Hostility and the risk of peptic ulcer in the GAZEL cohort

Health Psychol. 2015 Feb;34(2):181-5. doi: 10.1037/hea0000129. Epub 2014 Aug 11.


Objective: Evidence for an association between hostility and peptic ulcer mainly relies on cross-sectional studies. Prospective studies are rare and have not used a validated measure of hostility. This prospective study aimed to examine the association between hostility and peptic ulcer in the large-scale French GAZEL cohort.

Method: In 1993, 14,674 participants completed the Buss and Durkee Hostility Inventory. Participants were annually followed-up from 1994 to 2011. Diagnosis of peptic ulcer was self-reported. The association between hostility scores and ulcer incidence was measured by hazard ratios (HR) and 95% confidence intervals computed through Cox regression.

Results: Among 13,539 participants free of peptic ulcer history at baseline, 816 reported a peptic ulcer during a mean follow-up of 16.8 years. Adjusting for potential confounders, including smoking, occupational grade, and a proxy for nonsteroidal anti-inflammatory drug exposure, ulcer incidence was positively associated with total hostility (HR per SD: 1.23, confidence interval: 1.14-1.31), behavioral hostility (HR per SD: 1.13, confidence interval: 1.05-1.21), cognitive hostility (HR per SD: 1.26, confidence interval: 1.18-1.35), and irritability (HR per SD: 1.20, confidence interval: 1.12-1.29). The risk of peptic ulcer increased from the lowest to the highest quartile for all hostility measures (p for linear trend < .05).

Conclusions: Hostility might be associated with an increased risk of peptic ulcer. Should these results be replicated, further studies would be needed to explore the underlying mechanisms.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Female
  • Follow-Up Studies
  • France / epidemiology
  • Hostility*
  • Humans
  • Incidence
  • Male
  • Peptic Ulcer / epidemiology*
  • Peptic Ulcer / psychology*
  • Personality Inventory
  • Prospective Studies
  • Risk Factors