SAT0190 Risk for opportunistic infections in rheumatoid arthritis treated with bdmards in clinical practice, 10 years of follow up

• Published in: Annals of the rheumatic diseases Vol. 77; no. Suppl 2; p. 956
• Main Authors: Leon, L., Peñuela, M., Candel, F. J., Freites, D., Rodriguez-Rodriguez, L., Abasolo, L.
• Format: Journal Article
• Language: English
• Published: London BMJ Publishing Group LTD 01.06.2018
• Subjects: Candida; Corticosteroids; Epstein-Barr virus; Etanercept; Fungi; Herpes zoster; Infliximab; Leukopenia; Monoclonal antibodies; Multivariate analysis; Patients; Regression analysis; Rheumatoid arthritis; Rituximab; Steroid hormones; Tumor necrosis factor-α
• ISSN: 0003-4967; 1468-2060
• DOI: 10.1136/annrheumdis-2018-eular.5424

Background:Biologic disease-modifying anti-rheumatic drugs (bDMARDs) may be associated with opportunistic infections (OI).Objectives:Our purposes were to describe the incidence of OI in Rheumatoid Arthritis (RA) taking bDMARDs, and compare the risk of OI between TNF-targeted and non-TNF-targeted biologics.Methods:Retrospective longitudinal observational study from 2007 to 2017. We included RA patients, diagnosed and followed in our outpatient clinic, whom started treatment with a TNF-targeted bDMARD [etanercept (ETN), golimumab (GOLI), certolizumab (CTZ), infliximab (IFX), adalimumab (ADA)], or non-TNF-targeted bDMARD [rituximab (RTX), abatacept (ABA), or tocilizumab (TCZ)]. According to microbiologist criteria, we consider OI when there was a positive culture (virus, fungus, bacterial, parasitary) or compatible symptoms that respond to specific treatment. The independent variable was the Type of targeted bDMARD: TNF-targeted vs non-TNF-targeted. Secondary variables: sociodemographic; clinical and other therapies. We used survival techniques to estimate the incidence of OI, expressed per 1000 patient-year [CI 95 %]. The exposure time was defined from the start date of each bDMARD to the development of an OI, loss of follow up or end of study (01/02/2017). We performed Cox multivariate regression models to compare the risk of OI between the types of bDMARD used. Results were expressed in Hazard ratio (HR).Results:441 RA patients were included, starting 761 different courses of bDMARD treatment. 81% were women with a mean age at first bDMARD of 57.3±14 years. The median time from onset of bDMARD until onset of OI was 3.1 years [0.5–4.6]. More than 90% of patients were on steroids. 71.3% of the courses were TNF-targeted bDMARD (ADA 39%, ETN 33%, CTZ 15%, IFX 9%, GOLI 2%), and 28.7% non-TNF-targeted bDMARD (RTX 60%, ABA 21%, TCZ 18%). There were 38 OI [26 Virus (18 Herpes Zoster, 2 virus B reactivation, 3 virus C reactivation, 1 Epstein Barr, 1 Avian flu, 1 CMV), 6 Fungus (5 Candida, 1 Trichophyton), 5 Bacterian (1 Legionella, 1 Salmonella, 3 TB)] and 1 parasitary (Leishmania)]. 9 of them required hospitalization and one died (Candida). The global incidence of OI was 23.1 [16.8–31.8]. TNF-targeted bDMARD had 26 OI, with an incidence of 20.8 [14.1–30.6], and non-TNF-targeted bDMARD with 12 OI, incidence 30.4 [17.3–53.6]. We not find statistical differences in the rate of OI between TNF-targeted vs non-TNF-targeted in the multivariate model, adjusted by age, sex and calendar-time (HR 1.37, p=0.4). Male sex was found a predictor of OI in the multivariate analysis (HR 2.17, p=0.04). Age (HR 1.02, p=0.08), concomitant treatment with corticosteroids (HR 6.61, p=0.05) and leukopenia (HR 2.7, p=0.08) showed a tendency to increase the risk of OI.Conclusions:Incidence of OI due to bDMARDs was near 23 cases per 1000 patients–year. Crude incidence was higher for non-TNF-targeted bDMARD compared to TNF-targeted bDMARD. Nevertheless this difference was not maintained in the multivariate model, reflecting that many of the variability in patients' risk of OI development were driven by factors other than biological agent exposure. Close monitoring should be taken in those RA patients taking bDMARD, corticoids, and with leukopenia.Disclosure of Interest:None declared