The Dinner-Timing Trial Measured Blood Sugar, Not Weight Loss

The evening body produces melatonin on a schedule, and that hormone does something specific to digestion: it slows insulin's response to sugar. Feed the body a full meal while melatonin is already rising, and the blood glucose spike that follows takes longer to clear. That mechanism, not a personal habit passed between patients, is the actual science behind this year's push to move dinner earlier. It has been tested. Just not on the population currently being told to adopt it.

What the trials actually measured

In 2015, a team led by chronobiologist Marta Garaulet at the University of Murcia gave melatonin capsules to 17 normoglycemic women, average age 24, before a standard glucose tolerance test, then compared the result to a placebo session. The paper, published in Metabolism, found melatonin blunted insulin's response to sugar, more sharply in the eleven women carrying a common variant of the gene MTNR1B, a melatonin receptor roughly half of people of European descent carry a copy of. Five years later, a Boston team ran the more direct version: 20 healthy adults, average age 26, ate dinner at 6pm one night and 10pm another, with sleep fixed from 11pm to 7am both times. Published in The Journal of Clinical Endocrinology and Metabolism, the late dinner produced higher overnight blood glucose and left less of the meal's fat oxidized by morning.

Neither study weighed anyone before and after. Neither enrolled a single woman over 30, let alone over 50. What they measured was glucose handling, not kilos.

A French term doing more work than it should

The dinner-timing advice circulating this year echoes chrononutrition, a term the French physician Alain Delabos coined in 1986 to name a diet he built for himself after quitting smoking. His method is not "eat dinner before a cutoff." It assigns four meals to four enzyme windows he proposed, breakfast heavy in fat and protein, dinner reduced to white fish and vegetables, on the theory that the body processes each nutrient differently by hour. That original method has not itself been through a controlled trial. The melatonin and insulin research above is a separate, later body of work that Delabos's method now benefits from by association.

Where 50 actually fits, and doesn't yet

There is a real reason this advice targets people past 50: declining estrogen after menopause is linked to reduced insulin sensitivity, and a 2024 meta-analysis of 17 randomized trials covering more than 29,000 postmenopausal women found hormone therapy partly restores it. That makes the melatonin mechanism plausible for this group. It does not mean anyone has measured it in this group. No trial has isolated what an earlier dinner does to glucose tolerance in a body already navigating that hormonal shift, which makes a doctor's collective account of her patients a testimonial, not a finding.

What the data support is narrower and less dramatic: eating close to your own bedtime, whatever time that is, appears to blunt glucose control overnight. Losing weight while asleep is not what anyone measured.

Sources: Metabolism, via PubMed · The Journal of Clinical Endocrinology and Metabolism, via PubMed