Corticosteroids Linked to Cardiovascular Events in Rheumatoid Arthritis

Arthritis News > Corticosteroids Linked to Cardiovascular Events

Corticosteroids Linked to Cardiovascular Events in Rheumatoid Arthritis

Corticosteroids are associated with an increased risk of cardiovascular (CV) disease in the general population.  However, there are conflicting data on the CV risk of corticosteroids in RA, as their potent anti-inflammatory properties may result in a net beneficial effect.  Here, Davis et al (Arthritis Rheum 2007; 56(3): 820) investigate the effect of cumulative corticosteroids on the rate of CV events (myocardial infarction (MI), heart failure (HF), and CV mortality) in patients with RA.


Patients from Rochester, Minnesota with incident RA were enrolled over a 40 year period from 1955 to 1995.  Data on subject characteristics, CV events, and cumulative oral and parenteral corticosteroid exposures were retrospectively abstracted from medical records, with Cox proportional hazards regression used to model the risk of CV events according to glucocorticoid exposure, controlling for demographic and CV risk covariates.


603 RA subjects (73% female with mean age at baseline of 58 years) with a median follow-up time of 13 years were included.  Rheumatoid factor (RF) seropositivity was present in 65% at baseline.  Cumulative glucocorticoid exposure ranged from no exposure (42%) to more than 7 grams (median 5 grams).  RF positive subjects had significantly higher cumulative glucocorticoid exposure.  There were 232 CV events during follow-up; 41 MI, 172 HF, and 176 CV deaths.

After adjusting for demographic and CV risk factors, subjects with the highest cumulative glucocorticoid exposure (>7 grams) had an almost 2-fold greater risk of total CV events compared to those never exposed.  This effect was noted in RF positive, but not RF negative subjects, for the combined outcome, and individually for the outcome of HF.

In fully adjusted models, daily prednisone dosing of > 7.5 mg showed a trend to significance in the association with total CV events compared to subjects receiving lower daily doses (HR 1.75 (1.05 – 2.91); p=0.082).  Recent (within the past 3 months) exposure was significantly associated with CV events compared to no exposure (HR 1.66 (95% CI 1.14 – 2.41); p=0.029), but only in the RF positive group.  Past exposure (> 3 months) was not significantly associated with the combined CV outcome.


Corticosteroid use was associated with CV events in RA in a dose dependent manner.  The risk was highest for RF seropositive patients.

Editorial Comments

The findings paint a telling picture of CV risk associated with corticosteroid exposure and suggest that, when possible, their use should be limited in RA patients, particularly those with RF.  However, some caution should be used when interpreting the results.  Steroids are often used in subjects with the most residual inflammatory activity and, as these subjects are more likely to be RF positive, combined the known association of elevated systemic inflammation and CV events, it is possible that the effects noted here are due to the association of corticosteroid use with other risk factors and not the effect of the drug itself.  This confounding by indication can be difficult to parse in observational studies of clinic-based patients, particularly given changes in clinical practice over the four decade span of the study.  In fact, it is somewhat surprising that the prevalence of corticosteroid exposure was as low as reported (58%), given that the span of the study occurred over time periods in which corticosteroids were the norm for treatment in most patients.