Highlights of Significant Changes in the New Vancomycin Therapeutic Monitoring Guidelines

The Big Picture:

Previous 2009 guidelines recommended the use of trough monitoring as a surrogate marker for the area under the curve divided by minimum inhibitory concentration (AUC/MIC) of 400 mg*h/L to simplify vancomycin therapy management, but this came at the cost of increased nephrotoxicity and acute kidney injury (AKI).

Trough-based dosing is no longer recommended, and an AUC/MIC-guided dosing with a target of 400-600 mg*h/L within the first 24-48 hours is suggested for most invasive methicillin-resistant Staphylococcus aureus (MRSA) infections in adults and children. PK/PD and toxicodynamic studies show a significant reduction in vancomycin exposure and nephrotoxicity without compromising outcomes when using AUC/MIC vs traditional trough monitoring approaches.

Loading doses for patients who are critically ill, on renal replacement therapy, or on continuous infusion therapy should be based on actual body weight.

Introduction to the New Vancomycin Therapeutic Monitoring Guidelines1

Since the 2009 publication of the vancomycin consensus guideline, many others have evaluated the clinical safety and efficacy of trough-based dosing. In 2020, the American Society of Health-System Pharmacists (ASHP), the Infectious Diseases Society of America (IDSA), the Pediatric Infectious Diseases Society (PIDS), and the Society of Infectious Diseases Pharmacists (SIDP) revised their 2009 recommendations and published a new consensus on the optimization of vancomycin dosing and monitoring for serious MRSA infections such as bacteremia, sepsis, infective endocarditis, pneumonia, osteomyelitis, and meningitis. These changes arose from vancomycin-induced nephrotoxicity (VIN) and AKI associated with using 15 to 20 mg/L trough-based dosing without regard to AUC.
‍
Key elements of the revised 2020 vancomycin dosing guidelines include: the transition from a trough-based to an AUC-based dosing target, ‍the utilization of Bayesian dosing to accurately calculate daily AUC values with non-steady-state levels, and an AUC/MIC target of 400-600 mg*h/L for serious invasive MRSA infections.
‍
‍Bayesian-guided precision dosing platforms are essential tools to facilitate the implementation of AUC/MIC-based vancomycin dosing guidelines and to optimize clinical practice. PrecisePK is a Bayesian dosing software that has been shown to predict the future vancomycin AUC most accurately in critically ill patients, who are at increased risks of treatment failure due to altered antimicrobial pharmacokinetics2, according to a recently published peer-reviewed study3. Since 1986, we have worked with several renowned institutions such as UC San Diego, UC San Francisco, Sharp Healthcare, etc.

Why use target AUC/MIC versus trough when dosing Vancomycin?

• AUC/MIC is the best predictor of vancomycin activity against methicillin-susceptible S. aureus (MSSA), methicillin-resistant S. aureus (MRSA), and glycopeptide-intermediate S. aureus (GISA), according to an in vivo cidal and pharmacokinetic/pharmacodynamic (PK/PD) parameter study in neutropenic-mouse thigh infection models conducted by Ebert and colleagues in 1987.

Figure 1: The adapted plot above by Rybak 2006 demonstrates the correlation between efficacy and 3 different pharmacokinetic models in an experimental mouse infection study using data from Ebert 1987. Area under the curve divided by minimum inhibitory concentration (AUC/MIC) over 24 hours shows the strongest correlation with methicillin-susceptible S. aureus (MSSA)-cidal activity compared to peak/MIC and time above MIC.

• Trough-based dosing is a poor surrogate marker for AUC and efficacy endpoint overall. A study by Drennan5 et al. showed that predicted median total vancomycin concentration-time curves associated withdifferent dosing frequency yield can yield very different AUCs and peaks despite targeting the  recommended trough of 15mg/L.

Figure 2: Graph adapted from Drennan et al. showing two different dosing regimens (intermittent infusions every 6 hours vs every 12 hours) with with a fixed target trough concentration of 15 mg/L for an 80-kg person with a GFR of 120 mL/min based on a two-compartment population PK model by Thompson et al. 2009. With continuous dosing targeting a trough of 15mg/L, the expected AUC24 is predicted to be 370 mg/L*H, 500 mg/L*H with intermittent infusions Q6H, and 630 mg/L*H with intermittent infusions Q12H. With the same trough target, Q12H dosing is associated with much higher AUC24 and peak serum levels, increasing the risk of acute kidney injury (AKI).

Another study by Pai et al. 2014 demonstrated the relationship between trough levels and AUC.6. Trough concentrations correlate with significant inter- and intra-subject variability in  AUC24. Although the vancomycin 15-20 mg/L trough target concentration ensures clinical efficacy of treatment, it does not guarantee optimal AUC/MIC in patients with MRSA infections and may lead to AUC values associated with an increased risk of AKI.

Figure 3: The vancomycin trough target concentrations of 15-20 mg/L are associated with an AUC of ≥ 400 mg*h/L and ensures efficacy essentially all of the time. However, the Monte Carlo simulation by Pat et al.2014 demonstrated that inter-patient subject variability of a trough-based target range of 15-20 mg/L can lead to levels associated with acute kidney injury (AKI) and vancomycin-induced nephrotoxicity (VIN) due to considerable variability in the upper ranges of AUC/MIC.

• Trough concentration, when maintained above 15 to 20 mg/L, is associated with increased risk of VIN. The AUC-guided vancomycin monitoring strategy reduces the occurrence of AKI when comparing trough-guided monitoring8.

Figure 4: Forest plot indicating that the risk of VIN associated with AUC-guided monitoring was significantly lower than trough-based monitoring (OR=0.68, [95% CI 0.46-0.99).

How Does PrecisePK Apply Bayesian Algorithm to Compute Vancomycin Dosing?

• Vancomycin dosing guidelines mention over 20 different population models and provide definitive dosing recommendations for three main subgroups1:

• Adult and obese population
• Pediatric population
• Neonate population

• PrecisePK takes these unique population characteristics (age, weight, kidney) and extra drug factors (critically ill, burn, etc.) into consideration when building population pharmacokinetic model. The very first step PrecisePK takes is auto-selecting the most appropriate clinically validated population PK model based  on the patients’ characteristics, also known as the Bayesian priori. For example, critically ill adults have a very large volume of distribution and supranormal drug clearance compared to non-critically ill populations.9,10
• When combined with the individual patient’s measured drug level, Bayesian posteriori pharmacokinetic parameters are estimated. PrecisePK offers an individual recommendation after just one serum concentration input. In a robust Bayesian program like PrecisePK, the time of drug level acquisition is not necessary to obtain a prediction of individual pharmacokinetics, thus giving a very precise and specific dosing recommendation or AUC prediction.

Why is PrecisePK Capable of Predicting the Most Accurate AUC Among Its Competitors?

• PrecisePK has been used in clinical setting for over 35 years and has helped calculate dose for over a million patients since its inception. The reason for the software’s precision in estimating the Vancomycin AUC, as observed by Turner et. al3 is due to its:

• Detailed and accurate implementation of the two compartment PK models for critically ill population.
• Robust and effective Bayesian algorithm for identifying the individual pharmacokinetics using just one drug level. PrecisePK can identify the exact shape of the serum level graph, not just the terminal slope. This enables an accurate estimation of AUC since both the shape of the graph and the terminal slope determine AUC .
• Vancomycin pharmacokinetics is best described by a two-compartment model. The serum level graph associated with this model has a complex decay pattern that accounts for concomitant drug distribution into the peripheral compartment and elimination from the central compartment7. Targeting only trough levels runs the risk of ignoring all the intricacies of the shape of the graph and the fact that it involves a dual exponential decay curve. Due to this difference in the shape, the same trough target may result in vastly different drug exposure in different populations.