Background: Infection is an important cause of mortality and morbidity in patients with systemic lupus erythematosus (SLE) and a common cause for hospitalization. Glucocorticoids (GC) may contribute to increased mortality.

Objectives: We performed a 10-year retrospective study of SLE patients hospitalized for infection, and the clinical predictors of mortality, especially GC dose, in these patients.

Methods: Diagnosis codes for SLE were obtained from the electronic medical records for hospitalized patients from 2005 to 2015. Chart review was performed to ascertain the indication for hospitalization. The first hospitalization for infection (if any) was used as the index admission. Demographic and clinical characteristics, infection site and immunosuppressive drugs over the past month were abstracted. Multivariable logistic regression was used to determine predictors of all-cause mortality at 1 year.

Results: Diagnosis codes were obtained for 768 unique SLE patients with 3660 hospitalization episodes over 10 years, of which 689 had a physician diagnosis of SLE on chart review. Of these, 250 (36%) had an index admission for infection. 243/250 (97.2%) fulfilled the ACR 1997 criteria for SLE and were studied further ( Figure 1 ). Median (IQR) age was 45.1 (37.3, 56) years, 86% were female, 72% were Chinese, median (IQR) disease duration was 9 (4, 17) years. 53 (21.8%) patients had chronic kidney disease (CKD), 34 (14%) had diabetes mellitus (DM) and 12 (4.9%) had cancer. 231 (95.1%) patients were on immunosuppressive drugs and 210 (86.4%) were on GC. The median (IQR) GC dose was 8 (5, 15) mg oral prednisolone equivalent per day. The median (IQR) Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) score was 4 (1, 10). The most common organism was Escherichia coli, followed by Staphylococcus aureus and Salmonella enteritidis. The respiratory tract was the most common site of infection. 27.9% of patients who had blood cultures performed had bacteremia. There were 11 (4.5%) ICU admissions, 6 (2.5%) patients died in hospital and 1-year all-cause mortality was 13 (5.3%). SLEDAI had only a weak positive correlation with GC dose [R ² = 0.039/Pearson’s correlation = 0.197, p = 0.004)] ( Figure 2 ). Increased age (odds ratio (OR) 1.07, 95% CI 1.02-1.12, p = 0.005), average daily GC dose (OR 1.05, 95% CI 1.02-1.09, p = 0.002), bone infections (OR 42.24, 95% CI 2.76-646.8, p = 0.007) and CKD (OR 4.78, 95% CI 1.06-21.54, p = 0.04) were independent predictors of 1-year mortality, after adjusting for gender, SLEDAI, DM, and cancer ( Table 1 ).

Conclusion: Higher dose of oral GC was an independent predictor of mortality, even after adjusting for disease activity. It is important to prescribe the minimum effective dose of GC in SLE patients with infection, especially in older patients with CKD and bone infection.

Disclosure of Interests: Thurston Yan Jia Heng: None declared, Nicholas Chew: None declared, Kexin Amanda Choo: None declared, Aisha Lateef: None declared, Manjari Lahiri Grant/research support from: Manjari Lahiri is the site principal investigator for the Singapore National Biologics Register, which is a multi-pharmaceutical funded register, in which industry sponsors provide support through the Chapter of Rheumatologists, Singapore. Dr Lahiri does not personally receive any remuneration.