Background Septic arthritis is a multidisciplinary emergency which poses diagnostic problems due to the lack of biological and clinical specificity, and variable treatment modalities in the absence of formal consensus.

Objectives The objective of this study was to describe epidemiology, management and evolution of rheumatologic septic arthritis over the past 50 years.

Methods This epidemiological, retrospective, observational, single-center study, included patients hospitalized between 1997 and 2014 in one Rheumatology Department of a University Hospital, for septic peripheral arthritis of native joint. The diagnosis was determined according to Newman's criteria. After a first descriptive part, we tried to determine associated factors for joint morbidity (define by the use of a non-sparing surgery or functional sequelae) and mortality after a follow-up of 1 year. Finally, we compared the results with a previous study, conducted in the same Department between 1966 and 1993 over 176 cases.

Results On the 1997–2014 period, 198 patients were included corresponding to an annual incidence of 11 cases. The average age of patients was 56 years, with 57% men. Main comorbidities were arthropathy (37%), neoplasia (18%) and diabetes (17%). The most common presentation was monoarthritis (88%), of the knee (32%), caused by Staphylococcus aureus (56%). Nine percent of patients had received prior intra articular intervention. We noted 15% of arthritis with no identified pathogen. Endocarditis was found in 6% of patients. The support was multidisciplinary particularly between 2006 and 2014, with an increase of cooperation with bacteriologist (15% of cases between 1997–2005 versus 67% between 2006 and 2014, p<0.0001), and a decrease in antibiotic duration (97.1 days versus 75.5 days, p=0.002). The surgical treatment was required in 33% of cases. The death rate at 1 year was 6% and the joint morbidity rate was 20%. Mortality risk factors have been identified (endocarditis OR =12.6 [2.0 to 81.4], p=0.008); MRSA OR =13.8 [1.9 to 100.6], p=0.009). Patients receiving corticosteroids had an excess risk of joint morbidity (OR =3.0 [1.0 to 8.7], p=0.04). In comparison with the previous study, there were more comorbidities (p<0.01 for diabetes, neoplasia, rheumatoid arthritis), more bacterial identification and surgical care (p<0.01). Iatrogenic origin and mortality remain stable.

Conclusions In the absence of formal recommendations, multidisciplinary cooperation could impact directly on the diagnostic and therapeutic management of septic arthritis. The management seems better than in the 60–90's period, with no change in mortality. Infective endocarditis appears to be an important risk factor for mortality and its research should be systematic.

Disclosure of Interest None declared