Background and objectives: Septic arthritis of the knee is presumed to be the most frequent form of Kingella kingae -induced osteoarticular infection. This study aimed to report on the clinical course, biological parameters, and results of microbiological investigations among children with K. kingae -induced septic arthritis of the knee. It also assessed the modified Kocher-Caird criteria's ability to predict K. kingae -induced septic arthritis of the knee.

Methods: The medical charts of 51 children below 4 years old with confirmed or highly probable K. kingae -induced arthritis of the knee were reviewed. Data were gathered on the five variables in the commonly-used Kocher-Caird prediction algorithm (body temperature, refusal to bear weight, leukocytosis, erythrocyte sedimentation rate, and C-reactive protein level).

Results: Patients with K. kingae -induced arthritis of the knee usually presented with a mildly abnormal clinical picture and normal or near-normal serum levels of acute-phase reactants. Data on all five variables were available for all the children: 7 children had zero predictors; 8, 20, 12, and 4 children had 1, 2, 3, and 4 predictors, respectively; no children had 5 predictors. This gave an average of 1.96 predictive factors and a subsequent probability of ≤ 62.4% of infectious arthritis in this pediatric cohort.

Conclusions: Because the clinical features of K. kingae -induced arthritis of the knee overlap with many other conditions affecting this joint, the Kocher-Caird prediction algorithm is not sensitive enough to effectively detect K. kingae -induced septic arthritis of the knee. Excluding K. kingae -induced arthritis of the knee requires performing nucleic acid amplification assays on oropharyngeal swabs and joint fluid from those young children presenting with effusion of the knee, even in the absence of fever, leukocytosis, or a high Kocher-Caird score.