Prosthetic joint infections: bane of orthopedists, challenge for infectious disease specialists.

Prosthetic joint infections (PJIs) occur in approximately 1.5%-2.5% of all primary hip or knee arthroplasties. The mortality rate attributed to PJIs may be as high as 2.5%. Substantial morbidity is associated with a loss of mobility, although this is temporary. The costs associated with a single episode of PJI are approximately $50,000 per episode, exclusive of lost wages. Risk factors that increase the occurrence of PJI include revision arthroplasty, time in the operating room, postoperative surgical site infection, and malignancy. Pain is the most consistent symptom. Staphylococcus species are the most common organisms isolated from PJI sites. Two-stage revision is superior to single-stage revision or to debridement with prosthesis retention. Long-term antibiotic suppression and/or arthrodesis are useful for patients too frail to undergo extensive surgery. Using an optimal approach, recurrent infection occurs in <10% of previously infected joints.

Berbari et al. [2] performed a large case-controlled prospective study of patients undergoing THA at the Mayo Clinic between 1969 and 1991. They reported that the 4 most significant factors predictive of PJI were (1) postoperative surgical site infection (OR, 35.9), (2) a National Nosocomial Infection Surveillance (NNIS) score 12.0 (OR, 3.9), (3) concurrent malignancy (OR, 3.1), and (4) prior THA (OR, 2.0). NNIS scores are calculated from 0-3 points, with 1 point each for anesthesia scores 13, operating room time 13 h, and wound class у3. The mean time for follow-up in this prospective study was 512 days.
Despite the risks inherent in THA or TKA, both are much safer procedures now than in the past [11]. Four decades ago, 1 in 10 patients developed infection. Aggressive measures, including use of laminar air flow rooms, body exhaust suits, and judicious perioperative antibiotics, have reduced the risk of infection to 0.3% in THA and 0.5% in TKA. These results will be difficult, if not impossible, to improve.
Clinically, pain is the single most frequent symptom and is often exacerbated by motion. Local warmth, tenderness, drainage, and effusions are helpful in diagnosing infection. They are not as universally present as is pain [1]. A normal erythrocyte sedimentation rate (ESR), along with a normal C-reactive protein level, would suggest a very low risk of infection [1].
The most frequently recovered isolates are Staphylococcus aureus and Staphylococcus epidermidis [2,[5][6][7][12][13][14][15][16] in PJI infections. Table 2 presents data on the incidence of specific organisms recovered from infected THA or TKA. Gram-positive cocci predominate. Although there have been few evidencebased data to support a more aggressive surgical approach to infection with gram-negative bacilli, the impact of the recovered  etiologic agent on the approach to management is substantial. There exists a bias, based on older studies, that gram-negative bacilli in a PJI most often require total removal of the prosthesis and a 2-stage revision, whereas some PJIs due to gram-positive cocci can be treated with debridement and retention of the prosthesis [12] or a 1-stage revision procedure [6,7,17]. Various maneuvers to enhance the recovery of an isolate from the infected prosthesis have been attempted. These include direct inoculation of blood-culture vials with intraoperatively recovered joint aspirate [18] and immersion of the infected prosthesis into an anaerobic jar, followed by ultrasonography to loosen adherent organisms [19]. Histological evaluation of frozen section of intraoperatively obtained tissue from the prosthetic joint has been shown to have increased sensitivity and better negative predictive value than Gram stain alone in establishing infection [20].
In the current era, radiological evaluation includes not only plain x-ray and CT but MRI [21] and even positron emission tomography (PET) scans [22]. Table 3 reviews the radiological and nuclear medicine imaging techniques available or under study for their utility in diagnosing PJI [21][22][23][24][25][26][27][28][29]. With modifications, standard MRI can yield diagnostic-quality images, especially along the femoral component of THA. PET scan has a sensitivity of 90%, a specificity of 89.3%, and an accuracy of 89.5% in detecting THA infection. Its accuracy for TKA infection is 77.8% [22]. Table 3 should provide an overview of the developments within this area. The diagnosis of PJI requires clinical correlation.

APPROACHES TO THERAPY FOR THA INFECTIONS
The reader should note that the following is not an exhaustive review of the literature but represents my selections and reflects my opinions. Single-stage revision arthroplasties of the hip can be successfully performed with excellent outcomes. This approach has significant advantages that weigh in its favor: decreased duration of immobility, reduced cost, improved stability of the limb, and decreased patient morbidity [17]. However, to be successful, 1-stage exchanges require appropriate perioperative antibiotic treatment and meticulous operative techniques. Ure et al. [6] described 20 patients who had an average of 10 years free of infection in the postoperative period with the use of antibiotic-impregnated cement. It should be noted that a large number (45%) of the patients had a coagulasenegative Staphylococcus species PJI. Streptococcus species and S. aureus were recovered from 25% of the patients. Perioperative antibiotics were administered parenterally for an average of 5 weeks (range, 2-18 weeks) and then orally for 4.7 months (range, 2.5-6.0 months). These authors concluded that, with careful selection of patients, the 1-stage hip revision was as successful as a delayed exchange procedure. They specifically excluded patients who were immunosuppressed, had infection with known gram-negative or methicillin-resistant organisms, or had major skin, soft-tissue, or bone defects that made it impossible to obtain a closed wound or a stable implant. Miley et al. [17], in an earlier report, evaluated 101 patients with an 87% success rate after 1-stage revision. They limited 1-stage revision arthroplasty to those patients who had infections with monomicrobial gram-positive organisms susceptible to a minimum of 3 antibiotics, no evidence for superinfection or draining sinus tracts, and relatively healthy surrounding soft tissue and femoral cortical bone. Patients who did not meet these inclusion criteria were treated with girdlestone resection. Finally, Jackson and Schmalzried [7] reviewed 1299 1-stage hip revisions with an average duration of follow-up of 4.8 years (range, 0.1-17.1 years); 83% were deemed to be free of infection. Antibiotic-impregnated cement was used in 99% of the cases. Parenteral antibiotic use varied from 24 h to 8 weeks in length, and oral antibiotic use varied from none given to 8 months' duration after parenteral therapy. This is one of the largest studies of this procedure. Factors predicting a successful outcome included an absence of wound complications after primary THA, good general health of the patient, methicillinsensitive staphylococcal or streptococcal infection, and an organism sensitive to the antibiotic mixed into the bone cement. Predictors of failure included polymicrobial infection, gramnegative organisms, and methicillin-resistant staphylococci.
Two-stage THA revision was compared with 1-stage revision THA for infections caused by coagulase-negative staphylococci. [8]. Ninety-one patients with cemented prostheses and monomicrobial infection were evaluated for outcome after revision arthroplasty. Seventy-two patients had 1-stage revision, whereas 8 of 19 had a 2-stage revision. Although 83% of the patients with 1-stage revision were free of infection at a mean of 45 months, all 8 patients who had undergone the 2-stage procedure were had no infection after 21 months of follow-up. Problems with this study included the small number of patients who completed the 2-stage procedure and the short duration of follow-up. The authors found no differences in outcome of the patients who had undergone the 1-stage procedure on the basis of susceptibility of the isolate to gentamicin impregnated into the cement, presence or absence of sinus tract drainage, and elevated ESR (140 mm/h).
Debridement with retention of infected THA was advocated by Fisman et al. [3] on the basis of a mathematical model of clinical and cost effectiveness. With the increasingly elderly population at risk for THA revision, the benefits of reduced morbidity, less immobilization, and rehabilitation make debridement with retention an attractive choice. However, there are 4 cardinal assumptions made in this model for it to be cost effective. First, patients must not have unstable prostheses or gram-negative infections. Second, all patients receive 6 weeks of antibiotics after debridement. Third, debridement occurs within 30 days of the onset of clinical symptoms. Fourth, to make this approach cost effective, enough older patients die before needing resection arthroplasty. There was an annual relapse rate of 30% with all infections, and, for those patients with delayed debridement 130 days for S. aureus infection, the relapse rate was 80%. Brandt et al. [9] performed a retrospective study of S. aureus PJI. The 1-and 2-year treatment failure rates (defined as a relapse of infection) were 54% and 69%, respectively. A median of 4 procedures/patient was required to control the infection. Patients who had debridement delayed beyond 2 days after the onset of symptoms had a 12-fold (82% vs. 30%) increased risk of failure at 1 year after surgery. Although their study was small, with 33 infected hips treated by debridement, there were 9 patients who remained free of infection, whereas the majority (24/33) either failed to respond or relapsed within 2 years.
Long-term antibiotic suppression of infected THA without surgery can be successful in up to 60% of patients. Girdlestone resection of the hip is indicated for patients with severe bone loss, polymicrobial infection, virulent isolates, concurrent osteomyelitis, and unhealthy soft tissues. When a girdlestone resection becomes infected, long-term antibiotic administration is required to satisfactorily eradicate or suppress the infection. Long-term antibiotic suppression is an approach that should be reserved for the most medically fragile patients. There are no guidelines as to the duration or durability of such an approach. Drug toxicity, side effects, and potential antibiotic resistance are all possible consequences.
Two-stage revision of infected THA is the procedure most often used in the United States. The majority of patients who present with PJI for revision THA do not meet the inclusion criteria for debridement and retention proposed by Fisman et al. [3]. Rather, polymicrobial infection, a lack of positive cultures from the infected joint, delay beyond 30 days from onset to presentation and loosened prostheses make that option unrealistic [30]. The majority of patients present with several months of pain, extended periods of infection, and consequent implant instability, making it more likely that a 1-or 2-stage implant exchange will be required.
Success rates of 83%-87% can be achieved with 1-stage revision. Two-stage revision carries a 9% risk of reinfection or failure-that is, it has a 91% success rate [31]. The importance of antibiotic-impregnated cement is lessened with the 2-stage procedure because of the interval of 6 weeks to 6 months during which time the patient can receive specific antimicrobial therapy. Cementless THA revision with reimplantation beyond 1 year after resection was associated with a 4-fold decrease in reinfection (27% vs. 7%), compared with earlier reimplantation. The use of antibiotic-impregnated cement for THA improved the outcome. However, the study was smaller, with only 82 patients, and performed during an earlier era [10], when there was a more limited choice of antibiotics. Nonetheless, the 2-stage procedure remains the surgical option most frequently selected for the revision of an infected joint.

APPROACHES TO THERAPY FOR TKA INFECTIONS
In the case of 2-stage TKA revisions [30] success rates of 89%-100% are noted with follow-up periods of 2-7.5. years. Mont et al. [13] prospectively compared 35 patients who underwent 2-stage TKA reimplantation after 6 weeks of antibiotics with 34 patients who had reimplantation deferred for an additional 4 weeks of antibiotic therapy, at which time a repeat culture of the joint was performed. They found that 3 of 34 patients for whom implant was delayed had positive cultures and required an additional 6 weeks of antibiotics; 1 patient, despite having a negative culture, had a recurrent infection. In comparison, among the first group of patients, who had TKA reimplantation after 6 weeks of antibiotic therapy, 5 (14%) of 35 had recurrent infection. One way to look at these data is that, even with a 6-week-long course of therapy directed against the initial recovered pathogens, up to 9 (13%) of 69 had recurrent infection. Similar results were reported by Goldman et al. [14] in an earlier study. Among the 60 patients they monitored for up to 7.5 years, 6 had had reinfection; 6 additional patients had aseptic loosening.
Single-stage revision of infected TKA was noted by Bengston and Knutson [15] to have similar outcome to 2-stage procedures. Among 107 knees undergoing the procedure, 81 had clinical cure, regardless of the revision procedure selected. Goksan and Freeman [16] reported 1-stage revision to be reasonable, because only 1 of their 18 patients developed a recurrent infection. Patients were monitored for up to 10 years after procedure. Rand and Bryan [32], in an earlier study, reported a 36% success rate-5 of 14 patients remained free of infection when reimplantation was performed within 14 days of infected TKA removal. Recently, Fansa et al. [33] found that the 2-stage procedure was superior to the 1-stage replacement in patients who underwent concurrent muscle-flap coverage. The reader should note that the data are very limited. Most patients in the above studies had infection with low-virulence organisms, such as coagulase-negative staphylococci.
Debridement and retention of a TKA can be attempted by arthroscopic surgery. Proper selection of patients includes those with !7 days of symptoms and no radiological signs of osteitis or prosthetic loosening. Patients who are medically unstable or who are taking anticoagulation therapy can have a successful outcome with this approach. When combined with antibiotic suppression [34], ∼40% of patients do not require subsequent implant removal. Bengston and Knutson [15] did not find debridement and antibiotic therapy to be useful, resolving infection in only 37 (24%) of 154 knees.
Not all patients qualify for joint replacement. Windsor et al. [35] suggested that medically compromised patients with rheumatoid arthritis should instead have arthrodesis of the joint.
In summary, there is no single approach that is best. Patient Downloaded from https://academic.oup.com/cid/article-abstract/36/9/1157/312365 by guest on 15 March 2020 preference and health status need to be considered. Orthopedic surgeons assess the health of the infected bone and soft tissues. That consideration and the organism(s) recovered and their antibiotic susceptibility will dictate the approach to therapy. Infectious diseases specialists should advise surgeons on the optimal therapy, dosing, route of administration, and duration. We monitor these patients simultaneously and actively participate in their postoperative management. The diagnosis of PJI is often difficult, and treatment is prolonged, complicated, and expensive. Although its reported incidence is now !1% for all implanted joints [11], the occurrence of an episode of infection is a medical and surgical disaster for the individual patient. It is questionable whether randomized, prospective studies large enough to establish a difference in outcome between surgical approaches will ever be performed. We are, therefore, left to draw conclusions from small caseseries reports.
Although 2-stage delayed reimplantation seems to be the most efficacious in terms of infection-free status for both THA and TKA infections, other approaches may be of benefit to individual patients. It seems clear that even debridement and retention may be successful under optimal conditions.