Diet and Ischemic Heart Disease: What GBD Models Can—and Can’t Say

Dietary risk for ischemic heart disease (IHD) is not limited to what people eat too much of. Comparative-risk models repeatedly assign substantial burden to what diets lack: nuts and seeds, whole grains, fruits, vegetables, and fiber. Sodium and processed meat still matter. The useful lesson is broader than any one ranking: a diet that adds plant foods while cutting sodium and processed meat covers more evidence than a campaign against a single villain.

That conclusion needs a warning label. The Global Burden of Disease (GBD) framework estimates how much disease would occur under different population exposures; it does not randomly assign people to diets for decades. Its factor rankings are model outputs, shaped by exposure data, relative-risk estimates, and the theoretical minimum-risk exposure level chosen for each food. They can guide priorities, but they cannot settle whether low nuts and seeds outrank sodium everywhere on Earth.

A persistent dietary burden in a high-risk region

In North Africa and the Middle East, dietary risks ranked second among 12 modifiable risk factors for IHD deaths and disability-adjusted life years in both 1990 and 2019, behind high systolic blood pressure. Mohammadifard et al.'s 2025 regional GBD analysis also placed the region second globally for IHD burden in 2019. The rank did not shift over three decades.

That persistence does not prove that dietary risk failed to improve, or that treatment made no difference. It does show that, in this region, diet retained its relative place among major modifiable contributors despite large changes in cardiovascular care over the same period. The analysis reports uncertainty intervals intended to incorporate sources of bias, including inconsistent time trends, age trends, and data sources. Those intervals should temper any attempt to read an exact causal hierarchy from the ranking.

The regional study does not break its aggregate dietary category into sodium, processed meat, nuts, or whole grains. It therefore cannot answer the tempting "natural experiment" question of what low-burden regions eat differently. Country-level burden patterns alone are not dietary records. To make that comparison, researchers would need comparable food-intake data rather than a map and a hunch.

The U.S. model shows both excess and insufficiency

A 2017 analysis by Micha et al. applied a comparable comparative-risk approach to 10 dietary factors among U.S. adults in 2012. For combined deaths from heart disease, stroke, and type 2 diabetes, sodium, low nuts and seeds, processed meats, low omega-3 seafood, and low vegetables were the five largest modeled contributors. Sodium ranked first as an individual factor.

That finding is the needed correction to any story that crowns plant-food insufficiency as the universal leading culprit. In this U.S. analysis, excess sodium led the individual rankings. Yet when the authors' modeled insufficiency factors were considered together—low nuts and seeds, omega-3 seafood, vegetables, fruits, and whole grains—their combined contribution exceeded sodium's. Diet can fail in several directions at once, which is inconvenient for anyone selling one clean dietary enemy.

These estimates do not establish global 2023 rankings. They concern U.S. adults, 2012 exposures, 10 dietary factors, and a combined cardiometabolic mortality endpoint rather than IHD deaths alone. Differences in population, endpoint, and GBD cycle can change the order of factors. They are best read as evidence that the problem includes both harmful excesses and missing protective foods.

Fiber and plant foods: a plausible target, not a magic ingredient

China offers another example of the absolute burden that can accompany dietary insufficiency. Wang et al.'s 2021 GBD 2017 analysis found that IHD deaths and disability-adjusted life years attributable to a diet low in fiber increased in absolute terms from 1990 to 2017. Age-standardized rates showed partial moderation, so the rise in absolute burden should not be mistaken for a simple worsening of individual-level risk. Population growth and aging can increase total cases even when rates change more favorably.

Fiber also provides a practical way to think about the food categories that recur in these models. Whole grains, legumes, fruits, and vegetables often bring fiber along with other nutrients and bioactive compounds. A 2020 review by Aune et al. summarized observational evidence linking fruits, vegetables, whole grains, nuts, and legumes with cardiovascular outcomes and mortality. The available report supports the direction of those associations, but does not provide extractable dose-response estimates or intake thresholds. It supports eating more plant foods; it does not justify a precise prescription from this evidence alone.

Randomized evidence is narrower but useful. A 2024 systematic review and meta-analysis by Mao et al. pooled 13 randomized controlled trials of anthocyanin supplementation in people with type 2 diabetes. Across the trials, supplementation improved glycemic measures and related cardiovascular biomarkers. These are surrogate outcomes, not evidence that anthocyanin supplements prevent heart attacks or IHD deaths. The population had type 2 diabetes, and the intervention was supplemental anthocyanins rather than berries, vegetables, or a whole dietary pattern.

Still, the trial evidence makes the plant-food argument more credible than a model alone would. It shows that a class of compounds found in richly colored plant foods can alter measured metabolic and cardiovascular risk markers under controlled conditions in a high-risk population. It does not prove that a supplement can replace a diet built around fiber-rich foods.

The burden is also moving into younger populations

The future burden deserves attention because IHD is no longer a disease story confined to old age. Xie et al.'s 2025 GBD 2021 analysis reported that, among people aged 15 to 39 worldwide, IHD incidence rose about 16% and prevalence about 19% between 1990 and 2021. The largest increases occurred in low- and middle-income countries. This ecological analysis cannot assign those increases to diet; smoking, blood pressure, obesity, diabetes, access to care, and demographic change also matter. But it gives little reason to assume that dietary risk has quietly solved itself.

Related global analyses find that IHD clusters geographically with ischemic stroke across 203 countries and territories, consistent with shared risk factors. An et al.'s 2025 spatial analysis cannot identify which shared exposure caused the pattern, but it reinforces the practical value of measures that address several vascular pathways at once.

Why the rankings need skepticism

Comparative-risk assessment estimates each dietary factor separately against its own theoretical minimum-risk exposure. Real diets are correlated: someone who eats more whole grains may also eat more nuts and less processed meat. Adding independent attributable fractions can therefore double-count overlapping risk. The summed dietary burden should be treated as a model-based upper-bound estimate rather than a literal count of deaths that would each disappear after one dietary change.

Counterfactuals matter as well. The full text linked to the proposed 2023 primary citation describes a different study: a cohort of 43,272 U.S. male health professionals followed from 1986 through 2016. Its evidence note emphasizes that red-meat risk depends on the food replacing it: replacement with refined carbohydrates showed no excess risk, while replacement with plant proteins showed benefit. That does not validate a global GBD estimate, and it cannot verify any claimed 2023 global death total or factor ranking. It does make one methodological point sharply: "less red meat" is incomplete advice unless the replacement food is named.

What to do with this evidence

The reasonable individual strategy does not require betting on whether a model places low nuts and seeds above sodium. Build meals around whole grains, legumes, fruits, vegetables, and nuts or seeds; favor plant proteins when replacing processed or red meat; and reduce sodium. A 2025 French dietary-modeling study found that a moderate plant-forward pattern was associated with modeled health benefits and limited nutrient-deficiency risk in French adults, although modeling cannot prove what would happen after every individual changes diet.

The strongest claim the evidence supports is modest but actionable: both dietary excess and plant-food insufficiency likely contribute materially to IHD burden, and a plant-forward, lower-sodium pattern targets several plausible pathways at once. The exact global pecking order of dietary culprits remains less certain than the headlines often imply.