Abstract MP091: Healthy Eating Index and Food Based a Priori Diet Pattern Score and Mortality in Postmenopausal Women: The Iowa Women’s Health Study

Introduction Summary measures of food quality such as the Alternative Healthy Eating Index (AHEI) predict the risk of total mortality and cardiovascular disease (CVD). Most of the scores are impractical for public use in that they are based on selected nutrients rather than foods. We have created an a priori diet pattern score which is exhaustive and food based. Hypothesis We assessed the hypothesis that food quality scores (both AHEI and a priori diet score) are associated with a reduced risk of total and disease specific mortality. Methods We analyzed data from 24,859 postmenopausal women free at baseline in 1986 of diabetes, CVD and cancer and mean age 61.4 years in the Iowa Women’s Health Study. Food intake was assessed at baseline using a validated 127-food-item Harvard food frequency questionnaire. The AHEI score was calculated based on the values of 9 components; vegetables, fruits, nuts and soy, the ratio of white (seafood and poultry) to red meat, cereal fiber, trans fatty acids, the ratio of polyunsaturated fatty acids to saturated fatty acids, the use of multivitamins, and alcohol intake. Each component could contribute 0-10 points to the total AHEI score, except multivitamin use (2.5 points for non-users or 7.5 points for users). The a priori food score was based on intake categories for 34 food groups rated by expert judgment as positive (n=17), neutral (n=7) or negative (n=10); these judgments resulted in a plant-centered diet. The total score was the sum of the category scores (0-3) for positively rated food groups plus reverse scores (3-0) for negatively rated food groups. Through December 31, 2008, 8528 total, 2982 CVD, and 2675 cancer deaths were identified through the State Health Registry of Iowa and the National Death Index. Results Mean ± SD AHEI was 35.8 ± 9.5 and a priori diet score 38.5 ± 8.2; correlation between scores was 0.6. In proportional hazard regression models adjusted for age, energy intake, marital status, education, place of residence, high blood pressure, body mass index, waist-hip-ratio, hormone replacement therapy, physical activity and smoking, relative risk (RR) was computed for highest vs. lowest quartile of the diet score. For AHEI, the multivariable adjusted RR was 0.82 (95% CI: 0.77-0.88) for total mortality. For CVD and cancer mortality the multivariable adjusted RRs for AHEI were 0.73 (95% CI: 0.65-0.82) and 0.86 (95% CI: 0.76-0.97), respectively. The a priori diet score had a multivariable adjusted RR for total mortality of 0.76 (95% CI: 0.71-0.81). For CVD and cancer mortality, the multivariable adjusted RRs for a priori diet score were 0.77 (95% CI: 0.69-0.87) and 0.83 (95% CI: 0.73-0.92), respectively. Conclusions In conclusion, both the food-based a priori diet score and the food- and nutrient-based AHEI were associated with a reduced risk of total and disease specific mortality in older women. A food-based score may be more practical for public health policy.