Emerging evidence shows the benefits of using pharmacokinetics to adjust vancomycin dosing. What follows is a guide about how to do it. I will keep this up-to-date, so please let me know if you find any computer glitches or new evidence that should be included.
BLEEDING EDGE WARNING: The Excel spreadsheet still hasn’t been validated by multiple users across various clinical scenarios. Post-publication peer review is ongoing. For now, this technique should be used with caution (e.g. corroborate results with another pharmacokinetics program).
Getting started: Load, measure, and titrate.
Use whatever loading dose you would typically use. Pharmacokinetics should then be measured immediately following the first dose, for two reasons:
- This is the point at which pharmacokinetics are easiest to perform.
- Immediate assessment of pharmacokinetics facilitates achieving ideal drug levels as soon as possible.
After a dose is given, it takes a little while for the drug to fully distribute throughout the body (~1-4 hours). Drug levels shouldn’t be measured during this time period (they will be excessively high and won’t reflect the true volume of distribution). The exact amount of time it takes for drug to distribute is debatable (different sources list anywhere between 1-4 hours). Critically ill patients may have expanded volumes of distribution and poor perfusion, so it might be ideal to provide at least four hours for the drug to distribute before checking a level. However, if you wind up checking the level between 2-4 hours that’s probably fine.
A second level should be checked ~5 hours after the first level. If the two levels are measured too close together, it may be difficult to accurately calculate the half-life. However, if you wait too long to check the second level, this will delay administration of the next dose.
The precise timing of drug level measurement isn’t important, as long as the actual time the level is obtained is accurately recorded. For example, if a drug level is ordered for 6 PM but drawn late at 7:52 PM that’s fine, as long as the level is accurately time-stamped with the 7:52 PM sample time. The mathematical equations will sort this out effortlessly.
Microsoft Excel spreadsheet
I’ve created a free Excel spreadsheet to facilitate equation-based pharmacokinetics (download here). This should run off a computer or a mobile device with a spreadsheet app.
The spreadsheet is protected to prevent being inadvertently re-programmed. If you want to edit or re-program the spreadsheet, you can modify it using the password “pulmcrit.”
The math behind this spreadsheet is explained here. This spreadsheet is similar to many equation-based strategies, with the additional benefit that it can perform dosing re-calibration in non-steady-state conditions based on only two drug levels (not three).
Entering data into the spreadsheet
An example spreadsheet is shown above. This is fairly self-explanatory, but be careful when entering times. The program is only looking at differences between times, not absolute times. Make sure that there are no 24-hour discontinuities within the sequence of times. For example, imagine the last dose was given at 18:00, then the first level was obtained at 23:30 and the second level at 2:30 the following morning. If you enter {18:00, 23:30, 2:30}, the program won’t work. This must instead be entered as {18:00, 23:30, 26:30}. Adding “24” to the final timepoint keeps everything on a continuous relative timescale.
Results can be captured using a screenshot and pasted into the electronic medical record.
Interpreting results
The program is designed to determine a vancomycin dosing schedule that will achieve an AUC24 of 500 mg*h/L, based on emerging evidence that this is the most important pharmacokinetic target. Efficacy seems to require an AUC24 above 400 mg*h/L, but nephrotoxicity increases when the AUC24 is above 600 mg*h/L. Thus, an AUC24 around 500 mg*h/L seems to be the sweet spot. Estimated trough levels are also provided, but these don’t really matter.
The glomerular filtration rate (GFR) is estimated based on the clearance rate of vancomycin.1 This may be more accurate than glomerular filtration rates calculated based on the patient’s creatinine level. For example, if the GFR calculated from pharmacokinetic modeling is substantially below the GFR estimated from the patient’s creatinine level, this could be an early sign of acute kidney injury (serum creatinine will over-estimate GFR in the early phase of acute kidney injury because it hasn’t yet increased to reach equilibrium).
Can these equations be used if there was a dose adjustment?
The above spreadsheet is designed to model a scenario where the same dose has been given at a fixed time interval. If the dose was adjusted moderately >36 hours previously, that shouldn’t make a big difference and the equations should still work fairly well. However, if a substantial adjustment was made more recently, then the spreadsheet will be unreliable. The most rigorous way to deal with a recent drug adjustment is three-point pharmacokinetic modeling (including a trough, followed by two doses during the subsequent infusion using this calculator).
Re-calibration: Measure and re-titrate
Pharmacokinetic recalibration can be repeated periodically to ensure that levels are therapeutic (e.g. perhaps on day #3 of vancomycin administration). Instead of checking only a single trough level, two levels should be measured. Two levels provide considerably more information than a single level, because this allows for full pharmacokinetic analysis.
Re-calibration is performed similarly to the initial dose calibration. The only exception here is that if the patient is on a q8hr dosing schedule then there won’t be sufficient time to check two levels and re-calibrate the dose in under eight hours. Therefore, for patients on q8hr dosing it may be best to check one level ~4 hours, then check another level at ~8 hours immediately prior to the next dose (a trough level). This strategy won’t allow for an immediate dose adjustment, but you can adjust the dose subsequently.
Related references:
- Rationale for why we should use PK monitoring explored here.
- Download the free, open-source PulmCrit vancomycin calculator here.
- Equations & math behind the vancomycin calculator explained here.
- Another online vancomycin calculator is here: https://www.vancopk.com
Image credit: NASA
Somewhere through this vanco-related information, you mentioned that this spreadsheet should be validated by other clinics. Together with the Department of Infectious Diseases and Clinical Microbiology Department at Ankara University, we are willing to do this. I am in the process to put this work together as a research proposal. I will let you know as we go along. I am very excited. Best regards,