Article Text

SAT0571 Performing septic arthritis diagnosis in scenarios where synovial fluid is not available: multivariate logistic regression analysis
  1. CA Guillen-Astete1,
  2. GA Gabilondo-Alvarez2
  1. 1Rheumatology Department
  2. 2Microbiology, Ramon y Cajal University Hospital, Madrid, Spain


Background In order to establish the diagnosis of septic arthritis (SA) it is necessary to demonstrate the presence of bacteria into the synovial fluid. However, clinics should act according to its suspicious when the differential diagnosis included this possibility even when the microbiologic study of the synovial fluid in unavailable by any cause (small joints, lack of training or logistic deficiencies).

Objectives The purpose of our study is to determine the relative weight of other clinical or analytical variables that should be useful in these scenarios.

Methods A retrospective multivariate logistic regression analysis about the registries of monoarthritis assessed in our unit between 2013 and 2016. There were included only registries with cases of synovial joints with all parameters of interest available. The binary response variable was the microbiological demonstration of septic arthritis (culture/Gram stain of synovial fluid or tissue). Explanatory variables were age (stratified in <30, 30–39, 40–49, 50–59, 60–69 and >70 years old), gender, temperature above of 37.9°C, recount of neutrophils in peripheral blood sample, measure of procalcitonine (PCT), and measure of C reactive protein (CRP). Synovial fluid study, although was available in almost all cases was deliberately omitted for purpose of this study. The logistic regression analysis used the forward model strategy.

Results There were included 449 registries. One hundred an sisteen of them were SA. The fixed model showed a Chi-square=226.64 with 5 degrees of freedom and a P<0.0001. Explanatory variable gender was excluded from the forward model strategy due to its lack of impact on the binary response variable. The Odds Ratio for PCR, PCT >1.3 ng/dL, age, neutrofils recount and body temperature were 1.0175, 8.1588, 0.5135, 1.0001 and 3.4147, respectively. The model showed a sensibility of 68.1%, specificity of 94.8%, PPV of 82.2% and NPV of 89.5%. The following table shows the full results of the logistic regression analysis: Standard error, Wald index, p value and R coefficients.

Conclusions Previous studies have demonstrated the usefulness of PCT measure for the diagnosis of SA however all of them have been based on joints which synovial fluid is quite easy to obtain for further analysis. Present study allows laying the ground for creation of future diagnostic models based on five clinical variables that could be useful on scenarios where the assessment by joint puncture is not available.

Disclosure of Interest None declared

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.