Débridement and Implant Retention and Nonsurgical Options for Implant Salvage

Darin J. Larson, MD

Kevin L. Garvin, MD

INTRODUCTION

Total knee arthroplasty (TKA) has been one of the most successful procedures in the field of orthopedics over the course of the last several decades. High rates of patient satisfaction and ever-increasing implant survivorship have resulted in a substantial increase in demand for the operation. Although these operations have had a significant positive impact, complications still exist. As the volume of primary TKA continues to trend upward, rates of revision TKA due to a variety of causes have been projected to increase as well. Periprosthetic joint infections remain one of the most difficult of all of the complications to manage. While certain causes of revision have been decreasing, the relative rate of periprosthetic infections as a cause of revision has been increasing, now accounting for approximately 15% to 25% of all revision TKA procedures.¹,²,³,⁴ Current rates of prosthetic joint infection after primary TKA range from 0.4% to 2%, which has largely remained stable over the last decade.⁵,⁶,⁷,⁸,⁹,¹⁰,¹¹,¹²

The goals of treatment of TKA infections are threefold: elimination of infection, provision of a functional limb in order to allow the patient to perform their daily activities, and alleviation of pain. In order to accomplish these goals, multiple techniques have been described utilizing a combination of surgical intervention with a course of culture-specific antibiotics. These techniques include (a) irrigation and débridement with implant retention, (b) one-stage reimplantation, (c) two-stage reimplantation, (d) arthrodesis, and (e) amputation. Chronic antibiotic suppression may also be considered, although this treatment option does not eradicate infection. Indications for antibiotic suppression will be discussed later in this chapter. In the United States, the current gold standard for treatment of most TKA infections is a two-stage arthroplasty with placement of an antibiotic cement spacer. Success rates have been in the range of 72% to 100%.⁴,⁵,⁶,⁷,⁸,¹³,¹⁴,¹⁵ However, the potential for severe bone loss following explant and significant difficulties with mobility prior to reimplantation are causes for concern. This is even more troubling when considering the typical demographics of patients undergoing these large, staged procedures with a prolonged recovery, as these patients are more likely to be older with multiple medical comorbidities (i.e., diabetes, cancer, obesity). Two-stage reimplantation involves resection of all prosthetic components during the initial débridement with placement of an antibiotic-laden cement spacer, followed by a period of intravenous antibiotic administration, and eventually a return to the operating room for removal of the spacer, débridement, and reimplantation of new prosthetic components if there is evidence that the previous infection has been cleared.⁵,⁶,⁷

Prosthetic component retention, also known as débridement and implant retention (DAIR), has been studied as an alternative to two-stage reimplantation TKA. Indications for DAIR include clinically and radiographically well-fixed components, acute postoperative infections, or acute hematogenous infections as described by Tsukayama and Segawa,¹⁶,¹⁷ absence of a draining sinus tract, and infection with a single, low-virulence organism that is susceptible to antibiotics.¹⁸ These specific criteria will be discussed in depth later in this chapter. By retaining implants, bone stock may be preserved and the lengthy period of restricted weight-bearing and other functional limitations may be avoided. Management of periprosthetic infections involving component retention includes the following: (a) open irrigation and débridement with or without modular component exchange, (b) arthroscopic irrigation and débridement, (c) long-term antibiotic suppression without surgical intervention, and (d) serial joint aspirations combined with antibiotic administration (Table 75-1). While certain levels of success have been documented in the literature with component retention, these techniques have not matched the success rate of two-stage reimplantation. Outcomes with component retention can be improved if stringent indications for treatment are used.¹⁶,¹⁷,¹⁸,¹⁹,²⁰,²¹

OPEN IRRIGATION AND DÉBRIDEMENT

Open irrigation and débridement with modular component exchange (including placement of a new polyethylene tibial implant) is the most frequently described method of component retention in the management of TKA infections. This is accomplished with an arthrotomy, typically
utilizing the prior skin incision. Following the arthrotomy, a radical synovectomy of the knee including the medial and lateral gutters and posterior capsule as well as extensive lavage of the retained components is then undertaken. Care must be taken to perform a comprehensive débridement of all necrotic and infected tissues, as this can be a nidus for persistent infection. Removal of the polyethylene liner allows for improved exposure and subsequent débridement of the posterior capsule. Although a postoperative course of intravenous antibiotics for several weeks is a necessary part of treatment following débridement, antibiotics alone or in the setting of an insufficient synovectomy will be unable to eradicate the infection (Fig. 75-1). Thus, a thorough débridement is the essential component of this technique. No violation of the bone-prosthesis, cement-prosthesis, or bone-cement interfaces is necessary. However, these interfaces must be directly visualized to ensure that there is no evidence of loosening.⁴,²²

TABLE 75-1 Component Retention Methods and Indications in Periprosthetic Infections of the Knee

Treatment	Success Rate (%)	Indications	Comments
Open irrigation and débridement with or without polyethylene exchange	17.9%-100%	Symptom duration under 4 wk Low-virulence organism, susceptible to antibiotics Absence of draining sinus tract No evidence of radiographic or clinical loosening No radiographic or clinical evidence of infection	Combined with a postoperative course of culture-specific antibiotics Failure to exchange the polyethylene component is a predictor of failure
Arthroscopic irrigation and débridement	38%-100%	Same as open Consider in patients with coagulopathies or those with high potential for wound healing problems	Polyethylene component cannot be removed
Long-term suppressive antibiotics	18%-25%	Refusal of further surgery Patients unable to undergo an operation due to significant medical comorbidities Infection with low-virulence organism Ability to tolerate antibiotic side effects	Potential for issues with compliance May lead to antibiotic resistance Contraindicated in presence of other noninfected implants
Serial aspirations	10%-15%	Similar to suppressive antibiotics	Limited data available regarding its use

Historically, open irrigation and débridement with polyethylene exchange has not been as successful at eradicating infection as two-stage reimplantation. A wide range of success rates with this technique has been described in the literature (Table 75-2).²³,²⁴ Burger et al²⁵ followed 39 periprosthetic knee infections for an average of 4 years; only 7 of the 39 (17.9%) infections were successfully eliminated. Meehan and colleagues²⁶ reported a 100% retention rate in 13 patients with a mean follow-up of 5.8 years. The evidence is somewhat conflicting, as there is no standardized definition of “success” and “failure.” Multiple studies have considered a case to be successful after multiple débridements as long as the prosthetic components are retained, whereas others have recorded failures if more than one surgical procedure is required.⁴,⁷,²⁴,²⁵,²⁶

Certain cohorts do appear to be more amenable to irrigation and débridement with component retention. Multiple studies have discussed duration of symptoms prior to undergoing surgical intervention as the most important predictor of success. In general, the maximum amount of time allowed when considering retention of the prosthesis after symptom onset is 2 to 4 weeks. However, the true duration of infection may be difficult to ascertain, specifically within the first month following the index procedure. Thus, the duration of infection in the early postoperative period is typically quantified as the amount of time since the procedure was performed.⁴,⁷,²²,²⁴,²⁷,²⁸

When discussing the duration of symptoms, it is essential to differentiate infections in the early postoperative period and those that occur months to years after the initial arthroplasty. Tsukayama, Segawa, and colleagues have proposed a classification system based on the clinical presentation of periprosthetic joint infections. They describe four types of infections: (a) type-I infections are identified from positive cultures at the time of revision arthroplasty; (b) type-II infections are acute postoperative infections; (c) type-III infections are acute hematogenous infections; and (d) type-IV infections are chronic indolent infections that have been present for over 1 month. Acute postoperative knee infections occur within the first month following the index TKA; acute hematogenous infections are usually the result of a bacteremic event followed by acute symptom onset in the affected prosthetic joint within 48 hours. These acute hematogenous infectious can manifest after months to years without prior complication.¹⁶,¹⁷,²⁹ The duration of symptoms is of utmost importance in these situations. Gehrke et al⁴ reported a failure rate exceeding 60% for late-presenting cases. Triantafyllopoulos et al²⁴ evaluated 78 periprosthetic knee infections with successful retention in 43 of the 78 knees (55.1%) at final follow-up. Symptom
present for longer than 5 days was an independent risk factor for failure. Urish et al²⁷ noted a higher risk of failure with symptoms for longer than 4 weeks compared to those with symptoms present for less than 1 week. In this study, an overall failure rate of 57.4% following irrigation and débridement with component retention was noted at 4 years, with most of the failures (89.9%) occurring within the first year following DAIR. Failure was defined as the need for any subsequent operation. In a study by Wasielewski and colleagues,³⁰ seven acutely infected knees (defined as symptoms present for less than 2 weeks) were treated with open débridement and component retention. Eradication of infection was reported in five of seven knees (29% failure rate). The authors also reported failure of infection control in one of two chronically infected knees (symptoms present for longer than 2 weeks) treated with open débridement, noting that chronic infections were significantly less likely to be successful with attempted component retention.

FIGURE 75-1 A 65-year-old woman with osteoarthritis was treated with a left total knee arthroplasty. She had an uneventful recovery for 4½ mo postoperatively until she suddenly developed a fever and left knee pain. Inflammatory markers were mildly elevated (erythrocyte sedimentation rate of 26, C-reactive protein of 1.1). The knee was aspirated and sent for cell count and culture. Gross purulence was noted during the aspiration. Synovial white blood cell count of 15,000 with 93% neutrophils was reported. Cultures grew methicillin-sensitive Staphylococcus aureus. A and B: Radiographs of the left knee prior to the index arthroplasty. C and D: Radiographs taken within 24 h of developing symptoms. No evidence of osteomyelitis, osteolysis, progressive radiolucent lines, or loosening is noted. The patient underwent open irrigation and débridement with extensive synovectomy and polyethylene exchange. All metallic components were retained. Dakin’s solution irrigation was utilized. E and F: Postoperative radiographs obtained at 16-mo follow-up following the open débridement with no signs of recurrent infection.

TABLE 75-2 Summary of Studies on Irrigation and Débridement for Periprosthetic Infections of the Knee

Study	Number of Cases With PJI of the Knee	Mean Age of Patients (y)	Type of I&D	Pathogens	Follow-Up	Success Rate	Predictive Factors	Notes
Aboltins et al⁵²	7	72.1 (range, 58-81)	Open with exchange of the polyethylene liner when possible	MRSA, MSSA, CNS	29 mo (6-65)	71.4%	n/a	One patient died at 6 mo due to causes unrelated to infection.
Azzam et al³⁶	53	65 (range, 17-88)^a	Open with exchange of polyethylene liner in 26 patients and reimplantation of preexisting liner in 27 patients after immersion in betadine solution	Methicillin-resistant staphylococci, methicillin-sensitive staphylococci, polymicrobial infections^a	5.7 (2.4-10.4) y^a	45.3%	Adjusted analysis: staphylococcal infections, ASA score, and purulence around the joint. Unadjusted analysis: urinary tract infections during hospital stay and persistent drainage after débridement.^a
Barberan et al⁵³	28	74.6 ± 8.4	Not specified	Staphylococcus aureus, CNS	>1 y	57.2%	Duration of symptoms >6 mo, MRSA infection.^a
Bradbury et al³⁴	19	n/a	Open with polyethylene exchange	MRSA	43 (27-55) mo	16%	n/a	Prosthesis retention under chronic suppressive treatment with antibiotics was considered successful outcome.
Brandt et al⁵⁴	26	n/a	Not specified	S. aureus	Success group: 2345.5 (55-5221) d, failure group: 81 (15-614) d^a	38.5%	I&D performed >2 d after onset of symptoms and primary arthroplasty related with negative outcomes.^a
Buller et al⁵⁵	247	65 (range 12-94)^a	Open with polyethylene exchange	Four groups: (1) MRSA, vancomycin-resistant Enterococcus, and MRSE; (2) methicillin-sensitive CNS or MSSA; (3) other Gram-positive organisms or fungus; and/or (4) Gram-negative organisms^a	34 mo (range, 8 d-12.9 y)^a	50.6%	Duration of symptoms, preoperative ESR, pathogen (group 1), previous infection.^a
Burger et al²⁵	39	n/a	Open (fate of liner not specified)	Staphylococcal species, streptococcal species, Gram-negative bacteria, enterococci, polymicrobial	4.1 (1-13) y	17.9%	Duration of symptoms <2 wk, susceptible pathogens, absence of prolonged drainage or sinus tract, and absence of loosening were correlated with positive outcomes.	No clear statistical method reported for defining predictive factors.
Byren et al³²	51	n/a	Open with polyethylene liner exchange or arthroscopic	MRSA, MSSA, CNS^a	2.3 y^a	74.5%	S. aureus infection, previous revision, and arthroscopic washout related with negative prognosis.^a
Chiu et al⁵⁶	40	72.7 (range, 59-85)	Open with polyethylene liner exchange	MRSA, MSSA, CNS, Staphylococcus epidermidis, Gramnegative bacteria, streptococci, polymicrobial, Candida species	79 (36-143) mo	30%	Early postoperative (type I) and hematogenous (type III) infections correlated with better outcomes.	Only infected revision TKAs were studied. Chronic infections (n = 20) included in this cohort.
Chung et al⁴⁶	16	70 (56-78)	Arthroscopic	Streptococci, MSSA, MRSA, CNS, Mycoplasma hominis, negative cultures	47 (24-86) mo	62.5%	n/a	Only patients with duration of symptoms less than 72 h, previously well-functioning prosthesis, and no radiographic signs of loosening were included.
Duque et al⁴⁹	67	64.5 (36-82)	Open with polyethylene exchange and lavage with normal saline, betadine, Dakin’s solution, and bacitracin	Non-MRSA Staphylococcus, MRSA, Enterococcus, Streptococcus, Pseudomonas, Peptostreptococcus, Escherichia sp., Serratia sp., Proteus sp., Prevotella sp., Granulicatella sp., Enterobacter sp., Citrobacter sp., Aerococcus sp.	4.81 (2.04-9.40) y	68.66%	All MRSA and Pseudomonas infections failed.	100% success rate with Streptococcus sp. and anaerobes.
Estes et al²²	16	67 (range, 28-91)^a	Staged open procedure with use of antibiotic impregnated cement beads^a	MRSA, MSSA, CNS, Escherichia coli, streptococcal species, polymicrobial, Enterococcus faecalis, negative cultures	3.5 (1.2-7.5) y^a	87.5%	n/a
Fehring et al⁵⁷	46	61 (range, 17-89)^a	Open with polyethylene liner exchange in 98% of cases	Susceptible and resistant staphylococci, others^a	46 (24-106) mo^a	37%	n/a
Fink et al²⁸	39 (early periprosthetic) and 28 (acute hematogenous)	67.8 (30.0-80.0)	Open with modular component exchange and irrigation with octenidine	S. aureus, S epidermidis, other Staphylococcus species, Propionibacterium acnes, Streptococcus sp., others	41.8 (24-132) mo	71.6% overall, (82.1% early infections, 57.1% acute hematogenous)	Negative factors: longer time between procedure and first appearance of symptoms (>2 d), higher number of previous operations, higher ASA classification, and nicotine abuse.
Gardner et al⁵⁸	44	70 (range, 48-94)	Open with polyethylene liner exchange	S. aureus, S. epidermidis, other Gram-positive bacteria, Gram-negative bacteria	5 (1-9) y	43.2%	n/a
Geurts et al⁵⁹	20	69 (range, 27-93)^a	Staged open procedure with use of gentamicin-PMMA beads and without exchange of polyethylene liner	S. aureus, CNS, streptococci, Enterobacter, Pseudomonas aeruginosa, P. acnes, polymicrobial, negative cultures	52 (3-202) mo	85%	Interval between symptom onset and treatment (cutoff 4 wk).^a
Ilahi et al⁴³	5	60.2 (range, 49-70)	Arthroscopic	Streptococci, CNS	41 (36-43) mo	100%	n/a	Very small case series.
Konigsberg et al²⁹	22	60 (range, 25-86)^a	Open with exchange of polyethylene liner	Staphylococcal species, streptococcal species, others^a	56 (25-124) mo^a	77.3%	Staphylococcal infection was the only negative predictive factor.^a
Koyonos et al⁶⁰	78	64 (range, 18-89)^a	Open with exchange of polyethylene liner	Staphylococcal species, Gram-negative organisms, negative cultures^a	54 (12-115) mo^a	38.5%	Only staphylococcal infection independently predicted failure.^a
Kuiper et al⁶¹	29	70 (success group), 69 (failure group)	Open with or without exchange of polyethylene liner and with or without the use of gentamicin sponges or beads	CNS, S. aureus, streptococcal species, E. coli, Enterobacter cloacae, E. faecalis, others^a	³⁵ (0-79) mo^a	75.9%	Rheumatoid arthritis, symptoms >1 wk, late infection (>2 y from index procedure), ESR >60 mm, and CNS infection were correlated with negative outcome.
Löwik et al⁵⁰	86	73.2 y (standard deviation ± 11.5)	Open with optional exchange of modular components, gentamicin-impregnated beads, or sponges were inserted	S. aureus, Cutibacterium acnes, Bacteroides fragilis, Finegoldia magna, Proteus, Corynebacterium, others	Final follow-up not specified	62.8%	KLIC score used to predict failure in débridement and implant retention.
Marculescu et al³¹	52 cases	74 (range, 23-95)^a	Open with or without exchange of polyethylene liner^a	Staphylococci, streptococci, enterococci, Gram-positive bacilli, Gram-negative bacilli, anaerobic, polymicrobial, negative cultures, fungi, others^a	700 (1-2779) d^a	60% at 2 y^a	Univariate analysis: S. aureus, presence of sinus tract, and duration of symptoms ≥8 d correlated with adverse outcomes. Multivariate analysis: presence of sinus tract and duration of symptoms ≥8 d correlated with adverse outcomes.^a
Martinez-Pastor et al⁶²	32	70.7 ± 11.3^a	Open with exchange of polyethylene liner	Gram-negative bacilli	463 (344-704) d^a	75%	Negative predictors: CRP >15 mg/dL, treatment not including fluoroquinolones.^a
Meehan et al²⁶	13	70 (range, 44-86)^a	Open. Liner exchange was performed in four patients	Streptococcal species	2120 (672-4015) d	100%	n/a
Mont et al³⁸	24 joints in 22 patients	66 (range, 46-80)	Open with exchange of polyethylene liner in 21 cases and reimplantation of preexisting liner in 3 cases after immersion in betadine solution	Staphylococcal species, streptococci, Gram-negative bacteria, Aspergillus	45.1 (24-140) mo	83.3%	n/a
Narayanan et al⁵¹	55	60.7 (success group), 58.7 (failure group)	Open with exchange of polyethylene liner	S. aureus, S. epidermidis, Staphylococcus lugdunensis, Actinomyces meyeri, Group B Streptococcus, P. acnes, P. granulosum, Pseudomonas, polymicrobial	Minimum 1-y follow-up	82% (if treated within 2 wk), 50% after 2 wk	I&D after 2 wk was significantly more likely to fail.
Segawa et al¹⁷	17	n/a	Open with polyethylene liner exchange	Staphylococci, streptococci, enterococci	n/a	58.8%	n/a
Teeny et al⁶³	21	58 (30-74)^b	Open (fate of liner not specified)	Staphylococci, streptococci, Gram-negative bacteria, polymicrobial^b	4 (2-12) y^b	28.5%	n/a
Urish et al²⁷	216	65.9 ± 12.2	Open with exchange of polyethylene liner	Culture-negative, S. aureus, others	31.5 (IQR 14.4-67.0) mo	49.5%	Culture-negative infection had highest risk of failure, followed by infection with S. aureus.
Vilchez et al⁶⁴	35	70 ± 10.8^b	Open with exchange of polyethylene liner	S. aureus	879.3 ± 205 d^b	68.6%	Negative predictive factors: onset of infection ≤25 d after joint arthroplasty, CRP at admission >22 mg/dL, documented bacteremia and the need for a second débridement.^b
Waldman et al⁴⁵	16	72 (range, 57-82)	Arthroscopic	Staphylococci, streptococci, E. coli	64 (36-151) mo	38%	n/a	Only patients with symptoms ≤7 d and without radiographic evidence of loosening were included.
Zurcher-Pfund et al⁶⁵	21	80	Open in 11 patients, arthroscopic in 10 patients	MRSA, MSSA, CNS, streptococci, E. coli, Clostridium septicum, Pasteurella multocida	7 (4-20) y	33%	n/a
^aData include patients with PJI of the hip. ^bData include patients subjected to 2-stage revision.
ASA, American Society of Anesthesiologists; CNS, coagulase-negative staphylococci; CRP, C-reactive protein; ESR, erythrocyte sedimentation rate; MRSA, methicillin-resistant Staphylococcus aureus; MRSE, methicillin-resistant Staphylococcus epidermidis; MSSA, methicillin-sensitive Staphylococcus aureus; n/a, not available; TKA, total knee arthroplasty. Updated from Triantafyllopoulos GK, Poultsides LA, Zhang W, Sculco PK, Ma Y, Sculco TP. Periprosthetic knee infections treated with irrigation and débridement: outcomes and preoperative predictive factors. J Arthroplasty. 2015;30(4):649-657.