What happens to your body when you are chronically stressed: the biology of pressure

In 1967, epidemiologists began following eighteen thousand British civil servants working in London’s Whitehall district. Not coal miners. Not chemical plant workers. Not anyone whose occupation involved obvious physical risk. They sat at desks and processed paperwork for Her Majesty’s Government.

They died at strikingly different rates depending on where they sat in the hierarchy.

The lowest-grade workers — messengers, doorkeepers, porters — died at three times the rate of the most senior administrators. Same employer. Same access to the National Health Service. Same sedentary, non-dangerous work. When researchers controlled for every risk factor that medicine in the late 1960s knew how to measure — smoking, blood pressure, cholesterol, physical activity, body weight — those factors explained at most forty per cent of the gap. The remaining sixty per cent was blank.

Something was killing the people at the bottom of the organisational chart, and it wasn’t any of the things medicine had spent a century learning to measure. Not poverty — these were salaried government employees with pensions and job security. Not occupational hazard — the most dangerous object in a Whitehall office was arguably the lift. Not ignorance about health. The gradient wasn’t even between the poor and the rich. It was a smooth, continuous slope: each rung down the hierarchy corresponded to worse health outcomes than the rung above. Not a cliff between haves and have-nots. A ramp.

It took a generation and contributions from endocrinology, immunology, neuroscience, and social epidemiology before an answer arrived. It began with a hormone that was never meant to run this long.

The alarm system

Your body has an emergency protocol. It’s fast, ruthlessly efficient, and almost certainly older than your capacity for language. The sequence runs like this: the hypothalamus, a small region at the base of the brain, detects a threat and releases corticotropin-releasing hormone, CRH. CRH reaches the pituitary gland, which responds with adrenocorticotropic hormone, ACTH. ACTH enters the bloodstream and hits the adrenal glands, perched atop the kidneys, which flood the body with cortisol.

This is the hypothalamic-pituitary-adrenal axis — the HPA axis — and what it accomplishes in the next few minutes is genuinely impressive. Cortisol mobilises glucose into the bloodstream and muscles, providing immediate fuel. It sharpens alertness. Raises pain tolerance. And it shuts down everything the body doesn’t need for the next thirty seconds: digestion, immune surveillance, tissue repair, reproductive function. Not because those systems are unimportant — because if a predator is closing the distance between you, fighting off a virus is not the priority. Every suppression is a deliberate sacrifice, energy redirected from maintenance to survival. Each makes perfect sense on one condition: that the emergency will be brief. And it’s that condition — brief — that matters more than anything else about the system’s design.

The body triages. Cortisol is the triage order.

None of this is metaphorical. Cortisol physically diverts blood flow from the digestive tract to skeletal muscle, suppresses lymphocyte production, inhibits growth hormone and sex hormones. The body is not “feeling stressed.” It is executing a coordinated redistribution of physiological resources, and every redistribution is a bet that the current crisis is more important than whatever maintenance it’s interrupting. The bet pays off when the crisis ends in minutes. It bankrupts the body when it doesn’t.

Under normal conditions, once the threat passes, cortisol circulates back to the hippocampus and the hypothalamus, where it suppresses further CRH release. The pituitary stands down. The adrenals stop producing cortisol. Digestion, immunity, tissue repair — everything comes back online. This is the negative feedback loop, the braking system, and it’s elegant: the very hormone that launched the emergency eventually signals its own conclusion.

The whole design assumes a return to calm.

Cortisol follows a diurnal rhythm. It peaks shortly after waking — a burst of fuel and alertness to get you upright and functional — then declines steadily through the day. By evening it’s at its lowest. Sleep is when the suppressed systems do their heaviest work: tissue repair, immune consolidation, memory processing. The rhythm encodes an assumption about the architecture of a normal day: threat in the morning, recovery by nightfall.

The entire system was built for emergencies that are over in minutes. A lion charges, you run, and the episode resolves one way or another. What the HPA axis cannot do — what it was never built to do — is distinguish between a predator that will be gone in three minutes and a job that will grind on for thirty years. It has one protocol. It runs it. And if the threat doesn’t end, neither does the cortisol.

The fire that won’t go out

In the short term, cortisol suppresses the immune system. This is sensible triage: don’t fight the infection, fight the predator. Immune surveillance can catch up later.

But when the cortisol never drops, there is no later.

What happens next is not what most people expect. Immune cells chronically bathed in cortisol develop glucocorticoid receptor resistance — they downregulate their receptors and stop hearing cortisol’s anti-inflammatory signal. The brake goes soft. And when the brake fails, the immune system doesn’t return to its pre-stress baseline. It swings past it into chronic low-grade inflammation: elevated interleukin-6, tumour necrosis factor-alpha, C-reactive protein. These are the inflammatory markers now linked to cardiovascular disease, type 2 diabetes, depression, and autoimmune conditions. The inflammation is quiet — no fever, no acute symptoms — which is precisely what makes it dangerous. It runs in the background for years, doing damage the person never feels until the damage has a name.

In 1991, Sheldon Cohen and colleagues at Carnegie Mellon ran an experiment that would make the connection between psychological state and immune function impossible to dismiss as soft science. They gave 394 healthy volunteers nasal drops containing one of five live respiratory viruses — rhinovirus, respiratory syncytial virus, or coronavirus — assessed their psychological stress levels by questionnaire, then quarantined them and watched who got sick. The results were dose-response and unambiguous: the higher a subject’s stress score, the more likely they were to develop a verified clinical cold. Not just feeling worse. Actually infected, virus confirmed by lab isolation or an increase in virus-specific antibody titre. The most stressed subjects got sick at significantly higher rates than the least stressed, in a gradient that tracked stress level with almost mechanical precision. The psychological and the immunological were not two separate events with a mysterious connection between them. They were the same event, measured on two different scales.

The glucocorticoid resistance paradox

The central paradox of chronic stress immunology: cortisol is an anti-inflammatory hormone that, in sustained excess, produces a pro-inflammatory state. The mechanism is receptor downregulation. Immune cells repeatedly saturated with cortisol reduce the number of glucocorticoid receptors on their surfaces. Fewer receptors means a weaker regulatory signal. With cortisol's restraining influence diminished, pro-inflammatory cytokines — IL-6, TNF-α — rise without adequate opposition. The system doesn't collapse into immunosuppression. It tips into chronic activation. Duration inverts the mechanism: the longer cortisol stays elevated, the less it can do what cortisol is supposed to do.

The plumbing

Cortisol raises blood pressure. Useful, briefly — higher pressure drives more oxygen to muscles that need to run. Destructive, permanently.

Sustained cortisol elevation produces sustained hypertension. Clinical studies of exogenous cortisol at therapeutic doses show systolic blood pressure increases of roughly twelve millimetres of mercury. That’s not a large number in isolation — a bad day at work might do the same. But cortisol-driven hypertension is not a bad day. It’s a permanent condition. Twelve extra millimetres of mercury, every hour of every day, for years.

The structure that absorbs this is the endothelium — the single-cell layer lining every blood vessel in the body. The endothelium is not passive plumbing. It actively regulates blood flow, prevents clotting, controls the exchange of molecules between blood and tissue, and — critically — dilates vessels to accommodate changes in pressure. Chronic cortisol impairs that vasodilatory capacity, in part by inhibiting nitric oxide–mediated vasodilation. Vessels that can’t relax appropriately under load sustain repetitive mechanical stress on their inner walls. Inflammatory cells, already primed by the glucocorticoid resistance that turned chronic cortisol into chronic inflammation, infiltrate the damaged vessel wall. Lipids accumulate. Plaques form. The process is called atherosclerosis, and it is the leading cause of heart attack and stroke in wealthy countries. The chronic inflammation already running in the background and the mechanical pressure bearing down on the vessel walls are not independent pathologies. They converge on the same arterial wall, each amplifying the other.

And this is where the Whitehall mortality gradient acquires its biological explanation.

The study that couldn't explain itself

The original Whitehall study — now called Whitehall I — was designed by Geoffrey Rose and launched in 1967 with eighteen thousand male civil servants. The mortality gradient it uncovered was stark: threefold between the lowest and highest employment grades. But when the analysis controlled for every conventional risk factor available, those factors explained no more than forty per cent of the gap. Whitehall I produced one of the most important findings in twentieth-century epidemiology and had no framework to interpret it. That social position itself might be the mechanism — not through poverty, not through behaviour, but through what position does to the body — was not yet a credible hypothesis in a field that understood disease in terms of pathogens, genes, and personal choices.

In 1985, Michael Marmot launched Whitehall II — a new cohort of over ten thousand civil servants, now including women, and a question the original study had never thought to ask: job control. Not income. Not occupational prestige in the abstract. The specific, measurable degree to which a worker determined what they did, when they did it, and how they did it. Marmot’s team assessed job control at two separate time points and tracked cardiovascular outcomes over the following years.

The finding reframed the field. Workers with sustained low job control had an odds ratio of 1.93 for new coronary heart disease events compared to those with sustained high control — nearly double the risk of developing heart disease. And this held after adjustment for employment grade, smoking, cholesterol, blood pressure, body mass index, and negative affectivity. Every variable medicine traditionally blamed was already in the statistical model. After all of them were accounted for, job control still predicted who got heart disease and who didn’t. What remained, after every conventional explanation had been subtracted, was the job itself: specifically, whether you controlled it or it controlled you.

Robert Karasek had laid the theoretical groundwork in 1979 with his demand-control model of occupational stress. The toxic combination, he argued, was not high demand alone. A surgeon faces immense pressure but wields decision-making power over every cut. The toxic combination was high demand paired with low decision latitude: responsibility without authority, workload without autonomy. The condition of someone who must respond to everything and determines nothing. The permanent secretary chose the project, set the timeline, delegated the execution. The messenger carried instructions he had no part in forming, under a deadline he couldn’t negotiate, to a destination he hadn’t selected. Both were busy. Only one was running a chronic HPA response. And the Whitehall data showed, with a cohort of ten thousand, that the distinction between these two conditions predicted who developed coronary heart disease independently of every other measured variable.

The organ that remembers

The same cortisol circulating through the arteries also crosses the blood-brain barrier and reaches the hippocampus — the brain structure with the highest density of glucocorticoid receptors. This density is not a design flaw. The hippocampus encodes contextual memory: this place was dangerous, this situation caused harm. Doing that job fast requires sensitivity to the threat hormone. The organ that tags experiences as dangerous evolved to be exquisitely responsive to cortisol.

That responsiveness is now the vulnerability.

Robert Sapolsky’s work across the 1990s and 2000s documented what chronic glucocorticoid exposure does to this tissue: dendritic atrophy in CA3 pyramidal neurons — the branching connections between cells wither; suppressed neurogenesis in the dentate gyrus — fewer new neurons are produced; and in severe or prolonged cases, outright cell death. Not only in lab animals. Sapolsky found hippocampal atrophy in Cushing’s syndrome, major depressive disorder, and PTSD — three conditions that share, as their common biological feature, chronically elevated cortisol. The organ that evolved to remember danger is itself damaged by the chemistry of danger.

Sonia Lupien and colleagues, in a 1998 study in Nature Neuroscience, made the timeline visible. They tracked elderly subjects over several years and found that those with chronically elevated cortisol had measurably smaller hippocampi and worse performance on hippocampus-dependent memory tasks — declarative recall, spatial orientation, the kinds of memory that locate you in time and place. The degree of shrinkage correlated with cumulative cortisol exposure over the years of the study. The hippocampus was keeping a physical record of the total hormonal burden, even as that burden eroded its capacity to keep records of anything else.

Now comes the structural trap that makes chronic stress self-perpetuating rather than merely harmful. The hippocampus is not just a memory organ. It is part of the HPA axis’s own braking system — a critical node in the negative feedback loop that is supposed to end the cortisol cascade. When cortisol levels rise, a healthy hippocampus detects the elevation and sends an inhibitory signal back to the hypothalamus: enough. Stand down. Reduce CRH production. This is how the emergency ends. A hippocampus that has been damaged by years of excessive cortisol sends a weaker inhibitory signal. The brake is degraded. Animal studies confirm that hippocampal lesions prolong the HPA axis’s cortisol response, impairing the negative feedback that should bring the cascade back down. The cascade can persist even after the original stressor is removed, because the organ responsible for stopping it has been damaged by what it was supposed to stop. Cortisol erodes the brake on cortisol. The longer the stress lasts, the harder it becomes for the body to end the stress response, even if the external conditions change.

The overlap with depression is substantial and not coincidental. The hippocampal atrophy and HPA dysregulation documented in chronic stress conditions are also characteristic of major depressive disorder. Whether hippocampal damage causes depression, or both are downstream consequences of sustained glucocorticoid exposure, remains under active investigation. But the biological signatures overlap to the point where, in many clinical presentations, they are indistinguishable. What people experience as chronic stress and what clinicians diagnose as major depression are, at the level of the hippocampus and the HPA axis, often the same physiological event.

The chromosome

Every chromosome in your body ends with a telomere — a repetitive nucleotide sequence that functions as a protective cap, a buffer of expendable genetic material. Each time a cell divides, the telomere shortens slightly. When it reaches a critical length, the cell can no longer divide safely: it either enters senescence — a kind of functional retirement — or triggers its own death. Telomerase, the enzyme that can rebuild telomere length, was discovered by Elizabeth Blackburn, Carol Greider, and Jack Szostak, work that earned them the 2009 Nobel Prize in Physiology or Medicine. Telomere length is, in effect, a measure of how many divisions a cell has left. A biological countdown.

In 2004, Elissa Epel and Blackburn published a study in the Proceedings of the National Academy of Sciences that connected the cortisol cascade directly to that countdown. They recruited fifty-eight healthy premenopausal women: nineteen controls and thirty-nine who were primary caregivers for a chronically ill child — mothers whose daily lives involved sustained, unrelenting responsibility with limited control over outcomes. Within the caregiving group, the results were dose-response: more years of caregiving predicted shorter telomeres, lower telomerase activity, and higher oxidative stress in peripheral blood mononuclear cells, after controlling for the mothers’ ages. The women reporting the highest perceived stress had telomere lengths equivalent to those of women nine to seventeen additional years older.

That is not a metaphor. A measurement. The protective caps on their chromosomes had been shortened by the same amount that an additional decade or more of ordinary living would produce. Oxidative stress — the accumulation of free radicals, which chronic cortisol elevation accelerates — is one mechanism driving this accelerated erosion. Every subsequent cell division would begin from that shortened position. The biological clock had been physically advanced at the level of the nucleotide by years spent in a specific social condition: sustained caregiving without adequate control or relief.

A shorter telomere does not know why it is shorter. It does not distinguish between time and stress. It simply has less buffer remaining before the cell can no longer divide.

What the gradient proves

Return to Whitehall with the full machinery now visible. The lowest-grade civil servants died at three times the rate of the most senior administrators. Conventional risk factors explained at most forty per cent of that gap. The remaining sixty per cent is now legible: decades of sustained HPA activation deposited as endothelial damage in the arteries, chronic inflammation in immune tissue, atrophy in the hippocampus, shortened telomeres in the chromosome. The biology explains the gradient. The gradient names the cause.

What the hierarchy distributed was not just money or prestige. It was decision latitude — the degree to which a person determined the conditions of their own working day. Marmot’s finding, confirmed across both Whitehall cohorts, is that decision latitude is not a psychological comfort. It is a physiological variable. Low decision latitude is a chronic stressor in the technical HPA sense: inescapable demand without control, the precise combination that keeps cortisol elevated and prevents the negative feedback loop from completing its work.

The demand-control model

Robert Karasek's 1979 framework classifies jobs along two dimensions: psychological demand and decision latitude. The dangerous quadrant is high demand with low control — "job strain." A surgeon faces immense demand but wields immense control; a call-centre worker faces relentless demand with almost none. The model predicts that the call-centre worker carries greater cardiovascular risk regardless of income, and subsequent research — including Marmot's operationalisation in Whitehall II, where job control was measured directly for the first time in a large epidemiological cohort — confirmed it. Control predicted coronary events independently of every other variable in the study.

The permanent secretary chose his project and his timeline. The messenger carried the result under a deadline he couldn’t change, to a place he hadn’t picked, for reasons nobody told him. Both worked full days. One had a body whose emergency response resolved each evening. The other had a body running an emergency response that never received a termination signal.

Sapolsky found the same architecture in a different species entirely. Studying wild olive baboon troops in Kenya over decades, he documented that subordinate males showed chronically elevated basal glucocorticoid levels, suppressed immune function, elevated blood pressure, and higher rates of atherosclerosis than dominant males. And here is the finding that matters most for the Whitehall argument: the pathology tracked not with how hard the animal worked or how much physical danger it faced — subordinate baboons do not work harder than dominant ones — but with social position, the predictability of social interactions, and the degree of control the animal had over its daily experience. Dominant males could predict and control their social encounters. Subordinate males could not. The biology didn’t care about effort. It cared about control.

The Whitehall gradient is a specific instance of a broader pattern. Richard Wilkinson and Kate Pickett, comparing health data across wealthy nations, found that income inequality — the gap between top and bottom in a society, independent of absolute wealth — predicts worse outcomes across nearly every health measure: life expectancy, mental illness, infant mortality, obesity. The relationship held even among rich countries: it wasn’t about how much money a society had but about how that money was distributed. More unequal wealthy nations performed worse on health indices than less unequal ones. Marmot, who chaired the WHO Commission on Social Determinants of Health, concluded in its 2008 report that the conditions in which people are born, grow, live, work, and age are the primary drivers of health inequalities — not healthcare systems, not genetic variation, not personal choices about diet and exercise. The texture of daily life. The degree of control a person has over the circumstances they inhabit.

And conditions include the physical environment — the actual, literal spaces in which bodies spend their days. Roe and colleagues, studying socioeconomically deprived communities in Scotland in 2013, measured salivary cortisol in residents and found that those with greater access to green space had lower perceived stress and healthier diurnal cortisol patterns — the morning peak and evening decline that a properly functioning HPA axis produces.

Access to a park is not an amenity. It is a variable in the cortisol equation. The nervous system reads the environment at a level below conscious decision-making: noise, crowding, the absence of horizon, the presence or absence of trees, the length and unpredictability of a commute. These are not lifestyle preferences. They are structural conditions that modulate the same hormonal cascade that runs from hypothalamus to adrenal gland. A person living in a treeless estate beside a motorway is not choosing a stressful life. They are living in an environment that was designed — by someone — to sustain baseline cortisol elevation as a feature of daily existence.

The conditions that produce chronic HPA activation are structurally distributed. Precarious employment — zero-hours contracts, gig work without sick leave, the perpetual threat of termination — strips away the predictability the nervous system requires to allow recovery. Workplace hierarchies concentrate decision-making at the top and hold the majority in permanent states of demand without control. Built environments designed without green space or quiet or adequate housing sustain elevated cortisol as a background condition. These are not misfortunes. They are design decisions — made by people who possess decision latitude about the environments of people who do not.

The HPA axis does not process economic philosophy. It does not understand org charts or employment contracts. It has one input — is this threat going to end? — and it produces cortisol until the answer is yes.

For the civil servant on the lower floors of the Whitehall building, the one who couldn’t decide what he worked on, or when, or for what purpose, the answer was never yes. His body ran the same emergency cascade, day after day, that it would have run on a savanna when a predator broke cover. His endothelium took the pressure. His immune system lost its regulation. His hippocampus shrank, and the shrinking weakened the brake on cortisol, which made the hippocampus shrink further. His telomeres shortened. His organs didn’t know they were in a government office. They responded with perfect evolutionary fidelity to a signal that said: the danger is continuing.

It wasn’t the title on his door that his body was reading. It was the absence of a door he could close from the inside.

Avis de non-responsabilité de Gen AI

Certains contenus de cette page ont été générés et/ou édités à l'aide d'une IA générative.

Les médias

Pedro Figueras – Pexels

Principales sources et références

Cohen S, Tyrrell DAJ, Smith AP. Psychological Stress and Susceptibility to the Common Cold. New England Journal of Medicine. 1991;325(9):606-612.

Bosma H, Marmot MG, Hemingway H, Nicholson AC, Brunner E, Stansfeld SA. Low job control and risk of coronary heart disease in Whitehall II (prospective cohort) study. BMJ. 1997;314(7080):558-565.

Marmot MG, Smith GD, Stansfeld S, Patel C, North F, Head J, White I, Brunner E, Feeney A. Health inequalities among British civil servants: the Whitehall II study. The Lancet. 1991;337(8754):1387-1393.

Sapolsky RM. Glucocorticoids and hippocampal atrophy in neuropsychiatric disorders. Archives of General Psychiatry. 2000;57(10):925-935.

Lupien SJ, de Leon M, de Santi S, Convit A, Tarshish C, Nair NPV, Thakur M, McEwen BS, Hauger RL, Meaney MJ. Cortisol levels during human aging predict hippocampal atrophy and memory deficits. Nature Neuroscience. 1998;1(1):69-73.

Epel ES, Blackburn EH, Lin J, Dhabhar FS, Adler NE, Morrow JD, Cawthon RM. Accelerated telomere shortening in response to life stress. Proceedings of the National Academy of Sciences. 2004;101(49):17312-17315.

Karasek RA. Job Demands, Job Decision Latitude, and Mental Strain: Implications for Job Redesign. Administrative Science Quarterly. 1979;24(2):285-308.

Sapolsky RM. The influence of social hierarchy on primate health. Science. 2005;308(5722):648-652.

Wilkinson RG, Pickett KE. Income inequality and population health: a review and explanation of the evidence. Social Science & Medicine. 2006;62(7):1768-1784.

Commission on Social Determinants of Health. Closing the Gap in a Generation: Health Equity Through Action on the Social Determinants of Health. World Health Organization. 2008.

Roe JJ, Ward Thompson C, Aspinall PA, Brewer MJ, Duff EI, Miller D, Mitchell R, Clow A. Green space and stress: evidence from cortisol measures in deprived urban communities. International Journal of Environmental Research and Public Health. 2013;10(9):4086-4103.

Whitehall I study: Rose G, Marmot MG. Social class and coronary heart disease. British Heart Journal. 1981;45:13-19. Mortality gradient and risk factor analysis also reported in: Marmot MG, Shipley MJ, Rose G. Inequalities in death — specific explanations of a general pattern? The Lancet. 1984;323(8384):1003-1006.

Walker BR, Best R, Shackleton CHL, Padfield PL, Edwards CRW. Increased vasoconstrictor sensitivity to glucocorticoids in essential hypertension. Hypertension. 1996;27:190-196. Cardiovascular consequences of cortisol excess also reviewed in: Whitworth JA, Williamson PM, Mangos G, Kelly JJ. Cardiovascular consequences of cortisol excess. Vascular Health and Risk Management. 2005;1(4):291-299.