Haas v. MacDonald: Robotic Assisted TKA: The Future Is Now

This week’s Orthopaedic Crossfire® debate was part of the 34th Annual Current Concepts in Joint Replacement® (CCJR®), Winter meeting, which took place in Orlando. This week’s topic is “Haas v. MacDonald: Robotic Assisted TKA: The Future Is Now.” For is Steven B. Haas, M.D., Hospital or Special Surgery, New York, New York. Opposing is Steven J. MacDonald, M.D., F.R.C.S.(C), University of Western Ontario -London, Ontario, Canada. Moderating is Thomas S. Thornhill, M.D., Harvard Medical School, Boston, Massachusetts.

Dr. Haas: I’m debating Dr. MacDonald and he’s going to be saying “Bah humbug.”

Dr. MacDonald is someone who embraces technology. For example, he brought his car from Canada down here. He took some people out to lunch in his 1975 Pacer. Yes, Dr. MacDonald embraces new technology.

The rationale for robotics in total knee arthroplasty—and you can look at unicompartmental as an example which is associated with lots of failures related to technical error—like undercorrection, overcorrection, varus angulation, abnormal posterior slopes—which can be corrected with the use of computer-assisted technology for navigation and robotics.

With robotics positioning is much better and you can achieve the goals of alignment that you need to do. That was a major cause for failure of unicompartmental knee arthroplasty. Articles from Hospital for Special Surgery by Dr. Pearle actually have shown that survivorship and results are much better with the use of the robotic technology (J Arthroplasty 2018). His 5-year data is showing improved results with the use of robots.

What about the rationale for a total knee? The major causes for failure in the short term are malalignment and instability: 25% and 21%, respectively, as reported by Nam (J Arthroplasty 2014) and Sharkey (CORR 2002).

If you include accelerometer technology (Nam, J Arthroplasty 2014), it improves tibial component positioning dramatically. You can get over 95% alignment within 2-degrees of neutral with it as opposed to the 60% or 70% without it.

Now, let’s look at the patient population. Obesity is a huge epidemic in the United States and we are taking patients to the OR that have a BMI [body mass index] of 40, 50, even 60.

I will guarantee that aligning these knees is much more difficult and there is data to show that the alignment in those cases is much more likely to be varus, and that they have a much higher rate of failure if you place it in varus (Lustig, Knee Society 2016).

About 32% of the cases of total knees done without navigation or any type of aids are malaligned. Whereas less than 10% with enabling technologies of one variety or another are malaligned. Those are just the facts.

There are a bunch of enabling technologies. There’s computer navigation, gyroscopes, OrthAlign, patient specific robotic surgery with MAKO, NAVIO. Lots of different choices for enabling technologies that are available.

So, I’m sure Dr. MacDonald is going to tell us that navigation has never been shown to improve outcome, but most of these papers are short-term, most of them are underpowered, they’re all from major centers where there are very highly-skilled surgeons.

But on the other hand, in the generalizable surgeon population, enabling technologies may be necessary. If you look at registry data which is large, generalizable data, the Australian Registry showed a decrease in the rate of failure in younger patients (de Steiger, JBJS-Am 2015).

That’s who you care about because those are the ones that we worry about the most, those are the ones where you are going to see the problems right away. Navigation improved alignment, but what robotics adds is that it can do soft tissue balance which was always the thing that was missing from navigation. You can get sizing, soft tissue balance, and positioning.

Multiple articles have shown that the robotics do improve the early results and the robotic surgery can do lots of things. It can do planning, gap balancing, you can also use it to aid you in the execution of your operative plan.

In conclusion, innovation is good and necessary. Enabling technologies have improved clinical outcome and this is even more important as we go on and do many younger and obese patients and I think many new knee designs require a more individualized approach—as we learn what’s best, we can individualize it.

So, the future I would say is now in robotics and enabling technology.

Dr. MacDonald: My opponent has many accomplishments and accolades for certain. He’s a friend with impeccable fashion sense unquestionably. But since he got into the robotics scene something has changed, and I don’t know what exactly it is.

He’s confused. We need to be very clear about what issue we are debating. It’s not navigation. Instead, it is questionable added benefit of robotically assisted navigation.

We have many enabling technology choices today.

We can use: conventional instrumentation, patient-specific jigs, image-free navigation (where you intra-operatively do the registration with no pre-op imaging), image-guided navigation (where you get a pre-op CT or MRI or intra-operative fluoro), hand-held navigation and then lastly, there is robotic navigation—which is the topic of the debate.

With robotic navigation you generally get a pre-op CT and then guides that take you through intra-operative cuts or milling.

There is lots of hype regarding navigation so here’s a quote, “Long-term cost savings result from shorter hospitalization, decreased morbidity, improved joint stability and decreased rehab time.”

The problem is that’s from a sales brochure from 13 years ago so where have we moved this needle? The first navigated knee was 20 years ago. Yes, technologies evolve—as do techniques. In fact, almost everything changes over 20 years.

After 20 years of navigated total knees there is still no definitive proof. If we can at least agree to aim for a neutral mechanical axis, you’re trying to go down the middle of the runway, plus or minus 3 degrees (which is what we’ve generally aimed for historically) and we want to avoid the outlying. No question, navigation gets you there.

Multiple studies dating back 10 years or more show that you will have a tighter standard deviation and much fewer outliers with navigation.

Meta-analysis shows the same thing. Every paper favors navigation for achieving your alignment goals. Robert Barrack (CORR 2013) did an excellent review and found decreased outliers, improved mechanical alignment, no consistent difference in functional outcomes with navigation.

Also, as Steve mentioned, there are no long-term studies, just short- and medium-term studies. I’m a data-driven guy. The Australian Registry shows if you’re less than 65 years of age and if you navigate versus non-navigate a knee, there’s about a 1% difference at 9 years. About 30% of knees done in Australia right now are navigated compared to North America where we’ve gone from 5% to 3%, and most of our navigation machines look like they’re a glorified closet in our operating room.

So, I’ll freely concede that if you navigate every knee, you’ll have fewer outliers and a tighter distribution of alignment. But what does a robot add to this whole thing?

First of all, it adds an unbelievable cost. It’s hard to get the prices exactly but a typical robot assist unit costs between $1.5-2 million dollars. Annual service agreement of about $150,000; disposables per case, about $1,000. For that cost there must be a large body of evidence for the robot. Right?

If you PubMed search “robotic-assisted total knee,” you get the sum total of 30 papers so it doesn’t take you long to review it and a lot of these are review articles.

What is it show? It doesn’t ever show them comparing navigation to robotics, it’s always conventional to robotic. In one particular article (Bargar et al, CORR 2013), yes you have less outliers with robotics, but this is basically a navigation system. The robot took 25 minutes longer with no difference in PROMS [Patient Reported Outcome Measures].

Another series looked at 30 patients’ bilateral knees conventional vs. robotic (Song et al, Knee Surg Sports Traumatol Arthrosc 2011), the same thing as navigation, less outliers. The robot, again, took 25 minutes longer to do.

Another series retrospectively looked at it (Hansen et al, J Arthroplasty 2014), no difference in clinical or radiographic outcomes. The robot case took 20 minutes longer and the conclusion is further surgical, clinical, and economic study of this technology is necessary.

A review article (Jacofsky et al, J Arthroplasty 2016) [said], “further studies are needed to confirm the cost effectiveness of this” and a third article all published in the last year or two (Banerjee et al, Expert Rev Med Devices 2015) [said], “studies evaluating the projected cost-benefit analyses of this new technology are needed before widespread adoption.”

I could find one paper that supported the use of robotics over conventional with a cost basis (Moschetti et al, J Arthroplasty 2016). The assumptions in the paper were the following; the non-robotic UKA had to have triple the failure rate, unproven, you must do 100 UKAs per year, and then if you actually read the discussion it said, “robot assisted UKAs were more costly than traditional surgery but offered a slightly better outcome.” Again, unproven.

So, in conclusion, the robotic TKA, it’s got no evidence for superiority over navigation, none whatsoever. It does not appear to have any measurable clinical outcome benefits over navigation, but it does come with a substantial increased cost and OR time and I would say there’s really no valid indication for robotic total knee.

Moderator Thornhill: Okay so we may be in a little bit of confusion of semantics, so Steve define what “robotics” mean to you.

Dr. Haas: I think robotics are using some enabling technology to set alignment and then some aid to help you also prepare the bone as well. It doesn’t have to be a big robotic arm, there are different kinds of robots, so it doesn’t have to be narrowly defined. It has to help you with navigation and aligning the knee, it should help you with some balancing and it should aid you in how you prepare the bone.

Moderator Thornhill: Okay, so you saw the list that Steve was calling navigation. Are any of those robotics?

Dr. Haas: Well, they’re part robot. They’re half robot.

Moderator Thornhill: You must have a comment on that MacDonald.

Dr. MacDonald: It’s not really that complicated. You either have a robot or you don’t. So, I was going to offer my services to Haas, I’d come to his OR, and for not even $2 million, for a million dollars, I’ll put a retractor in to prevent cutting the MCL [medial collateral ligament]. I mean, that’s kind of what we’re talking about, so it’s got special saws, it cuts the bone for you, and there’s a new one coming out that will theoretically burr some entry holes and it’ll guide it from there. That’s really robotic surgery.

Moderator Thornhill: Okay so you’re defining this as anything that gives you enabling technology in addition to the classic way of doing knees?

Dr. Haas: Well, that’s right because the robot does not have to make the bone cut for you. If the robot sets your alignment guides and you make the cut, it’s still a robot.

Moderator Thornhill: Let’s talk about cost versus need. Sometimes the people that may benefit more from this are the ones that don’t do as many. And cost is a big issue.

Dr. MacDonald: I don’t think anyone benefits from robotic technology. I think there’s people who benefit from navigation, which is completely different, and the evidence is registry-based and it’s pretty sound. I think your point being that if you do 10 knees a year versus 200 you’re going to see the benefit from tightening your outliers. The evidence for that isn’t strong but that’s what logic would dictate.

Dr. Haas: There are navigation systems with accelerometers that give you very accurate results, and those are little tiny devices. The idea that there is only one model of this I think that’s where we were wrong.

Dr. MacDonald: I mean yeah, I guess I was just basing my debate on now—not where we’re going to be in 2020.

Moderator Thornhill: How many people use some form of navigation rather than just standard instruments for their total knees, hands up. Okay so it’s not that many. What is going to be the tipping point that’s going to make us go to robotics?

Dr. Haas: I think that the cost of the current technology is too expensive but a million-dollar robot I think is not going to be what everybody’s going to use. Cost has to come down.

Dr. MacDonald: The thing that’s going to drive it, like most technologies, is going to be cost, I would agree with Steve there. I mean if you can get this in your OR at a reasonable cost, but I do think that if you’re in the robotics field, you need to show the value proposition above the more inexpensive enabling technologies that we both talked about.

Moderator Thornhill: What percentage of this is marketing?

Dr. MacDonald: At the moment, 100%.

Dr. Haas: Right, I think that in the total knees, a fair amount is marketing.

Moderator Thornhill: Alright, okay so thank you gentlemen for a nice discussion.

Please visit www.CCJR.com to register for the 2019 CCJR Spring Meeting, – May 8-11 in Cleveland, Ohio.