Diets: The Science Behind the Best, Worst and Most Controversial

Food consumption, by supplying the body with essential and nonessential nutrients, supports growth and health. Individual food preferences and potential changes in eating habits are influenced by complex issues such as food availability and acceptability. Availability depends on physical, political and economic factors. Choosing among the many foods available or potentially available for consumption depends on socioeconomic, cultural and individual factors. Individual foods form part of diets, which are characterized by nutrient content, food combinations and timing. The variety of diets is continually expanding, and some have been extensively studied.

Plant-Based Diets

The Mediterranean Diet

Scientific research into Mediterranean eating habits and their potential health effects began after the Second World War with two scientists in the United States. Leland Allbaugh, field director of a Rockefeller Foundation-funded survey on Crete, described the Cretan diet as " surprisingly good" and " extremely well adapted to their natural and economic resources and needs". It consisted of olives, cereals, legumes (edible seeds), fruit—usually as dessert after dinner—vegetables and herbs, with limited amounts of goat meat, milk, game and fish. Whole-grain bread made from barley, wheat or both was important at every meal, and olive oil supplied a large share of energy intake.

The Mediterranean diet was supported by a second scientific investigation of its health-related characteristics: Ancel Keys' Seven Countries Study. The original design compared diets and lifestyles across seven countries using data from one or more male cohorts in each country. Cohorts using olive oil as their main dietary fat had lower all-cause and coronary heart disease mortality than those in Northern Europe and the United States.

Today, the Mediterranean diet is defined as an eating pattern dominated by plant foods—fruit, vegetables, minimally processed cereals, legumes, nuts and seeds—with moderate dairy intake, mostly fermented products such as cheese and yogurt; low to moderate fish and poultry intake; little red meat; and wine generally consumed with meals. It is the most extensively researched diet as a potential dietary modification of considerable interest for many health outcomes.

An umbrella review of meta-analyses of observational studies and randomized clinical trials, using data from more than 12800000 participants, suggested robust evidence for a protective association between Mediterranean diet adherence and the following outcomes (37 outcomes were examined): all-cause mortality, cardiovascular disease, coronary heart disease, myocardial infarction, cancer, neurodegenerative disease and diabetes. Several mechanisms may mediate its benefits, including reduced blood lipids and inflammatory and oxidative stress markers, improved insulin sensitivity, enhanced endothelial and antithrombotic function, and even reduced neurodegenerative brain changes.

Vegetarian Diets

Vegetarian diets have been adopted since antiquity for ethical or philosophical reasons and as part of religious observance. Since the final decades of the 20th century, however, their health effects and ecological benefits—lower greenhouse gas emissions and reduced water and land use—have attracted increasing attention. Vegetarianism today can encompass a range of eating behaviors characterized by different attitudes, perceptions, motivations, and social and health dimensions. A vegetarian diet can be defined as any eating pattern excluding meat, meat-derived foods and, to varying degrees, other animal products. Plant-based is a broader term describing patterns that do not exclude animal foods but rely largely on foods of plant origin. The Mediterranean diet is an example. Fruit, vegetables, cereals, nuts, seeds and legumes are important components of vegetarian and plant-based diets.

A vegan diet is a strict vegetarian diet consisting only of plant foods and excluding foods and beverages wholly or partly derived from animals. Dairy products are included in a lacto-vegetarian diet, and eggs may be consumed in an ovo-vegetarian diet; the most common version, lacto-ovo-vegetarian, permits both dairy and eggs. A pescatarian diet allows fish, seafood, dairy products and eggs. Finally, flexible eating patterns that include some animal products are broadly classified as flexitarian diets.

Given the number and multidimensional nature of vegetarian eating patterns, isolating a specific biological mechanism is difficult. Effects on many pathways have been suggested, including metabolic, inflammatory and neurotransmitter pathways, the gut microbiota and genomic instability. Greater adherence to vegetarian diets has been associated with lower risks of cardiovascular disease, ischemic heart disease and related death, dyslipidemia, diabetes, certain cancers and potentially all-cause mortality. Associations between adherence and body composition, anthropometric markers and weight-related measures have also been discussed.

The more restrictive the diet, the greater the risk of various nutrient deficiencies. Vegans are particularly vulnerable to vitamin B12 deficiency because vitamin B12 occurs only in animal foods. Lower intakes of other nutrients—including vitamin B2, niacin, iodine, zinc, calcium, potassium and selenium—have also been reported. Supplements or occasional low to moderate consumption of animal proteins (a flexitarian approach) may mitigate potential adverse health effects.

Low-Fat Diets

The balance of fats and carbohydrates—the two macronutrients contributing most to energy intake in modern diets—is a target of various dietary modifications intended to improve weight management and other health outcomes. Low-fat diets for weight loss were already in use before the medical community promoted them to reduce cardiovascular risk. During the 1980s, dietary fats were blamed for coronary heart disease and obesity. Low-fat diets, low-fat products and the low-fat philosophy became increasingly popular.

Although there is no universal definition, diets providing less than 30% of total energy from fats are considered low-fat. Very-low-fat diets provide less than 15% of total energy from fats (for example, a 2000-calorie plan contains <33 g of fat), with approximately 10% to 15% from protein and 70% or more from carbohydrates. The Ornish diet is a very-low-fat vegetarian diet in which fats supply 10% of daily calories (with a polyunsaturated-to-saturated fat ratio >1), and people eat as they wish. Nutritional adequacy depends greatly on individual food choices. Adherence can be challenging because many animal foods, as well as vegetable oils and fatty plant foods such as nuts and avocados, must be restricted. Supplementation with fortified products may be considered in these circumstances.

Low-fat diets have been evaluated in several well-designed studies and dietary interventions, including the Women's Health Initiative Dietary Modification Trial. Its results provided evidence of lower breast cancer mortality among women advised to follow a low-fat diet (20% of calories from fats) than among those following their usual diet, although the difference was small. Low-fat interventions can produce weight loss compared with usual diets but may be less effective than low-carbohydrate or other diets. They may also have various favorable effects on cardiometabolic factors, particularly low-density lipoprotein (LDL) cholesterol, alongside a debated lowering effect on high-density lipoprotein (HDL) cholesterol, and may reduce blood pressure and the risks of myocardial infarction, diabetes and all-cause mortality. However, the type of fat consumed may also matter. There is extensive literature on biological mechanisms linking different fatty acids to cardiovascular disease and total fat intake to obesity. Proposed mechanisms relating low-fat diets to cancer include immune changes, anti-inflammatory changes and fatty acid-related reactive oxygen species production.

Carbohydrate-Restricted Diets

Atkins, Ketogenic and Low-Carbohydrate Diets

Medical reports on the potential health effects of low- or very-low-carbohydrate diets without restrictions on fats, protein or total energy intake date back to the late 19th century. The initial therapeutic focus was weight reduction and management of drug-resistant epilepsy. These diets, which stimulate ketone body production, received their widest promotion through Dr. Atkins' 1972 book "Diet Revolution." Despite their popularity, concerns were raised. The American Medical Association's Council on Foods and Nutrition criticized the supporting scientific evidence and potential harmful health effects, particularly on the cardiovascular system.

In the 2000s, however, some randomized controlled trials found that participants advised to follow the lowest-carbohydrate diets, such as versions of the Atkins diet, lost more weight and had greater improvements in coronary heart disease risk factors than those assigned higher-carbohydrate diets. Although this advantage did not persist in every study's follow-up or maintenance phase and adherence varied, the scientific community began investigating the diet's clinical potential more extensively.

The term ketogenic describes various diets. For most people, consuming only 20 to 50 grams of carbohydrates daily produces measurable ketones in urine. These are called very-low-carbohydrate ketogenic diets. Another classification, used mainly to manage drug-resistant epilepsy, is based on the ratio of dietary fats to combined protein and carbohydrates. In the classic, strictest version, the ratio is 4:1 (less than 5% of energy from carbohydrates); the least restrictive option has a 1:1 ratio (the modified Atkins diet, with approximately 10% of energy from carbohydrates), with several options in between.

Diets containing 50 to 150 grams of carbohydrates daily are considered low-carbohydrate compared with usual intake, but may not trigger the metabolic changes of very-low-carbohydrate diets. In practice, any diet supplying less than 40% to 45% of total energy from carbohydrates—a range representing average carbohydrate consumption—can be classified as low-carbohydrate. Many popular diets fall into this category. The Zone diet provides 30% of calories from protein, 30% from fats and 40% from carbohydrates, with every meal following a protein-to-carbohydrate ratio of 0.75. The Zone, South Beach and other low-carbohydrate diets encourage complex carbohydrates to reduce postprandial serum insulin concentrations.

Carbohydrate-restricted diets induce metabolic changes similar to fasting, including changes in plasma free fatty acids, insulin, glucose and ketone bodies. Since reducing one macronutrient always involves a parallel increase in another, these diets should be evaluated not only for their low carbohydrate content but also for their high fat, protein or combined content.

Evidence from long-term studies (>6 months) suggests that very-low- and low-carbohydrate diets produce weight loss equivalent to, rather than greater than, other diets with more carbohydrates. However, randomized controlled trials show that low-carbohydrate ketogenic diets effectively reduce cardiovascular risk factors—blood glucose, glycated hemoglobin and some, but not all, blood lipids—particularly in patients with overweight or obesity and type 2 diabetes. Ketogenic diets have been reported to substantially reduce weight and fat mass in patients with cancer. Interest is growing in their potential effects on schizophrenia and mood disorders, and their relationship to cognitive function, Alzheimer's disease and other dementias.

The ketogenic diet's anticonvulsant effect may involve several mechanisms that stabilize synaptic function and increase resistance to seizures. Although not fully understood, they include reduced carbohydrates and inhibited glycolysis in the central nervous system; altered neuronal excitability through ATP-sensitive potassium channels following changes in mitochondrial function; inhibition of the mammalian target of rapamycin (mTOR) pathway; and inhibition of excitatory glutamatergic synaptic transmission. Low-carbohydrate ketogenic diets therefore appear to reduce seizure frequency in children with drug-resistant epilepsy. Benefits occur over the short to medium term and appear similar to those of available antiseizure medications. Less restrictive ketogenic diets show some, though not substantial, effectiveness in reducing seizures. These diets may also reduce seizure frequency in adults with drug-resistant epilepsy, although the evidence is uncertain; some promising results have been reported in adults with super-refractory status epilepticus. The most commonly reported adverse clinical effects include gastrointestinal symptoms such as constipation and dyslipidemia, generally associated with saturated fats and fat-rich animal protein sources.

Low-Glycemic-Index Diet

The glycemic index classifies how much a food raises blood glucose on a scale from 1 to 100, providing information beyond macronutrient composition. A food or meal's value depends not only on the nature of its carbohydrates but also on other food and nonfood factors affecting nutrient digestibility or insulin secretion. Nonstarchy vegetables, fruit, legumes and whole-grain cereals generally produce low glycemic responses and form part of a low-glycemic-index diet. Meat, fish, poultry, eggs and most dairy products also have a very low glycemic index. These diets may be a tool for managing diabetes and related cardiometabolic conditions and diseases, particularly glucose dysregulation.

Paleolithic (Stone Age) Diet

There was no universal diet followed by all Stone Age humans. However, isotope studies indicate that hunter-gatherer staples included uncultivated plants—fruit, roots and vegetables—some wild game, fish and occasionally honey. These diets were very high in fiber, with varying fat and protein content, and protein came mainly from animals. Most carbohydrates came from vegetables and fruit, with little from cereals and none from sugar-containing foods or dairy. The Paleolithic or Paleo diet has become popular in the media. Scientific and nonscientific literature uses various definitions, with lists of foods to include—vegetables, fruit, lean meat, fish, nuts and eggs—and exclude—grains or cereals, dairy, legumes, added sugar and salt, and refined oils. It is generally considered low-carbohydrate, although some versions provide up to 45% of daily energy from carbohydrates.

The limited high-glycemic-index and processed foods in a Paleolithic diet may favorably affect insulin resistance and inflammatory mechanisms. The diet has been associated with improvements in insulin resistance and diabetes, dyslipidemia, hypertension and inflammation, as well as body composition and weight-related measures. Evidence of its clinical effectiveness is less extensive than for other diets, and potential benefits remain a focus of research.

The DASH Diet

In the early 1990s, a multicenter randomized clinical trial called Dietary Approaches to Stop Hypertension (DASH) tested how eating patterns affected blood pressure control. Participants assigned to the experimental diet for 8 weeks had greater blood pressure reductions than those on the control diet—on average, 5.5 mm Hg systolic and 3.0 mm Hg diastolic. These findings established DASH as an effective strategy for preventing and treating hypertension. It is rich in fruit and vegetables—five and four servings daily, respectively—and includes two daily servings of low-fat dairy, little saturated fat and cholesterol, and relatively low total fat. Potassium, magnesium and calcium levels approach the percentile corresponding to 75% of consumption in the United States population, and the diet is high in fiber and protein. Basic and clinical research shows that changes in these and other nutrients influence many physiological mechanisms, including blood pressure control, glucose tolerance, inflammation, oxidative stress, fat absorption and adipogenesis.

Since the original report, DASH has been associated with many conditions beyond hypertension. Greater adherence has been linked to significantly lower all-cause mortality, and several observational studies suggest associations with reduced cancer incidence and cancer-related mortality. An umbrella review of meta-analyses, based on prospective cohort data totaling approximately 950000 participants, found that greater adherence was associated with lower incidence of cardiovascular disease, coronary heart disease, stroke and metabolic diseases such as diabetes. Controlled trials have demonstrated reductions in systolic and diastolic blood pressure, metabolic measures including insulin, glycated hemoglobin, total and LDL cholesterol, and body weight. Proposed modifications or refinements include DASH-sodium, combining the diet with reduced sodium intake, and OmniHeart, which partially replaces carbohydrates with protein—approximately half from plants—or unsaturated fatty acids, predominantly monounsaturated fats.

The MIND Diet

The MIND (Mediterranean–DASH Intervention for Neurodegenerative Delay) diet targets a specific health outcome: cognitive function. Based on previous findings about diet and cognition or dementia, it combines features of Mediterranean and DASH diets. It emphasizes plant foods—whole grains, vegetables, legumes and nuts—with particular attention to blackberries and green leafy vegetables. Red meat and other foods high in total and saturated fat—fast and fried foods, cheese, butter and margarine, pastries and sweets—are restricted, and olive oil is the main fat. Fish more than once weekly and poultry more than twice weekly are recommended. MIND has shown some potential cognitive benefits and is being actively studied in randomized clinical trials.

Time-Restricted Diets

Fasting—consuming no food or caloric beverages for periods ranging from 12 hours to several weeks—has been practiced for centuries. Many religions recommend fasting periods as expressions of devotion, self-discipline and renunciation of physical needs, while also offering opportunities for reflection, as in the Muslim month of Ramadan. Clinical research has focused mainly on its long-term effects on aging, metabolic dysregulation and energy balance. Fasting differs from calorie restriction, in which energy intake is reduced by a specified percentage, generally 20% to 40%, while meal frequency is maintained.

Intermittent fasting emerged as a less demanding alternative to continuous fasting. It is an umbrella term for regimens alternating fasting or restricted eating with normal or unrestricted eating. Approaches used so far fall into two main types. In the first, the reference period is a week. Alternate-day fasting involves fasting every other day, each fasting day followed by a day of unrestricted food intake; modified alternate-day fasting alternates a very-low-calorie diet with unrestricted eating. One regimen combines fasting with restricted eating on a very-low-calorie diet for 5 days (for example, the 5:2 diet), with normal eating on the remaining 5 days. Other approaches use 5 days of very-low-calorie restricted eating followed by at least 10 days of unrestricted eating. In the second type, time-restricted eating, the reference period is a day, with eating confined to a particular daily window, usually 8 or 10 hours.

When meals are arranged to accommodate intermittent fasting, a metabolic fuel switch from glucose to fatty acids and ketone bodies increases energy-production efficiency and has multiple effects on cell and organ function, including greater stress resistance. Evidence supporting long-term effects and clinical benefits compared with continuous energy restriction is limited and debated. Nevertheless, preliminary human and animal studies have associated intermittent fasting with improvements in glucose regulation and diabetes; body composition measures, abdominal fat and obesity; cardiovascular health indicators such as dyslipidemia, hypertension and inflammation; and various other conditions and disorders. Its clinical effects on cancer initiation and progression attract interest because it is thought to favorably influence energy metabolism, reduce cancer cells' growth potential and increase their sensitivity to treatment. Similarly, preclinical evidence concerning effects on neurodegeneration and other brain pathways has prompted recent research into potential effects on neuropsychiatric disorders.

Conclusions and Commentary

From a medical research perspective, it makes sense to clarify the effectiveness and biological mechanisms linking particular diets to health outcomes. Agreement on each diet's definition is essential, but the considerable diversity of diets studied makes standardization and assessment difficult. Other challenges include relatively limited randomized trial evidence, reliance on single-blind dietary interventions, variable adherence, dietary measurement errors, potential confounders and other frequently discussed limitations involving effect sizes. Evidence comparing one potentially healthy diet with another is also limited, although a recent study attempted such comparisons, primarily using structured dietary indices.

Diets developed within different historical and cultural settings and for different populations and patient groups can be difficult to translate into everyday clinical practice. Although health effects and transferability have been demonstrated for some ethnic groups and populations, adapting diets to other cultures and underrepresented groups may be challenging. Advances in nutritional biomarkers, imaging and other diagnostic techniques; understanding of the microbiome and food exposome; interactions with the genome; and multi-omics approaches may enable individual profiles for small groups or individuals and more personalized diets. Likely adherence must also be considered: it may decline over time, and more restrictive eating patterns have higher reported dropout rates.

Although many diets were initially developed for weight control, dietary composition has important health effects independently of adiposity. An overview of the literature suggests that plant-based diets with moderate fat content—rich in vegetables, fruit, whole grains, legumes, nuts and unsaturated fats, with low to moderate poultry and seafood, and little red meat and sugar—can provide substantial health benefits. The Mediterranean diet may be an option, at least for some population groups.

Healthcare professionals must be educated about diets and their potential effects on disease. Healthy-eating advice should be integrated into clinical practice and taught in medical curricula. Professionals should recommend variations of plant-based diets while considering individual preferences, cost, affordability and cultural issues.