Incidence, Risk Factors, and Prevention

Although the incidence rates of infection following TKR appear to have fallen over the past several decades, the reported incidence varies in many studies. Much of this variability is inherent in the relatively small numbers of patients studied and in the differences in patient populations described in these reports. As a result, the expected incidence varies widely according to risk factors present in the patient population and the expertise of providers being studied. For example, in 69,663 Medicare patients undergoing elective TKA, the incidence of PJI within 2 years was 1.55%, whereas the incidence occurring between 2 and 10 years was an additional 0.46%.⁵⁹ In another report, with a different patient population, the rate of infected knee arthroplasties was 0.92%; in this study, urban nonteaching hospitals experienced the highest burden of infection (1.26%) compared with those in a rural setting (0.69%) and urban teaching hospitals (0.77%).⁵⁸ Thus, any study reporting on the incidence of deep infection following TKR requires extensive stratification of patient risk factors and comorbidities, and an adequate description of the setting in which the procedures were performed.

Many inherent patient risk factors are known to predispose toward postoperative deep infection. Host factors include a diagnosis of rheumatoid arthritis,^12,^52,¹²⁰ skin ulcers,¹²⁰ diabetes mellitus,^31,⁸⁸ a history of malignancy,⁷ obesity,⁸⁸ a history of smoking,⁸⁸ renal or liver transplantation,¹⁰³ HIV-positive status,^87,¹⁰⁹ prior open knee surgery or periarticular fracture,⁵² and prior septic arthritis or adjacent osteomyelitis.⁵³ It is incumbent on the physician to be cognizant of these risk factors and to incorporate adequate screening measures and efforts to identify these risk factors into a disciplined preoperative evaluation process that facilitates optimization of host variables when possible.

Every effort should be made to incorporate differences in surgical technique and treatment algorithms that minimize the effects of these risk factors. For example, patients on immunosuppressive therapy should be optimized by altering or stopping their medications in the perioperative time period. A recent review summarized current recommendations for perioperative management of the more common antirheumatoid medications (Table 128-1).⁴⁹ Other risk factors, potentially under the influence of the surgeon in the perioperative time period, that seem to predispose toward an increased incidence of deep infection include an increased international normalized ratio (INR) in the postoperative period,⁷⁴ hematoma requiring reoperation,³⁵ early wound healing complications,^7,³⁶ recent intra-articular injection of corticosteroid,⁸⁵ and prolonged operative time.^7,⁸⁹

Table 128-1 Perioperative Antirheumatoid Medication Recommendations

Medication	Important Drug Interactions	Comments
Corticosteroids	Corticosteroid use with fluoroquinolones increases the risk of tendon rupture. Antifungal agents and clarithromycin may increase levels of corticosteroids.	Perioperative use depends on the level of potential surgical stress.
Methotrexate	Methotrexate along with intravenous penicillins may lead to neutropenia.	Continue perioperatively for all procedures. Consider withholding 1 to 2 doses of methotrexate for patients with poorly controlled diabetes; the elderly; and those with liver, kidney, or lung disease who are undergoing moderate or intensive procedures.
Leflunomide	Leflunomide may elevate levels of warfarin and rifampin.	Continue for minor procedures. Withhold 1 to 2 days before moderate and intensive procedures, and restart 1 to 2 weeks later.
Sulfasalazine	May increase INR in patients on warfarin	Continue for all procedures.
Hydroxychloroquine	None	Continue for all procedures.
TNF antagonists	Avoid live vaccines in patients taking these agents; otherwise, no significant perioperative drug–drug interactions are known.	Continue for minor procedures. For moderate to intensive procedures, withhold etanercept for 1 week, and plan surgery for the end of the dosing interval for adalumimab and infliximab. Restart 10 to 14 days postoperatively.
IL-1 antagonist	None	Continue for minor procedures. Withhold 1 to 2 days before surgery and restart 10 days postoperatively for moderate to intensive procedures.

IL, Interleukin; INR, international normalized ratio; TNF, tumor necrosis factor.

From Howe CR, Gardner GC, Kadel NJ: Perioperative medication management for the patient with rheumatoid arthritis. J Am Acad Orthop Surg 14:544–551, 2006.

Proper use of antibiotic prophylaxis represents the single most effective method of reducing infection in total joint arthroplasty.⁴⁶ Optimization of the surgical environment with proper surgical protocols is beyond the scope of this chapter but is clearly under the influence of the surgeon and operating room personnel (Table 128-2).^41,⁷⁸ Use of low-dose antibiotic-loaded bone cement (ABLC) represents an additional surgeon-directed prevention measure for reduction of infection in high-risk patients and in revision surgery.⁵² Frequent irrigation, careful surgical technique, and excellent wound closure are important variables under the surgeon’s control.

Table 128-2 Prevention of Deep Prosthetic Infection

Hematogenous infection of TKR in the early postoperative period, or many years after prosthetic replacement, is often influenced by the surgeon through education efforts made with arthroplasty patients. The rate of bacteremia after invasive procedures appears highest with oral procedures, followed by genitourinary manipulation, and lowest in association with gastrointestinal procedures.²⁹ In general, invasive procedures that potentially cause bacteremia should simply be avoided in the first 3 to 6 months after TKR. In a retrospective review of late infections, seven cases were strongly linked to a dental procedure.¹¹⁵ Predisposing factors in these patients included risk factors such as rheumatoid arthritis or diabetes mellitus with dental procedure duration greater than 75 minutes.

The American Academy of Orthopaedic Surgeons (AAOS) no longer has published guidelines for the use of prophylactic antibiotics for high-risk patients undergoing a procedure considered at high risk for bacteremia. Rather, the AAOS has provided this information in the format of an information statement because of the position that evidence-based literature on this topic is insufficient to meet the criteria of an established guideline. Information included within this information statement, entitled “Antibiotic Prophylaxis for Bacteremia in Patients With Joint Replacements,” can be obtained by request from the AAOS.

Microbiology and Diagnosis

An appreciation of microbial pathophysiology is essential to the accurate diagnosis and subsequent treatment of infected TKR. As strange as it may seem, no criteria for the definitive diagnosis of PJI have attained universal agreement.⁸² Although many authors have suggested that positive cultures are required to make the diagnosis, it is well recognized that some true prosthetic infections are culture negative. In one report of 897 episodes of PJI, 60 (7%) occurred in patients for whom this was the initial episode of culture-negative results.⁹ The 5-year estimate of survival free of treatment failure was 94% for patients treated with two-stage exchange and 71% for patients treated with débridement and prosthetic retention. Culture-negative PJI has been reported to be as high as 19%.⁸⁸

Our current definition of prosthetic infection includes a combination of clinical signs and symptoms, histologic analysis of tissue, and results of cultures. The diagnosis of definite infection is made if evaluation of the knee establishes at least one of the following criteria: (1) two or more cultures obtained by aspiration or deep tissue specimens obtained at surgery yield the same organism, (2) histopathologic evaluation of intra-articular tissue reveals changes in acute inflammation, (3) gross purulence is observed at the time of surgery, or (4) an actively discharging sinus tract is evident.⁴³

It seems reasonable to assume that optimal treatment outcomes are achieved by accurately identifying the offending microorganism(s) whenever possible and enacting directed treatment strategies. Multiple reports have documented the distribution of microorganisms involved.* In most series, gram-positive organisms predominate; however, it is important to remember that polymicrobial infections represent approximately 9% of cases of infected TKR.⁸⁸ It is also important to note that in the current era, many investigators are observing an increasing incidence of resistant organisms.⁸⁶

Methicillin-resistant Staphylococcus aureus (MRSA) and methicillin-resistant Staphylococcus epidermidis (MRSE) have emerged as common nosocomial pathogens often requiring complex antibiotics and potentially inferior treatment outcomes. In a study of 35 infected TKRs, those infected with methicillin-sensitive organisms demonstrated a treatment success rate of 89% compared with only 18% treatment success in those with methicillin-resistant infection.⁵⁵ In contrast, a multicenter study of 37 patients with a resistant organism found only 24% with reinfection, with most of these reinfections caused by a different organism.⁷⁵ These authors believe that two-stage reimplantation remains a viable treatment option for patients who have an infection with a resistant organism at the site of a TKR.

Periprosthetic fungal infections fortunately are rare, with Candida being the predominant species identified.⁹¹ A review of 10 cases documented successful treatment in 8 of 10 patients with appropriate antifungal therapy and two-stage reimplantation. Prosthetic joint infection due to Mycobacterium tuberculosis is also rare.⁸ To decrease the risk of reactivation of quiescent tuberculous infection, consideration should be given to preoperative or perioperative antituberculous prophylaxis.

Identification and diagnosis of biofilm organisms via traditional culture methods have lacked optimal sensitivity and specificity.¹⁰⁷ Basically, several strategies can be used to address these organisms. Culture-independent molecular methods have been developed to improve the diagnosis of prosthetic joint infection in the research setting. Detection of 16S ribosomal deoxyribonucleic acid by polymerase chain reaction is just one example.⁶⁹ One of the primary problems with this approach is the relative difficulty associated with excellent sensitivity of these techniques leading to false-positive test results.

An alternative strategy employs culturing of samples obtained from the prosthesis by sonication of explanted prostheses to dislodge adherent bacteria from the prosthesis.¹⁰⁸ In this study, culture of samples obtained by sonication of prostheses was more sensitive than that of conventional periprosthetic tissue culture for the microbiologic diagnosis of prosthetic hip and knee infection, especially in patients who had received antimicrobial therapy within 14 days before surgery. This technique is now used in our institution as a matter of common clinical practice to improve the detection of organisms causing PJI.

Definitive diagnosis requires a heightened sense of suspicion on the part of the clinician in both early and delayed presentations following surgery. A thorough history, physical examination, plain radiographs, arthrocentesis, and hematologic studies should be considered routine parts of the infection workup. Radionuclide studies are rarely required to make the diagnosis of infection.

Timing of the clinical presentation is a critical factor in the identification and implementation of the correct treatment strategy. These various clinical presentations have been characterized and classified as a useful guide to selecting the most appropriate treatment option for infected TKR (Table 128-3).⁹⁹

Table 128-3 Classification System of Prosthetic Joint Infection: Time to Onset of Infection Dictates Treatment

In this classification, postoperative infections diagnosed by positive intraoperative cultures after revision arthroplasty are generally low-virulence organisms such as coagulase-negative staphylococci and Propionibacterium spp.⁶⁸ Five-year survival free of treatment failure for the 16 episodes was 89%. These results suggest a favorable outcome of prosthetic joint infections caused by low-virulence pathogens initially diagnosed as positive intraoperative cultures after revision arthroplasty.

For the remaining clinical presentations, rapid and expedient diagnosis is essential to prevent the delay of diagnosis, which could result in diagnosis of a late or chronic infection that could have been identified and treated as an early infection. Pain is the most common presenting symptom. Persistent wound drainage is strongly suggestive of infection and probably should be treated with arthrotomy, débridement, and irrigation within the first several weeks after surgery.¹¹⁷ Cultures of serous wound drainage are difficult to interpret and potentially misleading and therefore are discouraged. Empirical antibiotic use for persistent wound drainage should be avoided, as this only suppresses the clinical symptoms of infection and potentially delays diagnosis, eliminating the possibility of treatment of the infection without removal of the prosthesis.^17,^73,^97,¹⁰⁵

Diagnosis of early postoperative infection is typically confirmed by joint arthrocentesis, as the erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) levels are nonspecific in the early postoperative period. CRP levels after TKR peak on postoperative day number two and decrease to preoperative baseline levels as early as 1 week but typically at 14 to 21 days postoperatively.^11,^60,¹¹⁸ In the early postoperative period, assertive management of delayed wound healing or marginal skin necrosis by débridement of necrotic skin and primary wound closure is preferable to empirical antibiotic treatment, prolonged observation, and eventual development of deep infection.⁶²

An acute hematogenous infection typically presents with sudden onset of pain or stiffness in a previously well-functioning arthroplasty.⁵ Specific risk factors for a hematogenous infection, such as a remote source of infection or a recent invasive procedure causing significant bacteremia, should be identified.⁶⁶ The severity of symptoms with pain, effusion, and restricted range of knee motion in this setting facilitates rapid diagnosis. Although ESR and CRP are typically elevated in these patients, the cornerstone of diagnosis is arthrocentesis with evaluation of the aspirate by gram stain, quantitative leukocyte count, and culture for aerobic and anaerobic bacteria. Empirical antibiotics for the unexplained painful prosthesis, without attempts at definitive diagnosis, unfortunately are commonly given; this approach only complicates subsequent efforts to diagnose deep infection.

A vast majority of patients with an infected total knee arthroplasty (TKA) are diagnosed in the subacute or chronic setting. Historical factors such as persistent pain since the arthroplasty, prolonged postoperative wound drainage (Fig. 128-1), antibiotic treatment for difficulties with primary wound healing, and knee stiffness despite extensive rehabilitation efforts may be indicative of deep infection. Sequential comparison of plain radiographs may reveal progressive radiolucencies, focal osteopenia or osteolysis of subchondral bone, and periosteal new bone formation (Fig. 128-2).⁷⁹ Additional studies should include CRP, ESR, and aspiration of the affected TKR. When elevated, CRP and ESR are obtained not only for diagnosis but to serve as baseline values for comparison with testing obtained during and after treatment.

Figure 128-1 Photograph of a chronic sinus tract over the anterolateral aspect of the knee joint in a patient with a chronically infected total knee replacement.

Figure 128-2 A, Anteroposterior and (B) lateral radiographs of a patient with a chronically infected total knee arthroplasty. Radiolucencies at the bone/cement interface, which typically are late findings in periprosthetic infection, can be seen inferior to the tibial component on both views.

In our opinion, arthrocentesis is considered an essential element of the workup and evaluation of a suspected deep periprosthetic infection. Synovial fluid analysis and culture enable identification of microorganisms, and knowledge of the synovial fluid cell count and differential provides an additional data point for diagnosis. Although these values may resemble a native joint infection in the acute hematogenous infection (neutrophil differential >90% and total nucleated cell counts >50,000/µL), it is important to note that most chronic TKR infections are associated with much lower values.^37,^70,¹⁰⁶

Based on our experience, a synovial fluid leukocyte differential of >65% neutrophils (or a leukocyte count of >1.7 × 10³/µL) is a sensitive and specific test for the diagnosis of prosthetic knee infection in patients without underlying inflammatory joint disease.¹⁰⁶ Patients should have all antibiotics discontinued several weeks prior to aspiration, as this oversight frequently accounts for the inability to isolate organisms.⁴ Use of molecular genetic diagnostic methods with joint aspirates, such as the polymerase chain reaction technique, is potentially promising, but these remain experimental modalities for diagnosis of infected joint arthroplasty.^24,⁶⁹ The use of radioisotope scans to facilitate diagnosis of the chronically infected knee prosthesis is only occasionally useful.^67,^94,⁹⁶

Despite all reasonable efforts to diagnose infection preoperatively, intraoperative evaluation of surgical tissue specimens may be necessary to confirm the diagnosis in difficult cases. The gram stain is notoriously unreliable, with a high percentage of false-negative results and an extremely low sensitivity.^2,²⁶ Frozen section testing to detect infection has been widely used with variable accuracy and results.^25,^33,^64,^80,⁸³ Variability is likely accounted for by differences in technique, sampling errors during retrieval of tissue samples, pathologist experience, and the definition used to declare the presence of infection. Analysis of intraoperative frozen section is a reasonable and reliable method when an accomplished and experienced pathologist evaluates appropriate tissue samples.

Although quantitative assessment is difficult, the surgeon’s intraoperative assessment is very helpful, and if there is any reason to suspect deep infection, one should obtain multiple tissue specimens for culture and sensitivity testing. In this setting, a careful review of the history, ESR, and CRP, when combined with the results of preoperative synovial fluid analysis and interpretation of tissue by the pathologist, will often allow determination of whether or not infection is likely.

Treatment

Variables that must be considered before treatment is initiated include (1) determination of whether the infection is superficial or deep; (2) the duration of time elapsed between arthroplasty and diagnosis; (3) identification of host factors that may adversely affect treatment of the infection; (4) appraisal of the soft tissue envelope surrounding the knee, specifically, the integrity of the extensor mechanism; (5) determination of whether the implant is loose or well fixed; (6) consideration of the pathogen(s) responsible for the infection; (7) the physician’s ability to provide the proper level of required care; and (8) careful assessment of the patient’s expectations and functional requirements.

Treatment goals for the infected TKA include eradication of infection, alleviation of pain, and maintenance of a functional extremity. Basic treatment options include (1) antibiotic suppression, (2) open débridement, (3) reimplantation of another prosthesis, (4) arthrodesis, (5) resection arthroplasty, and (6) amputation. With the exception of chronic antibiotic suppression, which does not eliminate infection, the cornerstone treatment principles of these treatment options include thorough surgical débridement combined with appropriate use of antibiotics and optimization of the host response. When confronted with an infected TKR, the treating physician should start by considering the question of whether or not the prosthesis can be retained, or if prosthesis removal is required to treat the infection.

Treatment Methods Whereby the Prosthesis Is Retained

Antibiotic Suppression

Antibiotic treatment alone will not eliminate deep periprosthetic infection but can be used as suppressive treatment when the following criteria are met: (1) prosthesis removal is not feasible (usually because of a medical condition that precludes an operative procedure), (2) the microorganism has low virulence, (3) the microorganism is susceptible to an oral antibiotic, (4) the antibiotic can be tolerated without serious toxicity, and (5) the prosthesis is not loose.¹¹² The presence of other joint arthroplasties or a cardiac valvular prosthesis is a relatively strong contraindication to this treatment approach.

In a multicenter study, antibiotic suppression was successful in only 40 of 225 knees (18%).⁶ Combining several series reveals that antibiotic suppression was successful in 62 (24 %) of 261 knees.⁴² Use of a combined regimen of rifampin with a quinolone has been reported to be more successful than treatment with a single antibiotic.²⁸ Despite the fact that most patients fail to meet all of these selection criteria, antibiotic suppression is commonly practiced, and this practice unfortunately prolongs the presence of infection and often complicates subsequent treatment attempts. Long-term antibiotic suppression should be initiated rarely and should be considered only when all treatment criteria are met.

Débridement With Prosthesis Retention

Open débridement may be indicated for the occasional acute infection in the early postoperative period (type II) or for acute hematogenous infection (type III) of a securely fixed and functional prosthesis. Suggested criteria for this treatment technique include (1) short duration of symptoms of infection (less than 2 weeks), (2) susceptible gram-positive organisms, (3) absence of prolonged postoperative drainage or a draining sinus tract, and (4) no prosthetic loosening or radiographic evidence of infection.¹⁸ Again, a relative contraindication for débridement and attempted salvage of the prosthesis is the presence of other joint replacements or of a cardiac valvular prosthesis.

Results of débridement are difficult to determine because of differences in microbiology and subsequent antibiotic management, variability in time to treatment, quality of the soft tissue envelope, extent and completeness of débridement, status of implant fixation, and the criteria for success in each report. A multicenter study reported a success rate of only 19.5% with open débridement of 154 knees.⁶ In another literature review, a success rate of 32.6% was reported in 530 infected TKRs treated with open débridement and component retention.¹⁰¹ Factors identified with failure of open débridement included postoperative drainage longer than 2 weeks in duration, existence of a sinus tract at débridement, hinged prostheses, and immunocompromised hosts.¹⁰¹

The importance of the timing of débridement in relationship to the onset of symptoms or the period of time elapsed since insertion of the prosthesis cannot be overemphasized.^17,^76,^97,^99,¹⁰⁴ In a study of 24 infected TKRs treated with open débridement and component retention, a success rate of 100% was reported in the early postoperative infection group (type 2) and 71% in the acute hematogenous group with duration of symptoms less than 30 days.⁷⁶ These authors emphasized the strict selection criteria, with patients demonstrating evidence of infection for less than 30 days.⁷⁶ It is clear that débridement with prosthesis retention should not be attempted in patients with chronic infection.^14,^54,⁹⁹