Infection: Diagnosis and Treatment Options

CHAPTER PREVIEW

CHAPTER SYNOPSIS

Early diagnosis of infection is the most important factor that affects prognosis. Treatment depends on the type of infection, isolated organism, extent of infection, and overall medical condition of the patient. Aggressive early medical and surgical management is paramount.

IMPORTANT POINTS

1

Early diagnosis and early initiation of treatment can greatly affect prognosis.
2

Diagnosis is made based on history and physical examination, laboratory tests, imaging studies, and intraoperative assessment.
3

Infections are classified into acute postoperative, late chronic, and acute hematogenous.
4

Acute postoperative infections may be treated by irrigation and debridement, intravenous antibiotics, and retention of prosthesis.
5

Late infections are best treated by irrigation and debridement, removal of prosthetic component, and two-stage revision reimplantation using a temporary antibiotic-impregnated cement spacer.

CLINICAL/SURGICAL PEARLS

1

Infection should be suspected with increased pain after shoulder arthroplasty.
2

Early intervention is essential to successful management of infected shoulder arthroplasty.
3

Two-stage revision arthroplasty provides the best results for managing chronic infections.
4

A minimum of three cultures should be obtained during debridement, including superficial and deep cultures, and incubated for at least 7 days.
5

At least one culture from the intermedullary canal should be obtained after removal of the prosthesis.
6

Propionibacterium acnes is a difficult organism to culture and can manifest in subacute infection.
7

A temporary antibiotic-impregnated cement spacer fashioned to fit the intermedullary canal should be used.
8

The cement spacer humeral head should be oversized by 25% to 50% in order to make up for deficiencies in the rotator cuff.

CLINICAL/SURGICAL PITFALLS

1

Negative aspiration results do not rule out infection.
2

Arthroscopic debridement has low success rates and should not be used.
3

Single-stage direct exchange procedures should not be used if virulent infective organism is suspected.

VIDEO AVAILABLE

Not applicable.

More than 1 million total joint replacements are performed annually worldwide; of these only a small percentage are shoulder arthroplasties. However, the number of total shoulder arthroplasties performed in the United States each year has increased substantially from fewer than 5000 in the early 1990s to more than 20,000 in the year 2000. As the population ages and new adaptable implants evolve, orthopedic surgeons can expect to see an increase in the number of shoulder arthroplasties performed, as well as their complications. Although sepsis is the second most common complication, it has proven to be the most devastating for both patient and surgeon alike.

Infection rates of total joint replacements have been reduced from between 5% and 10% in the late 1960s to a mere 1% presently. Infection after total shoulder arthroplasty may be due to problems with poor soft tissue healing and the difficulty of eradication of bacteria introduced during surgery or by hematogenous seeding long after surgery. The infection rate after shoulder arthroplasty is reported to fall between 0% and 3.9%. Sperling et al. reported on a large series from the Mayo Clinic of 2512 primary shoulder arthroplasties and 222 revision arthroplasties over 22 years. They found that 19 patients with with primary shoulder arthroplasties and 7 patients with revision shoulder arthroplasties were diagnosed with deep periprosthetic infection. Coste et al. retrospectively reviewed 2343 shoulder arthroplasties and found a 1.86% and 4% incidence of infection in primary and revision shoulder arthroplasty, respectively. The rate of infection is higher in patients with some degree of immunocompromise, such as chronic debilitating disease, diabetes mellitus, rheumatoid arthritis, systemic lupus erythematosus, systemic corticosteroid therapy, and intra-articular steroid injections after arthroplasty.

DIAGNOSIS

The most important step in managing suspected shoulder sepsis is the early diagnosis of infection. Early intervention may preempt the development of deep infection and avoid the need for implant removal. Diagnostic tests should be initiated as soon as suspicion of sepsis arises. Making the diagnosis of joint sepsis can be difficult at times, requiring careful evaluation of the patient’s signs and symptoms, laboratory tests, and imaging studies, as well as intraoperative assessment. The signs and symptoms of joint sepsis may be nonspecific but should prompt the physician to order other tests and studies, which can lead to the proper diagnosis. Increased pain is the most common presenting clinical symptoms of infection. An inordinate degree of periprosthetic pain should raise the suspicion of, and be attributed to, infection until proven otherwise. Other symptoms include stiffness, fever, malaise, chills, and night sweats. Initial clinical signs of infection include draining sinus, effusion, swelling, erythema, and septicemia.

Diagnostic laboratory tests are important in establishing the diagnosis and determining the severity of infection. These include an elevation in the white blood cell count (WBC), polymorphonuclear (PMN) cells, erythrocyte sedimentation rate (ESR), and C-reactive proteins (CRP). The WBC is considered abnormally elevated if it is more than 11.0 × 10 ⁹/L, and the number of PMN cells is considered to be increased if it is greater than 80% of the total WBC count. Both the ESR and CRP are nonspecific markers of inflammation and may be elevated in both loosening and infection. An ESR rate of more than 22 mm/h and a CRP level of more than 1 dL/L are considered high and indicate the need for further workup.

All these laboratory values can be elevated in the acute postoperative phase, but abnormalities in the subacute phase are suggestive of an infection or an inflammatory reaction. Exact levels indicating positivity of these laboratory tests are difficult to ascertain because of variable methods of reporting. In addition, a false negative result may occur when the patient has been treated with antibiotics or is on oral steroids. In one of the largest series looking at infection after shoulder arthroplasty, Sperling et al. reported on 26 cases of infected arthroplasty out of 2512 patients. The average time from arthroplasty to the diagnosis of infection was 3.5 years. The preoperative WBC count averaged 7.4 × 10 ³, and it was elevated above 10 × 10 ³in only two patients. The mean ESR was 47 mm/hr, and it was elevated in 14 of 24 patients. More recently there have been reports on the use of serum interleukin-6 as a marker of periprosthetic infection following total hip and knee arthroplasty. Its value for use as a marker of infection in shoulder arthroplasties has not been fully investigated.

It is strongly advised that preoperative and intraoperative cultures be taken to help identify the organism(s) that are the cause of sepsis. The value of preoperative joint aspiration has been debated. Traditionally an aspirate WBC count of more than 50 × 10 ⁹/L was considered evidence of joint infection. Recent studies indicate that, in the absence of an underlying joint inflammatory disease, a synovial fluid leukocyte count of greater than 1.7 × 10 ⁹/L is indicative of joint infection. Gram stain and culture of aspirated joint fluid can provide the definitive diagnosis. However, the sensitivity and specificity of these studies are variable, ranging from 28% to 92% for sensitivity and between 92% and 100% for specificity. Cultures of joint aspirate might yield negative results, especially if the patient has already been started on empirical antibiotics. Sperling et al. had organisms isolated on cultures of the aspirate in 14 of 18 patients, all of whom were deemed infected at the time of revision surgery. Preoperative aspiration of the joint, carried out in eight patients with infected arthroplasties, produced a positive culture in only four patients in a large study by Coste et al.

Over the course of the past few years, intraoperative cultures have become the gold standard for diagnosis and treatments of sepsis. Two studies are typically performed during the revision surgery. In the first, a biopsy specimen from the inflamed synovium is obtained and sent for histologic evaluation of frozen sections and Gram stain. Histologically, more than five polymorphonuclear leukocytes per high-power field are indicative of infection Secondly, intraoperative samples are sent for cultures (aerobic, anaerobic, acid fast, and fungal) and sensitivities. It is important to obtain several specimens for culture, with at least one of them from the intrameddulary canal after removal of the prosthesis. Atkins et al. prospectively evaluated criteria for microbiologic diagnosis of infection at elective revision arthroplasty. They found that five or six specimens were necessary to produce an accurate diagnostic test.

Low-grade infections may not be identified immediately by a routine culture but can be detected by immunoflurescence or molecular biologic techniques that can amplify small quantities of bacterial DNA. A newer form of intraoperative culture is polymerase chain reaction (PCR), which is used in determining the presence of bacteria in a sample. However, although PCR can be an effective way of determining the presence of bacteria in a sample, it does not measure the amount of bacteria in the joint, which is important for determining the appropriate treatment.

Imaging studies may provide additional information to help determine the presence and the extent of joint sepsis. Standard radiographs may show evidence of loosening around the prosthesis or the cement mantle evidenced by the presence of radiolucent lines at least 1.5 mm wide. In addition, evidence of osteolysis with bone erosion or a shift of the components may be present. In a large retrospective multicenter study of 2343 shoulder arthroplasty cases, Coste et al. found radiographic evidence of loosening in 37 of 49 patients with an infected prosthesis (75.5%). Loosening, however, may also be the result of polyethylene wear debris associated with aseptic osteolysis. Torchia et al. analyzed the results of 89 total shoulder arthroplasties at a mean of 12 years postoperatively (range: 5 to 17 years). They found radiolucent lines had developed at the bone–cement interface of 75 glenoid components (84%), and 39 components had definite signs of radiographic loosening.

Scintigraphy can provide additional information that can be helpful in making the diagnosis in difficult cases. Three-phase technetium ⁹⁹bone scintigraphy has high sensitivity but lacks specificity for infection. Technetium bone scan can remain positive for more than a year after prosthesis implantation because of periprosthetic remodeling. The use of bone scintigraphy in conjunction with leukocyte scintigraphy using indium ¹¹¹provides better diagnosis. The accuracy of diagnosis, sensitivity, and specificity using this combination is reported to be around 80% each. More recently, bone marrow scintigraphy using sulfur colloid has been used in some institutions to improve the diagnostic value of bone and leukocyte scintigraphy. The addition of sulfur colloid bone marrow scintigraphy is reported to increase the specificity and accuracy of the diagnosis to 94% and 91%, respectively.

Magnetic resonance imaging (MRI) may be effective when attempting to diagnose infection in a “virgin” joint. However, little has been published on its effectiveness in diagnosing infection in the presence of an arthroplasty implant. Some reports even suggest that MRI is not suitable for assessing arthroplasty complications due to the fact that metal-induced artifacts distort details. Diagnosis of infection is suspected on an MRI when juxtaosseous fluid appears on the image.

A variety of organisms have been implicated in infected shoulder arthroplasties; however, two of the most common causative agents are Staphylococcus aureus and Staphylococcus epidermidis . These organisms were the causative pathogens of infection in about 80% of cases from 1997 to 2003. In their review of 26 patients with infected shoulder arthroplasties, Sperling et al. found S. aureus in the shoulders of 13 patients, S. epidermidis in 9, and Propionibacterium acnes in 5. It is important to determine whether the bacteria in question is a virulent strain such as methicillin-resistant S. aureus (MRSA) or vancomycin-resistant Enterococci (VRE) because of their poor susceptibility to standard antibiotics, with diminished potential for eradication. The difficulty in removal of virulent bacteria is due to the formation of biofilm that hide and protect the infective organism from antimicrobial agents and host immune responses. In addition, microorganisms associated with biofilm formation have shown increased resistance to standard antibiotics. This is due to oxygen depletion within the biofilm, which causes some bacteria to enter a dormant or nongrowing state in which they are less susceptible to growth-dependent antibiotics. Because bacteria in periprosthetic infections are typically present in the biofilm on the surface of prosthesis, it has been hypothesized that culturing of samples obtained from the prosthesis would improve the microbiologic diagnosis of prosthetic-joint infection. Explanted prosthetic biofilm sonication (EPBS) has been proposed as a method for improving the microbiologic diagnosis of prosthetic-joint infection.

Trampuz et al. performed a prospective trial comparing cultures of samples obtained by sonication of explanted hip and knee prostheses to dislodge adherent bacteria from the prosthesis with conventional cultures of periprosthetic tissue for the microbiologic diagnosis of prosthetic-joint infection among patients undergoing hip or knee revision or resection arthroplasty. They found that cultures of samples obtained by sonication of prostheses were more sensitive than conventional periprosthetic-tissue culture (78.5% vs. 60.8%). Patients with virulent organisms such as MRSA require an additional 3 to 4 weeks of antibiotics during treatment. Some authors have even advocated for resection arthroplasty as a treatment option in these patients with resistant organisms.

CLASSIFICATION

Several classification systems have been proposed for staging periprosthetic infections. These classification systems attempt to provide general guidelines for treatment and an insight into whether the prosthesis should be removed or retained. In general, three broad categories of infection can be identified as complications following arthroplasty. These include early postoperative infection (superficial and deep), late chronic infection, and acute hematogenous infection. Figure 22–1 shows the characteristics and time frame related to each type of infection.

Acute infections usually occur within 4 weeks of the time of arthroplasty and are usually related to bacterial contamination during the time of surgery, infected hematoma, or problems with primary wound healing. They are divided into two groups, superficial and deep infections. Superficial infections present with localized swelling and erythema, without signs of joint involvement. Deep infections are characterized by prolonged or increasing postoperative pain. They are usually associated with clear clinical signs and symptoms, and laboratory findings are usually positive. In addition, these infections are commonly caused by virulent microorganisms such as S. aureus and gram-negative bacilli.

Delayed or chronic infections usually occur within 1 year after surgery but after the immediate postoperative period (more than 4 weeks). These infections are characterized by subtle clinical signs and symptoms, and laboratory studies are usually normal or slightly elevated. Caused by less virulent organisms such as P. acnes and S. epidermidis , these infections are difficult to distinguish from aseptic loosening, and scintigraphy studies are often helpful in making the diagnosis. Topalski et al. reviewed the results of 75 patients without overt infection who had positive intraoperative cultures at the time of revision arthroplasty. The most common cultured organism was P. acnes in 45 of 75, followed by S. epidermidis in 10 of 75.

Infections secondary to hematogenous spread are diagnosed if there is a sudden appearance of clinical signs and symptoms of infection after a symptom-free period of more than 1 year after arthroplasty. The manifestations are similar to those of an acute infection, and laboratory findings and imaging studies are usually strongly positive. The most common sources of bacteremia resulting in hematogenous seeding are skin infections, respiratory tract infections, dental work, and urinary tract infections.

MANAGEMENT

Infection of any kind around a prosthetic joint is a surgical emergency and should be treated aggressively. Treatment is determined based on type of infection, organism, previous treatment, extent of infection, and medical condition of the patient. Because relatively few shoulder arthroplasties are performed each year, knowledge gained through wider experience in managing sepsis in hip and knee arthroplasty has provided a basis for management of infection in shoulder. Options include antibiotic suppression therapy, wound debridement and prosthesis retention , one-stage reimplantation (immediate exchange of the prosthesis), two-stage reimplantation, resection arthroplasty, and shoulder arthrodesis. Treatment based on the type of infection is shown in Figure 22–2 .