Osteomyelitis can be categorized as acute hematogenous (AHO) , subacute (SO) in those patients presenting after 2 weeks of symptoms, or chronic (CO) in those patients with symptoms for greater than 3 months (Craigen et al. 1992; Beckles et al. 2010). The direct inoculation of bone, as occurs with open fractures, as well as hematogenous spread to an area of closed bony injury, is a well-recognized mechanism for developing an infection, and many authors believe that AHO is in many cases related to minor trauma or subtle fractures. Like septic arthritis, AHO has an increased incidence in young children, with 40 % of cases in one large study present in those children less than 3 years of age, with a gender ratio of nearly 2:1 of boys to girls affected (Goergens et al. 2005; Fig. 1).

The majority of cases of osteomyelitis in the industrialized world are AHO and SO , and the pathoanatomy is closely related to the vascular anatomy of the pediatric bone. In the metaphyseal region of the long bones of children, the nutrient artery perforates the metaphysis and terminates near the physis in a series of dilated, low-flow venous sinusoids, which, after making a sharp turn, become a venous capillary network in the medullary bone. The numerous dilated metaphyseal venous capillaries have been described as “blood lakes.” The low-flow state of this anatomical region leads to the increased susceptibility of bacterial infection transferred from the blood, as the organism “finds its ideal medium for development (Trueta 1959).” As the infection is initiated, thrombosis of the venous system is the initial vascular event, followed by thrombosis of the nutrient artery. Both thrombotic events further lead to an ideal milieu for the growth of bacteria in the pediatric metaphysis.

The vascular congestion and edema that occurs can lead to cortical breakthrough and elevation of the periosteum, which may then be followed by purulence in a similar tract down the diaphysis (Trueta 1959). The periosteal reaction caused by the infection then forms and can typically be seen on plain radiographs within several days. Over time, the new bone formation can lead to the formation of an involucrum, seen in chronic osteomyelitis. A sequestrum may also develop if avascular bone becomes walled off from the surrounding bone. The epiphysis garners its blood supply from a separate artery, and this supports the postulated vascular etiology of disease, in that the epiphysis and metaphysis are rarely both involved with AHO. Therefore, the physis and epiphysis are classically protected from the majority of osteomyelitis infectious processes.

In cases of osteomyelitis following an open fracture, the bone is directly inoculated, and the site of infection and subsequent development of infection correspond to the fracture site which is influenced by varying degrees of avascularity of the affected bone due to trauma. For closed fractures that are seeded by the bloodstream and develop osteomyelitis, a similar mechanism involving localized slowing of blood flow and/or formation of hematoma in the region is postulated to lead to infection via an analogous process to atraumatic metaphyseal osteomyelitis.

Children with acute and subacute osteomyelitis of the upper extremity classically present with fever, pain, and dysfunction of the affected extremity, but may exhibit a combination of only some or all of these findings. Localization of the affected region is typically possible by a thorough physical exam, as the patient can be tender to palpation, may exhibit swelling and/or erythema in the adjacent soft tissues, and/or may demonstrate pain with range of motion and/or weight bearing. At times, in particular with early acute hematogenous osteomyelitis, these findings may be subtle, without any visible changes in the surrounding soft tissues. A high index of suspicion is needed in particular in those children who are too young to adequately communicate with the physician, which is the demographic most commonly affected. This can be especially true in osteomyelitis of the upper extremity, as many times parents or caregivers note a reluctance to bear weight when the lower extremity is affected as the initial sign of pathology. This sign may be absent or delayed in the recognition of pathology in the upper extremity.

While always in the differential of musculoskeletal pain in children, the presence of osteomyelitis may still be missed by those treating physicians commonly involved prior to orthopedic consultation. Treatment for presumed cellulitis or an abscess without typical resolution can last for days before the true diagnosis of osteomyelitis is recognized and can compromise the outcome.

The initial imaging modality for suspected osteomyelitis is high-quality plain radiographs, and in addition to the bony findings, soft tissue findings should be assessed for. Evidence of a joint effusion and/or soft tissue swelling with acute osteomyelitis is important, in particular, as bony changes may not yet have occurred. When present, osseous findings can include bony destruction and/or periosteal reaction (Fig. 2).

MRI with and without contrast has emerged as the most sensitive and specific imaging modality in the diagnosis of osteomyelitis in children and has largely supplanted bone scintigraphy. In 1995, Mazur et al. reported a sensitivity of 0.97 and specificity of 0.92 with contrast-enhanced MRI (Mazur et al. 1995). In a subset of their cohort, bone scintigraphy was also obtained, with a sensitivity of 0.64 and specificity of 0.71. MRI also can aid in surgical planning, as it more precisely defines the anatomical extent of the infection (Fig. 3).

Bone scintigraphy may still be advantageous in suspected cases of osteomyelitis where the physical exam does not localize the process or in cases where multifocal osteomyelitis is suspected. In addition, ultrasound has been reported to be a viable mechanism to detect osteomyelitis prior to any plain radiographic findings being present (Riebel et al. 1996; Mah et al. 1994). However, where MRI is readily available, it has largely supplanted the use of both ultrasound and bone scintigraphy due to the accuracy and anatomic detail it affords. In addition, it is one of the first pieces of information that guides treatment.

Laboratory studies that should be routinely obtained during the initial evaluation of osteomyelitis include complete blood cell count with differential (CBC), erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP). An essential aspect in the interpretation of the laboratory studies is the consistent finding that the CBC can be normal in up to 80 % of children with osteomyelitis, and therefore the laboratory workup should be focused on the ESR and CRP (Pääkkönen et al. 2009). Multiple reports have noted the increase in both ESR and CRP in a very high percentage of children with acute osteomyelitis upon admission to the hospital (Pääkkönen et al. 2009; Khachatourians et al. 2003; Roine et al. 1995). In 1995, Unkila-Kallio et al. reported elevations of ESR and CRP upon admission in bacteriologically confirmed AHO in 92 % and 98 % of children, respectively. The mean values of the ESR and CRP were 45 mm/h and 71 mg/L, respectively. Similar rates of sensitivity and specificity have been reported by other authors more recently, with rates nearing 100 % when ESR and CRP are analyzed together (Pääkkönen et al. 2009). The CRP and ESR can therefore be very useful to the clinician for several reasons, including ruling out osteomyelitis, tracking the response to treatment, and making the decision to stop antibiotic therapy (Fig. 4).

Blood cultures should also be obtained, and, when possible, a tissue sample through aspiration should be obtained prior to the administration of antibiotics in order to facilitate accurate treatment of the causative organism. While these tests are standard in the evaluation of pediatric osteomyelitis, it is important to note that a negative culture result from the blood, aspiration, or direct biopsy is not uncommon. Arnold et al. noted a 49 % positive culture result (from blood, bone, or joint aspirate) in their series of 89 patients with osteomyelitis, which is similar to the report from Georgens et al. where they noted a rate of 45 % (Goergens et al. 2005; Arnold et al. 2006). For complex osteoarticular infections, however, the rate of positive blood cultures was 82 % (Goergens et al. 2005; Arnold et al. 2006). Persistently positive blood cultures (3 or more days) have been associated with MRSA infection (Arnold et al. 2006).

The most common causative pathogens in the postvaccination era are overwhelmingly varieties of Staph aureus (MSSA, MRSA, and coagulase negative), but other pathogens including Kingella kingae, Salmonella group D, Enterobacter, and Pseudomonas species have all been reported (Copley et al. 2013).

Kingella osteoarticular infection and its detection deserve a special note as an emerging pathogen in pediatric osteoarticular infections. In children less than 4 years of age, osteoarticular infection with Kingella has increasingly been recognized as an important and frequent pathogen – possibly the most common bacteria responsible for the infection in this age group. The clinical presentation and laboratory investigations can be mild and normal, respectively. Furthermore, Kingella has historically been a difficult pathogen to culture on solid medium, even when suspected. Newer polymerase chain reaction (PCR) techniques have been developed to allow for more accurate identification of this pathogen and have been applied clinically at some centers. In 2010, Ceroni et al. reported that 82.1 % of cases of pediatric osteoarticular infections in children less than 4 years of age were caused by Kingella, as determined by PCR identification (Ceroni et al. 2010). In each of these cases, traditional methods of staining and cultures were all negative. The authors’ findings led them to state, “Kingella kingae is an emerging pathogen that may be recognized as the most common bacteria responsible for osteoarticular infections (OAI) in young children.” Laboratory and specimen collection protocols that routinely attempt identification of this fastidious and common organism may alter the current treatment algorithms at many institutions (Yagupsky et al. 2011).

While rare in comparison to hematogenous osteomyelitis, direct inoculation of bone occurring as a result of open fractures, soft tissue trauma, or surgical interventions can lead to the development of osteomyelitis. In addition, closed fractures or minor injuries to the upper extremity may predispose the affected bone to the vascular stasis that underlies the development of spontaneous hematogenous osteomyelitis. Appropriate urgent management of open fractures with debridement and irrigation, parenteral antibiotics, and modern surgical care has led to a decreased incidence of osteomyelitis associated with open fractures in children.

A special note should also be made about chronic recurrent multifocal osteomyelitis (CRMO) – a rare, noninfectious condition characterized by the presence of bony lytic lesions that can affect one to multiple locations around the body. In isolation, lesions may mimic AHO in regard to their presentation (pain, swelling, redness), but typically multiple sites are affected. The differentiation of CRMO from AHO early in the course is important, as the recommended treatments for each disorder are significantly different.