Five Hundred Calories at One Lunch

Researchers sat 114 people down to a lunch they could eat as much of as they wanted, and weighed what they ate. Three weeks into tirzepatide, participants consumed 524.6 fewer calories at that meal than those on placebo.^[1] The more interesting result is what the drug did not change: their conscious effort to restrict what they ate.

Why an ad libitum lunch

Most of what is known about how much people eat comes from food diaries, which are the weakest instrument in nutrition research — people under-report, forget, and record what they intended to eat. A laboratory meal with unlimited access, weighed before and after, removes all of that.

It buys objectivity at the cost of realism: one meal, in a laboratory, on a single day. What it measures it measures accurately.

The row at the bottom is the finding

People ate five hundred calories less without trying harder not to.

Volitional restriction is the measure of deliberate effort — consciously holding back, choosing less, exercising restraint. It did not move. Everything that did move sits upstream of effort: how hungry people felt, how much they craved, how strongly food in the environment pulled at them.

It also cuts against the framing that people on these drugs are taking a shortcut around discipline. On this evidence they are not exercising more discipline or less — the thing discipline was being applied to got smaller.

The brain imaging, in the order it was found

A narrower finding did reach significance: activation to high-fat, high-sugar photographs specifically was reduced in four regions — the medial frontal and cingulate gyri, orbitofrontal cortex and hippocampus.

That is a sub-category within an imaging analysis, in named regions, after the aggregated outcome came back null. It is a reasonable finding to report and a poor one to lead with, and “brain scans show the drug reduces food reward” is how it will be summarized.

What limits it

• Phase 1, 114 people, six weeks, with the primary measurement at week three. This is a mechanism study, not an efficacy trial.
• One laboratory meal is not a day, a week, or a life. Whether a 500-calorie reduction at lunch persists across every meal is not what was measured.
• The design is asymmetric. Tirzepatide and placebo were blinded; the liraglutide arm was open-label. Participants who know what they are taking report subjective outcomes differently, so that arm cannot be read alongside the other two on appetite measures.
• Everyone was without diabetes, at a BMI between 27 and 50.

For the mechanism proposed for why this effect fades over months — in mice — see why the appetite comes back.