Skip to main content
Open Access Publications from the University of California


UCLA Previously Published Works bannerUCLA

A clinical decision tool for septic arthritis in children based on epidemiologic data of atraumatic swollen painful joints in South Africa



In settings with limited access to specialist services, differentiating septic arthritis-a surgical emergency-from non-infectious atraumatic arthropathy in paediatric patients is challenging, especially in a setting with a high burden of tuberculosis (TB). We aimed to investigate the aetiologies of swollen, painful joints in an urban setting in South Africa and determine how clinical and laboratory findings varied with diagnosis.

Patients and methods

A retrospective review of patients aged 12 or younger presenting to a paediatric hospital in Cape Town, South Africa, with atraumatic swollen, painful joints was conducted over a two year period from 2013 to 2015. Children were excluded if they did not have tissue culture or analysis conducted at our facility. Aetiology was classified as non-infectious, TB septic arthritis, or pyogenic arthritis from other bacterial causes.


One hundred and four children met inclusion criteria. Arthritis was classified as non-infectious in 43 (41%), TB in 15 (14%), and pyogenic in 40 (38%), with six (6%) patients never receiving a final diagnosis. Mean C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), and white blood cell count (WCC) were all significantly higher in pyogenic infectious arthritis compared with TB and non-infectious arthritis. There were no significant differences in these parameters between non-infectious and TB arthritis. Using cut-point analysis, thresholds were identified predictive of the presence of pyogenic arthritis versus TB or non-infectious arthritis; these included the presence of fever, CRP > 50 mg/L, ESR > 65 mm/h and WCC > 12x109/L. The absence of all of these criteria resulted in a negative predictive value of 100% for pyogenic infection; the presence of three to four criteria resulted in a positive predictive value of 71%.


Despite insignificant differences in their clinical presentation compared with non-infectious arthidities, 15% of children were diagnosed with tissue-confirmed TB infection. Predictive values of clinical criteria are reduced in our population due to elevated levels of inflammatory markers in all patients. Synovial biopsy to rule out TB is recommended in all patients in a high-burden setting given clinical similarity to non-infectious aetiologies.

Many UC-authored scholarly publications are freely available on this site because of the UC's open access policies. Let us know how this access is important for you.

Main Content
For improved accessibility of PDF content, download the file to your device.
Current View