Sleep

What the research found.

• One study looked at three inflammatory diseases at once, in people who were mostly in remission or close to it. They slept badly anyway, at 63.7 percent with rheumatoid arthritis, 61.5 percent with psoriatic arthritis, and 66.7 percent with axial spondyloarthritis. Their inflammation was under control and their sleep was not.

  Polak and colleagues, Clinical Rheumatology, 2026

• Bad sleep correlated with higher C-reactive protein at an effect size of 0.12, and with higher interleukin 6 at 0.20, while short sleep gave 0.09 and long sleep 0.17. Those small effects all point one direction, and they come from a review of 72 studies and more than 50,000 people. Keeping people awake on purpose raised neither marker, and it didn't raise TNF-alpha either, which showed no link to any sleep measure at all.

  Irwin and colleagues, Biological Psychiatry, 2016

• Poor sleep predicted getting rheumatoid arthritis at a hazard ratio of 1.50, from 1.20 to 1.87, with middling sleep quality at 1.23 and a run of poor or short sleep at an odds ratio of 1.70. Each hour less sleep added about 7 percent. Then a genetic study tested whether sleep causes it, and frequent insomnia came in at 0.99, any insomnia at 1.09, and sleep length at 0.65. Not one of those counted, so there was no genetic sign of cause for any sleep trait it looked at. That mix usually means something else explains both.

  Chen and colleagues, Arthritis Research and Therapy, 2026

• One trial tested group talking therapy for insomnia in rheumatoid arthritis, with sleep efficiency on a sleep study as its main outcome. It came in at 88.7 percent against 83.7 percent, a difference of 5.03 percentage points, from minus 0.37 to 10.43, at a P value of 0.068. So the machine reading missed the mark, and what patients reported about their sleep and their arthritis improved a great deal, which is a split worth stating instead of smoothing over.

  Latocha and colleagues, Rheumatology, 2022

• Talking therapy for insomnia improved sleep efficiency by 0.77 standard deviations across 67 trials and 5,232 people with a long-term illness. Insomnia severity improved by 0.98 and time to fall asleep by 0.64. That research covers long-term illness broadly, including chronic pain, heart disease, and cancer, and not joint disease specifically.

  Scott and colleagues, JAMA Internal Medicine, 2025

How common it is, and why treating the disease doesn't fix it

Sleep is one of the most common complaints in joint disease and one of the least often addressed. Pooled figures put sleep problems in rheumatoid arthritis at 60.9 percent, ankylosing spondylitis studies range from 35 to 90 percent depending on the questionnaire, and lupus patients score 3.45 points worse than controls on the Pittsburgh scale. The most useful study here compared three diseases side by side in people who were mostly in remission or close to it, and two thirds slept badly anyway. Their inflammation was under control and their sleep was not, so waiting for your treatment to fix your sleep means waiting for something that may never come.

Whether bad sleep drives inflammation

Here the usual story and the evidence part company, and it's worth following slowly. One large review compared sleep against blood markers, and in the studies that only watched people, bad sleep correlated with slightly higher C-reactive protein and interleukin 6. That paper looked at experiments too, where researchers decide who sleeps and who stays awake, and keeping people awake raised nothing at all, not C-reactive protein, not interleukin 6, and not TNF-alpha. When an experiment disagrees with watching, the experiment usually wins, because it strips out everything else that differs between good and bad sleepers.

Whether it causes disease

That split appears again, harder, on whether bad sleep brings on rheumatoid arthritis. A large study followed people over time and poor sleep quality predicted getting the disease at a hazard ratio of 1.50, from 1.20 to 1.87. Then a genetic study came at it a different way, using gene variants that affect sleep as a natural experiment, and it found no sign of cause for any sleep trait it looked at. Both studies can be right, because bad sleep may mark something else that raises risk, such as pain, low mood, or disease that has started and not yet been found.

Sleep apnea, and what this page doesn't tell you

Sleep apnea deserves its own mention, because it's treatable and missing it costs more than tiredness. One population study followed people with rheumatoid arthritis, and after an apnea diagnosis heart disease followed at a hazard ratio of 1.9 and stroke disease at 2.4, with the added risk looking much like anybody else's. Snoring, waking up tired, and daytime sleepiness are the things to mention, and the job is finding it.

There's no target number of hours on this page, no sleep hygiene checklist, and nothing about melatonin. Sleep hygiene has never been tested on its own in these diseases, and no trial of a sleeping pill, melatonin, or a sedating antidepressant with sleep endpoints has been run in any of them. Anyone selling you a disease-specific sleep protocol is going past what's been studied.

Common misconceptions.

Questions patients ask.