Heaping the Pelion of Vancomycin on the Ossa of Methicillin-resistant Staphylococcus aureus: Back to Basics in Clinical Care and Guidelines

(See the Viewpoints Article by Dilworth et al on pages e675–81.)

In Greek mythology, the Titans piled Mount Pelion and Mount Ossa upon Olympus to scale the dwellings of the Gods but were overwhelmed by Zeus. “Heaping Pelion upon Ossa” thus refers to making monumental efforts that nevertheless fail to achieve their objective. This aphorism applies well to the never-ending struggle of having medical guidelines accelerate the translation of research into clinical practice, improve the quality and safety of healthcare, and reduce inappropriate practice variation [1]. The Viewpoint article by Dilworth et al that appears in this issue of Clinical Infectious Diseaseshelps underscore this dilemma [2].

Some years ago, national practice guidelines were published recommending that vancomycin dosing be adjusted to maintain trough (Cmin) serum concentrations between 15 and 20 µg/mL [3]. The evidence supporting this recommendation was based on hypothesis-generating, not hypothesis-confirming, data (eg, mouse neutropenic thigh data and a retrospective study of Staphylococcus aureus pneumonia [3, 4]). Once guidelines are published, they tend to set a medical–legal standard of care that drives everything from national regulatory reviews to payer reimbursement to hospital protocols and order sets, and, unfortunately, to lawyer behavior. Thus, setting such guidelines based on hypothesis-generating data creates a ripple effect that can lead to unwanted side effects, especially if subsequent findings do not confirm the hypothesis-generating data on which the guidelines were based.

More recently, vancomycin dosing guidelines have been revised, switching from a focus on trough-level monitoring to use of the 24-hour area under the concentration-time curve (AUC₂₄) [5]. The new guidelines describe an optimal vancomycin serum AUC₂₄ range between 400 and 600 in patients with serious “suspected or definitive” methicillin-resistant S. aureus (MRSA) infections (eg, bacteremia, infective endocarditis, osteomyelitis, meningitis, and pneumonia) [5]. This guideline revision was driven mainly by concerns over increasing rates of vancomycin-associated acute kidney injury (VA-AKI) in patients with AUC₂₄ concentrations greater than 650, which occurred as a result of the previous recommendation to increase dosing to maintain higher Cmin. The change was also driven by a concern expressed in the guidelines that Cmin values in some studies may not have correlated with AUC₂₄ [5, 6]. However, numerous other studies have demonstrated good to excellent correlations of Cmin with AUC₂₄ in both adults and children [7–22], so troughs may perform better as a surrogate for AUC₂₄ than is presumed in the new guidelines.

Ironically, the new guidelines appear to be limited by concerns that are similar to those of the prior guidelines: they may be based on hypothesis-generating, rather than hypothesis-confirming, data. More data, and higher-quality data, are needed to determine whether vancomycin dosing based on AUC₂₄ improves patient outcomes. In this regard, Dilworth and colleagues present an interesting viewpoint regarding the everyday practical aspects of implementing AUC₂₄-based monitoring, optimal use of vancomycin, and adherence to the recently released guideline recommendations.

The viewpoint underscores multiple challenges in approaching this emerging problem. Who is at most risk for serious MRSA infections? What is the best vancomycin management regimen for suspected or definitive MRSA infections in the setting of early infection? How does one limit the emergence of nephrotoxicity while trying to achieve the best clinical outcome? And, given the limits of available data that support altering vancomycin dosing, are there higher-priority, higher-reward clinical practice changes we could focus on to improve clinical outcomes until better data are available to inform vancomycin dosing?

Unfortunately, randomized, controlled trial data that pertain to vancomycin therapeutic drug monitoring (TDM) are scarce. In the recent guidelines, most recommendations regarding vancomycin TDM in adults are based on graded moderate- to good-quality evidence (A-II). While Bayesian dosing to target optimal AUC₂₄ might lower the risk of VA-AKI compared with trough-based dosing, it remains unclear if dosing to AUC₂₄ targets actually achieves better clinical and bacteriological response rates. Furthermore, 2 high-quality, randomized, controlled trials that compared vancomycin therapy to alternatives create doubt as to the wisdom of adjusting vancomycin dosing to improve outcomes. In those trials, adjusting vancomycin dosing to achieve higher levels not only did not improve cure or mortality, it worsened the frequency of VA-AKI and thus likely led to net harm [23, 24]. The fact that clinical cure or mortality were not improved by higher, trough-based dosing raises 2 important questions. First, if raising vancomycin to clearly toxic levels does not improve efficacy, why would raising it to lower, somewhat safer levels (as suggested by revised guidelines based on AUC₂₄ rather than troughs) improve efficacy? And, if it does not improve efficacy, why do it at all?

In their Viewpoint article, the authors make 2 fundamental, very important points about vancomycin safety and efficacy. With respect to safety, the drug is often overused in hospitals. Inappropriate or unnecessary vancomycin cannot, by definition, improve clinical outcomes but can add to toxicity. Thus, a more parsimonious, more effective approach to minimizing toxicity may be to focus on effective antibiotic stewardship efforts to decrease prescribing overall, rather than optimizing dosing based on limited, hypothesis-generating data. Furthermore, if vancomycin de-escalation is triggered by negative cultures within the first several days of therapy, the drug will be stopped before even the limited, hypothesis-generating data suggest any potential benefit to altered dosing (which appears to be after day 3 or 4 of therapy).

Similarly, with respect to efficacy, perhaps rather than seeking dosing alterations, we could focus on more promising avenues. There is reason to believe, for example, that a beta-lactam that is active against MRSA (eg, ceftaroline, ceftobiprole) might be more effective than vancomycin in treating MRSA bloodstream infections, but no such trial has yet been completed [25]. Other technologies, such as bacteriophages, immune modulators, and antibodies, may also merit investigation [24, 26, 27]. Perhaps the most fundamentally important point to focus on with respect to improving efficacy (akin to stewardship to limit toxicity) is earlier, more regular, and more complete achievement of source control [28–31].

Overall, as suggested in the Viewpoint article, perhaps the time has come for the vancomycin world to get back to basics and adopt a “less is more” approach to many serious MRSA infections. Better antibiotic stewardship and better source control are more likely to feasibly improve outcomes (safety and efficacy) in the near term than the adoption of technically sophisticated dosing alterations or even new technologies.

Finally, the infectious diseases community must come to grips with the manner in which we have adopted clinical guidelines that have driven practice. Despite the best intentions of academics to keep guideline documents from creating standards that are treated by regulators, payers, practitioners, and lawyers as if stone tablets descended from the heavens, this is indeed what tends to happen. Guidelines that set such standards should be based strictly on high-quality prospective clinical trials. Summaries of lower-quality data can certainly also be presented (even in the same documents, possibly) but should be clearly marked as consensus opinion of experts, separate and distinct from formal guidelines. It would perhaps be best if such opinion and hypothesis-generating data were not used to set formal guideline standards.

Let us remember the words of that most famous of experts, Dr William Osler: “Half of everything I’m teaching you today is wrong; the only problem is, I don’t know which half.” Guidelines must be based on more than expert opinion, and experts must be humble enough to acknowledge that at any moment in history, we are frequently wrong and future scientific research will teach us paths to better care.

Notes

Financial support. This work was funded by the National Institute of Allergy and Infectious Diseases at the National Institutes of Health (grants R01 AI130060, AI1117211, R41 AI155423, and R42 AI106375 to B. S.).

Potential conflicts of interest. In the last 24 months, W. F. W. had no conflicts. S. C. J. J. has received consulting fees from Sunovion Pharmaceuticals. B. S. has received consulting fees from Pattern Diagnostics, Acurx, and Iqvia. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.