Glucocorticoids and cardiovascular events in rheumatoid arthritis: A population-based cohort study

John M. Davis, Hilal Maradit Kremers, Cynthia S. Crowson, Paulo J. Nicola, Karla V. Ballman, Terry M. Therneau, Véronique L. Roger, Sherine E. Gabriel

Research output: Contribution to journalArticle

162 Scopus citations


Objective. To determine the relationship between glucocorticoid exposure and cardiovascular (CV) events in patients with rheumatoid arthritis (RA). Methods. A total of 603 adult residents of Rochester, Minnesota with incident RA between 1955 and 1995 were followed up through their medical records for a median of 13 years (total of 9,066 person-years). Glucocorticoid exposure was defined 3 ways: tertiles of cumulative exposure; recent use (≤3 months) versus past use (>3 months); and average daily dosage (≤7.5 mg/day or >7.5 mg/day). CV events, including myocardial infarction, heart failure, and death from CV causes, were defined according to validated criteria. Cox regression models were adjusted for demographic features, CV risk factors, and RA characteristics. Results. Rheumatoid factor (RF)-negative patients with exposure to glucocorticoids were not at increased risk of CV events, irrespective of the glucocorticoid dosage or timing of use, as compared with the reference group of RF-negative patients who had never been exposed to glucocorticoids. In contrast, RF-positive patients were at increased risk of CV events, particularly with higher cumulative exposure, higher average daily dosage, and recent use of glucocorticoids. RF-positive patients with high cumulative exposure to glucocorticoids had a 3-fold increased risk of CV events (hazard ratio 3.06 [95% confidence interval 1.81-5.18]), wliereas RF-negative patients with high cumulative exposure were not at increased risk (hazard ratio 0.85 [95% confidence interval 0.39-1.87]). Conclusion. RF-positive but not RF-negative patients were at increased risk of CV events following exposure to glucocorticoids. These findings suggest that glucocorticoids interact with RF status to modulate the occurrence of CV events in patients with RA. The mechanisms underlying this interaction are unknown and should be the subject of further research.

Original languageEnglish (US)
Pages (from-to)820-830
Number of pages11
JournalArthritis and rheumatism
Issue number3
StatePublished - Mar 1 2007


ASJC Scopus subject areas

  • Immunology and Allergy
  • Rheumatology
  • Immunology
  • Pharmacology (medical)

Cite this