Saturday, October 4, 2008

Un-picking the Mediterranean diet [score]

The Mediterranean diet has been in the news again with the release of a study that pits it against a low-fat diet (the standard AHA recommendation) and a low-carb diet. But, I'm not going to talk about that. I'm actually a bit obsessed with the whole idea of the Mediterranean diet score. I think it's got something to do with the way it makes it all seem simple - as in a diet book approach you adjust your diet by tweaking your diet score - yet it's scientific - because some researchers have used this score to prove that if you keep your score high enough you will live longer!
The Mediterranean diet is usually described as emphasizing legumes, fish, olive oil and lots of fruit and vegetables and is supposed to represent the kind of healthy diet typical of those who live in the sun-drenched regions around the Mediterranean sea. In fact, of course, there are lots of different real Mediterranean diets: French, Spanish, Italian, Greek and so on. So what is the Mediterranean diet as it exists in the mind of nutrition researchers? The Mediterranean diet concept was launched with a study that declared that
A higher degree of adherence to the Mediterranean diet was associated with a reduction in total mortality (adjusted hazard ratio for death associated with a two-point increment in the Mediterranean-diet score,
The Mediterranean diet score described in that study was based on a Greek diet. The researchers studied the dietary intake of elderly Greeks using Food Frequency Questionnaires and based on the results from that came up with 14 "food groups" : potatoes, vegetables, legumes, fruits and nuts, dairy products, cereals, meat, fish, eggs, monounsaturated lipids (i.e. olive oil), polyunsaturated lipids (vegetable-seed oils), saturated lipids and magarines, sugar and sweets, nonalcoholic beverages.

These food groups already reflect the preoccupations of researchers. Why for instance are potatoes considered a separate group from vegetables? Aren't potatoes a vegetable? Is it because potatoes are starchy tubers? Well, presumably not, because it's only potatoes; other starchy root vegetables such as carrots and parsnips are not similarly singled out. Is it because potatoes contain enough nutrients apart from the starch to act as a staple crop just like cereals (in fact better than cereals) ? Probably not. My personal answer to this question is that it is because potatoes can be made into chips and crisps and therefore potatoes are (as every nutritionist knows) most definitely not a vegetable.

The diet score itself was based on only 9 out of these 14 "groups": vegetables, fruits and nuts, legumes, meat, fish, dairy products, cereals, monounsaturated to saturated fat intake ratio and alcohol consumption. (Why did they pick these and leave out potatoes, eggs, sugar and sweets and nonalcoholic beverages? Who knows? I can hazard a guess though: it reflects the obsession of fat bad: carbs good) and the actual diet score was worked out as follows:
  • for the items considered beneficial or 'good' (vegetables, fruits and nuts, legumes, fish, cereals, high monounsaturated fat ratio) a score of +1 was given if the individual consumed more;
  • for the items considered harmful or 'bad' (meat, dairy products, alcohol) a score of +1 was given if the individual consumed less.
More or less than what? What was the dividing line? Well, that is the interesting bit: the dividing line was the 'middling' consumption of the cohort, i.e. the median. [Given the whole range of consumption values for each item, the median is simply the value exactly in the middle, with 50% of values above it and 50% below.] This of course has the obvious result, that for every item, immediately, the study population separates out into sheep and goats: the sheep who are on the right side of the median (even by 1 gram!) and the goats who are on the wrong side!

So what was the 'middling' Mediterranean diet?
You can see it for men and women in this table.

To investigate what this means in terms of macronutrient ratios, I used Fitday to simulate these median diets as follows: for vegetables I included a combination of zucchini, peppers, aubergines and salad vegetables: lettuce, cucumber, tomatoes; for fruits: apricots, apples, oranges and bananas and nuts: almonds and pistachios. Fruits and nuts were combined as 95% fruit and 5% nuts – this was adjusted to get the resulting fat intakes close to the actually recorded values. For dairy, I used 60% wholemilk plain yoghurt and 40% (full-fat) feta cheese. For meat, a 50/50 combination of lamb and poultry and for fish, red mullet. The legumes were kidney beans and the representative sweets I chose was halvah. (After all what does sweets represent? Desserts? Or just confectionary?) Nonalcoholic beverages were a mix of fruit juices, soft drinks and tea. Cereals included were bread, crackers, cornflakes and filo pastry, oh and a representative biscuit – as it was explicitly stated that biscuits were included under cereals (and not under sweets for example).
As a check of my methods I compared my 'results' with the recorded fat intakes: I came out pretty close to the actual results:








Total kcal Sat Fat (g) PUFA (g) Monounsat (g)
Men
Calculated2290.232.5 15.757.7
actual 2354.5 33.115.055.9
Women
Calculated 1933.9 28.7 13.3 49.2
actual 1863.0 27.0 12.6 46.5


And the macronutrient ratios of such a diet?


Protein Fat Carb
Men 14.56% 43.22% 42.23%
Women 14.23% 44.00% 41.77%

Grams of carbs per day were were 220g for men and 183g for women – not low-carb. Most of the carb intake does come from the cereals group – but how does this compare to the typical diet pyramid recommendation? The diet pyramid recommendation is for 5-6 "oz grain equivalents" per day for women and 6-8 per day for men (the variation is due to age). A "1 oz grain equivalent" is a slice of bread, a cup of breakfast cereal, ½ cup of cooked rice or pasta. Calculating a combination of these for our median individuals gives a requirement to eat about 450g of cereals per day for men and about 340g per day for women – as you can see our Mediterraneans (177.7 g/day for men and 139.7g/day for women) are quite far from this ideal!

One final thing to note is that the proportion of fat in the actual Mediterranean diet as evidenced above is considerably more than was recommended in the recent diet study where it was restricted to 35% or less.