• osteomyelitis
• septic arthritis
• paediatrics

Short individualised treatment of paediatric bone and joint infections, based on clinical and laboratory response, has the potential to reduce the duration of antibiotic therapy, but only few data exist on this treatment.1–4 In Denmark, short individualised treatment was recommended from 2012.

This is a retrospective study of all children aged 3 months to 16 years with bone and joint infections treated between 2012 and 2016 at two paediatric departments in Copenhagen. Children with osteomyelitis (OM) were included if the diagnosis was confirmed by MRI, positron emission tomography-CT, technetium bone scintigraphy or X-ray. Children with septic arthritis (SA) were included if the diagnosis was confirmed by isolation of a microorganism from blood or synovial fluid, or by relevant findings on ultrasound examination or MRI. If the child had both OM and SA, it was treated and registered as OM. Children with osteoarticular infections secondary to other disease, traumatic injuries, surgery, or immunosuppression therapy, were excluded.

The recommended empirical treatment was intravenous cefuroxime for at least 3 days, with change to oral amoxicillin-clavulanate when the child was afebrile and showed clinical improvement and declining C reactive protein (CRP). Amoxicillin-clavulanate was recommended for 3 weeks in children with OM and for 1 week in children with SA but was discontinued only if the child had no symptoms and had normal CRP. The treatment was adjusted according to the aetiology when possible.

A total of 82 children was included (table 1). A microorganism was isolated in 60% and 55% of children with OM and SA, respectively. Detected microorganisms included Staphylococcus aureus (51%), Kingella kingae (36%), streptococci (11%) and meningococci (2%). The median age for patients with S. aureus and K. kingae was 11.3 years (IQR 7.4–12.8) and 1.3 years (IQR 1.3–1.4), respectively. All isolates of S. aureus were methicillin susceptible.

The antibiotic and surgical treatment are shown in table 2. The median follow-up time was 320 days (IQR 146–484). Seventy-eight children (96%) had no relapse or sequelae. Two children (2%) with OM had relapse after 2½ months and 6 months, respectively. Both had symptoms for more than 5 weeks before the initial diagnosis. An additional six children had symptoms for more than 14 days before the diagnosis but experienced no complications. Two children (2%) had permanent sequelae: one child had shortening of one finger due to epiphysial plate destruction. The other child had total hip replacement due to necrosis of the femoral head following S. aureus SA. He received antibiotics for 12 weeks.

In conclusion, it seemed safe to shorten the course of antibiotic treatment of bone and joint infections in this population when the duration of the treatment was individualised based on clinical and laboratory responses. However, caution should be taken in children with prolonged symptoms before diagnosis.