Stress steers what you eat; cortisol explains less of it than you would guess#
Cortisol reaches appetite, and the direction is the one everybody expects: under stress some people eat more, and what they add skews sweet and fatty. Cortisol also genuinely builds central fat — at high enough doses it relocates body fat so visibly that doctors diagnose from it. Both halves of the folklore rest on something real.
What the folklore gets wrong is the scale. In people without a glucocorticoid disorder, the blood cortisol supposed to explain a thickening waist is not elevated; reviews of the human data describe it as normal or low1. The best long-term measure of cortisol exposure we have tracks body-mass index at about r = 0.10 across 26,527 people2. And the two cleanest laboratory tests of whether a stress-induced cortisol spike predicts how much someone then eats disagree with each other. The mechanism is real. "Cortisol belly" as something you have, and can test for, and can buy a fix for, is not what the measurements say.
This page takes cortisol as a stress hormone. Its other job in this cluster — the evening climb after a lost night, and how it compares with ghrelin and leptin — belongs to the appetite-hormone pillar, which audits all four together.
At disease doses, cortisol does exactly what the folklore says#
Start where the evidence is strongest, because it is genuinely dramatic. When cortisol runs high for months — from an adrenal or pituitary tumor, or a long course of prescribed steroids — fat physically relocates. Central and especially visceral depots increase two- to five-fold while the subcutaneous fat of the arms and legs wastes1. It is one of the few endocrine conditions recognizable across a room.
There is even a tidy reason the belly gets it rather than the thighs. Glucocorticoid receptor binding and receptor messenger RNA run two- to four-fold higher in omental (abdominal, around-the-organs) fat than in subcutaneous fat. More receptors mean more response to the same circulating signal, and the depot with the most receptors is the one that expands.
So the mechanism from cortisol to central fat is established, specific, and dose-dependent. Everything sold on the back of it depends on you being somewhere on that dose curve. The rest of this article is about whether you are.
The blood test measures the wrong compartment#
Here is the result that should have killed the popular version and didn't. If common obesity were driven by cortisol the way Cushing's syndrome is, people carrying more fat should show more cortisol in their blood. They generally don't — circulating cortisol in obesity is normal or low1.
The obvious objection is that a single blood or saliva draw is a terrible measurement. Cortisol runs on a steep daily rhythm, peaking shortly after waking and bottoming out late at night, so one sample captures a point on a moving curve rather than your exposure. That objection is fair, and the field answered it: cortisol deposits into growing hair, so a few centimetres of it integrates months of exposure into one number.
Pooling that literature across 146 cohorts and 34,342 people, hair cortisol correlates with body-mass index at r = 0.10 (95% CI 0.08–0.13; 122 studies, 26,527 participants), and the strongest signal in the whole analysis — hair cortisone against waist circumference — reaches r = 0.18 (95% CI 0.11–0.24) from just seven studies2. Square that headline correlation and long-term cortisol accounts for something like 1% of the variation in BMI between people; that squaring is our arithmetic on their figure, not a number the paper reports.
Read the direction and the size together, because they pull opposite ways. The association is consistent and positive — the authors' own summary — which means it is not nothing. It is also small enough that knowing a stranger's months-long cortisol exposure would barely improve your guess at their weight.
| What you can measure | What it captures | What it shows in ordinary obesity |
|---|---|---|
| Morning blood or saliva cortisol | One point on a steep daily curve | Normal or low, not elevated1 |
| Hair cortisol | Months of systemic exposure | r = 0.10 with BMI across 26,527 people2 |
| Cortisol generated inside fat tissue | Local regeneration by 11β-HSD1 | Elevated with obesity — and not visible in blood3 |
The cortisol that shapes fat is made inside the fat#
That table's last row resolves the paradox, and it is the part almost nobody repeats. Fat tissue does not passively receive cortisol from the bloodstream. It manufactures its own, using an enzyme called 11β-hydroxysteroid dehydrogenase type 1, which converts inert cortisone into active cortisol right where the fat cells are. Local activity, not blood concentration, is what the receptors actually see.
The demonstration is elegant. Researchers engineered mice to overexpress that enzyme in adipose tissue only, at roughly the level measured in fat from people with obesity. The mice developed visceral obesity, exaggerated by a high-fat diet, plus insulin-resistant diabetes, hyperlipidemia, and increased eating despite high leptin3. The whole metabolic-syndrome picture, produced by moving cortisol production into one tissue.
Fat tissue makes its own cortisol. That is why a blood test can read normal while the receptors in your abdomen are seeing plenty — the hormone is being generated in a compartment the blood draw never reaches.
Hold the size of that claim carefully. It is a transgenic mouse, and the enzyme's elevation in human obesity is an association whose direction of cause is unsettled — fat tissue may be responding to obesity rather than producing it. What the experiment does establish is that local cortisol is sufficient to produce the phenotype, which is enough to explain why the systemic measurements come back so quiet.
There is a familiar shape here. A named condition, a mechanism people find intuitive, and no measurement in an individual that can confirm they have it — the same structural problem that leaves leptin resistance undiagnosable. When the mechanism is real but the biomarker is missing, the marketing has room to move in.
Two experiments asked whether a cortisol spike makes you eat, and split#
Now the appetite half. The founding study put 59 premenopausal women through a stress session and a control session on separate days, with snacks available afterward. Women classified as high cortisol reactors ate significantly more calories on stress days than low reactors, while eating similarly on control days, and high reactors ate more sweet food across both days; rising negative mood after the stressor also tracked more eating4. That is the origin of nearly every "cortisol makes you crave sugar" headline written since.
A 2024 study ran the same basic paradigm — a standardized psychosocial stressor, then food — in 63 healthy men and women. Cortisol reactivity did not predict how much anyone ate, at all (P ≥ .90). What it did predict ran backwards: lower immediate cortisol reactivity went with higher perceived hunger (β = −0.38; P = .004). The variable that actually separated eaters was sex, with men consuming 147.4 g of ice cream against women's 95.8 g5.
This is a real disagreement rather than two studies talking past each other, and it is worth naming what separates them. Epel split her sample into high and low reactors and measured spontaneous snacking during stress recovery in women only; Degroote treated reactivity as a continuous predictor of intake from a single test food in a mixed-sex sample. Those designs ask slightly different questions, and the between-group comparison is the more forgiving one — but they are close enough that the second failing to reproduce the first should lower your confidence in the mechanism. The behavior everyone recognizes — a stressful day ending in something sweet — is on firmer ground than cortisol's claim to be causing it.
One design does give cortisol a forward-looking role. Following adults for six months, higher baseline morning cortisol predicted greater subsequent weight gain — but so did baseline insulin and self-reported chronic stress, while ghrelin was what predicted cravings6. That is observational, and cortisol arrives in it as one of three correlated markers of a stressed life rather than as an isolated cause.
What to do with a hormone you can't usefully measure#
The practical consequences are narrow, and they are almost all subtractions.
Don't order the test. A morning cortisol reading in someone without symptoms of adrenal disease has no threshold to act on and, at r = 0.10 against BMI, almost no information about your body composition. Don't buy the blockers: nothing sold as a cortisol supplement has been shown to change fat distribution, and given that the operative cortisol is generated inside adipose tissue, a product aimed at the blood level is aimed at the compartment where the signal already looks normal.
What survives is unglamorous and mostly upstream. Sleep is the input with the cleanest experimental effect on cortisol itself: after a night of partial or total sleep loss, evening cortisol ran 37% and 45% higher respectively, with the nightly quiet period delayed at least an hour7 — the fuller context for which is in sleep and belly fat. And the pathway from a stressful week to a heavier body still runs through your mouth, which means it is countable: cortisol does not create a calorie surplus on its own, so the arithmetic of the deficit is unchanged by how stressed you were when you ate. If your weight has stalled, cortisol belongs well down the list of explanations worth ruling out first.
Where stress genuinely earns its place in this cluster is not as a hormone panel but as a state that changes decisions — the mood-driven, comfort-seeking version of eating that a short night makes worse, which is the subject of poor sleep and emotional eating. That is a behavior you can see happening. The hormone underneath it is not.
FAQ#
Is "cortisol belly" a real thing?#
At disease levels of cortisol, unambiguously — visceral fat increases two- to five-fold in glucocorticoid excess while limb fat wastes1. As a description of an ordinary stressed person's waistline, it is far weaker: circulating cortisol in common obesity is normal or low, and long-term hair cortisol correlates with BMI at only r = 0.102. The pattern is real at the extreme and close to undetectable in the middle.
Should I get my cortisol tested if I can't lose belly fat?#
There is nothing useful to do with the result. Cortisol swings across the day, so a single draw measures timing as much as exposure, and even the months-long hair measure explains roughly 1% of the variation in BMI between people. Worse, the cortisol that acts on fat is regenerated locally inside adipose tissue by 11β-HSD13, so a blood level can be entirely normal while the tissue is well supplied. Testing for adrenal disease is a different question, and one for a clinician.
Does stress actually make you eat more sugar and fat?#
Some people, some of the time — and cortisol's role in it is contested. In the founding experiment, women with high cortisol reactivity ate more after a stressor and ate more sweet food generally4. A 2024 study using the same kind of stressor found cortisol reactivity predicted intake not at all, with sex the stronger predictor5. Stress-driven eating is a well-recognized behavior; treating it as a hormone problem rather than a situational one is what the evidence doesn't support.
Sources#
- Lee MJ, Pramyothin P, Karastergiou K, Fried SK. Deconstructing the roles of glucocorticoids in adipose tissue biology and the development of central obesity. Biochim Biophys Acta. 2013;1842(3):473-481.
- van der Valk ES, Abawi O, Mohseni M, et al. Cross-sectional relation of long-term glucocorticoids in hair with anthropometric measurements and their possible determinants: a systematic review and meta-analysis. Obes Rev. 2022;23(3):e13376.
- Masuzaki H, Paterson J, Shinyama H, et al. A transgenic model of visceral obesity and the metabolic syndrome. Science. 2001;294(5549):2166-2170.
- Epel E, Lapidus R, McEwen B, Brownell K. Stress may add bite to appetite in women: a laboratory study of stress-induced cortisol and eating behavior. Psychoneuroendocrinology. 2001;26(1):37-49.
- Degroote C, Renner B, Wickl J, Leven A, Wirtz PH. Eating after acute psychosocial stress in healthy men and women: sex differences and endocrine mechanisms. J Clin Endocrinol Metab. 2024;109(2):e543-e551.
- Chao AM, Jastreboff AM, White MA, Grilo CM, Sinha R. Stress, cortisol, and other appetite-related hormones: prospective prediction of 6-month changes in food cravings and weight. Obesity (Silver Spring). 2017;25(4):713-720.
- Leproult R, Copinschi G, Buxton O, Van Cauter E. Sleep loss results in an elevation of cortisol levels the next evening. Sleep. 1997;20(10):865-870.
