Background In RELESSER (Spanish Society of Rheumatology Systemic Lupus Erythematosus-SLE-Registry) bacteremia is the main cause of death by infection. The available information about this severe infection in SLE patients is scarce.

Methods Retrospective nested case-control study of SLE patients (ACR97 criteria) with at least a bacteremic episode and random controls from RELESSER. Descriptive, bivariate and multivariate analysis (logistic regression).

Results 114 bacteremic episodes in 83 patients were found. Incidence rate: 2.7/1,000 patient-years (total n:3658). At the time of the bacteremia: median age: 40.5 (8Ð90)years, 88.6%female, disease duration:9.7 (IR16.7), medianSELENASLEDAI:4 (IR8), severe flare (SFI criteria):66%, active nephritis:16.7%, median SLICC/ACR DI: 3 (IR4), any comorbidity:64% (McCabe-Jackson criteria: 28.1% rapidly/ultimately fatal), more frequently renal failure (15.8%) or diabetes (11.4%). SLE treatment at bacteremia: 88.6%corticosteroids (68,6%>10 mg/day), 57%immunosuppresors (mycophenolate17.5% and cyclophosphamide (CYP)12.3%), 27%antimalarials. 44.7% suffered invasive procedures, more frequently intravascular catheter (24.6%). Nosocomial bacteremia in 35.1%, more frequently urinary source (27.2%). 64% developed systemic inflammatory response syndrome, 35% needed intensive care unit admission, 22.8% had multiorganic failure. The most frequent microorganisms were E.coli (29.8%), Staphylococcus aureus (16.7%) (22% methicillin-resistant) and Salmonella spp. (10.5%). 16% of the gram-negative enteric bacilli were extended- spectrum b-lactamase positive. 17.5%were multidrug resistant. 68.4%started the antibiotherapy before blood culture results, resulting finally active in susceptibility testing in 56 (71.8%), indicating an appropriate empirical antibiotic therapy in 49%. The bacteremia-related mortality was 14%. Risk of death was higher in patients with severe sepsis (Pitt index >8) (OR: 13, 95% CI: 3.71 to 45.17). Bacteremia was recurrent in 26.3%. Associations with bacteremia in bivariate analysis (114 bacteremias vs 688 controls) are shown in table 1. Antimalarials were protective. In the multivariate analysis (adjusted for disease duration), only elevated creatinine (OR: 1.31 (95% CI: 1.01 to 1.70) p=0.045), diabetes (OR 6.01 (95% CI: 2.26 to 15.95) p=0.000), cancer (OR: 5.32 (95% CI: 2.23 to 12.70), p=0.000), immunosuppressors (OR: 6.35 (95% CI: 3.42 to 11.77) p=0.000), CYP (OR: 9.37 (95% CI: 5.12 to 17.14) p=0.000) and SLICC/ACR DI (OR: 1.65 (95% CI: 1.31 to 2.09) p=0.000) remained statistically significant.

Conclusion Bacteremia occurred mostly in active SLE, frequently in the context of a severe flare. Gram negative bacilli predominated, with high rate of multidrug resistance. The empiric treatment was inappropriate in a half of the cases. Recurrence and mortality were high. Immunosuppressors use, comorbidity and damage were all associated to bacteremia.