Evaluation of the Painful/Failed Total Knee Arthroplasty

Chapter Preview

CHAPTER SYNOPSIS:

Successful evaluation of the painful total knee arthroplasty directs treatment by providing the clinician with a specific diagnosis. Common diagnoses include loosening, infection, instability, component failure, patellofemoral disorders, and osteolysis. The less common causes of pain are numerous and may present a diagnostic challenge. The pattern of pain may provide the clinician with insight into underlying etiology. Timely diagnosis of nonarticular causes of pain after total knee arthroplasty is essential as it may save the patient from unnecessary delay in appropriate treatment while avoiding inappropriate diagnostic and therapeutic intervention.

IMPORTANT POINTS:

1

Establishing a specific diagnosis is paramount.
2

Evaluation is incomplete until bacterial contamination has been correctly diagnosed or excluded.
3

All painful total knee arthroplasty evaluations must include investigation for extra-articular sources of pain such as hip, spinal, or vascular pathology.
4

The role for advanced imaging is evolving and includes providing information on component position and the likelihood of a loosening, infection, or complex regional pain syndrome etiology for the patient’s symptoms.
5

Surgical intervention without a clearly defined indication results in poor outcomes.

CLINICAL PEARLS/PITFALLS:

1

Symptom pattern may provide insight into etiology.
2

Late postoperative development of symptoms in the cruciate-retaining knee may indicate flexion instability.
3

Small amounts of malrotation in the tibial and femoral components can be cumulative and clinically relevant.
4

Late postoperative symptoms in association with effusion may be suggestive of polyethylene wear synovitis.
5

Incorrect exclusion of infection.
6

Exploratory surgery or revision without clearly established etiology.
7

Isolated liner exchange for multidirectional instability or flexion-extension gap mismatch.

VIDEO AVAILABLE:

Not available.

HISTORY/INTRODUCTION/SCOPE OF THE PROBLEM

The success of modern total knee arthroplasty (TKA) in providing effective, predictable pain relief is well supported by more than 20 years of clinical follow-up. The mean effect size of a total knee—as defined as the number of standard deviations of change from baseline scores—has been shown to be an improvement of 1.62 and 2.35 at 0 to 2 years of follow-up in WOMAC and Knee Society scores, respectively. These positive results have also proved to be durable, with long-term follow-up studies consistently showing contemporary implant survivorship of greater than 90% at 10- to 13-year follow-up. Problems with this procedure, however, persist. While a meta-analysis of 9879 total knee patients from 130 studies recently demonstrated an 89.3% good/excellent result, these successes must be considered in light of the 10.7% of patients who had a fair/poor result. Physicians who perform TKA, therefore, will frequently be required to evaluate patients with postoperative pain who are unsatisfied with their outcome.

The clinician’s paramount consideration when presented with symptoms at the level of a TKA is the establishment of a specific etiology for these symptoms. Several studies have demonstrated that poor results can be expected when intervention without specific justification is attempted. In a series of 27 patients revised for a prerevision diagnosis of pain, 59% had a poor or fair result. Even in the subset of patients who at the time of surgery were found to have an apparently identifiable cause of symptoms, only 25% benefited from the procedure. Similarly, in a series of 28 failed primary total knees, none of the patients revised without a diagnosis had a good or excellent result.

INDICATIONS/CONTRAINDICATIONS

The indication for diagnostic evaluation of the painful TKA is pain at the site of a total knee that is concerning to the patient and/or interferes with the patient’s daily life. Upon presentation, the patient’s description of their pain pattern should be carefully considered. It has been suggested that the pattern of symptoms may be suggestive of the symptom’s etiology. While pain patterns may frequently exhibit crossover between different modes of failure, they nonetheless provide a useful means for producing a differential diagnosis and directing the course of the patient’s evaluation. Close attention to the patient’s symptom pattern may be especially relevant as the physical examination is often nonspecific.

Laskin has described seven types of pain after a total knee: start-up, weight-bearing, full-extension, full-flexion, stair, rest, and day 1 onward. Start-up pain is typified by pain upon first rising from a seated position. Outside of the perioperative period, when it may suggest normal rehabilitation and, for cementless designs, ongoing bony ingrowth, start-up pain is a pattern that may be suggestive of component loosening. Weight-bearing pain that is relieved with rest and can be seen with or without start-up pain may be indicative of mechanical causes such as loosening or instability. Full-extension pain, in which the patient is comfortable during activities until the knee is brought into terminal extension, may be experienced by patients with an overstuffed extension space. Full-flexion pain, on the other hand, occurs when the patient attempts to flex past their knee’s soft-tissue, bony, or prosthetic limit; posterior impingement and overstuffing of flexion or patellofemoral spaces should be considered. Pain with climbing or descending stairs should raise suspicion for issues with patellar tracking, tibial slope, and flexion instability. Pain at rest can be indicative of inflammation due to infection or synovitis or may be indicative of complex regional pain syndrome. So-called Day 1 onward pain, in which the patient experienced no postoperative period of pain relief, should also raise clinical suspicion for infection. In such cases, a full workup for extra-articular etiology becomes especially relevant.

CLASSIFICATION SYSTEM

Postoperative periprosthetic knee pain can be simply classified by the common modes of failure responsible for the patient’s symptoms. These include the following:

* Infection
* Loosening
* Instability
* Stiffness
* Osteolysis
* Component failure
* Extensor mechanism and patellar-femoral complications
* Soft-tissue impingement
* Unexplained pain

Infection

While a full review of resources available to establish the diagnosis of periprosthetic infection is beyond the scope of this chapter certain characteristics of the chronically infected total knee merit inclusion. The literature endorses four types of periprosthetic infection: type I, patients with a positive intraoperative culture; type II, early postoperative infection, type III, acute hematogeneous; and type IV, late chronic infection. The patient describing an extended period of pain at the site of a TKA must raise the concern for a type IV infection.

A review of the patient’s past medical history is a prerequisite for proper evaluation because a number of comorbidities have been identified as risk factors for periprosthetic infection. These include compromised immune status, immunosuppressive therapy, poor nutrition, hypokalemia, diabetes mellitus, and tobacco use. A history specific to the affected knee is also relevant because previous surgery and prolonged surgical time (>2.5 hours) have also been shown as risk factors for subsequent infection. The postoperative course of the index procedure should also be reviewed for periods of prolonged wound drainage, delayed healing, or hematoma as these have been shown to lead to increased risk for infection. Early return to the operating room and/or extended periods of antibiotics may also be suggestive of a prior type I or II infections. Body habitus may also play a role; a report of 50 TKAs done in morbidly obese patients showed a 5-fold increase in infection rates compared with age-matched control subjects.

The physical examination findings for chronic sepsis are nonspecific. They may include moderate swelling, decreased range of motion, and warmth of the joint compared with the non-involved joint.

Diagnostic tools available for establishing or excluding the diagnosis of periprosthetic infection include hematological studies including C-reactive protein, erythrocyte sedimentation rate, and white blood cell count; joint aspiration for Gram stain, culture, cell count, and differential; and technetium scans, particularly when paired with an indium-labeled leukocyte scan.

Loosening

In a series of 5760 knee arthroplasties reviewed through the use of a community registry, aseptic loosening was responsible for 40.8% of revisions. Radiostereometric analysis (RSA) has provided an accurate picture of implant behavior both for failed implant and for the stable prosthesis. In a randomized study of 26 knees using RSA to compare migration of the tibial baseplate, similar migration rates at 1 year of 1 ± 0.2 mm for cemented implants and 1.4 ± 0.22 mm for uncemented designs have been observed. Elevated early migration, however, has been associated with late prosthetic loosening. This suggests that what is thought of as an abrupt loss of fixation is instead a process initiated very early in the service life of the arthroplasty, perhaps intraoperatively, and represents a prolonged adverse progression that only later manifests as pain.

Nonetheless, patients with a loose knee prosthesis will commonly describe an initial postoperative period of pain relief and improved function; only after this asymptomatic period will symptoms develop. The patient may endorse start-up pain and/or pain with weight bearing. The physical examination is quite nonspecific and may include swelling and a painful, reduced range of motion in comparison to the presymptomatic state. Tenderness at the bone–implant interface, while often suggested as a sign of component loosening, is not an uncommon finding in the well-fixed, normally functioning knee arthroplasty.

The shortcomings of conventional radiography continue to make the routine monitoring of implant fixation problematic. Subsidence and change in implant position are two conventional criteria used to define component loosening. Unfortunately, only relatively large changes in position are detectable with conventional radiographs, which have been shown capable of detecting a 2-mm change in component position and a 3-degree change in angulation under study conditions and are likely a great deal less reliable under routine clinical conditions.

A complete, progressive radiolucent line at the bone–cement interface is also suggestive of loss of fixation ( Fig. 15-1 ). Less clear, however, is the significance of an incomplete radiolucency. In an examination of 123 cemented total knees, no statistically significant correlation was found between the occurrence of incomplete radiolucent lines in any location and the eventual postoperative clinical result. The same investigation demonstrated that a change in only 6 degrees of beam angulation was sufficient to obscure a 1-mm radiolucent line. There may, therefore, be a role for fluoroscopically guided radiographs to correctly orient the beam. The identification and interpretation of incomplete radiolucencies on routine radiographs, however, may be of little import. The senior author has discontinued reporting of incomplete or nonprogressive radiolucent lines in his TKA studies as there has been no correlation noted between their existence and late aseptic failures.

Evidence exists that prosthetic loosening rates are influenced by component alignment. This finding, however, has not been universally replicated. Examination of radiographs for gross malalignment may, however, alert the clinician to technical considerations from the time of the index procedure that may have a negative impact the bone–implant interface. This includes issues pertaining to the effectiveness of cementation technique and the adequacy of exposure for component positioning.

Advanced studies performed to distinguish loosening from infection and complex regional pain syndrome (CRPS) are discussed below.

Instability

Instability is a substantial source of pain after TKA. In a recent series of 212 consecutive revision total knees 21.2% of early and 22.2% of late revisions were performed for a diagnosis of instability. This is similar to the 28% instability seen in a cohort of 440 knees referred for revision within 5 years of the index arthroplasty. Instability may occur in the coronal and/or sagittal planes and may be grossly obvious to the clinician and patient or may be subtle enough to prove elusive on initial evaluation. Depending on the underlying source of instability, symptoms may be apparent immediately after implantation or may not become manifest until late in the postoperative course.

Coronal Plane Instability

Coronal plane instability may be due to ligament imbalance, component malalignment, or component failure. The patient’s preoperative radiographs and physical examination results should be reviewed for significant and/or fixed deformity that may have led to incompetent preoperative ligamentous restraints or that required extensive soft-tissue release at the time of arthroplasty. Patient characteristics such as morbid obesity should be noted as such patients have been shown to have a significantly increased risk for intraoperative medial collateral ligament avulsion. Observation of the patients gait may reveal a varus or valgus thrust for patients with gross coronal plane instability. Examination of coronal plane stability should be performed with the knee in both full extension and in flexion with attention paid to the reproduction of the patient’ symptoms by the application of a varus or valgus stress. Radiographs should be examined for malalignment, outlying insert sizes suggesting problems with intraoperative balancing, and overhanging components that may have resulted in ligamentous attenuation over time. In patients with suspected coronal instability stress radiographs may be warranted.

Sagittal Plane Instability

Sagittal plane instability, typically manifest with the knee in flexion, can be due to a disproportionately increased flexion gap secondary to component size or position, an improperly balanced or attenuated posterior cruciate ligament in cruciate retaining designs, and issues related to the tibial post in posterior stabilized designs. It is routinely referred to as flexion instability and mid-flexion instability.

Flexion instability has been described as a distinct clinical entity seen in cruciate retaining knees, an observation which expanded on earlier case reports of the phenomena. In these patients, an improperly balanced posterior cruciate ligament or its delayed rupture leads to a painful subluxation of the tibia on the femur. This may be exacerbated by a disproportionately large flexion gap due to an undersized femoral component, an anteriorly placed femoral component, or overresection of the posterior tibia resulting in an increased tibial slope. Patients with flexion instability will endorse a sense of instability especially in the descent of stairs and may describe recurrent effusions. The physical examination will commonly reveal an above average range of motion, tenderness over the pes anserinus tendons, and a positive posterior drawer or posterior sag sign ( Fig. 15-2 ). Standard radiographs will often be normal; posterior stress radiographs with the knee in 90 degrees of flexion, however, may provide radiographic evidence of posterior subluxation of the joint.