The dietary literature is famously contested. Where exercise research has converged on a fairly stable picture, nutrition produces a new headline most months and a new bestseller most years. The signal is harder to extract than in other domains because the studies are harder to run — long-term randomised trials of diet are expensive, hard to blind, and complicated by the fact that what people eat for thirty years is not what they eat in a four-week study.
But the signal is there, if you read carefully. A small number of dietary patterns have accumulated substantial evidence across cohort studies, large randomised trials, and mechanistic research. They look more similar than they look different. And they all point at variations of the same picture — plants first, fish present, olive oil and fermented foods present, ultra-processed foods rare, modest portion control, enough protein.
This guide covers what those patterns are, where the evidence is strongest, what's worth paying attention to in midlife and beyond, and where the recent research has changed the conversation.
The most-replicated dietary pattern in the longevity literature is the Mediterranean diet, and the strongest single piece of evidence is the PREDIMED trial. Originally published in the New England Journal of Medicine in 2013 (and re-analysed and re-published in 2018 after methodological corrections, Estruch et al.), PREDIMED randomised approximately 7,500 high-cardiovascular-risk Spanish adults to one of three groups: Mediterranean diet supplemented with extra-virgin olive oil, Mediterranean diet supplemented with mixed nuts, or a low-fat control diet. Both Mediterranean groups showed roughly 30% lower risk of major cardiovascular events than the low-fat control over the median follow-up.
This is, by the standards of the dietary literature, an unusually strong finding. It is one of the few dietary RCTs of meaningful size, duration, and clinical endpoint to produce a clear positive result. Subsequent cohort and observational research has consistently supported the broader picture: higher adherence to Mediterranean dietary patterns associates with lower cardiovascular disease, lower type 2 diabetes incidence, lower cognitive decline rates, and lower all-cause mortality.
What does the pattern actually look like? The components, in roughly the order of consistency:
- Vegetables, legumes, fruit, whole grains, nuts — most of every meal.
- Olive oil as the primary added fat, used liberally.
- Fish and seafood weekly, often more than once.
- Modest dairy, mostly fermented (yoghurt, cheese).
- Modest poultry, eggs.
- Red meat rare (a few times per month, not per week).
- Wine with meals, modest, optional — the alcohol component has aged less well in newer research.
- Sweets and ultra-processed foods rare.
The portion sizes are not large. The eating is unhurried, often shared. The pattern is as much about what is not eaten — frequent ultra-processed snack food, sugary drinks, large meat portions — as what is.
Dan Buettner's Blue Zones work (Buettner & Skemp, Am J Lifestyle Med 2016) identified five geographic regions with unusually high concentrations of centenarians: Sardinia, Okinawa, Nicoya (Costa Rica), Ikaria (Greece), and Loma Linda (California's Adventist community). The dietary commonalities he identified are observational, not randomised, but they line up with the Mediterranean signal in striking ways:
- A predominantly plant-based diet, with legumes (beans, lentils, soy variants) as a centrepiece.
- Modest portions and stop-eating-when-80%-full as a cultural norm.
- A daily glass of wine in some regions, none in others — alcohol is not the load-bearing variable.
- Frequent shared meals, slow eating, in family or community contexts.
- Fish or modest meat in some regions; no animal protein in Loma Linda.
The honest read on Blue Zones: the data is observational, the regions are geographically and culturally heterogeneous, and the methodology has been critiqued in ways worth taking seriously. Some claims about specific regions have been challenged on demographic record-keeping grounds. But the converging dietary picture is a real signal, and it overlaps substantially with what PREDIMED tested in a controlled setting.
A meaningful update to the dietary picture in the past decade has been on protein. The standard dietary reference intake (DRI) of 0.8 g per kg body weight per day was set decades ago and is increasingly understood to underestimate need, particularly in older adults. The PROT-AGE Study Group's 2013 consensus (Bauer et al., JAMDA) recommended at least 1.0–1.2 g/kg/day for older adults, with higher amounts (1.2–1.5 g/kg/day) for those with acute or chronic illness, and 1.5+ g/kg/day for those recovering from severe illness or injury.
Why this matters for longevity: muscle mass is one of the most-predictive biomarkers of healthspan in older cohorts. Sarcopenia — the gradual loss of muscle mass that begins in the third or fourth decade and accelerates after 60 — predicts frailty, falls, and mortality. Adequate protein intake combined with resistance training is the only reliable defence. Adequate protein without resistance training does less; resistance training without adequate protein produces less adaptation. The two work together.
Practically, for most adults in midlife and beyond, this means roughly:
- 1.0–1.2 g protein per kg body weight per day baseline.
- Distributed across meals (the body uses a single large bolus less efficiently than 25–40 g distributed over 3–4 meals).
- From varied sources — fish, legumes, eggs, dairy, modest meat — rather than concentrated in one food group.
- Accompanied by resistance training to translate the protein into actual muscle preservation.
This is one of the meaningful adjustments to make from "general healthy eating" guidelines if you're thinking about healthspan specifically.
The literature on time-restricted eating (TRE) has matured in the past five years. The strongest signal is for gentle protocols — 12 to 14-hour overnight fasts, eating windows of 10 to 12 hours — which align with circadian biology and capture most of the documented metabolic benefits.
A representative recent study published in Cell Metabolism found that a 10-hour eating window (Wilkinson et al., 2020) improved cardiometabolic markers in adults with metabolic syndrome compared to a wider eating window, even with similar caloric intake. The mechanism is thought to involve circadian alignment, glucose regulation, and time for autophagy and metabolic reset overnight.
The aggressive end of the TRE spectrum — 18+ hour daily fasts, alternate-day fasting — has thinner evidence and worse adherence in healthy populations. The 14:10 / 16:8 windows have the most-supported benefit-to-effort ratio.
Practically: finishing dinner by 19:00 and breakfast at 09:00 (a 14:10 pattern) captures most of what TRE has to offer for healthspan-oriented adults without trade-offs against protein adequacy or social eating patterns. Most people who try this for two weeks report sleeping better — which is itself a longevity-relevant downstream effect.
Ayurveda has the most explicit dietary tradition. The framework is the six tastes (sweet, sour, salty, pungent, bitter, astringent), all of which should be present in a balanced meal; warm, freshly cooked food preferred; eating in a calm setting; finishing meals with a sense of ease rather than fullness. Spices used liberally — turmeric, cumin, coriander, ginger — many of which have independent research signals for digestion and inflammation. The convergence with the Mediterranean pattern is more substantial than it first appears: variety of plants, moderation, freshness, the eating environment as part of the practice.
Traditional Chinese Medicine treats the "spleen system" and digestive fire as foundational to all other body systems. The strongest digestive fire is at mid-morning; lighter eating in the evening. Warm, easily digested foods — soups, broths, steamed vegetables — preferred over cold or raw. A modest evening meal completed several hours before sleep. Again, the convergence is striking: TRE and lighter evening meals are an explicit prescription, predating the modern circadian biology research by millennia.
The Mediterranean cultures themselves are a tradition in their own right. The southern Italian, Greek, Cretan, and Spanish food cultures are not so much specific recipes as a structural pattern: fresh ingredients, olive oil, plants in abundance, fish, modest meat, wine with meals, eating slowly, eating with people. The 1960s Seven Countries Study that first identified the Mediterranean signal was looking at populations whose dietary culture had not yet been disrupted by industrialised food.
Blue Zones populations vary in their specific practices — Loma Linda is essentially vegetarian, Okinawa is a sweet-potato-based pattern, Sardinian is closer to Mediterranean — but converge on the structural elements: plants first, modest portions, frequent legumes, slow shared meals, ultra-processed foods rare or absent.
The honest convergence: every long-evidence dietary tradition with a healthspan signal points at variations of the same pattern. The macronutrient ratios differ. The specific foods differ. The structural pattern doesn't.
A few areas where the picture has changed meaningfully in the past decade:
The protein floor has been raised. The 0.8 g/kg DRI is widely understood to underestimate need for older adults; ranges of 1.0–1.5 g/kg/day are now the consensus position for healthspan-oriented adults in midlife and beyond.
Saturated fat has become less central to the dietary discourse than it was. The Mediterranean trial signal happens despite the inclusion of olive oil (which is monounsaturated but high-fat) and dairy. The current evidence weighs ultra-processed food as a stronger signal than saturated fat per se.
Ultra-processed foods have become a distinct category in the research. The NOVA classification (used in many recent studies) treats ultra-processed foods as a separate dietary feature with consistent associations across mortality endpoints, independent of macronutrient composition.
Glucose variability has gained attention. Continuous glucose monitoring research has shown that even non-diabetic individuals show meaningful glucose excursions that may be metabolically relevant. The healthspan implications are still being worked out, and the practical guidance — eat earlier in the day, walk after meals, limit ultra-processed snacks — is unspectacular and consistent with everything else above.
The microbiome's role is increasingly recognised. Plant fibre diversity, fermented foods, and dietary patterns that support a diverse gut microbiome have become a substantial research area. The practical guidance again converges on what was already known: eat varied plants, include fermented foods, limit ultra-processed.
A few claims worth flagging:
Specific superfoods carry less weight than dietary patterns do. The cohort signal is for the pattern, not for the discrete inclusion of any one item.
Personalised macro-prescriptions based on genotype or microbiome testing remain weakly supported. The variability in individual response is real; the predictive value of the testing isn't yet demonstrated at the level the marketing implies.
Aggressive caloric restriction as a longevity intervention has good support in animal models but mixed support in human trials, particularly for adults already at healthy weight. The trade-offs against muscle mass, mood, and adherence are real.
Specific anti-inflammatory diets marketed for chronic disease tend to converge on what the Mediterranean pattern already prescribes. The branded packaging adds little.
Nutrition is the area of healthspan most affected by individual context. Pregnancy, breastfeeding, kidney disease, diabetes, eating disorder history, food allergies, religious or cultural practice, food access — all change what's appropriate. The general guidance above is a starting place, not a prescription. If you have a medical condition, take prescription medication, are pregnant or breastfeeding, or have a history of disordered eating, the conversation belongs with your physician or a registered dietitian, not with a guide on the internet.
If you wanted to take the most-defensible offerings in the dietary literature and assemble a starting place for healthspan-oriented eating, the shape is roughly:
- Plate composition — half the plate plants, a quarter protein, a quarter complex carbohydrate. Olive oil for cooking and dressing. Fermented foods regularly.
- Sufficient protein — roughly 1.0–1.2 g/kg/day for adults in midlife and beyond, distributed across meals, from varied sources.
- A 14:10 overnight eating window — finish dinner by 19:00, breakfast at 09:00.
- Limit ultra-processed foods — not zero, just rare.
- Eat slowly, often with people — the meal environment is part of the practice in every tradition that's accumulated evidence.
That's most of what the literature converges on. Done consistently for years, it captures the substantial majority of what nutrition can contribute to healthspan. The remainder — exotic protocols, specific superfoods, personalised plans — adds modest marginal returns at best, and most of them require the basics first to express any benefit at all.
The traditions agreed on this picture long before the cohort studies caught up. Plants first. Plenty of fish where it's available. Olive oil and fermented foods. Modest portions. Eaten with people. Done daily. Held over decades.
It is, almost surprisingly, that simple.