• Summary
  Click here to print a Quick Reference Guide
• Background
• Clinical Diagnosis
• Investigation
• Treatment
• Non-Antimicrobial Treatment
• Empirical Antimicrobial Treatment
• Directed Antimicrobial Treatment (when microbiology results are known)
• Duration of Treatment
• Switch to oral agent(s)
• Treatment Failure

These guidelines apply to children with acute haematogenous osteomyelitis but excluding those with spinal infection. Click for links to guideline for osteomyelitis in neonates and deep spinal infection.

Diagnosis is clinically based:
History: Usually rapid onset and progressive pain in affected bone. Fevers and systemic upset is variable. Infants and younger children – parent may report reduced use of limb or limping.
Ask about: trauma [e.g. nail penetration injury in sole of foot], penicillin [and other] allergies; Haemophilus influenzae vaccination status; information on previous MRSA carriage/infection.

Examination: Look for swelling and bony tenderness to palpation. Assess systemic signs of infection. An infant may not be using the affected limb.

Investigations

• Routine blood tests: Blood Cultures [before antimicrobials] FBC and CRP.
• A plain x-ray of affected bone[s].
• Ultrasound scan – if abscess/septic arthritis suspected or severe sepsis.
• MRI – if infection of spine or pelvis suspected.

Non-antimicrobial treatment
Pain relief and splintage of adjacent joint in a cast as determined by orthopaedic consultant
Fluid resuscitation as necessary

Antimicrobial treatment

The following is a summary of initial empirical intravenous antibiotic treatment. Antibiotics should not be given until Blood Cultures have been taken.

The recommended antibiotics doses are in agreement with BNFC 2020. Dose reductions may be required in patients with hepatic or renal impairment, consult product literature or pharmacy for further information.

Children and infants age 1 month to 5 years:
Intravenous Cefuroxime (50–60 mg/kg (max. 1.5 grams) every 6)

Children age 5 years to 16 years:
Intravenous Flucloxacillin (50 mg/kg (max. 2 grams) every 6 hours.)
add Cefotaxime in severe sepsis (50 mg/kg every 6 hours (max. 12g daily)) pending Blood Cultures and PCR.

Children and infants age 1 month to 16 years with a true immediate-type penicillin allergy:
Intravenous Vancomycin (15 mg/kg every 6 hours (do not exceed an initial maximum of 750mg every 6 hours). Adjust dose according to plasma concentrations, aim for ‘trough’ levels of 10-20mg/L) PLUS
Intravenous Ciprofloxacin (10 mg/kg (max. 400 mg) every 8 hours).

Children and infants age 1 month to 16 years with a previous MRSA infection or colonisation:
Intravenous Vancomycin (15 mg/kg every 6 hours (not exceed an initial maximum of 750mg every 6 hours). Adjust dose according to plasma concentrations, aim for ‘trough’ levels of 10-20mg/L) PLUS
Intravenous Cefuroxime (50–60 mg/kg (max. 1.5 grams) every 6 hours).

Duration
6 weeks (intravenous initially and completed orally) is the mainstay of treatment.

Oral switch
See full guideline

Back to top

Pathogenesis
Osteomyelitis can occur via haematogenous seeding of bone, by direct spread from an adjacent focus of infection or as a complication of open fractures or surgery. In children osteomyelitis is mostly commonly acute and caused by seeding of the metaphyses of long bones [proximal tibia and distal femur most often] via the bloodstream (Berbari et al., 2005)

Osteomyelitis in children is usually haematogenous but with the exception of skin rashes e.g. chickenpox a primary focus is seldom identified. It is believed that the anatomy of the metaphysis may predispose to infection (Berbari et al., 2005). Loops of capillaries extend from the nutrient artery under the growth plate and feed into venous sinusoids where blood flow is sluggish. Low oxygen tension may retard white cell phagocytosis. Naturally occurring bacteraemic episodes result in bacterial seeding in this environment and when the virulence/host resistance balance tips then infection develops. Antecedent local minor trauma is associated with the onset of osteomyelitis possibly due to local bone bruising and further capillary circulation compromise but a convincing trauma history is not particularly common in the author’s experience. Infection is usually in a single site but occasionally multifocal disease occurs.

Contamination of open fractures can lead to deep fracture site infection if there is not urgent and adequate surgical debridement. This form of bone injury/infection requires specialized orthopaedic/plastic surgical management and is not considered further here.

Microbiology
The causative organism of bone infection varies with the age of the child and this difference in epidemiology influences the choice of empirical antimicrobial therapy.

The most common causes of osteomyelitis in children >1 months up to 5 years are: Staphylococcus aureus, Kingella kingae, Streptococcus pneumoniae and beta-haemolytic streptococci [mostly A and B] (Chometon et al., 2007; Rasmont et al., 2008; Yagupsky et al., 1992).

The most common causes of osteomyelitis in children >5 to 16 years are Staphylococcus aureus and streptococci(Yagupsky et al., 1992).

Neisseria meningitidis, Haemophilus influenzae, Enterobacteriacaea [“coliforms”], and other streptococci are uncommon causes of osteomyelitis (Chometon et al., 2007). Salmonella species should be considered in children with haemaglobinopathies.

Pseudomonas aeruginosa should be considered when puncture wounds to the sole of the foot are the cause of infection.

Culture negative osteomyelitis
Osteomyelitis is a very rare manifestation of cat-scratch disease [caused by Bartonella henselae] (de Kort et al., 2006). Since Bartonella henselae is not culturable using routine microbiological techniques, PCR on bone biopsy is required to make a diagnosis. Tuberculosis should be considered in more insidious infections with epidemiological risk factors or typical radiological findings.

Back to top

The typical history is of a previously well child who develops a febrile illness and malaise with localized pain in an extremity. In infants the history is necessarily sketchy but systemic features as above and “pseudoparalysis” of the affected limb draws the parents’ attention. Older children will provide an account of increasing localized pain and demonstrate a limp. It is important to ask about: penicillin [and other] allergies, Haemophilus influenzae vaccination status and information on previous MRSA carriage/infection because these will influence empirical therapy. Any relevant history of recent trauma should be elicited e.g. stepping on a rusty nail etc.

Examination shows swelling and bony tenderness to palpation especially in subcutaneous bones [e.g. around the knee and ankle]. A fever is usual [>90%] but not universal. On occasion the child will present with septic shock reflecting delayed diagnosis or an unusually aggressive organism. These are medical and possibly surgical emergencies [e.g. if a collection is seen on scanning].

Infection is usually acute but a more insidious sub-acute variant is recognized [Brodie’s abscess] reflecting an altered balance of bacterial virulence and host resistance.

Panton Valentine Leukocidin Staphylococcus Aureus (PVL-SA)
Where a child has bone and/or joint infection and has severe systemic disease and in particular an associated chest infection then PVL (Panton-Valentine Leucocidin) variant SA infection should be suspected. An aggressive approach to imaging, surgical drainage and anti-microbial medication is required. General supportive treatment is needed and such cases often require ICU care. If PVL-SA is suspected please Discuss with Microbiology.

Back to top

1. Blood tests
   Recommendation: Blood Cultures should be sent prior to starting antimicrobials in all patients.
   [Evidence level C]
   Bone biopsies are rarely performed in children therefore the microbiological diagnosis of osteomyelitis usually relies on blood cultures. Blood Cultures provide a microbiological diagnosis in ~60% of cases.
   
   Recommendation: Full blood count and CRP should be sent from all patients being treated with intravenous antimicrobials at baseline.
   [Evidence level D]
   CRP is more sensitive than ESR, increases and responds to treatment earlier. NB CRP is occasionally normal, at least initially, in acute haematogenous osteomyelitis. Baseline white cell count [WCC] can be helpful, again the WCC may be normal at least initially in acute haematogenous osteomyelitis.
   
   Recommendation: Bone biopsies should be undertaken in immunocompromised children, when tuberculosis is suspected, when the clinical and radiological diagnosis is uncertain, after treatment failure on first line therapy or if surgery is planned anyway.
   [Evidence level D]
   Bone biopsies are not essential in most children and Blood Cultures are the microbiological investigation of choice, however, in some situations obtaining microbiological confirmation is critical to successful therapy.
   
   Recommendation: Bone biopsies should be collected using aseptic technique and placed in sterile universal containers, where possible a portion of sample or fluid should be directly inoculated into a paediatric blood culture bottle.
   [Evidence level C/B]
   Kingella kingae can be difficult to grow and its recovery can be enhanced using blood culture media.
   
   Recommendation: Send a blood sample for Bartonella henselae serology in children with osteomyelitis, negative Blood Cultures and a history of cat exposure, such cases should be discussed with microbiology first.
   [Evidence level D]
   Bartonella henselae [the cause of cat scratch disease] is not culturable using routine microbiological techniques, PCR on bone biopsy or serology is required to make a diagnosis.

2. Imaging
   Recommendation: plain x-rays of the affected bone[s] should be undertaken in all cases of suspected osteomyelitis.
   [Evidence level C]
   
   Bone lysis and periosteal elevation are the hallmarks of acute haematogenous osteomyelitis but are usually absent during the initial 10 days or so of the evolution of the disease. Children usually present within 2 or 3 days of symptom onset but plain Xrays are still helpful as a baseline and may occasionally demonstrate the painful limb is in fact broken or has some other pathological lesion.
   
   Recommendation: Ultrasound scanning of the affected bone[s] should be undertaken in children with clinical evidence of severe systemic infection and suspected osteomyelitis or if concomitant septic arthritis is suspected.
   [Evidence level D]
   
   Ultrasound is useful in selected cases to exclude involvement of adjacent joints [NB the proximal femoral metaphysis is one of the few intra-articular metaphyses and contiguous spread is not uncommon]. Ultrasound will also demonstrate subperiosteal abscess collection in the late presenting case or where medical treatment is failing. Urgent surgical drainage is required in both cases. Useful in the initial assessment of the septicaemic child.
   
   Recommendation: Magnetic resonance imaging [MRI] is imaging method of choice in all cases of suspected osteomyelitis of the spine and pelvis.
   [Evidence level C]
   MRI is very specific but only needed in the minority where diagnostic doubt remains. MRI is especially useful in axial disease of the pelvis and spine. General anaesthesia is often required to obtain diagnostic quality MRI in younger children.
   
   Recommendation: Isotope bone scanning is not recommended for routine investigation of osteomyelitis.
   [Evidence level D]
   Isotope bone scanning has a very limited place in modern management but may help in cases where multifocal disease is suspected [or MRI scanning not available].

Back to top

Resuscitation
Fluid resuscitation as necessary

Analgesia
Pain relief and splintage of adjacent joint in a cast as necessary, according to existing protocols.

Surgery
Surgery has a limited place in the management of acute haematogenous osteomyelitis. Drainage of superiosteal abscesses and contiguous joint collections is essential. Suspicion is raised by gross swelling, persistent spiking temperatures and failure to improve with medical treatment. Ultrasound is an excellent highly specific, easily obtained and well-tolerated examination in such cases. It can also be repeated very easily.
Surgery may also be required for particularly virulent pathogens e.g. PVL-positive staphylococci.

Surgery may also be required to obtain bone biopsies in immunocompromised children, when tuberculosis is suspected, when the clinical and radiological diagnosis is uncertain, or after treatment failure on first line therapy.

Back to top

DO NOT START ANTIMICROBIALS UNTIL BLOOD CULTURES HAVE BEEN TAKEN.

1. Recommended therapy for children and infants age 1 month to 5 years:
   Intravenous Cefuroxime (50–60 mg/kg (max. 1.5 grams) every 6 hours)
   [Evidence level C]
2. Recommended therapy children age 5 years to 16 years:
   Intravenous Flucloxacillin (50 mg/kg (max. 2 grams) every 6 hours.),
   [Evidence level C]
   add Cefotaxime in severe sepsis (50 mg/kg every 6 hours (maximum 12grams per day))
   [Evidence level D]
3. Recommended therapy for children and infants age 1 month to 16 years with a true immediate-type penicillin allergy:
   Intravenous Vancomycin (15 mg/kg every 6 hours (do not exceed an initial maximum of 750mg every 6 hours). Adjust dose according to plasma concentrations, aim for ‘trough’ levels of 10-20mg/L) PLUS
   Intravenous Ciprofloxacin (10 mg/kg (max. 400 mg) every 8 hours)
   [Evidence level D]
4. Recommended therapy children and infants age 1 month to 16 years with a previous MRSA infection or colonisation:
   Intravenous Vancomycin (15 mg/kg every 6 hours (not exceed an initial maximum of 750mg every 6 hours). Adjust dose according to plasma concentrations, aim for ‘trough’ levels of 10-20mg/L) PLUS
   Intravenous Cefuroxime (50–60 mg/kg (max. 1.5 grams) every 6–8 hours).
   [Evidence level D]
   
   There is little high level evidence to direct the choice of appropriate antimicrobial therapy and some data is extrapolated from treatment of adults. Ampicillin-subactam has been compared with Ceftriaxone in a randomised study of treatment for skin, joint and bone infections in children (Kulhanjian et al., 1989) Ceftriaxone had a satisfactory clinical and microbiological response in 93% compared to 100% with Ampicillin-subactam. Notably, Ceftriaxone failed in two cases of Staphylococcus aureus infection and is not therefore recommended as a routine empirical agent. Although aztreonam was considered by the guideline development team to be the most appropriate option in children with a true immediate-type penicillin allergy, current supply issues mean that it is not reliably available. It was agreed that Ciprofloxacin was the most appropriate alternative.
5. Recommendation: Initial antimicrobial therapy for osteomyelitis should be intravenous.
   [Evidence level D]
   
   Reliable IV access is a prerequisite and whilst initial access using a peripheral cannula is adequate if the course is longer than 3-4 days repeat cannulation is almost inevitable and experience shows that in unusual/protracted cases it is better to arrange insertion of a PICC [peripherally inserted central catheter] in theatre at an early stage rather than try and persevere with painful, serially inadequate peripheral cannulae. If a child requires a general anaesthetic for surgical exploration and evacuation of pus a PICC should be inserted at the same time.

Back to top

In culture negative cases directed therapy is not possible but empirically in >90%+ cases monotherapy with Flucloxacillin is appropriate. In a child commenced on Flucloxacillin and Cefotaxime for severe sepsis and suspected meningococcal infection, Cefotaxime can usually be stopped if Blood Cultures do not grow meningococci and meningococcal PCR on blood samples is negative.

Directed antimicrobial regimens, when a pathogen is identified from Blood Cultures or bone samples are given in Table 1.

Table 1: Directed antimicrobial regimens for osteomyelitis in children and infants. Provided isolate is susceptible.

Note: Dose reductions may be required in patients with hepatic or renal impairment, consult product literature or pharmacy for further information. When prescribing high dose oral antibiotics (e.g. Amoxicillin , Flucloxacillin , Cefalexin Clindamycin ) patients may experience an increase in gastrointestinal upset (nausea and diarrhoea)

Back to top

Surprisingly little scientific evidence exists on the length of the treatment course. Animal model work has demonstrated that treatment of osteomyelitis with Clindamycin is more effective at three weeks than after one week of treatment (Mayberry-Carson et al., 1986). Treatment failures are more frequent in adults following shorter courses of treatment e.g. less than 4-6 weeks but equivalent data for children is not available.

The standard duration of therapy is usually 6 weeks though this seems to be largely on the basis of common “custom and practice” and not on validated human studies. It has been suggested that if the infection is diagnosed and treated during the initial bone “cellulitic” phase before bone destruction occurs [as manifest by a bone defect on X-ray] then a three week course of antibiotics will suffice. We have little experience of this management and have concerns about the validity of a diagnosis of osteomyelitis without radiological evidence of bony involvement. We therefore prefer to adhere to the time honoured six week course. On the whole a long course of high dose antibiotics is reasonably well-tolerated by children.

Back to top

Initial IV treatment in acute haematogenous osteomyelitis is mandatory. If the clinical response is good [usual case] as indicated by the child’s general well-being [feeding etc.], rapidly resolving temperature and rapidly falling CRP [preferably<20] then oral antibiotics can be substituted.

In such rapid responders there is evidence that long IV courses are unnecessary (Zaoutis et al., 2009). Our experience is that a minimum of 3-4 days IV treatment is needed and in favourable cases oral treatment can then be substituted.

Hospital discharge should be at least 24 hours after switch to confirm continuing improvement. On discharge from hospital patients will be supplied 14 days of antibiotics and GP will provide the remainder of the course.

Back to top

Slow/non-responders: In patients that failure to respond by 96 hours, the causative pathogen may not be covered by empirical regimens or abscess formation/adjacent joint sepsis may be compromising antimicrobial activity. Surgical drainage may be required.

Recommendation: Discuss with senior orthopaedic surgeon and microbiology.

Back to top