Why Cravings Get Worse at Night — and What Your Body Is Doing

The fridge opens at 10:47pm. The person standing in front of it ate dinner three hours ago and is not, by any reasonable definition, hungry. They know this. They open the fridge anyway, and the contents of the freezer become disproportionately interesting in a way that the same freezer at 10:47am never quite manages.

This is one of the most consistent findings in appetite research. Across study designs and across populations, the desire to eat — and specifically the desire to eat calorically dense food — rises in the late evening and peaks somewhere between 8pm and midnight. It is not imagination, and it is not a question of insufficient daytime planning. The body's appetite system runs on a clock, and that clock tilts toward hunger after dark.

The circadian system has a hunger schedule

Frank Scheer, who runs the Medical Chronobiology Program at Brigham and Women's Hospital, has spent more than a decade documenting how the circadian system regulates appetite. In a controlled inpatient protocol published in Obesity, his team kept participants on a 28-hour day for 13 days — long enough that meals and sleep cycled through every clock position relative to their internal circadian phase. The variable being measured was hunger.

The result was clean. Independent of when participants had last eaten, independent of how much they had eaten, hunger was lowest in the biological morning and peaked roughly twelve hours later, in the biological evening. Appetite for sweets, starchy foods, and salty foods followed the same curve, with the largest swings in the evening positions.

The implication is unflattering to willpower-based explanations. The desire to eat at 9pm is not the residue of decisions made earlier in the day. It is what the appetite system does on its own schedule, regardless of the day's caloric arithmetic.

Why the system evolved this way

The evolutionary logic is plausible if not provable. Humans are diurnal, and the period of greatest food availability across most of our ancestral history was the late afternoon and early evening — after the day's foraging or hunting concluded and before the long fasting window of sleep. An appetite system that ramped up in the evening, encouraging substantial intake before the overnight fast, was adaptive when calories were uncertain.

That same system, in an environment of continuous food availability and after-work stress, produces a different outcome: a powerful pull toward food at exactly the time of day when the kitchen is closest, the day's restraint has thinned, and the calories will be least efficiently used.

The cortisol dip changes how rewarding food feels

Cortisol follows a predictable diurnal pattern in healthy adults: a sharp peak about 30 minutes after waking, a gradual decline through the day, and a low point in the late evening. The morning peak is associated with mobilising energy for the day. The evening trough is associated with preparing for sleep.

Cortisol also modulates the brain's reward sensitivity. Research by Elissa Epel and colleagues at UCSF has shown that cortisol responses correlate with cravings for calorie-dense food, particularly under stress. When cortisol is low — as it is in the evening — the reward system is not less active. If anything, it appears more responsive to palatable food cues, partly because the inhibitory braking systems that operate alongside cortisol are also dialled down.

The net effect is that the same slice of cheesecake, encountered at 10pm, produces a stronger anticipatory dopamine response than it would at 10am. The food has not changed. The brain receiving the cue has.

Melatonin tilts glucose handling toward fat storage

Melatonin begins rising about two hours before habitual sleep time, peaks in the middle of the night, and declines toward morning. Its job is to signal darkness to the body, coordinating sleep, body temperature, and a wide range of metabolic processes.

One of those processes is insulin secretion. When melatonin binds to receptors on pancreatic beta cells, it suppresses insulin release. The system is designed for a body that is asleep and not eating — the overnight fast does not require active glucose handling, so the machinery is partially shut down.

If a person eats during the period when melatonin is high, the same meal produces a larger glucose excursion than it would earlier in the day. Marta Garaulet and Scheer demonstrated this in a randomised crossover trial: identical meals consumed during high-melatonin conditions caused significantly worse glucose tolerance. The body is, in effect, less prepared to deal with the food.

For people consistently eating in the late evening, this represents a subtle metabolic penalty applied to every meal in that window. Repeated over years, it contributes to the association between late-night eating and weight gain that observational studies have documented for decades.

Decision fatigue and behavioural drift

The biological factors are not operating in isolation. By the late evening, most adults have made hundreds of decisions across the day — work, family, logistics, social interactions. Self-regulation, treated as a finite resource in the psychological literature, has been drawn on heavily. The capacity to exercise effortful restraint is at its weakest.

Add to this the structural features of the evening: the day's external accountability has ended, the kitchen is physically accessible, screen time often dominates, and the food environment in most homes is heavily weighted toward palatable, calorie-dense options. Each of these would pull behaviour in the same direction even without the circadian and hormonal contributions. Combined with them, the evening is the time of day when staying out of the kitchen requires the most active management.

Night-eating syndrome: when the pattern crosses a threshold

The clinical end of this spectrum is night-eating syndrome, first described by Albert Stunkard at Penn in 1955. Patients consume a disproportionate share of their daily calories after the evening meal — often more than 25% — wake during the night to eat, and report difficulty falling asleep without food. The pattern is associated with mood symptoms, sleep disruption, and elevated obesity risk.

Most people who notice strong evening cravings do not have night-eating syndrome. But the syndrome demonstrates that the same biological pressures that produce ordinary evening cravings can, in some individuals, generate a clinically significant disorder. The continuum is real, and it is biological.

What actually helps in the evening window

Effective interventions tend to work upstream of the craving rather than against it in the moment. Front-loading the day's protein intake — getting 30–40g at breakfast and lunch — appears to attenuate evening hunger more reliably than equivalent calories distributed differently. Earlier dinner timing, with most calories consumed before 7pm, gives the circadian and melatonin systems more aligned conditions. Adequate sleep matters substantially: the literature on partial sleep restriction shows next-day ghrelin rises and leptin falls, amplifying evening hunger the following day.

A few less obvious levers deserve mention. Light exposure in the morning anchors the circadian system more firmly, which sharpens the morning hunger trough and modestly reduces the evening peak. Caffeine after about 2pm can erode sleep quality even when subjective sleepiness feels unaffected, which then feeds back into next-day appetite dysregulation. Alcohol in the evening lowers blood glucose during the night, which can manifest as nocturnal awakening with hunger or as a particularly intense morning-after craving pattern. These are not heroic interventions, but each removes a small headwind.

The strategies that tend to fail are the ones that try to override the evening pull through resolve in the moment. The biological tilt is structural enough that effortful resistance, repeated nightly, depletes faster than the system requires. The interventions that work best change the conditions under which the evening arrives.

For chronic evening cravings that persist despite these adjustments, the question shifts from behaviour to biology. GLP-1 receptor agonists modulate both the homeostatic and reward components of appetite, and patients on semaglutide and tirzepatide consistently report that evening cravings are among the first things to recede. The mechanism is not magical: the medications dampen the reward-sensitivity tilt that the evening biology amplifies. The fridge stops being interesting at 10:47pm not because resolve improved, but because the signal pulling toward it was reduced at the receptor level.

For a wider picture of how evening intake interacts with sleep, our piece on how sleep deprivation drives weight gain covers the bidirectional loop. For the related phenomenon of post-meal sweet cravings, see why sugar cravings spike after eating.

Key takeaways

• Scheer's controlled circadian work shows hunger and cravings rise predictably in the biological evening, peaking roughly twelve hours after the morning low — independent of meal timing or caloric intake earlier in the day.
• Cortisol falls in the evening, which appears to amplify the reward system's response to palatable food cues; the same dessert feels more compelling at night than in the morning.
• Melatonin suppresses insulin release. Meals consumed during high-melatonin windows produce significantly larger glucose excursions than identical meals earlier in the day.
• Self-regulation capacity is depleted by evening, and the home food environment is most accessible — biology and behaviour pull in the same direction.
• Night-eating syndrome, first described by Stunkard in 1955, sits at the clinical end of a real biological spectrum.
• Effective interventions work upstream: protein front-loading, earlier dinner timing, adequate sleep. For persistent evening cravings, GLP-1 medications target the reward-circuit tilt directly.