Low-friction metal-on-polyethylene primary total knee arthroplasty has been utilized for forty years, and anterior knee pain has been an issue for all of that time. The first implants either ignored the patellofemoral joint or even compromised it by means of a transverse metal bar that joined the medial and lateral condylar femoral components. Anterior knee pain was common, developing in 40% to 58% of patients1. As a result, attempts were made to address the patellofemoral joint in the design of future total knee arthroplasties. The first initiative was to add an anterior flange to the femoral component, and this was followed shortly by the addition of a polyethylene resurfacing button to the articular surface of the patella that articulated with the anterior flange of the femoral component2. The prevalence of anterior knee pain was greatly reduced by these changes, but it was not completely eliminated.
Since that time, continued patellofemoral-related changes have been made, such as a move to specific right and left femoral components, closer replication of normal patellar tracking, and deepening of the patellar groove in the femoral component3. Despite these improvements, consensus was not reached with regard to how to deal with the patellofemoral joint during total knee arthroplasty. Complications (rates of 4% to 50%) associated with patellofemoral resurfacing began to emerge, prompting some surgeons to abandon patellofemoral resurfacing4. As a result, the management of the patellofemoral joint at the time of primary total knee replacement has been inconsistent, with some surgeons always resurfacing the patella, others never resurfacing the patella, and others selectively resurfacing the patella. Proponents of not resurfacing the patella have influenced femoral component design, prompting development of more anatomic designs and designs with support of the patella through 90° of knee flexion, the so-called "patella-friendly" total knee arthroplasty5.
In order to shed additional light on this complex issue, Pavlou et al. performed a meta-analysis of eighteen Level-I randomized controlled trials in which patellar resurfacing during total knee replacement (n = 3463) was compared with not resurfacing the patella during total knee replacement (n = 3612). The review of the current literature is excellent. The authors did acknowledge a number of potential weaknesses of their analysis that they thought should be considered when interpreting the study conclusions. The weaknesses were that their study only assessed the "always resurface" versus the "never resurface" operative approaches, without addressing selective resurfacing; that their meta-analysis depended on the quality of the randomized studies included; that many implant designs were combined together with the assumption that all were the same; and that the follow-up times varied.
Pavlou et al. found no significant differences in the prevalence of anterior knee pain or functional outcome scores between patients who had had the patella resurfaced during the total knee arthroplasty and those who had not. Higher rates of reoperations were noted in the not-resurfaced cohort (p = 0.012), but the authors thought that this might be attributed to the fact that patients who did not have the patella resurfaced and had postoperative anterior knee pain had the option of secondary patellar resurfacing, an option not available to patients in whom the patella had already been resurfaced. The authors noted that the literature suggested that secondary resurfacing was often not successful in relieving anterior knee pain.
Analysis of "patella-friendly" versus "non-patella-friendly" designs demonstrated no differences in reoperation rates among patients who had not had the patella resurfaced. Considerable variations were noted among studies with regard to anterior knee pain and functional outcomes, but the aggregate findings did not suggest a significant difference between the "patella-friendly" and "non-patella-friendly" designs.
My assessment of this meta-analysis is that it is well done and provides a good description of current thinking with regard to patellar resurfacing during primary total knee arthroplasty. Unfortunately, after reading this study, I remained uncertain about whether the patella should or should not be resurfaced during contemporary primary total knee arthroplasty. This confusion is largely due to a lack of information in the literature. First, most studies have not addressed the fact that anterior knee pain following total knee arthroplasty may be caused by factors other than whether the patella was resurfaced or not. Other etiological factors for anterior knee pain include incision discomfort, neuromas, loss of sensation, bursitis, tendinitis, patellar instability, and fracture1. Second, most of the total knee replacements analyzed in this study are no longer used. Whether the findings of this study can be generalized to contemporary implants is debatable. Third, there is a paucity of validated outcome tools with which to assess patellofemoral pain and function. As a result, the outcome tools used in this meta-analysis might not have been able to accurately detect important differences in patellofemoral pain relief and functional improvement among the implant types studied.
As we move forward, whether or not to resurface the patella during total knee arthroplasty will remain a subject of much debate. From my point of view, future randomized controlled trials on this subject will need to address not only the strategies of "always resurface" and "never resurface," but also "selective resurfacing" of the patella. We will also need better patellofemoral-specific outcome tools and implant-specific data for contemporary implants used in these randomized controlled trials. Finally, greater use of National Registry joint replacement data should be encouraged to provide implant-specific data on patella-related revision rates for those surgeons who "always resurface," "never resurface," or "selectively resurface" the patella during primary total knee arthroplasty.