
This study was to show that adherence to a “Mediterranean Diet” (Med diet), as reflected by a high “alternate Mediterranean (aMed) diet score”, is associated with reduced risk of chronic heart disease (CHD) and stroke and stroke mortality.  The aMed diet is a modified version of the Med diet and is characterized by higher consumption of plant foods, including plant proteins (excluding potatoes, but including fruits, nuts, whole grains and legumes), monounsaturated fat (compared to saturated fat) and fish, and lower consumption of animal products and saturated fats.  Moderate intake of alcohol is allowed.
  	

Prior prospective studies documented an inverse association between adherence to a Med diet and risk of total and cardiovascular (CVD) mortality and an intervention trial showed the Med diet to be more effective than a low-fat diet, in promoting weight loss and lowering the ratio of total to high-density lipoprotein cholesterol in obese individuals.  Prior studies did not focus on stroke or stroke mortality.


This epidemiological study followed 74,886 women, 38 to 63 years of age, without a history of CVD or diabetes, who are participants in the Nurses' Health Study, a cohort study of 121,700 female nurses in 11 US states.  Six times, from 1984 to 2004, the authors computed an aMed diet score from self-reported “food frequency questionnaires” (“FFQs) designed to assess nutrient intake during the preceding year.  More than 95% of nurses continued participation throughout the 20 year period. FFQs were subjected to validation studies.  Reports of CHD, fatal CHD and stroke were confirmed by review of medical or hospital records, autopsy or death certificates. Where records were not available, cases were considered “probable” and included, with about 17% of all strokes and 24% of nonfatal myocardial infarcts so classified. Body mass index, smoking status, aspirin use, multivitamin intake, menopausal status and use of postmenopausal hormone were calculated for each FFQ, and beginning in 1986, leisure-time physical activity was assessed. 

To reduce random variation for an individual and better represent long-term dietary intake, cumulative averages of the aMed score from the repeated FFQs were calculated (averaging together scores from two prior years).  Adjustments were made for “confounders” (factors that could confuse results) by adjusting for age, smoking, BMI, menopausal status, postmenopausal hormone use, multivitamin intake, alcohol intake, physical activity and aspirin use.

Cox proportional-hazard modeling was used to assess the association between an aMed score and risk of CHD and stroke, using separate models for fatal and nonfatal CHD and stroke types. 

Greater adherence to the Mediterranean diet, as reflected by a higher aMed diet score, was associated with a lower risk of incident CHD and stroke in women.  During 20 years of follow-up, 2391 incident cases of CHD, 1763 incident cases of stroke and 1077 cardiovascular disease deaths (fatal CHD and strokes combined) were found. Women in the top aMed quintile were at lower risk for both CHD and stroke compared with those in the bottom quintile (relative risk [RR], 0.71). 