Deep infections are among the most severe complications in total knee arthroplasty. Infection rates described in the first series published were as high as 23%.

In more recent papers, the figure ranges between 1% and 2%

. In a recent review of 6,489 knees, this figure was brought down to 0.4% in primary surgeries and 1% in prosthetic revisions.

Assessment of risk factors

Appropriate patient selection and a correct preoperative assessment are fundamental to determine whether the benefits of the surgery may be outweighed by its risks. Rheumatoid arthritis, diabetes mellitus, obesity, immunosuppression treatment, psoriasis and previous infections in the knee are the risk factors most clearly related with TKA infection.

Some risks are universal: Lack of vascularization in the prosthesis itself hampers the arrival of antibiotics and cells from the immune system. Tissue-initiated dissemination is the only mechanism through which antibiotics can reach the infected area. Also, interaction between the prosthetic components and the host seems to lower the patient’s defense mechanism. The combination of metal particles, polyethylene and polymethyl-methacrylate (PMMA) reduces the destructive and chemotactic activity of polymorphonucleocytes.



The most pervasive pathogenic germs in infected TKA’s are gram-positive cocci. In several studies, the pathogens causing more than 50% of the infections were Staphylococcus aureus and Staphylococcus epidermidis

. Other, less frequent pathogens can also be found, such as Gram-negative bacilli, streptococci, coagulase-negative staphylococci, anaerobic Gram-positive cocci, candida and mixed flora. Once the infection develops, bacteria commonly form a glycocalyx film on the implant. This layer provides a provides a protected environment for the bacteria, decreasing their exposure to host immune system components and antibiotics.

Classification of Periprosthetic Infections

Several classifications of infections have been proposed. The original classification of deep infection after total hip reported by Coventry

has been expanded to four categories and modified according to current treatment guidelines

Type I

This group includes patients with positive intraoperative cultures found at the time of revision for presumed aseptic failure.

Type II

Type II infections, or early postoperative infections, are thought to be seeded intraoperatively, although can also be caused by hematogenous spread in the early postoperative period. By definition, these infections are diagnosed within 28 days of the index procedure.

Type III

Type III infections, or acute hematogenous infections, are characterized by an acute presentation in a previously well-functioning joint replacement and may be associated with a documented or suspected bacteremia. By definition they are less than 28 days in duration.

Type IV

Type IV infections, or late chronic infections, are those that present indolently, 28 days or more after the index arthroplasty. They are usually low-grade infections thought to originate perioperatively. Late chronic infections also include missed acute infections (early postoperative or hematogenous) that are now greater than 28 days in duration.


History, physical exam, laboratory data and imaging studies must all be considered in the decision making process.  In addition to cultures, the most commonly used laboratory tests include serum inflammatory markers and synovial fluid cytology.


Plain films

The appearance of rapidly progressive radiolucent lines surrounding an implant may be present during an infection. The resorption of subchondral bone and patchy osteoporosis can also be elements of suspicion.

Bone scan

A bone scan can help us confirm a diagnosis. However its high cost and its inability to provide acceptable levels of sensitivity and specificity have restricted its use.

Laboratory tests

Serum white blood cell count

WBC levels are typically normal. In a study of 73 infected knees, only 28% had counts over 11,000, the average was 8,300.

Serum inflammatory markers – ESR & CRP

ESR and CRP are valuable markers for both diagnosing and monitoring periprosthetic infection. Greidanus et al have recently reported optimal ESR and CRP cutoff values of 22.5mm/hr and 13.5mg/L, respectively, in their review of 151 knees presenting for revision surgery. Using these values in combination produced a sensitivity of 88%, specificity of 93%, positive predictive value of 84% and negative predictive value of 95%

Serum inflammatory markers – Interleukin-6

In a prospective, case-control study of 58 patients undergoing revision surgery of total hip and knee arthroplasties, serum Interleukin-6 values >10 pg/mL was reported to have a sensitivity of 100%, specificity of 95%, positive predictive value of 89%, negative predictive value of 100% and accuracy of 97%

Synovial inflammatory markers – CRP

At the 2011 AAOS Annual Meeting, Parvizi et all presented a clinical study (Click here to read abstract) examining the accuracy of synovial CRP concentrations in diagnosing periprosthetic infection. The study population included 53 patients undergoing revision total knee arthroplasty, 17 of whom were infected. The area under the curve was 0.965 for the individual ELISA, 0.931 for the multiplex assay and 0.872 for the serum CRP assay. Sensitivity and specificity were 87.5% and 87.5% for the individual ELISA, 93.3% and 88.9% for the multiplex assay and 80% and 89.7% for the serum CRP assay ?(1).

Knee aspirate cell count and differential

Synovial fluid cell count and differential is a very useful diagnostic test. False negatives are not uncommon in patients who received antibiotics prior to aspiration. For this reason antibiotics should be suspended, if possible, for 10 to 14 days before carrying out the aspiration. Ghanem et al have suggested a leukocyte threshold value of >1100 cells/microliter and a neutrophil percentage of >64% for the diagnosis of periprosthetic knee infection. Using these values in combination yields sensitivity of 85.0%, specificity of 99.2%, positive predictive value of 98.6% and negative predictive value of 91.6%

A retrospective study presented at the 2010 AJBS reported that the leukocyte cutoff value should be higher for patients with acute postoperative (Type II) infections (this group defined acute infections as those less than 42 days in duration). Receiver operator curves (ROC) identified an optimal threshold synovial fluid WBC count of 27,750 WBC/uL with a sensitivity of 84%, specificity of 100%, positive predictive value (PPV) of 100%, and negative predictive value (NPV) of 97%; the area under the curve was 97%. The optimal cut-off point for the differential was 89% PMN and for the CRP was 95 mg/dL (3).

Traumatic aspirations will result in falsely evevated leukocyte counts. In the case of a blood-tinged aspiration, the following formulas can be used to arrive at a corrected neutrophil count

Expected WBC = (WBC blood  ÷  RBC blood) x RBC synovial fluid

Adjusted WBC = observed WBC – expected WBC

Synovial leukocyte esterase

The technique of placing a drop of synovial fluid on the leukocyte esterase assay strip (e.g. similar to that found on a urine dipstick) was introduced at the AAOS 2010 Annual Meeting by Parvizi et al (click here to read abstract). If both ++ and + patients were considered positive, the test was 100% sensitive and 83.9% specific; if only ++ patients were considered positive, the test was 93.8% sensitive and 100% specific (2).

Polymerase chain reaction (PCR)

This method is used to detect and amplify the presence of bacterial DNA. It is thought to be a quick method since it is not affected by whether the patient takes antibiotics or not. However, a high percentage of false positives has been detected, which might have been caused by any type of contamination. In short, it is a technique that can be used as a complement to the previous ones and whose usefulness may increase in the future.

Intraoperative Frozen Section

The analysis of frozen histological sections is a valuable tool for diagnosing infection. It most often used to assist decision-making in cases with equivocal serum inflammatory makers and aspirate cytology. The cutoff value of >5 neutrophils per high power field is most commonly used for the diagnosis of infection

Intraoperative Gram Stain

This modality is unreliable (sensitivity = 27%) and should not be used routinely

Management of TKA infection

We have several options when it comes to managing an infected TKA. But before we select any of these, we must take into consideration a series of factors. These factors include the amount of time elapsed from infection, host-related factors, condition of the soft tissues, condition of the implant, type of microorganism present and its degree of sensitivity and, last but not least, the patient’s expectations and functional needs.

Planning for any one option requires having detailed clinical records, cultures, x-rays and information of previously received treatment. It is important to identify high-risk patients, i.e. those receiving immunosuppressive treatment or suffering from malnutrition or systemic disease, trying to improve their general condition as much as possible before surgery. Physical examination should provide information about the patient’s neurovascular situation, their articular mobility, the condition of their extensor mechanism and their soft tissues as well as about any previous incisions or the need of skin coverage by a plastic surgeon. The final goal of treatment is to eradicate infection, ease the pain and preserve the limb’s function.

Antibiotic suppressive therapy

Infection eradication with antibiotic therapy only is highly unlikely due to the presence of bacterial biofilm, therefore its use should be restricted to specific circumstances. Indications for this type of treatment are as follows: 1) High operative risk due to medical comorbidities; 2) Presence of low-virulence microorganisms susceptible oral antibiotics that can be tolerated by the patient; 3) Mechanically stable prosthesis.

Antibiotic treatment should follow 3 basic principles: 1) Use of antibiotics of proven intracellular efficacy (rifampicin, fluorated quinolones, quatrimoxazole, clindamycin, phosphomycin, macrolides (azitromicine, clarithromycin), and the new antistaphylococcal agents (linezolid, quinupristin-dalfopristin). 2) Antibiotics should be combined, using a minimum of two to enhance the possibility of therapeutic success. 3) Long-standing administration, i.e. treatment should last a minimum of 6 months.

It would seem that the use of new antibiotics could improve results for resistant bacteria. The oxazolidinone linezolid is a new wide-spectrum antibiotic with very attractive pharmacokinetic and activity profiles. It is an antibiotic that acts against methicillin-resistant staphylococci and vancomycin-resistant enterococci. Its oral presentation makes its administration very convenient, reaching serum levels similar to those of parenteral administration. Its most significant – albeit reversible – side effect is myelosuppression. The usual dose is 600 mg BID for an average of 6 weeks. The studies available show that linezolid subsequent to an appropriate surgical treatment could be a good method to treat osteomyelitis and prosthetic infections.

Debridement and prosthesis retention

If diagnosed early, it is possible to successfully treat acute postoperative and acute hematogenous infections with debridement and retention of the prosthesis. Prerequisites include: 1) Short duration of symptoms (<2-4 weeks); 2) Mechanically stable components; 3) Organism susceptible to antibiotics that can be tolerated by the patient; 4) Absence of excessive scar tissue from prior surgical procedures. The average historical success rate of retention debridement for acute periprosthetic knee infections is approximately 55%. The use of rifampin-combination antibiotic therapy, antibiotic beads, and serial debridements have shown improved treatment results for the treatment of Staphylococcal infections.

Surgical treatment may be carried out in the form of a single open debridement, multiple debridements pending serial culture results

, or as a two-stage debridement with the use of antibiotic-impregnated cement beads. The surgical technique

for the latter consists of performing a wide-exposure arthrotomy so as to access every area in the joint. Subsequently, the PE insert is removed, which will make it possible to subluxate the joint and access the posterior area.  The PE insert is either flashed in the autoclave or scrubbed with betadine. A synovectomy must be performed, making an effort to make it as complete as possible, and multiple samples should be taken for the pathological study. The implants should be carefully inspected and their degree of fixation assessed. Retained components are scrubbed with betadine or castile soap. Antibiotic-impregnated bone cement beads can be implanted, using vancomycin combined with tobramycin for wide-spectrum coverage. Sterilized modular components are replaced. If beads are used, the wound is then closed and immobilized with a bandage. A drain is typically employed. After a week of IV treatment, and once the microorganism has been identified, a new procedure is carried out to remove the cement beads and to perform a new debridement. Fresh samples must be taken and a new PE insert implanted. IV antibiotics should be administered over the next 4-6 weeks. Patients may then be put on a variable course of oral antibiotics, dependent on associated comorbidities.

Prosthesis exchange

Prosthetic revision surgery can be performed in a single stage or in two surgical stages, i.e. the old prosthesis is removed in a first procedure, and the new one is implanted in a second one.

Single-stage exchange

Indications for a single-stage exchange are not clearly defined. Most authors seem to agree that if the infection has been present less than 28 days, the microorganism is sensitive to antibiotic treatment, the soft tissues are in appropriate condition and the patient’s general health is good, then a single-stage surgery can be considered. A meta-analysis published in 2002 reported successful treatment in 33/37 (89%) knees using this technique

. Single-stage debridement and component exchange is used much more commonly in Europe than in the U.S.

Two-stage exchange

Two-stage exchange remains the gold standard for the treatment of infected knee prostheses. The first surgical stage entails removal of all components, an aggressive debridement (including total synovectomy and removal of all cement), and the placement of a high-dose antibiotic-imprgnated cement spacer. After the stage-I procedure the patient receives a course of antibiotic therapy, most commonly a six-week course of IV antibiotics. Antibiotics are then withdrawn for a period of time and the serum inflammatory markers are monitored. Typically patients are taken back to the operating room for the Stage II procedurewo to three months after the Stage-I procedure, if serum inflammatory markers and knee aspirate cell count have normalized while off antibiotics (i.e. the infection appears eradicated). At the stage-II procedure the cement spacer is removed, a repeat debridement is performed, and a new prosthesis is placed.

In methicillin-resistant germ infections, whose prevalence has significantly increased in the last decade, orthopedic surgeons should follow the current treatment guidelines. In a recent study, Volin et al.

compared prosthetic infections from methicillin resistant and non-resistant bacteria that received the same type of treatment, i.e. two-stage replacement. This author reported a 94% success rate in non-resistant bacterial infections vs. 88.9% for resistant bacteria, although difference was not statistically significant. The study concludes that two-stage replacement is a successful treatment both for methicillin-sensitive and methicillin-resistant organisms.

Enteral and parenteral antibiotic therapy used in conjuntion with surgical treatment

Improved results have been reported with the use of rifampin-combination antibiotic therapy for the treatment of implant-associated staphylococcal infections

including a randomized, controlled trial which has shown rifampin-combination therapy to be superior to treatment with vancomycin or ciprofloxacin alone for the treatment of acute, implant-associated staphylococcal infections

Use of antibiotic-loaded cement spacers

Cement beads or spacers are used in order to prevent soft tissue contractures. Additionally, high-dose antibiotic-impregnated PMMA spacers release very high concentrations of antibiotic locally, while systemic levels remain low

. Elution characteristics depend upon the cement type, the characteristics of the antibiotic, the antibiotic dose used, and the the porosity and surface area of the cement. Numerous studies have examined the use of various fillers (or “poragens”) in order to improve elution characteristics.

The PMMA polymer and antibiotc powder should be mixed before the monomer is added in order to improve antibiotic elution characteristics. A study of antibiotic elution concentrations from cement spacers in vivo reported highest elution concentrations when at least 3.6 grams of tobramycin and 1 gram of vancomycin were used per 40 gram batch of cement

. Some recommend using 3.6 grams of tobramycin, 2-3 grams of vancomycin, and 2-3 grams of cefazolin per 40 gram batch of cement for most bacterial infections (cefazolin acts as a porogen, increasing the elution of the other antibiotics).

Ever since spacers came to be used in the 80’s, results have improved dramatically, with some series achieving 90% of successful results

. Although results are excellent, the spacer block-based system does have its drawbacks, mainly related to wound-healing and joint mobility problems that make the posterior approach difficult.

Articulating Spacers

For that reason, articulated spacers have been designed that allow the knee to move. They are of two types: the first type is entirely made of cement; the second attempts to replicate a knee prosthesis using both polyethylene and metal (PROSTALAC system). The goal of these systems is to permit mobility and partial weight-bearing in the period between surgeries, preserving soft tissue tension and articular stability

. An articulating spacer can be fashioned by hand, molded with vaccuform molds simply made by molding trial implants, commercially available disposable molds, premade antibiotic laden implants, or the PROSTALAC system which has yet to receive FDA approval hence is not readily available. The functional results of an articulating spacer provide significant advantages to both the patient and the surgeon making the subsequent second stage easier because of retention of function, range of motion and the absence of scar tissue seen with a static spacer as demonstrated by this patient. There are other articulating space options available, molds to fashion a cement on cement articulation made by the surgeon in the operating room or premade antibiotic impregnated implants are available off the shelf.  Today articulating spacers are the preferred method, static spacers are used in situation where articulating spacers are not technically possible, for example in patients with a failed distal femoral replacement or a significant bony defect around the knee as demonstrated by this patient with an infected TKA and an infected supracondylar non-union

Vaccuform Femoral Mold Vaccuform Tibial Mold

These molds were made by occupational therapist using a vaccuform mold technique also used to make simple splints and orthotics.  They were molded from a set of trial implants.  They are gas sterilized and reusable.

All cement femur (Palacos) All cement tibia unmolded (Simplex)
Cement on cement

Cement on cement articulations are certainly an improvement over simple spacer blocks.  The patients find these more uncomfortable than a metal on plastic articulation.  They tend to be “sticky”

The PROSTALAC knee system is a metal on polyethylene device.  It allows for a smooth articulation and the results are comparable to other spacer techniques

The advantages of an articulating spacer is that it allows for the maintenance of motion, it decreases scar tissue, and is often functional enough that the second stage can be delayed for months and the second stage reimplantation is greatly facilitated when the joint is not stiff

Knee arthrodesis

This method is used in few cases as first-line treatment. In general, arthrodesis can be considered a therapeutic option when some other of the techniques described above have failed, especially in young patients with high functional demands or in patients with large deformities, alterations of the extensor mechanism, damaged soft tissues, immunosuppression or infections by extra virulent bacteria.

Arthrodesis provides a stable and pain-free joint. However, the function of the knee is sacrificed, with all that this means for the carrying out of day-to-day activities. Note should be taken of whether other joints (the ipsilateral hip or ankle or the contralateral knee) are involved and of whether the contralateral leg has been amputated. Such circumstances will be contraindications for arthrodesis.

Resection arthroplasty

The purpose of this technique is to create a false joint that may allow a certain range of motion. The procedure consists in explanting the components and, after a wide debridement followed by antibiotic treatment the area is immobilized between 3 and 6 months. The purpose is to get the soft tissues to retract so that they can provide the area with a certain degree of stability. Candidates for this type of treatment are patients with polyarticular disease and few functional demands.

Limb amputation

This technique should be considered a last resort when treating a prosthetic infection. Its indications are as follows: an uncontrollable infection that threatens the patient’s life, massive bone defects and severe soft tissue loss. Functional results tend to be extremely poor and patients often end up in a wheelchair.


Two-stage revision arthroplasty is currently the gold standard for treating the infected knee prosthesis. Articulated spacers seem to be better than static ones. The survival rate free of implant removal for any reason is 90% at 5 years and 77% at 10 years.



1. Parvizi J. CRP in joints: A Molecular Marker for Periprosthetic Joint Infection? 2011. Available at:

2. Parvizi J, Jacovides C, Azzam K, Antoci V, Ghanem E. Diagnosis of Periprosthetic Joint Infection: The Role of a Simple, yet Unrecognized, Enzyme. 2010. Available at:
3. Saxena A, Jacovides C, Siad J, Ghanem E, Ketonis C, Azzam K, Kelly C, Parvizi J. Cell count and differential of aspirated fluid in acute potoperative period total knee arthroplasty. Presented at the Musculoskeletal Infection Society 19th Annual Open Scientific Meeting. San Diego, CA. August 2009.


02020025.JPG (image/jpeg)

all cement tibia gross.jpg (image/jpeg)

AP xray.jpg (image/jpeg)

cement on cement.jpg (image/jpeg)

both components unmolded.jpg (image/jpeg)

Dowkes.jpg (image/jpeg)

gross ap in place green.jpg (image/jpeg)

metal on plastic xray.jpg (image/jpeg)

Gross photo knee flexion.jpg (image/jpeg)

clinical full extension.jpg (image/jpeg)

poly femur mold.JPG (image/jpeg)

poly tibial mold.JPG (image/jpeg)

Prosthesis unmolded and assembled.JPG (image/jpeg)

tibia unmolded.JPG (image/jpeg)

AAHKS image.gif (image/gif)

AAHKS image.gif (image/gif)

AAHKS image.gif (image/gif)

AAHKS image.jpg (image/jpeg)

AAHKS.gif (image/gif)

poly femur mold.JPG (image/jpeg)

poly tibial mold.JPG (image/jpeg)

02020025.JPG (image/jpeg)

tibia unmolded.JPG (image/jpeg)

cement on cement.jpg (image/jpeg)

cement on cement.jpg (image/jpeg)