Diagnostic algorithm in aseptic TKA failure - What is evidence-based?

1. Introduction

Total knee arthroplasty (TKA) is a very successful and standardised surgical procedure for the treatment of high-grade osteoarthritis. Despite improved therapeutic standards, new surgical techniques and alignment philosophies, as well as an indication guideline for TKA, up to today it has not been possible to match the success achieved by total hip arthroplasty (THA).¹ With a patient satisfaction of as high as 95–98%, THA represents the most successful surgical procedure of the past century. In comparison, around 20–25% of patients are dissatisfied with the outcome after TKA.^2,3

According to the EPRD 2019 (German Arthroplasty Registry), 119131 primary implantations and 13378 follow-up interventions after TKA were performed in Germany. This corresponds to a revision rate of around 11% (EPRD 2019).⁴ The most common reasons for revision after primary TKA according to the EPRD 2019 are aseptic loosening, with a percentage share of 25%, and infection, with a share of 14.5%. Further causes of failure are instability (8.9%), PE wear (5.7%), arthrofibrosis (4%), as well as periprosthetic fracture (3%) and malalignment (1.8%).⁴

The data from the EPRD 2019 are consistent with the current registry data for 2019 from Sweden, the USA and Australia.4, 5, 6, 7 The same could be confirmed by current studies from 2013 to 2018. Again, the most common reasons for revision include aseptic loosening, infection, instability, patella problems, but also malalignment or arthrofibrosis.8, 9, 10, 11

The following contribution aims to provide an overview of the most common causes of failure after primary TKA. In addition, a diagnostic algorithm for the treatment of a painful TKA is to be formulated on the basis of the current registry and literature data.

2. Rendering the diagnosis (finding the reason for failure)

2.1. Medical history

The medical history is the most important element in rendering the diagnosis of painful TKA.

The most important questions for evaluating pain include (Table 1):

• •Determination of the exact pain localisation
• •Difference in intensity or nature of the pain pre-to postoperative
• •Pain at rest/Pain upon weight-bearing
• •Pain-free/Low-pain interval after surgery
• •Pain provoked by specific movement patterns
• •Pain radiation
• •Nocturnal pain (sign of possible infection)
• •Accompanying pathologies (lumbar spine, hip)
• •Accompanying systemic diseases (CVI, diabetes, rheumatological diseases)
• •Pain improved by topical or systemic administration of NSAIDs

Table 1.

General diagnostics.

anamneses	clinical examination	imaging (x-ray)
type of physical complaints	inspection	knee in 2 planes
pain character	palpation	patella axial
pain pre- to postoperative	ROM (range of motion)	possibly long-leg view
pain at rest/pain upon weight-bearing	stability of ligaments
nocturnal pain	patella tracking
increased pain after surgery	examination of the hip/spine
feeling of instability	possibly aspiration
wound healing disorder
additional revision surgery

Apart from asking about the pain symptoms, the medical history involves enquiring about other complaints, such as a feeling of instability or “sounds made by the prosthesis”. Further, the patient should always be asked about the operation and the postoperative course, the duration of wound secretion, injuries, or other surgical interventions.

A post-hoc review of the indication for total joint replacement is also helpful, as persistent pain of unclear origin without high-grade osteoarthritis is common. Pre-OP deformity and limitations sometimes also explains typical post-OP problems, such as valgus deformity, hyperlaxitiy or very limited ROM.

2.2. Clinical examination

The patient should be asked to localise the point of maximum pain as accurately as possible, as this may provide an indication as to the cause.

The clinical examination should comprise the following points (Table 1):

• •Inspection (redness, swelling, effusion, surgical scar)
• •Palpation (overheating, effusion, oedema, pain on pressure: patella, Gerdy's tubercle, pes anserinus, head of the fibula)
• •Clinical assessment of the leg axis (orthograde, varus, valgus)
• •ROM (range of motion)
• •Stability of the lateral ligaments (varus-valgus stress at 20°, 60°, 90°)
• •Sagittal stability testing in CR prostheses (posterior drawer test)
• •Evaluation of patellar tracking (tilt and shift)
• •Examination of hypoesthesias and paraesthesias
• •Identification of neurinomas

2.3. Imaging procedures

2.3.1. Standard images (Table 1)

• •Knee radiograph in 2 planes a.p. and strictly lateral (alignment, overstuffing, underhang, loosening, post. offset, ant. offset, fractures, necrosis, osteolysis, stress shielding, joint line, patella baja, pseudo patella baja, wearing)
• •Axial patellar radiograph (tilt, shift, osteoarthritis, necrosis, osteolysis, stress shielding)

2.3.2. Special images (Table 2)

Table 2.

Special diagnostics.

knee radiograph in 2 planes	long-leg view	patella axial or	stress images varus/valgus or	computer tomography	3-phase skeletal	magnetic resonance imaging	sonography	local infiltration	aspiration	biopsy
		sunrise view	sagital (CR design)	(CT)	scintigraphy	(MRI)
loosening	component malpositioning	patella tracking	medio-lateral instability	component rotation	loosening	insufficiency of the ext. apparatus	insufficiency of the ext. apparatus	soft tissue problems	infection	infection
joint line	mechanical axis	tilt and shift	anterior-posterior instability	loosening (dual energy)	fracture	muscular imbalance	effusion	infection	gout	allergy
malpositioning	osteoarthritis of hip/ankle	osteolysis		TTTG	hot patella	neurinoma	Baker's cyst
fracture	joint line	stress shielding		fracture/osteolysis	infection		capsule
osteolysis				impingement	synovial inflammation		popliteus tendon
PE wear				stress shielding	CRPS

• •Long-leg view (for evaluation of the mechanical axis)
• •Patella axial or sunrise view in 30°, 60° and 90° (for evaluation of patellar tracking)
• •Stress images varus/valgus at 30°, 60°, and 90°
• •Stress images sagittal (posterior instability)
• •CT (component rotation, TTTG, impingement, loosening, fracture, osteolyses, stress shielding)
• •Sonography (quadriceps, patellar tendon, capsule, Baker's cyst, popliteus tendon)
• •3-phase skeletal scintigraphy (at the earliest 12 months after surgery: loosening, CRPS, fracture, patellar hyperpression, synovial inflammation, infection)
• •Leukocyte scintigraphy (to rule out infection)
• •PET-CT (to detect focus points, to rule out infection)
• •Duplex sonography (thrombosis, aneurysm, arterial occlusion)
• •Electrophysiological examination (nerve conduction velocity)
• •If necessary, radiograph of the lumbar spine standing and pelvis
• •MRI (extension apparatus, quadriceps, neurinoma)

Notice: Scintigraphy, PET-CT and other expensive studies play an exceptional role in the evaluation and should not be performed as standard.

3. Reasons for revision

3.1. Periprosthetic infection

For the sake of completeness, periprosthetic infection is to be mentioned briefly in this section. With a share of up to 30% in the registers and up to 36% in the studies, it is the second most common cause of failure as a whole and the most common cause of early failure after TKA within 2 years.^4,5,8,9,13 Therefore, infection should be ruled out in every case of a painful knee prosthesis.

3.2. Aseptic loosening

Several studies have demonstrated that pain with an onset 1 year after primary TKA and in which periprosthetic infection was excluded is most probably caused by aseptic loosening. Lim et al. investigated 122 cases of aseptic TKA pain and showed that they were predominantly caused by aseptic loosening (n = 40).¹⁴ In a further article, the rate of aseptic loosening was 23.1%, again making it the main reason for revision surgery.¹⁵ Kenney et al. determined that, in the revisions performed 2 years after primary TKA, aseptic loosening of the components accounted for 51.4% of the patients, with a loosening rate of 22.8% in the first two postoperative years.¹⁶ Similar results were achieved by the research group of Thiele et al., who followed up 358 revision operations after primary TKA in their study. They attributed 16% of the revisions in the first 3 years to aseptic loosening, and even 34.7% in revision operations after more than 3 years.⁹

Generally, loosening of the prosthesis components is detected on the basis of already clearly visible or progressive radiolucent lines as well as changes in the position of the prostheses.¹⁷ In addition, osteolyses and osteolytic defects in combination with stress-dependent pain may be indicative of aseptic loosening. In this respect, for projectional radiography, the research group of Marx et al. described a sensitivity of 77% for the femoral and 83% for the tibial component, with a specificity of 90% and 72%, respectively.¹⁸ In contrast, 3-phase skeletal scintigraphy can be used as a diagnostic procedure to confirm loosening at the earliest 1 year postoperatively, since bone re-modelling still takes place during this time.¹⁹ Here, Smith et al. state a sensitivity von 92.3% and a specificity of 75.9%.²⁰ However, over the past years, PET-CT has increasingly shown distinct advantages in the diagnosis of aseptic implant loosening. Depending on the tracer used, the diagnosis of early aseptic loosening and a differentiation from periprosthetic infection is also possible.^21,22 In a current meta-analysis by Barnsley et al., the data suggest that the most accurate method for detecting an aseptic TKA loosening is SPECT/CT arthrography.²³

In conclusion most cases of aseptic loosening will detected by standard radiograph. In rare doubtful cases SPECT/CT can be a useful additional imaging.

3.3. Instability

Ligamentous instability is the second most common cause of early failure (up to 23.9%) of TKA within 2 years.^9,11 According to Graichen (2007), 77.3% of instabilities are acquired primarily and are directly connected to the primary implantation.²⁴ Here, it is important to distinguish between a sagittal instability (in CR prostheses), a medio-lateral instability (in extension, midflexion or flexion) in CR, PS or CC prostheses, or a multidirectional instability. The most common instabilities are flexion instability and multidirectional instability.

A medio-lateral instability is tested for by varus or valgus stress on the collateral ligaments. For this purpose, the affected knee joint is examined in extension (in approx. 30° flexed position to relieve the dorsal capsule) and in 60° flexion, and the instability is evaluated. If clinical evidence of instability is obtained, this should be documented by imaging using a hand-held device under the image intensifier or a mounted device.²⁵

To test for sagittal instability, the drawer test is performed in approx. 90° flexion. Here, above all the posterior drawer is assessed in the case of CR prostheses. An anterior drawer of up to 5 mm is within the normal range, from 7 mm upwards this may also be symptomatic.²⁶

3.4. Malalignment

Malalignment refers to causes of mechanically related pain after TKA. Axis deviations in the rotational plane, termed malrotation, are clinically the most relevant here.²⁷ Malrotation of the tibial or femoral component can be a cause for pain, instability or patella maltracking. The rotation of the implant components can be evaluated best using normal computed tomography of the knee. Additional in a case of patella maltracking the determination TTTG make sense by using CT. The value of rotational CT has come down over the last years, so it is recommended only for specific cases. In contrast, long-leg views as well the lateral radiograph are sufficient for evaluation of coronal and sagittal alignment. Implant malalignment or the imbalance between flexion/extension gap may result in continuous complaints and cause a feeling of instability.²⁸

3.5. Patellar maltracking/pathology (patella baja, aseptic necrosis, patellar clunk syndrome, secondary RPA)

Disturbances of patellofemoral mechanics, such as patellar maltracking or increased pressure caused by the extension apparatus, may lead functionally to a local overloading of the lateral retinaculum.²⁹ A static or dynamic valgus malalignment can result in lateral maltracking of the patella.³⁰ The incidence of a patellofemoral instability, which usually involves maltracking, is reported as being between 1 and 20%.³¹ From a diagnostic point of view, tangential patellar images (sunrise view) are beneficial here. Attention should be paid to lateralisation, patellar tilt, an overhang, or the presence of lateral osteophytes.

Contractures of the patellar tendon or scarification above the tibial tuberosity can lead to a postoperative patella baja. A so-called pseudo patella baja develops due to an elevated joint line and is the result of an over-resection of the distal femur or under-resection of the proximal tibia, with a high insert. A lateral radiograph of the affected knee, combined with determination of the Insall-Salvati index or the Caton-Dechamps index, is sufficient for diagnosis.^32,33

Other causes of TKA failure include chondrolysis or necrosis of the patella. In this connection, the term “overstuffing” implies a narrowing of the patellofemoral joint, with increased contact pressure resulting from the implantation of a femoral component that is relatively too large or implanted too far anteriorly, or an insufficient degree of patella resection accompanied by a too thickly applied retropatellar replacement. Here, too, an increased tension of the retinacula occurs, with resulting pain above the patella upon weight-bearing and the potential development of avascular necrosis.^34,35 The gradual development of a stress fracture of the patella may also be an explanation for persistent anterior knee pain and may also be attributable to patellar necrosis. If this is suspected, beside conventional imaging, CT scans play a crucial role.

Patellar clunk syndrome involves a fibrous thickening at the transition from the proximal patellar pole to the quadriceps tendon, which causes a “painful snapping” of the patella and is described above all in the case of PS prostheses.³⁶ The incidence of patellar clunk syndrome is reported as 0–18% in the literature.³⁷ If there is clinical-radiological evidence of secondary retropatellar osteoarthritis, diagnostics should be extended to include 3-phase skeletal scintigraphy, in order to verify the findings and, if necessary, implant a secondary retropatellar replacement.³⁸

3.6. Arthrofibrosis/CRPS

Arthrofibrosis involves the formation of excessive scar tissue, leading to a painful restriction of movement. This includes an extension deficit of at least 10° and a limited flexion capacity of less than 90°.³⁹ It also represents a common reason for revisions after TKA. The registers and the current literature contain differing rates of revision due to arthrofibrosis, ranging from 3 to 15.8%.^11,13,40 The diagnosis of arthrofibrosis is primarily rendered clinically and can be confirmed histologically in the case of a revision operation.³⁹ Here, a primary form with unclear origin is distinguished from secondary arthrofibrosis with an identifiable cause. Whereas it is recommended that primary arthrofibrosis be treated conservatively, in the case of secondary arthrofibrosis the cause should be rectified surgically.^39,41

In the following table, the most important risk factors for arthrofibrosis (according to Thompson et al., 2019) are summarised.³⁹

3.7. Preoperative risk factors

• •Impaired ROM
• •Previous surgery on the knee
• •Nicotine abuse
• •Systemic diseases (diabetes mellitus)
• •Depression
• •Genetic factors

3.8. Intraoperative risk factors

• •Excessive soft-tissue balancing
• •Malpositioning/Malrotation of the components
• •Overstuffing of the components
• •Incorrect joint line
• •Patellofemoral malrotation

3.9. Postoperative risk factors

• •Long-lasting immobilisation
• •Infection
• •CRPS

3.9.1. CRPS (complex regional pain syndrome)

As a differential diagnosis for arthrofibrosis, a CRPS (complex regional pain syndrome) should be ruled out. The AWMF (Association of the Scientific Medical Societies in Germany) provides an aid to rendering the correct diagnosis in the form of a guideline on the diagnosis and therapy of complex regional pain syndrome (CRPS), which was revised in 2018.⁴² Therapy is conservative by means of physiotherapy, ergotherapy and systemic pain therapy. If therapy is unsuccessful, steroids or bisphosphonates are recommended. Further medical options are N-acetylcysteine, gabapentin or ketamine.⁴²

3.10. Impingement syndrome

Impingement belongs to the rare complications after TKA. One reason for this might be the more difficult and unclear diagnosis. The term impingement syndrome may bring together several different pathologies. At present, a clear explanation of the term is lacking. By definition, it is a compression or irritation of soft-tissue structures, which may lead to pain and impaired mobility.43, 44, 45, 46, 47 All soft-tissue and ligament structures may be affected in the knee. Commonly affected structures are Hoffa's fat pad, the popliteus muscle, the patellar ligament or the iliotibial tract.43, 44, 45, 46, 47, 48, 49

3.10.1. Popliteus impingement

One reason for a popliteus impingement is that the PE insert does not match the convex structure of the lateral tibial plateau of the natural knee joint.⁴³ This problem is observed in the case of ultra-congruent inserts in particular. Therefore, more recent studies have recommended that the size of the tibial component should not only be orientated according to the cortical structure but that soft tissues should also be respected.^43,44 The diagnosis is generally rendered clinically. The patient typically describes dorsolateral knee pain. For confirmation of the diagnosis, a CT scan should be performed to detect a possible dorsolateral overhang. With regard to therapy, the cause must be treated by re-positioning of the tibial component or, if necessary, the prosthesis must be replaced completely.

3.10.2. Hoffa's impingement

Hoffa's impingement appears above all to be a problem of mobile and more commonly of floating inserts, as a rotation of the insert upon flexion may occur here, leading to irritation of Hoffa's fat pad.^46,47 Hoffa's impingement may result in the case of a patella baja in particular. A clear recommendation with regard to the treatment of Hoffa's fat pad in TKA is lacking. Whereas some recommend a complete resection in order to avoid Hoffa's impingement, others are of the opinion that only a partial resection of the fat pad is necessary or no resection at all.^50,51 The diagnosis is primarily rendered clinically. In the lateral radiograph, a patella baja or an anterior overhang of the tibial component can be visualised. The therapy options include infiltration or, if conservative treatment is unsuccessful, a revision operation with therapy of the cause of symptoms.

3.10.3. Iliotibial band impingement

Iliotibial band syndrome or impingement describes a friction of the iliotibial band on the tibial or femoral component. A common cause is the escape of cement in the area of the femur or a lateral overhang of the tibial component.⁴⁹ The incidence is higher in the case of implants with guided kinematics.⁵¹ The patient describes lateral knee pain, predominantly upon flexion.^53,54 The diagnosis is rendered clinically and radiologically. A strict ap radiograph should be performed to confirm the diagnosis. An additional CT scan may be helpful to detect an overhang of the prosthesis.

3.11. Soft-tissue problems (insertional tendinopathy, insufficiency of the extension apparatus, neurinoma, muscular imbalance, kinematic problems)

If there is clinical evidence of soft-tissue symptoms after implantation of a knee replacement, diagnostic procedures should be extended to include sonography or tomographic imaging. If, for example, a pronounced peripatellar synovitis is present, a synovectomy should be performed. With a systematic review, van Jonbergen et al. showed that this can have a positive effect on pain after TKA.⁵⁵

3.11.1. Insertional tendinopathy

In addition, insertion tendinopathy occasionally occurs in the first months after TKA. Typical localisations are the pes anserinus and Gerdy's tubercle anterior-laterally on the tibial head, as well as more rarely the tip of the patella. Palpation reveals isolated pain on pressure. Further diagnostics are performed with a probatory local infiltration of local anaesthetic, in order to confirm the diagnosis.

3.11.2. Insufficiency of the extension apparatus

If the patient complains of a lack of strength and lifting of the stretched leg is impaired, sonography should be performed to check for a patellar and quadriceps tendon rupture. If the findings are positive, surgical therapy is performed with reconstruction of the extension apparatus.⁵⁶ In a systematic review from 2019, the complication rate after repair of the patellar tendon (63.16%) was higher than the complication rate after repair of the quadriceps tendon (25.37%). However, the complication rates in the case of patellar tendon and quadriceps tendon ruptures after auto-, allotransplantation or mesh reconstruction were similar (18.8% versus 19.2%).⁵⁷

3.11.3. Muscular imbalance

If only a quadriceps weakness is present, this may be compensated for by a forward leaning of the trunk, which leads to an increased pelvic tilt and a flexion position in the knee joint, with a high loading of the femoropatellar joint.⁵⁸

3.11.4. Neurinoma

In his studies, Sanchis-Alfonso found nerve changes in the lateral retinaculum in the sense of a neural myxoid degeneration with neuromas, and he was able to demonstrate a clear connection between neuromas and patellofemoral pain.²⁹ Dellon et al. treated such patients, who reported a clear improvement of the symptoms in response to a probatory nerve blockade using local anaesthetic, with a denervation. In all cases, the authors achieved an improvement lasting at least 6 months.⁵⁹

3.11.5. Kinematic problems (paradoxical translation roll forward in CR)

Investigations of TKA kinematics have shown an anterior translation of the femur relative to the tibia in flexion movements, which is referred to as paradoxical translation or roll forward. Delport et al. reported that this phenomenon exists in the majority of CR prostheses, compared with 0% in PS knee prostheses.⁶⁰ This paradoxical translation can therefore have a negative influence on the outcome of certain CR prostheses. The tibiofemoral range of motion can thus be reduced, polyethylene wear is increased and the moment of force of the extension apparatus is reduced.⁶¹ Published possibilities of diagnostics regarding this kinematic problem are the fluoroscopic analysis and marker-based motion detection of the gait analysis (1. Femoral rollback of cruciate‐retaining and posterior‐stabilized total knee replacements: In vivo fluoroscopic analysis during activities of daily living; Fantozzi S. et al. JOR 2006.

2. Modifications of femoral component design in multi-radius total knee arthroplasty lead to higher lateral posterior femoro-tibial translation, Pfitzner T. et al., KSSTA 2018.) If the knee pain can be attributed to this phenomenon, the implant should be changed to a posterior stabilized prosthesis.^62,63

3.12. Referred pain in the hip and spine, insidious fracture

Osteoarthritis patients are often additionally suffering from disorders of the hip-stabilising musculature. As a result of weakness of the external rotators, a valgus position of the knee joint may occur (functional malalignment).⁶⁴ If the functional weakness of the external rotators is a result of hip osteoarthritis, therapy of the hip osteoarthritis is of primary importance to alleviate the femoropatellar pain. Degenerative comorbidities in the area of the lumbar spine also occur in rare cases. In their study, Al-Hadithy et al. reported that spinal complaints and osteoarthritis of the hip were the second most common causes of painful TKA.⁶⁵

3.13. Allergy (PMMA, chrome, nickel)

Knee replacements are mainly composed of metals and acrylate cements, which are subject to corrosion and wear.^66,67 The metal ions and other molecules thus formed can induce a delayed-type hypersensitivity reaction (DTH). Nickel, cobalt and chrome are the most common sensitisers, whereas reactions to polymethyl methacrylate (PMMA) tend to be rare.68, 69, 70 The patch test is the gold standard for confirming the diagnosis of a contact allergy, although its sensitivity is stated as 100%–77%, at a specificity of 71%–64%.⁷¹ This casts doubts on its validity, as there are also few clinical studies available that provide relevant data about the prevalence of metal sensitivity and the clinical outcome of the implant.⁷² In a level-II cohort study conducted by Bravo et al., no causal connection could be detected between a positive patch test regarding metal hypersensitivity and implant failure.⁵⁶ In contrast to this, there are studies suggesting that, on the basis of the high proportion of metal hypersensitivity reactions in patients with prosthesis loosening and the shorter lifespan of the implant in patients with a positive patch test, allergy may be a factor contributing to failure of the implant.^68,73 Other in-vitro tests for a potential metal allergy include the lymphocyte transformation test (LTT), which measures the proliferative reaction of lymphocytes after activation and is considered more suitable for testing implant-related reactions, with higher sensitivity rates, and the leukocyte migration inhibition test (LIF), which measures the restriction of the migration of leukocytes.72, 73, 74 A positive LIF correlates well with pain, swelling and dermatological reactions.^72,73 However, these two tests should be reserved for those patients who have a negative patch test but show high-grade clinical signs of an implant allergy. As a last resort, a biopsy may be performed to confirm the diagnosis of a metal/cement allergy, with the typical T-lymphocyte-rich immunohistopathology being detectable in the periarticular tissue.⁷⁵

A clear risk factor for metal/cement allergy is the presence of a positive medical history for general delayed-type hypersensitivity reactions, which increases the probability of prosthesis failure by a factor of four. Other in-vitro tests have shown a reduction of metal ion release and superior tribological properties of coated TKA implants and thus fewer polyethylene wear particles.⁷⁶ However, it is not known whether such implants bring a patient-relevant benefit in vivo with regard to hypersensitivity reactions, metal ion release and improved long-term results. Middleton et al. propose that a metal hypersensitivity does not justify the revision of a TKA and that implants made of standard material should continue to be used in all patients.⁷⁷ In a further review article, the authors come to the conclusion that, on the basis of the evidence currently available, there may be a connection between TKA failure and hypersensitivity reactions.⁷⁸ On the other hand, however, there are currently no recommendations for obligatory, preoperative metal hypersensitivity tests. Therefore, the tests are primarily reserved for patients with signs and symptoms of hypersensitivity towards metals after TKA implantation.⁷⁹

4. Conclusion

As a result of the numerous different causes of pain after TKA, a systematic diagnostic approach with the aid of an algorithm is desirable. Its order of ranking derives from the invasiveness of the diagnostic measures and the probability of the suspected diagnosis, and is therefore flexible.