No single diet is best for everyone. In head-to-head trials, low-carbohydrate, low-fat, Mediterranean and other patterns produce broadly similar average weight loss when calories are matched, and the difference lies in adherence. The useful question is which pattern a person can sustain while it also meets their medical needs.
Why the trials converge
Diet debates promise a winner. The evidence is quieter: when total calories are similar, most named diets produce similar average weight loss, and the strongest predictor of success is not the macronutrient ratio but whether a person can keep eating that way. Trials pitting low-carbohydrate against low-fat, or one branded plan against another, tend to converge by 12 months, with wide individual variation inside every group.
Four separate outcomes sit behind the word "works": short-term weight loss, long-term maintenance, metabolic markers such as glucose, insulin, lipids and liver fat, and cardiovascular events. A pattern can succeed on one and fail on another, so the comparison below reports them separately. → The WeightExperts Clinical Approach
The patterns, compared
Effect sizes here are qualitative, and individual response varies widely. Randomised evidence on cardiovascular events exists for very few dietary patterns.
| Pattern | Weight loss and maintenance | Metabolic and cardiovascular | Grade |
|---|---|---|---|
| Mediterranean — vegetables, legumes, whole grains, fish, olive oil, nuts; low in ultra-processed food and red meat | Modest loss; relatively good maintenance, being palatable and sustainable | Improves lipids, glucose and liver fat. The only pattern with strong randomised cardiovascular-event evidence → Obesity & Cardiovascular Health | Grade A 🟢 |
| Low-carbohydrate — reduced carbohydrate, higher fat and/or protein | Often faster early loss, partly water; converges with low-fat by ~12 months. Maintenance depends on adherence | Lowers triglycerides and can improve glycaemia in type 2 diabetes. No clear event benefit over other patterns → Obesity & Diabetes | Grade B 🟡 |
| Ketogenic — very low carbohydrate (roughly <20–50 g/day) forcing ketosis | Brisk early loss and strong appetite suppression for some; hard to sustain, with adherence falling | Can markedly improve glycaemia and triglycerides short-term; LDL rises in some people. No outcome evidence, and long-term safety uncertain | Grade C 🟠 |
| Low-fat — reduced dietary fat, higher carbohydrate | Comparable to low-carbohydrate at 12 months when calories match; maintenance variable | Can lower LDL, and may raise triglycerides if refined-carbohydrate heavy. Older cardiovascular evidence mixed | Grade B 🟡 |
| High-protein — protein emphasised across the day | Protein is satiating and preserves lean mass in a deficit, which also supports maintenance. A supporting strategy more than a standalone diet | Neutral to favourable; no direct event data → Exercise, Muscle & Body Composition | Grade B 🟡 |
| High-fibre — fibre-rich whole foods | Modest but real loss through satiety and lower energy density; supports maintenance | Improves glucose, lipids and gut health. Fibre intake tracks with lower cardiovascular risk in cohorts | Grade B 🟡 |
| Plant-based, vegetarian and vegan — centred on plants, minimising animal and ultra-processed foods | Modest loss, reasonable maintenance. Whole-food quality matters more than the label | Improves lipids and often glycaemia; favourable when whole-food based. B12, iron, omega-3 and protein adequacy need attention | Grade B 🟡 |
| Intermittent fasting — alternate-day or 5:2 periodic days of very low intake | Works by cutting total calories and is not superior to daily calorie restriction head-to-head. Workable for some, difficult for others | Improvements largely track the weight lost. No outcome data | Grade B 🟡 |
| Time-restricted eating — eating confined to a daily window, for example 8–10 hours | Small loss, mostly through reduced intake; trials are mixed and often modest beyond calorie effects | Follows the weight lost | Grade C 🟠 |
| Meal replacements and structured calorie restriction — portion-controlled shakes or meals, or formal programmes | Among the more effective behavioural approaches, since structure removes decisions. Maintenance depends on the transition back to food | Tracks weight lost, and total diet replacement can drive diabetes remission → Obesity & Diabetes | Grade A 🟢 |
| Calorie restriction (any pattern) — the shared active ingredient beneath most diets | Reliable while sustained. Maintenance is undermined by the body's set-point defence rather than by weak will → Why Is Weight So Hard to Lose? | Improves with loss | Grade A 🟢 |
| Volumetrics — low-energy-density, high-volume foods for fullness on fewer calories | Modest loss, sustainable for many. A principle more than a brand | Favourable, mirroring high-fibre whole-food eating | Grade B 🟡 |
What actually distinguishes them
Strip away the branding and three features separate good nutrition from the rest.
- Adherence. A diet someone keeps outperforms a theoretically better one they abandon, so the pattern is fitted to preferences, culture, budget and schedule.
- Food quality, not only macronutrients. Whole foods ahead of ultra-processed ones, whatever the ratio; diet quality predicts health beyond calorie count.
- Medical fit. Diabetes, kidney disease, MASLD, high triglycerides, pregnancy and medications all change what is appropriate, so nutrition is matched to the person like any other treatment. → The WeightExperts Clinical Approach
In our practice no universal diet is promoted, because the trial evidence does not support one. Claims that a single eating pattern suits everyone run ahead of what has been shown.
Common nutrition claims and what the evidence shows
| Claim | What the evidence shows | Evidence |
|---|---|---|
| "Carbs make you fat." | Excess calories drive fat gain rather than carbohydrate itself. Low-carbohydrate and higher-carbohydrate diets produce similar loss at matched calories, and whole-grain, high-fibre carbohydrates are tied to better metabolic health. | 🟢 A · Established |
| "Insulin is the cause of obesity" (the carbohydrate-insulin model). | Controlled feeding trials do not show the metabolic advantage the model predicts, and calorie balance dominates. Insulin is involved in fat storage without being the master switch of obesity. | 🟡 B · Promising |
| "Eating fat makes you fat." | Low-fat and higher-fat diets yield similar loss at matched calories. Fat type, unsaturated against trans, matters more for health than fat quantity. | 🟢 A · Established |
| "Sugar is toxic or addictive like a drug." | Added sugars, particularly in sugary drinks, promote weight gain and metabolic harm, which is reason enough to limit them. The addiction label is not established in humans the way it is for drugs. | 🟡 B · Promising |
| "Eating small meals every 2–3 hours stokes metabolism." | Meal frequency barely affects total energy expenditure, since the thermic effect scales with how much is eaten rather than how often. | 🟢 A · Established |
| "Fruit is bad because of the sugar." | Fibre, water and micronutrients blunt the sugar, and fruit intake tracks with lower diabetes and cardiovascular risk. Fruit juice differs: harm from fructose is dose-dependent and driven by sugary drinks and added sugar. | 🟢 A · Established |
| "You can't lose weight after 40 because of metabolism." | Large studies show total energy expenditure is fairly steady from about 20 to about 60, and menopause, muscle loss and activity shifts explain much of the perceived change. Harder, sometimes; impossible, no. → Why Is Weight So Hard to Lose? | 🟡 B · Promising |
| "Detoxes and juice cleanses remove toxins and reset metabolism." | The liver and kidneys already perform this work. No credible evidence shows cleanses remove toxins or improve metabolism, and the weight lost is water and glycogen, quickly regained. → Supplements | ⚫ E · Unsupported |
| "Superfoods and negative-calorie foods burn fat." | No single food meaningfully accelerates fat loss, and no food costs more energy to digest than it provides, since the thermic effect is a fraction of a food's calories. Low-calorie, filling foods do help, through satiety rather than this mechanism. | ⚫ E · Unsupported |
| "'Natural Ozempic' foods work like GLP-1 drugs." | Fibre and protein modestly increase satiety, but no food produces drug-level appetite suppression or weight loss. → GLP-1 hub | ⚫ E · Unsupported |
Questions patients ask
So what actually is the best diet?
The one a person can sustain that also meets their medical needs. In trials, matched-calorie diets produce similar average loss, and adherence and food quality decide the outcome. For heart-attack and stroke prevention specifically, a Mediterranean pattern has the strongest evidence. Grade A 🟢 → Obesity & Cardiovascular Health
Does intermittent fasting or a 16:8 window melt fat faster?
No. Fasting and time-restricted eating help by reducing total intake rather than through a special fat-burning mode, and head-to-head they do not beat daily calorie restriction. They are useful where the structure makes eating less easier. Grade B 🟡
Is calorie counting necessary?
A sustainable calorie deficit is necessary; gram-level tracking is not. Portion control, higher protein and fibre, fewer ultra-processed foods, or meal replacements all create a deficit without a spreadsheet. Grade B 🟡
Can food replace a GLP-1 medication?
No. Protein and fibre modestly increase fullness, but "natural Ozempic" foods do not reproduce the appetite and weight effects of the medications. Nutrition and medication work together rather than interchangeably. Grade E ⚫ → GLP-1 hub
Which eating pattern fits a particular person, given their preferences, comorbidities and medications, is decided individually at the practice.
Written by Darius A. Schneider, MD, PhD · Board-Certified Endocrinologist (ECNU) · Last updated: August 2026 · References: DIETFITS (JAMA) low-carb vs low-fat trial; PREDIMED (Mediterranean diet, CV outcomes); Cochrane and network meta-analyses of named diets; Hall et al. controlled feeding studies; American Diabetes Association nutrition guidance; DiRECT (total diet replacement, diabetes remission). Educational; not individualized advice.
