Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Mediterranean diet evidence

Observational study [1]
Results
Randomised trial [2]
Results
Comment

To some summer visitors to the shores of the Mediterranean, a Mediterranean diet might be thought to consist of lager and burgers, just in larger quantities than usual. To most of us it is more like a Greek salad with lots of olive oil, grilled fish, and some fruit, all washed down with the odd bottle of vino. There are big differences in the diet, and, since we are what we eat, big differences in what that diet does for us.

Much has been written on the benefits of Mediterranean diets, and the particular components that make it better. Recent large observational [1] and randomised [2] studies make understanding just a bit easier.

Observational study [1]

This was an examination of 22,000 Greek adults aged 20 to 86 years from all parts of Greece. Dietary intake for the year before enrolment was obtained by a questionnaire delivered by trained interviewers, and examined frequency and amount of food with photographs for estimation of usual portion sizes. The daily intake of 14 food groups or nutrients was obtained in grams per day for each participant.

Adherence to a Mediterranean style diet was determined with reference to some of these food groups, by the simple expedient of whether an individual consumed more or less than the median for their sex (Table 1). Possible scores were 0-9.


Table 1: Components of Mediterranean diet score for men and women


Median (grams/day)
Score
Diet components
Men
Women
Above median
Below median
Beneficial components
Vegetables
550
500
1
0
Legumes
9
7
1
0
Fruits and nuts
360
360
1
0
Cereal
180
140
1
0
Fish
24
19
1
0
Monosaturated:saturated lipid ratio
1.7
1.7
1
0
Detrimental components
Meat and poultry
120
90
0
1
Dairy products
200
190
0
1
Alcohol consumption
10-50
5-25
1
1
Note that 25 grams of alcohol is about two drinks a day, and that higher ratios of mono-unsaturated to saturated fat indicate larger consumption of olive oil


Figure 1: Death rate by diet score




Figure 2: Olive oil consumption by diet score




Individuals were followed up for a median of 44 months, and the date and cause of death for any participant obtained from death certificates and other sources. Observers blinded to the diet score of individuals adjudicated outcomes.

Results

There were 22,000 people with full details available. There were significantly fewer deaths among women than men, and more deaths in people over 55 years than in those under 55 years, and in current smokers, but fewer deaths in people who took more exercise.

Death rates in women and men were higher in those with a low diet score of 0-3 (1.7%) than in those with diet score of 4 and above (1.0%). Figure 1 shows the death rates per 10,000 person-years at different Mediterranean diet scores. Adherence to the diet involved large differences for the proportions above and below the median. For example, Figure 2 shows olive oil consumption among women with different diet scores, and Figure 3 shows vegetable consumption among men with different diet scores. Similar results obtained for legumes, fruit and nuts, fish, cereals, dairy products and monounsaturated to saturated fat ratios, but not for meat.

Figure 3: Vegetable consumption by diet score




A two point increment in the Mediterranean diet score reduced the risk of death by about 25%. Effects were important for older people, those taking less exercise, and any level of BMI, as well as cause of death or coronary heart disease or cancer.

Over an average of 3.7 years, a population aged 20 to 86 years with a Mediterranean diet score of 4 or more would have one fewer death for every 140 people (95% confidence interval 94 to 276) than a similar population with a score of 3 or below.

Randomised trial [2]

A randomised trial examined the dietary intervention of an Indo-Mediterranean diet consisting of a control group using the National Cholesterol Education Programme step 1 diet and the same diet with additional recommendations to consume every day at least:

The aim was to provide plenty of phytochemicals, antioxidants and alpha-linoleic acid.

Patients were Indians with a documented history of coronary artery disease, and randomisation was stratified by risk factors. Follow up was for two years, and the principle outcomes were fatal or nonfatal myocardial infarction, sudden cardiac death, and the combination of these outcomes.

Results

Each group contained almost 500 patients, and the groups were comparable at baseline for risk factors and treatments. Their average age was 49 years. Most patients had serum cholesterol between 6.2 and 6.7 mmol/L, and about half were smokers at entry. About 30% were overweight or obese.

The combined outcome occurred in 39/499 (8%) patients on the Indo-Mediterranean diet and 76/501 (15%) on the NCEP diet (Figure 4). The adjusted rate ratio was 0.5 (0.3 to 0.7), and reductions were similar for all components of the combined outcome. For every 14 (95% confidence interval 9 to 29) patients randomised to an Indo-Mediterranean diet for two years, one fewer had a fatal or nonfatal myocardial infarction or sudden cardiac death than similar patients using a standard NCEP diet.


Figure 4: Total cardiac endpoints for Indo-Mediterranean diet and control diet




While significant reductions in total and LDL cholesterol and triglycerides occurred in both groups, the reductions were much larger in those on the Indo-Mediterranean diet. After two years, these patients had reduced their total cholesterol by 0.7 mmol/L, compared with 0.2 mmol/L in control patients.

Other secondary clinical outcomes occurred significantly less frequently with the Indo-Mediterranean diet, including angina, a positive exercise test, left ventricular hypertrophy and heart failure. There were 49 cardiovascular events. Angiography, angioplasty or bypass surgery happened in 49 (10%) of people on the Indo-Mediterranean diet compared with 96 (19%) on the control diet, with an NNT of 11 (7-20).

Comment

These are two terrific studies, with different study architectures, and conducted in different people in different parts of the world. They come to much the same conclusion, that a diet rich in fruits, nuts, vegetables, fish and mono-unsaturated vegetable oils, and with limited dairy products and meat, is good for the heart. The odd glass of wine helps those of us who want to partake. The fact that they come to much the same conclusion is important, because it gives us more confidence in the results and the conclusion that a Mediterranean diet is a good thing.

The gains are not trivial. The setting of the randomised trial was that of secondary prevention, where we expect NNTs over five years for statins of about 10-20 (Bandolier 47). The Indo-Mediterranean diet had an NNT of 14 over two years, equivalent to an NNT of 5-6 over five years if the effect continued over that time. And in the observational Greek study, an NNT of 140 over 3.7 years comes down to about 10 over our 50 years or so of life as an adult.

Nor is this a particularly hard goal for most of us. The randomised trial [2] claims that the Indo-Mediterranean diet is economical, and can be produced by farmers at a cost of about US$1 per day. It may cost some of us a tad more at our local supermarket, but even so it is hardly likely to break the bank.

The take home message is that diet is a powerful agent for reducing our chances of dying young, or having heart disease. That message is just as important for those who already have heart disease as those who do not. For some folk access to components of a Mediterranean diet will be difficult, as it was for most of us before the advent of supermarkets and year-round fruit and vegetables. For most of us there's just no excuse any more.

References:

  1. A Trichopoulou et al. Adherence to a Mediterranean diet and survival in a Greek population. New England Journal of Medicine 2003 348: 2599-2608.
  2. RB Singh et al. Effect of an Indo-Mediterranean diet on progression of coronary artery disease in high risk patients (Indo-Mediterranean diet heart study): a randomised single-blind trial. Lancet 2002 360: 1455-1461.

previous or next story