Etiology

Bacteria are the most common pathogens in acute skeletal infections. In osteomyelitis, Staphylococcus aureus (Chapter 174.1) is the most common infecting organism in all age groups, including newborns. Community-acquired methicillin-resistant S. aureus (CA-MRSA) isolates account for >50% of S. aureus isolates recovered from children with osteomyelitis in some reports. The USA300 clone of S. aureus is the most common among CA-MRSA isolates in the USA and is more likely to cause venous thrombosis in children with acute osteomyelitis than other S. aureus clones or other bacteria for reasons that are not known.

Group B streptococcus (Chapter 177) and gram-negative enteric bacilli (Escherichia coli, Chapter 192) are also prominent pathogens in neonates; group A streptococcus (Chapter 176) constitutes <10% of all cases. After 6 yr of age, most cases of osteomyelitis are caused by S. aureus, streptococcus, or Pseudomonas aeruginosa (Chapter 197). Cases of Pseudomonas infection are related almost exclusively to puncture wounds of the foot, with direct inoculation of P. aeruginosa from the foam padding of the shoe into bone or cartilage, which develops as osteochondritis. Salmonella species (Chapter 190) and S. aureus are the two most common causes of osteomyelitis in children with sickle cell anemia. S. pneumoniae (Chapter 175) most commonly causes osteomyelitis in children <24 mo of age or children with sickle cell anemia. Bartonella henselae (Chapter 201.2) can cause osteomyelitis of any bone but especially in pelvic and vertebral bones.

Kingella kingae may be the second most common cause of osteomyelitis in children <5 yr of age in some parts of the world. K. kingae is a slow-growing, gram-negative, β-hemolytic bacterium in pairs or short chains of short bacilli. The organism, once thought to be rare, is increasingly recognized as a cause of osteomyelitis, spondylodiskitis, septic arthritis and bacteremia, and, less commonly, in endocarditis. It has been identified as the causative agent in pneumonia and meningitis. Nearly 90% of identified K. kingae infections have been in young children.

Infection with atypical mycobacteria (Chapter 209), S. aureus, or Pseudomonas can occur after penetrating injuries. Fungal infections usually occur as part of multisystem disseminated disease; Candida (Chapter 226) osteomyelitis sometimes complicates fungemia in neonates with or without indwelling vascular catheters.

A microbial etiology is confirmed in ∼60% of cases of osteomyelitis. Blood cultures are positive in ∼50% of patients. Prior antibiotic therapy and the inhibitory effect of pus on microbial growth might explain the low bacterial yield.

Epidemiology

The median age of children with musculoskeletal infections is ∼6 yr. The incidence of osteomyelitis in children is estimated to be 1 : 5,000. Bone infections are more common in boys than girls; the behavior of boys might predispose them to traumatic events. Except for the increased incidence of skeletal infection in patients with sickle cell disease, there is no predilection for osteomyelitis based on race.

The majority of osteomyelitis cases in previously healthy children are hematogenous. Minor closed trauma is a common preceding event in cases of osteomyelitis, occurring in ∼30% of patients. Infection of bones can follow penetrating injuries or open fractures. Bone infection following orthopedic surgery is uncommon. Impaired host defenses also increase the risk of skeletal infection. Other risk factors are noted in Table 676-1.

Table 676-1 MICROORGANISMS ISOLATED FROM PATIENTS WITH OSTEOMYELITIS AND THEIR CLINICAL ASSOCIATIONS

Modified From Lew DP, Waldvogel FA: Osteomyelitis, Lancet 364:369–379, 2004.

Pathogenesis

The unique anatomy and circulation of the ends of long bones result in the predilection for localization of bloodborne bacteria. In the metaphysis, nutrient arteries branch into nonanastomosing capillaries under the physis, which make a sharp loop before entering venous sinusoids draining into the marrow. Blood flow in this area is thought to be “sluggish,” predisposing to bacterial invasion. Once a bacterial focus is established, phagocytes migrate to the site and produce an inflammatory exudate (metaphyseal abscess). The generation of proteolytic enzymes, toxic oxygen radicals, and cytokines results in decreased oxygen tension, decreased pH, osteolysis, and tissue destruction. As the inflammatory exudate progresses, pressure increases spread through the porous metaphyseal space via the haversian system and Volkmann canals into the subperiosteal space. Purulence beneath the periosteum may lift the periosteal membrane of the bony surface, further impairing blood supply to the cortex and metaphysis.

In newborns and young infants, transphyseal blood vessels connect the metaphysis and epiphysis, so it is common for pus from the metaphysis to enter the joint space. This extension through the physis has the potential to result in abnormal growth and bone or joint deformity. During the latter part of the 1st year of life, the physis forms, obliterating the transphyseal blood vessels. Joint involvement, once the physis forms, can occur in joints where the metaphysis is intra-articular (hip, ankle, shoulder, and elbow), and subperiosteal pus ruptures into the joint space.

In later childhood, the periosteum becomes more adherent, favoring pus to decompress through the periosteum. Once the growth plate closes in late adolescence, hematogenous osteomyelitis more often begins in the diaphysis and can spread to the entire intramedullary canal.

Clinical Manifestations

The earliest signs and symptoms of osteomyelitis, often subtle and nonspecific, are generally highly dependent on the age of the patient. Neonates might exhibit pseudoparalysis or pain with movement of the affected extremity (e.g., diaper changes). Half of neonates do not have fever and might not appear ill. Older infants and children are more likely to have fever, pain, and localizing signs such as edema, erythema, and warmth. With involvement of the lower extremities, limp or refusal to walk is seen in approximately half of patients.

Focal tenderness over a long bone can be an important finding. Local swelling and redness with osteomyelitis can mean that the infection has spread out of the metaphysis into the subperiosteal space, representing a secondary soft-tissue inflammatory response. Pelvic osteomyelitis can manifest with subtle findings such as hip, thigh, or abdominal pain. Back pain with or without tenderness to palpation overlying the vertebral processes is noted in vertebral osteomyelitis.

Long bones are principally involved in osteomyelitis (Table 676-2); the femur and tibia are equally affected and together constitute almost half of all cases. The bones of the upper extremities account for one fourth of all cases. Flat bones are less commonly affected.

Table 676-2 SITES OF OSTEOMYELITIS IN CHILDREN

Modified from Gafur OA, Copley LA, Hollmig ST, et al: The impact of the current epidemiology of pediatric musculoskeletal infection on evaluation and treatment guidelines, J Pediatr Orthop 28(7):777–785, 2008.

There is usually only a single site of bone or joint involvement. Several bones are infected in <10% of cases; the exception is osteomyelitis in neonates, in whom two or more bones are involved in almost half of the cases. Children with subacute symptoms and focal finding in the metaphyseal area (usually of tibia) might have a Brodie abscess, with radiographic lucency and surrounding reactive bone.

Diagnosis

The diagnosis of osteomyelitis is clinical; blood cultures should be performed in all suspected cases. Depending on the results of imaging studies (see later) aspiration or biopsy of bone or subperiosteal abscess for Gram stain, culture, and possibly bone histology provides the optimal specimen for culture to confirm the diagnosis. These specimens are often obtained by the interventional radiologist or at the time of surgical drainage by the orthopedic surgeon. Direct inoculation of clinical specimens into aerobic blood culture bottles can improve the recovery of K. kingae, particularly if held for 1 wk. Polymerase chain reaction appears to be the most sensitive technique to detect K. kingae, with detection up to 6 days after antibiotics are initiated.

There are no specific laboratory tests for osteomyelitis. The white blood cell count and differential, erythrocyte sedimentation rate (ESR), or C-reactive protein (CRP) are generally elevated in children with bone infections but are nonspecific and not helpful in distinguishing between skeletal infection and other inflammatory processes. The leukocyte count and ESR may be normal during the first few days of infection, and normal test results do not preclude the diagnosis of skeletal infection. However, most children with acute hematogenous osteomyelitis have elevations in the ESR and/or CRP. Monitoring elevated ESR and CRP may be of value in assessing response to therapy or identifying complications.

Radiographic Evaluation

Radiographic studies play a crucial role in the evaluation of osteomyelitis. Conventional radiographs, ultrasonography, CT, MRI, and radionuclide studies can all contribute to establishing the diagnosis. Plain radiographs are often used for initial evaluation to exclude other causes such as trauma and foreign bodies. MRI has emerged as the most sensitive and specific test and is widely used for diagnosis. The sequence of radionuclide studies or MRI is often determined by age, site, and clinical presentation.

Computed Tomography and Magnetic Resonance Imaging

CT can demonstrate osseous and soft-tissue abnormalities and is ideal for detecting gas in soft tissues. In selected children who cannot remain still or tolerate sedation, CT is a valuable imaging modality. MRI is more sensitive than CT or radionuclide imaging in acute osteomyelitis and is the best radiographic imaging technique for identifying abscesses and for differentiating between bone and soft-tissue infection. MRI provides precise anatomic detail of subperiosteal pus and accumulation of purulent debris in the bone marrow and metaphyses for possible surgical intervention. In acute osteomyelitis, purulent debris and edema appear dark, with decreased signal intensity on T1-weighted images, with fat appearing bright (Fig. 676-1). The opposite is seen in T2-weighted images. The signal from fat can be diminished with fat-suppression techniques to enhance visualization. Gadolinium administration can also enhance MRI. Cellulitis and sinus tracts appear as areas of high signal intensity on T2-weighted images. MRI can also demonstrate a contiguous septic arthritis, pyomyositis, or venous thrombosis.

Figure 676-1 MRI of an 8 yr old girl with acute pelvic hematogenous osteomyelitis. A, Axial T1-weighted contrast-enhanced MRI with fat saturation reveals a nonenhancing fluid collection adjacent to the inflamed pubic synchondrosis. B, The fluid collection appears hyperintense on the corresponding T2-weighted image (arrowheads). In addition, a contrast enhancement within the adjacent internal obturator muscle is seen (arrow), indicating pelvic acute pelvic hematogenous osteomyelitis with complicating adjacent abscess formation and soft tissue inflammation.

(From Weber-Chrysochoou C, Corti N, Goetschel P, et al: Pelvic osteomyelitis: a diagnostic challenge in children, J Pediatr Surg 42:553–557, 2007.)

Differential Diagnosis

Distinguishing osteomyelitis from cellulitis or trauma (accidental or abuse) is the most common clinical circumstance. Myositis or pyomyositis can also appear similar to osteomyelitis with fever, warm and swollen extremities, and limping; tenderness to palpation of the affected soft tissue area is generally more diffuse than noted in acute osteomyelitis. Nevertheless, distinguishing myositis and pyomyositis from osteomyelitis clinically may be difficult. Myositis and pyomyositis often are found adjacent to an osteomyelitis on MRI. Pyomyositis is often caused by S. aureus or group A streptococcus. The pelvic muscles are a common site of infection and can mimic a pelvic osteomyelitis. MRI is the definitive study to identify and localize pelvic pyomyositis. An iliopsoas abscess can manifest with thigh pain, limp, and fever and must be considered in the differential diagnosis of osteomyelitis. The iliopsoas abscess may be primary (hematogenous: S. aureus) or secondary to infection in adjacent bone (S. aureus), kidney (E. coli) or intestine (E. coli, Bacteroides spp.). M. tuberculosis has been reported in patients with HIV infection.

Appendicitis, urinary tract infection, and gynecologic disease are among the conditions in the differential diagnosis of pelvic osteomyelitis. Children with leukemia commonly have bone pain or joint pain as an early symptom. Neuroblastoma with bone involvement may be mistaken for osteomyelitis. Primary bone tumors need to be considered, but fever and other signs of illness are generally absent except in Ewing sarcoma. In patients with sickle cell disease, distinguishing bone infection from infarction may be challenging. Chronic recurrent multifocal osteomyelitis (CRMO) and synovitis, acne, pustulosis, hyperostosis, and osteitis syndrome are rare noninfectious conditions in children characterized by recurrent osteoarticular inflammation and different skin conditions, palmoplantar pustulosis, psoriasis, severe acne, neutrophilic dermatosis (Sweet syndrome, Chapter 163), and pyoderma gangrenosum.

Treatment

Optimal treatment of skeletal infections requires collaborative efforts of pediatricians, orthopedic surgeons, and radiologists. Obtaining material for culture (blood, periosteal abscess, bone) before antibiotics are given is essential. Because most patients with osteomyelitis have an indolent, non–life-threatening condition, cultures should be obtained even if there is a delay of a few hours in initiating antibiotics.

Prognosis

When pus is drained and appropriate antibiotic therapy is given, the improvement in signs and symptoms is rapid. Failure to improve or worsening by 72 hr requires review of the appropriateness of the antibiotic therapy, the need for surgical intervention, or the correctness of the diagnosis. Acute-phase reactants may be useful as monitors. The serum CRP typically normalizes within 7 days after start of treatment, whereas the ESR typically rises for 5-7 days, and then falls slowly, dropping sharply after 10-14 days. Failure of either of these acute-phase reactants to follow the usual course should raise concerns about the adequacy of therapy. Recurrence of disease and development of chronic infection after treatment occur in <10% of patients.

Because children are in a dynamic state of growth, sequelae of skeletal infections might not become apparent for months or years; therefore, long-term follow-up is necessary with close attention to range of motion of joints and bone length. Although firm data about the impact of delayed treatment on outcome are not available, it appears that initiation of medical and surgical therapy within 1 wk of onset of symptoms provides a better prognosis than delayed treatment.