AHEI and DASH Lead Diet-Score Evidence—But Labels Matter Less

AHEI and DASH have the broadest evidence linking higher diet quality to lower risk of death, cardiovascular disease, and type 2 diabetes. That is the practical ranking from a 2023 umbrella review. But it is not proof that either score produces a larger effect than the Mediterranean diet or the Healthy Eating Index. The review graded the consistency and quality of observational evidence, not a head-to-head causal contest with a winner's medal.

The more useful finding is less glamorous: the leading scores reward much of the same food pattern. Vegetables, whole grains, legumes, nuts, unsaturated fats, and less processed food, refined carbohydrate, and saturated fat keep appearing under different names. For people deciding what to eat, adherence matters more than the diet-score brand.

What the umbrella review actually ranked

Brlek et al.'s 2023 umbrella review combined seven systematic reviews of prospective cohort studies. It assessed five diet-quality indices—Alternate Healthy Eating Index (AHEI), DASH, Dietary Inflammatory Index (DII), Healthy Eating Index (HEI), and Mediterranean diet scores—against all-cause mortality, cardiovascular disease incidence or mortality, and type 2 diabetes incidence or mortality. Its search ran through October 2021.

Using NutriGrade, a tool for judging the quality of nutritional meta-evidence, AHEI and DASH were the only indices with moderate-quality evidence across all three outcome categories. DII reached that threshold for all-cause mortality and cardiovascular outcomes, but not type 2 diabetes. Mediterranean diet scores reached moderate evidence for all-cause mortality only. HEI reached moderate evidence for cardiovascular incidence and mortality only.

That supports an evidence-quality hierarchy of AHEI and DASH first, then DII, with Mediterranean diet scores and HEI behind them in overall breadth. It does not establish that AHEI cuts mortality more than DASH, or that DASH prevents more diabetes than a Mediterranean pattern. The published abstract does not report the individual risk ratios, hazard ratios, or 95% confidence intervals for every score-and-outcome pairing. A quality grade and an effect-size ranking answer different questions.

The apparent winners eat the same way

AHEI, DASH, Mediterranean diet scores, and HEI differ in their point systems and food-group cutoffs, but their overlap is substantial: more plant foods and fiber-rich staples, more nuts and unsaturated fats, and fewer highly processed foods and refined carbohydrates. DASH also emphasizes lower sodium; Mediterranean scores commonly emphasize olive oil and other monounsaturated fats. Those details can matter for an individual's preferences, blood pressure, culture, budget, and medical needs. They do not turn these scores into separate biological species.

Within-cohort comparisons support that interpretation. In the 2021 Physicians' Health Study analysis, researchers compared Mediterranean, DASH, and AHEI patterns with all-cause mortality among U.S. male physicians under the same outcome ascertainment conditions. The study adds useful comparability, although its participants were health-conscious, relatively affluent men with limited dietary variation. That profile can shrink observed differences and limits generalization.

A 2023 Multiethnic Cohort Study examined 10-year changes in diet quality and later cardiovascular mortality using multiple indices. Its contribution was temporal: sustained or improving diet quality tracked with lower subsequent cardiovascular mortality. A one-time food questionnaire is a blurry photograph; repeated diet assessment is closer to a short film, though still not a randomized trial.

Other pattern scores point in the same direction. The Portfolio Diet study across three prospective cohorts linked a plant-based, cholesterol-lowering score with cardiovascular risk. The Portfolio pattern emphasizes plant protein, viscous fiber, and unsaturated fats. Separate randomized trials of that pattern reported lower LDL cholesterol by 17%, non-HDL cholesterol by 14%, apolipoprotein B by 15%, and high-sensitivity C-reactive protein by 32% versus a low-saturated-fat diet; each comparison was statistically significant. Those are surrogate outcomes, not proof that the diet prevents heart attacks, but they make the cardiovascular association biologically plausible.

Why associations are credible but still not causal verdicts

Observational nutrition studies cannot fully separate food choices from the lives that make those choices easier. A 2024 JAMA Network Open cohort study found that socioeconomic status, diet, and behavioral factors were interrelated in cardiometabolic disease and mortality. Education, income, healthcare access, neighborhood conditions, smoking, physical activity, and baseline health can all travel with a higher diet score. Statistical adjustment reduces this problem; it does not erase it.

The concern is especially relevant to diabetes. In the 2023 Sister Study analysis of 40,243 women, socioeconomic disparities were tied to dietary patterns and type 2 diabetes risk. This does not make diet irrelevant. It means an observational association between a score and diabetes risk contains both dietary signal and social context, in proportions the study design cannot perfectly divide.

Index design adds another complication. A 2025 scoping review found inconsistent definitions, heterogeneous components, and suboptimal reporting across diet-quality-index studies. AHEI and DASH may have more consistent evidence because their food patterns are superior, because they have been studied more often or better, or because both are true. The present evidence cannot disentangle those explanations cleanly.

DII illustrates the uncertainty. It had moderate-quality evidence for mortality and cardiovascular outcomes but not type 2 diabetes in the umbrella review. Since inflammation is involved in diabetes biology, that gap could reflect a weaker relationship with glucose metabolism, fewer or more heterogeneous diabetes studies, or limitations in how the DII captures diet. It does not validate a simple story in which one inflammation score explains every chronic disease outcome.

Mechanisms and higher-risk populations add context

Mechanistic work offers clues, not a causal receipt. In the Framingham Heart Study proteomics analysis, 6,360 participants received AHEI, modified Mediterranean-style, and modified DASH scores. Thirty of 71 cardiovascular-related proteins were associated with at least one diet score after multiple-testing adjustment and internal validation. Protein mediation analyses connected diet quality with incident cardiovascular disease and all-cause mortality. These biomarkers may sit on relevant pathways, but mediation analysis in an observational cohort cannot prove that changing a particular protein causes the clinical benefit.

Newer cohorts extend the association into populations with more illness. A 2025 study of 9,101 people with established cardiovascular disease linked dietary patterns with survival. A 2026 NHANES-based cohort study included 16,589 adults aged 30 and older with cardiovascular-kidney-metabolic syndrome; over a mean 8.2 years, it recorded 2,811 deaths from any cause and 882 cardiovascular deaths while evaluating AHEI, DASH, alternate Mediterranean, and HEI-2020 scores. It reported lower all-cause and cardiovascular mortality with healthier dietary patterns. These findings matter because people with established disease still eat, but they remain associations rather than evidence that assigning a score will improve survival.

For diabetes, a 2025 UK Biobank prospective study found that the Planetary Health Diet index and other plant-based indices were associated with lower type 2 diabetes incidence. The authors described similar risk reductions across several frameworks, which fits the broader convergence story without directly reranking AHEI or DASH.

How to use the ranking without overreading it

Choose AHEI or DASH if you want the score with the most consistently moderate-quality observational evidence across mortality, cardiovascular disease, and type 2 diabetes. Choose a Mediterranean-style or HEI-based approach if it better fits your food culture and you can follow it for years. The evidence does not support abandoning a high-quality Mediterranean pattern because an umbrella review assigned it a lower evidence grade for some endpoints.

• Build meals around vegetables, legumes, whole grains, nuts, and other minimally processed plant foods.
• Use unsaturated-fat sources such as olive oil, nuts, and fatty fish when they fit your diet.
• Reduce ultra-processed foods, refined carbohydrates, and saturated-fat-heavy foods.
• For DASH, pay particular attention to sodium if blood pressure is a concern.

The best-supported conclusion is modest and useful: high adherence to a shared whole-food dietary pattern tracks with lower chronic-disease risk and mortality across many cohorts. No randomized trial has assigned people to competing diet-score targets and measured death, incident cardiovascular disease, or incident type 2 diabetes as primary outcomes. The labels are useful measurement tools. They are not competing metabolic religions.