Skipping Breakfast: What Randomized Trials Show About Weight

“Breakfast is the most important meal of the day” is one of nutrition’s most repeated lines, and for weight it rests on thin ground. The claim comes almost entirely from observational studies, which find that people who skip breakfast tend to be heavier. When researchers have instead randomly assigned adults to eat or skip breakfast, the difference has run the other way, by a small margin, in short trials. This post separates what the randomized trials show about weight from what they show about everything else.

Where the reputation came from

The observational record is large and consistent. A 2020 meta-analysis pooled 45 observational studies (36 cross-sectional, 9 cohort) and found that people who ate breakfast infrequently had 48% higher odds of overweight or obesity in the cross-sectional data (odds ratio 1.48, 95% CI 1.40 to 1.57) and a 44% higher relative risk in the cohort studies (RR 1.44, 95% CI 1.25 to 1.66).

The catch is that observational studies cannot separate breakfast from everything else that differs between skippers and eaters, so they establish an association, not a cause. Brown, Bohan Brown and Allison made this argument in a 2013 paper titled “Belief beyond the evidence”: their cumulative meta-analysis showed the association had been re-confirmed so many times (final p below 10^-42) that further observational studies added nothing, while the causal question stayed open. Confidence in the effect, they concluded, had outrun the data, helped along by uninformative studies and overstated reporting.

What happens when people are randomized

The 2019 BMJ review by Sievert and colleagues included 13 trials from high-income countries. In the seven trials that reported weight, participants assigned to eat breakfast weighed 0.44 kg more at the end than those assigned to skip it (95% CI 0.07 to 0.82 kg). In the ten trials that reported energy intake, breakfast eaters consumed about 260 kcal per day more (mean difference 259.79 kcal, 95% CI 78.87 to 440.71). The weight trials lasted an average of seven weeks and the intake trials two weeks. Every one of the 13 trials carried a high or unclear risk of bias somewhere. The authors’ conclusion was still that adding breakfast might not be a sensible weight-loss strategy, whatever a person’s existing habit.

A 2020 meta-analysis in Obesity, restricted to seven randomized trials with 425 participants and a mean length of 8.6 weeks, found that skipping breakfast reduced body weight by 0.54 kg (95% CI -1.05 to -0.03) with no difference in body-fat percentage across the five trials that measured it. It also flagged one metabolic cost: LDL cholesterol was 9.24 mg/dL higher in skippers in the three trials that measured it (95% CI 2.18 to 16.30). Blood pressure, fasting glucose, insulin, HOMA-IR, triglycerides, C-reactive protein, leptin and ghrelin did not differ.

Half a kilogram over two months is not a weight-loss strategy in either direction; it shows that adding a meal does not make people eat correspondingly less later.

Three trials worth knowing by name

The Alabama recommendation trial (Dhurandhar 2014). The largest trial in the set randomized 309 adults with overweight or obesity (BMI 25 to 40) to instructions to eat breakfast daily, to skip it, or to a control group, for 16 weeks. Weight change ranged from -0.53 to -0.76 kg across the groups, with no significant effect of the assignment and no interaction with whether people had been breakfast eaters at the start. Compliance exceeded 92% in both instructed groups.

The Bath Breakfast Project (Betts 2014, Chowdhury 2016). These two six-week trials prescribed either at least 700 kcal before 11:00 or nothing until noon, first in 33 lean adults and then in 23 adults with obesity. In the lean group, breakfast eaters consumed 539 kcal more per day (95% CI 157 to 920) but also burned 442 kcal more per day through physical activity (95% CI 34 to 851), and neither body mass nor fat mass differed. In the group with obesity, morning activity was higher with breakfast (188 kcal per day, 95% CI 40 to 335), but the 24-hour activity and intake differences were not significant and there was no treatment effect on weight or body composition.

The oldest trial in the set (Schlundt 1992). Fifty-two women with moderate obesity followed a 12-week weight-loss programme with or without breakfast, at identical calorie targets. Habitual breakfast eaters assigned to skip lost 8.9 kg against 6.2 kg for those who kept eating it, and habitual skippers lost more when assigned to eat (7.7 kg against 6.0 kg). The interaction fell just short of statistical significance (p below 0.06), and the authors’ reading was that those who had to change their habits most did best. The much larger 2014 trial found no such interaction, so treat it as a hypothesis rather than a rule.

Why eating breakfast does not cut later intake

The intuition behind the breakfast rule is that skipping it drives overeating later. Trials find that compensation happens but is partial. In a crossover study by Levitsky and Pacanowski that tracked a full day’s food, omitting breakfast raised lunch intake by about 144 kcal and hunger ratings rose, yet daily intake still finished roughly 408 kcal below the breakfast days.

The other half of the rule, that skipping breakfast “slows your metabolism”, has not survived measurement either. Resting metabolic rate in the Bath trials was stable to within 11 kcal per day in lean adults and 8 kcal per day in adults with obesity across six weeks in either direction. In a respiration-chamber crossover in 17 adults, Nas and colleagues found that a breakfast-skipping day raised 24-hour energy expenditure by 41 kcal and fat oxidation by 16 g compared with three meals, with dinner skipping raising expenditure by 91 kcal. These differences are too small to matter for weight, but they run opposite to the claim.

Where breakfast does change something is movement. In the lean Bath trial, breakfast eaters were markedly more physically active over the day, and that extra activity went a long way toward offsetting their extra calories; in adults with obesity, the difference was confined to the morning.

Blood sugar is a different question

Glucose handling is where skipping shows a cost, and it belongs apart from weight. In the Nas crossover, the glucose response after lunch was 46% higher and the HOMA index 54% higher on the breakfast-skipping day than on the dinner-skipping day, although 24-hour glucose and insulin secretion were unchanged. In the lean Bath trial, afternoon and evening glucose was more variable in the fasting group by the final week (difference in coefficient of variation 3.9%, 95% CI 0.1 to 7.8). In the Bath trial in adults with obesity, the insulin response to an oral glucose tolerance test fell with breakfast and rose with fasting, at a borderline p value of 0.05.

The clearest signal comes from a small crossover study in 22 adults with type 2 diabetes. On the day without breakfast, the glucose response after lunch was 36.8% higher and after dinner 26.6% higher than on the day with breakfast, with lower insulin and GLP-1 responses and an insulin peak delayed by 30 minutes. This is one day in one small group, but it matches the other crossovers: a long morning fast worsens the response to the meals that follow (the post on meal timing and the gut’s circadian rhythm covers the wider picture). Anyone with diabetes or taking glucose-lowering medication should discuss meal patterns with their clinician rather than experimenting.

Training before breakfast

For morning exercisers the question becomes “eat before or after”. A randomized crossover in 12 physically active young men compared breakfast then rest, breakfast then 60 minutes of cycling, and fasted cycling followed by breakfast. Twenty-four-hour energy balance was +492 kcal after breakfast and rest, +7 kcal after breakfast and exercise, and -400 kcal after fasted exercise (95% CI -571 to -230). Some compensation at lunch occurred, but it did not close the gap. That is a one-day result in lean, fit men, and whether the deficit persists over weeks has not been shown. The same logic underlies the modest results from intermittent fasting: the eating window matters less than total intake.

Bottom line

Randomized trials do not support the idea that eating breakfast helps with weight. Two meta-analyses found that people assigned to eat breakfast ended about half a kilogram heavier than those assigned to skip it, the larger one found they ate around 260 kcal more per day, and the largest trial in the set found no effect of either instruction over 16 weeks. The trials are short (2 to 16 weeks) and none is free of bias concerns, so the safe reading is “no advantage” rather than “skipping works”. In practice:

• Anyone who eats breakfast and is not hungry for it can stop without expecting a drop in resting metabolism; anyone who wakes hungry can eat without expecting a weight benefit.
• Habit change may matter more than the meal itself: the one trial that hinted at an effect found it in people doing the opposite of what they were used to, and even that result was borderline.
• Glucose handling and LDL are the outcomes where skipping showed a cost in short trials. People with diabetes, high cholesterol, or on related medication should ask a clinician before changing meal patterns.
• For how sleep timing and the daily cortisol rhythm shape morning appetite, see the post on the morning cortisol rhythm.

Preference, schedule and how you feel by mid-morning are better guides than the meal’s reputation.