Incidence, associated factors and clinical impact of severe infections in a large, multicentric cohort of patients with systemic lupus erythematosus

Abstract

Objectives To estimate the incidence of severe infection and investigate the associated factors and clinical impact in a large systemic lupus erythematosus (SLE) retrospective cohort. Methods All patients in the Spanish Rheumatology Society Lupus Registry (RELESSER) who meet ≥4 ACR-97 SLE criteria were retrospectively investigated for severe infections. Patients with and without infections were compared in terms of SLE severity, damage, comorbidities, and demographic characteristics. A multivariable Cox regression model was built to calculate hazard ratios (HRs) for the first infection. Results A total of 3658 SLE patients were included: 90% female, median age 32.9 years (DQ 9.7), and mean follow-up (months) 120.2 (±87.6). A total of 705 (19.3%) patients suffered ≥1 severe infection. Total severe infections recorded in these patients numbered 1227. The incidence rate was 29.2 (95% CI: 27.6–30.9) infections per 1000 patient years. Time from first infection to second infection was significantly shorter than time from diagnosis to first infection (p < 0.000). Although respiratory infections were the most common (35.5%), bloodstream infections were the most frequent cause of mortality by infection (42.0%). In the Cox regression analysis, the following were all associated with infection: age at diagnosis (HR = 1.016, 95% CI: 1.009–1.023), Latin-American (Amerindian-Mestizo) ethnicity (HR = 2.151, 95% CI: 1.539–3.005), corticosteroids (≥10 mg/day) (HR = 1.271, 95% CI: 1.034–1.561), immunosuppressors (HR = 1.348, 95% CI: 1.079–1.684), hospitalization by SLE (HR = 2.567, 95% CI: 1.905–3.459), Katz severity index (HR = 1.160, 95% CI: 1.105–1.217), SLICC/ACR damage index (HR = 1.069, 95% CI: 1.031–1.108), and smoking (HR = 1.332, 95% CI: 1.121–1.583). Duration of antimalarial use (months) proved protective (HR = 0.998, 95% CI: 0.997–0.999). Conclusions Severe infection constitutes a predictor of poor prognosis in SLE patients, is more common in Latin-Americans and is associated with age, previous infection, and smoking. Antimalarials exerted a protective effect.