Eating too fast

Pooled across 23 studies of adults, people categorised as eating quickly had 2.15 times the odds of being obese as people categorised as eating slowly (95% CI 1.84 to 2.51), and their body mass index averaged 1.78 kg/m2 higher (95% CI 1.53 to 2.04). At a height of 1.70 m that difference works out to roughly five kilograms - our arithmetic, not the study's. Two honest qualifications belong in the same breath. An odds ratio of 2.15 is not "twice as likely"; for an outcome as common as obesity the odds ratio runs ahead of the real difference in risk. And the comparison is the fastest self-rated group against the slowest, the two ends of the range rather than typical eaters.

Ohkuma T et al. 2015

The review deliberately excluded interventional studies, so what it pools is people observed, mostly at a single moment in time. That leaves the arrow pointing both ways. Being obese may change how fast someone eats, or how fast they judge themselves to eat, and that fits the data as well as the reverse. The exposure compounds the problem: the pooled studies sorted people into eating-speed categories rather than timing meals. In the review that pooled the metabolic evidence, participants were asked to compare themselves with other people, using questionnaires it describes as designed by each study without any shared standard, with anywhere from two to five speed categories. The individual studies also disagreed with each other a great deal. This is a real and consistent signal about a real behaviour - and it is not, on its own, evidence that slowing down changes anyone's weight. The authors' own closing line is that further studies are warranted to determine whether interventions to slow eating are effective for weight control.

Ohkuma T et al. 2015 · Yuan SQ et al. 2021

Twenty-two studies did the harder thing: they manipulated eating rate directly, comparing a slow eating-rate condition with a fast one in the same experiment. Pooled, the slower condition went with lower energy intake in that meal (standardised mean difference 0.45, 95% CI 0.25 to 0.65), and the effect held regardless of which technique was used to slow people down. The second result is the interesting one: there was no significant relation between eating rate and hunger at the end of the meal, or at any point up to three and a half hours later. So the slow condition was not people leaving the table hungry and holding out - the appetite bill did not arrive. What these studies cannot tell you is whether it adds up. They are single meals in laboratories, with no follow-up, so the leap from "ate less at this meal" to "weighs less in a year" is a leap.

Robinson E et al. 2014

A separate review of self-reported eating speed reported higher odds of metabolic syndrome among faster eaters (1.54, 95% CI 1.27 to 1.86, pooled from seven studies), along with higher odds of each of its components: central obesity 1.54, raised blood pressure 1.26, raised triglycerides 1.29, low HDL cholesterol 1.23 and raised fasting blood glucose 1.16. Treat this as supporting texture rather than a second headline. The studies disagreed with each other enormously, most were snapshots, the exposure was again self-rated, and the evidence is overwhelmingly from east Asia - the single non-Asian study in the pool found no association at all. Metabolic syndrome is also a cluster of risk markers rather than a diagnosis, and nobody in this review was followed to an actual illness.

Yuan SQ et al. 2021

You will have heard that a meal should last twenty minutes, or that you should chew each mouthful thirty-two times. Neither number comes from the evidence above. None of these reviews establishes a target meal duration, a chew count, or any threshold at all - the experiments varied the rate and did not hand back a clock. That is worth knowing for two reasons. First, a number you cannot trace is a number you should not obey. Second, a target duration turns a meal into a task, and the behaviour that actually slows people down does not need one.

Robinson E et al. 2014 · Ohkuma T et al. 2015

Nobody has shown that yet, and the anchor review's own authors say so - their closing line is that further studies are needed to determine whether interventions to slow eating are effective for weight control. What is established is narrower and still useful. In experiments where eating rate was deliberately slowed, people ate less at that meal without reporting more hunger afterwards, and in observational data fast eaters are consistently heavier. Treat it as a lever worth pulling, not as a weight-loss method with a result attached.

There is no evidence-based answer, which is why we do not give a number. None of the reviews here establishes a meal duration, a chew count or any threshold. What the experiments manipulated was the rate, and the effect held across the different methods used to slow people down. Putting the fork down between bites slows the rate on its own, and it does not require you to watch a clock while you eat.

That is a live possibility and we are not going to pretend otherwise. The obesity review pooled studies that measured eating speed and body size at the same time, and it deliberately excluded interventional studies, so the direction is genuinely open. This is exactly why the experimental evidence matters - there, researchers changed the eating rate themselves and measured what happened to intake within that meal. Those experiments run in the direction people hope for, but they are single meals in laboratories.