Editor’s Spotlight/Take 5: Can Topical Vancomycin Prevent Periprosthetic Joint Infection in Hip and Knee Arthroplasty? A Systematic Review

Let’s imagine a clinical problem that can cause real harm; periprosthetic joint infection (PJI) will do. Now imagine there’s an available treatment that seems sensible; let’s call it vancomycin powder. By available, I mean you have a surgical nurse who will put it into your gloved hand if you ask for it. By sensible, I mean it doesn’t cost an arm and both legs, it treats most of the bacteriological bad actors, you haven’t seen anyone get hurt by it, and maybe there’s even a little evidence suggesting it may reduce the risk of PJI after THA [1].

This is the proverbial no-brainer.

Now, imagine that eight other studies find no evidence of efficacy of that same treatment. And while there was no evidence of harm either, you can imagine it causing harm—we know vancomycin carries the risk of real toxicity—without stretching your imagination too much.

This is a much harder problem. As much as we’d like things to be otherwise, orthopaedic surgeons can’t fall back on a wealth of randomized trial evidence to support all of our treatment decisions. In fact, we find ourselves basing our choices on observational research and other forms of storytelling about as often as not. And from what we know about how positive-outcome bias influences our base of available evidence, we can confidently surmise that what we think we know about the benefits of our interventions is grossly overestimated [4, 5, 10]. Indeed, proving safety—or even getting a rough handle on the frequency with which uncommon harms occur—is all but impossible with the sample sizes we see in common orthopaedic studies [6].

No longer a no-brainer, right?

Such were the main findings in a wonderful systematic review by Dr. Nicholas Desy’s group from the University of Calgary in Canada in this month’s Clinical Orthopaedics and Related Research® [11]. This should command our attention because vancomycin powder is in widespread use.

Dr. Desy’s team shared that eight of nine studies found no reduction in PJI risk associated with the use of vancomycin powder after THA or TKA, with odds ratios tightly embracing the line of no difference. Although they found no apparent harms in the six studies that reported on side effects, those studies were far, far too small to expect to turn up differences in the kinds of severe-but-uncommon vancomycin-related complications we care about, like ototoxicity or nephrotoxicity.

Their work deserves your interest, no matter your subspecialty, for two reasons: First, the main findings probably apply to other major-but-common orthopaedic interventions, since if anything, arthroplasty presents one of the stronger use-cases for vancomycin powder that we have, given the terrible morbidity PJI causes. If the risk-reward calculation doesn’t come down clearly in favor of vancomycin powder for preventing PJI, it may not be very useful for other routine procedures either. And second, because the general question (“What do you do with a treatment that seems sensible but has not been proven effective, and may have real risk?”) resonates across our specialty, as do Dr. Desy’s answers, both in the paper [11] and in the Take 5 interview that follows below.

Would you use vancomycin powder, given these concerns and the discoveries in this new systematic review [11]? Read Dr. Desy’s interview here, and share your thoughts in a letter to the editor at eic@clinorthop.org.

Take 5 Interview with Nicholas M. Desy MD, FRCSC, senior author of “Can Topical Vancomycin Prevent Periprosthetic Joint Infection in Hip and Knee Arthroplasty? A Systematic Review”

Seth S. Leopold MD: Congratulations on this study, the findings of which I think really should change practice. But we’ll come to those in a moment. First, to get the attention of a broader group of readers—nonarthroplasty surgeons—can you think of a few examples of orthopaedic interventions that might fit under the definition I mentioned in the commentary, namely, treatments that seem sensible but have not been proven effective, and may have real risk?

Nicholas M. Desy MD, FRCSC: Thank you for those kind words. Some studies on the spine point to the success of vancomycin powder in reducing the incidence of perioperative infections [2, 8, 9]. Based on my experience with this review, I would encourage those in other orthopaedic subspecialties to examine the quality of the evidence that may support the use of vancomycin powder to reduce infections. In addition to spine surgery, some examples that come to mind include various lower extremity fractures (open or closed), pelvic and acetabular fractures, and foot and ankle procedures. Most papers looking at vancomycin powder in these settings are retrospective studies

Figure 1.

Nicholas M. Desy MD, FRCSC

[7, 9].

Dr. Leopold: One issue that is commonly missed but you covered so well is the idea that something as seemingly obvious as throwing an antibiotic powder into a wound to prevent an infection may have a counterintuitive (and disturbing) effect. You pointed out that because the odds ratios bracketed the line of no difference, “vancomycin powder may have, in fact, been associated with an increase or a decrease in the risk of PJI” [11]. The unintended consequences of our surgical choices may be more common than people realize, and this possibility should put our antennae up. In light of findings like yours, can you offer readers some tips about how not to be misled by what they read in orthopaedic journals?

Dr. Desy: I agree that our seemingly benign or obvious surgical choices may have unintended consequences. As I reflect on my own journey with vancomycin powder, I recall first using it based on my limited knowledge of the evidence from spine studies. I also learned of other trauma surgeons who would use it for patients with complex fractures that were either open, took a lot of operating time, or had a substantial soft tissue injury component. This was mostly based on anecdotal evidence. I used it in patients with revision hip and knee arthroplasties, and it then slowly crept into complex primary procedures, and after that, I started to use it in procedures I performed on patients with known risk factors for PJI in the simple primary setting; eventually, I took to using it for all arthroplasties.

After performing this review, however, I have stopped its routine use. Performing this review was eye opening. The evidence was not nearly as robust as it seemed before we began, and more of my choice than I realized was based on anecdotes and papers that were not of good quality. I urge readers to look for systematic reviews like the one we performed to help guide decisions of this sort, and to pay careful attention both to whether complications are being adequately reported and whether treatment benefits are as clear as they need to be to justify a clinical intervention. I find it challenging to do this on my own on a routine basis, which is why I also rely on journal clubs, conferences, and clinical practice guidelines.

Dr. Leopold: You point out that some research that has pooled data has arrived at different conclusions than yours [3], perhaps because they included patients with established PJI rather than as a means to prevent it in patients with no history of infection. This should cause us to read studies’ inclusion criteria carefully and generalize their findings only with care. Still, I think your take-home messages may carry a bit outside the boundaries of primary joint arthroplasty. How do you see it?

Dr. Desy: Yes, it is important to make sure that the studies we read really are answering the questions we want answered. It is possible—though I don’t believe conclusively proven—that there are benefits to vancomycin powder in patients who are being treated for PJI. And the fact that we found no benefit to vancomycin for PJI prevention doesn’t exclude the possibility that it could have a small benefit. However, after reading carefully the papers we included in this review, I would say it seems unlikely that vancomycin powder will ever be shown to have anything more than a small effect size, if any, for that purpose. The key message outside the boundaries of primary joint arthroplasty here would be to continue to ask questions and look for better evidence, and not jump straightaway to a large practice change, such as the one I made prior to this review, when I incorporated vancomycin powder into my practice for primary total hip and knee arthroplasty based on paper with lower levels of evidence and anecdotes.

Dr. Leopold: Let’s feed the joints surgeons just a bit here at the end: Given the gaps in what we know, and the chance that there may be some benefit to vancomycin powder in patients with established infections [3], in what situations might you might still use vancomycin powder?

Dr. Desy: Regarding both THA and TKA, vancomycin powder is often included in antibiotic spacers as part of a two-stage exchange treatment strategy for established PJIs. Putting antibiotic spacers aside, I also consider using vancomycin powder in the setting of a debridement, antibiotics, and implant retention (DAIR) procedure for acute PJI; during the second stage of a two-stage exchange; and in primary and revision total hip and knee arthroplasty in patients with particular risk factors for PJI, such as obesity, smoking history, and poorly controlled diabetes where patients are just not able to adequately improve these modifiable risk factors preoperatively. I think about incorporating powdered vancomycin in these circumstances to provide some local antibiotic delivery to treat an established infection during a DAIR procedure or because the risk of infection is much greater compared to patients undergoing straightforward primary THAs or TKAs without any risk factors for PJI.

Dr. Leopold: Both you [11] and others [3] have called for multicenter and/or randomized trials to answer the question of whether vancomycin powder is effective to prevent PJI. Let me challenge you a bit: If PJI occurs in about 1% of patients without using vancomycin powder, it would take several thousand patients in each arm of a randomized trial—perhaps more—to demonstrate even a modest benefit in reduction of PJI. You’d probably still not be able to detect a difference if the effect were smaller. Since it’s not possible to fund nor find the resources to test every plausible question, why shouldn’t we see your study as sufficient to say this avenue of inquiry does not justify that kind of time, energy, and great expense? Why not deploy that same effort to investigate more-promising treatments? Not every research question deserves our attention, right?

Dr. Desy: While the risk of a patient experiencing a PJI is quite low, it is still a devastating complication to the patient and puts a much larger burden on the medical system. When I treat a patient with PJI, I always ask myself how we can bring the risk to 0%. We’ll never get there, but I believe it is worth exploring potential options that may further decrease the risk of PJI without increasing the risk of other complications, such as the incorporation of topical antibiotic powder.

Yes, it would require quite a large sample size to detect a possible difference with the use of vancomycin powder, but our study alone is not enough to conclude whether vancomycin powder is worth using to prevent PJIs. And while not every research question deserves or can be answered with a rigorous randomized trial, I believe such an approach for vancomycin powder would be simple to perform, as far as randomized trials go. The surgical procedure would not change much, and we would not be randomizing patients to either operative or nonoperative treatment, which can be difficult for patient enrollment. The challenging part is the sample size as you alluded to, which can be solved with multicenter collaboration.

I also believe that a good use of resources would be to fund projects that look at preventing PJI in conjunction with studies that look to improve its treatment. A similar scenario is the use of perioperative intravenous (IV) antibiotics. It is quite clear that the administration of a certain IV antibiotic given within a definite period prior to skin incision reduces the risk of a surgical site infection. What remains controversial is whether there is a need for any postoperative IV or oral antibiotic doses which has led to various practice patterns and deserves further study as well.