A common counter-argument to this thesis is that more severe patients are treated with steroids chronically, thus the increased mortality seen is more a reflection of RA severity. The evidence, however, suggests otherwise: some investigators have found that it is a characteristic of the physician rather than the patient that determines the use of steroids.8 Some time ago, the ARAMIS database showed steroids to be associated with increased mortality, whereas azathioprine, at that time reserved for more severe patients, was not.9 In our own study, patients who were initially matched for severity were either treated with low-dose steroids or none at all.10 Those patients on steroids showed a significantly worse outcome at the five-year follow-up.
The side effects of steroids can be broken down to specifics. Osteoporotic fractures are the best understood. They can occur remarkably quickly, within three months of initiating steroids, but the likelihood of fracture can be reduced with antiresorptive therapy.11 However, I wonder what percentage of patients with RA on steroids are actually receiving effective antiresorptive therapy. Hospitalization rates for pneumonia increase with steroid use and, in those patients receiving anti–tumor necrosis factor agents, concomitant steroids may be more important risk factors for infection than the biologic drug itself.12,13 The rate of Herpes zoster infection is increased with the use of steroids, as is tuberculosis.14,15 A recent study confirmed the prior impression that steroid use in rheumatoid patients is associated with serious lower gastrointestinal events and death.16
The good news is that in North America, physicians recognize the hazards of long-term steroid use and patients are more frequently started on limited, short-term therapy. Over a brief period, many, if not most, patients are tapered off steroids.17 Indeed, it is a measure of biologic therapy’s success that this can be achieved.
In continental Europe, the DMARD evidence favoring steroids seems to have persuaded doctors to use this drug routinely. Yet, it was interesting to observe that in Bakker’s study, about one-third of those patients who were solicited to participate in the study declined to do so.3 Eighty percent of these patients stated that they would not participate because they did not want to take steroids.
In my opinion, the most troublesome “side effect” of steroids is that, if rheumatologists are seen to regularly prescribe oral steroids (compared to joint injections), this may provide cover for the primary care physician (PCP) to do the same without recognizing the importance of concomitant DMARD therapy. Unfortunately, even among some rheumatologists, the importance of concomitant DMARD therapy is not always recognized.18