A study out of Korea (Kim, et al., JBJS-Am, 2012) was a prospective randomized trial of 520 patients undergoing bilateral knee replacement. One knee had computer navigation, the other with conventional instrumentation, all targeting neutral alignment. They had similar survivorships of 98% and 99% between the two cohorts for mechanical loosening at 10 years and they had no difference in knee function, pain, and WOMAC scores.
Do new implant designs make a big difference?
We looked at 527 total knee replacements from 4 centers (Nunley, et al., Knee Proceedings 2014). They were all administered as a telephone survey from a third party, blinded, independent center. We found no difference in patient satisfaction or symptoms with new implant designs such as high flexion, gender-specific, rotating platform versus an old design (10-year cruciate retaining).
What are some of the newer alignment concepts? Well, we talk about constitutional varus, and Johan Bellemans, et al. (CORR, 2012) found that 32% of men and 17% of women have a natural, mechanical alignment of greater than 3 degrees at skeletal maturity. About 2 degrees of valgus of the distal femur and about 3 degrees of proximal tibia varus is their native alignment and thus, they hypothesize that a neutral alignment will be unnatural in a significant proportion of patients and to me that concept makes sense. You’ve been in varus your whole life, if you get corrected to slight valgus or neutral maybe that’s not going to feel normal for you.
In a retrospective study by VanLommel, et al. (KSSTA, 2013), 132 patients with a preoperative varus deformity were followed for a mean of 7 years. Researchers found that patients with mild varus had improved Knee Society scores with no impact on survivorship at that time point. I agree this is a midterm report and we need longer follow up, but a lot of the data that Peter showed are on older implant designs that may not show these same types of results.
Kinematic alignment…where the goal is to align and set the joint line to that of the native knee is another new concept. In these cases you’re building off of the femur and whatever your overall mechanical hip-knee-ankle alignment becomes is really a secondary outcome. So, measured resections are used to restore the joint line based in the thickness of the femoral component.
As Peter said, what you end up with is really an oblique joint line. But if you look at the mean hip-knee- ankle alignment, it’s actually very similar between kinematically-aligned knees and neutral mechanically aligned knees. The difference is that your joint line is more anatomic, or supposedly more anatomic, with slight obliquity.
You also end up with a mean internally rotated femoral component relative to the transepicondylar axis and I would argue that targeting the transepicondylar axis in every single patient may lead to a knee that does not recreate the normal flexion/extension axis of that knee. And 3D CT studies demonstrate that actual flexion/extension axis of the knee is actually slightly internally rotated relative to the transepicondylar axis in a large proportion of patients.
Looking at the clinical results, a prospective-randomized trial (Dossett, et al., BJJ, 2014) looked at mean hip-knee-ankle alignment—which was exactly the same between the kinematically aligned knees and the mechanically-aligned knees. The main difference was that the joint line was more oblique in the kinematically-aligned knees and at 2 years the mean Oxford, WOMAC, Knee Society, and flexion were all improved in the kinematically aligned cohort.
Most surgeons still aim for a neutral component and overall hip-knee-ankle alignment in knee replacement. I’m not going to argue that that is not unreasonable.
I do think, however, that questions remain regarding the optimal target for each individual patient. I think we need to question the norm in stating that just making everybody look the same is going to be okay.
I think the concepts of constitutional varus and kinematic alignment definitely have merit.
