Medical disclaimer: This article is for informational purposes only and does not constitute medical advice. Dietary interventions can have significant effects on blood glucose, lipid profiles, medication requirements, and other health parameters. Always consult a qualified healthcare professional before making substantial changes to your diet, particularly if you have diabetes, cardiovascular disease, kidney disease, or any other chronic condition.
Few topics in nutrition science have generated more heat and less light than the contest between low-carbohydrate and low-fat diets. For four decades, researchers, clinicians, and commentators have argued as though one side must be categorically correct and the other fundamentally dangerous. The clinical evidence tells a more nuanced story — and understanding it properly can meaningfully change how you approach your own dietary choices.
The modern debate traces its roots to physiologist Ancel Keys and his landmark Seven Countries Study, published in 1958. Keys observed a correlation between dietary saturated fat, serum cholesterol, and coronary heart disease across seven nations and concluded that dietary fat — particularly saturated fat — was the primary driver of cardiovascular mortality. The study was influential, methodologically contested, and ultimately overextended into policy.
In 1977, the McGovern Senate Select Committee on Nutrition and Human Needs codified the fat hypothesis into federal dietary guidelines, advising Americans to reduce total fat intake to 30% of calories and increase carbohydrate consumption. The food industry responded with extraordinary efficiency: by the mid-1980s, supermarket shelves were stocked with fat-free cookies, low-fat salad dressings, and reduced-fat snack products that replaced dietary fat with refined sugars, modified starches, and corn syrup.
The result was paradoxical. As the population dutifully reduced fat intake throughout the 1980s and 1990s, obesity rates accelerated sharply. Between 1980 and 2000, the prevalence of obesity in American adults roughly doubled. This epidemiological incongruence did not disprove the fat hypothesis overnight — causality in nutrition is never that clean — but it planted the first serious doubts that the simple macronutrient reduction model was too crude to explain the crisis.
Into that doubt stepped the low-carbohydrate advocates. Robert Atkins had published his first diet book in 1972, largely ignored during the fat-phobic decade that followed, but by the late 1990s a new wave of researchers — Gary Taubes, Jeff Volek, Stephen Phinney, and others — were constructing a rival hypothesis. The insulin-carbohydrate model proposed that dietary carbohydrate, by driving insulin secretion, was the primary hormonal cause of fat storage, and that reducing carbohydrate would unlock adipose tissue for oxidation in a way that caloric restriction alone could not achieve.
Both hypotheses contained real biology. The question was: what did the controlled clinical evidence actually show when these diets were put head-to-head?
The most rigorous evidence on macronutrient effects on fat loss comes from metabolic ward studies, where participants live in a controlled facility and every calorie consumed and expended is measured. These studies are expensive, logistically demanding, and typically short — but they eliminate the confounding variables that plague free-living dietary trials.
Kevin Hall and colleagues at the National Institutes of Health conducted one of the most consequential of these studies in 2015. Nineteen adults with obesity were admitted to the NIH metabolic ward and fed, in randomised crossover order, a baseline diet followed by either a low-fat diet (reducing dietary fat by 33% with calorie deficit held constant) or a low-carbohydrate diet (reducing carbohydrate by 30% with the same calorie deficit). The diets were isocaloric — total energy intake was identical.
The result directly challenged the insulin-carbohydrate model's strongest prediction. When calories were equated, fat loss was similar between conditions. The low-fat arm produced a modest but statistically measurable advantage in body fat loss over six days. The low-carbohydrate arm produced greater total weight loss — but this was primarily through water and glycogen depletion, not adipose tissue. The insulin-carbohydrate hypothesis predicted that carbohydrate restriction should produce substantially greater fat oxidation even at the same calorie level. The metabolic ward data did not support that prediction with any meaningful effect size.
Hall's team followed this in 2021 with a crossover trial comparing a plant-based low-fat diet against a ketogenic diet in twenty adults with obesity. Participants were housed at the NIH facility for four weeks on each diet and could eat ad libitum — as much as they wished. The low-fat plant-based diet produced spontaneously lower calorie intake: participants consumed approximately 700 fewer calories per day than on the ketogenic diet, largely because of the higher fibre content and lower energy density of the food. The ketogenic diet produced lower insulin levels and higher fat oxidation rates, consistent with the low-carb hypothesis — but participants compensated by eating more protein and fat to meet hunger. Net fat loss over four weeks did not differ dramatically between conditions.
These metabolic ward findings do not settle every question — four to six weeks of controlled feeding cannot tell us what happens over two or three years in a free-living population. But they do establish that the mechanistic case for carbohydrate restriction as uniquely fat-liberating, independent of calorie balance, is not well supported when you measure it tightly.
If metabolic ward studies tell us what happens under controlled conditions, large randomised trials tell us what happens in the real world. The DIETFITS trial, published by Christopher Gardner and colleagues at Stanford in JAMA in 2018, is the most rigorous long-term head-to-head comparison conducted to date.
Six hundred and nine adults with overweight or obesity were randomised to either a healthy low-fat diet or a healthy low-carbohydrate diet and followed for 12 months. Crucially, both groups received identical levels of dietary counselling — 22 group sessions with registered dietitians — and both were instructed to maximise vegetable intake, minimise added sugars, refined grains, and ultra-processed foods, and choose whole-food versions of their respective macronutrient approach. This was not Atkins versus the Standard American Diet; it was a fair, well-matched comparison.
At 12 months, weight loss in the healthy low-fat group averaged 5.3 kilograms, compared to 6.0 kilograms in the healthy low-carbohydrate group. The difference of 0.7 kilograms was not statistically significant. Both groups showed substantial individual variability — some participants lost more than 25 kilograms on each diet, while others gained weight. Secondary metabolic outcomes were mixed: the low-carbohydrate group showed greater reductions in triglycerides and fasting insulin; the low-fat group showed greater reductions in LDL cholesterol.
The DIETFITS team had pre-specified two hypotheses for why outcomes might differ between individuals: insulin secretion status (measured by 30-minute post-glucose-load insulin) and a panel of genetic variants previously linked to macronutrient response. Neither hypothesis was confirmed as a reliable predictor. Genotype did not predict differential weight loss response. Baseline insulin secretion showed a trend toward better response to low-carbohydrate in high-insulin secretors, but this interaction did not reach statistical significance in the primary analysis.
A subsequent study by the same group — the Ebbeling 2018 Framingham trial — did find that participants with the highest baseline insulin secretion showed superior cardiometabolic outcomes on a low glycaemic load diet, suggesting the biology is real but the effect size is modest and context-dependent. The headline conclusion from DIETFITS remains: when diet quality is held constant and both diets are followed with equivalent support, weight loss outcomes are nearly identical at one year.
Acknowledging the null result on total weight loss does not mean low-carbohydrate diets lack meaningful clinical advantages. The evidence is clear in several specific domains.
Glycaemic control in type 2 diabetes is the strongest case. Eric Westman's 2008 trial at Duke enrolled 84 participants with type 2 diabetes and randomised them to either a low-carbohydrate ketogenic diet or a low glycaemic index diet with a 500-calorie daily deficit. After 24 weeks, 95.2% of participants in the ketogenic diet group had reduced or eliminated their diabetes medications — a rate dramatically higher than in the comparison arm, and a finding that has been replicated across multiple independent trials. For individuals managing blood glucose with insulin or sulphonylureas, carbohydrate restriction can rapidly and meaningfully reduce pharmacological burden.
Visceral adipose tissue — the metabolically active fat stored around the abdominal organs, strongly associated with insulin resistance, type 2 diabetes, and cardiovascular risk — appears to respond preferentially to carbohydrate restriction in the early months of an intervention. Multiple trials have shown greater visceral fat reduction with low-carbohydrate diets at three to six months, even when total weight loss is equivalent, likely through the interaction between insulin suppression and enhanced lipolysis in visceral depots.
Triglyceride reduction is consistently more pronounced on low-carbohydrate protocols, reflecting reduced hepatic de novo lipogenesis when dietary carbohydrate is lowered. For individuals with hypertriglyceridaemia — which frequently accompanies metabolic syndrome — this effect can be clinically significant and rapid, sometimes visible within two to three weeks.
Non-alcoholic fatty liver disease (NAFLD) regression has been documented on low-carbohydrate and ketogenic diets in several trials, with reductions in liver fat measurable by MRI within two to four weeks. The mechanism involves both reduced fructose-driven lipogenesis and enhanced hepatic fat oxidation in the low-insulin state.
The case for low-fat diets is less fashionable in current nutrition discourse but is supported by real evidence in specific contexts.
LDL cholesterol response is the most consistent advantage. A subset of individuals — sometimes called hyper-responders — show significant increases in LDL-C on high-saturated-fat ketogenic diets. Whether this translates to greater cardiovascular risk is actively debated, since particle size and ApoB concentration matter as much as LDL-C, but for individuals with familial hypercholesterolaemia or the APOE4 genotype — which confers heightened sensitivity to dietary saturated fat — a lower saturated fat intake may be advisable. Plant-based low-fat diets consistently reduce small dense LDL, which is considered more atherogenic than large buoyant LDL.
Athletic performance at high intensities depends on glycolytic metabolism, and severe carbohydrate restriction can impair power output during explosive or sustained high-intensity efforts. While fat adaptation during endurance exercise is well-documented, the performance penalty for carbohydrate-restricted athletes in team sports, interval training, or strength sports is real and well-established in exercise physiology literature. Glycogen availability matters for efforts above roughly 70% of VO₂ max.
Population-level adherence data also favour moderate or lower-fat approaches for many people. Bread, rice, fruit, legumes, and root vegetables are staple foods across most of the world's dietary traditions. For individuals whose food culture is built around these foods, eliminating carbohydrates imposes social and hedonic friction that pure metabolic modelling ignores. A diet that cannot be maintained at family dinners, in restaurants, or during travel is a diet that will fail in practice regardless of its theoretical properties.
This brings us to what may be the most important variable of all: the diet you can sustain is more important than the diet that is theoretically optimal.
George Foster's randomised trial (2003, NEJM) was one of the first large-scale comparisons to show low-carbohydrate diets outperforming low-fat at six months — a finding that generated enormous attention at the time. The same research group's longer-term follow-up (2010) showed that by 12 months, the weight loss gap had largely disappeared as low-carbohydrate adherence eroded back toward baseline eating patterns. The pattern has been replicated across multiple trials: low-carbohydrate diets produce faster initial weight loss, partly through glycogen and water depletion, and partly through spontaneous calorie reduction driven by protein satiety and the novelty of dietary restriction. Over 12 to 24 months, adherence typically declines and differences narrow or disappear entirely.
This is not an argument against low-carbohydrate diets. It is an argument against treating any single dietary approach as universally correct for all people over all timeframes. The best diet is the one that creates a sustainable caloric environment without excessive restriction-driven deprivation, supports your individual health markers, and is compatible with your social and cultural life over years, not weeks.
The dichotomy between low-carb and low-fat is partly a marketing artifact. Most evidence-supported dietary patterns occupy a moderate position on the macronutrient spectrum. The Mediterranean diet — consistently associated with cardiovascular benefit in large observational cohorts and in at least one large randomised trial (PREDIMED) — is neither low-carb nor low-fat. It includes bread, legumes, and fruit alongside olive oil, fish, and nuts, with limited saturated fat and high fibre. It is, in essence, a diet that emphasises food quality over macronutrient targets.
Carb cycling — alternating lower-carbohydrate days with moderate-carbohydrate days, particularly around training sessions — attempts to capture the metabolic benefits of carbohydrate reduction while preserving performance and long-term adherence. Protein-first strategies, which prioritise adequate protein at each meal before filling the remaining plate with fat or carbohydrate according to preference and tolerance, leverage the satiety-per-calorie advantage of protein without requiring dogmatic macronutrient restriction.
For a detailed breakdown of how protein intake interacts with weight loss outcomes, see our protein intake and weight loss evidence guide. For a broader review of common dietary misconceptions that persist despite the evidence, the weight loss myths evidence review covers several claims not addressed here.
Rather than asking which diet is better in the abstract, the more clinically useful question is: which approach is likely to work better for you given your specific metabolic profile?
If you have two or more features of metabolic syndrome — elevated fasting glucose, high triglycerides, low HDL, elevated blood pressure, or central adiposity — the evidence most strongly supports trialling a low-carbohydrate or ketogenic approach, at minimum for the first three to six months. The combination of triglyceride reduction, insulin sensitisation, and preferential visceral fat loss makes the metabolic case compelling in this population. For a detailed look at the intervention evidence on visceral fat specifically, see the visceral fat reduction evidence guide.
If your primary concern is cardiovascular risk reduction — particularly LDL-C — and you carry the APOE4 genotype or have familial hypercholesterolaemia, a Mediterranean or plant-forward dietary pattern with limited saturated fat is likely the more appropriate starting framework. The APOE4 variant is carried by roughly 25% of the population and meaningfully amplifies the LDL response to dietary saturated fat.
If you are physically active and training regularly at moderate to high intensities, severe carbohydrate restriction may compromise performance and adaptation. A moderate-carbohydrate, high-protein approach typically serves this population better than either extreme, supporting both glycogen availability for performance and protein synthesis for recovery.
If none of the above risk factors are prominently present, adherence becomes the primary determinant of outcome. In that case, the choice between low-carb and low-fat should be driven largely by which dietary pattern you can maintain, enjoy, and integrate into your life over years rather than weeks.
Those interested in how research on metabolic health and body composition interventions continues to evolve can explore this peptide and metabolic research library, which covers peptide and lifestyle intervention literature relevant to these outcomes.
The diet wars have produced a clear finding — it just is not the decisive victory either camp was hoping for. When dietary quality is controlled and sufficient support is provided, low-carbohydrate and low-fat diets produce nearly identical long-term weight loss in most individuals. The insulin-carbohydrate model is not supported as a general mechanistic explanation for obesity in the general population, but carbohydrate restriction does confer specific and meaningful metabolic advantages in individuals with insulin resistance, type 2 diabetes, and elevated triglycerides. The fat hypothesis produced genuinely distorting dietary policy, but fat quality and food processing context matter far more than total fat intake as a percentage of calories.
What the evidence does support, emphatically, is the primacy of food quality over macronutrient ratios. Diets built around minimally processed whole foods — whether they happen to be low in fat or low in carbohydrate — consistently outperform diets built around ultra-processed, hyper-palatable, energy-dense products, irrespective of macronutrient profile. The most powerful dietary intervention available to most people is not choosing the correct side in a macronutrient debate, but systematically reducing their consumption of industrially processed food that neither side of the diet war has ever had anything to say in favour of.
The macros matter less than the food. The food matters less than the consistency. The consistency matters more than almost everything else.
Eli Lilly's orforglipron is the first non-peptide oral GLP-1 agonist with full Phase 3 data. This guide covers the ATTAIN-1 obesity trial, the ACHIEVE-1 T2D trial, weight loss versus injectable semaglutide and tirzepatide, side-effect profile, and Australian timeline implications.
A detailed breakdown of the TRIUMPH-1 Phase 3 trial results for retatrutide — the triple-hormone agonist that achieved up to 28.3% mean weight loss at 80 weeks and ~30.3% in a higher-BMI subgroup at 104 weeks.
Is body recomposition possible? The evidence says yes — under specific conditions. Here's the science behind losing fat and gaining muscle at the same time, including protein targets, training requirements, and realistic timelines.