Sleep and suicide: an analysis of a cohort of 394,000 Taiwanese adults

David Gunnell, Shu-Sen Chang, Min Kuang Tsai, Chwen Keng Tsao, Chi Pang Wen

Research output: Contribution to journalArticle (Academic Journal)

42 Citations (Scopus)

Abstract

BACKGROUND: Sleep problems may lead to, or be symptomatic of, depression and other mental illnesses yet few studies have investigated their association with suicide risk.

DESIGN: Prospective cohort study.

SETTING: Taiwan.

PARTICIPANTS: 393,983 men and women aged 20 or above participating in the MJ health check-up programme.

RESULTS: There were 335 suicides over a mean of 7.4 years follow-up. There was a reverse J-shaped association between sleep duration and suicide risk. When compared with those sleeping 6-8 h per night the adjusted hazard ratios (95% confidence intervals) for suicide associated with 0-4, 4-6 and >8 h sleep were 3.5 (2.0-6.1), 1.5 (1.1-1.9) and 1.5 (1.1-2.0), respectively. People requiring sleeping pills to get to sleep (1.2% participants) were at over 11-fold increased risk; difficulty falling asleep (11.5% participants), frequent dreaming (16.7%) and being easily awoken (30.6%) were associated with a 2.0-, 1.6- and 1.3-fold increased risk of suicide, respectively.

CONCLUSIONS: Less than 6 h sleep duration, sleep disturbances and reported use of sleep medicines are markers of suicide risk. Sleep problems should be assessed when evaluating suicide risk.

Original languageEnglish
Pages (from-to)1457-65
Number of pages9
JournalSocial Psychiatry and Psychiatric Epidemiology
Volume48
Issue number9
DOIs
Publication statusPublished - Sep 2013

Keywords

  • Adolescent
  • Adult
  • Aged
  • Depression
  • Female
  • Humans
  • Incidence
  • Male
  • Middle Aged
  • Odds Ratio
  • Prospective Studies
  • Psychiatric Status Rating Scales
  • Questionnaires
  • Risk Factors
  • Sleep
  • Sleep Disorders
  • Suicide
  • Taiwan

Fingerprint Dive into the research topics of 'Sleep and suicide: an analysis of a cohort of 394,000 Taiwanese adults'. Together they form a unique fingerprint.

Cite this