Bone infection.  Caused either by haematogenous spread (from distant site), direct innoculation (trauma) or spread from adjacent focus (eg otitis media).

Can be acute or more indolent eg 2/52+ symptoms (Brodie abcess). Mostly femur, tibia, humerus but can be anywhere. Up to 20% multifocal. Pelvic disease often presents with abdo/lumbar pain!

Septic arthritis – similar to OM, usually due to bacteraemia, same organisms. Potential for growth plate or joint consequences. Can lead to OM and vice versa.

Discitis – same bugs (probably), generally under 3-5yrs only, due to persistence of blood vessels in cartilage that subsequently atrophy – can mimic vertebral OM. Well, blood culture neg (cf vertebral osteomyelitis – >3yr, toxic, blood culture pos). Rx clindamycin for 2/52 but no evidence! Likely that some of these would resolve spontaneously if untreated.

Note also Chronic recurrent multifocal Osteomyelitis (CRMO) – clue is in the name! Sterile. Treat with NSAIDS.


Neonatal cases have wider range of organisms and generally involve both bone and joint.  Approx 50% have no systemic features, present with pseudoparalysis alone.


Epidemiology varies. MRSA predominant in US etc.

Staph aureus is the main cause, plus:

  • Neonates – GBS, candida, enterobacteriae
  • Infant – kingella, rarely pneumococc, GAS, haemophilus (type B and others), MRSA.
  • Penetrating injuries – risk of contamination, so odd organisms

Kingella, incidentally, is often resistant to vanc and clinda!

Beware salmonella (sickle cell), meningococcal (gonococcal too), TB and non tuberculous mycobateria.  But unknown in up to half, PCR vs culture…

Panton Valentine Leucocidin toxin positive associated with higher fever, higher inflammatory markers, multiple sites, chronic/recurrent.


X-ray shows mixture of fast (cortical breach, lysis) and slow (sclerosis, periosteal reaction) changes – but non-specific.

Abscess with sinus extending into soft tissue is indicative of infection.

Fistula and sequestrum (isolated bit of dead bone) more suggestive but take 1-2 weeks to develop.

“Onion skin appearance” suggests chronic infection but is also seen in normal infants, malignancy and in retinoid therapy.

USS may show sub-periosteal oedema early on.

MRI standard for defining collections and guiding surgery: 92-97% sens cf 64-71% for bone scan (latter is also operator dependent).


CRP better than ESR.  But can (like WBC) be normal.

Bone biopsy rarely done unless cancer suspected. Joint/pus aspiration can be useful, but generally diagnosis relies on blood culture, hence low identification rates. New PCR technique for Kingella from conventional culture media.


Surgical drainage if collection identified or poor response after 48-72hrs in absence of resistance. But still not well defined – over 1 yr and delayed presentation seems to have worse prognosis but will also depend on virulence and resistance.

A septic hip should always be drained promptly, but less clear for other sites – some still advocate early intervention given potential for rapid joint destruction. Aspiration and irrigation probably adequate for most, with second line open arthrotomy (laparoscopic view is limited)

Immobilization +/- traction are considered good practice. For SA, window of 4hrs considered acceptable to get samples before starting antibiotics.

Poor evidence for antibiotic treatment.  Current recommendations:

Cefuroxime under <6yrs (non-neonatal) else Fluclox  or clinda. Beware PVL staph, Hib unimmunized and risk factors for MRSA.  Broad spectrum if neonate, immuncompromised or sickle cell disease.

For all antibiotics use the maximum dose. Never give rifampicin monotherapy as resistance quickly induced. Beware erythromycin resistant MRSA and clinda (inducible resistance). Teico good for bone (keep level>10).

Switch to oral when clinically improved +/- CRP response (eg<20 or down more than 2/3 of peak), median 4 days.  Under 3 months, give full 4/52 treatment course IV.

Failure to improve –

  • consider metastatic infection, DVT, unusual organism eg Fusobacterium. PVL pos
  • IV for 21 days minimum, total 6/52 minimum (even SA)
  • HPA recommend max dose clinda, rifamp AND linezolid. Max 4/52 linez – neuropathy.

Total duration 4-6 weeks for OM, in textbooks, 3 weeks for SA.

Recent HPA guidelines suggest initial therapy for deep seated PVL positive Staphylococcus aureus infections in children as intravenous clindamycin plus rifampicin and linezolid (linezolid maximum of 4 weeks due to the risk of development of peripheral neuropathy), followed by clindamycin plus rifampicin.  Beware thrombosis.

Shorter courses eg 10 days? Peltola’s Finnish studies of shorter courses may not be applicable, as possibly selected for milder infection, exclude culture negative but have high rates of positive diagnosis (staph). [Current Opinion in Pediatrics. 25(1):58-63, 2013 Feb. UI: 23283291] Howard-Jones & Isaacs in JPedsCH however feel evidence for 3/52 courses (grade 2B), emphasize QDS dosing. 1 small old study suggested nafcillin/methicillin inferior to cefuroxime, so Clinda preferable to Fluclox. [Jourrnal of Paediatrics and Child Health 49 (2013) 760–76 UI: 23745943]

[Saul Faust Arch Dis Child 2012;97:545–553. doi:10.1136/archdischild-2011-301089]