Answers are grounded in this paper and cite their sources.
Expert commentary reviews pharmacokinetic dosing of vancomycin in MRSA infections, highlighting the shift from trough levels to AUC-guided therapeutic monitoring.
(See the Major Article by Lodise Jr. et al on pages 1536–45.) Vancomycin remains the mainstay of therapy for treatment of serious methicillin-resistant Staphylococcus aureus (MRSA) infections in spite of the continuing deliberation on whether it is obsolete compared to newer agents [1, 2]. Most clinicians use vancomycin for both empiric and definitive therapy of systemic MRSA infections as outlined in current Infectious Diseases Society of America (IDSA) treatment guidelines [3]. Vancomycin should be continued as long as the patient is clinically and microbiologically responding to therapy, other foci of infection have been adequately debrided and removed, and susceptibility testing shows a minimal inhibitory concentration (MIC) of 2 μg/mL or less (Clinical and Laboratory Standards Institute [CLSI] and US Food and Drug Administration susceptible breakpoint). Alternative treatment options should be considered when an isolate of MRSA has a vancomycin MIC that is >2 μg/mL. Monitoring of serum drug concentrations and optimal dosing of vancomycin continues to be examined despite more than 60 years of clinical use of this antibiotic [4–7]. Numerous in vitro, animal, and human studies provide support that the most appropriate pharmacokinetic-pharmacodynamic target for vancomycin is the 24-hour area under concentration-time curve (AUC0–24) to MIC ratios (AUC0–24/MIC) of ≥400 based on total (ie, bound and unbound) blood concentrations [4, 6, 8–11]. The 2009 consensus guidelines for therapeutic monitoring of vancomycin suggested that trough-based serum vancomycin concentrations of 15–20 µg/mL could serve as a surrogate of an AUC0–24/MIC ratio of at least 400 when a MIC value was ≤1 µg/mL, although clear clinical evidence to support this exposure target was lacking [4]. Numerous studies and publications continue to be generated regarding different strategies for the dosing and monitoring serum concentrations of vancomycin [5–7, 12]. Several key observations reported in the past decade have confirmed the following: (1) vancomycin MIC values are method specific (eg, broth microdilution vs Etest) and vary among commercial MIC testing methods, which requires consideration when determining pharmacokinetic-pharmacodynamic target values for AUC0–24/MIC [6, 9, 13–16]; (2) higher vancomycin MIC values for isolates of methicillin-susceptible S. aureus (MSSA) and MRSA appear to influence clinical outcomes associated with beta-lactam agents and vancomycin, respectively [17, 18]; (3) clinical outcomes with vancomycin are a function of bacterial load, site of infection, and presence of biofilm [19–22]; (4) targeting of higher vancomycin trough concentrations (eg, 15–20 μg/mL) is associated with modest improvements in clinical outcomes of selected patients [11]; (5) vancomycin trough concentrations are an imprecise surrogate for AUC0–24 such that patients with trough concentrations <15 μg/mL can have AUC0–24 ≥ 400 μg h/mL [7, 23–26]; (6) higher rates of nephrotoxicity and acute kidney injury (eg, 20%–30%) are more commonly associated with trough concentrations of ≥15 μg/mL compared to <15 μg/mL, and this becomes even more evident when trough concentrations are >20 and 30 μg/mL [27–30]; (7) vancomycin nephrotoxicity may be better determined by AUC values instead of trough concentrations [30–35]; (8) higher rates of nephrotoxicity occur when vancomycin is combined with an aminoglycoside, piperacillin-tazobactam, and other nephrotoxic agents [27, 36, 37]; (9) evidence is increasing that these pharmacokinetic-pharmacodynamic relationships are applicable to pediatric patient populations [38–41]; and (10) use of loading doses and continuous infusion versus the traditional intermittent dosing of vancomycin is another option in selected patients [6,7,19]. Obviously, no single factor can serve as a predictor of nephrotoxicity and/or success or failure for vancomycin therapy [11]. Applications of these findings to specific patient populations (eg, obesity, renal disease, critically ill patients with and without renal replacement therapy), Gram-positive pathogens other than MRSA, and/or different sites or types of infections other than bacteremia further limit our current knowledge in the optimal dosing and monitoring of vancomycin [6,7,19]. In this issue of Clinical Infectious Diseases, Lodise and colleagues present results from the observational study, PROVIDE, evaluating the association between clinical outcomes and vancomycin exposure in adult patients with MRSA bacteremia [42]. The authors provide support to the accumulating literature in MRSA bacteremia for dosing vancomycin early in therapy and being based on an AUC0–24 target value for patients with a MIC value of 1 µg/mL or less. This study, along with several other recent publications, suggest that an AUC0–24 range between 400 and 600 µg•h/mL was most commonly linked to clinical success, whereas an increased risk for vancomycin-associated acute kidney injury or nephrotoxicity occurs when an AUC0–24 threshold exceeds a range of 600–800 µg•h/mL [24, 31–35, 42]. These ranges should serve only as guidelines and not be considered as a panacea for ensuring clinical success and minimizing toxicity. The DOOR risk-benefit analysis in this current study illustrates this, in part, by observing that the best global outcomes occurred in the 2 lowest AUC exposure quintiles (AUC ranges of 94–515 µg•h/mL) versus the 3 higher exposure quintiles (AUC ranges of 515–1755 µg•h/mL). Other pharmacological and physiological variables such as the wide range of protein binding (ie, 10%–55%) observed in patients may be contributing to why the most desired outcomes are occurring over a wide range of total drug concentration exposure [43–45]. The 2 most common dosing methods currently being recommended for achieving target AUC0–24 ranges include Bayesian forecasting computer software programs (using 1 or 2 vancomycin concentrations) and log-linear pharmacokinetic-based equations (requiring 2 vancomycin concentrations) [7, 12, 24, 35, 42]. Bayesian dosing programs and optimal sampling strategies for monitoring serum vancomycin concentrations to improve the accuracy and precision of predicting vancomycin exposure have been used in clinical practice for over 3 decades [46–49]. There are several Bayesian software programs currently available with various degrees of flexibility to allow the use of different pharmacokinetic models, population pharmacokinetic parameters, assay variability, dosing and serum concentration history, and covariates such as weight, creatinine clearance, and age to be incorporated [50–52]. Although Bayesian forecasting is one of the most accurate dosing techniques (with as few as 1 vancomycin concentration), it requires access to computer software, accurate collection of patient information, and pharmacometric knowledge about modeling and choice of a priori parameters in order to maximize vancomycin dosing with and without serum concentrations [7, 12, 53, 54]. An alternative method for AUC-based vancomycin dosing includes the collection of 2 serum vancomycin concentrations during a single dosing interval and the use of several pharmacokinetic equations to determine the elimination rate constant and AUC by a linear trapezoidal method [7, 12, 35]. The dosing method can be easily applied at the bedside with a handheld calculator and can serve as a reasonable alternative to Bayesian forecasting software. The major disadvantages of this method are the accurate recording of dosing and sample collection times, optimal sampling window of the 2 required serum vancomycin concentrations, and that the data will not account for acute physiologic changes that may occur after its collection time. Both AUC-based vancomycin dosing methods have shown to be accurate and precise to reaching the desire target ranges of AUC0–24 compared to trough-based dosing strategy or dosing nomogram [12, 24, 55]. Other additional benefits that have been observed include lower daily doses of vancomycin, fewer vancomycin concentrations being collected, less subsequent dosage adjustments, and a shorter duration of therapy [24, 56]. Several institutions have already reported their experience and/or implementation plans for switching to AUC-based vancomycin dosing and moving away from monitoring trough-based monitoring of vancomycin concentrations in the range of 15–20 µg/mL [35, 56–59]. Although this dosing method is gaining momentum, a recent survey of a national health consortium reported that the majority of their 124 academic medical centers were unlikely to change from trough-based dosing of vancomycin over the next year(s) due to unfamiliarity of AUC-based vancomycin dosing [60]. As more literature and updated consensus guidelines for therapeutic monitoring of vancomycin become available, it is likely that more institutions will consider AUC-based vancomycin dosing as part of their monitoring practice. The individualization of vancomycin based on pharmacokinetics equations and/or Bayesian forecasting software is “déjà vu” for many of us. Although Lodise and colleagues suggested that this as the “Emperor’s New Clothes,” it looks more like we are dusting off the Emperor’s finest garments from the back of the closet instead of spending a lot of money on new-fangled magical clothing [42]. Just as in the 1980s and 1990s [61–64], the controversies and debates about the value of serum vancomycin concentrations for dosing of vancomycin will continue mainly because of the wide dose-response relationship of vancomycin observed in various types of infections and patient populations, as well as the difficulties to clearly establishing dose-toxicity relationship in seriously ill patients with MRSA infections [6, 7, 19, 27]. In view of the limitations of existing reports, sound clinical judgment must always rule the kingdom of vancomycin use. Potential conflicts of interest.
No takes yet. Share an insight, caveat, or question.
Keith A. Rodvold (2019) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: