In the early 1980s, posterior cruciate-stabilizing implants were introduced to improve knee flexion by controlling femoral rollback. An additional benefit anticipated with this more congruent implant was a reduction in polyethylene wear. Previous authors have reported excellent intermediate and long-term results in association with both cruciate-retaining and cruciate-substituting designs in outcome studies with limited sample sizes and without direct comparisons1-5. Deliberation of the benefits of posterior cruciate-stabilizing designs as compared with cruciate-retaining designs continues today, with the debate now in its fourth decade.
The strengths of the study by Abdel et al. include the unbiased look at registry information from a single institution and the significance of the data. Surprisingly, the posterior cruciate-stabilizing implants in this study had a worse outcome in terms of survivorship than did the implants that retained the posterior cruciate ligament. From the conclusions and supporting data in this paper, we must question why the posterior-stabilizing implants had a higher failure rate than the posterior cruciate-retaining designs under review.
The failure of total knee implants by aseptic loosening can be impacted by patient variables, surgeon variables, and implant materials and design variables. In this study, the patient variables were accounted for as the groups were similar in age, sex, and preoperative diagnosis. The surgeon variable was controlled as the data were from skilled surgeons who performed at least fifty knee arthroplasties per year. Therefore, the plausible explanation for the differences in survivorship was differences in either the implant design or materials between retaining and stabilizing components.
Design differences may have contributed to implant failures. The most common mechanism of failure in this study was aseptic loosening. Knee stability and kinematics with cruciate-retaining implants are achieved with minimally congruent articular surfaces and surgical techniques that balance and retain the supporting ligaments. Posterior cruciate-stabilizing implants have more congruent articular surfaces and a post-and-cam mechanism that provides stability and controls sagittal plane kinematics. Stresses at the fixation interface would have been higher with posterior-stabilized implants because knee stability is achieved through interaction between the components. The higher aseptic loosening rate associated with the posterior-stabilized implants might have been caused by greater stress at the bone-cement interface.
The second and third most common mechanisms of failure were wear and osteolysis, respectively. The contact stresses would have been lower and wear debris should have been less with the more congruent posterior-stabilized implants. However, polyethylene debris might have been created by the interaction between the tibial post and the cam mechanism that creates femoral rollback. The wear debris generated from a more congruent implant articulation may have been smaller, more bioreactive, and more likely to create osteolysis. A more congruent implant also increases stress on the modular tibial locking mechanism. This may have increased tibial insert backside wear, contributing additional debris.
Material differences might also explain the lower rate of survival of stabilized implants. A major weakness of this study is the lack of information on the method of sterilization and the shelf age of the different implants under investigation. Oxidation secondary to shelf aging of gamma-irradiated-in-air polyethylene dramatically impacts the survival of both unicondylar and tricompartmental knee implants6-8. This study transcends a time interval when implant manufacturers were eliminating sterilization by gamma irradiation in air. However, through the 1990s, many implants in inventory were sterilized with this method and were used as long as the shelf life was less than five years. In the early 1990s, posterior cruciate-retaining designs were more popular and were implanted in large numbers. These implants may have had a considerably shorter shelf age and minimal oxidation at the time of implantation. The cruciate-stabilizing implants may have become oxidized by remaining on a storage shelf until later in the decade when the posterior-stabilized designs gained popularity at this institution. The authors of this study apparently did not have access to shelf-age information. Examination of the failed components to evaluate fatigue wear secondary to oxidation may have answered this question as well as questions about post and backside wear.
In the discussion, the authors point out that most outcome studies suffer from small sample size with lack of extensive follow-up and the limitations of a retrospective review. The same is true of this study. At fifteen years, only fifty-two (2%) of the 2728 stabilizing implants were reviewed retrospectively. This accounts for the large confidence interval in this subgroup. Many, if not most, of these knees may have been followed only by a questionnaire or phone call as only 42% had an actual office visit.
Further investigation and survivorship analysis of a larger number of implants with long-term follow-up is needed to evaluate the variables that contributed to the higher failure rates with the posterior cruciate-stabilizing total knee implants in this study. If the method of implant sterilization and the shelf age of the implants were similar for the two groups in this study, the data would fully support the argument that failure was secondary to implant design.