TKA is a successful surgery with over 90% 10 year survival, 80% 20 year survival.  Yet complications occur in up to 10% of patients.  We will review some of the most common issues affecting TKA.

1. TKA Pain

Pain is a common symptom associated with many complications following TKA. Identifying the cause is challenging because many are inter-related and self-reinforcing, for example, stiffness and knee pain can be the cause and the result of the other.  Working through the web of differential diagnosis and cause and effect is critical because performing revision TKA to treat pain without an underlying cause has only 40% improvement. 

The timing of pain is critical for diagnosis. 

A knee that is painful since surgery is associated with instability (such as component malalignment or unbalanced gaps), acute infection, soft tissue impingement (from component malpositioning or incorrect sizing), or from an external source (such as the ispilateral hip or back, which is hopefully a secondary source of pain, and not the true pain misdiagnosed as an arthritic knee). 

A knee that initially improved and then became painful is associated with component loosening or infection.  Additional information, such as start up pain suggest component loosening while continuous pain or night pain suggests infection, tumor, or ARDS.  A recent history of trauma may also make one think of loosening or periprosthetic fracture. 

Physical examination offers other insights.  Point tenderness can indicate an area of impingement (such as pes bursa inflammation from medial overhang of the tibial component), or soft tissue reaction from retained cement.  Point tenderness over the medial metaphysis (at least 1 cm below the joint line) can also indicate component loosening or stress fracture.  Joint effusion is very nonspecific and results from hemarthrosis, synovitis, ligament instability, poly wear, infection…you get the idea).  Evaluating TKA stability and patellar tracking helps to identify issues with malpositioning. 

Laboratory and radiographic work up should always be part of the evaluation for TKA related pain.  Inflammatory markers should be ordered in most cases unless there is an obvious etiology for pain. The ESR and CRP are the standard initial tests to evaluate for infection. AP and Lateral X-ray will provide information about component positioning (see post-op x-ray evaluation), loosening, and periprosthetic fracture.  Comparison to prior x-rays is helpful when looking at osteolysis and loosening.  A high quality AP and Lateral x-ray is critical to evaluate component positioning and alignment. Further imaging studies such as CT, MRI, and bone scan can be considered under certain circumstances and will be discussed in the followingsections.  

2. Periprosthetic Joint Infection

Periprosthetic Joint Infection (PJI) is a cause of significant morbidity and mortality.  The preoperative patient optimization protocols (see preop optimization) were created in a large degree to minimize infection risk due to its correlation with obesity, diabetes, metabolic syndrome, malnutrition, smoking, and s.aureus colonization. While PJI in THA (0.3 – 1.3% primary, and 3% in revision THA) is less common than TKA (1 – 2% primary, and 6% in revision TKA), it remains an important complication.  Additionally, PJI should always be a part of the differential diagnosis when evaluating postop patients for pain, loosening, instability or even periprosthetic fracture. 

What is a PJI?  This seems obvious: the prosthesis is infected with a microorganism. Yet diagnosis of PJI is far more challenging in reality because positive cultures are not a reliable means of diagnosis, with reports suggesting that cultures are only 60% sensitive. There are other complicating factors.  For example, what if you see “gross purulence” around the implant, can you just say that its infected?  The answer is that “gross purulence” alone (with other tests negative) is not enough to diagnose an infection because many of the MoM soft tissue reactions appear very similar to “gross purulence” and osteolysis from poly wear can also appear as purulence, an thus will fool a surgeon’s into treating an infection.  So if you cannot trust your eyes and you cannot trust cultures, what can you hang your hat on to say that a joint replacement is infected?

The Musculoskeletal Infection Society (MSIS) Diagnostic Criteria was developed to formalize the diagnosis process [75], It offers a good foundation for PJI diagnosis.  Yet controversy remains even with this algorithm, specifically regarding what cutoffs should be used for certain lab values (cutoff values are essentially a compromise between sensitivity and specificity and there are no absolutes, outliers always exist).

Infection is diagnosed as 1 major criteria (either sinus tract or 2 cultures of the same bacteria), or 3 out of 5 minor criteria (elevated ESR/CRP, elevated synovial cell count or Leukocyte Esterase, elevated PMN%, one culture, positive histology).


Identifying symptom duration and time from index procedure are two critical forks in the diagnostic and treatment pathways [76].  PJIs are best understood by separating them at these time points.  First see what category the patient get sectioned into and then work them up accordingly. For example, an early infection (surgery < 4 weeks prior) compared to a late infection, the CRP cut off jumps from 10 to 95 mg/L, while cell count jumps from 3,000 to 10,000, and PMN% increases from 60-70% to over 80%.  This is why it is critical to know the date of surgery before interpreting any results. 

When patient presents with concern for infection, knee swelling, fevers, pain, erythema, drainage, etc.  The first step is to obtain blood work including ESR, CRP.  If inflammatory markers are elevated, then obtain aspiration.  If inflammatory markers are not elevated, but there is notable concern based on the h&p, then aspirate as well (remember the blood work is only about 90% sensitive).  In cases of acute onset of symptoms, it is believed that the bacteria is limited to the joint fluid, while chronic symptoms suggest the bacteria has had time to adhere to the prosthesis (biofilm) and invade the interface between bone and implant.

ACUTE EARLY.  Acute postoperative infection is onset of symptoms 4 weeks from the index procedure. The distinction of an early infection is important because the cutoff values for many of the tests used to diagnose a PJI change in the early postoperative period.  Synovial leukocyte levels do not normalize for about 6 weeks, and therefore WBC cell counts for aspiration are elevated at baseline.  Furthermore, the systemic inflammatory markers are elevated.  It takes around 3 weeks for CRP to normalize, and over 6 months for ESR to normalize.

Diagnostic values.  Elevated CRP in the early postoperative period is > 95 mg/mL.  Aspiration cell count over 10,000 and 89% PMN are considered elevated.

Treatment see below.

LATEOnset of symptoms more than 4 weeks from the index procedure, but usually it occurs years after surgery.   A late infection occurs when the inflammatory phase of the primary surgery has resolved and is no longer a confounding variable when interpreting the lab results. The lab values can also be used in patients with inflammatory arthritis. Acute Late (Hematogenous).  This is acute onset (< 3 days) of symptoms long after the index procedure (> 4 weeks). Patients seen in the office after acute onset of symptoms need to be worked up quickly to maximize the benefit of surgical intervention if they are showing signs of PJI. Acute onset of symptoms suggest the bacteria has not formed a biofilm, although bacteria form the biofilm at different rates (with all bacteria forming biofilm by 4 weeks).

Diagnostic. Blood work: CRP > 10 mg/L, ESR > 30. Aspiration: WBC count > 1,700; PMN > 80%

Chronic LateThis is gradual onset of symptoms (> 3 days), long after the index procedure (> 3 weeks).  In the case of chronic symptoms it is believed that bacteria has formed a biofilm (all bacteria form biofilm by 4 weeks).  Biofilm is a layer 15% cells and 85% glycocalyx (formal name: exopolysaccharide glyocalyx) that makes the infection 1,000 – 1.5k more resistant to antibiotics. Furthermore, there is no reliable way to remove biofilm once formed.  Thus, chronic infections require removal of the infected implants


Inflammatory markers. The ESR is an acceptable marker for late infections, however, it cannot be used in the immediate postop period (“acute early” infections) because ESR requires up to 6 months to return to normal.  It is therefore falsely elevated.  CRP in contrast returns to normal around 3 weeks, and can often be used in the decision making algorithm [77].  Even in possible late PJI, the ESR is only a questionably valuable marker, and therefore, it must be positive in conjunction with the CRP to achieve a sensitivity and specificity over 90%.

ESR and CRP levels are important precursors to joint aspiration for a few reasons.  ESR and CRP are highly sensitive and therefore, if they fall within a normal range, its ok to stop the work up for infection unless highly suspicious. There is the risk of introducing bacteria into a THA with aspiration and therefore, every patient with a fever and a THA should not get an aspiration.  Additionally, no test is perfect and some aspirations can be falsely positive.  However, aspirations that are preceded by inflammatory markers influences the positive predictive value and thus reduces the risk of unnecessary major surgery [78].

Aspiration. What is a significant cell count upon aspiration.  This remains controversial.  A WBC count suggestive of infection is considerably lower than for a native knee because there is less synovial lining for neutrophils to penetrate the joint.  Studies have suggested a cell count > 3,000 is indicative of a late infection [79]. Other studies suggest an even more sensitive cutoff of 1,700 cells, with anything over 65% PMNs [80].  These values however can only be used in late infections (> 6 weeks from index procedure) because synovial leukocyte levels do not normalize until then.  Therefore in Early infections the recommended cutoff is a cell count of 27,800, 89% PMN (and CRP > 95 mg/mL) [81]

Notice that gram stain is not part of the MSIS criteria.  The sensitivity is too low to be helpful and should not be included in the work up [82].

Intraoperative culture: In cases of revision surgery, a single positive culture is insufficient to diagnose an infection.  These findings should be used in conjunction with other tests for diagnosis.


The goal of treatment is eradication.  This goal is challenging because bacteria form a biologic matrix around the hardware components that prevents antibiotics from reaching the bacteria.  The duration of infection (time since symptom onset) and type of bacteria both determine how advanced this glycocalyx matrix has become, and thus whether the components need to be removed.  The success of differing treatments depends on the type of bacteria and the duration of infection.  Lets look at treatment for each of PJI groups.

Acute Early Infection.  Consider I&D with poly exchange, followed by 6 weeks of IV antibiotics [83].  Studies suggest a 50% cure rate in the acute period.  However, there is a high failure rate with MRSA [84]reported around 85% failure, and therefore, 2-stage revision should be considered based on bacteria[85, 86]. 

Acute Hematogenous Infection.  Approached the same as an acute early infection (due to similar impact of bacteria and timing on matrix formation). It is best to prevent late infections by giving antibiotics before dental procedures (although the correlation between dental work and acute hematogenous infection is unclear because the organisms cultured in the synovium are rarely the same ones commonly found in the mouth).  [87]

Chronic late infection. The standard treatment is a 2-stage exchange.  The emphasis in the first stage is removing all infected material, performing an extensive debridement, opening the tibial and femoral canals, and placing an antibiotic spacer [88].  The second stage is implanting hardware that offers stable, functional knee.  The cure rate is about 80-95% (depending in part on the organism). 

STAGE 1. Dr. Duncan performed a lot of the primary groundwork investigating the elution of antibiotics in cement [64, 89-95]. Currently there is significant variability in the literature with regards to antibiotic spacer dosing, but these incremental changes derive from the initial work by Duncan et al.

The antibiotic spacer contains variable types and amounts of antibiotic.  The standard is 3-6 g of vancomycin and 1.2 g of tobramycin per package of cement.  The combination increases the rate of elution into the knee.  Thicker cement, like palacos, elutes the abx more rapidly, and thus it creates a higher concentration of antibiotic in the joint, and also in the blood stream and it therefore must be monitored closely.  Simplex does not elute as well and therefore it can have higher concentration of abx without worrying about toxic levels.  Spacers also come in two forms: static vs. dynamic. Static spacers should be held in place with a styman pin that goes up the femoral and tibial canals to prevent the spacer from being extruded from the joint and eroding through soft tissue, such as the patellar tendon.  The more popular type is a dynamic spacer.  This has the advantage of allowing for knee ROM to prevent stiffness once replant occurs.   

The spacer remains in place for 6 - 8 weeks.  Following Stage 1, the patient receives IV antibiotics for 6 weeks. Serial ESR/CRP is performed and should trend down (but often fails to normalize) [96].  

A repeat knee aspiration is performed around 8 weeks (after 2 weeks off antibiotics), and a cell count of 3,000 should be utilized as a cutoff for response to intervention.   

STAGE 2: Reimplantation of TKA.  Often times a more contrained design is required due to soft tissue attenuation or destruction during the process of irradicating the infection, and the bone loss associated with removing implants. 

One-stage exchange for chonic late infections is commonly used in Europe, but is not the standard in the United States.  In this technique, the components are explanted and a vigorous I&D is performed (read: significant tissue debridement, almost skeletonizing the remaining bone: collateral ligaments are often resected, and the revision TKA is a rotating hinge). The patient is then re-prepped, re-draped, and new components are placed.  In these cases, the bacteria is always known preoperatively, its typically not performed for MRSA infections, and it is performed in generally healthier patients.

Avoid I&D in chronic late infections due to high failure rate and added damage to the soft tissue envelope with multiple I&Ds, which may risk outcomes for 2-stage [97].  One study found worse outcomes in patients with failed I&D who then required a 2-stage procedure, suggesting that either an I&D independently impedes recovery from further surgical procedures or it merely selects patients with a more virulent organism (already resistant to I&D) who are therefore more likely to fail 2-stage. 

The entire process of revision surgery with abx spacer and final replant takes its toll on the patient.  Revision for TJA has higher mortality rate than other revisions [98]Overall, the results for revision TJA for infection isn’t terrific, but functional spacers have improved results. [99]

JAAOS Review Infection TKA:[103][104]


Infections are clearly very challenging and thus the key is prevention.  Optimize patients.  Take precautions in the OR to minimize risk.  Give preoperative antibiotics.  Number needed to treat to prevent 1 infection is only about 50 patients. [105] About 10% of Americans believe they are allergic to penicillin, while only 10% of those patients have any true reaction, and most can safely receive any beta-lactam antibiotic [106].  This is an important fact because Ancef (cefzolin) has a proven track record of offering good coverage [107]. Drug distribution in obese patients decreases due to the greater volume of tissue.  It may lead to unacceptably low blood levels, and thus doses should be increased in the obese patient.  High doses of ancef can be administered rapidly, however, vancomycin must be infused slowly (90 – 120 minutes) to avoid the potential complication of red man syndrome, and thus may complicate the timing of a surgical day.  The question of whether vancomycin should be used in high risk patients? 


1.         Bozic, K.J., et al., Patient-related risk factors for postoperative mortality and periprosthetic joint infection in medicare patients undergoing TKA. Clin Orthop Relat Res, 2012. 470(1): p. 130-7.

2.         Bozic, K.J., et al., The epidemiology of revision total hip arthroplasty in the United States. J Bone Joint Surg Am, 2009. 91(1): p. 128-33.

3.         Parvizi, J., et al., New definition for periprosthetic joint infection: from the Workgroup of the Musculoskeletal Infection Society. Clin Orthop Relat Res, 2011. 469(11): p. 2992-4.

4.         Fehring, T.K., et al., Early failures in total knee arthroplasty. Clin Orthop Relat Res, 2001(392): p. 315-8.

5.         Aalto, K., et al., Changes in erythrocyte sedimentation rate and C-reactive protein after total hip arthroplasty. Clin Orthop Relat Res, 1984(184): p. 118-20.

6.         Barrack, R.L. and W.H. Harris, The value of aspiration of the hip joint before revision total hip arthroplasty. J Bone Joint Surg Am, 1993. 75(1): p. 66-76.

7.         Schinsky, M.F., et al., Perioperative testing for joint infection in patients undergoing revision total hip arthroplasty. J Bone Joint Surg Am, 2008. 90(9): p. 1869-75.

8.         Trampuz, A., et al., Synovial fluid leukocyte count and differential for the diagnosis of prosthetic knee infection. Am J Med, 2004. 117(8): p. 556-62.

9.         Bedair, H., et al., The Mark Coventry Award: diagnosis of early postoperative TKA infection using synovial fluid analysis. Clin Orthop Relat Res, 2011. 469(1): p. 34-40.

10.       Ghanem, E., et al., Periprosthetic infection: where do we stand with regard to Gram stain? Acta Orthop, 2009. 80(1): p. 37-40.

11.       Segawa, H., et al., Infection after total knee arthroplasty. A retrospective study of the treatment of eighty-one infections. J Bone Joint Surg Am, 1999. 81(10): p. 1434-45.

12.       Bradbury, T., et al., The fate of acute methicillin-resistant Staphylococcus aureus periprosthetic knee infections treated by open debridement and retention of components. J Arthroplasty, 2009. 24(6 Suppl): p. 101-4.

13.       Deirmengian, C., et al., Open debridement of acute gram-positive infections after total knee arthroplasty. Clin Orthop Relat Res, 2003(416): p. 129-34.

14.       Deirmengian, C., et al., Limited success with open debridement and retention of components in the treatment of acute Staphylococcus aureus infections after total knee arthroplasty. J Arthroplasty, 2003. 18(7 Suppl 1): p. 22-6.

15.       Chen, A., et al., Prevention of late PJI. J Orthop Res, 2014. 32 Suppl 1: p. S158-71.

16.       Burnett, R.S., et al., Technique and timing of two-stage exchange for infection in TKA. Clin Orthop Relat Res, 2007. 464: p. 164-78.

17.       Wentworth, S.J., et al., Hip prosthesis of antibiotic-loaded acrylic cement for the treatment of infections following total hip arthroplasty. J Bone Joint Surg Am, 2002. 84-A Suppl 2: p. 123-8.

18.       Penner, M.J., B.A. Masri, and C.P. Duncan, Elution characteristics of vancomycin and tobramycin combined in acrylic bone-cement. J Arthroplasty, 1996. 11(8): p. 939-44.

19.       Masri, B.A., et al., Effect of varying surface patterns on antibiotic elution from antibiotic-loaded bone cement. J Arthroplasty, 1995. 10(4): p. 453-9.

20.       Kendall, R.W., C.P. Duncan, and C.P. Beauchamp, Bacterial growth on antibiotic-loaded acrylic cement. A prospective in vivo retrieval study. J Arthroplasty, 1995. 10(6): p. 817-22.

21.       Jackson, J., et al., The use of bone cement for the localized, controlled release of the antibiotics vancomycin, linezolid, or fusidic acid: effect of additives on drug release rates and mechanical strength. Drug Deliv Transl Res, 2011. 1(2): p. 121-31.

22.       Haddad, F.S., et al., The PROSTALAC functional spacer in two-stage revision for infected knee replacements. Prosthesis of antibiotic-loaded acrylic cement. J Bone Joint Surg Br, 2000. 82(6): p. 807-12.

23.       Duncan, C.P. and B.A. Masri, The role of antibiotic-loaded cement in the treatment of an infection after a hip replacement. Instr Course Lect, 1995. 44: p. 305-13.

24.       Brady, O.H., et al., The reliability and validity of the Vancouver classification of femoral fractures after hip replacement. J Arthroplasty, 2000. 15(1): p. 59-62.

25.       Ghanem, E., et al., Staged revision for knee arthroplasty infection: what is the role of serologic tests before reimplantation? Clin Orthop Relat Res, 2009. 467(7): p. 1699-705.

26.       Sherrell, J.C., et al., The Chitranjan Ranawat Award: fate of two-stage reimplantation after failed irrigation and debridement for periprosthetic knee infection. Clin Orthop Relat Res, 2011. 469(1): p. 18-25.

27.       Berend, K.R., et al., Two-stage treatment of hip periprosthetic joint infection is associated with a high rate of infection control but high mortality. Clin Orthop Relat Res, 2013. 471(2): p. 510-8.

28.       De Man, F.H., et al., Infectiological, functional, and radiographic outcome after revision for prosthetic hip infection according to a strict algorithm. Acta Orthop, 2011. 82(1): p. 27-34.

29.       Emerson, R.H., Jr., et al., Comparison of a static with a mobile spacer in total knee infection. Clin Orthop Relat Res, 2002(404): p. 132-8.

30.       Willis-Owen, C.A., A. Konyves, and D.K. Martin, Factors affecting the incidence of infection in hip and knee replacement: an analysis of 5277 cases. J Bone Joint Surg Br, 2010. 92(8): p. 1128-33.

31.       Daines, B.K., D.A. Dennis, and S. Amann, Infection prevention in total knee arthroplasty. J Am Acad Orthop Surg, 2015. 23(6): p. 356-64.

32.       Shirwaiker, R.A., et al., A clinical perspective on musculoskeletal infection treatment strategies and challenges. J Am Acad Orthop Surg, 2015. 23 Suppl: p. S44-54.

33.       Hill, C., et al., Prophylactic cefazolin versus placebo in total hip replacement. Report of a multicentre double-blind randomised trial. Lancet, 1981. 1(8224): p. 795-6.

34.       Frumin, J. and J.C. Gallagher, Allergic cross-sensitivity between penicillin, carbapenem, and monobactam antibiotics: what are the chances? Ann Pharmacother, 2009. 43(2): p. 304-15.

35.       Ponce, B., et al., Surgical Site Infection After Arthroplasty: Comparative Effectiveness of Prophylactic Antibiotics: Do Surgical Care Improvement Project Guidelines Need to Be Updated? J Bone Joint Surg Am, 2014. 96(12): p. 970-977.

3. TKA Instability

Instability is associated with almost 50% of early revisions and 10-20% of all revisions. Instability is caused by a multitude of underlying issues.  Gap imbalance, coronal imbalance (failed soft tissue tensioning, or ligament rupture/attenuation), component malpositioning (ie internal rotation), extensor mechanism failure, or component loosening. 

Patients feel their knee “giving way” or “buckling”, particularly with stairs.  Patients may also complain of stiffness because of secondary guarding.  A locking sensation is more common with patellar maltracking.  A varus thrust (the knee buckles laterally upon heel strike) with lateral instability or a valgus thrust with medial instability.  It is important to rule out muscle weakness before homing in on instability.  Persistent quad weakness due to poor rehab or inability to fully extend the knee that prevents a patient from "locking-out" their knee while standing can appear similar to instability.  

The onset of instability is important for diagnosis.  Gap Imbalance, iatrogenic collateral ligament injury, and component malpositioning (size, internal rotation) present early postop, while aseptic loosening occurs late, as can attritional ligament instability (such as PCL rupture in a CR-TKA).  Extensor mechanism injury presents as acute onset of instability.  

Instability is more of a symptom than a diagnosis,and identifying the underlying cause is key to successful treatment. 

Gap IMBalancE & ligament instability

Bone cuts and ligament stability influence each other so it can be somewhat challenging to isolate one cause.  Examining the varus/valgus stability in 0°, 30°, and 90° can give insight about cause of instability.

Isolated collateral ligament insufficiency can be identified when either the medial or lateral side opens in flexion and extension. A CCK “contrained condylar knee” (the POST is wider and therefore has greater congruence with the CAM, thus limiting coronal motion) is used.  Most CCK implants require a larger box cut to accommodate the wider post and thus implant revision is needed.  Many surgeons also add stems to the CCK implant because as it is more constrained, it assumes greater stress during ambulation, and therefore transmits more force to the bone-cement interface.  Adding a stem will increase the surface area for this stress and decrease aseptic loosening.  However, some implants are “partially constrained” meaning the constrained post is wider than a standard post but not completely congruent with the cam and therefore there is some wiggle room (maybe 5° as opposed to only 1° of motion for the standard CCK)  , and therefore, only the poly liner needs to be exchanged.  The CCK offers stability in the coronal plane, but not in the saggital plane, therefore balanced flexion and extension gaps must be maintained.  The CCK is sometimes used with the superficial MCL is damaged during surgery (ie released off the tibia) and it acts as an internal splint, giving stability in the first 6 weeks until the soft tissues heal and then give further stability.  A hinge implant is used with complete ligament instability.  If the MCL is completely missing, a CCK will experience excessive stress over time and will eventually fail due to poly wear or fracture.  Therefore a hinge is recommended for cases of complete MCL deficiency.  Current hinge designs are a rotating platform, which adds some increased motion at the articular surface.  A fixed hinge transmits excessive force to the bone-cement interface and causes early failure.  Additionally there is a segmental type hinge for large bone defects versus a condylar-type hinge for use in cases of good bone stock and deficient soft tissue, which is preferred because of better load distribution. 

Mid-flexion instability is a recognized problem with a controversial underlying etiology.  The soft tissues in TKA are balanced during surgery at 0° and 90°, but think about the more complete arc of motion, and the potential for changes in soft tissue tensioning between these two points.  Some believe that stability should be checked not only at 0° and 90° but also at 30° and 60°.  Its possible that the greater articular congruence at 0° may mask ligament instability in early flexion, which can be identified by checking knee stability in more points of motion.  Other theories on mid-flexion instability see an association in TKA cases that had a normal extension gap and loose flexion gap, which were corrected by increasing distal femoral resection.  When such a bone cut is used to balance soft tissue tension mismatch at 0°, the soft tissue mismatch is only solved at 0° but persists in mid-flexion because the true problem was a loose flexion gap, not a tight extension gap.  

Midflexion instability is seen in people that never feel confident in their knee.  They report that it gives way rising from chair, going down stairs or with walking.  They often experience recurrent knee effusions, complain of anterior knee pain, and wear some type of knee brace.  The key to fixing is restoration of posterior offset.   This can be issue with anterior referencing guide.  The flexion gap is particularly at risk in valgus knees due to an attenuated MCL and bony erosion of the posterior condyle.  Its also a concern in PCL retaining knee (CR-TKA) due to changes in the tibial slope cut. Treatment: Revision surgery to restore the flexion gap typically with posterior condylar augments +/- distal femoral augments.

Patellar Instability & component malposition 

Typically caused by malrotation of components. For a complete explanation of component rotation, see the Basic Concepts Chapter, but in brief review, overly internally rotating or medializing the components will change the Q angle, thus increasing the lateral subluxing force. The Q angle is the direction of patellar tracking (which change based on component positioning) relative to the vector of pull from the quad muscle (remember the quad is fixed from a more proximal position so it does not change).  If the patella tracks in line with the pull of the quad, the Q angle is small (there is not conflict in trajectory).   Patients with patellar instability typically report a locking sensation or a deficient extensor mechanism (both can present as buckling) as well as vague anterior knee pain with intermittent swelling, patellar crepitus.  The best method to diagnose patellar maltracking is with CT scan.  While there is no numerical indication for revision surgery, some studies have correlated degree of Internal Rotation with risk for symptoms.  1-4° of IR cause lateral tracking and lateral patellar tilt; 3-8° cause subluxation, 7-17° may cause patellar dislocation or fracture.   Reports on early TKA designs cited patellar complications of 12-24%, yet developments in femoral component design deepened the trochlear groove, created asymmetric femoral condylar flanges (augmenting the lateral side) and graduated the anterior-to-distal transition.  Currently, reports of patellar issues are rare. 

Treatment: Identify malrotated component and revise if significant discomfort that fails to improve with physical therapy. Alternatively consider lateral release. 


see below.  

AsEptic Loosening

see below.


1. Azzam, K., et al., Revision of the unstable total knee arthroplasty: outcome predictors. J Arthroplasty, 2011. 26(8): p. 1139-44.

2. Babis, G.C., R.T. Trousdale, and B.F. Morrey, The effectiveness of isolated tibial insert exchange in revision total knee arthroplasty. J Bone Joint Surg Am, 2002. 84-A(1): p. 64-8.

3. Hofmann, A.A., et al., Posterior stabilization in total knee arthroplasty with use of an ultracongruent polyethylene insert. J Arthroplasty, 2000. 15(5): p. 576-83.

4. Vince, K.G., A. Abdeen, and T. Sugimori, The unstable total knee arthroplasty: causes and cures. J Arthroplasty, 2006. 21(4 Suppl 1): p. 44-9.

5. Parratte, S. and M.W. Pagnano, Instability after total knee arthroplasty. J Bone Joint Surg Am, 2008. 90(1): p. 184-94.

6. Eisenhuth, S.A., et al., Patellofemoral instability after total knee arthroplasty. Clin Orthop Relat Res, 2006. 446: p. 149-60.

4. Stiffness


Clinically relevant stiffness occurs in about 5% of TKA cases.  It can be caused by pain, component malpositioning/malsizing, or arthrofibrosis. Flexion contracture (ie loss of terminal extension) is the least tolerated limitation in motion because >5-8 ° affects the gait pattern and prevents patients from locking out their knee while standing, leading to increased muscle fatigue bc the quad never gets a chance to relax.  Terminal flexion is important for daily activity: 67° flexion is needed for the swing phase of walking, 100° to descend stairs, 105° to rise from chair.   

A TKA rarely achieves the terminal flexion of a native knee…so when do we say the TKA is clinically stiff?  The definition of stiffness has been a moving target over the past few decades with limitations to knee motion becoming less and less acceptable. Generally speaking, flexion less than 90° is considered stiff.  Stiffness is often very painful and causes significant functional disability.  It is closely related to patient satisfaction after TKA. Risk factors include: young age, limited pre-op ROM, female, poor post-op pain control.


The key is to first identify the underlying cause.  Soft tissue contractures should be a diagnosis of exclusion. Before associating the limited motion with soft tissue, its important to rule out infection, complex regional pain syndrome, and gross malpositioning or malsizing of the components.  Malrotation causes an asymmetric flexion gap, and can lead to lift-off and issues with patellar tracking that decreases flexion.  Furthermore, a tight flexion gap due to oversized femoral component or a tight PCL with a CR-TKA, or a tight patellofemoral joint that causes the retinaculum to tighten with flexion leading to anterior knee pain, or patella baja (the patella articulates with the PE liner causing a block to motion) can all prevent terminal flexion and require separate intervention.  If these causes are ruled out, and soft tissue appears to be the cause of limited motion, then a number of treatments are recommended in a sequential manner (see figure).  To assist in understanding the best treatment, consider the severity of limited motion, if there are technical errors from surgery (best evaluated with a CT scan), if there are patient factors that cannot be modified (ie morbid obesity, major limitation to motion preop, RSD, etc). 

Physical Therapy.  Many patients fail to progress due to poor physical therapy, or poor postoperative pain control.  Improving pain control and putting patients onto a strict PT regimen can often lead to meaningful gains, when the issue is identified early.  The sooner the better, and aggressive PT is best within 6 weeks for flexion or extension bracing for incomplete extension.  Gains from therapy are best in cases of good pre-op motion, small contractures, and relatively good pain control. 

Continuous Passive Motion: Gains from a CPM machine are not well supported in the literature.

Manipulation under anesthesia (MUA). MUA is considered for those that fail to achieve 90° by 6-12 weeks postop.  Beyond 12 weeks, MUA has not proven an effective treatment and it increases the risk for fracture or extensor mechanism injury. The main indication is for limited flexion, yet limited extension is a more common and more refractory soft tissue contracture (MUA for > 10° lost extension is associated with notable risk for supracondylar fracture and should be avoided). The goal of MUA is > 115° flexion with gravity alone.  The technique involves gradually increasing pressure to the knee over 5 to 15 minutes.  Intra-op, gravity alone ROM should be documented with a goniometer and lateral photograph.  After the procedure and before discharge, check that straight leg raise is intact.  Consider a femoral or epidural catheter for postop pain relief because all patients should begin rigorous physical therapy immediately postop.  Outcomes are encouraging.  At an average of 9 weeks post-intervention, the average knee gained 35° flexion and 90% maintained this improvement and continued to improve over the first few months.  One study that compared MUA to patients that refused and elected for therapy alone showed a difference in improvement of 33° to 3° (MUA is 10x better).   Remember that the best outcomes are for painless, stiff TKA, and that significant pain is often an indication of another underlying diagnosis that needs to be addressed before treating the stiffness. Diabetes or obesity are negative predictors for improvement with manipulation. Furthermore, patients with significant preoperative stiffness show regression after a manipulation, with motion approaching preoperative values at one year.

Arthroscopic Lysis of Adhesions (LOA).  This surgical option is typically reserved for recurrent stiffness after failed MUA, or stiffness beyond 3 months (by this time, scar becomes very strong and manipulation increases the risk for fracture or extensor mechanism injury).  Results are best for a painless and stiff TKA.  Resection of excessive scar tissue in the suprapatellar pouch, and medial/lateral gutters can lead to significant gains.  Outcome studies report improvement of up to 36° flexion, and 7° extension (which is often better than MUA).  


1.         Namba, R.S. and M. Inacio, Early and late manipulation improve flexion after total knee arthroplasty. J Arthroplasty, 2007. 22(6 Suppl 2): p. 58-61.

2.         Fitzsimmons, S.E., E.A. Vazquez, and M.J. Bronson, How to treat the stiff total knee arthroplasty?: a systematic review. Clin Orthop Relat Res, 2010. 468(4): p. 1096-106.

3.         Berger, R.A., et al., Malrotation causing patellofemoral complications after total knee arthroplasty. Clin Orthop Relat Res, 1998(356): p. 144-53.

4.         Kim, J., C.L. Nelson, and P.A. Lotke, Stiffness after total knee arthroplasty. Prevalence of the complication and outcomes of revision. J Bone Joint Surg Am, 2004. 86-A(7): p. 1479-84.

5.         Gandhi, R., et al., Predictive risk factors for stiff knees in total knee arthroplasty. J Arthroplasty, 2006. 21(1): p. 46-52.

6.         Bedard, M., et al., Internal rotation of the tibial component is frequent in stiff total knee arthroplasty. Clin Orthop Relat Res, 2011. 469(8): p. 2346-55.

7.         Keating, E.M., et al., Manipulation after total knee arthroplasty. J Bone Joint Surg Am, 2007. 89(2): p. 282-6.

8.         Esler, C.N., et al., Manipulation of total knee replacements. Is the flexion gained retained? J Bone Joint Surg Br, 1999. 81(1): p. 27-9.

9.         Cates, H.E. and J.M. Schmidt, Closed manipulation after total knee arthroplasty: outcome and affecting variables. Orthopedics, 2009. 32(6): p. 398.

10.       Baker, P.N., et al., The role of pain and function in determining patient satisfaction after total knee replacement. Data from the National Joint Registry for England and Wales. J Bone Joint Surg Br, 2007. 89(7): p. 893-900.

Osteolysis & Aspetic Loosening

Aseptic loosening is mainly caused by bone loss secondary to osteolysis (osteolysis is caused by polyethylene wear).  Osteolysis not only causes implant loosening but is commonly associated with periprosthetic fractues (as bone is lost, it becomes weak and more susceptible to fracture).

Osteolysis is less common in TKA as compared to THA, and while its still a common issue for TKA, particularly in the second and third decade of a TKA lifespan, it is not the most common complication (as in THA).  What is the difference between TKA and THA poly wear that leads to different rates of osteolysis?  TKA has different kinematics than THA, which leads to different poly wear patterns.  In THA, abrasive wear is the cause of poly debris.  In TKA, poly debris is generated by fatigue and delamination wear.  Delamination wear is repeated loading that causes subsurface cracks over time that propagate to form large debris particles (which are considerably less in number and less immunogenic).  

A TKA poly must be at least 6 mm thick to prevent fracture or significant delamination wear.  The only time you see abrasive wear in TKA is with backside wear (the backside of the poly rubs against the tibial tray that its locked into: however, the modern tibial trays are highly polished and therefore generate little debris).  While cross-linked poly is believed to have revolutionized THA wear and improved longevity, it is unlikely to have the same effect on TKA because cross-linked poly is most resistant to abrasive wear but there is no improvement to delamination.  In fact, there is concern that because cross-linked poly is more brittle (less ductile and therefore less resistant to fatigue crack propagation), it may be more susceptible to delamination wear.

TKA bone defects are described by the AORI Classification system (Anderson Orthopedic Research Institute).  There are three types.  Type 1 is a contained metaphyseal defects that doesn't impact the stability of a revision implant.  Type 2 is a metaphyseal defect in one femoral/tibial condyle (Type 2A) or both femoral/tibial condyles (Type 2B) requiring cement only (< 5 mm), cement + rebar (5 - 10 mm), or metal augments (>10 mm) to stabilize the implant. Type 3 is a significantly deficient metaphyseal region requiring cones or sleeves, or requiring the implant to bypass the metaphysis completely for stability, and often risks causing soft tissue instability (ie MCL deficiency). 

6. Extensor Mechanism

Patellar Tendon Rupture

Patellar tendon rupture is a serious complication with generally poor results. Direct repair of the tendon has very poor outcomes (<50% success with suture or staple)[65, 66], and therefore a rupture requires reconstruction (like a torn ACL). Reconstruction options include autograft, allograft, or synthetic mesh.  Overall the functional outcome results are inconsistent and highly variable.

Autograft.  Similar to an ACL autograft reconstruction, the semitendinosis can be harvested and used as the patellar tendon reconstruction [67]

Technique. Instead of removing the tendon at its insertion on the pes, the tendon is mobilized but not detached, and a tendon stripper is then used to release the tendons from their muscular attachment proximally.  The tendons are then wrapped through the patella (typically if just the SemiT is used, it can fit into a 6 mm tunnel in the distal 1/3 of the patella) and it is then re-attached onto the proximal tibia with the knee at 90° flexion to ensure proper patellar height.  The knee is then immobilized in extension for 6 weeks, followed by limiting knee flexion to 60° for 12 weeks, followed by progressive knee motion. This is the standard surgical treatment for a chornic patellar tendon rupture in the native knee, which good long term outcomes.  The technique however is less effective in the TKA due to the patellar bone loss from its resurfacing. 

Allograft.  The allograft technique uses either an Achilles (with calcaneal bone block) or Whole Extensor Mechanism (quad tendon-patella-patellar tendon-proximal tibia) graft.  The graft is an afibrous network that allows for fibrous ingrowth.

Technique. The Whole Extensor Mechanism graft requires 5 cm of proximal bone stock and 5 cm of quad tendon. When implanted, the graft must be placed under maximal tension with the knee in extension (and don’t test the graft afterward).  Tensioning majorly decreases the residual extensor lag, which is the primary mode of failure for these grafts.  The graft is either placed over the anterior surface of the extensor mechanism (“onlay technique”) or threaded into the extensor mechanism through the lateral retinaculum (like a pulver-taft weave).  The knee is then immobilized in a cast for 6 weeks. The allograft has good integration into host tissue, yet there are many failures (about 35% early failure and close to 50% at 10 years) [68, 69]Most are functional failures as defined as an extensor lag > 30°.

Mesh. Browne and Hanssen reported a technique using Marlex Mesh (a high density polyethyene fiber typically used for hernia repairs in abdnomial surgery) to recreate the extensor mechanism after a patellar tendon rupture [70]. The concept is that mesh promotes extensive scar tissue formation that provides strength and resists creep, which commonly leads to failure in the allograft technique.  The initial paper reports good outcomes in 9/13 patients, with 3 early re-ruptures and 1 infection. 

Technique.  Fold the mesh several times until it’s a width of 2 – 2.5 cm, and then secure to the tibia by creating a trough proximal and slightly lateral to the tibial tubercle and fixed the mesh with a screw, and then cemented it in place.  If the tibial component is being revised at the time, the mesh can be placed into the tibial canal and then implant cemented over it.  The mesh then runs under the remaining patellar tendon, through the lateral retinaculum, is then fixed to the patella, and then sutured to the quad tendon and vastus lateralis with non-absorbable sutures with the knee in extension.  Then VMO is then mobilized and sutured on top of the mesh to create a "mesh sandwich".  The knee is kept in extension for 6-8 weeks and then 10° of knee flexion is advanced each week.  

Patella Fracture

Patellar fracture is an uncommon problem, occurring less than 0.5% of cases.

Fracture is closely associated with patellar bone stock.  Increased risk of fracture is associated with < 12 mm of native patella after resurfacing.  The incidence of fracture in non-resurfaced patella is only 0.05%. Patella fracture is also associated with AVN.  Disruption of the anastomotic ring around the patella can lead to AVN and then fracture.  Increased risk for AVN is associated with lateral release for patellar tracking and the V-Y turndown for exposure in revision cases.  The prosthetic designs more commonly associated with fracture are the single peg versus the tri-peg fixation.  Also, femoral components with a thicker anterior flange.  Malrotated femoral component can also place excessive force across the patella.

Treatment is determined by whether the extensor mechanism is intact, and whether the implant is stable. [58].

Type 1: Stable implant and extensor mechanism is intact.  Treat nonoperatively with protective weight bearing and a cylinder cast for 6 weeks. 

Type 2: Disrupted extensor mechanism but stable implant, which requires surgical treatment to ensure return of extensor function. These are challenging because the tension band is less effective as when used in native patella fractures.  The patellar component prevents the conversion of the tension force to a compressive force so you cannot get the same level of healing.  Need to cast for 6 weeks after fracture regardless of fixation technique. 

Type 3: Loose components, surgery is recommended but treatment decision further depends on bone stock. With good bone the implant can be revised (possibly using a tantalum patella), while in poor bone (< 10 mm thickness or comminution) may require patellar excision. 

Patella Instability

Patellar mal-tracking is caused by tibia or femoral mal-rotation (internal rotation). Its symptoms are often vague: anterior pain, instability, stiffness, weakness, locking and giving out.  It is most commonly observed in valgus knees (large Q angles), obese patients, malrotated componenets, asymmetric patellar resection.  If surgery is required, the internally rotated components should be revised to maximize improvement.  Lateral release is often used to decrease lateral directed tension to decrease dislocation risk.

Patellar Clunk

The Patellar Clunk Syndrome is a painful clunk sensation when the knee is brought into extension and is associated with scar tissue in the suprapatellar pouch that becomes trapped in the femoral notch at 30° flexion. The problem has mainly been resolved with newer designs that alter the trochlear groove, however, patients with older design TKA can be treated effectively the arthroscopic debridement.


1.         Ortiguera, C.J. and D.J. Berry, Patellar fracture after total knee arthroplasty. J Bone Joint Surg Am, 2002. 84-A(4): p. 532-40.

2.         Rand, J.A., B.F. Morrey, and R.S. Bryan, Patellar tendon rupture after total knee arthroplasty. Clin Orthop Relat Res, 1989(244): p. 233-8.

3.         Lynch, A.F., C.H. Rorabeck, and R.B. Bourne, Extensor mechanism complications following total knee arthroplasty. J Arthroplasty, 1987. 2(2): p. 135-40.

4.         Cadambi, A. and G.A. Engh, Use of a semitendinosus tendon autogenous graft for rupture of the patellar ligament after total knee arthroplasty. A report of seven cases. J Bone Joint Surg Am, 1992. 74(7): p. 974-9.

5.         Burnett, R.S., et al., Retrieval of a well-functioning extensor mechanism allograft from a total knee arthroplasty. Clinical and histological findings. J Bone Joint Surg Br, 2004. 86(7): p. 986-90.

6.         Burnett, R.S., et al., Extensor mechanism allograft reconstruction after total knee arthroplasty. A comparison of two techniques. J Bone Joint Surg Am, 2004. 86-A(12): p. 2694-9.

7.         Browne, J.A. and A.D. Hanssen, Reconstruction of patellar tendon disruption after total knee arthroplasty: results of a new technique utilizing synthetic mesh. J Bone Joint Surg Am, 2011. 93(12): p. 1137-43.

7. Periprosthetic Fracture

The incidence of periprosthetic fracutres (PPF) of the femur is ~ 1.3%, increasing as patients live longer and stay active with their TKA.  There is a close association with osteoporosis, with 4 of 5 occurring in elderly women, and the vast majority occurring after low-energy fall in patients over 60 years old [1].   


The algorithm to evaluate the PPF begins with determining the implant stability (as noted in the Rorabeck classification system) [2-4].

If the implant is stable, then the fracture should be addressed with ORIF using either the submuscular bridge plating via a minimally invasive plate osteosynthesis, MIPO, technique or a retrograde intramedullary nail.

The submuscular plates offer a minimally invasive approach with locking screws to improve fixation in bone that is virtually always osteoporotic. Advantages of a plate includes reliability in cases of significant comminution and distal extension due to the ability to achieve distal fixation with many screw options.  The plate contour further facilitates fracture reduction and maintenance of reduction. However, studies suggest there is a risk of varus collapse when the fracture involves the medial femoral condyle.  Furthermore, patients are unable to weight bear following fixation, and similar to patients with hip fractures, delayed mobilization can lead to significant complications such as pneumonia and DVT.

Retrograde nails similarly offer a minimally invasive approach, but can only be used in implants that have an open notch to allow nail insertion, and no stem.  The notch size determined by the knee implant, which can be quite variable, and surgical planning to identify the implant is necessary to avoid surprises during the case. In CR-TKA, the notch diameter is between 11 – 20 mm, in PS-TKA, the notch is narrower.  The space should be at least 1 mm larger than the planned IMN.  Some common implant notch sizes: S&N Genesis II: 16.5, Zimmer NexGen 14.1 – 21.6, Omnifit – closed box, DePuy TC3 – closed box, DePuy Sigma CR and PS (17.4 vs. 11.9). 

If the implant is unstable, then the fracture and the loose component must be addressed.  Depending on the remaining bone stock, this requires a stemmed femoral implant vs. a hinge or a distal femoral replacement. The fracture is also addressed with cables.  


Meta-analysis of locking plate and retrograde IMN union rates is around 90% [5,6].

However, meta-analysis suggests locking plate complication rates (35%) were lower than IMN (53%), albeit rates were not statistically different, and appear high in both groups, highlighting the severity of this complication (and the underlying morbidity of the patients that suffer this complication).  This complication is most common in the frail elderly population, and therefore, it remains unclear whether one fixation type has a significant impact on the risk for complications.  The most common complications were nonunion and malunion (6% each, yet these rates were 2x higher in patients treated nonoperatively).


1.         Berry, D.J., Epidemiology: hip and knee. Orthop Clin North Am, 1999. 30(2): p. 183-90.

2.         Rorabeck, C.H. and J.W. Taylor, Classification of periprosthetic fractures complicating total knee arthroplasty. Orthop Clin North Am, 1999. 30(2): p. 209-14.

3.         Rorabeck, C.H. and J.W. Taylor, Periprosthetic fractures of the femur complicating total knee arthroplasty. Orthop Clin North Am, 1999. 30(2): p. 265-77.

4.         Su, E.T., H. DeWal, and P.E. Di Cesare, Periprosthetic femoral fractures above total knee replacements. J Am Acad Orthop Surg, 2004. 12(1): p. 12-20.

5.         Shin, Y.S., H.J. Kim, and D.H. Lee, Similar outcomes of locking compression plating and retrograde intramedullary nailing for periprosthetic supracondylar femoral fractures following total knee arthroplasty: a meta-analysis. Knee Surg Sports Traumatol Arthrosc, 2016.

6.         Ebraheim, N.A., et al., Periprosthetic Distal Femur Fracture after Total Knee Arthroplasty: A Systematic Review. Orthop Surg, 2015. 7(4): p. 297-305.