Malalignment in total knee arthroplasty (TKA) is a major source of failure. Concern exists as to the acceptable window for overall coronal alignment in TKA. We evaluated the anatomical coronal alignment of 6070 primary TKAs using standard-length knee radiographs. The mean postoperative alignment was 4.8° of valgus, with 1 standard deviation within the mean defining a range of 2.4° to 7.2° of valgus. The revision rate not related to infection for this well-aligned group was 0.5% (3°=0.47%; 4°=0.54%; 5°=0.47%; 6°=0.39%; 7°=0.62%). In comparison, the failure rate for TKAs aligned in relative varus (<2.4° of valgus) was 1.8% (P=.0004) and for those in valgus (>7.2° of valgus) was 1.5% (P=.0027). Kaplan-Meier survival analysis confirmed these findings. The failure rates were statistically higher for the valgus and varus groups, compared to the well-aligned group within 1 standard deviation of the mean. Moreover, varus-aligned knees failed primarily by medial tibial collapse, whereas valgus-aligned knees failed because of ligamentous laxity. Restoring coronal alignment to between 2.4° and 7.2° of anatomical valgus is the most important surgeon-controlled factor in TKA.
Malalignment has been noted to be the major source of failure in total knee arthroplasty (TKA); however, most of the literature is based on studies using older prosthesis designs. At the 2007 American Association of Hip and Knee Surgeons annual meeting, the Mayo Clinic presented a series of approximately 400 modern TKAs with 15 years of follow-up and full-length standing hip-to-ankle radiographs.1 They found that there was no difference in revision rate based on the coronal alignment. Moreover, they suggested that there was a trend toward better survivorship in the outlier group, although this was not statistically significant.
To answer the question of whether coronal alignment is important in TKA, we must first determine if we need full-length radiographs. Petersen and Engh2 found a 1.4° difference between the tibiofemoral angles measured on standard knee radiographs compared to long-leg radiographs. Although this result was statistically significant, the difference of 1.4° was not clinically significant in light of a standard deviation of 2.2°. There was also a direct relationship between the mechanical axis and the anatomical axis, represented by the tibiofemoral angle.
We looked at a series of approximately 200 TKAs on both long and short radiographs (D.M. Fang, M.A. Ritter, K. Davis, unpublished data, 2009). The 0.8° difference, which was statistically significant (P<.0001), was not clinically significant. A difference of <1° cannot be accurately measured using a handheld goniometer with single-degree markings. Therefore, we feel that standard-length knee radiographs accurately represent overall coronal alignment after TKA. This is therefore the review of 6070 primary TKAs in approximately 4000 patients.
Berend et al3 reported that the failure rate of the tibial component statistically increased when it was positioned in >3.9° of varus. This was accentuated statistically if the body mass index was >33. Our data includes 6 additional years of information than Berend et als3 study and examines all sources of failure, not just tibial-sided failure. The demographics of our patients are similar to most patients who have TKAs.
Mean postoperative alignment was 4.8° of anatomical valgus. One standard deviation above and below the mean is 2.4° to 7.2° of valgus. The failure rate for those within 1 standard deviation was 0.5%, compared to those in <2.4° of valgus at 1.8% and those >7.2° of valgus at 1.5%. The most important finding was that the failure rate was identical for knees aligned at 3°, 4°, 5°, 6°, and 7° of valgus, with a revision rate of <0.5%, as long as they fell within this range.
The Kaplan-Meier survival curve illustrates the same findings. The failure rates were higher for those in the valgus and varus groups, as would be expected. In the varus aligned group, there was 7 times the chance of failure by collapse on the medial proximal tibia. On the valgus side, there was 4 times the risk of failure by ligamentous instability. These results were statistically significant.
The overall coronal alignment and tibial component positions are both important. When we looked at these factors, we found the overall anatomic alignment was better than the tibial component alignment in predicting overall survival (P=.005.) Restoring coronal alignment to between 2.4° and 7.2° of valgus is the most important surgeon-controlled factor in TKA. Alignment can be achieved without computer navigation, as long as the tibial component is perpendicular and the anatomic alignment is within this 5° range. A recent study suggested that computer navigation may improve the accuracy of alignment, but this comes with higher expenses, increased operating room times, and more complications.4 Clinical experience and sound judgment cannot be replaced by a computer.
- Parratte S, Pagnano MW, Trousdale RT, Berry DJ. The mechanical axis may be the wrong target in CAS TKA: 15-year survival of 399 modern TKA: somewhat better for so-called outliers. Paper presented at: American Association of Hip and Knee Surgeons Seventeenth Annual Fall Meeting; November 2-4, 2007; Dallas, Texas.
- Petersen TL, Engh GA. Radiographic assessment of knee alignment after total knee arthroplasty. J Arthroplasty. 1988; 3(1):67-72.
- Berend ME, Ritter MA, Meding JB, et al. Tibial component failure mechanisms in total knee arthroplasty. Clin Orthop Relat Res. 2004; (428):26-34.
- Bonutti PM, Dethmers D, Ulrich SD, Seyler TM, Mont MA. Computer navigation-assisted versus minimally invasive TKA: benefits and drawbacks. Clin Orthop Relat Res. 2008; 466(11):2756-2762.
Dr Fang is from the Department of Orthopedics, Indiana University School of Medicine, Indianapolis, and Dr Ritter is from the Center for Hip & Knee Surgery, St Francis Mooresville, Indiana.
Dr Fang has no relevant financial relationships to disclose. Dr Ritter belongs to Joint Replacement Surgeons of Indiana, which receives monetary support from Biomet for information that can be extracted from this article.
Presented at Current Concepts in Joint Replacement 2008 Winter Meeting; December 10-13, 2008; Orlando, Florida.
Correspondence should be addressed to: Merrill A. Ritter, MD, Center for Hip & Knee Surgery, 1199 Hadley Rd, Mooresville, IN 46158.