Stress

What the research found.

• A Swedish record study followed 106,464 people diagnosed with a stress disorder, 1,064,640 matched people, and 126,652 of their own brothers and sisters. Autoimmune disease came in at 9.1 per 1,000 person-years in the stress group, against 6.0 in the matched group and 6.5 in the siblings, at a hazard ratio of 1.36. The sibling comparison is what gives that its weight, because brothers and sisters share an upbringing and much of their genetics. Post-traumatic stress disorder came in higher still, at 1.46 for any autoimmune disease and 2.29 for having three or more.

  Song and colleagues, JAMA, 2018

• Mindfulness in rheumatoid arthritis gave a DAS28-CRP effect, a change on the combined disease-activity score that uses CRP as its blood marker, of minus 0.44, and a pain effect of minus 0.58, across five trials and 399 people. Neither of those counted as real. Mood improved in all four trials that measured it, which is the pattern across this whole literature.

  DiRenzo and colleagues, Current Rheumatology Reports, 2018

• One trial tested talking therapy for long-term stress in lupus, over ten weekly sessions with fifteen months of follow-up. Low mood, worry, daily stress, and quality of life all improved. The authors report no real change in any immune measure they took, and they took several of them.

  Navarrete-Navarrete and colleagues, Psychotherapy and Psychosomatics, 2010

• One review of stress in rheumatoid arthritis screened 11,115 records and kept just 16 studies. Those 16 used 13 different questionnaires between them. So no pooled figure was possible at all, which is a finding about the state of the field rather than about stress.

  De Cock and colleagues, Seminars in Arthritis and Rheumatism, 2022

• Some lupus patients in a three-year study had their felt stress rise by half a standard deviation or more, and their scores followed. Doctor-rated disease activity came in at 3.7, against 2.5 in the people holding stable or improving. Self-rated activity was 12.2 against 8.2, and pain and fatigue ran in that direction too.

  Patterson and colleagues, Arthritis Care and Research, 2023

• A separate study followed 666,269 Iraq and Afghanistan veterans. Post-traumatic stress disorder came out at a relative risk of 2.00 against veterans with no mental health diagnosis, and 1.51 against veterans with other mental health diagnoses. That second figure is the useful one, because it points at something specific to post-traumatic stress and not to mental illness in general.

  O'Donovan and colleagues, Biological Psychiatry, 2015

• A 24-year group of 54,763 women found trauma linked to new lupus at a hazard ratio of 2.83, and probable post-traumatic stress disorder at 2.94, from 1.19 to 7.26. Only 73 women in the whole group developed lupus, so the direction is clear and the size isn't. A range that wide is a reason to hold the figure loosely.

  Roberts and colleagues, Arthritis and Rheumatology, 2017

• One review pooled 27 studies covering 8,728 cases. Hard childhood experiences were associated with adult autoimmune disease at a standardized mean difference of 0.30. Then the authors tested for publication bias and found it, at an Egger z of 2.17 and a P value of 0.003, and filling in 14 missing studies dropped the pooled figure to 0.11, from 0.02 to 0.20. The smaller number is the one to remember.

  Jesuthasan and colleagues, Brain, Behavior, and Immunity, 2025

What the strongest evidence shows

The strongest study here is a Swedish record comparison, and what makes it strong is that it set people with a stress disorder against their own brothers and sisters. Siblings share an upbringing and much of their genetics, so a difference between them is a good deal harder to explain away. Post-traumatic stress disorder shows a bigger effect than other stress disorders, in that study and in a large veteran study built differently. Trauma has been linked to new lupus as well, in a group followed for more than two decades, and the numbers with their ranges are in the list above.

Childhood adversity, and what publication bias does to it

Hard childhood experiences have been linked to adult autoimmune disease across dozens of studies, and the headline figure from the largest review is the one everybody quotes. That review tested whether the published studies lean one way, found that they do, and corrected for it. Correcting shrank a moderate link to a very small one, and the corrected figure is the honest one. The study designs point that way too, because asking people to recall their childhood after they're already ill gives a bigger answer than watching them from the start.

The flare claim

This is where writing on stress goes furthest past its evidence, and the absence underneath it is worth saying plainly. One review screened 11,115 records and kept 16 studies, and those 16 used 13 different questionnaires between them, so nothing could be pooled at all. No trial has defined a flare by proper criteria and then tested stress ahead of it as the trigger. That's a flat absence and not a mixed result, and the claim gets repeated with confidence anyway.

What treating it achieves, and which direction it runs

The trials agree with each other, and what they agree on isn't what most people expect. Mood improves reliably and the immune numbers don't follow it. Mindfulness in rheumatoid arthritis moved neither disease activity nor pain by enough to count, and a lupus trial of talking therapy improved low mood, worry, and quality of life while reporting no real change in any immune measure.

Most of this evidence can't tell stress causing disease trouble apart from disease causing stress, and in axial spondyloarthritis disease activity predicted low mood at a correlation of 0.607 and a regression coefficient of 0.711. So getting your disease well controlled is one of the more reliable stress treatments you have. You also shouldn't be told your disease would be quieter if you managed your stress better, since the trials built to show that didn't show it.

Common misconceptions.

Questions patients ask.