The Longevity Diet: What a Randomized Trial Tells Us About Eating to Live Longer

We've read a lot of "longevity diet" content. Most of it is a listicle (turmeric, blueberries, olive oil, seven more things) with no honest signal about what moves mortality numbers. It's the kind of content that makes you feel informed without changing anything you eat.

Read the actual nutrition epidemiology, though, and almost every large study converges on the same unsexy conclusion. It's not a superfood. It's not a supplement. It's a pattern.

The clearest way to think about diet and lifespan comes from Fadnes et al. (PLOS Medicine, 2022), a modeling study that translated decades of cohort data into years of life. A sustained shift from a typical Western diet to a healthier one is associated with several added years (more the earlier you start). A follow-up UK Biobank analysis (Fadnes et al., Nature Food, 2023) reached the same place: a 40-year-old who sustains a shift toward a healthier pattern gained up to roughly a decade. That's the estimate our calculator anchors to.

The strongest randomized evidence for the underlying mechanism is the PREDIMED trial (Estruch et al., NEJM, 2018): 7,447 high-cardiovascular-risk adults assigned to either a Mediterranean diet supplemented with olive oil or nuts, or a low-fat control. Over ~5 years, the Mediterranean group experienced roughly 30% fewer major cardiovascular events. One honest caveat: PREDIMED measured cardiovascular events, not lifespan directly, so we anchor the years estimate to Fadnes and use PREDIMED as confirmation that the effect is causal (randomized trials in nutrition are rare and expensive). Observational work agrees: Sofi et al. (2010), pooling ~2M participants, found ~9% lower all-cause mortality with greater Mediterranean adherence.

The important part is the pattern, not the specific cuisine: mostly whole plant foods, some fish, limited ultra-processed foods. That same pattern shows up across the longest-lived populations in the world (Okinawa, Sardinia, Ikaria, Loma Linda California) with dozens of local variations. It isn't the olive oil. It's the overall shape of what's on the plate.

The short answer

• A Mediterranean diet reduced major cardiovascular events by ~30% in the PREDIMED randomized trial (Estruch 2018)
• Observational data (Sofi 2010, 2M+ participants) found ~9% lower all-cause mortality with greater adherence
• The longest-lived populations in the world don't follow any single "diet" — they share a pattern: mostly plants, some fish, minimal ultra-processed food, low-to-moderate alcohol
• The plants matter more than any individual "superfood." Studies show diet quality correlates more with diversity of plant intake than with any specific nutrient
• Perfect adherence isn't required: 80/20 captures most of the benefit

Want to see how your diet is affecting your lifespan estimate? Take the Longlevity assessment →

PREDIMED: a rare randomized trial in nutrition

PREDIMED is unusual in nutrition science because it's a randomized controlled trial: participants were assigned to a diet rather than self-selecting. That distinction matters: it largely eliminates the "people who eat healthy probably also exercise and don't smoke" confounder that plagues observational nutrition research.

The pooled findings:

• ~30% reduction in major cardiovascular events (composite endpoint of myocardial infarction, stroke, or cardiovascular death)
• Effects observed within ~5 years of dietary change
• Benefit appeared whether the Mediterranean diet was supplemented with olive oil or with nuts, both worked

What counts as "Mediterranean-style" in these studies:

• High consumption of: vegetables, fruits, legumes, nuts, whole grains, fish
• Moderate consumption of: dairy, poultry, moderate alcohol (usually wine with meals)
• Low consumption of: red meat, processed meat, refined grains, added sugars, ultra-processed foods

Notably, the research doesn't specify olive oil as the magic ingredient. It's the overall pattern (especially the plant-forward shift away from ultra-processed foods) that drives most of the effect.

The observational evidence base behind PREDIMED is also large: Sofi's 2010 meta-analysis pooled 7 prospective cohort studies covering 2M+ participants and found ~9% lower all-cause mortality, ~9% lower cardiovascular mortality, and ~6% lower cancer incidence/mortality with greater Mediterranean adherence. The direction and magnitude are consistent across the literature.

The counter-intuitive insight: it's about plants, not proteins

Most modern diet debates are framed around macronutrients: "eat more protein," "cut carbs," "add fat back." The longevity data doesn't care much about those. What the research consistently identifies as the strongest signal is plant diversity and minimally processed foods.

The American Gut Project (analyzing microbiome data from thousands of participants) found that people who ate 30+ different plant foods per week had dramatically more diverse gut microbiomes, which correlates with better metabolic markers, better immune function, and lower inflammation. "Plants" here is broad: vegetables, fruits, legumes, nuts, seeds, whole grains, herbs, spices. A bowl of chili with beans, tomatoes, onions, peppers, garlic, and cumin counts as six plants in one meal.

Compare this to the typical "longevity diet" framing which focuses on specific foods (kale, quinoa, blueberries). Those foods are fine. But the research says variety beats any specific superfood. Eating 30 different plants a week matters more than eating 30 kale smoothies.

This is why Blue Zones populations (who eat different foods on different continents) share similar longevity outcomes. Okinawans eat a lot of sweet potato and seaweed. Sardinians eat a lot of beans and whole grain bread. Adventists in Loma Linda are often vegetarian. The specific plants differ; the plant-forward pattern is the same.

Ultra-processed food is the other half of the story

The flip side of "eat more plants" is eat fewer ultra-processed foods (UPFs). A large French cohort (Schnabel et al., JAMA Internal Medicine, 2019) found that a 10-percentage-point increase in the share of ultra-processed food in the diet was associated with a 14% higher all-cause mortality rate.

What's an ultra-processed food? Foods engineered for palatability and shelf stability, typically containing ingredients you wouldn't have in your kitchen: emulsifiers, hydrogenated oils, industrial flavorings, high-fructose corn syrup, preservatives. Things like: most packaged snacks, sugary drinks, fast food, packaged meals.

The mechanism isn't just "UPFs have more calories." They also:

• Disrupt satiety signaling (you eat more before feeling full)
• Drive faster glucose spikes
• Shift gut microbiome composition unfavorably
• Contain additives associated with inflammation

You don't need to avoid them entirely: the Mediterranean research explicitly allows for cultural indulgences. But if UPFs make up more than half of your calories (the U.S. average is around 57%), cutting that to 20-30% is probably higher-leverage than any other dietary change.

How many actual years?

Honest answer: a 9% reduction in all-cause mortality is a relative-risk figure that doesn't cleanly map to a fixed number of years. The best longitudinal estimates suggest a shift from a typical Western diet to a Mediterranean-pattern diet is associated with roughly 2-3 additional years of life expectancy on average, concentrated in healthspan (fewer years lost to cardiovascular disease, diabetes, and some cancers). The Fadnes modeling puts the ceiling higher, up to about a decade, for a larger shift sustained from a younger age. That's the range in the hero stat above: a modest change earns a few years, a big early change can earn far more.

For perspective: that's roughly the same gain as going from inactive to meeting exercise guidelines. Diet is that impactful. The catch is that dietary change is harder to sustain than exercise change for most people: the habit loop is deeper.

The highest-leverage change

The principle that emerges from the mortality data, simplified as much as it can be: shift the defaults. Not the occasional special-occasion meals, but the meals you make on autopilot when you're tired and not thinking about food.

Two shifts do most of the work.

The first is cutting ultra-processed food intake. Not to zero, but from the U.S. average (around 57% of calories) down to 20-30%, for all the reasons in the section above.

The second is adding plant variety. You already have the target: 30 different plants a week, herbs and spices included, and whatever is cheap and in season counts.

Notice what's not in this list: keto, carnivore, intermittent fasting, superfood rituals. Those are individual tools that work for specific people; they don't drive the population-level mortality signal that PREDIMED and the broader Mediterranean diet literature consistently identify.

(If you have a history of disordered eating, please work with a healthcare provider before making dietary changes. This post is educational, not prescriptive.)

Which of these shifts would move your estimate most depends on where you're starting. Longlevity works out the magnitude, and the interaction with sleep and exercise often matters more than the dietary move in isolation.

What doesn't work (or doesn't work as well as people think)

• Extreme elimination diets (strict keto, carnivore, raw vegan) produce short-term metabolic changes but have limited long-term adherence data. Blue Zones populations don't eliminate food groups. They just emphasize different ones.
• Supplements have weak evidence for mortality benefit in healthy adults. The PREDIMED and broader Mediterranean-diet effects come from food patterns, not individual nutrients in pill form.
• "Detoxing" has no scientific basis. Your liver and kidneys handle detoxification; no juice cleanse improves on that.
• Calorie obsession without diet quality. 1,800 calories of ultra-processed food produces worse health outcomes than 2,200 calories of Mediterranean-pattern food, at the same weight.
• Expensive specialty foods. The longevity benefit comes from beans, vegetables, olive oil, and fish, not acai bowls, organic exotic grains, or gold-leaf smoothies.

The bottom line

Diet is one of the three or four highest-leverage longevity behaviors. The research is robust: meta-analyses spanning millions of participants. The pattern is simple: mostly plants, some fish, minimal ultra-processed food. The specific cuisine varies by culture; the pattern doesn't.

For most people starting from a Western default, the biggest wins come from cutting ultra-processed food intake and adding plant variety. You don't need to overhaul everything at once. You need to shift the defaults.

Frequently asked questions

Citations

• Fadnes LT, Økland JM, Haaland ØA, Johansson KA. Estimating impact of food choices on life expectancy: A modeling study. PLOS Medicine. 2022;19(2):e1003889.
• Fadnes LT, Celis-Morales C, Økland JM, et al. Life expectancy can increase by up to 10 years following sustained shifts towards healthier diets in the United Kingdom. Nature Food. 2023;4:1083-1095.
• Estruch R, Ros E, Salas-Salvadó J, et al. Primary Prevention of Cardiovascular Disease with a Mediterranean Diet Supplemented with Extra-Virgin Olive Oil or Nuts (PREDIMED trial). New England Journal of Medicine. 2018;378(25):e34.
• Sofi F, Abbate R, Gensini GF, Casini A. Accruing evidence on benefits of adherence to the Mediterranean diet on health: an updated systematic review and meta-analysis. American Journal of Clinical Nutrition. 2010;92(5):1189-1196.
• Schnabel L, Kesse-Guyot E, Allès B, et al. Association Between Ultraprocessed Food Consumption and Risk of Mortality Among Middle-aged Adults in France. JAMA Internal Medicine. 2019;179(4):490-498.