Longevity

Time-restricted eating: what the trials show

Time-restricted eating means fitting all your food into a fixed daily window, usually 8 to 10 hours, with only water, tea or coffee without milk and sugar outside it. In randomised trials the results are modest: a 16:8 pattern with no other change lost 0.26 kg more than a control group over 12 weeks in 116 adults, which was not statistically significant, while an early 8-hour window from 7:00 to 15:00 combined with weight-loss counselling lost 2.3 kg more than a window of 12 hours or more over 14 weeks in 90 adults with obesity. Over 12 months, adding an 8:00 to 16:00 window to a calorie-counted diet in 139 adults with obesity made no significant difference (8.0 kg lost against 6.3 kg), and no trial to date has run long enough to measure heart attacks, strokes or deaths.

What time-restricted eating is

Time-restricted eating (TRE) sets a daily window for food and keeps the rest of the day free of calories. Patterns are named by fasting hours to eating hours: 16:8 means an 8-hour window, 14:10 a 10-hour window. Unlike 5:2 or alternate-day fasting, which cut calories on particular days, TRE prescribes when you eat, not how much.

The starting point is wider than most people assume: in the TIMET trial, adults with metabolic syndrome ate across more than 14 hours a day at baseline. Most trials therefore ask for a cut of 4 hours or more, down to 8 to 10 hours. The case for an early window is circadian: glucose handling is thought to be better earlier in the day, as our guides on chronotype and morning light explain.

What the trials found for weight

The one trial that tested the window alone is TREAT (JAMA Internal Medicine, 2020). It randomised 141 adults aged 18 to 64 with a body mass index of 27 to 43 to eat what they liked between 12:00 and 20:00 or 3 structured meals a day for 12 weeks, with no dietary counselling; 116 provided data (70 of them men) and 105 finished. The 16:8 group lost 0.94 kg (95% confidence interval 0.20 to 1.68 kg), the control group 0.68 kg, and the 0.26 kg difference between them was not significant. In the 50 measured in person, the 16:8 group did lose 1.70 kg, but its appendicular lean mass index fell 0.16 kg/m2 more than the control group's, a hint that part of the loss was muscle.

The 12-month trial adds calorie limits. In 2022 Liu and colleagues in Guangzhou (New England Journal of Medicine) randomised 139 adults with obesity to 12 months on 1,500 to 1,800 kcal a day (men) or 1,200 to 1,500 kcal (women), eaten within an 8:00 to 16:00 window or at any time. 118 participants (84.9%) completed the year. The window group lost 8.0 kg (95% CI 6.4 to 9.6), the calorie-only group 6.3 kg (4.7 to 7.8); the 1.8 kg gap was not significant (P = 0.11), and waist, body fat, blood pressure, metabolic markers and adverse events followed suit.

A Cochrane review from January 2021 pooled 18 randomised trials of intermittent fasting of all kinds, 1,125 participants followed for 4 weeks to 6 months. Against eating at any time, fasting patterns lost 2.88 kg more (224 participants, 7 studies, low-certainty evidence). Against daily calorie restriction the difference was 0.88 kg in the short term (719 participants, 10 studies, very low certainty) and 0.56 kg beyond 3 months (279 participants, 4 studies, low certainty); both of those intervals include no difference at all, and the reviewers judged neither change clinically significant. The pattern across the trials: a window helps mainly by trimming calories, and once calories are matched little is left over.

Whether the timing of the window matters

Of these trials, the one with the largest weight difference placed the window early. Jamshed and colleagues at the University of Alabama at Birmingham (JAMA Internal Medicine, 2022) enrolled 90 adults with obesity from a weight-loss clinic (mean BMI 39.6, age 43, 80% women). All received the same energy-restriction counselling; half were also asked to eat only between 7:00 and 15:00, the rest over a window of 12 hours or more. The early group kept to its window on 6.0 days a week, eating over 7.6 hours on average.

After 14 weeks the early group had lost 6.3 kg (5.7% of body weight) against 4.0 kg (4.2%), a difference of 2.3 kg (95% CI 0.9 to 3.7), equivalent to eating about 214 kcal a day less. Diastolic blood pressure fell 4 mm Hg more (95% CI 0 to 8), and scores for fatigue, vigour and low mood improved. Body fat fell 1.4 kg more, short of significance, and food intake, activity and sleep were similar. Dropout was high, 24% in the early group and 20% in the control group, so only 59 completed.

Set against TREAT's 12:00 to 20:00 window, when you eat may matter as much as how long, but the trials differ in counselling, length and participants, so no head-to-head result exists. The snag is social: a last meal at 15:00 rules out family dinner. The TIMET compromise, a window of 8 to 10 hours that ended at 19:00 on average, is easier to keep.

Blood sugar, blood pressure and cholesterol

The TIMET trial (Annals of Internal Medicine, 2024) tested a window in people already on treatment. It randomised 122 adults with metabolic syndrome, medication allowed; 108 (89%) completed, 54 in each group, mean BMI 31, 56 of them women. The TRE group was 4 years younger (56.6 against 60.6), the one baseline difference, so the analysis was adjusted for age. Both groups received standard nutritional counselling; the TRE group was also given a personalised window of 8 to 10 hours, at least 4 hours shorter than before, and logged every intake in a smartphone app for 3 months, cutting its eating day from 14.38 to 10.09 hours.

The primary result was modest: HbA1c fell 0.10 percentage points more than with standard care (95% CI 0.003 to 0.19), enough to take a reading from 6.0% to 5.9%. The TRE group lost 2.98 kg (3.3%) against 1.32 kg (1.5%), a gap of 1.66 kg (95% CI 0.32 to 3.00), and its LDL cholesterol fell 9.11 mg/dL further (95% CI 1.87 to 16.36). Fasting glucose fell 4.83 mg/dL inside the TRE group, HOMA-IR 0.89 and systolic pressure 3.48 mm Hg, but the gaps against standard care, 3.33 mg/dL, 0.51 and 1.76 mm Hg, all cross zero and the authors call them imprecise. One participant reported irritability, fatigue and trouble concentrating; no other adverse events were recorded. The authors list the limits: 3 months is short, diet was self-reported, several elements changed at once.

What is not known, and who should be careful

No trial has run long enough to measure what most people take up TRE for. The Cochrane reviewers found no data at all on deaths, heart attacks, strokes or heart failure in their 18 trials; the longest run 12 months, the Guangzhou trial among them, and most last 12 to 14 weeks. Side effects were reported in only 4 studies, mainly mild headaches. Nearly all participants have been overweight, middle-aged adults, and most pooled trials excluded anyone who already had cardiovascular disease, so safety and benefit over years, or for people at a healthy weight, with a heart condition, the young or the old, are unknown. A 2024 conference abstract linked windows under 8 hours to more cardiovascular deaths in a US survey, but it rested on 2 days of self-reported meal times, was not peer reviewed and could not separate the window from the kind of person who eats that way.

TREAT's 0.16 kg/m2 loss of limb lean mass argues for enough protein inside the window and for strength training. The trials also came with meal logging and clinic support: the Alabama group kept its window 6.0 days a week, and the results belong to that consistency.

Talk to a doctor before shortening your window if you take insulin or other glucose-lowering drugs (a long gap without food can cause hypoglycaemia), if you are pregnant or breastfeeding, under 18, underweight, have heart disease or another cardiovascular condition, have kidney disease, or take medicines that must go with food. The Cochrane authors name diabetes and eating disorders as the groups where benefit and harm are still unknown; a fixed fasting clock is a poor fit after disordered eating. The Eating Window Planner shows a 12:12, 14:10 or 16:8 pattern on a 24-hour clock, and healthspan versus lifespan puts a 1 to 2 kg gain in context.

Starting a shorter eating window

Frequently asked

Does 16:8 work if I change nothing else?

Barely, in the one trial that tested it. TREAT split 116 adults between a 12:00 to 20:00 window and 3 structured meals a day for 12 weeks; the window group lost 0.94 kg, the control group 0.68 kg, and the 0.26 kg difference was not statistically significant.

Is an early or a late eating window better?

The early window has the stronger trial result: 7:00 to 15:00 plus counselling lost 2.3 kg more than a window of 12 hours or more over 14 weeks in 90 adults with obesity, while a 12:00 to 20:00 window without counselling lost only 0.26 kg more than control. The two trials differ in design, so there is no direct comparison yet.

What can I have during the fasting hours?

Water, tea and coffee without milk or sugar. Anything with calories, including milk in coffee, a small snack or a sweet drink, ends the fast; the TREAT trial asked participants to take no calories at all for 16 hours a day.

Does time-restricted eating lower blood sugar?

Modestly. In 108 adults with metabolic syndrome, an 8 to 10 hour window for 3 months lowered HbA1c by 0.10 percentage points and LDL cholesterol by 9.11 mg/dL more than standard care, with 1.66 kg more weight lost. The differences in fasting glucose, insulin resistance and blood pressure all crossed zero, so they cannot be called a result.

Is an 8-hour eating window bad for the heart?

Nobody knows either way. Randomised trials of up to 12 months found no difference in adverse events, but none has measured heart attacks, strokes or deaths, and a preliminary 2024 survey analysis linking short windows to cardiovascular deaths relied on 2 days of self-reported meal times and could not show cause.

Who should not try time-restricted eating?

Anyone on insulin or other glucose-lowering medication without medical advice, women who are pregnant or breastfeeding, people under 18 or underweight, anyone with a history of an eating disorder, and people with heart disease, kidney disease or medicines that must be taken with food. The trials excluded most of these groups, so ask your doctor first.

Sources

• Lowe et al. 2020, JAMA Internal Medicine: Effects of time-restricted eating on weight loss and other metabolic parameters in women and men with overweight and obesity, the TREAT randomized clinical trial · 141 randomised, data from 116, 105 completed the 12 weeks; aged 18 to 64, BMI 27 to 43, 60.3% men, eating 12:00 to 20:00 against 3 structured meals; TRE lost 0.94 kg (95% CI 0.20 to 1.68), control 0.68 kg, difference 0.26 kg (95% CI from 1.30 kg more to 0.78 kg less, P = .63); in-person cohort of 50: TRE 1.70 kg, appendicular lean mass index 0.16 kg/m2 lower than control; no difference in estimated energy intake. Open access copy on PubMed Central
• Liu et al. 2022, New England Journal of Medicine: Calorie restriction with or without time-restricted eating in weight loss · 139 adults with obesity, Guangzhou, 12 months, 8:00 to 16:00 window plus 1,500 to 1,800 kcal (men) or 1,200 to 1,500 kcal (women) against calorie restriction alone; 118 (84.9%) completed; 8.0 kg (95% CI 6.4 to 9.6) against 6.3 kg (4.7 to 7.8), net 1.8 kg (95% CI from 4.0 kg more to 0.4 kg less, P = 0.11); waist, BMI, body fat, lean mass, blood pressure and metabolic risk factors consistent with the primary outcome; adverse events similar; registered as NCT03745612. Linked to the Europe PMC record of the paper because the journal page answers 403 to automated requests
• Jamshed et al. 2022, JAMA Internal Medicine: Effectiveness of early time-restricted eating for weight loss, fat loss, and cardiometabolic health in adults with obesity, a randomized clinical trial · 90 adults with obesity (mean BMI 39.6, age 43, 80% women), University of Alabama at Birmingham, August 2018 to April 2020, 14 weeks; 7:00 to 15:00 window plus energy restriction against a window of 12 hours or more plus the same counselling; adherence 6.0 days a week, mean window 7.6 hours; weight 6.3 kg (5.7%) against 4.0 kg (4.2%), difference 2.3 kg (95% CI 0.9 to 3.7), equivalent to 214 kcal a day; body fat 1.4 kg (not significant); diastolic blood pressure 4 mm Hg (95% CI 0 to 8); mood scales improved; dropout 24% and 20%, 59 completers. Open access copy on PubMed Central
• Manoogian et al. 2024, Annals of Internal Medicine: Time-restricted eating in adults with metabolic syndrome, a randomized controlled trial (TIMET) · 122 randomised, 108 completed (89%), 54 per group, mean age 56.6 (TRE) and 60.6 (standard care), BMI 31.50 and 30.95, 56 women, medication allowed; standard counselling against counselling plus a personalised 8 to 10 hour window (at least 4 hours shorter) for 3 months, logged by smartphone app; baseline window 14.38 hours (TRE) and 14.00 (standard care), TRE window 10.09 hours across the 3 months and 9.83 hours in the final 2 weeks (4.55 hours shorter than baseline); HbA1c 0.10% lower (95% CI 0.003 to 0.19); weight 2.98 kg (3.3%) against 1.32 kg (1.5%), between-group 1.66 kg (95% CI 0.32 to 3.00); LDL 10.48 mg/dL within the TRE group against 1.37, between-group 9.11 mg/dL (95% CI 1.87 to 16.36); fasting glucose 4.83 against 1.50 mg/dL but between-group only 3.33 (95% CI from 7.32 more to 0.65 less); HOMA-IR between-group 0.51 (95% CI from 1.66 more to 0.65 less); systolic between-group 1.76 mm Hg (95% CI from 7.76 more to 4.24 less); diastolic between-group 2.54 mm Hg (95% CI from 6.39 more to 1.30 less), all crossing zero and described by the authors as imprecise; one participant reported irritability, fatigue and trouble concentrating; limits: short duration, self-reported diet. Open access copy on PubMed Central
• Allaf et al. 2021, Cochrane Database of Systematic Reviews: Intermittent fasting for the prevention of cardiovascular disease · Search 12 December 2019, 39,165 records, 26 studies met criteria, 18 pooled with 1,125 participants over 4 weeks to 6 months; no data on all-cause mortality, cardiovascular mortality, stroke, myocardial infarction or heart failure; fasting against ad libitum 2.88 kg (95% CI 1.80 to 3.96; 224 participants; 7 studies; low certainty); against continuous energy restriction 0.88 kg short term (95% CI 0.00 to 1.76; 719 participants; 10 studies; very low certainty) and 0.56 kg medium term (95% CI from 1.68 kg more to 0.56 kg less; 279 participants; 4 studies; low certainty); glucose 0.03 mmol/L; changes in weight and glucose not deemed clinically significant; side effects reported in 4 studies, mild headaches; most included studies excluded people with a history of cardiovascular disease; further research needed in diabetes and eating disorders. Open access copy on PubMed Central

Read on

Tools that fit

This guide is for general information. It does not replace medical advice or an examination.