Reliability and validity of the Wrightington classification of elbow fracture-dislocation

Listen to Mr Andrew Duckworth interview Professor Adam Watts, from the Wrightington Upper Limb Unit, about his paper entitled 'Reliability and validity of the Wrightington classification of elbow fracture-dislocation'.

Click here to read the article

Transcript

Copy Transcript Copied

Reliability and validity of the Wrightington classification of elbow fracture-dislocation 

[00:00:00] Welcome everyone to our BJJ podcast for the month of August, I'm Andrew Duckworth and a warm welcome from your team here at The Bone & Joint Journal. 

As always, we'd like to thank our readers and listeners for the comments and support we continue to receive as well as to our many authors and guest interviewers who have taken part so far. 

Over this year, we continue to build on the range of topics we have covered through our series so far with a continuing aim to improve the accessibility and visibility of the studies we publish. Both you as our readers and listeners as well as for our many authors. 

For this month study as you know, over the next 20 minutes or so we'll cover a range of aspects for the chosen paper, emphasizing the important points of how the study is being put together, as well as the key findings from the paper and how these potentially fit into each of our day-to-day clinical practices.

We also hope to give you a behind the scenes insight into how the authors have developed the study and give them an opportunity to put forward the key findings of the work. 

So today I have the pleasure of being joined by my editorial board colleague here at the journal, Professor Adam Watts, from the Wrightington Upper Limb Unit to discuss their study, entitled Reliability and validity of the Wrightington classification of elbow fracture-dislocation, [00:01:00] which has been published in the August edition of the BJJ.

Welcome Adam and a big thank you for taking the time to join us today. 

Thanks very much, Andy and many thanks to The BJJ for this opportunity. 

Absolute pleasure. It's great to have you here, Adam. So if we go straight onto the paper, obviously the aim of the study was to assess the reliability, the reproducibility, and the validity  of the classification system that you've developed  at Wrightington. So if you can give us a brief introduction to the paper and some background to these injuries and how they've traditionally been classified. 

Thanks Andy. So first of all, if I can acknowledge my co-authors, Jag Singh, Mike Elvey, and particularly Zaid Hamoodi, who put a huge amount of work into bringing this study together. 

When we treat elbow fracture-dislocations it can get quite confusing with lots of different types of injury, terrible triads, Monteggia, Monteggia variants, and post-remedial fracture-dislocations and each of these have their own [00:02:00] classification system.

And understandably for the occasional elbow surgeon this can lead to a lot of confusion. This study was designed to assess a new classification tool that was designed to identify patterns of elbow fracture-dislocation, which is based on the important stabilizing structures of the elbow. 

All too often, we assess the radiographs of the elbow fracture-dislocations, our eye is drawn to the radial head fracture, the most obvious injury that we see, or the injury to the electron or ulnar shaft. And we tend to ignore the most important element, which is the coronoid process, often hiding in the shadows behind the radial heads on the lateral projection, but actually the coronoid is the key to understanding these injuries.

Historically, we've used the Regan and Morrey Classification which assesses the coronoid injury based on the loss of height on the [00:03:00] lateral projection. Problem is that it doesn't discriminate between those injuries where the coronoid fracture leads to instability, and it doesn't differentiate between medial and lateral coronary injuries.

O'Driscoll introduced a really importantclassification, which was a good advance in describing the coronoid as a three dimensional structure of fans shaped structure with different elements of this bony structure that can be injured. It has many strengths, but people do struggle to remember the components of the system. And that can lead to challenges with reliability. And it also doesn't consider what's happened to the radial head or to the important stabilizing ligaments. And that's why we felt there was room for a more universal classification systems guide management. 

That's a great overview, Adam. I think like you say, I think the [00:04:00] key I found with the system and after reading the paper and as you know I've talked about it before, is that idea of the coronoid being the key, you know, that that's what we need to build everything around. And also with that appreciation of the soft tissue component, which is often with all these injuries, what we're really chasing is stability to the elbow. Isn't it? So that we can initiate early range of motion and prevent stiffness in the elbow, which is what we're always really fighting. Would you agree with that?

I think that's absolutely right. I think that you should, view the bony injury as the tell.  It indicates the likely soft tissue injury. So if you can understand the patterns of injury you can infer from that the likely soft tissue injury that also needs to be addressed. Cause we can't restore stability just by addressing the bony elements, we also have to restore integrity of the soft tissues. 

Absolutely. So if we move on to the classification system and it's often difficult on a podcast, but can you describe how that was sort of developed and what it sort of entails and how it potentially can guide treatment as well?

Yeah. So, it is hard [00:05:00] to describe without illustrations and so I'd encourage people to read the paper where the illustrations will make a lot of this clearer, but essentially the Wrightington classification is based on a three-column concept of elbow stability. And in brief, this considers the stability to come from the anteromedial facet of the coronoid, which is the medial column. The anterolateral facet of the coronoid, which is the middle column. And then the radial head, which is the lateral column. The fulcrum for varus and valgus sits between the anteromedial and the anterolateral facets of the coronoid process. 

As we've already talked about Andy, the stability also comes from the ligaments and the muscles, the neuromuscular elements around the elbow. And these are particularly important on the lateral side of the elbow, around the radial head, where we've got relatively little bony congruity and therefore the soft tissue elements become much more important compared to the medial side [00:06:00] where the distal humerus has covered through an arc of almost 180 degrees and so the osseous elements on the medial side are much more important. 

So if the elbow collapses into varus, around that fulcrum between the anteromedial and anterolateral facet, the medial column is what becomes broken. So that's the anteromedial facet. The middle and lateral column remain intact because the humorous moves away from those structures. But the lateral ligament gets pulled off as a sleeve from the humerus. 

So this is a type A, and anteromedial from Wrightington classification and anteromedial facet fracture, or what's also known as a post-remedial fracture-dislocation, O'dricoll described as a post-remedial rotatory instability.

So you can see how the nomenclature becomes quite confusing for people, but essentially this is an injury where you got an isolated coronoid fracture. It will often look like a Reagan and Morrey Type [00:07:00] I injury. So what people often refer to as a tip of coronoid fracture, but actually it's a really important injury because instability will often follow from this injury pattern because of the associated soft tissue injury. The posterior band of the medial ligament may be torn too. And the treatment for these injuries is to fix the important lateral soft tissue stabilizers and to restore the medial column in whatever way. And there, you know, there are various ways to address that, which is probably beyond the scope of this discussion.

Sure.

 If you apply, so we're moving on to the type B pattern. So if you apply a direct axial force, you'll break both the anteroremedial and the anterolateral facet of the coronoid. So that's a bi-facet or basal fracture, which is a type B in the Wrightington classification. The radial head may be fractured or it may escape.

So in [00:08:00] an apex anterior Monteggia, or Monteggia-like lesion, the radial head moves anteriorly away from the humerus and will remain intact. And you can just have the bi-facet coronoid element. And in this situation, because both facets of the coronoid are injured, particularly the involvement of the anteromedial facet, it's really important to fix the coronoid and all of our treatment should be aimed at restoring the integrity of the coronoid.

The radial head needs to be addressed as well, either with, if it is fractured either with plating or withheadless screws or with replacement and any lateral ligament injury needs to be addressed if it is torn. But in some of the B-plus type of injuries, your ligaments will be intact because the whole forearm has actually moved closer to the humerus. And so the working length for the ligaments has actually shortened if you'd like, so those ligaments [00:09:00] can be intact. 

The posterior elements of the ligament complex is maybe torn, but the anterior bands will be intact. And so if you restore the bony anatomy, then actually stability can be restored in those situations.

Moving on to the type C. So this is where the elbow collapses into valgus it's often a valgus external rotation injury, the anteromedial facet or medial column escapes injury in this situation. And you get an injury to the anterolateral facet and radial head. So this is a combined anterolateral facet and regular head fracture Wrightington C, which is also what we would consider to be a terrible triad injury. 

The lateral ligament we can refer will be torn in these injuries and will need to be addressed. And so this injury is treated by the store and the radial head lateral ligament, but the coronoid because of its adjustment  anterolateral facet of [00:10:00] middle column injury it doesn't need to be addressed, as long as we've restored the lateral column and that'ssometimes difficult for people to understand. And I know that a lot of people do try and put sutures around these anterolateral facet fractures, but as long as you assure that it doesn't extend to involve the anteromedial facet, and as long as you've done a good job in fixing lateral ligament complex and addressing the radial head, then actually that anterolateral facet fracture coronoid doesn't need to be individually fixed. 

And the common muted radial head with an intact coronoid also fits into this type C because the treatment is the same. So it's a matter of addressing the radial head and addressing the lateral complex. 

And then the last group of injuries are those which are distal to the coronoid process. So that's a diaphyseal or distal injury type [00:11:00] D in the Wrightington classification where the contribution of the coronoid to stability is essentially not affected, but the ulnar is fractured and the radial head will be either fractured or dislocated. And so the ulnar needs to be addressed to bring the radius back into alignment. And then the radial lateral column needs to be addressed and the lateral ligament may need to be fixed in those two again to restore stability. 

So end up with a classification system with type A, which is an anteromedial facet fracture, type B a bi-facet fracture with or without a radial head fracture. Type C common use of radial head or combined radial head, and then anterolateral facet fracture and type D, which is a diaphyseal fracture, which is distal to the coronoid  with or without a radial head fracture.

 So I hope that I havent confused things too much for people with that. But as I say, I think the illustrations in [00:12:00] the manuscript probably make things a lot clearer, but that's an overview of what we've described.

No, I think that's brilliant, Adam. And I think like you say it's difficult without the diagram, which is obviously figure two in the paper when it comes out. But I think that was a really great description. As you know, I'm also a big believer in what you described in particular, those, the terrible triad, the type C.

And I think certainly there has been a vogue of moving away if they're a true, terrible triad, and it's just the anterolateral facet of just fixing or replacing the radial head and repairing the lateral ligament.  And they're  probably one of the most common types  as we'll see coming on, but that's a really good overview. And I think the great, great thing about it as well is it's not just a classification system that describes the injury pattern. It also helps guide treatment, which I think is the limitation of a lot of our classifications that are out there, not just in the elbow, but throughout orthopedics. 

So if we then move on to the study, it was obviously a retrospective review of your prospective trauma database you have there for your elbows. All patients were over 16 years of age and they had a fracture-dislocation of the elbow between 2010 to 2018. So just for the listeners could you give a brief overview of the database that was used at the sort of inclusion [00:13:00] and exclusion criteria to be included in this study?

Yeah. So, we included all elbow fracture-dislocations that were treated surgically in the daily space. So there are injuries where the patients may not have proceeded to surgical interventions for various reasons. So these are only the surgically managed patients over the age of 16 years. 

For inclusion in this study, the patients had to have a complete set of preoperative radiographs and CT scans, from which we could construct 3D reconstructions, and a full operative record detailing the pattern anatomy of the injury for reasons that we'll come on to talk about probably later on.

One of the issues that we had in the study, cause we had a total of 60 fracture-dislocations identified within the database. But many of the patients were referred from external centres. And with the PACS system, when CT scans are [00:14:00] transferred they're not always transferred in a format that enables 3D reconstruction. And so they had to be excluded from the study.

Yeah. And how many was that Adam? It was a relatively small number, wasn't it I believe, in the end? 

That were excluded?

From the total number, I think it was, but there was no particular difference between the characteristics from those included and excluded, I believe. Is that right?

There were no differences in the characteristics of the two. So we looked at age, gender. We looked at fracture patterns between those that were included and excluded, and we found no statistical differences between the two. We included 49 patients out of 60, within the database. 

Yeah.

Yeah, it was just 11 were excluded. 

Excellent, excellent. So if we move on to the observers, obviously there were seven observers in the study that looked at the various imaging modalities. So why were they chosen and what sort of information were they provided with? 

So Andy we wanted to ensure [00:15:00] that the classification could be used by all those who treat elbow fracture-dislocations, not just those with a specialist interest in the elbow and therefore we approached trainees,  we approached fellows in upper limb surgery, and we approached surgeons from district general hospitals with the general practice and specialist trauma surgeons, so that we tried to get a breadth of experience across the surgical field. The type of people who may be addressing these types of injuries. 

We gave  the surgeons a description of the classification and an illustrative guide and they were provided with anonymized CDs, containing images of the patients for the elbows that were included in the study. And they were then asked to classify the injury  and initially from the x-rays alone. And then we asked them to classify based on the 2D CT scan and finally with [00:16:00] the 3D reconstructions. And they were asked to do this at two separate time intervals at least a month apart. 

Okay. Right. That's a really good description, obviously the various imaging modalities use. 

And what was your sort of gold standard for confirming the diagnosis in those 48 patients included? So, this is where the operative records came into, into important. So the operative record of the pathoanatomy was taken as the gold standard. So that was record of the injury to the radial head, the coronoid process, but also the important soft tissue injury as well, and the description of that from the operation note. 

Yeah. So it's a very robust sort of gold standard to make sure that the classification was right. And before we move on to the results, can you just give a brief overview of the analysis that you performed and remind me more probably than our listeners, the difference between the reliability, reproducibility and validity when assessing these things, as well as the Landis and Koch criteria, which you used in the study.

Yeah. So, so as I've said, we had seven surgeons who assess the imaging on [00:17:00] two separate occasions and at least a month apart. And we assess the  inter-observer reliability. That's how consistent an assessment tool is at measuring at what it is meant to measure. 

Hmm. 

And this was measured using the multi-observer Fleiss Kappa statistic.

The intra-observer reproducibility, which is a measure of whether you'll get the same answer if you repeat the assessment, was assessed using the Cohen Kappa statistic, and then the validity of the classification, which is essentially the accuracy of  that measurement, whether the measurement tool is measuring what it's supposed to measure, which in this instance is the agreement between the classification of the octave findings was assessed using percentage of that agreement, between the x-rays and the CT scans and the operative record.

Hmm.

In order to maintain consistency in the [00:18:00] nomenclature when describing the relative strength and agreement between their Capus statistics at Landis and Koch in 1977, fairly arbitrarily, but in a way that has been widely adopted, assigned, parameters to the K, the Kappa value. So if the Kappa is less than zero it's poor, 0.2 there's slight agreement, 0.21 to 0.4 there's fair agreement, 0.41 to 0.6, moderate, 0.61 to 0.8 substantial, and 0.81 to one almost perfect agreement. And that really has been widely adopted amongst the orthopedic community. 

Perfect. Great. And so if we move on to the results, 48 patients had a mean age of 49 years, 300 type A of the Wrightington classification injury, 11 are type B, 16 are type B plus, 16 are type C, which is, like you say, like the terrible triad and then type two type D pluses, but none had a [00:19:00] type D injury. And there were no significant differences in the demographics of fracture classes between those patients including those excluded that we've already said. 

Just before we move on to the key findings Adam no type D injuries was there a reason for that in particular that you can think of, was it just one of those things or...?

So we did have type D injuries in the series, but unfortunately the cases were from external tertiary referrals essentially. And so we weren't able to do 3D reconstructions of those injuries. 

Yeah.

 It's just a less common injury, hence the low numbers. I think it's unfortunate to admit that it is a weakness of  the study that we weren't able to include those, but unfortunately that was the material that we had to work with.

Oh, certainly, like you said, they're relatively rare injuries, aren't they? If you could just move on to those key findings in relation to the performance of the classification,  what do you feel the key findings of the study in relation to particularly the inter- and intraobserver reliability.

Well, the main observation was that the  Wrightington classification is reliable, it's reproducible, [00:20:00] and it's valid across the seven observers. The spread of observations was very narrow, and therefore, given the numbers of observers we had, we weren't able to determine whether there was an effect of surgeon grade on accuracy or of surgeon specialty on accuracy.

But what we found was actually that it was pretty consistent across the observers. And everyone was able to achieve a very, a pretty consistent outcome using this classification, which was reassuring. 

Absolutely. No, and like you say, it means it's more generalizable for a variety of surgeons of various experience with these types of injuries. 

And how did the various imaging modalities compare in terms of, you know, you had x-rays, the 2D CTs and the 3D CT reconstructions.

 Well, the main observation was that the, the use of CT improve the accuracy of the classification. We only had moderate reliability based on plain radiographs, but when we include CT scans [00:21:00] in the assessment of the injury, then the inter-observer reliability increased to two substantial.

Yeah. So really a strong argument for the use of CT scans. The intra-observer reproducibility was substantial based on plain radiographs and CT scans. Inclusion of a 3D reconstruction improve that even further to almost perfect, with a significant difference based on the assessment  on the x-rays alone.

So it looks like the use of CT scans is valuable in assessing these injuries certainly for this classification. 

Absolutely. Sure. Absolutely. 

And then lastly, just the validity of the classification, which was confirmed with over 70% agreement, with the surgical findings based on plain x-rays. And again, that increased further when we based our assessment on the CT scan, so over [00:22:00] 85% agreement.

 I say we, I should emphasize that I was not involved in, as the developer of the classification, I was not involved in assessing any of these radiographs within this study. This was all done by independent people. 

And just finally, before we move on to the implications of the study. What did you find regards the sensitivity testing that you used to examine the ability of the observers to identify each sort of subtype of the classification?

Yeah. So again our observations here were slightly limited with the low numbers for some types of injury. And I think it would be an interesting thing to look at in a larger cohort of patients, but we found that the observers could quite accurately identify type C injury. So that's a terrible triad injury, or a comminuted radial head fracture, even from plain radiographs, we had an 88% agreement. Use of CT with 3D [00:23:00] reconstruction, improve their accuracy to 100%. Yeah. So again, really strengthening the case for the use of CT.

On the other hand, there was only 69% agreement between the operative findings and plain x-rays for a B-plus type injury. But this rose to 94% when a CT scan was used. So, I think the important take home message from that is that the assessment of coronoid fracture configuration is difficult for most observers from plain x-rays.

The CT scan can substantially improve our assessment for that, and for that reason, I would strongly encourage the, the routine use of CT for these injuries. 

Absolutely. Sure. And just before we move on, I'm just looking at table six there, I know there's only three type A injuries, but even with the 2D and 3D CT, it's still nowhere near comparative is it in terms of, you know, these are difficult, these PMRI or whatever type it is you want to [00:24:00] describe them as, they're rare but difficult to identify, but very important we do so it's a difficult problem that almost isn't it?

It is. I think that that it's about education, it's about recognition. So very often these type A injuries, injuries to the anteromedial facet of the coronoid process will be dismissed as simple dislocations of the elbow with a flake avulsion of the coronoid process.

And we really need to get the message out there very clearly that this is not a simple dislocation. This is an elbow fracture-dislocation with an important injury to an important bony stabilizer of the elbow, the anteromedial facet. And once you start to recognize that and change the thought process,  then it'll prompt you to get the CT scan and to better understand what's going on.

So if you see a, what may be considered to be a tip of coronoid fracture in isolation, [00:25:00] then this is a type A anteromedial facet fracture until proven otherwise. And definitely I would get a CT scan or an MRI scan in that situation. 

I know  I couldn't agree with that more. And I have to say, you know, from discussion with you, I think it's something like you say, I think it’s often dismissed as an elbow dislocation with a simple tip fracture and carry on, but actually we need to look into those more and more detailed and not to be missed because they are relatively rare as well.

So if we move on to the key findings, what do you feel that they are Adam in terms of positioning the literature and it obviously considering any potential limitations, but what it really adds. 

Well, I think the Wrightington classification gives us a valid, reliable, and reproducible classification system that encompasses all elbow fracture- dislocations that we're faced with.

And importantly, for a classification, this also gives us a guide as to how to manage the injuries. So we have algorithms assigned to each category that can give people a route map [00:26:00] to managing these things. And the routine use of CT scan can clearly strengthen the accuracy of our assessment of these injuries, and will hopefully prevent missteps in their management and the later dislocations that can present after management of these injuries, that can be difficult to salvage.

Yeah, no, absolutely.  What's your interpretation of how it compares and fits in with the other historical classification systems that we discussed at the beginning? 

Well, this is the first system that is a universal classification system. It's the first one that really includes all of the injury patterns of fracture dislocation around the elbow. And it helps us, hopefully to understand or to organize our thoughts and to guide our treatment. But I suppose most importantly, it's the first classification that's actually been validated. 

Yeah, no, I totally agree. And in terms of, you know, moving forward,  [00:27:00] what would you feel the next steps are? I mean, obviously this is the major, first paper that's come out related to the classification system, you've talked about it a lot before, but what would you feel the next steps are for it? 

Well, I think that, you know, we've looked within our own practice at how this can be applied to the management of these injuries. And we've shown that in, in our hands, following these algorithms, which are not my algorithms, these are algorithms taken from the literature from managing these entry injury patterns. And so it's based on many expert surgeon's experience with these injuries, but by following this classification, by following these algorithms, we've been able to demonstrate reliable and consistent outcomes for these injuries.

But we don't know if that's going to be reproducible and generalizable, and that needs to be demonstrated for others. That is, we can either establish a [00:28:00] multicentre study to assess that, but certainly other centres need to explore whether these pathways work in their hands. 

No. I totally agree, Adam. Absolutely.

Well I think that's actually all we have time for today, Adam, but thank you so much for taking the time to join us. I really enjoyed that and congratulations on a really excellent study, which, that is without doubt and evolve, addition to the literature, and certainly has a big influence on my practice. But thank you so much for joining us. 

Well, thank you very much, Andy. And I'd also just like to take the opportunity to thank all of the surgeons and trainees who contributed to completing this study. Thanks very much, indeed. 

Thank you, Adam. And to our listeners, we do hope you've enjoyed joining us and we encourage you to share your thoughts and comments through Twitter, Facebook, and like feel free to post a tweet about anything we've discussed here today. And thanks again for joining us.