Patient-Dependent Factors

Surgeon-Dependent Factors

Laboratory Studies

A complete blood cell count, including differential and erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP), should be obtained during initial evaluation of bone and joint infections. The white blood cell count is an unreliable indicator of infection and often is normal even when infection is present. The differential shows increases in neutrophils during acute infections. The ESR becomes elevated when infection is present, but this does not occur exclusively in the presence of infection. Fractures or other underlying diseases can cause elevation of the ESR. The ESR also is unreliable in neonates, patients with sickle cell disease, patients taking corticosteroids, and patients whose symptoms have been present for less than 48 hours. Peak elevation of the ESR occurs at 3 to 5 days after infection and returns to normal approximately 3 weeks after treatment is begun. CRP, synthesized by the liver in response to infection, is a better way to follow the response of infection to treatment. CRP increases within 6 hours of infection, reaches a peak elevation 2 days after infection, and returns to normal within 1 week after adequate treatment has begun. Other tests, such as the S. aureus surface antigen or antibody test and counterimmunofluorescence studies of the urine, are promising, but their usefulness in clinical situations has not been proved. Material obtained from aspiration of joint fluid can be sent to the laboratory for a cell count and differential to distinguish acute septic arthritis from other causes of arthritis. In septic arthritis, the cell count usually is greater than 80,000/mm³, with more than 75% of the cells being neutrophils (Table 20-3). Jung et al. devised an algorithm to predict the probability of septic arthritis in children (Table 20-4). A Gram stain also should be obtained. Gram stains identify the types of organisms (gram-positive or gram-negative) in about a third of bone and joint aspirates. Intraoperative frozen section also should be obtained in cases in which infection is suspected. A white blood cell count greater than 10 per high-power field is considered indicative of infection, whereas a count less than 5 per high-power field all but excludes infection.

TABLE 20-3 Synovial Fluid Analysis

From Morrissy RT: Septic arthritis. In Gustilo RB, Genninger RP, Tsukayama DT, editors: Orthopaedic infection: diagnosis and treatment, Philadelphia, 1989, WB Saunders.

Imaging Studies

Radiographic studies are helpful but are not as useful in the diagnosis of acute bone and joint infections as they are in following responses to treatment. Plain radiographs show soft tissue swelling, joint space narrowing or widening, and bone destruction (Fig. 20-1). Bone destruction is not apparent on radiographs, however, until an infection has been present for 10 to 21 days. In addition, 30% to 50% of the bone matrix must be lost to show a lytic lesion on radiographs (Fig. 20-2). Wheat found that fewer than 5% of plain radiographs were initially abnormal in bone and joint infections, and fewer than 30% were abnormal at 1 week; however, 90% were abnormal at 3 to 4 weeks. If initial radiographs are normal in the evaluation of bone and joint infections, other imaging methods that show soft tissue swelling and loss of normal fat planes around the involved bone or joint should be used.

FIGURE 20-1 A, Radiographic evidence of acute bone and joint infection. B, Treatment with antibiotic cement spacer.

(Courtesy of Andrew Crenshaw.)

FIGURE 20-2 Anteroposterior (A) and lateral (B) radiographs showing bone destruction.

Conventional tomography can be useful in identifying a sequestrum or subchondral bony plate destruction, although it has largely been replaced with more conventional radiographic methods. Arthrography helps document proper aspiration of a suspected septic joint. Dye should be injected only after fluid is obtained from the joint because the bactericidal effect of iodinated contrast material can cause a false-negative culture result. CT can help determine the extent of medullary involvement. Pus within the medullary cavity replaces the marrow fat, causing an increased density on the CT scan. Adjacent soft tissue abscesses also are seen easily (Fig. 20-3). CT diagnosis of acute osteomyelitis is based on detection of intraosseous gas, osteolysis, soft tissue masses, abscesses, or foreign bodies. Additionally, increased vascularity after administration of a contrast agent also can aid in the diagnosis. Narrowing of the medullary cavity by granulation tissue and new bone is readily shown during the healing phase of osteomyelitis. CT identifies sequestra in chronic osteomyelitis (Fig. 20-4). It also is helpful in identifying alterations in areas poorly seen on plain films, such as the sternoclavicular joint, sacroiliac joint, and spine. Contrast material can be used to delineate abscesses in necrotic tissue that does not enhance from surrounding hyperemic tissue.

FIGURE 20-3 CT of pelvic abscess in a child.

FIGURE 20-4 A and B, CT scans showing sequestra.

(Courtesy of Todd Williams.)

Ultrasonography can also be used to localize an abscess cavity, detect joint effusion, or guide a physician in the proper placement of the needle when obtaining aspirate from a bone or joint.

Radionuclide scanning has become a useful imaging adjunct in the diagnosis of osteomyelitis. Although radiography and CT give a structural or anatomical picture, radionuclide scanning gives a more physiological picture. Bone scintigraphy does not detect the presence of infection but, instead, reflects inflammatory changes or the reaction of bone to the infection. Radionuclide scanning also is useful in patients with metallic implants in whom CT and MRI are of limited value because of contraindications and metallic-generated artifact, although metal subtracting software is improving imaging in these patients. The three most commonly used radioisotopes are technetium-99m (^99mTc) phosphate, gallium-67 (⁶⁷Ga) citrate, and indium-111 (¹¹¹In)–labeled leukocytes. The most common is ^99mTc phosphate, which can detect osteomyelitis within 48 hours after clinical onset of infection. The uptake of this compound is related primarily to osteoblastic activity, although regional blood flow also plays a role in skeletal uptake. After intravenous injection, the technetium is distributed rapidly throughout the extracellular compartment. Bone uptake is rapid, with more than 50% of the administered dose being delivered to bone within 1 hour. The remainder of the dye is excreted by the kidneys into the urine.

The standard technique of ^99mTc phosphate imaging is to perform a three-phase study. Although this does not increase the sensitivity of the test significantly, it does increase specificity from 74% to 94%. The three-phase bone scan consists of images taken in (1) the flow phase, (2) the immediate or equilibrium phase, and (3) the delayed phase. The flow-phase image is similar to a radionuclide angiogram in that it shows blood flow. The equilibrium or blood pool image shows relative vascular flow and distribution of the radioisotope into the extracellular space. The delayed-phase image generally is obtained 2 to 4 hours after injection when renal excretion has eliminated most of the isotope except that taken up by osteoblastic activity. This image shows osteoblastic activity and is positive in numerous disease states, including osteomyelitis, tumors, degenerative joint disease, trauma, and postsurgical changes. Usually a focus of osteomyelitis appears as an area of increased tracer uptake on delayed images. To have a “hot spot” on a bone scan, the vasculature to the involved bone must be intact. If blood flow to the involved area is decreased by subperiosteal pus, necrosis (i.e., sequestrum), joint effusion, vasospasm, or soft tissue swelling, a “cold” scan may result.

A major disadvantage of three-phase ^99mTc phosphate bone scintigraphy is that the increased uptake caused by osteomyelitis is difficult to distinguish from that caused by degenerative joint disease or posttraumatic or postsurgical changes. The relative activity in each of the three phases may be helpful in differentiating other causes of increased uptake. Cellulitis causes increased activity during the flow and equilibrium phases and a decreased or normal uptake in the delayed phase. Osteomyelitis causes increased uptake in all three phases (Fig. 20-5). Increased uptake in the delayed phase but not in the flow or equilibrium phase suggests degenerative joint disease (Table 20-5). ^99mTc phosphate bone scans are unreliable in neonates (<6 weeks old) and usually are negative in 60% of these patients with bone or joint infections.

FIGURE 20-5 Three-phase bone scan showing osteomyelitis.

(Courtesy of Michael Fleming.)

⁶⁷Ga citrate is the oldest tracer and has been used to localize inflammatory lesions as well as malignant tissue. The mechanism of gallium deposition is controversial; it seems to be related to increased endothelial permeability or diffusion by transportation as gallium-transferrin. The specificity of a ⁶⁷Ga citrate scan alone is poor (82%). ⁶⁷Ga citrate scanning can be useful in osteomyelitis when it is used in conjunction with ^99mTc phosphate scanning. In purely reactive bone formation (posttraumatic or postsurgical), the intensity on the ^99mTc phosphate scan is proportionally greater than that on the ⁶⁷Ga citrate scan. In areas of inflammation, however, gallium uptake either exceeds that of technetium in relative magnitude or displays a different spatial configuration of activity. A disadvantage of ⁶⁷Ga citrate imaging is its slow clearance after injection, which requires a delay in imaging ranging from 24 hours after injection for the appendicular skeleton to 72 hours for the axial skeleton. Specificity decreases in ⁶⁷Ga citrate scintigraphy when the lesion is located peripherally rather than centrally. With the combination of ⁶⁷Ga citrate and ^99mTc phosphate scans, sensitivity and specificity are 70% and up to 93%, respectively, for the detection of osteomyelitis.

¹¹¹In–labeled leukocytes have been suggested for differentiating between osteomyelitis and reactive bone formation. This scan is positive at earlier stages of osteomyelitis than ^99mTc phosphate scintigraphy. The leukocyte scanning technique involves in vitro radionuclide labeling and injection of autologous leukocytes, predominantly polymorphonuclear neutrophilic leukocytes, followed by imaging 24 to 48 hours later. Fifty milliliters of the patient’s venous blood is obtained, separated from the other blood elements in vitro, and labeled with ¹¹¹In. The labeled leukocytes are reinjected into the patient, and scans are obtained at 24 hours. A scan is positive if focal accumulation of activity exceeds adjacent normal bone activity. Scintigraphy with ¹¹¹In has been reported to be helpful in the diagnosis of acute osteomyelitis, but there is disagreement about its efficacy in chronic osteomyelitis because the latter is predominantly lymphocytic and may give a negative or “cold” scan. The ¹¹¹In scan also is unreliable for differentiating between aseptic and septic loosening of a painful arthroplasty. Teller et al. did not recommend the routine use of ^99mTc phosphate with ¹¹¹In-labeled scans for detecting aseptic or septic loosening of arthroplasty because of the high cost associated with these tests and the low specificity and sensitivity of 78% and 64%, respectively. Prandini et al., in a meta-analysis, found that ^99mTc-labeled white blood cells had a greater sensitivity (89%) and specificity (90.1%) than ¹¹¹In-labeled white blood cells. Several authors have recommended prolonging the scan time until 24 hours for the ^99mTc-labeled white blood cell scan to improve detection. ¹¹¹In-labeled monoclonal immunoglobulin is a substitute for ¹¹¹In-labeled leukocytes. It seems to be as effective as ¹¹¹In-labeled leukocytes, does not require phlebotomy, and avoids the risk of radiation to white blood cells and the perceived risk of malignant transformation. According to Hakki et al., compared with ¹¹¹In-labeled leukocytes and ^99mTc phosphate scintigraphy, monoclonal antibody fragment (LeukoScan, Granuloscint, NeutoSpect) has better sensitivity, specificity, and diagnostic accuracy. In addition, these researchers suggested that LeukoScan is a stronger diagnostic tool in patients with a low leukocyte count (i.e., human immunodeficiency virus [HIV]–infected patients) and in patients with chronic osteomyelitis. However, meta-analysis studies show that these agents are less accurate than in vitro–labeled white blood cells in most patients, and a risk of allergic reaction (some fatal) does exist especially when repeated scans are necessary. These agents have limited availability in the United States; however, the quest for a more perfect monoclonal antibody fragment continues. The detection of chronic osteomyelitis, especially of the central skeleton, can be enhanced with a ^99mTc/ciprofloxacin (infection) scan and fluorine-18 (¹⁸F)-fluorodeoxyglucose-labeled positron emission tomography (FDG-PET). FDG-PET is the most accurate test (92%) with the most positive predictive value (94%). It is extremely useful for chronic infections and infections already treated with antibiotics. However, it is the most expensive and is not readily available in all health care centers.