Episode Transcript
[00:00:02] Speaker A: Welcome to your Cases on Hold, the JVGS podcast, hosted by Andrew Schoenfeld and Aisha Adkeen.
[00:00:08] Speaker B: Here we discuss the best of what each issue of JBJS has to offer with the usual dose of entertainment and pop culture.
[00:00:16] Speaker A: Take us with you in the gym, on the commute, and, as ever, whenever your case is on hold.
[00:00:26] Speaker B: Hello and welcome to the next episode of our podcast, your Case is on hold, where we delve into the best research from the next issue of the Journal of Bone and Joint Surgery. This is episode number 108 and if you are listening on the day we drop, this is June 16th for the June 17th issue. My name is Aisha Abdeen. I'm an orthopedic surgeon and Chief of the Division of Hip and Knee Arthroplasty at Boston Medical Center. I'm an Associate professor of Orthopedic Surgery at Boston University. I have with me my co host.
[00:00:55] Speaker A: I'm Andrew Schoenfeld, professor of Orthopedic Surgery and Vice Chair of Education for the Department of Orthopedic Surgery, Harvard Medical School.
[00:01:04] Speaker B: Just as a reminder, the opinions expressed here are exclusively our own and do not represent those of jbjs, the editorial board, the Board of directors, nor the affiliate journals of jbjs. This episode of your Cases on Hold is brought to you by JBGs, CME and the Clinical Classroom. It's an efficient online orthopedic learning system that uses adaptive learning technology, observes how you learn and then individualize content with precise focus on knowledge, skill and competence deficits. Orthopedic specialists worldwide use JBGS Clinical Classroom and CME for clinical practice, lifelong learning and exam prep.
So get your learn on at the top of the pile. We'll start with our first article, which is an obituary. We pay tribute to The Legacy of Dr. Aaron Rosenberg, 1953-2026. Truly a giant in adult reconstructive surgery.
We have what's New in Spine Surgery by Chung, which is permanently free.
We are at an inflection point, but not the one we expected by Young Comparative efficacy of Surgical versus Non Surgical management of Acute Achilles Tendon Rupture in a novel mouse model by way and there's a commentary a blinded analysis of quality and fidelity in Orthopedic patient education materials simplified by ChatGPT and humans by Nasser. And this is 30 days free a Humanitarian Crisis Care of the Migrant Patient by Yang. And this is a highlight paper Advances in the Management of Sternoclavicular Joint Injuries by Nielsen.
We have non traumatic osteonecrosis of the Femoral Head An International Evidence Based Clinical Practice Guideline by Cheng. This is a highlight article as well.
We'll get into our headlines first. We have Rebuilding the Research Pipeline. We why Medical Students Struggle in Orthopedic Scholarship and How the System Can Improve by Wu.
Take it away, Andrew.
[00:03:02] Speaker A: Okay, so this is a Viewpoint article on something that I am intimately involved in and very passionate about.
I fear here that I don't I don't want to be called out for picking on anyone because I realized subsequently, not not until I kind of got into things that we a few months ago in the March episode covered a viewpoint from Dr. Wu and another colleague about the residency selection process tonight.
So, you know, I'm getting a little bit concerned that Dr. Wu and I can't be friends.
It's not personal, I assure you.
But I will pull out my best Doc Holliday from the movie Tombstone and say I'll be your huckleberry.
Because in you know what, I don't this is a viewpoint and of course everyone is titled entitled to their view. And obviously Dr. Wu is very interested in, in making sure that medical students have a quality research experience as part of their portfolio.
And who wouldn't want that? I mean that we all want those things.
But I also was a little bit concerned, similar to how I was in March and why I selected that, that some of the information that's imparted here can be both misleading and in certain contexts, probably not helpful to the end user, particularly if they are a medical student.
And so I wanted to use this time for those and I know a lot of people who listen are younger, some medical students, some in residency, some in early career that are our audience tends to be on the younger side and intentionally so.
So there's there's a lot of information here about residents and about medical students and about how research should be constructed.
And if I'm only providing a counterfactual, I wouldn't even say going back to the Tom Cruise movie A Minority Report, because I actually think that Dr. Wu's points I think I would characterize as more not necessarily aligned with the mainstream, certainly in terms of how research is done. I will say because of my interest and my passion for this, I'm afraid that I'm going to go long.
I want you to exercise the host's prerogative here and cut me off when maybe I've gone too far. I'm going to try and limit myself. But there's so many points here that I think are very important to call out and bring out.
And it's not just Dr. Wu who's come up with these ideas. This is misinformation and bad information that I think permeates our environment and our ecosystem. So I want to start with like the, the, the basic premise here that he starts off with in his first couple of paragraphs, which is that doing research is high stakes for medical students and that there's just a market inflation in terms of research expectations. He sets the premise right off the bat by, by quoting a study that also I think has a lot of problems. I'm very familiar with that work because it's done by one of my colleagues, but it's essentially a modeling study where they estimated that matched applicants by 2040 would average over 160 research items.
You know, that is just not, I can guarantee you, I would bet the house on that is not going to pan out in 2040. And I will bet it for you for a number of different reasons right off the bat. So I don't want medical students to have a heart attack reading this and being like, how am I going to come up with if it's 160 in 2040? I mean, granted that's still more than a decade away, but you know, what is it now? Is it 80? Is it? You get all of these questions all the time about oh, how much research should I do? And I think that we're already starting from the wrong place.
The idea of research is I think that it provides a toolkit and a skill set that you can acquire earlier than you can surgical skills.
And it conveys a lot of the things in an objective way that residency programs are looking for. The residency programs don't know if you're going to be able to suture skin to skin in two minutes or less. They don't know how good you're going to be or how quickly you're going to be able to put in a total knee replacement by yourself. They can't determine that from any medical student, but they can determine, are you a hard worker? Are you dedicated?
Are you task oriented? Are you somebody who can take something, an assignment or a series of integrated assignments and go from A to Z and get it done. And that's why there's this proxy metric for research performance that has, has some degree of value.
That value is lost if it just becomes, can you plug yourself into a research shop that just spits out a bunch of, you know, data mine, Pearl Diver, Trinet, X Nisquip, just sort of research Products where you as the medical student or the resident for that matter, are one of 15 or 20 other authors in the middle of a paper.
So why I say that this 160 average that we're starting from, the foundational conversation as the rationale for this work is problematic is one, if you think that, if you think about it, it takes the estimated and the estimated average for like a quality research project from start to finish is usually like three to six months.
So if we give it the benefit of the six months and you have 160 research projects, well, I guess you've been working for 80 years already before you even submitted your ERAs application.
Second, this is considering a very unique time and space.
If we blew it out to let's start in 1950 and look how many medical students have publications and then went all the way through to the present, we would have a different number altogether. This is a very truncated timeline, a very narrow timeline that's focused on a unique time point in our existence where medical students are able to, where everyone really is able to produce research at a pace that's much faster than it used to be. And again, that's relying primarily on generating products from pre existing data that's either acquired or purchased. And it tends to be these data mine type of studies which have an issue in and of themselves. One, many people are tired of seeing them. Two, I wrote an article about this about 10 years ago which was that they're kind of like a natural resource, like you will, they will get tapped out and it creates a moral hazard because you can only ask the Oracle of the Pearl Diver or the Trinet X or Epic Cosmos.
Your question one time next year there might be some more data, but you can't really. It's not sufficient enough to revisit that question entirely. And not only are you working on it, but there's 15, 20, 30, 50, 100 other researchers in other places who are also working on it.
There's this arms race to make sure that you get enough publications out of your investment in Trinet X or whatever it is. It just creates a terrible moral hazard. And it's a moral hazard for the medical students, it's a moral hazard for the residents, and it's a moral hazard for the literature because you get. I've been in, I've been at meetings, in sessions where medical students are presenting data from Pearl Diver on the same topic and are coming up with different answers. You had one that said smoking isn't a problem or vaping isn't a problem for Spine fusion. And the other one said using the same data, the same substrate, the same data set. And it's incredibly embarrassing indeed.
So when we start talking about medical students as the pipeline for research, I think that's a fundamentally bad place to start.
And some of what Dr. Wu is bringing to the forefront is, you know, he says, like we need to, he wants better training for the medical students up front, he wants orientation, he wants a lot of different process measures to make sure.
And these are laudable goals, but they're also somewhat misplaced because let's just change the framework here. If this was rebuilding the total joint arthroplasty pipeline, why medical students struggle in total joint replacement and how the system can improve, we would be like, that's patently ridiculous. You wouldn't, you know, how can we get it to the point where medical students can do the total joint arthroplasty themselves? How can we get it right? Or so you know, maybe that's too extreme for you all. Let's flip it to postdocs. Okay. Rebuilding the grant writing pipeline, why postdocs struggle to get R01 grants and how the system can improve also fundamentally misplayed. That's not the role of postdocs. Yes, the role of medical students is not to really create research.
It's not to really be the main driver of a research program, certainly, or a research portfolio. At times there may be a very experienced medical student who through their background, maybe they have a PhD previously, maybe they've been involved in research for a long time, maybe their parents are researchers and have, you know, through osmosis or just being interested and involved. They've had them in the lab from an early age. So there are certain one offs who could potentially do this. But your average medical student, the expectation should not be that they do the research themselves or that they're going to. The expectation should be that they're participating in some way that's meaningful that allows them to have a published product that they can bring to the residency interviews and present that either as an experience that's going to help inform what they might want to do or has informed how they think about being and attending, or just create a conversation piece and something that sets them apart from somebody else.
Once we start talking about residents and medical students as being the drivers of research, we've lost the plot that, that you wouldn't talk about them being the, the drivers of surgical productivity or grant writing productivity.
It's inappropriate to talk about them as the drivers of research productivity. They are there to learn. They are learners, they can't be producers. And so then the process measures that he's talking about, which are all really good practices. But I would say that what he's presenting as kind of a, a novel, a novel new paradigm in the way. This is what all the best research shops do. You set goals, you have tracking sheets, you have meetings at our research shop meets every two weeks, every Monday at 5pm everybody who's working on research in our spine division, we all get together and we all go over, you know, these are the action items from two weeks ago. This is where these projects are. These are who's on board. We have a spreadsheet.
We've been doing that since 20 years now. We've been doing that. This is what the programs.
This is just scalable. It's scalable in certain contexts. But you have to have that, you know, just like again, to go back to the analogy to do total joint replacements, well, you have to have someone who's experienced at doing total joint replacements and, and then you have to have the team that supports them to get good R01 level funding for your department. You have to have somebody who's experienced in writing R01s and is able to navigate the NIH system.
You can't crowdsource it or just grow it from. We'll just train the medical students to do this.
It's just so the take home points are, I think that the outlook for research, the forecasts are incorrect.
You won't get to a point where you have these many publications because the journals will just will not publish them.
You need a journal to actually accept your work at the end of the day.
And there's only 24 hours in a day. The final point that I will make is that there's only 24 hours in a day.
That doesn't change.
So if you're spending all your time doing research, you're not spending your time learning clinical stuff that's actually more valuable to getting you into a residency than I think the research is.
[00:17:47] Speaker B: That is a crucial point.
I love how passionate you are about this topic and I think you've made some very, very good points.
But it's true that the amount of research and the focus on research doesn't necessarily make a good resident. Right. The two are not equated one to one in all instances.
And I think the article does raise some crucial about having structure and it depends on who's reading this and who their audience is. And perhaps, you know, surgeons that are dabbling in research could use this as a framework to think, well, no, I need more. I need a research project manager to keep tasks on hand to be able to guide the medical students.
I'm always thinking that, you know, they're putting a lot on the orthopedic department to provide the initial understanding and foundation for understanding research. Some of that should come from the medical school as well.
There should be perhaps an elective or an extracurricular where those interested in research could refine those skills and understand the basics of how to develop an abstract, how to do a literature review. Because sometimes they're coming in very green and the expectation is that we teach them these things, and we do teach them these things when they're interested. But that's where the paucity of time and other things and resources to be able to teach the medical students is limited.
So I wonder if we could put some of the onus on the medical schools to provide that not necessarily as mandatory within the curriculum, but as an elective. Has mentioned, because people are doing this on the side, and as surgeons, we tend to do this on the side more so than medical physicians, who typically have an FTE protected time to do their research.
As surgeons, this is something we do out of interest. It's an extracurricular, if you will, for us as well. And therefore the time becomes limited.
[00:19:31] Speaker A: And the last thing that I would say, all the points that you raise are 100% on board. But the other point that I want people to hear, those who are listening to the podcast to, really, when I look at somebody, a medical student or even a resident who's applying to fellowship, and they've got like 80, 90, 100 publications, I don't look at the total number of publications.
I look at how many of these. Are you the first author?
The ones that you are the first author, what journal did it publish in and what was the topic?
The residency game is the same as the medical student game and maybe even more so. If you have a hundred articles published in residency, what were you doing other than like, when were you in the. Or when were you learning surgical skills? Right.
[00:20:19] Speaker B: That's a really good question.
[00:20:20] Speaker A: Right. And so it's like if, if you. I.
It's 24 hours in a day. Either you weren't doing the clinical stuff to. To let you just do your research thing, or you were getting your name put on a bunch of stuff and you actually weren't contributing in the way that.
Not, not saying that, that you don't meet the ICMJE criteria, but just like being a middle author on 50 papers doesn't really like. So you did some chart abstraction, so you did some literature reviews, so you formatted some tables. That's not the essence of the research really.
[00:20:59] Speaker B: Exactly.
Well said.
So we'll move on to the next article.
So the article I'll be discussing is entitled 18 Fluoride PET CT Analysis of post operative bone mineralization adjacent to acetabular cups in total hip arthroplasty A randomized control trial. This is by Soteriu and there is a commentary. I actually wrote the commentary so you'll get a bit of a sneak peek of my opinion on the paper as well.
This study was performed to compare bone ingrowth on two different acetabular cup surfaces and total hip arthroplasty, including a hydroxyapatite and trabecular porous titanium. This was measured with a fluoride positron emission tomography combined with computed tomography or fpetct, which captures metabolic events and is sensitive to bone metabolism. The study was an rct in which 28 Swedish patients aged 5069 with unilateral hip osteoarthritis were randomized to receive one of two models of press fit acetabular components during primary total hip arthroplasty. One group received a trabecular titanium what they refer to as a TT cup and this is by ANOVA Surgical and made of titanium and the other group received a hydroxyapatite or HA coated cup made of titanium with a layer of HA coating and this was also by Innovis Surgical.
Neither dual mobility nor constrained liners were used.
Primary implant stability was achieved without the need for screw fixation.
All surgeries were performed by a single surgeon via a posterior approach.
The contralateral healthy periacetabular bone served as the reference group and all FPET scans were performed using a GE discovery or a GE discovery MI. Both were from GE. Obviously, clinical outcomes were assessed preoperatively and at the 36 week follow up with the use of the Merle, Daubignier and Postell score modified by Charnley.
Radiographic assessment was performed by comparing the radiographs made immediately after surgery and those performed 36 week follow up.
To evaluate bone metabolic activity in the acetabulum, the authors used a modified polar map technique. The SUVs for all regions of interest or ROIs were first analyzed together and then the three ORIs with the highest SUVs were identified.
No complications were observed between the preoperative and 36 post operative time points. The mean functional outcome score improved from 13.2 to 17.7 in the TT group and from 14.2 to 17.7 in THE HA group, so no differences. The score final follow up was better than that preoperatively for all patients, indicating that there was no deterioration in clinical results during the follow up period.
On radiographs. At 36 weeks postoperatively, no radiolucent lines had developed in any of the Dalee and Charnley zones, indicating that all cups had stable fixation and to evaluate the FPET results and the different ROIs with respect to the actual cup orientation, the standard radiographic analysis of cup position was performed.
The prostheses and fluoride uptake were well visualized. At four weeks postoperatively, the metabolic activity for all ROIs analyzed together was 17% higher in the TT group compared with the HA group, although the difference was not significant at that time point. The SUV for all ROIs analyzed together was significantly higher in both the HA and the TT group compared with their respective healthy reference groups, that is the contralateral hip.
The three ROIs with the highest SUVs were the lateral inner, the central, and the lateral outer region of interest, and at four weeks the metabolic activity for these three ROIs analyzed together was significantly higher than that for the other six regions of interest analyzed together. In both the HA and the TT groups, lower activity was observed in the medial inner and the medial outer.
The analysis of the secondary outcomes demonstrated no significant difference in metabolic activity between the two study groups. At 16 weeks at any ROI and at that same time point, the SUV for each ROI was significantly higher in both the HA and the TT groups compared with their respective reference groups. At 36 weeks, only one ROI had significantly higher SUV in the TT compared to to the HA group. Then at that time point, activity in eight of the nine regions of interest in the HA group and five out of nine regions of interest in the TT group was significantly higher than then in the reference groups. From 4 to 16 weeks, metabolic activity decreased more in the TT group than in the HA group, with decreases observed in eight of the nine ROIs in the TT group and five of nine ROIs in the HA group.
From 16 to 36 weeks, the SUV decreased more in the HA group than in the TT group.
Ultimately, the study found that specific regions of interest had higher metabolic activity, including lateral inner, central and lateral outer groups. The authors conclude that FPETCT is an exciting new tool in orthopedic research. They suggest that this study validated the use of this technology in studying implant fixation and Osseous integration.
The study was limited by its small sample size. Furthermore, the two implant designs actually have additional morphological differences.
For instance, the HA coated cup also had an equatorial retentive grooves designed to enhance fixation, while the TT cup was hemispheric. These structural variabilities obviate the direct comparison of osseointegration of the TT versus the HA surfaces because they were slightly different in morphology.
Ultimately, the most valuable and clinical meaning aspect of the study for me is that it serves as a proof of concept investigation that identifies the foam PETCT as an emerging technology to study osseointegration implants in total hip arthroplasty.
Unlike conventionally used imaging including plain radiographs and technetium 99 labeled bone scans, the FPET CT can quantify metabolic activity of bone adjacent to the prosthesis with great precision and accuracy.
The pitfalls of this technique include greater radiation exposure and cost.
The study revealed a decline in bone turnover adjacent to the acetone component at 16 weeks, the time point following cup implantation. That findings of periacetadler remodeling are expected to completely normalize based on this imaging technique remains elusive.
So more to come on that technique.
Nonetheless, this technology remains a promising option, I think, for quantitative assessment for osseous integration of total hip arthroplasty. And it may be particularly valuable in instances where conventional methods provide sort of equivocal results in this assessment of aseptic loosening. So it may be an additional tool that we can use.
Furthermore, targeted mapping of ingrowth regions by FETF PETCT might be helpful in the revision setting to determine, you know, regions of interest for implant extraction of a, well, fixed acetabular component which can be, you know, particularly challenging in total arthroplasty. So I think potentially this is a very interesting technology that we could use in the revision setting as well.
Andrew, what are your thoughts on this paper?
[00:28:12] Speaker A: Well, I mean I agree with everything you said.
You know, very interesting technology, a lot of potential hypothetical advantages and applications going forward.
Not like it was the level of evidence appropriately as great as level two based on the limited sample, restricted event rate, issues with variation, restricted variation.
I'm just not sure why like the randomized trial was, was really needed in this. Like I, I feel like it could have just been done and maybe you could have had more patience or something if you obvious eliminate the cost associated with doing the randomization part and all this because at the end of the day you're just showing what this technology can do. It isn't it isn't proving anything beyond that.
[00:29:04] Speaker B: Yeah, exactly. It would have been nice to just longitudinally follow patients with 12 and see when does the uptake diminish and completely resolve. Would also be helpful, but that would require longer term follow up. But okay. And so for now for our your cases on hold featurette, we have Periprosthetic joint infection following total knee arthroplasty is associated with a significantly elevated risk of mortality. A population level database study by Ikhtiari and associates. This is the lead article, there is a commentary, an infographic and it is 30 days free.
This is a study out of Toronto that aimed to Compare long term 10 year mortality rates in patients who developed prosthetic joint infection within one year following total knee arthroplasty to those that did not develop PJI within one year following the index total knee arthroplasty.
In the background, the authors lay the groundwork for why this topic is so important. PJI is known to be quite devastating to patients as it is associated with significant burden of illness and is also very costly.
As the volume of primary arthroplasty continues to increase and patients live longer with implanted prostheses, the economic and social burden of PJI is projected to continue to increase and place a substantial burden on healthcare systems around the world.
This was a retrospective population database study of a single payer public system in Canada.
The data came from the ICES data repository, which is a set of linked administrative databases capturing all healthcare interactions for the population of Ontario, Canada's most populous province.
The studies reported according to the strength and reporting of observational studies in epidemiology or the STROBE guidelines and the relevant extension reporting guidelines conducted using using observational routinely collected health data or the record statement.
Those eligible to participate were adults undergoing their first primary elective total knee arthroplasty in the Province of Ontario between April 1, 2002 and March 31, 2021, so a large time frame.
Procedures were identified using the Canadian Institute for Health Discharge Abstract database or the same day Surgery database and these data were linked with the Ontario Health Insurance Plan or OHIP physician building codes for the same patient.
Patients who died within one year postoperatively were excluded to minimize the risk of immortal time bias and all patients were followed until they were censored because of mortality or until the end of the study.
All cause cause mortality was captured from the OHIP registered person's database. The main exposure was PJI requiring revision surgery on the Same joint within 365 days of the index total knee arthroplasty.
This case definition was used because it had a sensitivity of 92% and a positive predictive value of 78%.
Previously validated algorithms were used to identify patients with a history of range of comorbidities including copd, chf, hypertension and diabetes. Frailty was determined using the Johns Hopkins ECG system and the modified Charleston Dayo comorbidity score was also calculated.
Covariates were compared between the matched groups to confirm adequacy of the match and the primary outcome was 10 year mortality among patients who developed PJI within one year following the index total nery arthroplasty compared with those who did not develop PJI within that timeframe.
Patients in the overall cohort were grouped on the basis of whether or not they developed PJI within the first index surgery.
The demographics for the entire cohort and infection and match controls are reported descriptively.
Standardized differences for all covariates were calculated after matching and a standardized difference of greater than or equal to 0.1 demonstrated a significant difference or poor matching.
Mortality rates at 5 and 10 years postoperatively were calculated for the full cohort and by group and compared using hazard ratios. Mortality rates were plotted using Kaplan Meier survival and compared using log rank testing with the proportional hazards assumption formally tested and met.
So three separate preplanned sensitivity analyses were performed in addition to the main analysis. Number one matching patients by primary surgeon in addition to the above variables to control for variations in surgeon decision making reporting number two analyzing only patients less than 60 years of age to reduce the likelihood of natural or expected mortality and three repeating the main analysis for pneumonia within 30 days of surgery to compare the effect of another serious infection in the acute post operative period on mortality rates.
Given that different treatment strategies for pji debridement, antibiotics, implant retention or DARE versus single stage revision or two stage revision have substantially different impact on patients.
The authors compared the risk of mortality between early post operative PJI which they defined as less than or equal to 40 days after the index surgery and and late PJI which they defined as greater than 48 days after surgery as a surrogate for treatment strategy. They state that early PJIs are more likely to have been treated with a DARE compared to late PGIs and 48 days was selected as a cutoff as it was the median time to prosthetic joint infection.
Of the 263,000 patients, 263204 patients who underwent primary total knee arthroplasty during the study period the mean age was 67.9 years, 0.5 subsequently developed PGI. That's 1,228 patients subsequently developed PJI within one year.
Across the entire sample, the group that developed PGI within one year had a significantly higher rate of male patients, significantly higher rate of CHF and COPD and a significantly lower rate of patients with modified Charleston Dayo score of 0 compared with the group that did not develop PGI within one year.
In their cohort of over 260,000 patients, the incidence of PGI was low.
Patients who developed PGI within 1 year had significantly higher unadjusted 5 year and 10 year mortality rates compared with those who did not develop PGI within the following index total knee arthroplasty after matching total nearthroplasty, recipients who developed PGI in the first year had a significantly higher 5 year mortality rate and 10 year mortality rate rate.
In the sensitivity analysis for the same surgeon, 400 patients with PGI within one year were compared with 400 matched patients who did not develop PGI within one year and those who developed PGI within one year had a higher 10 year mortality risk of 7.25 versus 1.25. This was also the case for comparison of patients less than 60 years of old. 60 years of age they had 360 patients within one year that developed PJI compared to 60. 360 match patients who did not develop PJI that were all less than 60 years old and the 10 year mortality risk was 5.4% compared to 0.8%.
And in the analysis of 849 patients who developed pneumonia within the 30 days following total neurotoplasty compared to those that did not develop pneumonia within the 30 days after the index surgery there was no difference in 10 year mortality. So very interesting that another comparable infection there was no difference in 10 year mortality but PJI did elicit a difference.
The result of secondary post hoc analysis demonstrated that both early and Late POST OP PGIs were associated with significantly higher risk of mortality compared with their respective matched cohorts. However, the effect size was larger for late PJIs.
The study found an astounding greater than fourfold increase in mortality at 10 years in the patients with PJI within one year post op and compared to those that did not sustain a prosthetic joint infection within one year following surgery. To put this in context, this is comparable to the 10 year mortality of several adult cancers including prostate cancer and melanoma. It is very conserving given that the volume of total joint arthroplasty being performed is increasing and the most common reason for revision after total knee arthroplasty is pji.
The study is a retrospective database study, so it establishes an association between PJI within one year of surgery and mortality, but not a definitive causal association.
The authors speculate that physiologic burden of revision surgeries or loss of limb and salvage procedures following prosthetic joint infections may detrimentally affect overall health to the point of increasing mortality.
Limitations of the study pertain to it being a database with data collected from billing codes, which can result in inaccurate data.
The data also spanned a very long time period for more than 20 years, and there could be secular trends that impacted outcomes.
Another limitation of the study that I found was the assumption that all cases that the authors defined as acute by somewhat arbitrary cutoff of 48 days were treated with DARE and all chronic were revised. This was an interesting assumption because it's not the duration from time of index surgery that defines acute versus chronic infection, but rather the time from the onset of symptoms, and that timeframe is typically defined as four weeks, not 48 days. And while the acuity often does guide decision for DARE versus revision, there are other factors that play a role, such as the virulence of the organism, the presence of soft tissue compromise, or a draining sinus.
They cannot be certain by any stretch of the imagination that the cases that they define as acute acute all received a DARE and the ones that they defined as chronic received a revision. And that's one of the shortcomings of the database study is that they don't have that data of how the infections were treated.
There was also a lack of analysis on the basis of treatment type or pathogen, where likely data was not available to them in this database.
But I would say despite the limitations, this is a very well conducted study with very meaningful findings. It adds to the growing body of literature that it's a cautionary tale about health how bad PJI is. This adds to the importance of the work that we do to optimize patients prior to elective total knee arthroplasty, which I'm particularly passionate about, especially when some of the modifying risks can seem punitive. We have patients quit smoking before a total joint arthroplasty. Oftentimes the patients, including physicians and healthcare providers, are not aligned with why we need to optimize them so aggressively when it comes to diabetes, weight management, substance misuse use, tobacco use, and so on. And so the association with PJI and mortality rates comparable to cancers is quite compelling.
And I say, you know, this is a very important study and this case is not on hold for me. What do you think Andrew?
[00:39:52] Speaker A: So I think I agree with everything that you said.
I think that this is a very interesting natural history study with some very startling findings.
I think that this study for our interested readers is absolutely testable material.
And the association with PGI and 10 year mortality, I think it's just a matter of time before that's on the oite. I could see this paper for the web based longitudinal assessment, some of the figures in terms of the fourfold increase that you quoted.
Also in and of itself, not just the association but the actual magnitude of effect.
Where I'd like to start with this, I would say that the study itself I would not put on hold.
The clinical relevance is where my question regarding whether or not it should be on hold is and I would lean toward yes, it should be on hold.
And this is because I think as, as a community, what surgeons, you know, infections are amongst the most feared complications in, in any type of surgical procedure but in spine and in joints they really can.
An infection after joint replacement, after spine surgery can have catastrophic life changing effects. And for a variety of reasons, including the surgeon's own ability to sleep at night, we as a community want to convey the patient. Yes, this is an, you know, this is an event that we would wish would never happen, but we know it does happen.
We're going to treat this, we're going to get this taken care of, we're going to take it seriously and after it's done, once we have you infection free, this is not going to have long term ramifications on your outcome. And this study says. No, actually it is.
This study says, you know, a very extreme stance, but it's saying, you know, if an infection happens with a patient, you're contributing to their ultimate mortality risk in, in a very remarkable way.
And so I would want to push back on that a little bit. I think that it's not questioning what they found, but it's just questioning in the way that they found it. You know, they're doing a propensity score match but they're matching patients without infection to patients with infection. Right. And those are potentially two different populations at baseline.
So that's my first real like some of the other. There are a lot of points here, the, everything that I'm not going to delve into the, the use of like the 48 day cut off for. In fact I think there's there's more to it that I'm concerned about even than that. But just for the sake of time, and I know I went long on the, the first thing, I'm not going to go into that. I think your points absolutely stand and I would add to that. But, but my real area of concern here is they're doing a propensity score match, which is a causal inference test. And if you sort of drill that down to the basics, it's like you're randomizing patients to infection versus not infection.
And that's, you know, what they have accessible to them through the database. I don't think you can really distill. And they haven't proven to me, they, they don't, they're not particularly robust on the methodology in that domain, but they certainly haven't proven to me that these are patients who had the same baseline risk of infection at baseline and mortality at baseline. Yes, I understand they did their post match assessment. Standardized differences. That is a best practice. And it's, you know, they have comorbidities in there and they have a DAY score in there and they have rheumatoid arthritis in there and all that is well and good.
This is not the be all end all in terms of like infection risk and especially mortality risk, the baseline mortality risk. I think that that's, I would say that the patients who develop an infection are a different, are in a different part of the Venn diagram. And there is some intersection, but there's a piece that's outside the circle of the folks who did not develop infection.
And so they presented this pneumonia case as sort of a contrast to two, to further emphasize and shore up their argument that getting an infection after total knee arthroplasty is, you know, it's pneumonia. They didn't find the findings with pneumonia, but pneumonia after total joint arthroplasty could also be a different group of the Venn diagram than the people that have the, the periprosthetic joint infection and pneumonia isn't as bad as a periprostate. A real, you know, significant fulminate periprosthetic joint infection. We're not talking about like a little suture abscess or something.
And pneumonia is probably easier to treat. I, I would say on several fronts. I'm not a medical doctor in terms of like intern, I'm not a pulmonologist and I'm not I.D. or you know, maybe some of those disorders. Twins would take umbrage with that statement. But we'll let it stand as it is. That's not our listening audience.
What I would say a better comparison would be would be look at patients not with toe joint arthroplasty, but patients who develop sepsis and see what their timeline for recovery is versus the PJI group.
That, I think, would be a more meaningful comparison regarding how the event itself of the PJI is a foundational and fundamental shift in the survival trajectory of these patients.
[00:46:08] Speaker B: Got it.
I agree with you. I think, however, the plight of the arthroplasty surgeon is to identify and really put it out there that the consequence of a PJI is of such great magnitude.
And so we're taking another thing that people think, generally speaking, is bad pneumonia and saying the mortality associated with that is lower. Now, whether their methodology was sound enough to really prove that. But I see where the impetus is coming from, why they chose pneumonia versus sepsis, because sepsis is likely going to be much, much worse than pji. If the patient was not septic, for instance, or comparable. We don't know the difference. But I think that might have been their rationale for selecting something like pneumonia.
[00:46:52] Speaker A: Yeah, I mean, they very well, what they're trying to say is some other kind of infection after pji.
But the people who are developing that may be a completely separate group from the people who are developing pji. And that's what they have not shown me. They have not shown me that there was clinical equipoise in the risk of PJI at baseline. There wasn't. Clinical equipoise.
Yeah, I'm not using that term exactly correctly and I know that, but just go with it. I'm not. There's not clinical equipoise in terms of the five year mortality risk. I want two patients who have the same five year mortality risk, one of them gets a PGI and then we see actually how that impacts.
[00:47:33] Speaker B: Got it. And now we can't do that from a database though, right? Correct. No, we have to have prospective data and that's really hard to conduct in this, to have the numbers to be able to follow that.
[00:47:45] Speaker A: But there's also a moral obligation, I think, on the part of researchers to contextualize their work and not over sell the magnitude and the impact like what they're showing here. This has the potential to. Some family gets this after their family member had a periprosthetic joint infection and says, well, look what you did made their mortality risk go up significantly over the next 10 years. I mean, I think that that's another burden that's going to be placed on arthroplasty, folks, after this research, that's interesting
[00:48:19] Speaker B: how you read that accountability that it's put on the surgeon. I'm almost thinking, well now these are the things we really need to hone in for modifying risk prior to surgery, you know what I mean?
And engaging the PCPs and the patient themselves to invest in that.
[00:48:36] Speaker A: And you're going to get it. I mean like it's less than 1% of infection of the total number. I mean the, the, the incidence rate is quite low. Right?
[00:48:47] Speaker B: And theirs was even lower than, you know, typical rates, typically 1 to 2%. Theirs was 0.5%. Yeah, excellent. Great points.
Okay, so now we'll move on to our honorable mention. This is entitled A Fully automated Multi State Stage Deep Learning System for Lengthy Classification Enhanced Diagnostic Precision in Adolescent Idiopathic Scoliosis by zoo. There is a commentary and this is permanently free.
Adolescent idiopathic scoliosis, or AIs, is a complex spinal deformity for which surgical correction remains key strategy to halt progression. The lengthy classification is a cornerstone for surgical planning, categorizing curves into six types with optional lumbar and sagittal thoracic modifiers.
The system accounts for three dimensional deformity facilitating accurate classification and guiding intervention.
The traditional classification relies on clinical expertise and imaging assessment, but it introduces subjectivity, inter observer variability and inefficiency. Deep learning has emerged as a powerful tool for medical imaging analysis and predictive modeling. This paper demonstrates a novel, validated and fully automatically automated deep learning system for precise lengthy classification using spinal x rays. It was a retrospective study of 650 patients from China and Tibet with 183 healthy controls and 467 patients with AIs. Patients were aged 10 to 18 years with a COBB angle less than or equal, greater than or equal to 25 degrees and less than 90 degrees with full spine radiographs. Multi stage deep learning system consisting of SWIN unit segmentation of vertebrae for automated COBB angle measurement as well as a deep lab 3 localization of lumbar pedicles to determine modifiers via the centroid to CSVL or the central sacral vertebral line distance and a fusion model integrating features to curve types and lumbar and sagittal thoracic modifiers was designed to perform end to end lengthy classification automatically.
Validation used an independent test set. The system achieved 95.6% overall accuracy in lanky classification and had A macro averaged F1 score of 0.862. Vertebral segmentation attained dice coefficients of 0.917 or 0.917 AP and 0.942 lateral Cobb angle measurements showed excellent agreement with those of expert modifier assignments achieved F1 scores 0.912 for the lumber and 0.928 for sagittal, exceeding clinical acceptability thresholds. In conclusion, the fully automated system was able to perform rapid, objective, interpretable and clinically reliable classification by the Lenki type directly from radiographs with performance comparable to that of expert assessment. And it demonstrates the potential for standardizing AIs surgical planning, reducing diagnostic variability, and improving surgical workflow efficiency.
So there you have it. That concludes our episode 108 for your cases on hold. Stay tuned for our next episode, where my colleague Dr. Schoenfeld will be hosting.
