SAT0698-HPR SYSTEMATIC LITERATURE REVIEW AND META-ANALYSIS OF INFECTION TYPES AND THE EFFECT OF MEDICATIONS ON INFECTION IN SYSTEMIC LUPUS ERYTHEMATOSUS

Background The infection incidence rates are high in systemic lupus erythematosus (SLE) patients1 and the medical literature documents broad categories of specific bacterial infection risks for SLE patients.2-4 However, there has not been a detailed incidence of infection types and infection predictors. Objectives In a systemic literature review (SLR) and meta-analysis, to describe the types and incidences of infections in SLE patients and to examine the effect of medications used to treat SLE. Methods An SLR of SLE and infection articles generated through a search of MeSH terms using PubMed and Medline generated 1,211 articles. Using predefined Incl/Exc criteria, data from 32 accepted articles were double-extracted and descriptive and multivariable analyses were conducted. Relative risks between drug classes were estimated using Arm-Based Network Meta-Analysis. Hypothesis tests were two-sided and a p-value Results 4,130 patients were considered, 91% females, average age: 36.8 (10.9) years, mean disease duration: 6.9 (6.1) years, SLEDAI mean: 11.7 (4.6). In drug trials, 775 used conventional synthetic DMARDs (csDMARDs), 1,809 (43.8%) used biologic DMARDs (bDMARDs), 691 (16.7%) used placebo and the rest were in non-drug specific observational studies. 90.4% used background corticosteroid therapy. Bacterial infections occurred most commonly, being 61% of the total of 315 non-serious infections, of which 28.4% were gram negative and 15.3% gram positive. 135 (42.9%) had viral infection – H zoster, HPV, and CMV were most common. 36 (11.4%) were opportunistic infections - 12 candida, 11 mycobacterial, 4 pneumocystis jiroveci pneumonia, and 9 unspecified infections. There were 182 serious infections (SIE), of which 110 were bacterial (gram negative were the most common subtype (36.3%)). Serious opportunistic infections accounted for 31 (17%) of the SIEs, similar to serious viral infections 41 (22.5%). Patients using csDMARDs had an absolute risk to develop H Zoster of 15.2% (95% CI: 7.7-28.2), while it was 8.2% (CI: 1.9-29.2%) when using bDMARDs. The next most likely infection for which these SLE patients were at risk was for gram negative infections, where csDMARDs demonstrated a 7% absolute risk (CI: 0.4-36%) of developing a gram negative infection, compared to 5.6% (CI: 0.1-52.6%) for bDMARDs. The attributable risk for developing gram positive infections when using csDMARDs was 3.9% (CI: 0.2-25%) and 3.5% (CI: 0.1-41.6%) when using bDMARDs. Conclusion In this SLR and meta-analysis in SLE, the frequency of infections was bacterial>viral > opportunistic, in that order, although some details were unavailable. csDMARDs were associated with more infections than bDMARDs. References [1] Ramos M, Cuadrado MJ, Abla P, et al. (2008) Acute viral infections in patients with SLE: description of 23 cases and review of the literature. Medicine (Baltimore) 87:311-318. [2] Feldman, et al. (2015) Serious infections among adult Medicaid beneficiaries with SLE and lupus nephritis. Arthritis Rheumatol 67: 1577-1585. [3] Dubula T, Mody GM (2015) Spectrum of infections and outcome among hospitalized South Africans with SLE. Clin Rheumatol 34: 479-488. [4] Bouza E, Moya JG, Munoz P (2001) Infections in SLE and RA. Infect Dis Clin North Am 15: 335-361. Acknowledgement J. Grotts Disclosure of Interests Nashla S Barroso: None declared, Aly M Aly: None declared, Ethan Zaccagnino: None declared, Wendy Li: None declared, Harsh Agrawal: None declared, Sarah M Doaty: None declared, Daniel Furst Grant/research support from: F. Hoffmann-La Roche, Genentech