Development and validation of a modified Dietary Approaches to Stop Hypertension (mDASH) index in the community-based...

The Dietary Approaches to Stop Hypertension (DASH) intervention was developed in the 1990s to study blood pressure-lowering effects of a diet high in potassium, calcium, and magnesium, as well as fiber and protein. Various DASH Indices were developed in the 2000s to study the health effects of such a DASH eating pattern in observational studies. Here, we propose a modified DASH Index (mDASH) incorporating relevant updated evidence. We evaluated the reliability and validity of the new index and its ability to predict hypertension (HTN) risk in the prospective Framingham Heart Study Second Generation cohort and cross-validated its predictive ability in the younger Third Generation cohort.

Food frequency questionnaires were used to derive a food-based mDASH diet index. We began with the key components of the original DASH diet: fruits, vegetables, and low-fat dairy and then modified this base model by replacing low-fat dairy with total dairy. We then iteratively added other dietary components to develop the final mDASH index. Test-retest reliability was assessed across four examination cycles (exams 5-8), and construct validity was determined by comparing the target nutrient composition of mDASH with that of other indices. Finally, predictive validity through Cox regression models was assessed by comparing the ability of mDASH (versus earlier DASH indices) to predict incident HTN.

The final mDASH index ranged from 8 to 40 points and included: total fruit; total vegetables (including potatoes); total dairy (both full-fat and low-fat); fish/seafood; whole grains; legumes, nuts, and seeds; red and processed meats; and sugar-sweetened beverages. Moderate-to-strong Pearson correlations (r = 0.56-0.70) were found amongst mDASH scores across four exams, while construct validity was supported by similarly strong correlations (r = 0.51-0.71) between mDASH scores and key DASH nutrients-calcium, potassium, magnesium, fiber, and protein. Lastly, the mDASH pattern was strongly predictive of HTN risk. In the Second Generation cohort, those in the highest (versus lowest) mDASH quintile had a 27% lower risk of HTN (p-trend across quintiles = 0.0056), a finding that was comparable to that of an earlier DASH index by Fung (p-trend = 0.0255) and similar to a 35% lower risk (p-trend = 0.0004) associated with the highest adherence to the DASH index by Günther. A final DASH index by Dixon was not statistically significantly (p-trend = 0.1434) associated with HTN risk. In the validation cohort (FHS Third Generation), highest adherence to mDASH index was associated with a 20% lower HTN risk (95% CI: 0.61, 1.04).

We developed an updated, valid, and reliable eight-component mDASH index that performed as well as or better than existing indices in predicting HTN. This modified diet pattern provides more food choices in a DASH pattern without compromising its beneficial associations with blood pressure.