Under Strain

Stress makes us ill: almost everyone believes it. What the stress hormones actually do, and how strong the evidence is for colds, wounds, the heart, autoimmune disease and cancer. Strong in places, weak in others.

One of the recent successes against ovarian cancer, described in Line by Line, is a drug that blocks the receptor for cortisol, the best-known stress hormone. Added to chemotherapy in women whose cancer no longer responded to platinum, it raised median survival from 11.9 to 16.0 months.[1] It is tempting to read that as proof of something many people already believe: that stress makes us ill, and might even feed a cancer.

The truth is more interesting, and more uneven. Stress hormones do act on the immune system, the blood vessels and the heart, and for some diseases the link with stress is well measured. For others, cancer above all, the evidence is weak, and the idea that a patient’s worry shapes her prognosis does real harm. This essay sorts the strong evidence from the weak. The immune terms are explained in the glossary of The Second Lock.

47%of the most stressed volunteers given a cold virus developed a cold, against 27% of the least stressed

21×the rate of heart attack in the 24 hours after the death of someone close

1.36the hazard ratio for autoimmune disease after a stress-related disorder, in 106,464 Swedes

0.97the hazard ratio for cancer with high job strain, in 116,056 Europeans: no link

Two alarms

What stress does in the body

When the brain perceives a threat, it sets off two systems (Figure 1). The fast one works through nerves: within seconds, the core of the adrenal glands releases adrenaline and noradrenaline, the heart speeds up, blood pressure rises and sugar is released for energy. The slow one works through the blood: the pituitary gland sends a hormone, ACTH, to the outer layer of the adrenal glands, which release cortisol over minutes. Cortisol mobilises energy, puts non-urgent work on hold and, crucially for this series, brakes inflammation. It also feeds back to the brain to switch the alarm off.

Fig. 1 — The two stress systems, fast (nerves and adrenaline) and slow (hormones and cortisol), and the feedback by which cortisol ends the response. A schematic of textbook physiology.

Cortisol’s effects are not simple. A classic review by Robert Sapolsky and colleagues found that, depending on the organ, it can permit, stimulate, suppress or prepare the stress response; one of its main jobs is to stop the response from overshooting.[2] Its brake on the immune system is strong enough that synthetic versions, the steroid drugs such as prednisone, “have been used to treat various immune-related disorders” since 1948;[3] they are still widely used in the autoimmune diseases described in One Signal at a Time.

For a few hours, all of this is useful. The trouble starts when the alarm does not switch off: when stress lasts for months, as it does for a carer, a person in an impossible job or someone living with a trauma.

Measured

Colds, vaccines and wounds

The most persuasive evidence comes from experiments in which stress is measured first and the body is then challenged in a controlled way. In 1991 Sheldon Cohen and colleagues asked 394 healthy volunteers about their stress, gave them nasal drops containing one of five cold viruses, and kept them in quarantine. Infection rose from about 74% to about 90% across the levels of stress, and clinical colds from about 27% to 47%, in a dose-response pattern that was not explained by smoking, drinking, exercise, diet, sleep or personality.[4]

Two studies of people caring for a spouse or relative with dementia, a heavy and long stress, point the same way. Among elderly carers, only 16% of 50 had a strong antibody response to the flu vaccine, against 39% of 67 people of similar age who were not carers; the carers also had higher cortisol.[5] In another study, a small standard skin wound took 48.7 days to heal in 13 carers against 39.3 days in 13 matched controls.[6] Figure 2 puts the three side by side.

Fig. 2 — Three experiments on stress and immunity. Colds: Cohen et al. 1991 (the percentages are approximate, as reported in the abstract). Flu vaccine: Vedhara et al. 1999, share with a fourfold rise in antibody titre. Wound healing: Kiecolt-Glaser et al. 1995, mean days to heal.

Why would stress, which raises cortisol, sometimes lead to more inflammation rather than less? Cohen’s group proposed an answer in 2012: when stress lasts, immune cells become less sensitive to cortisol, so the brake slips. In two virus-challenge studies, people who had recently lived through a long, threatening experience showed this resistance, those with resistance were more likely to develop a cold, and greater resistance predicted more inflammatory signals in the nose after infection.[7] Figure 3 shows the idea.

Fig. 3 — When the brake stops working: the glucocorticoid receptor resistance hypothesis. A schematic after Cohen et al. 2012, not data.

The heart

Strain, grief and a broken heart

For the heart, the evidence is large and consistent, though modest in size. Pooling individual records from 13 European cohorts and 197,473 working people, Mika Kivimäki and colleagues found that those reporting job strain had a 23% higher risk of coronary heart disease. The link held after excluding the first years of follow-up, a check against illness causing the strain rather than the reverse. But the authors were careful about its weight: removing job strain would prevent about 3.4% of cases, “a much smaller effect” than tackling standard risk factors such as smoking.[8] In the INTERHEART study, 11,119 people with a first heart attack in 52 countries reported permanent stress at work or at home about twice as often as controls (odds ratio 2.17), though such studies ask people about their stress after the event.[9]

Acute shocks are sharper. Among 1,985 people interviewed after a heart attack, the rate of heart attack was 21 times higher in the 24 hours after the death of someone close, falling day by day. The absolute risk stays small: about one extra heart attack in the following week for every 1,394 bereaved people at low risk, and one for every 320 at high risk.[10] And there is a condition named for it. In stress cardiomyopathy, sometimes called broken-heart syndrome, a sudden emotional shock stuns the heart muscle; in a 2005 series, 95% of the patients were women, adrenaline levels were about three times higher than in patients with a severe heart attack, and the heart’s pumping recovered fully within weeks.[11]

Self and non-self

Stress and autoimmune disease

The closest link to the rest of this series comes from Sweden. Huan Song and colleagues followed 106,464 people diagnosed with a stress-related disorder (post-traumatic stress disorder, acute stress reaction or adjustment disorder) for about ten years, alongside more than a million unaffected people and 126,652 of the patients’ own brothers and sisters. Autoimmune disease appeared at 9.1 cases per 1,000 person-years in the patients, against 6.0 in the general comparison group: a hazard ratio of 1.36. For post-traumatic stress disorder it was 1.46, and 2.29 for developing three or more autoimmune diseases. The link held when patients were compared with their own siblings, who share much of their genes and upbringing, and it was stronger in younger patients.[12]

This does not prove that stress breaks the locks described in The Second Lock; people with stress disorders differ in many ways, and an emerging autoimmune disease can itself cause distress. But it is a large, careful study, and it fits the biology: a system whose brake on inflammation is worn down is a system more likely to attack the wrong target.

Life and company

Company and childhood

Two broader findings belong here, with caution. A meta-analysis of 148 studies and 308,849 people found that those with stronger social relationships were 50% more likely to survive over the follow-up periods studied, an effect the authors judged “comparable with well-established risk factors for mortality”.[13] And in the Adverse Childhood Experiences study, adults who had suffered four or more kinds of abuse or household dysfunction as children had markedly higher rates of smoking, alcoholism, depression and later heart disease, cancer and lung disease, in a graded relationship with the number of adversities.[14] Both are observational, and much of the effect runs through behaviour: smoking, drinking, inactivity, missed check-ups. They show that a hard life is bad for health, not that stress hormones alone are the route.

The weakest link

Stress and cancer

Cancer is where belief runs furthest ahead of evidence. On whether stress causes cancer, the best answer is no, or not detectably. Pooling 12 European cohorts with 116,056 people followed for a median of 12 years, Katriina Heikkilä and colleagues found that high job strain was not associated with cancer overall (hazard ratio 0.97) or with colorectal, lung, breast or prostate cancer.[15] A wider 2008 meta-analysis did find associations between stress-related factors and cancer, but stressful life events were linked to survival rather than to incidence, and the authors warned of publication bias.[16]

On whether stress makes an existing cancer grow, the laboratory evidence is real. In mice with ovarian tumours, chronic stress raised adrenaline-type hormones in the tissue and made the tumours larger and more invasive, through the growth of new blood vessels and the signal VEGF, the target of the drug bevacizumab.[17] In 113 women with ovarian cancer, a flattened daily cortisol rhythm before treatment went with shorter survival, an estimated average of 3.3 against 7.3 years between the high and low night-cortisol groups, and with more inflammation near the tumour; but the authors noted that the abnormal rhythm may reflect the disease disturbing the hormone system rather than the other way round.[18] And if adrenaline fed ovarian cancer in people as it does in mice, beta-blockers, which block its effects, should help. A meta-analysis of 11 cohort studies with 20,274 women found no association between taking them after diagnosis and survival.[19]

The test that matters most is whether reducing distress lengthens life. In 1989 David Spiegel reported that women with metastatic breast cancer who joined weekly support groups lived a mean of 36.6 months against 18.9 in controls.[20] It became famous; his own larger repeat trial, published in 2007, did not replicate it.[21] Another trial, of a psychological programme for women with breast cancer, reported fewer deaths from breast cancer after 11 years (hazard ratio 0.44),[22] but a single positive trial against a failed replication is not a settled answer. The case for psychological support rests on how people live with their illness, which is reason enough, not on survival.

That is why the cortisol-blocking drug needs reading carefully. Relacorilant was not designed to treat the patient’s stress. Its rationale is molecular: when cortisol reaches its receptor inside tumour cells, it “has been shown to suppress the apoptotic pathways used by cytotoxic agents”, helping the cells survive chemotherapy.[23] The drug works on that signal whatever the patient feels, and the trial did not select women by stress or by any biomarker. It is a discovery about cancer cells, not about worry.

The balance

How strong is each link?

Put the large studies side by side (Figure 4) and the pattern is clear. For the heart and for autoimmune disease, the risks are clearly raised, by roughly a quarter to a half for long-term stress, more for acute shocks. For cancer, the estimates sit on the line of no effect.

Fig. 4 — How strong is the link? Risk ratios with confidence intervals (95% unless marked) from large studies; pink squares are clearly above 1. Hazard ratios from cohorts (Kivimäki et al. 2012; Song et al. 2018; Heikkilä et al. 2013; Wen et al. 2021) and an odds ratio from a case-control study (Rosengren et al. 2004, INTERHEART). The beta-blocker estimate is for the drugs, not for stress itself, and tests whether blocking adrenaline helps.

Known. Stress hormones change immune responses, measurably: more colds after exposure to a virus, weaker vaccine responses, slower healing. Long-term strain modestly raises the risk of heart disease, and acute grief or shock sharply raises it for a few days. Stress-related disorders go with more autoimmune disease, even within families. In cancer cells, cortisol’s receptor can protect against chemotherapy, and blocking it helped in one large trial.

Not known. How much of the link with heart and autoimmune disease runs through hormones, and how much through sleep, smoking, drinking and missed care. Whether reducing stress reduces these diseases; the trials are few. Whether glucocorticoid resistance, so far measured mainly in virus experiments, explains the autoimmune link. And, for cancer, there is no reliable evidence that how a patient feels changes how long she lives.

So what

What follows

Two practical things follow, and they pull in different directions. The first is that chronic stress, loneliness and trauma are worth treating as health problems, because for the heart and the immune system the links are real, and because depression, isolation and post-traumatic stress are treatable in their own right. The second is that the stress story must not become a story of blame. Nobody’s cancer came back because she worried, and nobody’s survival depends on staying positive; the evidence does not support it, and believing it adds guilt to illness. The cortisol that matters in ovarian cancer is the cortisol inside the tumour, and the drug that blocks it works for the anxious and the calm alike.

The series on the immune system

Seven pieces, best read in this order; the terms they share are explained in the glossary of The Second Lock.

1. The Second Lock — the immune system’s brakes, autoimmune disease, prevention and resets; with the series glossary.
2. Why So Many Locks — a short explainer: why the immune system checks and rechecks before it attacks.
3. Other Locks — could it be designed better? Theory, bacteria, plants, lampreys and engineered cells.
4. The Brakes Released — the same brakes released against cancer: checkpoint drugs, CAR-T cells, vaccines.
5. One Signal at a Time — the biological medicines most patients with autoimmune disease take today.
6. The Disease in the Middle — mixed connective tissue disease, between lupus, scleroderma and myositis.
7. Under Strain (this essay) — what stress hormones do, and how strong the evidence is linking stress to disease.

A companion on cancer: The Inner Lining — cancers of the peritoneum, and what surgery with heated chemotherapy can do.

On method and tools

This essay was researched and written with Claude Opus 5.5 (Anthropic): human specification, editorial direction and critical review; machine research, drafting and figure generation. The question came from Luis, who asked whether the link between stress and disease deserved more than a line after reading about relacorilant in Line by Line. Every reference was checked against its PubMed record, and the figures and quotations come from the abstracts. The description of the two stress systems in the first figure is textbook physiology, and the figure on glucocorticoid resistance is a schematic of the hypothesis, not data. The forest plot and the bar charts use the published estimates named in their captions, listed in docs/understrain-results.json; they are drawn by scripts/understrain.py. There is no model in this essay. The cover is a drawing, not data. This essay describes what studies have found; it is not medical advice.

Authored by: Luis Matos Ferreira — Physicist, Developer, Writer

Sources

1. D. Lorusso, L. Gladieff, D. M. O'Malley et al., “Overall survival with relacorilant and nab-paclitaxel in patients with platinum-resistant ovarian cancer (ROSELLA): a phase 3 randomised controlled trial”, Lancet (London, England) 407, 1513–1524 (2026), doi.org.
2. R. M. Sapolsky, L. M. Romero & A. U. Munck, “How do glucocorticoids influence stress responses? Integrating permissive, suppressive, stimulatory, and preparative actions”, Endocrine Reviews 21, 55–89 (2000), doi.org.
3. D. W. Cain & J. A. Cidlowski, “Immune regulation by glucocorticoids”, Nature Reviews Immunology 17, 233–247 (2017), doi.org.
4. S. Cohen, D. A. Tyrrell & A. P. Smith, “Psychological stress and susceptibility to the common cold”, The New England Journal of Medicine 325, 606–612 (1991), doi.org.
5. K. Vedhara, N. K. Cox, G. K. Wilcock et al., “Chronic stress in elderly carers of dementia patients and antibody response to influenza vaccination”, Lancet (London, England) 353, 627–631 (1999), doi.org.
6. J. K. Kiecolt-Glaser, P. T. Marucha, W. B. Malarkey et al., “Slowing of wound healing by psychological stress”, Lancet (London, England) 346, 1194–1196 (1995), doi.org.
7. S. Cohen, D. Janicki-Deverts, W. J. Doyle et al., “Chronic stress, glucocorticoid receptor resistance, inflammation, and disease risk”, Proceedings of the National Academy of Sciences of the United States of America 109, 5995–5999 (2012), doi.org.
8. M. Kivimäki, S. T. Nyberg, G. D. Batty et al., “Job strain as a risk factor for coronary heart disease: a collaborative meta-analysis of individual participant data”, Lancet (London, England) 380, 1491–1497 (2012), doi.org.
9. A. Rosengren, S. Hawken, S. Ounpuu et al., “Association of psychosocial risk factors with risk of acute myocardial infarction in 11119 cases and 13648 controls from 52 countries (the INTERHEART study): case-control study”, Lancet (London, England) 364, 953–962 (2004), doi.org.
10. E. Mostofsky, M. Maclure, J. B. Sherwood et al., “Risk of acute myocardial infarction after the death of a significant person in one's life: the Determinants of Myocardial Infarction Onset Study”, Circulation 125, 491–496 (2012), doi.org.
11. I. S. Wittstein, D. R. Thiemann, J. A. Lima et al., “Neurohumoral features of myocardial stunning due to sudden emotional stress”, The New England Journal of Medicine 352, 539–548 (2005), doi.org.
12. H. Song, F. Fang, G. Tomasson et al., “Association of Stress-Related Disorders With Subsequent Autoimmune Disease”, JAMA 319, 2388–2400 (2018), doi.org.
13. J. Holt-Lunstad, T. B. Smith & J. B. Layton, “Social relationships and mortality risk: a meta-analytic review”, PLoS Medicine 7, e1000316 (2010), doi.org.
14. V. J. Felitti, R. F. Anda, D. Nordenberg et al., “Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults. The Adverse Childhood Experiences (ACE) Study”, American Journal of Preventive Medicine 14, 245–258 (1998), doi.org.
15. K. Heikkilä, S. T. Nyberg, T. Theorell et al., “Work stress and risk of cancer: meta-analysis of 5700 incident cancer events in 116,000 European men and women”, BMJ (Clinical Research Ed.) 346, f165 (2013), doi.org.
16. Y. Chida, M. Hamer, J. Wardle et al., “Do stress-related psychosocial factors contribute to cancer incidence and survival?”, Nature Clinical Practice Oncology 5, 466–475 (2008), doi.org.
17. P. H. Thaker, L. Y. Han, A. A. Kamat et al., “Chronic stress promotes tumor growth and angiogenesis in a mouse model of ovarian carcinoma”, Nature Medicine 12, 939–944 (2006), doi.org.
18. A. Schrepf, P. H. Thaker, M. J. Goodheart et al., “Diurnal cortisol and survival in epithelial ovarian cancer”, Psychoneuroendocrinology 53, 256–267 (2015), doi.org.
19. Z. Y. Wen, S. Gao, T. T. Gong et al., “Post-Diagnostic Beta Blocker Use and Prognosis of Ovarian Cancer: A Systematic Review and Meta-Analysis of 11 Cohort Studies With 20,274 Patients”, Frontiers in Oncology 11, 665617 (2021), doi.org.
20. D. Spiegel, J. R. Bloom, H. C. Kraemer et al., “Effect of psychosocial treatment on survival of patients with metastatic breast cancer”, Lancet (London, England) 2, 888–891 (1989), doi.org.
21. D. Spiegel, L. D. Butler, J. Giese-Davis et al., “Effects of supportive-expressive group therapy on survival of patients with metastatic breast cancer: a randomized prospective trial”, Cancer 110, 1130–1138 (2007), doi.org.
22. B. L. Andersen, H. C. Yang, W. B. Farrar et al., “Psychologic intervention improves survival for breast cancer patients: a randomized clinical trial”, Cancer 113, 3450–3458 (2008), doi.org.
23. A. B. Olawaiye, J. W. Kim, A. Bagameri et al., “Clinical Trial Protocol for ROSELLA: a phase 3 study of relacorilant in combination with nab-paclitaxel versus nab-paclitaxel monotherapy in advanced platinum-resistant ovarian cancer”, Journal of Gynecologic Oncology 35, e111 (2024), doi.org.