Largest Analysis Yet Defines Infection Risk from RA Biologics

May 19, 2015

Singh JA, Cameron C, Noorbaloochi S, et al. Risk of serious infection in biological treatment of patients with rheumatoid arthritis: a systematic review and meta-analysisThe Lancet. Online: 11 May 2015. DOI:

Dixon WG

Comment: Rheumatoid arthritis: biological drugs and risk of infection

The Lancet. Online: 11 May 2015. DOI:

A new meta-analysis gives the best estimates yet for the risks of serious infection among patients using biological drugs for rheumatoid arthritis (RA).

Biological drugs have about a 30% increase in serous infection, compared to disease-modifying anti-rheumatic drugs (DMARDs).

The risk was significantly greater for methotrexate-experienced patients. It was greater for standard- and high-dose, but not low-dose, biological drugs, within the power of this study.

Previous meta-analyses have been inconsistent. This is the largest meta-analysis, and the large sample size, with higher statistical power, resolves some of those inconsistencies.

This study covers 106 trials, with 42,330 patients using nine biological drugs. In all, 965 patients had serious infections, usually defined in each study as infections associated with death, hospitalization, or intravenous antibiotics.

Patients taking traditional DMARDs had an annual risk for infection of 2%, or 20 patients per 1,000.

Patients taking standard doses of biological drugs had an increase in the relative risk of 1.31 (95% credible interval 1.09-1.58). That’s equivalent to six additional cases per 1,000 patients per year.

Patients taking high doses of biological drugs had an increase in the relative risk of 1.90 (95% CrI 1.50-2.39). That’s the equivalent of 17 additional cases per 1,000 patients per year.

Patients taking combination biological therapy had an increase of 55 per 1,000 patients per year.

For methotrexate-naïve patients, the risk of infection was not increased (odds ratio 1.08, CI 0.75-1.53). This could help recently-diagnosed patients who are considering starting treatment.

However, methotrexate-experienced patients had an increase in relative risk of 70, with a much wider confidence interval.

Other studies have shown that the risk is greatest when treatment begins, and then declines.

The authors tried to interpret their results in terms that physicians could use to explain the risk to their patients.

The goal, says the editorialist, should be to provide personalized estimates of harm, and benefit, for each drug, to allow doctors to choose the right treatment for each patient.

Network meta-analysis is a relatively new technique that allows comparisons among multiple treatments simultaneously, even if they were not compared directly in head-to-head trials.