Does Intermittent Fasting Work for Weight Loss? Evidence Review (16:8, 5:2, TRE)

Modestly yes — but only because it reduces calories. Intermittent fasting (16:8, 5:2, ADF, TRE) produces about 3–5% total body weight loss across pooled trials^[2][4]. Head-to-head RCTs find no advantage over plain daily calorie restriction^[1][3][5]. The active ingredient is the deficit, not the clock. Magnitude is far below GLP-1 receptor agonists (15–21% TBWL).

At a glance

• Modest weight loss: ~3–5% TBWL across pooled IF trials. Welton 2020 systematic review^[2] of 27 IF trials in Canadian Family Physician reported 0.8%–13.0% weight reduction (most clustering 3–8%); Patikorn 2021 umbrella review^[4] of 11 meta-analyses spanning 130 RCTs in JAMA Network Open found IF protocols reduced body weight modestly versus ad libitum eating.
• No advantage over daily calorie restriction in head-to-head RCTs. Lowe 2020 TREAT trial^[1], Trepanowski 2017 ADF trial^[3], Liu 2022 NEJM 12-month trial^[5], and Carter 2018 T2D trial^[6] all reached the same conclusion: when calories are matched (or when calorie intake is tracked), the timing window does not produce extra weight loss.
• Compliance is the real differentiator. For some patients, “don't eat before noon” is easier to execute than “eat 500 kcal less every day.” If IF helps you produce a sustained deficit, it works through that mechanism. If it doesn't, switching protocols won't fix the underlying calorie math.
• Lean-mass concern in ADF. Trepanowski 2017 ^[3] documented that alternate-day fasters lost a higher proportion of lean mass than daily calorie restrictors at 12 months. Adequate protein (1.6–2.2 g/kg/day) plus resistance training is mandatory if ADF is the chosen protocol.
• Magnitude gap vs GLP-1s is large. Best-case IF loss (~5% TBWL over months) is one-third to one-quarter of what Wegovy (−14.9% at 68 weeks)^[7] and Zepbound (−20.9% at 72 weeks)^[8] deliver. IF is a lifestyle pattern, not pharmacotherapy.
• GLP-1 + TRE is compatible. Centering the eating window during nausea-tolerant hours (often midday) and front-loading protein is a common pragmatic combination — there is no pharmacokinetic interaction.
• Not for everyone. Pregnancy, eating-disorder history, T1D or insulin-dependent T2D, frailty in older adults, and any sulfonylurea or insulin regimen are contraindications or require clinician supervision.

The protocols: 16:8, 18:6, 20:4, 5:2, ADF, OMAD, eTRE

“Intermittent fasting” is an umbrella term covering several distinct schedules. The names matter because the RCT evidence attaches to specific protocols, not to the general idea of “skipping meals.”

Two points are easy to miss. First, “TRE” (time-restricted eating) is the academic literature's preferred term for the daily window protocols (16:8, 18:6) because it more precisely describes the intervention than the popular “intermittent fasting” label. Second, the 16:8 schedule — by far the most popular protocol on social media — has the thinnest RCT evidence base for superiority over standard calorie restriction. The largest dedicated 16:8 trial (Lowe TREAT 2020) was a null result on the weight-loss superiority question.

The evidence, study by study

Lowe 2020 TREAT: the headline 16:8 RCT

The TREAT (Time-Restricted Eating on Weight Loss) trial^[1], published November 2020 in JAMA Internal Medicine, randomized 116 overweight or obese adults to either a 16:8 time-restricted eating window (eating noon to 8 pm, fasting 8 pm to noon) or a control regimen of three structured meals per day. Participants in both arms were not given an explicit calorie target. At 12 weeks, the TRE arm lost a mean of 0.94 kg (about 2.1 lb) and the three-meal control arm lost 0.68 kg (about 1.5 lb). The between-arm difference was not statistically significant.

DXA body-composition analysis added a secondary finding: participants in the TRE arm lost a higher proportion of lean mass as a fraction of total weight loss than controls. The authors flagged this as a signal for clinical attention — though the absolute lean-mass loss was small in both arms.

TREAT is the trial most often cited when clinicians say “the evidence for 16:8 is weaker than the marketing suggests.” The protocol that dominates Instagram, TikTok, and most consumer wellness coverage has, in its dedicated head-to-head RCT, produced weight loss indistinguishable from a control diet without explicit timing rules.

Welton 2020: the pooled IF magnitude

Welton and colleagues^[2] published a systematic review of intermittent fasting and weight loss in Canadian Family Physician in February 2020, summarizing 27 trials covering 2–104 weeks of IF intervention. The pooled magnitude across protocols was approximately 0.8% to 13.0% body-weight reduction, with the bulk of trials clustering between 3% and 8% TBWL — broadly comparable to what continuous calorie restriction produces over similar timeframes.

The review's clinically useful framing: IF is “an appropriate approach to weight loss that is roughly equivalent to continuous calorie restriction” and may be a reasonable option to offer patients who prefer a structured eating-window schedule over daily calorie tracking. The review does not position IF as superior to standard caloric restriction.

Patikorn 2021: the umbrella review across 130 RCTs

Patikorn and colleagues^[4] published an umbrella review in JAMA Network Open December 2021 — a meta-analysis of meta-analyses, pooling 11 existing systematic reviews and 130 randomized trials examining intermittent fasting and obesity-related outcomes. The umbrella review separated the analysis by protocol family (TRE, ADF, 5:2) and by comparator (ad libitum eating vs active calorie restriction).

The aggregate findings: IF protocols produce statistically significant reductions in body weight, BMI, fat mass, waist circumference, and several cardiometabolic markers when compared to ad libitum (unrestricted) eating. When compared to active calorie restriction, IF's body-weight effect is small and not consistently superior. The review's authors conclude that intermittent fasting is “an effective and safe approach” for weight loss in adults with overweight or obesity, with effect sizes “comparable to those of continuous energy restriction.”

Trepanowski 2017: alternate-day fasting is not superior to daily restriction

Trepanowski and colleagues^[3] randomized 100 metabolically healthy obese adults to alternate-day fasting (25% energy needs on fast days, 125% on feast days), daily caloric restriction (75% of energy needs), or a no-intervention control, for 12 months (6 months weight loss + 6 months weight maintenance), published in JAMA Internal Medicine July 2017. This is the single most rigorous head-to-head RCT of ADF vs daily restriction.

At 12 months, both interventions produced similar weight loss (~5.5% TBWL ADF vs ~5.0% TBWL daily restriction), with no statistically significant difference. The ADF arm had higher dropout (38% vs 29% in daily restriction) — patients found the feast/fast cycle harder to sustain than a flat daily deficit. DXA analysis showed a higher proportion of lean-mass loss in the ADF arm than in the daily-restriction arm at the 6-month weight-loss endpoint.

Net interpretation: ADF works because it produces a weekly calorie deficit, not because it produces some special metabolic adaptation. And it is harder to stick with than daily restriction for most people.

Liu 2022 NEJM: 12-month confirmation

Liu and colleagues^[5] published a 12-month RCT in The New England Journal of Medicine April 2022, randomizing 139 adults with obesity in Guangzhou, China, to either calorie restriction (1,500–1,800 kcal/day men, 1,200–1,500 kcal/day women) with an 8-hour eating window (8 am to 4 pm), or to the same calorie restriction without an eating-window constraint.

At 12 months, both groups had lost a comparable amount of weight: the TRE arm −8.0 kg, the calorie-restriction-alone arm −6.3 kg. The between-group difference (−1.8 kg, 95% CI −4.0 to 0.4) did not reach statistical significance. Secondary endpoints (waist circumference, BMI, body fat percentage, lean mass, blood pressure, glucose, lipids, insulin sensitivity) showed no significant differences between the groups.

The authors' conclusion: “Among patients with obesity, a regimen of time-restricted eating was not more beneficial with regard to reduction in body weight, body fat, or metabolic risk factors than daily caloric restriction.”

Carter 2018: intermittent vs continuous restriction in T2D

Carter, Clifton, and Keogh^[6] randomized 137 adults with type 2 diabetes to either intermittent energy restriction (two 500–600 kcal “fast days” per week, normal eating the other five — a 5:2 protocol) or continuous energy restriction (1,200–1,500 kcal/day men, 1,000–1,200 kcal/day women) for 12 months, published in JAMA Network Open July 2018. Both regimens were designed to be calorically equivalent over a week.

The HbA1c reduction was equivalent in both groups (−0.5% intermittent vs −0.3% continuous, not significantly different). Weight loss was also equivalent (−6.8 kg vs −5.0 kg, P = 0.25). The trial was designed as a non-inferiority study, and intermittent restriction met that bar: not worse than continuous restriction for glycemic and weight outcomes, but not better either.

Important T2D safety caveat the authors emphasized: patients on insulin or sulfonylureas need active medication adjustment on fast days to avoid hypoglycemia. This is the population where the “don't do IF without a clinician” warning is loudest.

The fasting-is-calorie-restriction finding

The convergent finding across Lowe 2020^[1], Trepanowski 2017^[3], Liu 2022^[5], and Carter 2018 ^[6] is the single most important takeaway from the intermittent-fasting literature: the active ingredient is the calorie deficit, not the timing of the eating window. Across four head-to-head RCTs of different protocols (16:8 TRE, ADF, 5:2, eTRE) compared against matched calorie-restriction controls, none has shown a clinically meaningful advantage for the fasting protocol.

This contradicts the marketing claim — popular in books, podcasts, and supplement adjacency — that fasting works through “metabolic switching,” ketogenic shift, autophagy amplification, or some other mechanism beyond simple energy balance. There may be modest metabolic effects in some protocols (Patikorn 2021^[4] documented small improvements in insulin sensitivity and lipid markers in some IF arms versus ad libitum eating), but for the central body-weight question, the timing window is functionally a behavioral adherence tool, not a metabolic intervention.

Practical translation: if eating only between noon and 8 pm helps you naturally cut a few hundred calories per day (because the window forecloses the late-night snack window, or because skipping breakfast removes a 400-kcal meal), then 16:8 will produce weight loss for you — but the weight loss is happening through the calorie reduction, not through some unique fasting biology. If 16:8 makes you ravenous and you overeat at noon to compensate, it won't work.

The lean-mass concern

The Trepanowski 2017 ADF trial^[3] documented a higher proportion of lean-mass loss in the ADF arm than in the daily-restriction arm at 6 months. The Lowe 2020 TREAT trial^[1] flagged the same concern at 12 weeks. The mechanism is straightforward: feast-fast cycles, especially when protein intake on fast days falls below the threshold required to suppress muscle proteolysis (~1.6 g/kg/day per the muscle-protein literature), can push the body into a higher lean-mass-loss fraction than a steady daily deficit.

For most adults this is a small absolute amount of muscle. For older adults (60+), post-menopausal women, and anyone already at sarcopenia risk, it is a meaningful clinical concern. The mitigation hierarchy is the same as the GLP-1 lean-mass-preservation playbook: hit adequate daily protein (1.6–2.2 g/kg/day), front-load protein in the first meal of the eating window, and pair the fasting protocol with resistance training 2–3 times per week. See our companion review of exercise pairing for lean-mass preservation and our protein calculator for the per-kilogram target by body weight and goal.

Magnitude: intermittent fasting vs GLP-1 receptor agonists

Cross-trial caveat: figures above are from independent trials with different populations, durations, and adherence-monitoring methods. They are not head-to-head. IF magnitudes vary widely by protocol, baseline weight, and adherence; GLP-1 magnitudes are at the guideline dose with active titration support.

Intermittent fasting on a GLP-1: pragmatic combination

There is no pharmacokinetic interaction between TRE/IF and any GLP-1 receptor agonist (semaglutide, tirzepatide, liraglutide, orforglipron). The Wegovy and Zepbound FDA labels do not list intermittent fasting, eating windows, or meal-timing patterns as contraindications. For patients on a GLP-1 who want to layer a time-restricted eating window, the practical considerations are behavioral, not pharmacologic:

• Center the eating window during nausea-tolerant hours. For many GLP-1 patients, the worst nausea is in the morning post-injection (for once-weekly semaglutide and tirzepatide, that's often Monday morning). A noon-to-8-pm window often aligns naturally with the period of best appetite and lowest GI side-effect intensity.
• Front-load protein. Hit the protein target in the first meal of the window. The reduced appetite from a GLP-1 makes it easy to under-eat protein and over-skew toward simple carbohydrates and small portions — the worst combination for lean-mass preservation. See our GLP-1 protein guide for per-meal protein targets.
• Watch hydration. A 16-hour fast plus the modest dehydration that some GLP-1 GI symptoms produce is a setup for fatigue, headache, and orthostatic symptoms. Water, electrolytes, and unsweetened tea/coffee during the fasting window are allowed and recommended.
• Don't break the fast with sugar. A first meal of refined carbohydrate + low protein (toast and jam, breakfast pastries, fruit-only smoothies) will spike glucose and insulin sharply after a 16-hour fasted state and often produces a reactive hunger crash 90 minutes later. A protein-anchored first meal (eggs, Greek yogurt, cottage cheese, protein shake with frozen berries) avoids the swing.
• Stop if it feels punitive. The GLP-1 is already doing most of the appetite-suppression work. Layering an aggressive 20:4 or OMAD window on top often produces enough calorie deficit that lean-mass loss accelerates. If the combination is uncomfortable or your weekly weigh-in shows rapid drop with strength loss, ease back to a longer eating window (10–12 hours) and prioritize protein.

Who should not do IF

Intermittent fasting is contraindicated or requires clinician supervision in several populations. These are the standard exclusion criteria across the trial literature and major obesity guidelines:

• Pregnancy and lactation. Caloric and nutrient needs are elevated; restriction is contraindicated.
• Type 1 diabetes. Insulin dosing assumes meal carbohydrate intake; fasting creates hypoglycemia risk.
• Type 2 diabetes on insulin or sulfonylureas. Same hypoglycemia concern. Carter 2018^[6] required active medication adjustment on fast days under clinician supervision.
• History of eating disorders. Restrictive eating-window protocols can reactivate disordered patterns (anorexia, bulimia, binge-eating disorder). Clinician consultation is mandatory.
• Children and adolescents under 18. Caloric and growth needs preclude restriction-based interventions.
• Adults >75 years or with frailty. Sarcopenia risk and the lean-mass-loss signal from Trepanowski 2017^[3] argue against ADF in particular.
• Active eating-context medications. Some medications must be taken with food (metformin, certain antibiotics, NSAIDs); long fasting windows can interfere with tolerable dosing.

Bottom line

• Intermittent fasting produces modest weight loss (~3–5% TBWL on average), comparable to plain calorie restriction. Welton 2020^[2] and Patikorn 2021 ^[4] are the canonical syntheses.
• No advantage over daily calorie restriction in head-to-head RCTs. Lowe 2020 TREAT^[1], Trepanowski 2017^[3], Liu 2022 NEJM^[5], and Carter 2018^[6] all reached null between-arm conclusions.
• The mechanism is calorie restriction, not metabolic magic. The eating window is a behavioral adherence tool, not a unique metabolic intervention.
• Lean-mass loss is a real concern in ADF and aggressive TRE. Adequate protein (1.6–2.2 g/kg/day) plus resistance training is mandatory.
• Magnitude is far below GLP-1 receptor agonists. Wegovy −14.9% (STEP-1^[7]), Zepbound −20.9% (SURMOUNT-1^[8]). IF is a lifestyle pattern, not pharmacotherapy.
• Compatible with GLP-1 therapy. Center the eating window during nausea-tolerant hours; front-load protein; watch hydration.
• Contraindicated in pregnancy, T1D, insulin/sulfonylurea T2D, eating-disorder history, children, and frail older adults. Clinician consultation is mandatory in medication-dependent patients.

Further reading

• GLP-1 protein diet cheat sheet