Robotic arm-assisted versus manual unicompartmental knee arthroplasty

Listen to Andrew Duckworth, Fares Haddad, Nick Clement and Mark Blyth discuss the paper 'Robotic arm-assisted versus manual unicompartmental knee arthroplasty' published in the May 2022 issue of The Bone & Joint Journal.

Click here to read the paper.

Find out as soon as the next episode is live by following us on Twitter, Instagram, LinkedIn or Facebook!

Transcript

[00:00:00] Welcome everyone to our BJJ podcast for the month of May. I'm Andrew Duckworth and a warm welcome from your team here at The Bone & Joint Journal, as always, we'd like to start by thanking all of you for your continued comments and support as well as a big, thanks to our many authors and colleagues who have taken part.

We hope you're continuing to enjoy our podcast and all the knowledge translation work we're delivering from your team here at the journal. Our podcast continue to focus on papers published each month here at the BJJ, as well as our special edition podcast series. That includes our Insights from the US series, along with our Specialty Editor series with our invaluable Speciality Editors here at the journal, both of which we are continuing this year.

So today for our monthly podcast, I have the pleasure of being joined by three authors from a paper published in this month's edition of the BJJ entitled 'Robotic arm-assisted versus manual unicompartmental knee arthoplasty: a systematic review and meta-analysis of the MAKO robotic system'. So firstly, I'm very pleased to be joined by Mr.

Nick Clement, who is a fellow editorial board member here at the BJJ. And one of my great colleagues and friends here in Edinburgh. Thanks for joining us, Nick. It's great to have you with us. [00:01:00] Brilliant, thanks for inviting me back Andrew. I didn't think you'd invite me back after last time. Secondly, we have the me and Nick have the pleasure of being joined by one of our great colleagues from the west, Mr.

Mark Blythe from Glasgow, Mark. Great to have you with us today. Delighted to be here. Thank you for having me. Rounding off our lineup lastly, but certainly not least. We are delighted to welcome back a great Editor-in-Chief here at the BJJ, Professor Fares Haddad. Great to have you back. Andrew, thank you for having me and thank you for doing this.

So I thought we'd kick off guys with, actually with yourself Prof, if that's alright. I, I thought you maybe could give us a brief overview of your experience so far with robotic knee replacements and how you feel the literature has sort of evolved and developed over the past few years. Yeah. Thanks.

It's been a fa- it's really been a fascinating area, which has grown exponentially looks like it really is here to stay. So certainly from a London and England setting my experience of the, the current systems, if you like of M AKO dates back to 2016, when we got the first clinical system [00:02:00] in the UK and which after Mark's research system that which has had in ed in Glasgow and done some very nice research with, so we've got now five and a half years of experience with it, starting with the uni.

And then adopting the total knee and increasingly using the total hip with the modern software. It's, it's been a fascinating learning curve in terms of understanding the power of this technology and quite what it can deliver in terms of the three-dimensional planning in terms of understanding individualized targets

and instead in terms of being able to execute them. And I think the literature's basically paralleled in my experience, it's been initially started with studies, focused on accuracy and precision, and the ability to hit the. Then there were sort of short term clinical studies looking at the perioperative pathway, early patient outcomes, patient recoveries, the pain response, the need for fewer opiates than standard procedures.

So [00:03:00] documenting some signals and as we've gone along, the studies have increased in number and become more sophisticated. So now we're seeing more granular patient based outcomes. We're seeing longer term outcomes, both clinically and radiographically. We are, I can tell you in the background starting to see five-year PROMs, we're starting to see some registry data.

And we're starting to see the randomized studies beyond the, the excellent marked life studies starting to report, including one of our own pilot studies that we published in the journal last year, but the others now closing to recruitment, and we've also seen some bigger studies start to recruit like the RACER study.

So I think the literature here is becoming richer. For this one device, I think it's important to stress that the robotic explosion goes beyond this one device. It's really the whole enhanced technology field is growing. But each device needs its own data, its own evidence and its own literature. And certainly for this one particular device, the MAKO, we've seen an explosion of interest and [00:04:00] a great deal of data now.

Yeah, absolutely Prof. Like you say, it's, it's very important to say that this is very much on the MAKO robotic system. We're going to be talking about for this paper, but like you say, it's so important that each of the systems have their own evidence and they have their own build up of literature. But I agree with you, it like a lot of, maybe not maybe potentially false dawns that we see in our, in our field where you see something exciting comes along, but actually it doesn't stay all the evidence and support it.

This really has just kicked on. Hasn't it really. I mean, it's been incredibly impressive. I mean, I think it's interesting because we saw navigation come and to a certain extent go. And I think there is a difference in parallel here in that we're looking more three-dimensionally and we're starting to understand the target and particularly in the knee, that may be the big change that's happening and that we're recognizing the target that we're chasing may not be the target we chase.

Yeah, absolutely. And Mark, if I could come to you next, you know, we all know you've as Prof has alluded to a huge amount of literature in this area with some excellent RCTs you've done, many of which are published in the journal. Can you give us [00:05:00] your take on a similar question to profit on the current evidence in robotic knee replacement, how that's mirrored your own experience as well?

So Andrew, I think, look, I think the evidence base, is, is growing. Although, you know, if it'd be really critical, robotic systems have been around in orthopedics for the last 30 years and it's been a pretty slow burn. It's it's actually said that there are more review articles at written about robotic systems than actual articles on original research which is, which is a bit damning.

Yeah, a lot of the original studies were done of course, on the original, fully autonomous systems. Sure relatively high complication rates particularly with with the soft tissues. And I think that experience more recently with these semi-active systems, which really focused on in this paper are probably much more relevant to modern practice.

And the, the data that we've got on unicompartmental knees, I think is much more mature. The system MAKO system was developed and became available for clinical use in [00:06:00] 2008 but it wasn't really, until it was acquired by Stryker that it became widely available for use with total knee replacement.

So that total knee replacement data is definitely not nearly as mature. So I think it is developing. It's really been very interesting to see the results of all the different studies that have been coming through. I think the only thing I would say about the evidence that from the literature and as we look at it, as we always have to be mindful with these new technologies about the risk of bias, we're all invested in this as surgeons, as the nurses and physios.

The porters, the managers, industry, but especially the patients are all, we all want this to work and that might affect the way that we report how our knee feels or how we interpret the results of the studies and the data that we are presented with. So I think we always need to be careful. That we look and we search out for [00:07:00] bias in case we feel that there are perhaps undeserved benefits from this technology, but very exciting to see the data that's coming through now.

Yeah. I think that's a really interesting point. Like you say about bias and even. I like you say from the patient point of view in that certainly if you say to a patient you're going to get a robotic knee, there's probably already a bias there that they're going to get something a bit shiny or in a bit better.

And that's a really interesting point. Isn't it? About how they interpret it afterwards? Yeah. And I think it's, it's, it's interesting because that bias, I think has evolved too. So we've we embarked on our first unicompartmental knee study probably back in 2010 was our first recruited at patient. And actually there was quite a lot of suspicion around at that time.

And you fast forward five years, eh, or, or plus two, there were more recent studies and patients are definitely would look at that technology as providing them a benefit. So patients are certainly not in equipoise, but we need to be really careful, even in a [00:08:00] randomized control trial type studies and that we try and maintain that blinding where possible and really look for, for, for protecting at the studies from bias.

Absolutely. And Mark, before I move, I move on to the paper. It just, just for those who are not familiar with it, you mentioned the type of system. So the MAKO is a in contrast to some other systems is a semi-active systems that allows a surgeon sort of interact with the robot during the bone preparation, alignment, and rebalancing is that.

Yeah. So it's under, it's under surgeon control throughout the whole process. And what the robot does is, is, is create this haptic, this boundary based on your on, the preoperative plan beyond which it's impossible for you to cut. It provides a physical resistance and also gives you some audible, clues as well

when you're at the edge of the boundary. And will eventually shut down if you, if you push it beyond those boundaries. So very much under surgeon control, whereas these fully autonomous systems urgent. And you can take your hands [00:09:00] off. And the the robot would do all of the milling at for you and much more difficult, therefore to control that from a safety perspective.

Yeah, that's really interesting. So Nick, if I, if I come to you, can you sort of give us a brief introducing introduction Nick to the paper itself and maybe some background on the current systematic reviews that are out there and what made you look at. Yes, try and defend the review after Mark's comments, that it was more reviews than pay of as this is obviously the best review, of course.

And that's the mark for others. So this review really came on the back of a paper that we opened in 2019 ish when I was just learning robotic surgery, I haven't I've haven't done any on a patient just yet, I've done a few on cadavers, but I'm still just looking at the evidence rather than doing the surgery myself and, and we amalgamated all different types of robots.

Semi-automated and, and pulled results together. But prof mentioned that at that time, [00:10:00] when we did that review is that there's obviously different robots out there. Then obviously the more reading that we've done, we then looked at the recent systematic reviews that are out there and everybody's done the same

sort of thing. Put put them all together, see what comes out the other end. So we first did this for total knees, looking at the they robotic arm assisted i.e. the MAKO image based system compared with gold standard being manual. And then we looked at total hip that was published last year

in the journal, then we moved on to looking at the uni compartmental knee, and that was the only paper, that actually I didn't actually show functional difference. So that was the aim really just to, just to have a, a like-for-like comparison rather than a homogeneous robotic. Yeah, no, that makes sense.

And would you, in terms of the quality of the literature, when you look at sort of total knee uni, knee and hip is, is w which one would you say has the best quality of evidence currently? Mark Blyth has obviously published his randomized trial for them. And that's a high quality study, but there's lots, as Prof said at the very beginning,

[00:11:00] there's lots of coming through now, the Risser study with just total knees and Risser two total hips. So there's a lot more data coming out now. And Prof's own studies are I'm sure will be coming out shortly. Yeah, absolutely. So Nick, if we just move on to the methods of the paper, I thought we talked about this a very briefly, I don't want to take us away from time for discussion, but can you just maybe give a quick overview of how the paper was set up the review was set up and what studies you included.

There wasn't enough randomized controlled trials, sadly. Cause it was only Mark's, four times. So, so we included level two and level three studies only robotic arm assisted and any study that was reporting learning curve. Implant alignment, functional outcome complications, and cost.

Cause we thought those were kind of the main five outcomes. And they were collected by two fellows that were with us at the time we did all the work Junren Zhang, who's the main author and Nathan [00:12:00] Ng. So, but thank them for all their work. Absolutely. And just before we move on to the results, Nick, in terms of the outcome measures, the problems particular that, that varied in what was reported and it wasn't reported or was it, is that correct?

Well, so it wasn't reported in them all those that obviously followed PROMs up report a different PROM. So like some had the Oxford score some had the Knee Society score and some had the WOMAC in all the different time points sometimes there weren't, there was different variations on that score report. So it was very difficult

to, to, to put them together. So we used to report on the WOMAC and the Knee Society score, one option would have been to convert all the scores to zero to a hundred and then just put them all together. But we felt that was a bit, maybe it's not perfect. We report on them in two outcomes, being the WOMAC and the Knee Society score.

Brilliant. Okay. So if we, if we move on to the results, Nick, the 179 articles were identified from the initial screening, 14 were [00:13:00] included eventually. So of the 14, the one reported on learning curve, five on implant positioning, six on outcomes, five on complications, six on survivorship and three on cost and of the type of studies four, were

were from the same randomized trial, which, which you've mentioned. Two were Markov decision cost economic analyses, and three were retrospective and five being retrospective. So in terms of the main findings, could you sort of summarize what you found with regards learning curve and an implant alignment? I think the learning curve is fascinating because it's the same for the total knee as well.

I can't remember for the hip off the top of my head, certainly, but the learning curve for position is zero, which I find this fascinating. Right. So even if you're starting to learn, like if I don't do unicompartmental knees, but if I did, I have a learning curve of 25 50, probably because I wouldn't be optimal where with this your learning curve is zero for implant positioning.

Obviously I'll take a bit longer to do anyway. And e v en for operating room staff confidence. I was around with six cases. And I'm almost [00:14:00] sure now that you can apply a MAKO to total knees and unicompartmental, those six cases are probably going to be spread between the two. I would've thought so all of a sudden, after a week, you probably going to be flying.

It's really interesting. The start thing, like you say, when you compare it to the manual uni as well, it's, it's quite remarkable in it. And in terms of sort of, what about the functional outcomes report too, we've, we've mentioned them, but what did they show and what about complications? There's no difference in functional outcomes.

And Mark showed no difference in his I mean, and, but, but, but, but it did show a ceiling effect. I can't remember the exact number now you'll have to keep me right. I'm sure you presented the Oxford score, a score or something like, I think was like 30% had a, had a maximum score. No, no, that's right.

The score within the MCID of the maximum score. That that's what it was. So a third of patients had that. So whether we're using the wrong scoring system and we just don't have something sensitive enough yet. I know some people have used the Forgotten Joint Score. Less of a ceiling effect, but anyway, there was no difference in functional outcome.

[00:15:00] But the main difference we, you mentioned that the complications there, but within the complications, it was reoperations. And there's no difference between deep infection, superficial infection and reoperation. But when you put them all together, the overall rate of complications being all those things together were higher. Within that group

we also looked also those one registry study. Cause obviously you can't include registry data in a systematic review cause it ain't on PubMed, you know? But there was an Australia registry study that showed it was about, it was 0.7 versus 3% revision rate at around about two years between robotics and manual unis.

So that's certainly given the hint of that improved accuracy. Might not be shown in, in, in, in, in functional outcomes, but it might be shown in improves survivorship. Yeah, absolutely. Absolutely. So Nick, I think sort of drawing that all together, what do you feel are the take home message of the, of the study, but I suppose caveating it with [00:16:00] any, any limitations you felt that were of there, of the data?

I think it was as Mark says is the bias of bias or anything else than a randomized trial. Trial isn't it. When, when a patient knows that they're already going to get a MAKO. It would be nice to do that study I suppose, if I don't know whether Mark's got that data to ask what knee you thought you had and whether their outcomes are better, depending on what they thought they might've had.

I don't know whether that would be quite a nice anyway. But I think that's certainly a big thing. And, and, and, and I think the risk comes from the total knee side of things. That'd be brilliant to let us know whether. Hopefully, it's all blinded to see whether there is a real difference with robotic surgery.

But when you speak to more surgeons, because as I said before, I don't do this, but, but when surgeons see these patients, post-operatively, they all say it's fantastic outcomes. Good. But again, whether that's bias of their own.

Yeah, absolutely. And it would be fair to say Nick as well. A lot of these studies in the grand scheme of things are still relatively short-term followup [00:17:00] compared to what we need, or is that, would that be harsh. Mark's the longest that, was five years. I don't know what, you're coming up to ten now Mark probably are you? I am.

So five is probably sort of five would probably come the 10 year follow-up but it would be nice to have a larger cohort, I suppose, with maybe different PROMs. Yeah, absolutely. So Prof if I could come back to you, what's your take on this? How do you feel this adds to the current data and what is your sort of, what's your experience with drinking that data about the learning curve in the context of your own experiences with robotic surgery?

Thanks. I think the learning curve is the most interesting thing here in some respects, because we've looked at it within our own unit for a total hip for uni and for total knee. And you hit the target from case one, but the learning is all about the team. As Nick said, it's all about time. It's, it's, it's really, and it's different for it.

It'll be [00:18:00] different for different people because if you've navigated, then your learning curve is going to be a lot shorter in terms of time and putting pins in and so on. So I think that's really going to be interesting because we will also get better at teaching this. So I think the learning curve doesn't seem to be harming patients and that's really reassuring and that's certainly been, our experience has been easy to educate people in it.

Trainees come in, they love it. They learn a lot from it and that's something I think the literature hasn't yet possibly articulated is what a phenomenal teaching tool it is, to be able to plan something in three dimensions and then execute it and think about it. So that's sort of cognitive work that takes place is a critical part of training in knee surgery, which, which is quite important.

Yeah, that's really interesting Prof, the idea that actually that it can actually, it maybe improve your understanding of what, of what you're trying to do. So that's a really, that's really interesting. And we shouldn't get away from the fact that she would probably still don't know exactly what the target is.

We still don't know what the proper balance of the knee is. There's a [00:19:00] long journey that we're going on here. But my suspicion is that where this paper sits is it gives us enough of a signal that this is a good journey to be on and that we need to look beyond, you know, it's fascinating for me. Randomized studies are great and we must do them and we are doing them and Mark's done some great ones, but the problem is Mark's a superb surgeon.

So actually Mark's manual unis, probably are way better than everybody else's. And hence for him to detect a signal may be quite hard, whereas, you know, take it out into the masses, bring something that reduces the error, you know, get Clement, doing it for God's sake. And then you, you may see a fairly dramatic signal.

And, and that'd be on the other hand, you may not because there's a bit of cognitive work that goes some decision-making to do. So there really is an unknown here as to how this is going to translate, but this is a stepping stone, looking at the data from enthusiasts, from big centers that are doing it.

And then over the next five to 10 years, we're going to see longer term data, [00:20:00] but we're also going to see more pragmatic data. You know, the registries really start to pick this up. Yeah, no, I think that's interesting is actually when, when it spreads to the masses, like you're saying you get the data from everybody, then that's really the test

isn't it, Mark. If I, if I come to you, then, you know, what your thoughts as well into, you know, similar question to that of Prof's, but also, you know, it was interesting reading about, you know, the, you know, the big reduction, if you're allowing alignment outliers with, with the robotic surgery and the complications, is that your experience with it as well

and do you think that's where the benefit of this lies in the longer term? Yeah, I think that's absolutely right. I mean, that's the one thing we've got a real clear signal out across all of the studies that have been done that it's much more accurate at delivering that preoperative surgical plan. And the pre-op surgical plan is probably best worked out for unis because really what are we trying to do?

We're trying to do. So to reconstitute that constitutional alignment by retensioning the ligament on the, on the, on the involved side of the, of the joint, and maybe there's a little bit of work to do on what the coronal, eh, [00:21:00] alignment is of the tibial component, but everything else is really a resurfacing procedure.

I think that the pre-op surgical plan as Fares has suggested is really not that well worked out for the total knee replacements. And so we've got this reproducibility and precision that we can perhaps be used to then define what that target is, whether it is mechanical axis, restricted, or fully kinematic, or this functional alignment that's being done now, which is a kind of hybrid between kinematic and mechanical access alignment.

And so I think so I think that that's, that's certainly the way ahead, but that's going to be required a lot of a lot of data and a lot of experience that over time. But I think the other thing we can do is to use that those that, that safety that we w we're demonstrating with the lower complication rate, and perhaps it really will be of advantage to people who are lower, not low volume surgeons, because we're hopefully sort of gradually eliminating those perhaps people in early years of their career to give them a [00:22:00] little bit more confidence to do that, to do the procedure with with a little bit more safety.

Yeah, no, absolutely. Mark. I think that's, that's, that's a nice summary of, of, of those potential benefits and how we can move forward. I thought maybe I'd finish off asking a question to all of you, which is the same and sort of maybe being devil's advocate a bit, but given the equivalent findings, you know, with regards to PROMs and reintervention infection will robotic uni knee replacement ever be cost-effective in the NHS setting.

Maybe I'll start with, with yourself. That's a very good question. Cost-effective depends on what you mean by cost-effective. So, so cost-effective from a cost utility point of view is difficult after working with Mark currently on, on his data using the EQ-5D and the recent paper that he published.

There was no difference in the EQ-5D. So by definition from a cost utility point of view, it's difficult, but there is that decreased revision risk. [00:23:00] Potentially in the longer term and the decreased complication risk return to theater in Mark's paper, there was, I can't remember off the top of my head.

It's gone. I think there was one arthroscopy. There was two and one revision in the manual group. So all of that mounts up. So cause those patients will have had scans, CTs, whatever else, been to the clinic a few times, seeing the GP went the there. So all that kind of cost mounts up. But whether that balances against the cost of a MAKO

procedure to prevent that morbidity. And within the NHS, I don't know, but I think, I think certainly if it was a member of my family, I'd want a knee replacement that was probably done as accurately as possible. So we probably probably go onto Mark or Prof to do them manually. But certainly, in a normal say, I just wonder whether, whether accuracy has to be our golden end or our gold standard.

So whether it is a cost effective thing or whether it's just best care. Yeah. That's interesting. And what do you think Prof, do [00:24:00] you think that's that's, that's the take it all. I think, you know, the studies are being designed, including our studies, including racer to re-look at that in terms of the uni intuitively.

It is so the right thing to do to plan that, that way in balance it to the soft tissues. And if you look at the failure mode of unis, you know, we we've touched the surface with this. We've had, you know, Mark's too good, a surgeon, quite frankly, to have done that study. So the problem is the failure mode tend to be things like overcorrection, malposition, et cetera.

So I think when we get into the longer term, when we get into wider application, I would imagine it will be the enhanced technology of some form will be the way to do unis because the, the, the error it's too easy to get it wrong. Otherwise that would, would be my view, but you're right, getting that data.

That's going to be a lot of work. And that's why these, you know, sometimes we need more research. No, absolutely Prof. No, that's an interesting take. Isn't it like when you say, like you said before, if we do spread it, you spread it to the masses, [00:25:00] everybody, and actually the, the money you will probably save potentially on revisions and other problems.

It could all very easily add up in favor couldn't it. Mark. If I come to you last, what do you think you've done a lot of the research in this area. Do you think there will be a day when it, it will just be standard of care on the NHS? Yeah, I think that's, it's inevitable. And Fares has made the point.

I've heard him say this many times, you know, that robots, once they're in an industry, they're there once they're there, they're there. They never go. And I think that it look, the cost will come down. I mean, it's the same with all new tech. They're expensive. And I think that the costs will come down over time.

And particularly if you've got wider applications of the technology, then you can use it in more situations. So the cost per case, and will, will come down and maybe we're asking too much of it to, to think that we're going to see the clinical improvements just in the milling or bone cutting. And with the marginal gains that you'll get with other new technologies.

So about, you know, [00:26:00] ligament, tension sensors about augmented reality, AI and all those sorts of things, which incrementally will improve the the, the outcomes. Then we'll be able to, to, to justify those additional costs. One of the things that if you look at it, though, you can kind of turn costs in its head a little bit by saying, well, look you get about six or seven

QALYs for a unicompartmental knee replacement, the additional cost for us in our study on the basis that we were going to be doing a hundred cases a year, which is pretty low volume for, for the use of a robot, what was about a thousand pounds per case. And you kind of do the math in your head. That's about 130 pounds per QALY.

So given that, you know, NICE will fund, whatever it is, 30,000 pounds per QALY, it's pretty, it's pretty small change. Isn't it? So I think all we need to do is to demonstrate some some, some benefit and we, we get obsessed with it being a, you know, a minimally important clinical difference in a PROM that we need to demonstrate, but we [00:27:00] need to just demonstrate a societal benefit.

And I think those costs will actually a) go down and then they will become insignificant in the grand scheme of things. I that's a very true Mark. And I think that's a really nice final note actually, in terms of, like you say, I think it's, the costs will come down eventually, like with all, with all technology, as you say, and that it will almost, as you say, probably become a fait accompli.

That that's what we need to do. Well guys, I'm afraid that's all we've got time for today, but thank you so much for taking time to join us. All of you and congratulations on a great study and all the work that you've all done in this area. It was great to have you all with us, and I really enjoyed that. And to our listeners, we do hope you've enjoyed joining us, and we encourage you to share your thoughts and comments through social media and like feel free to tweet or post about anything we've discussed today.

And thanks again for joining us. Take care of everyone.