Physiology

Stress Hormones And A Fasting Number: What The Endocrinology Literature Says

“Stress raises blood sugar” is one of the most confidently repeated lines in wellness writing, and the mechanism behind it is real and well documented in a laboratory. Whether an ordinary bad week actually moves tomorrow morning's fasting reading is a separate question, and the cohort studies that have tried to answer it do not all agree with each other.

By GlucoPril Team · Last updated September 27, 2026 · 11 min read · Physiology

The mechanism: what stress hormones actually do to glucose

The body's acute stress response involves several hormones working on different timescales, and the two most relevant to a glucose reading are epinephrine (adrenaline), released within seconds by the adrenal medulla, and cortisol, released over minutes to hours by the adrenal cortex. Both are sometimes lumped together in popular writing as “stress hormones that raise blood sugar,” as though they worked the same way. The physiology literature says they do not, and the difference matters for how much a single stressful evening should actually be expected to move a fasting reading the next morning.

Epinephrine: a direct, fast, well-documented effect

A study published in the American Journal of Physiology in 2009 infused epinephrine into seven fasting, healthy young volunteers at a rate designed to mimic a real physiological stress response, while holding insulin and glucagon artificially constant so the epinephrine effect could be isolated. Plasma epinephrine rose roughly twentyfold, and plasma glucose rose from 4.3 to 13.3 mmol/L (about 77 to 240 mg/dL) within 90 minutes — a 2.5-fold increase in the body's rate of glucose production. Using magnetic resonance spectroscopy to watch liver glycogen directly, the researchers showed this happened in two phases: the first 60 minutes came mostly from the liver breaking down stored glycogen, and the second phase, as glycogen ran low, shifted to the liver manufacturing new glucose from scratch (gluconeogenesis), which had roughly doubled by the end of the infusion.

This is about as direct as physiological evidence gets: raise epinephrine on its own, with everything else held constant, and glucose production rises sharply and quickly. It is also an extreme, controlled laboratory dose, roughly the concentration seen in a genuine fight-or-flight response, not the low-grade elevation of an ordinary annoying day. What it establishes is the mechanism, not the everyday magnitude.

Cortisol: permissive, not direct

Cortisol's role is a genuinely different story, and one the popular framing usually gets wrong. A 2017 study from the University of Edinburgh infused cortisol into 20 healthy men at low, medium and high concentrations, in combination with controlled insulin and adrenaline levels, and measured glucose and fat metabolism directly. Even at a high dose, cortisol on its own did not alter glucose kinetics. Its clear, measurable acute effect was on lipolysis — the breakdown of fat into circulating fatty acids — and that effect only appeared when insulin or adrenaline was also elevated. The authors' own description of this relationship is precise: cortisol's acute effects on fuel metabolism are permissive, meaning it sets the stage for other hormones to act more strongly, rather than direct, meaning it does not move glucose by itself.

This is the opposite of the epinephrine finding in an important way. The hormone most often blamed for a “stress belly” of high glucose is, acutely, the one that did not move glucose on its own in a controlled human trial. The hormone that actually did the moving was the fast-acting one that gets less attention in popular writing.

Three cohorts, three different answers

The laboratory studies above isolate a mechanism under artificial, controlled conditions. Whether everyday psychosocial stress, over months or years, actually predicts a worse glucose outcome in real populations is a separate and much harder question, and the three largest cohort studies to ask it directly do not agree.

Three real cohorts, three different sample sizes, three different conclusions. This is not a case of one study being right and two being wrong.

Why these studies do not agree with each other

The disagreement is informative rather than simply a mess. The French IPC study is the clearest illustration of why: pooled across its entire 22,567-person cohort, perceived stress showed no overall link to incident diabetes at all. Split by occupational status, though, the picture split in opposite directions — stress predicted more diabetes in workers with less job control, and less diabetes in workers with more of it, a statistically significant interaction the authors could not explain with standard risk factors. Averaging opposite effects together produces a null result that hides two real, opposing signals underneath it. The Beijing cohort, which found a clear positive association, used a validated occupational-stress questionnaire (the Copenhagen Psychosocial Questionnaire) in a single, more occupationally homogeneous population, and specifically measured cortisol as part of its mediation analysis rather than inferring it. The Danish cohort was, by its own authors' admission, too small (355 people) for its estimates to reach statistical significance, even where a real underlying tendency might exist.

Put together, the honest reading is not “stress doesn't affect glucose” or “stress definitely causes diabetes.” It is that the relationship is real enough to show up clearly in some populations and study designs, small enough to wash out in others, and dependent on factors — like how much control a person has over the source of their stress — that a single fasting reading cannot capture.

What this means for a morning reading

The mechanism section above explains why a single very bad night, or an acute fright, could plausibly nudge a morning glucose reading through an epinephrine surge, in the same family of physiology covered on this blog for sleep. It does not establish that an ordinary stressful week, without an acute adrenaline spike attached to it, reliably does the same thing at the individual level — the cohort evidence above is genuinely split on that larger claim. What both the mechanism studies and the cohort studies agree on is that cortisol, the hormone most commonly blamed by name, is not the one doing the direct work.

Who the cohort split actually describes

The French IPC finding deserves one more pass, because “it depends on occupational status” is easy to read as a throwaway line rather than the headline result it actually was. Workers with lower occupational status — less autonomy over their own schedule, less control over how their work gets done — showed a real, statistically significant increase in diabetes risk as perceived stress rose. Workers with higher occupational status showed the opposite: rising stress was associated with a lower diabetes risk in that group, a direction that seems counterintuitive until it is read as a proxy for control. A demanding job a person has real authority over may carry a different physiological signature than an equally stressful job they have no say in. Neither the French cohort nor the Chinese cohort measured control directly in a way that let them test that explanation head-on, so it remains a plausible reading of the pattern rather than a proven mechanism, and it is offered here as exactly that.

Where the honest answer lands

“Stress raises blood sugar” is not wrong, but it compresses two different hormones with two different jobs into one sentence, and it skips past three cohort studies that could not agree on how much everyday stress matters outside the laboratory. The mechanism is real. The size of its effect on an ordinary person's ordinary week is genuinely still an open, actively studied question.

References