EAT-Lancet and health

How many lives could the Planetary Health Diet save?

The diet was built from health evidence first. Choose how closely the world’s diets follow it and see how many early deaths a year could be prevented, according to a 2024 study in PNAS.

How close to the diet?

Population estimates, not a personal prognosis. Check your own plate

Most of the way The study's healthy reference

14.7 million early deaths prevented each year

If diets in 171 countries reached 120 out of 140 on the Planetary Health Diet Index, about 27 in every 100 deaths could be avoided.

Every day
40,200
Every minute
28
Share of all deaths
27% (24–30%)

Planetary Health Diet Index, out of 140

World average today: 85 This scenario: 120

Where the number comes from

A model built on decades of real diets

  • Diets scored in 171 countries

    The study scored what adults ate in 2018 on the Planetary Health Diet Index, from 0 to 140. No country came close to the reference: the average was 85.

  • Risks from people followed for decades

    How much a better score lowers the risk of dying comes from 206,404 American adults whose diets were recorded again and again over about 30 years. Combined with each country’s deaths in 2019, it gives the deaths that better diets could prevent.

  • Up from about 11 million

    The 2019 EAT-Lancet Commission estimated about 11 million deaths a year from diet-related chronic diseases. The newer estimate counts deaths from every cause; the 2025 Commission reports it as about 15 million, around 40,000 a day.

Cause by cause

Lower risk across the main causes of death

In the three American cohorts, the fifth of people eating closest to the diet had a 23% lower risk of dying from any cause than the fifth eating furthest from it, after accounting for smoking, activity and other factors.

Risk of death, closest vs furthest fifth from the diet (hazard ratio, 95% confidence interval; 1 = no difference)
Cause of death Risk Hazard ratio
All causes 23% lower 0.77 (0.75–0.80)
Respiratory disease 47% lower 0.53 (0.48–0.59)
Neurodegenerative disease 28% lower 0.72 (0.67–0.78)
Cardiovascular disease 14% lower 0.86 (0.81–0.91)
Cancer 10% lower 0.90 (0.85–0.95)

Source: Bui, L. P. et al. American Journal of Clinical Nutrition 120, 80–91 (2024), “Planetary Health Diet Index and risk of total and cause-specific mortality in three prospective cohorts”.

Beyond one country

What cohort studies found

Studies in other populations point the same way. Pooling 17 studies of more than a million people, those eating closest to the diet had a 20% lower risk of dying from any cause.

Risk of death from any cause, closest vs furthest from the diet (hazard ratio, 95% confidence interval; 1 = no difference)
Study Hazard ratio
Nurses' Health Studies and Health Professionals Follow-up Study United States · 206,404 people · Top vs bottom fifth of the index 0.77 (0.75–0.80)
NHANES United States · 42,947 people · Top vs bottom quarter of the index 0.77 (0.71–0.84)
Malmö Diet and Cancer Sweden · 22,421 people · Highest vs lowest adherence group 0.75 (0.67–0.85)
CoLaus|PsyCoLaus Switzerland · 3,866 people · High vs low adherence 0.70 (0.49–0.98)
All studies pooled 17 studies · 1,063,289 people · meta-analysis 0.80 (0.76–0.85)

Not every study agrees: in the British EPIC-Oxford cohort, a higher score went with less heart disease and diabetes but was not clearly linked with total mortality. Each study scores the diet in its own way.

Ten years of estimates

How the estimate has grown

Successive models have put the deaths that better diets could prevent each year at 5 to 8 million, then about 11 million, then about 15 million. They compare different diets, years and causes of death, so they are not strictly comparable, but they all point to millions of lives a year.

Modelled deaths prevented each year by a global shift in diets, in millions
Estimate Deaths a year
2016 Springmann et al., PNAS Global dietary guidelines, up to fully vegan diets, by 2050 5.1–8.1 million
2019 EAT-Lancet Commission Everyone eating the reference diet, three different models 10.9–11.6 million
2024–25 Gu et al., PNAS; EAT-Lancet Commission Diets everywhere raised to a score of 120 out of 140 14.7 million

How to read these numbers

  • They are modelled, not counted. They assume that the lower risks seen in cohort studies are caused by the diet, which observational studies cannot prove on their own, even when they account for smoking, exercise and income.
  • The estimate covers every cause of death among adults. Leaving out deaths from injuries and accidents, which diet does not explain, more than 13 million deaths a year would still be preventable.
  • The risks come from American cohorts and are applied to every country, where diets, health care and the causes of death differ.
  • Estimates of this kind have been questioned: a 2020 reanalysis of the 2019 Commission’s figures for the United States found that the diet would prevent no more deaths there than eating the right amount of energy to avoid under- and overweight would on its own.
  • These are population figures. They say nothing certain about any one person; if you have specific health needs, talk to a health professional.

Sources: Gu, X. et al. PNAS 121, e2319008121 (2024), “Global adherence to a healthy and sustainable diet and potential reduction in premature death”; EAT-Lancet Commission 2025, Rockström, J. et al. The Lancet 406, 1625–1700 (2025); Knuppel, A. et al. The Lancet 394, 213–214 (2019); Zagmutt, F. J. et al. Journal of Nutrition 150, 985–988 (2020); pooled results: Zhu, R. et al. Advances in Nutrition 17, 100666 (2026). Feast on the Planet is not affiliated with the authors, EAT or The Lancet.

Start with one meal

More vegetables, fruit, whole grains, legumes and nuts. Animal foods in flexible amounts. Less, not never.

EAT-Lancet-inspired recipes Check your own plate The climate side