Use of glucocorticoids — both current and past — heightens the risk for serious infections in older patients with rheumatoid arthritis, especially if taken for long periods, a case-control analysis determined.

For instance, a patient who has taken 5 mg of prednisolone daily for 3 months has a 30 percent increase in risk of being hospitalized for an infection, according to William G. Dixon, MBBS, PhD, of the University of Manchester in England, and colleagues.

But if that dose — considered low by many clinicians — was taken for 3 years, the increased risk of a serious infection reached 100 percent, the researchers reported in the July Annals of the Rheumatic Diseases.

It has long been recognized that extended use of glucocorticoids has many potential adverse effects, including weight gain, diabetes, cardiovascular disease, and disturbances in immune function.

"Surprisingly, after over 60 years of experience, it is not clear how the risk of these outcomes depends on treatment regimens, what degree of risk is conferred, or what happens to risk on stopping therapy," observed Dixon and colleagues.

Infections, too, are an acknowledged hazard associated with steroid use because of interference with cell-mediated immune responses.

Past efforts to quantify the infection risks have been hampered by the use of statistical methods that are unable to account for patterns of drug use over time, which is a salient factor with steroids because doses typically are increased and decreased in response to symptoms and side effects.

So Dixon's group constructed a series of analytical models based on weighted cumulative doses, which reflect not only the quantity of drug and duration of use, but also how recently the drug was used.

Their study population included 1,947 patients 65 and older from an administrative database in Quebec who had at least one prescription for a disease-modifying antirheumatic drug from 1985 to 2003, and who were hospitalized for a first serious infection.

A matched control group consisted of 9,735 patients who had received treatment for rheumatoid arthritis but who had never developed a serious infection.

Prednisone doses were averaged from a prescription drug database and weighted for importance, with adjustments for the additional use of injectable steroids along with the oral formulations.

Potential confounders in the analysis included disease severity, comorbidities, and co-administration of other immune-suppressing agents, so these factors were adjusted for in the analytic models.

The researchers first calculated risks according to conventional binary approaches, using classifications such as current use or never use, and while these models showed significantly increased risks, the risks estimated in the weighted cumulative dose models were considerably more accurate, they found.

Among the weighted estimated infection risks according to dose and duration were the following:

• 5 mg per day for the past week
• 5 mg for the past 6 months
• 30 mg for the past week
• 30 mg for the past 6 months

Discontinuing the glucocorticoid also influenced risk. For a patient who had taken 5 mg daily for 6 months but discontinued use 6 months ago, the infection risk fell to 1.06.

However, risks remained even when drug exposure was relatively distant, such as in a patient who had stopped the steroids more than 2 years ago after taking the drugs for several years.

In such a case, the distant doses were weighted lower than recent doses would have been, but the long duration of use would yield increased risk.

This type of finding would not be seen with conventional statistical models, according to the researchers.

In explaining the persistent risks on infection with steroid use, they pointed to effects on both innate and adaptive immunity.

"Impaired innate immunity would be expected to have an immediate impact, but may be predicted to recover following glucocorticoid discontinuation. Our data show that 'remote' therapy has a lesser, but still important, impact," they wrote.

But the effects on the adaptive immune system may be less responsive to drug withdrawal, in that influences on T-cell function may result in "failure to generate pathogen-specific adaptive immune responses," they observed.

Limitations of the study included the possibility of confounding, although adjustments were made for concomitant drug therapy and disease severity, and a lack of information about treatment adherence.

Source: Infection Risks Persist with Steroid Use in RA