Therapeutic Vancomycin Monitoring Using AUC

During our 30 minute webinar on vancomycin monitoring, we received a number of questions from attendees. We answered most of these during the session, but still had a few left to cover. Sanford Guide editor Douglas Black PharmD prepared answers to these questions.

 

Q. Does the 400-600 µg x hr/mL target apply to all serious MRSA infections, or is a different target recommended for endocarditis, osteomyelitis, meningitis, etc.?

A. At present, the data allow us only to identify 400-600 µg x hr/mL as the appropriate AUC24 target for serious MRSA infection. There is no compelling evidence to support a different target for a specific MRSA infection. We are in need of more data supporting AUC monitoring for meningitis (since S. pneumoniae is the likely pathogen in adults, not S. aureus).

 

Q. What about the use of AUC monitoring for other infections?

A. The revised guidelines are based on improved efficacy and safety data for AUC24/MIC monitoring for serious MRSA infections. The guidelines exclude nonserious infections such as urinary tract infections and skin and soft tissue infections. In addition, they do not apply to other pathogens such as methicillin-susceptible S. aureus, coagulase-negative Staphylococcus, or streptococci (we need data on the appropriate AUC24 target for these). The guidelines also do not apply to MRSA infections for which the vancomycin MIC is known to be greater than 1 µg/mL by broth microdilution.

 

Q. When doing AUC monitoring, why is it recommended to assume a vancomycin MIC of 1 µg/mL?

A. Vancomycin MIC depends on the method used to measure it. For example, the vancomycin MIC determined by E-test methodology tends to be 1.5-2x higher than the MIC measured by broth microdilution (BMD). There is known disagreement between commercially available MIC automated testing methods, E-test, and BMD, complicated by ±1 doubling dilution variability in results. We also know that at most institutions, the vancomycin MIC by BMD is 1 µg/mL or less (and there hasn’t been much MIC creep). Therefore, for monitoring purposes the vancomycin MIC for MRSA should be assumed to be 1 µg/mL (unless it is known to be something different using BMD).

 

Q. Can we use AUC monitoring during AKI?

A. We believe the AUC target monitoring range (400-600 µg x hr/mL) also applies to patients with renal dysfunction, although outcome studies validating this target for patients receiving hemodialysis, a hybrid hemodialysis therapy such as SLED, or CRRT are not available. Patients with unstable renal function are certainly going to be a challenge. We need more published data to help us with AUC monitoring in renal dysfunction; for now, trough-based dosing or dosing by concentration in some patients might be more practical. For example, the guideline recommendation for patients on hemodialysis is to monitor by targeting a predialysis concentration of 15-20 µg/mL, which will serve as a reasonable proxy for an AUC of 400-600 µg x hr/mL.

 

Q. Does the timing of the peak blood draw matter significantly with the extrapolation technology in the calculator?

A. The timing of the peak blood draw doesn’t matter as long as 1) it is postdistributional, and 2) the elapsed time from the start of infusion to the peak blood draw is known. The calculator is able to back-extrapolate to the “true” peak as long as these conditions are met.

 

Q. How often do you redraw peak and troughs for the calculator?

A. I would suggest the same thought process we would use if monitoring with troughs. In a stable patient, weekly monitoring of AUC is reasonable. You might determine AUC (or trough) more frequently in an unstable patient, but no monitoring approach works particularly well in unstable patients.

 

Q. Can the peak be calculated via pharmacokinetic calculations versus a blood draw?

A. It is possible to estimate the peak concentration (without measuring it) using population pharmacokinetic parameters, but using that peak in the AUC calculator will degrade its performance. The accuracy of the AUC returned by the calculator would depend on how closely the patient in question matches population PK parameters. If you are using a measured trough, it’s better to use a measured peak too.