Background Infections are an important cause of morbidity and mortality in the rheumatoid arthritis (RA) population. Infection risk in biological-exposed RA is often compared to the risk in patients exposed to non-biological disease modifying anti-rheumatic drugs (nbDMARDs). An understanding of the baseline risk of infection in nbDMARD-exposed RA and whether this has changed over time will inform analyses examining infection risk in the RA population.
Objectives To examine infection risks in the nbDMARD-exposed RA population and compare them to the general population over time and with different exposure definitions.
Methods Adult, unselected patients with prevalent RA were identified from the Swedish National Patient Register and required to have ≥2 visits in inpatient or outpatient care with ≥1 to a specialist and ≥1 in outpatient care. Serious infection (SI) was defined as any hospitalization with an ICD code for infection as main or contributory diagnosis listed in the inpatient register through Dec 2013.
Information on DMARD treatment was obtained from the prescription drug register (PDR; Jul 2005-Dec 2012). Patients ever treated with a biologic DMARD were excluded/censored. We investigated several different ndDMARD exposure definitions, all employing a new-user approach: any change (2005-2012) to a new nbDMARD, 1st recorded change, 2nd change and 3rd change. We also varied the number of days between the 1st and 2nd nbDMARDs: 60 days between the 1st and 2nd, 180 days or 365 days.
General population comparators were selected at random from the total population register matched to each unique patient with RA on sex, year of birth, county and required to be alive and living in Sweden at the time of start of follow-up of their index RA patient.
The 1-year risk of SI counting from ndDMARD treatment change was calculated overall and by calendar year. Modified Poisson regression with sandwich estimators were used to estimate risk ratios overall and by calendar year comparing the nbDMARD-exposed to the general population.
Results The risk of SI within 1 year in nbDMARD-exposed RA was 5.9% compared to 2.2% in the general population. The risk of SI did not vary over calendar time in 1st switchers (Figure). The risk of SI within 1 year of switching was 6.1% for 1st switchers, 6.0% for 2nd switchers and 6.8% for 3rd switchers. Overall, regardless of nbDMARD exposure definition, the nbDMARD-exposed RA population had a 3-fold increased risk of SI compared to the general population (1st switchers RR 2.96; 95%CI 2.69, 3.36). The amount of time required for individuals to be on the nbDMARD before switching to the next did not change the observed risks.
Conclusions The risk of infection with nbDMARD in RA is 3 times the risk in the general population. Serious infection risk did not vary appreciably depending on the exposure definition used and has remained constant over calendar time.
Disclosure of Interest None declared