Bone loss in total ankle arthroplasty (TAA) can have a significant impact on outcome. There is no validated classification system to date for bone loss in TAA. A simple classification system needs to be adopted and validated by the surgical community so that patterns of bone loss can be linked to outcomes, ultimately guiding the surgeon’s treatment options. The authors propose a classification system that is based on patterns of bone loss in critical areas that have independently been shown to result in poor outcomes.
Bone loss can occur in TAA without classic findings of osteolysis or lytic lines, as seen in hip and knee arthroplasty, such as in the case of subsidence.
Bone loss associated with aseptic loosening is usually associated with pain. Significant bone loss can result in the need for arthrodesis.
There are no studies to date that can reliably guide the surgeon as to when an arthrodesis is required, versus a revision TAA.
The assessing surgeon must be familiar with the morphology of the implant he or she is studying to more accurately assess patterns of bone loss.
Painful TAAs should be considered to have aseptic loosening with associated bone loss until proved otherwise.
Bone loss involving the malleoli or critical areas of the plafond that would result in collapse of the prosthesis into a malaligned position is significant.
Collapse of the talus is also significant.
HISTORY/INTRODUCTION/SCOPE OF THE PROBLEM
Total ankle arthroplasty (TAA) has seen a resurgence in interest, with some authors now recommending it as the treatment of choice for end-stage arthritis of the ankle. This is a significant departure from early results of TAA, which were suboptimal, partly because of design issues but also because of the use of cement, both of which lead to significant bone loss. Several papers report on the salvage of these prostheses, but, unfortunately, no standardized reporting system was used to describe the extent and location of the bone loss, let alone the predictive nature of defining bone loss before or during revision surgery (i.e., what were the patterns and amounts of bone loss and how did they lead to a change in management and outcomes).
A second generation of implants has addressed some of the issues relating to poor outcomes and bone loss. Most prostheses are now uncemented and designs are more sophisticated with respect to their kinematics and more appropriate sizing, as well as better intraoperative instrumentation. Early to midterm results are promising, but significant issues are still unsolved regarding modes of failure and their consequences. The most common modes of failure reported in the literature include aseptic loosening, with or without migration and subsidence, infection, stiffness, wound problems, impingement, and bone fractures, both perioperative and postoperative. Of these, aseptic loosening has been reported as the most significant mode of failure, with some authors reporting failure rates as high as 33% for this cause.
Aseptic loosening is not necessarily synonymous with bone loss, although the two terms are often used interchangeably. This is particularly relevant in TAA as unique patterns of osteolysis and subsidence can be seen in follow-up that do not behave in exactly the same pattern as would be seen in hip and knee arthroplasty. More specifically, subsidence of the talar component has been documented to occur without osteolysis or other signs of bone loss, leading Buechel et al. to conclude that the issue of component subsidence and loosening may be more complicated than initially was considered. Furthermore, osteolytic, or radiolucent, lines seen early around components do not necessarily lead to progression and can be present in well-functioning prostheses. Subsidence has specifically been related to undersizing of the talar component in rheumatoid patients. Still, for a prosthesis to become aseptically loose, especially in the uncemented scenario, failure of in-growth or loss of maintenance of in-growth leads to osteolytic or radiolucent lines. Subsidence occurring in the talus without evidence of a radiolucent line must have bone loss by definition through collapse of the supporting bone. Removal of failed components will add an iatrogenic component of bone loss, out of necessity.
From the literature, it appears that the most relevant and common treatment options for failed TAA are revision to another TAA, conversion to an arthrodesis, or amputation. Fortunately, amputation is a rarely reported outcome and is almost always associated with chronic infection as the mode of failure. Several researchers have reported on the success rates for treatment of aseptically loose prosthesis. Factors that influence the type of treatment included the quality of bone, degree of bone loss, previous ipsilateral foot and ankle surgical procedures, patient co-morbidities, and patient choice. Of these, a lack of available bone stock was the most important, since it precluded any type of revision. However, the authors did not report in a standardized fashion specifically on the patterns of bone loss or the actual and/or relative amounts of bone loss. Subsequently, the less experienced surgeon cannot fully benefit from the experience of the surgeons reporting in such papers.
To date, there have been no standardized metrics for assessing or reporting bone loss in TAA. For it to be of value, a classification system should be intuitive and easy to apply, and more important, it should have clinical significance. Clinical significance can be in several forms. First, the classification system should be able to describe in an accurate and valid sense the patterns or types of bone loss so that discriminant outcomes can be assessed. Providing a valid and reliable classification metric would then allow for a predictive outcome metric or, more specifically, a metric that would guide the treating surgeon regarding possible treatment options and their relative success. Also, it would be ideal if the classification system had an evaluative aspect, such as whether the TAA components are loose. Because of the increasing interest in TAA and the ongoing development of implants and techniques, it is important that the surgical community agree on a simple classification system for bone loss. In this chapter, a classification system is proposed that is simple and based on known determinants of poor outcome.
Aseptic loosening of TAA components, with associated osteolysis and bone loss, has been reported to be universally associated with pain, swelling, and difficulty with gait. As such, it is important to keep these diagnoses in the differential when assessing the painful TAA. Obviously, other modes of failure can present in a similar fashion, making the diagnosis of aseptic loosening and associated bone loss dependent on radiography.
Plain radiographs remain the mainstay of determining bone loss in TAA, especially when considered and interpreted in a series of examinations over time. Anteroposterior (AP) and lateral standing radiographs of the ankle should be obtained. Serial radiographs are critical to determine not only whether there is obvious osteolysis or bone loss but also whether there is change in component positioning over time, which is suggestive of component loosening. To determine if there are subtle changes in component position, the examining surgeon should be familiar with the morphology of the implant he or she is studying. This is particularly germane given that there are now more than 20 designs of implants available worldwide, with studies published on over 40 designs. In addition, a standardized reference orientation of the components is important. Several authors have reported on similar methods for assessing change in component position. To assess a change in alignment of the tibial component, a line is drawn on the AP and lateral radiograph parallel to the long axis of the tibia. A line parallel to the undersurface of the tibial component is then drawn and the resulting angles (alpha and beta) are recorded and followed with time ( Fig. 19-1 ). Subsidence of the tibial component can be followed by noting the distance between a fixed osseous reference, such as the tip of the lateral malleolus and a line drawn parallel to the base of the tibial component ( Fig. 19-2 ). Assessing change in angular position and subsidence of the talar component is more difficult in that fixed osseous references are more subjective in interpretation. It has been proposed that angular changes in talar component position can be assessed by drawing a line on the lateral radiograph that connects the anterior and posterior distal most aspect of the prosthesis and then subtend an angle from the posterior reference point anteriorly down the long axis of the talar neck ( Fig. 19-3 ). Subsidence can again be assessed by examining the relationship of the talar component to a fixed osseous reference. This can be complicated. Bestic et al. recommend drawing a line on the lateral radiograph between the anterior and posterior fins of the talar component. A second line is drawn from the calcaneal tubercle to the dorsal aspect of the talar navicular joint. Relative change in the distances between these two lines is suggestive of subsidence. To accurately and reliable assess these references, it may be necessary to use fluoroscopy first to most optimally orientate the ankle for plain radiographic assessment. The interobserver and intraobserver reliability of these measurements has not been reported.
Osteolytic or radiolucent lines have been reported to be significant with respect to poor outcomes. However, the threshold values for pathologic and nonpathologic findings are somewhat subjective and certainly have not been validated or standardized. Osteolytic lines greater than 2 mm have been reported around uncemented tibial components between 10 and 14 years of follow-up, without any of these prostheses being deemed loose. Similarly, osteolytic lines of 2 mm or more were seen around uncemented talar components in the same series. Revisions for severe polyethylene wear revealed the components with osteolytic lines to be stable at time of revision, 8 to 19 years postoperatively. This has been supported by other studies that report that progressive enlarging osteolytic lines greater than 2 mm are significant of pending failure. Still other authors suggest that a circumferential radiolucent line greater than 1 mm is significant. It is difficult to determine the extent of osteolytic lines around the talus because the component often obscures the interface.
Doets et al. report that subsidence of greater than 3 mm was significant for possible aseptic loosening, while others suggest a significant subsidence value of greater than 5 mm of either component. While these numbers are quite subjective, it is intuitively obvious that subsidence leads to bone loss, both through the physical collapse of bone and the associated loss of bone as a result of necessitated revision surgery. Subsidence has been reported more commonly in inflammatory patients, with undersizing of the component, and in those with preexisting osteonecrosis of the talus. As such, patients undergoing TAA for osteonecrosis of the talus should be followed especially closely for talar subsidence.
It is recommended from previous studies that a patient presenting for TAA who has malalignment in any plane at the ankle of greater than 10 degrees should have a corrective osteotomy as the results of TAA in this uncorrected scenario are poor. Post TAA, a serial angular change in component position between 3 and 5 degrees or greater is considered significant. As such, bone loss that could affect the support and alignment of the components, especially the tibial component, should be accounted for. Examples of such bone loss would occur in the medial or lateral plafond.
Some prosthetic systems are designed to be supported by a syndesmotic fusion. Interestingly, delayed union of the syndesmosis (28 ankles) and nonunion of the syndesmosis in these systems have been reported to be associated with the development of lysis around the tibial component. Nonunion of the syndesmosis was also associated with migration of the tibial component and circumferential radiolucency around that component. Cases of impressive ballooning osteolysis around the tibial component in the face of syndesmotic nonunion have been shown by other authors to lead to frank loosening of the tibial component. As this area of bone is critically important in supporting the tibial component and preventing it from subsiding, tilting into valgus, and loosening, it is important to assess the status of the syndesmotic joint in serial follow-up. As not all systems rely on fusion of the syndesmosis, it is again illustrated that it is important for the clinician following the patient to be familiar with the TAA systems he or she is examining. Bone loss specifically of the medial or lateral malleoli has been reported to be significant when it leads to fracture and loss of constraint of the components, especially in certain prosthetic systems.
The role of computed tomography (CT) scanning in assessing bone loss has been examined. Hanna et al. examined 19 TAAs in 17 patients at 2- to 4-year follow-up with plain radiographs and CT scanning with metal-artifact minimization algorithms. The purpose of their study was to determine whether CT was more sensitive than plain films in detecting the presence and extent of periprosthetic lucency. CT scanning identified significantly more lesions of bone loss (29 compared with 18 lesions). CT scanning revealed smaller lesions, and of the lesions that were identified with both CT and plain films, the CT-acquired lesions were three times larger than those seen on plain film. This led the authors to conclude that CT is a more accurate method for early detection and quantification of periprosthetic lucency than plain radiographs.