Professor Edward Davis from the Royal Orthopaedic Hospital in the United Kingdom reviews published literature comparing computer-assisted THA versus conventional THA.
thank you very much for the invitation to speak. My name is Ed Davis. I'm speaking to you from Birmingham. I work here at the Royal Orthopedic Hospital In the next 15 minutes. I'm just going to try and share my thoughts on what I think the data shows at the moment around the use of computer assistance and hip replacement. I do have some personal disclosures here. Some of the work that I'm going to talk about was funded by smith and nephew So there's absolutely no doubt hip replacement is a phenomenal operation and phenomenally successful as referred to back in 2007 in this Lancet article where it was coined the operation of the century. That's one of the reasons why I think a lot of people say, what are you trying to improve when it is such a successful operation. But if we look back at this paper in 2007, even then it said that computers were probably going to be part of the future and would come into widespread use to try and improve the operation still And yet I suppose here we are in 2021 and still not widely up taken, but we are seeing big moves to the increasing use of computer navigation and robotics. Going to talk a little bit now about some of the evidence we already have around hip replacement and navigation. And this is a systematic literature review on the use of computer navigation. 746 studies identified ending up with 49 comparative studies in the final review and for those two key main factors from hip navigation, which is a component position in leg length discrepancy and that analysis you won't be surprised to hear that a large proportion of these studies were retrospective reviews with a few randomized control studies and some observational studies as well. So what are the main key factors? The evidence is really very strong and has been for a long time that if you use computer navigation and you know the target you want to hit as far as your A. C tablets concerned, you will significantly reduce the deviation from your planned as a tabular anti version and inclination angles. And of course, probably more importantly you'll reduce those outliers and I hope maybe we're going to discuss later around the target, which is still debatable of where we want to put our asset tabular component. But I think if we define a position we know and the literature would support that navigation can very accurately take you there. What about leg length and offset? Well, leg length discrepancies, The literature would certainly support that leg length discrepancies were smaller if you use navigation and the literature would support that you get reduced leg length discrepancies. So less of those patients that may be dissatisfied because they come to you saying their legs feel a different length. But also of course those patients who are very unhappy and unhappy enough to take legal action and reducing the number of patients that may sue you because their leg length discrepancy is wrong. I think the literature is quite strong as you can see now one of the biggest reasons that surgeons say that they don't want to navigate hip replacements or take on this technology is because of the increased surgical time. So what does the literature say about this? Well, the vast majority of studies found that actually there was no significant difference in timing whether use navigation or you don't. Some studies showed that it took slightly longer and the two studies showed that it was shorter from me as a long term navigation user. I certainly support the literature that seems to suggest that it makes no difference. And why is that? So I think when we look at all your cases overall there is no significant difference. However, those really straightforward, easy hip replacements do take a few minutes longer with navigation because you've got to do the registration and you've got to put some pins in. But actually the time saving comes in those cases that are slightly more challenging. Well you might not be quite sure whether you've got it perfect whether you may feel that you're not sure on the leg length and therefore you've got a retrial and then retrial again or change your a c tablet component. You don't need to do that when you use navigation because you just look at the screen and it gives you the numbers and you know that you're right and you can move on. So I think overall when you look at the total number of cases then there is no significant difference in timing. So why aren't we all using navigation then? Well, a lot of surgeons go back to the point. Well that's fine. So you can hit a target. Great. But what about the long term outcome for patients? What about proms? What about survivorship and hip replacement when you use navigation? And a lot of papers that you read go back and they quote this article back in 2016 in core Which said that at 10 years following hip replacement, comparing computer assistance and conventional instrumentation, there was no difference in problems and no difference in survivorship. However, when you look at this paper there was only 30 patients in each group and I think it was probably underpowered. But also these were clearly very good surgeons. So if you look at the Kaplan meier plot here, which is revision for mechanical reasons and dislocation. They had no revisions in either of those two groups. The only one revision they did have was prepared for aesthetic fracture in the control group. So clearly very good surgeons and we know that navigation comes into its own in preventing those outliers. So it may be in this study particularly we had very good surgeons but also it was underpowered. So when we have issues like this we can look to real world data. And we've seen a big interest in this more recently. It really started in my mind from this great paper that was published in the J. B. J. S. American back in 2019. It was using the Medicare dataset and teasing out of that. Those patients who have benefited from computer navigation and those that hadn't. And they showed very nicely that there was a reduction in dislocation rate and a reduction in tabular component revision. When computer assistance was used prior to Covid, I was really lucky and managed to travel down to the Australian meeting back in 2019 in Canberra in Australia and Bill Water stood up at the podium and presented a beautiful paper looking at the Australian joint registry and looking at how computer navigation in hip replacement affected survivorship. And he showed at that meeting a very very significant improvement in survivorship. When you went to computer navigation in your hips versus conventional instrumentation. It really made me sit back and think, well we've got the paper the Medicare data set from the U. S. And then Bill Water talking about the data from the Australian joint registry showing such significant improvements with navigation and hips that really when you compare that to what we've been trying to show in Nis. I'll be honest over the years we've struggled using this real world data to show a very big difference when you use navigation in knees and actually the difference in hips is vastly better. And that got me thinking and came back from Australia Australian meeting and thought, well why don't we look at our national joint registry here in England Wales northern Ireland and they are a man and tease out of that to see if it fits with that Australian and that Medicare data sets. So I'd just like to talk you through that study that we had published earlier this year. And if you hit that QR code there, it'll take you to the full paper. So the our aims of this study were really to look at the effect of computer guidance on the survivorship but also patient satisfaction after hip replacement using the N. J. R. Data set. So we're able to get data going back to April 2003 with the last entry in February 2020. Now we know when we're considering data like this, particularly registry data that there's a huge amount of contracting fact contracting, sorry confounding factors that start to creep in. So initially we started off by saying, well what's one of the issues with hip replacement? It's the components. So we wanted to limit our analysis down to single suppliers un cemented as a tabular component. Now I know that reduces the numbers hugely but I think it's really important to take out that confounding factor of the implant itself. We also isolated it to osteoarthritis the same reason. And we also exclude metal metal bearings for the same reason. We were able to access some pre and six month postoperative proms for a subset of these cases. And importantly the statistical analysis was all performed independently by the N. J. R. Standard survival analysis was done using capitalism our plots. But importantly again, when you're using this data, you have to be very careful in the cox proportional hazard models are really good because we can adjust for the significant confounding factors particularly and in this model we adjusted for gender A. S. A group approach, head size, yearly cohort effect, age group B. M. I. And bearing. So we ended up with a total number of cases of 42.5000 in this analysis. Unfortunately in the UK only about 2% of cases benefited from computer guidance. The mean implantation time was good at five years but our maximum implantation time was around 16 years. We have 36 surgeons implanting in this cohort And actually this was one of my biggest concerns. When we started this analysis that we may only have one or two surgeons using computer guidance but actually we ended up with 36. So here's our headline Kaplan meier plots. You can see here in red. We've got the computer guidance survivorship and in blue we've got conventional instrumentation and the hazy areas around the confidence intervals which are clearly very narrow in the standard conventional instrumentation but slightly wider in the computer guided due to the numbers at 10 years, we saw a statistical difference between the survivorship with computer guidance With revision rates of only 1.1%. Yet in the conventional implementation we were seeing revision rates of about 3.9. But again we need to be careful and we need to adjust that for the factors that we know can bias these types of results. So in in red here, you see the fully adjusted Cox model with a hazard ratio of .45 that did come to significance. So that's a 55 Reduction in revision rate if you use computer guidance over conventional instrumentation. So I'll say that again, it's huge. A 55% reduction in revision rate in this group of patients when computer guidance was used. What about patient reported outcome measures? We've got success and satisfaction plotted here. On the left is the computer navigation. In the right is standard instrumentation. If you concentrate on the right hand plot, which is satisfaction, there was a significant improvement in satisfaction rates at six months. You can see the upward trend there. If you look at the blue box which is the patients reporting excellent satisfaction rates, you can see it shifted up. But I think probably more importantly, you can see that the red box has disappeared in the computer guidance plot and that's those patients reporting poor satisfaction. So in this analysis we saw a significant reduction in the revision rate of of patients who benefited from computer navigation over conventional instrumentation for those patients at six months that we were able to look at. They reported improved satisfaction rate when they had computer navigation over conventional instrumentation. We have 36 surgeons in the navigated group. The statistical analysis was all performed independently by the N. J. R. When we compare these results to those in the Medicare analysis that I talked about earlier. I think we were seeing slightly improved survivorship rates because we were able to look at the total age group. So of course in the Medicare we're excluding those younger high demand patients. Whereas we were able to include those in our analysis which I think collect the slightly improve survivorship in the computer computer navigated group. One of the limitations. Of course this is a registry study with real world data and all the concerns that that may bring. We only have 2% of cases with computer guidance and yes the possibility for unaccounted, confound us. Despite all the effort we went to it selecting a single manufacturer of adjusting for all the parameters that we know cause problems and maybe we can talk a little bit about that in the discussion. But I suppose my main concern here is the type of surgeons that will be using computer navigation. Maybe the biggest confound er here and maybe they're good at surgery because they're perfectionists and they're striving for the best and maybe even without navigation they'd be better which is why we need these randomized control studies to back this type of data up satisfaction rates a single satisfaction measure and we know the concern about that was somebody rude to the patient. Could they not park their car? Is that why they were dissatisfied rather than the outcome of the surgery? So conclude our study showed a 55% reduction in revision rates with with the use of computer navigation. And we saw significantly improve patient satisfaction at six months. But of course it is registry data and therefore cause arctic can't be inferred. We really need those randomized controlled studies. I'm involved in two very large R. C. T. S. That are running with hips at the moment. One in robotics and image this navigation which we've been talking about today which I hope will start to feed through introducing that level one data on whether it makes a difference to survivorship and improves as we're starting to see from these type of register and real world datasets. Since we published our paper the Australian joint registry have also published the outcome of the work that was presented with Bill by Bill Waters at their Australian meeting and they showed very nicely in this paper that if you use computer navigation in their total cohort, they showed reduced rates of revisions for dislocation. But when they did there component combinations they also showed that there was a reduction in revision rate for all reasons. So to conclude, I think now we're seeing three papers from the U. S. From the U. K. And from Australia, all backing up the evidence now that actually if you use computer navigation, you can improve your outcomes for your patients and there may be a very significant improvement in revision rates with this type of technology. Thank you very much.