the less well off groups had more variety in their fruits and vegetables and consumed a smidge more of the mono and poly-unsaturated fatty acids.
The top income group also ate more “nutritionally dense” foods.2:
Dietary intake of polyphenols, antioxidants, fibre and indicators of TAC [total antioxidant capacity] were all increased in the higher income group as compared with the lowest.
Lambert had an interesting theory: that olive oil is often adulterated, and therefore the lower income groups might not be getting the “monounsaturated fatty acid” benefit that the researchers thought they would from their reported oil oil consumption. I doubt that would swing the results but it could be a contributing factor.
The authors admit they are groping for explanations:
On the basis of our results and by accounting for all the limitations inherent to our approach, we suggest that at comparable levels of adherence to the MD, higher SES groups actually select foods with increased nutritional value as those higher in antioxidant content or capacity, and are more keen on reporting a larger variety in fruit and vegetable consumption, thus obtaining more adequate intake of essential nutrients. Such nutritional gaps may partly explain the observed socioeconomic pattern of protection derived from apparently similar scores of adherence to the MD. Of note, the interaction between diet and SES on CVD health outcomes was found both for cultural (education) and financial resources (income), likely indicating that healthier choices are driven either by a good set of knowledge and skills or greater financial resources.
In light of our findings, we speculate that standard dietary scores, although useful and valid parameters to quantify the adherence to the Mediterranean diet, may not fully capture the complexity of this diet, leaving out a number of additional dietary details mainly related to the quality of the products. If so, we would be dealing with a methodological limitation which may be overcome by the use of other tools to better appraise the dietary behaviours of a given population.
Needless to say, this finding is intriguing but leaves a lot of questions unanswered. Again, the most likely hypothesis would seem to be that diet can’t explain the difference in results. The most likely non-diet factor would be stress, either overall or greater incidence of high stress events in the lower 2/3 groups played a role.
Is it possible that mere concerns about food safety could be a significant factor? More diet-conscious Americans are nervous, to the degree of neuroticism, about what they eat, so you get self-undermining behaviors like people drinking bottled water when it turns out sitting in plastic means the water picks up chemical nasties from the plastic. However, given how powerful placebo effects are, I wonder if better off-people are more confident in their food choices, like buying more organic food, and that helped contribute to these results.
Needless to say, this is an important indicator if disheartening indicator of yet another way those at the top of the heap come out ahead.
1 My view, consistent with that of a relative who had a healthy cooking venture and later coached health coaches, was that people in the US eat organic food to avoid things like pesticides and artificial hormones, and not because they think they have more “nutrients”.
2 The reason I am signaling some skepticism is “nutritional density” has become a fad in the US. While some foods clearly have more nutritional value than others, such as dark leafy greens versus, say cucumber, the “supernutrient” fetish implies that you can get enough nutrients without supplementation. The USDA said in 1938 that soils were so depleted that it was no longer possible to get enough nutrients from one’s diet. It’s also pretty much impossible for women to get enough calcium on a caloric intake that won’t make them overweight (note this presumably was possible when women were eating a lot more because they were engaged in strenuous physical activity on a regular basis, like helping out in the fields, beating rugs out, and churning butter by hand). So while there is some merit in the line of thinking, it’s been overhyped.