By Stan Deresinski, MD, FACP, FIDSA

Clinical Professor of Medicine, Stanford University

Dr. Deresinski reports no financial relationships relevant to this field of study.

SYNOPSIS: The prolonged use of low-dose oral corticosteroids (including < 5 mg prednisone equivalent doses) in rheumatoid arthritis patients is associated with an increased risk of infection.

SOURCE: George MD, Baker JF, Winthrop K, et al. Risk for serious infection with low-dose glucocorticoids in patients with rheumatoid arthritis: A cohort study. Ann Intern Med 2020; Sept. 22. doi: 10.7326/M20-1594. [Online ahead of print].

George and colleagues set out to quantify the risk for hospitalized infection associated with the use of low-dose glucocorticoid therapy in adults with rheumatoid arthritis who were also receiving a disease-modifying antirheumatic drug (DMARD) for more than six months. A hospitalized infection was considered to have occurred if there was an ICD-9 diagnostic code indicating infection among all the discharge diagnoses from an acute care hospital. The primary outcome was the time to first such infection. They used the Medicare national claims data from 2006 to 2015 and the Optum Clinformatics Data Mart from 2001 to 2015 covering individuals with commercial health insurance. The investigators identified 247,297 and 44,118 qualifying medication courses in the Medicare Optum databases, respectively. At six months of stable DMARD use, 47.1% and 39.5%, respectively, were receiving glucocorticoids, most at doses of < 5 mg per day.

The one-year cumulative incidences of hospitalized infection in Medicare patients receiving glucocorticoids in daily doses (as prednisone equivalents) of < 5 mg, 5 mg to 10 mg, and > 10 mg were 11.0%, 14.4%, and 17.7%, respectively, while it was only 8.6% in those not receiving glucocorticoids. The incidences were lower in the Optum database cases involving a younger population: 5.2%, 8.1%, and 10.6% at the three doses, respectively, while it was 4.0% in those not receiving glucocorticoids. Each of the differences achieved statistical significance. The most frequent infections involved the urinary tract, as well as pneumonia, bacteremia or septicemia, and skin or soft tissue infections.


Glucocorticoid administration is an important element of the treatment for many patients with rheumatoid arthritis and is believed to be effective for durations of as much as six months. The risk-benefit ratio of longer durations is uncertain, at least in part because of the risk of drug-associated adverse effects, such as osteoporosis and infection. The study reviewed here demonstrates an increased risk of dose-dependent infection.

The massive number of observations in this analysis lend strength to the result, but the nature of the study leaves open the possibility of several confounders. For instance, the diagnosis of hospitalized infection relied on discharge ICD-9 codes, raising issues of accuracy. Thus, it is certain that many coded as a urinary tract infection were, in fact, instances of asymptomatic bacteriuria. Nonetheless, the authors pointed out that the overall results were concordant with similar studies. They also noted that the risk associated with a daily dose of < 5 mg is similar to that reported with biologic therapies in previous studies.

Thus, there is an infection risk associated with prolonged use of even very low doses of glucocorticoids. These results provide support to recommendations that their long-term use be minimized, to the extent possible. They also inform decisions regarding choices between chronic prednisone administration and the use of biologics.