TEC Assessment Index

Computer-Assisted Navigation for Total Knee Arthroplasty

Assessment Program
Volume 22, No. 10
November 2007

Executive Summary

Background

Total knee replacement, or total knee arthroplasty (TKA), has achieved widespread success for the treatment of degenerative arthritis. However, a small percentage of patients who undergo TKA experience poor outcomes, often requiring TKA revision surgery. Improper alignment of the prosthetic implants has been proposed as one reason for poor outcomes following TKA. Computer-assisted navigation for TKA is a refinement of conventional TKA that is intended to improve the accuracy of implant alignment.

Objective

To determine whether computer-assisted navigation improves the accuracy of implant alignment for TKA, and whether the amount of improvement in alignment results in meaningful improvements in health outcomes, such as pain, function, or revision surgery.

Search Strategy

A MEDLINE® search was conducted for the periods of January 1980 to August 2007. For randomized, controlled trials evaluating implant alignment, the following search was used: ((computer or navigation OR navigated) AND total AND knee) AND ((clinical[Title/Abstract] AND trial[Title/Abstract]) OR clinical trials[MeSH® Terms] OR clinical trial[Publication Type] OR random*[Title/Abstract] OR random allocation[MeSH® Terms] OR therapeutic use[MeSH® Subheading]). For studies of the relationship between alignment and clinical outcomes the keywords (“total knee arthroplasty” OR “TKA” OR “total knee replacement”) were cross-referenced with the terms (“alignment” OR “malalignment” OR “revision”).

Selection Criteria

For evaluation of the accuracy of implant alignment, selection criteria were: 1) randomized, controlled trials (RCTs) with at least 25 patients per treatment arm, 2) used tricompartmental knee replacement done by the open approach, and 3) compared alignment of prosthetic implants between groups. For the association of malalignment with clinical outcomes, selection criteria were: 1) cohort studies that evaluated postoperative alignment as a predictor of future poor outcomes, such as the need for revision surgery; 2) case-control studies that evaluated differences in alignment between patients with poor outcomes and patients with good outcomes; and 3) case series of patients with poor outcomes following TKA that reported the percentage of patients who had component malalignment.

Main Results

For the outcome of overall limb alignment, 4 of 7 trials reported a statistically significant reduction in the number of patients with malalignment for the computer-assisted navigation group. A fixed-effect, inverse-variance-weighted pooling of data from these 7 trials suggests that malalignment of greater than 3 degrees will be avoided in approximately 15.5% of patients (95% CI: 9.4–21.7% when computer-assisted navigation is used; number needed to treat of 6.45 to avoid one instance of malalignment [95% CI: 4.61–10.69]). A similar pattern of results was reported for other measures of alignment (e.g., alignment of tibial components alone or alignment of femoral component alone).

Operating time was compared between groups in 7 of the reports from the RCTs, and 5 of the 7 reported significantly increased operating room times for the computer-assisted navigation group. The increase in time ranged from 5–34 minutes. Five trials compared the amount of blood loss between groups, and 2 of the 5 reported a statistically significant decrease in blood loss for the computer-assisted navigation group. The difference in blood loss in those two trials was approximately 100–200 mL. None of the trials meeting the article selection criteria reported on other postoperative complications such as thromboembolism or fat embolism.

Seven publications included comparisons of functional outcomes, including 3 that followed patients for up to 2–3 years. In these studies, there were no differences between groups on any of the functional status outcomes. These data establish that there are not large differences in functional outcomes between conventional TKA and TKA performed with computer-assisted navigation at 2–3 years of follow-up. However, none of these trials was adequately powered to detect small differences that might be expected on these functional outcome measures.

For the question of whether improper alignment leads to meaningful differences in clinically relevant outcomes, 5 cohort studies were identified. One of these evaluated postoperative alignment prospectively as a risk factor for revision surgery using multivariate analysis. This study reported a strong relationship between malalignment and outcomes, with malalignment of greater than 3 degrees associated with a greater than 17-fold risk for revision surgery. However, a number of methodologic limitations reduce confidence in the validity of these reported results. First, the study covered a long time period from 1983–2000 and thus was subject to maturation effects arising from changes in TKA surgery over this period. There were only a relatively small number of patients undergoing revision surgery (n=41), thereby limiting the power to evaluate the many potential predictors of revision surgery. Finally, the analysis was done using data-mining techniques and iterative modeling that likely resulted in an overestimate of the true predictive ability of malalignment.

The other 4 cohort studies divided patients into groups based on postoperative alignment (e.g., varus, neutral, and valgus alignment) and compared revision rates among the group. In 3 of the 4 studies, there was a significantly increased revision rate for the group of patients who had varus postoperative alignment. Two small case-control studies compared rotational alignment among patients with poor outcomes and those with good outcomes. Both studies reported significant differences in rotational alignment between groups, with excess internal rotation noted in the group of patients with poor outcome. Four case series of patients undergoing TKA revisions were identified: these studies were consistent in reporting that approximately 10% of patients had evidence for implant malalignment (range: 9.4–12%).

Author's Conclusions and Comments

The RCTs allow conclusions on postoperative alignment outcomes. These RCTs are relatively consistent in demonstrating a reduction in alignment outliers for the computer-assisted navigation group across different studies and different measures of alignment. It is possible to conclude that approximately 15.5% of patients may avoid malalignment of more than 3 degrees in overall limb alignment when computer-assisted navigation is used. It is also possible to conclude that computer-assisted navigation is associated with longer operating times. The evidence on blood loss is less certain, in that some studies report a small reduction in blood loss, but others do not.

There is no direct evidence to evaluate whether long-term functional outcomes are improved by computer-assisted navigation. The RCTs that report on short- to intermediate-term functional outcomes do not report improvements associated with computer-assisted navigation. This establishes that there is not a large difference in functional outcomes associated with computer-assisted navigation over a 2- to 3-year follow-up period. However, these small RCTs are inadequately powered and have an insufficient length of follow-up to detect the smaller differences in functional status scores that might be expected from computer-assisted navigation when evaluating the entire population of patients undergoing the procedure. Since the vast majority of patients do well following TKA and only a small minority has poor outcomes, larger RCTs with longer follow-up will be required to demonstrate improved outcomes as measured by standardized knee rating scores.

The available evidence on the relationship between malalignment and clinical outcomes consists of observational studies of various designs. Only one of these observational studies was a prospective cohort study that used postoperative alignment as a predictor of future poor outcomes in a multivariate analysis, and this study had only a relatively small number of outcomes (n=41). This study did show a strong relationship between malalignment and poor outcomes. Four other cohort studies, which were generally much older studies, reported that outcomes were worse for patients who had postoperative varus alignment.

The case-control studies and case series that evaluated the association of alignment with outcomes represent weak study designs to answer the specific questions. These studies did not use postoperative malalignment to predict outcomes, but rather evaluated alignment at the time of the poor outcomes. The case-control studies did report an association between rotational malalignment and clinical outcomes, and the case series suggested that approximately 10% of patients undergoing TKA revision surgery have malalignment at the time of revision surgery.

Thus, while RCTs suggest that approximately 15.5% of patients may avoid malalignment of greater than 3 degrees by this analysis, it is not possible to conclude that all these patients benefit from computer-assisted navigation. The threshold definition for malalignment is derived from older studies. It is not certain that this threshold is the most clinically relevant definition of malalignment in the current era of TKA. There have been many advances in TKA that might mitigate the impact of malalignment seen in earlier studies. In addition, there is a lack of clinical studies that define the threshold for malalignment, and there may be an interaction between alignment and other risks for poor outcome, such as obesity.

The positive associations reported across different types of observational studies suggest that there is a relationship between malalignment and poor outcomes. However, as a result of deficiencies in the available evidence, it is not possible to test this hypothesis or to determine whether the degree of improvement in alignment reported in the RCTs leads to meaningful improvements in health outcomes, such as pain, function, or revision surgery. Therefore, it is not possible to conclude that the use of computer-assisted navigation with TKA leads to improved health outcomes.

Based on the available evidence, the Blue Cross and Blue Shield Association Medical Advisory Panel made the following judgments about whether the use of computer-assisted navigation for total knee arthroplasty (TKA) meets the Blue Cross and Blue Shield Association’s Technology Evaluation Center (TEC) criteria.

1. The technology must have final approval from the appropriate governmental regulatory bodies.

TKA is a surgical procedure that is not subject to U.S. Food and Drug Administration (FDA) regulations. Several systems for computer-assisted navigation have been cleared for marketing by the FDA via the 510(k) process (e.g., PiGalileo™ Computer-Assisted Orthopedic Surgery System, PLUS Orthopedics; OrthoPilot® Navigation System, Braun; Navitrack® Navigation System, ORTHOsoft).

2. The scientific evidence must permit conclusions concerning the effect of the technology on health outcomes.

The evidence is sufficient to conclude that the use of computer-assisted navigation with TKA results in more accurate implant alignment. This conclusion is derived from RCT evidence comparing TKA using computer-assisted navigation with conventional TKA and reporting on the number of patients in each group with malalignment. Approximately 15.5% of patients may avoid malalignment of greater than 3 degrees in overall limb alignment with the use of computer-assisted navigation.

The evidence is not sufficient to conclude that the improvement in alignment associated with computer-assisted navigation leads to meaningful differences in health outcomes, such as pain, function, and revision surgery. Long-term evidence from RCTs is not available to answer this question. Observational studies that evaluate the association between alignment and clinical outcomes consistently report an association between malalignment and poor outcomes. However, these studies have a variety of methodologic limitations and, by their nature, are hypothesis generating. Thus, the evidence is not sufficient to determine that the degree of improvement in alignment reported in the RCTs leads to a meaningful benefit in health outcome.

3. The technology must improve the net health outcome; and
4. The technology must be as beneficial as any established alternatives.

Evidence is not sufficient to permit conclusions as to whether computer-assisted navigation improves the net health outcome or is as beneficial as conventional alignment techniques.

5. The improvement must be attainable outside the investigational settings.

It cannot be determined whether any improvement is attainable outside the investigational setting since the evidence is not sufficient to permit conclusions on the effect of computer-assisted navigation on health outcomes.

For the above reasons, the use of computer-assisted navigation for total knee arthroplasty does not meet the TEC criteria.

TEC Assessment Index

KEYWORDS: alignment; angle; anteroposterior; arthritis; biomechanics; CAN; cohort; computer assistance; computer-assisted navigation; condyle; degenerative joint disease; device; DJD; embolism; femoral; implant; intramedullary guides; jig; kinematics; knee rating scores; malalignment; malrotation; navigated; orthopaedic; orthopaedics; osteoarthritis; pain function; patella; patellar; prosthesis; prosthetic; revision; rotation; surgery; thromboembolism; tibia; tibial; tibiofemoral; tissue balancing; TKA; TKR; total knee arthroplasty; total knee replacement; tracking; tricompartmental; valgus; varus;