Pediatric Septic Arthritis of the Knee: Predictors of Septic Hip Do Not Apply

Background:The early diagnosis of pediatric septic arthritis is paramount to the prevention of long-term sequela. The purpose of this study is to investigate if the commonly used criteria developed by Kocher and colleagues for hip septic arthritis can be used for screening children with suspected septic knee.Methods:We retrospectively reviewed the charts of patients under the age of 19, between June 2002 to June 2017, who presented to a major tertiary-care children's hospital with septic knee. Diagnostic criteria included either a positive synovial culture from the knee, synovial white blood cell (WBC)>50,000 cells/mm³, or synovial WBC count >25,000 cells/mm³ and clinical agreement of diagnosis from Infectious Disease and Orthopaedic colleagues. Collected data included the initial criteria described by Kocher and colleagues: history of fever, non-weight-bearing, erythrocyte sedimentation rate, and serum WBC as well as a recently modified criterion: C-reactive protein. Univariate analysis was used to determine the quality of these variables in ruling out septic knee.Results:One hundred four patient charts were found to meet our inclusion criteria demonstrating C-reactive protein>20 mg/L (75%), fever (65%), non-weight-bearing status (64%), erythrocyte sedimentation rate>40 mm/h (60%), and WBC>12,000 cells/mm³ (49%). With the 25 different combinations of these predictors adjusted for, in an escalating manner, 0 predictors suggested a sensitivity of 0.02, 1 predictor a sensitivity of 0.06, 2 predictors a sensitivity of 0.2, 3 predictors a sensitivity of 0.32, 4 predictors a sensitivity of 0.3, and 5 predictors a sensitivity of 0.11.Conclusions:According to the Kocher criteria of the hip, at 3 or more criteria the probability of septic arthritis becomes 93% with a sensitivity of 0.84 provoking many physicians to use this cutoff in their assessment of hip pain. This study suggests that if these criteria were applied to the knee, 52% of septic knee cases could be missed. There is a need for further investigation of specific criteria of the knee as the markers of the hip septic arthritis are not necessarily applicable in the knee.Level of Evidence:Level III.