Listen to Andrew Duckworth, Prof Whitehouse, Prof Haddad, Alex Liddle & Nick Clement discuss the paper 'Patellar resurfacing during primary total knee replacement is associated with a lower risk of revision surgery' published in the May 2021 issue of The Bone & Joint Journal.
Click here to read the paper
[00:00:00] Welcome everyone to one of our BJJ Podcasts 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 thank you to our many authors and colleagues who've taken part so far.
We hope that you're continuing to enjoy our podcast and all of the knowledge translation work we've delivered so far this year. And before we move on to the papers today, we also just wanted to highlight our new, special edition podcasts that are with our Specialty Editors here at the journal. These started in March with Sam Oussedik, our Specialty Editor for Knee and continue in April with Mr. Dan Perry, our specialty editor for children's orthopedics. We do hope these are providing some insight into the fantastic work our specialists do here at the BJJ. And just for everyone listening, we wanted to make sure you're aware that the highlight papers we discussed on those special edition podcasts will have temporary free access following their release.
So for this study today, we're going to take a slightly different format and approach, predominantly given the area in question. It's a really hot topic at the moment in the knee world, controversial, excitement maybe. Antibodies are often [00:01:00] raised amongst our new colleagues in particular when it's discussed. And I've had a few good luck emails before this podcast by taking on this topic itself.
So we'll have a brief overview of the chosen paper from the senior author who'll be joining us and then we're going to have a civil, but debated discussion, maybe on the topic from a series of special guests. We've got a really excellent line up to join us today.
So, firstly, I have the pleasure of being joined by the senior author for the selected paper, Professor Mike Whitehouse, from Bristol, to discuss his study entitled Patellar resurfacing during primary total knee replacement is associated with a lower risk of revision surgery and announced this using the National Joint Registry for England Wales, Northern Ireland and the Isle of Man, which has been published in the May edition of the BJJ.
Welcome back, Mike, and thanks for joining us today.
Thanks very much, Andrew .
And joining Mike and I are some special guests. We have the pleasure of being joined by one of my editorial board colleagues here at the journal and the specialty editor for BJO, Mr. Alex Little from Imperial. Alex. Great to have you with us today.
Thanks for having me on.
I also have the great pleasure today to be joined by Mr. Nick Clement, [00:02:00] who is a specialty editor for the knee at the BJR, or one of my great colleagues here at Edinburgh. Thank you for joining us, Nick. It's great to have you with us.
Thanks Andrew. A pleasure to be here today and thanks for the invite.
And lastly, but not least we are delighted to welcome back our editor-in-chief at the BJJ, Professor Fares Haddad. Great to have you back.
Thanks. And thank you for doing this.
So Mike, if I can kick off with you. Your paper opens up by saying the debate remains as to whether the patella should be resurfaced during total knee replacement. And you aim to try and add further data to this through a retrospective observational study, using NJR data that compared all cause revision rates following primary TKRs, according to whether the patella was resurfaced or not. And for the non resurfaced TKRs you estimate what the revision rate would have been if the patella had been resurfaced. And then finally looked at the risk of rerevision, following secondary patella resurfacing.
So Mike often we spent 20 or 30 minutes talking about this, but if you can give us a relatively brief overview of the paper, highlighting really the key findings and the take home messages that we can then maybe go [00:03:00] on to discuss in more detail.
Yeah, no problem Andrew. So as you say, this is an NJR based study. So the big power of this was the numbers. So we were able to include just over 842,000 total knees . In terms of what proportion had patella resurfacing, 36% had a patella resurfacing performed at the primary operation. And very much the kind of main study design was as you described.
So really when you are kind of considering that range of different outcomes, you need to bring to bed a couple of different approaches in order to analyze this so the kind of primary survivorship analysis use kaplan-Meier estimates. And we then built a flexible parametric survival model that could account for the differences between the groups and the reason for using that rather than something like a Cox regression people may be more familiar with is it allows us to model for predictors whose effect on baseline hazards will vary over time. So it's really got that kind of dynamic element that lets you fully capture that.
So the prediction model, we're obviously able to bring in some [00:04:00] predictive variables into that. So we incorporated age, sex, ASA grade, the year of primary and the TKR fixation and constraints as well. And basically that modeling worked by we took a random 90% sample of the total knees and built the model based on that. And then were able to validate it against the remaining 10% to prove that it is really predicting the outcome that we see. And we did that on the knees that had had patella resurfacing. So that 36% of the population.
So for the risk of the secondary patella resurfacing, we have to use slightly different approach. So there we did use the, probably more familiar, Cox regression model, and we had to sensor within that for competing risks. So if you'd already been revised for another reason, for example, then that changes your risk of secondary patella resurfacing, and obviously mortality.
So really drill down into this issue of whether this is a constraint problem or whether this is a variation between brands and trochlea geometry. We've [00:05:00] ended brand level comparisons for brands where we had more than a thousand implantations of both the CR and PS bearings.
So, what we found was the risk of all cause revision was significantly higher if the patella was not resurfaced at primary. And to put that in context the differences were 3.54% risk of revision in the non patella resurface group versus 3% at 10 years.
The prediction model that we developed performed well and was validated on the 10% test set. The excess revisions in the non-resurface group equated to one excess revision per 189 primaries at 10 years. So within the cohort of this size, that would represent approximately 2,841 additional revisions that could potentially have been avoided if those knees had received patella resurfacing.
And interestingly within the knees that didn't have a patella resurfacing at the primary, 19.5% of them, so just over 2,000 of them, their first recorded revision was the [00:06:00] secondary patella resurfacing.
When we looked at the risk of secondary patella resurfacing by constraint, we found it was 17% higher in PS variance overall compared to cruciate retaining. And when we accounted for those competing risks the final gray model showed a similar result with a 16% higher rate. So that's when we broke it down to really look at the brand level analysis. And as I said, we only included ones with over a thousand implantations of each variant. And that meant we could consider 390,000 total knees and 1,424 of those underwent secondary patella resurfacing.
There was a significant interaction between constraints and brand. So for example, with the PSC Sigma, which was by far the most common within that group, the risk of secondary patella resurfacing was significantly higher for the PS variance. The hazard ratio was 2.28. But that wasn't consistent across all the brands. And particularly in the Vanguard, it was actually lower with a hazard ratio of 0.37, and for the Genesis two Genesis two, next gen, [00:07:00] Scorpio, Scorpio NRG and the triathlon, there was no difference in the risk of secondary patella resurfacing, according to constraint.
However, when a secondary patella resurfacing was performed, that knee did not behave like a primary procedure anymore. So the five-year revision rate was much higher for a primary TKR than for a primary TKR where the patella had been resurfaced. And the five-year revision rates we saw were 8.8% in the knee that had undergone the secondary patella resurfacing versus 1.9% for a total knee where the patella was resurfaced at the primary.
That's great, Mike really great overview. The papers fantastic. And it's really a lot of data in there. That's a really good summary. Before we move on to maybe how that fits with the registry data. What do you feel are the key limitations of the data as a whole? And you discussed them nicely in your paper, but if you could just summarize them.
Well as ever, this is observational data. So there is selection. There's that X factor that happens at the moment of selection for a surgeon as to whether they [00:08:00] resurface or not. So we can't determine causality directly from the state that even with the kind of methods we've used here. With all observational data, there's a risk of under-reporting. So there may be under-reporting of revisions. However, we're fairly confident that wouldn't actually affect our results because we don't think that would differ according to whether you had the patella resurfaced or not at the primary procedure.
And our analysis of secondary patella resurfacing relied on only a patella component being added. So if, for example, the knee had been in for a few years and the surgeon took the opportunity to exchange the tibial poly while they were there, that would appear as a more complete revision. And therefore was excluded from that secondary patella resurfacing.
We were limited by brand level comparison. And as I say, I think, you know, we all have a perception that you get a patella friendly trochlea design, patella unfriendly designs. However, we haven't got any form of description of that geometry here, and we're just relying on the brand level data for that. We don't have any data on [00:09:00] the condition of the native patella. So obviously it's not a free choice in all cases. And some it's just not safe to resurface, which then, you know, you'd assume those don't undergo a secondary patella resurfacing, but I think we've all got an experience of seeing a patella's left because it was considered unsafe. And then actually go on to have a secondary resurfacing later.
And obviously, you know, much as there's sort of some complex stats and big numbers in here we don't have data on non-revision procedures. We haven't included PROMS in this, and this is not a formal health economic analysis.
Yeah, no, I think that's very fair Mike. It puts that sort of findings into context. And before we move on to one of our guests, how does this data really fit with the other registry data that is out there?
So there is different practice across the world. So our rates of patella resurfacing are much lower than they are in the US currently and similar for Australia and possibly slightly higher than they are in some of the Scandinavian countries.
Now, this is the largest registry study on this topic today and it's bigger than all the other [00:10:00] studies combined. Smaller studies from the New Zealand, Sweden and Norway didn't show a difference in revision rates, according to patella resurfacing or not. But the New Zealand study did report better PROMS at six months and five years in the patella resurface group.
There's been a fairly recent Australian study that there was a big one as well incorporated just over 136,000 TKRs. And they did find a higher revision rate in TKRs, but they did a slightly different analysis and they looked at whether surgeons performed no patella resurfacing or infrequent patella resurfacing and found that patients treated by that group were at higher risk.
Yeah, no, that's great. And it's nice to put it into context of those other registries. So that's sort of nicely moving on. I'll come back to you Mike.
But Prof, if I could come to you next. So first of all, I suppose, how do you interpret the findings of this study?
I mean I really liked this study because I think it's the best we've got to interpret the data we've got available here. And I also think it's a great example of how the registry community and, you know, [00:11:00] thanks to the leadership, Professor Whitehouse and others around him have really moved on in looking at that data, in analyzing it, recognizing its limitations and in being, you know, incredibly clever about how they're using it to inform us where we need to be informed. You know, the RCTs in this area are tiny relative to this. You know, we can take some learning from them. But in reality, I think this is an incredibly valuable study that's going to help us to guide our thinking moving forward. As Professor Whitehouse said, there are some things in terms of the selection of the patients in terms of what we really don't know about whether there was a patelloplasty or not, you know, something we've published on in the journal before. You know, how is the patella treated if it's not resurfacing. And we also don't know if it was resurfaced, how well it was resurfaced. We also don't know the alignment of these patients. We also don't know really what kinematics or function they end up with.
So there are a whole load of unknowns, but as an overall message to our community, that there is a higher reoperation stroke revision rate. [00:12:00] That's really quite interesting and it opens up a whole load of other questions around that. I mean, for me, there's an interesting factor that if you do not have your patella resurfaced and you present with pain, that could be related to the patellofemoral joint or it could not be, but in reality, it puts you automatically under a microscope.
We've talked before and comparing unis and totals, how there's a kind of bias. If you see a uni in pain people will tend to revise it. Less so with a total. Having an unresearched patella to a surgeon who's looking to do something is a red flag to a bull. And it also encourages that surgeon to start looking at that two degrees of varus, that little malrotation of the femoral component and interpreting that in a slightly different light. So I think to go back to your question, I really liked this paper. I liked the analysis and I liked the trend and the direction, which our registry colleagues are going in, looking at their data, linking it and analyzing it so cleverly.
No definitely Prof. As Mike says in his study and as you just said, you know, [00:13:00] the RCTs are relatively small with short-term followup, you know. I think you said in your paper Mike, most are under 150 patients and the follow-ups under three years. So, I mean, in terms of what this adds in terms of number of patients and length to follow up, it's very unique.
Well, how'd you interpret the implant sort of findings Prof? What, what do you take home from them?
No, I think it's difficult. Implants change over time. You know, Professor Whitehouse has described the most commonly used implant in this study is an implant that's being phased out quite frankly, for a new generation, from the same company, or at least they're trying to do that. So I think that, you know, that's something we should look at very cautiously because not everyone does a PS the same, not everyone does a CR the same. The selection criteria may be different. You know, we talked about selective resurfaces well we've got selective PS users, selective CR users, and we've got those who do PS or CR for everything.
So I think I struggle slightly. I think it's great to have that data in there, but the numbers get smaller. And what [00:14:00] we can make from that data gets less. So although you start with 800,000, by the time you're looking at a thousand cases, with so much variability, I'm not sure how valuable that is. And the whole concept of, you know, everyone tells us their patellas, they're neutral, clear shape is patella friendly. I'm not sure that we really know or can interpret that from this.
Sure. And in terms of, you sort of led into it then. In terms of, you know, the newer designs, you know, we've got the robotic , you know, knee replacements now. What do you feel... where are the current trends going? You know, maybe not just as well, we've obviously seen the trends in the UK with Mike's paper, but maybe in like the US or other places. What's your feel for that?
That's interesting. So if you look at where knee surgery is going, and we could talk about a number of facets, I think cementless knee arthroplasty is on the rise, certainly in North America, but also elsewhere. It's attractive. We know the results of cemented knees, although excellent, aren't as good in the younger population. And we've got a big young, heavy population needing knee arthroplasty. So [00:15:00] once you're moving away from cement you're less likely I suspect to resurface the patella because the one bit of cementless implants that really worry surgeons is a cementless patella and it's slightly in congress to do a cementless knee replacement and then cement a patella button on.
Likewise we've published a lot of robotic assisted data in the journal over the last few years. And that's a really exciting technology to optimize alignment, to really get us thinking about knee arthroplasty. But all these enhanced technology tools help us with the tibia femoral joint, with the femur and the tibia, but they don't really address the patella. They may improve kinematics. That may make it easier not to resurface the patella, but in principle, I think they'll drive people to do less patella resurfacing because the device does not cut the patella for you. And in reality, you know, there's about 10 different ways you can get it wrong. That's the thing about, you know, one of the things I take away from Professor Whitehouses study is that the thing we need [00:16:00] to do is educate people about how to resurface the patella.
Well, training needs to focus on, you know, looking at patellofemoral offset, measuring the thickness, having devices that will help you cut the patella appropriately, not create asymmetry and decide what you're going to do with it. So I think there's lots here that we can learn in terms of the knock on effects of such a study.
Absolutely. No, that's great. So Alex maybe if I come to you next .Maybe just to again ask you the same question as Prof, what's your sort of main sort of interpretation of the paper? And maybe focusing on some of the arguments about which we've touched on already is how we define revision and how it's interpreted. What is your take home from that?
Well, I'll echo Prof in saying that it was a really high quality study. And as you'd expect from this group the analysis was very comprehensive and and really took into account sort of minimized a lot of the difficulties you see with registry type data . As is often the case with this kind of [00:17:00] data set the the initial crude comparison of revision rate, which was a 3.5% in the nonresurface group and 3% in the resurface group was almost exactly the same as the more complex adjusted analysis 10 years which was 3.55 compared to an expected rate of 3.02 in that group.
The thing that struck me most about this was how rare revision is as an outcome overall anyway . And the numbers that we're talking about here really reflect how individual surgeons can convince themselves that what they're doing is the right thing. So if you're looking at one excess revision per 189 cases by the 10 yea r timepoint you know, someone with an average revision, sorry, an average primary knee practice might have enough cases to fill the fingers of one hand by 10 years. And that type of thing can certainly be overlooked in your individual practice. And so this is why we have to be mindful of the [00:18:00] literature out there. We need to look at these population level studies to guide our practice, rather than just doing what we think works in our hands.
As well as all of the new data that this brings and the data that we have already there are some unknowns still I think . we know now that the revision rate is higher in non resurfaced knees .We saw that in the knee arthoplasty trial as well, which was a large pragmatic, randomized controlled trial which showed something similar. We're still not clear about the PROMS. And again, this reflects the paucity of instruments that we have when we're looking at what is overall a very successful procedure and a very skewed outcome.
We don't know about revision threshold and clearly there's a lower threshold for resurfacing the patella cause it's a straightforward procedure. On the other hand there are obviously some people out there with painful patellas who don't go on to to have secondary resurfacing. And whether they would have been helped is something we don't know.
We don't have very good quality data on the [00:19:00] success rates of secondary surfacing. I think the best we have is a meta-analysis in 2016, which said that the patient satisfaction rate was 64% after secondary patella resurfacing, which for primary procedure is pretty poor. But actually if you've got a procedure that six out of 10 people get a benefit from then it suggests actually that this is a real effect rather than just people doing what they can.
And the other thing we don't really know is, is what the downside of primary resurfacing is. So we've got a good number needed to treat, but we don't really have a number needed to harm. So we don't know how many intraoperative patella fractures there are. We don't know how many complications related to the patella resurfacing as was touched on before . from the technical point of view, how many recent resurfaced patellas have been overstaffed or have malpractice as a result of the resurfacing that would have done fine otherwise. And so there's still a bit of uncertainty about this. But overall, this adds to a pretty strong body of evidence that resurfacing is the right thing to do. And there really [00:20:00] isn't a body of evidence that says that explicitly that resurfacing is the wrong thing to do. I don't know of any studies and the others might correct me on this, that there's a real downside. And the parallel that was brought with unicompartmental knee replacement. There's clearly an upside to unicompartmental knee replacement that in some way outweighs the disadvantage of that higher rate of revision and that lower threshold for revision. And I just don't think there is an upside in not resurfacing the patella again, on a population basis.
That's interesting. And if we move, maybe in relation to that, so how does does the findings of this study really fit with NICE guidance and other guidance at the moment, you know, that's been quite a hot topic recently. How do you think this adds to that?
So it really supports where we are. I think NICE, with these controversial subjects , they do tend to to shy away from being overly prescriptive, but the NICE guidance has really evolved [00:21:00] towards supporting patella resurfacing. And I think the current guidance is that we should offer patella resurfacing to patients. We should offer the option of patella resurfacing. And I suspect that's as strong as it's going to get. And what this paper does is that it really supports that at that point of view and supports the body of literature that led to that recommendation in the first place.
Would you agree with that Prof as well? Do you think that's the, the way we're going?
I hope that's the way we're going. I think the issue we face specifically in a UK setting is a number of commissioning groups who've taken the cost at time of surgery is the issue and tried to bar patella resurfacing, and make it an exceptional thing to do. And this is strong evidence to refute that.
And we know that over the lifetime for prosthesis, the prosthesis costs are absolutely negligible in terms of the overall health economics.
So Nick, if I could maybe come to you following on from Alex . well, I'll ask you the same question that I asked Prof . What your [00:22:00] final thoughts on the study in terms of the interpretation and in the context of the current literature?
Well, I commend the paper to begin with. I think it certainly shines some light on the dark art of patella resurfacing and the current literature would suggest that there's no real difference in clinical outcomes between resurfacing and non-resurfacing. So it really comes down to revision rates and cost economics side of things. And I think this paper sums it up quite nicely. Goes on to show that obviously there's around a half a percent difference in revision rate. If you don't resurface versus resurfacing and there may well be a cost benefit in that which was shown in the discussion, but there's also other parts to the paper that were highlighted with the different types of implant that was associated with a different revision rate according to whether you resurfaced or didn't resurface the patella. And I just wonder whether that could be associated with [00:23:00] the complexity of the knee with a posterior stabilized knee being used maybe more in the valgus part of arthritis, as opposed to a CS knee being used in a standard varus knee, which might have patellofemoral problems.
There's also some evidence more recently of position of the femoral component with more flection and decrease in the anterior coronal offset decreases the pressure in the patellofemoral joint and therefore decreases the anterior knee pain. So even position of the component more recently has been discussed as opposed to patella resurface in itself. But nonetheless, I think it's a good bit just to give us some idea, o f once you adjust for confounding variables, because we know from the registry that patella resurfacing is associated with a lower risk of revision, but there's also a whole lot of other things that influenced that. I think this paper goes somewhere trying to adjust for all those confounding factors.
That's great, Nick. And I think like the [00:24:00] others have said, it sort of does add a lot to literature in this area. If I could just maybe just slightly different track though, but in terms of patellar resurfacing itself, what are the potential complications associated with that though?
That is a great question. And the authors kind of highlight that in the discussion. The only complications they address is revision of the component they're off, but they're suggesting if there's been any washout or extensive mechanism problem or fracture that hasn't been picked up. And potentially one of the risks is the extensor mechanism. If you resurface the patellar to the patellar tendon, quads tendon or specifically to patellar fracture and certainly looking at the literature to round up to 1% rate of patellar fracture and about half a percent for patellar tendon and quadriceps tendon. So there's certainly a real risk to the extensive mechanism, which has disastrous consequences for the [00:25:00] patient, of course, which may not have been picked up and that wouldn't negate any revision benefit, which was observed in this study, about half a percent. If potentially there is half a percent to 1% of the patients that's going to have extensive mechanism problems. However those papers are quite old and techniques have evolved since then. And there's different patella r fixations and techniques now so that may not be the same, but that should be recognized. And that may not be recognized in this paper.
No, that's very interesting.
Yeah. So Mike that actually probably leads me back to you quite nicely. Obviously just quoting one of the sentences in your discussion, you say that if all just over 530,000 TKRs without a patellar resurfacing from the NJR were actually resurfaced, it is possible that savings of over 55 million pounds could be made over 10 years. I mean, you've obviously caveated, you know, it's not a true economic analysis, but how would you explain that data to our [00:26:00] listeners?
So the estimate there comes from the numbers, as you've said, in terms of the volume of procedures that didn't have patellar resurfacing and the findings of the CAT trial. So in CAT, they found that over 10 years, the cost saving was 104 pounds per patient who had a patellar resurfacing compared to not. So once you've accounted for those implant costs, because of the change in the profile of what happens in low revision rate, you actually accumulate a cost saving for having made that investment upfront.
So at the moment, you know, it's an estimate just based on extrapolating that number out to numbers that we've got in the registry, but it comes, because we do a lot of knee replacements, it comes to a very big number.
Yeah, absolutely. I think sort of, if you're looking at maybe the number needed to treat from your data, you know, how many people are you going to have to resurface to avoid one revision?
So that we haven't got within there. I mean, as I say, the kind of number of excess revisions by not doing it as the way we chose to kind of tackle it as a [00:27:00] calculated number. And as Alex has pointed out, you know, for the kind of volumes of procedures that we see for surgeons, recording procedures in the NJR, that means that's an awful long time between accumulating a revision hence, you know, it doesn't clock up on the radar.
And you know, when we talk about the kind of the decision to resurface, I think there's absolutely no doubt that, you know, we all have in mind the patellar thickness and the risk to the extensor mechanism because failed extensor mechanism and the knee replacement is not a nice place to be. And is a very difficult problem to sort out.
So I suppose, sort of moving forward then, Mike. I mean, this is obviously very robust data. What else do we need then? You know, you've mentioned a lot of things that, you know, you couldn't look at such as things like PROMS or maybe a more robust health economic analysis. What's the next step forward do you think?
So I think it's really around this question of always versus selective and that's what we saw highlighted in the NICE guidance although they made that [00:28:00] recommendation default position should be operate. They said that they couldn't find data to answer that question of what's better always doing it or selectively doing it. And certainly when we've done kind of survey work around this , the most common practice for surgeons who completed our survey was to selectively resurface and the majority of those that selectively resurfaced said they did so in less than 50% of their primary total knees and the next most common practice was always resurfacing. So at the moment, despite the data we've got available, we still have that variation in practice. So I think really the next question for us to answer is that one that NICE specified as to what's better always reserve single doing so selectively. And, you know, I think the the difficult part of that is picking apart that selection process. So, you know, is the hard and fast things within there or is it just that kind of decision making by the surgeon at the time on the basis of what they see with the patellar thickness to safety. [00:29:00] And as a Prof has pointed out, what else do surgeons do to the patellar at the time that may actually influence outcomes?
That's very interesting. And do you think, I suppose it's a difficult question to answer just to finish up, but do you think if we had maybe a robust trial that showed that, do you think that would convince people, or do you think there will always be people who say I'm going to do it selectively or I'm not going to do it at all. I know it's a difficult question
I'm not aware of a single RCT that's converted the entire world to the conclusions about RCT, but, you know, as ever we're looking to evolve and drive practice, and if we kind of, you know, influenced practice, then that then filters down to trainees as they're coming through and you see those kind of shifts over time. So I don't think even a definitive RCT would cause an immediate revolution, but I think it would drive practice in the right direction. The question being, of course, what is the right answer?
Yes, absolutely. Absolutely. Well, Mike, I think that's a pretty good spot to finish on and I'm afraid [00:30:00] that's probably all we have time for as well. So thank you so much to all of you for joining us. That was a really interesting discussion and congratulate Mike on an excellent study that is without doubt an invaluable addition to the literature in this area I'm sure will lead to much much debate and discussion going forward.
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 a like. Feel free to post or tweet about anything we've discussed here today. And thanks again for joining us. Stay safe, everyone.