Here we present a tandem mass spectrometry-based protocol for the quantification of frequently used antibiotics in intensive care units, namely cefepime, meropenem, ciprofloxacin, moxifloxacin, linezolid, and piperacillin.
There is an ever-increasing demand for the therapeutic drug monitoring of antibiotics in many clinical facilities, particularly with regard to the implementation of hospital antibiotic stewardship programs.
In the current work, we present a multiplex high-performance liquid chromatography-tandem mass spectrometry (HPCL-MS/MS) protocol for the quantification of cefepime, meropenem, ciprofloxacin, moxifloxacin, linezolid, and piperacillin, commonly used antibiotics in intensive care units. The method was previously comprehensively validated according to the guideline of the European Medicines Agency.
After a rapid sample cleanup, the analytes are separated on a C8 reverse-phase HPLC column within 4 minutes and quantified with the corresponding stable isotope-labeled internal standards in electrospray ionization (ESI+) mass spectrometry in multiple reaction time monitoring (MRM). The presented method uses a simple instrumentation setting with uniform chromatographic conditions, allowing for the daily and robust antibiotic therapeutic drug monitoring in clinical laboratories. The calibration curve spans the pharmacokinetic concentration range, thereby including antibiotic amounts close to the minimal inhibitory concentration (MIC) of susceptible bacteria and peak concentrations (Cmax) that are obtained with bolus administration regimens. Without the necessity of the serum dilution before the sample cleanup, the area under the curve for an administered antibiotic can be obtained through multiple measurements.
Although antibiotics have revolutionized the practice of medicine, severe bacterial infections remain a leading cause of morbidity and mortality in critical illnesses1. In this regard, the prompt administration of a suitable anti-infective in an adequate dosage is of the uppermost importance for disease control2.
A growing body of evidence demonstrates that the empirical treatment with broad-spectrum antibiotics is becoming increasingly problematic with the complexity of patient populations. This is especially true for intensive care units (ICU), where a tremendous inter-individual variability of key pharmacokinetic (PK) parameters is frequently observed3,4. Accordingly, ICU patients are at imminent risk of sub-therapeutic levels with the danger of an insufficient therapeutic success5,6. Then again, patients are unnecessarily exposed to excessively high antibiotic concentrations that may result in serious adverse events with no clinical benefits7. Both the antibiotic misuse and the insufficient dosing have also fueled the dissemination of antibiotic resistance, which is becoming an ever-growing threat to public health8.
To improve the use of antibiotics and to preserve their effectivenessas long as possible, the World Health Organization has launched a global action plan on antimicrobial resistance in 20159. Antibiotic stewardship programs constitute an essential cornerstone of prudent antimicrobial use in national public health strategies10, helping clinicians to improve the quality of patient care11 and, at the same time, significantly reducing the antibiotic resistance12. Antimicrobial dosing in individual patients through the application of therapeutic drug monitoring (TDM) is a key instrument in this context13.
To date, commercially available TDM assays are only available for the glycopeptide antibiotics and aminoglycosides. The quantification of substances from other classes commonly requires an in-house method development or validation that can be cumbersome. We, therefore, present in detail the protocol for a robust mass spectrometry-based assay that can be used for the quantification of the most relevant antibiotics in ICU within their clinical relevant concentration ranges14. The method was recently established in our mass spectrometry facility and has been applied for the routine TDM in ICU since then. The procedure uses a straightforward and simple analytical setting with a uniform sample cleanup, allowing for the rapid implementation of antibiotic TDM in many facilities with mass spectrometry capabilities.
The protocol described here was optimized for the quantification of cefepime, meropenem, ciprofloxacin, moxifloxacin, linezolid, and piperacillin in human serum, using isotope dilution liquid chromatography (LC) in combination with a tandem mass spectrometry (MS/MS). For the isotope dilution LC-MS/MS methodology, stable isotope-labeled compounds are added to a sample of interest with a specific matrix (e.g., serum). Isotope-labeled standards can be distinguished from their unlabeled counterpart, namely the analyte of interest, due to different molecular weights of the natural molecule and their fragmentation products, termed a parent-ion-to-daughter-ion transition. As isotope-labeled compounds have an almost identical overall physicochemical behavior compared to their unlabeled counterpart, they are ideal internal standards for the MS/MS, allowing a nearly matrix-independent analyte quantification with a high degree of accuracy15. Nowadays, many stable isotope-labeled internal standards that can be used for small-molecule quantification, including the TDM of antimicrobials, are commercially available.
The chromatographic separation of the antibiotic analytes in the described protocol is performed with an analytical C8 alkyl-chain-length reverse-phase column (100 mm x 2.1 mm, 3 µm particle-size). During the method development, the internal standard normalized matrix factors for all analytes was between 94.6% and 105.4%, with a coefficient of variation of ≤8.3%14.
NOTE: It is recommended to work in a fume hood when handling organic solvent, such as methanol. Prepare all buffers and mobile phases in volumetric flasks. If not otherwise specified, the solutions can be stored at room temperature for up to 1 month after preparation.
1. Preparation of the Calibrators and Quality Control Samples
NOTE: A corresponding data analysis sheet for the preparation of stock and spike solutions is given in the Supplemental File. For reasons of traceability, insert the manufacturer, catalog number, and a lot number of each antibiotic in the corresponding columns. Dissolve all antibiotics in a cold storage at 4 °C and keep the working time as short as possible.
2. Preparation of the Internal Standards Mix
NOTE: Internal standards are isotope-labeled counterparts of the analytes of interest that are added to a sample during sample cleanup. As the internal standards have almost identical overall physicochemical properties to their unlabeled counterparts, they compensate for the matrix effects of a given sample.
3. Patient Sample Storage
NOTE: Ensure that the serum is obtained as fast as possible and that the cold chain of frozen samples is maintained.
4. Buffer Preparation for Chromatography
5. Instrument Tuning
NOTE: This step is performed for the set-up of the method on a specific mass spectrometer.
6. HPLC-MS/MS Set-up
NOTE: Features of the mass spectrometer, HPLC system (including the autosampler), and the corresponding software depend on the manufacturer. Adapt the mass spectrometer parameters and the wash procedure according to the manufacturer’s recommendations.
7. Sample Measurement Master File
NOTE: With the 'sample measurement master file', the patient samples are specified, the HPLC-MS/MS analysis is started, and the data evaluation is performed. Two separate template files including a low- and high-quality control pair are generated; one template includes QC pair A and C, the other one QC pair B and D.
8. Sample Cleanup and HPLC-MS/MS Analysis
NOTE: For each sample batch, a paired quality control set with a low and high antibiotic concentration (QC A/C or QC B/D) is processed and analyzed. Between different batches, the paired QC samples are used in an alternate sequence (e.g., on day 1, select the 'sample measurement master file' including QC pair A/C; on day 2, select the one including QC pair B/D. The processing of the serum samples is illustrated in Figure 1.
Figure 1: Schematic representation of the sample cleanup. Protein precipitation at the high centrifugal force gives a dense pellet and clear supernatant, indicating that protein precipitation was complete. The entire processing time is approximately 30 min, including the sample cleanup, the chromatographic separation, and the MS/MS analysis. Please click here to view a larger version of this figure.
9. Quality Assessment and Quantification
Using the described protocol, a typical chromatogram is depicted in Figure 2. According to the United States Pharmacopeia (USP) chromatography guidelines16, the column dead volume in the present system was determined with ~0.22 mL and the extra-column volume (including the injector, tubing, and connectors) with ~0.08 mL, giving a hold-up volume of ~0.30 mL. The calculated retention factors for all analytes were 2.8 (for cefepime) – 4.2 (for piperacillin).
Figure 2: Typical analytical chromatogram with normalized signal intensities. The antibiotics are eluting in the following order: cefepime (green), meropenem (brown), ciprofloxacin (red), moxifloxacin (black), linezolid (orange), and piperacillin (purple). The retention times, which are given in minutes, and the analyte peak symmetries vary, depending on the exact composition of the mobile phases, the flow-rate, the chromatography tubing, and the analytical column age. Please click here to view a larger version of this figure.
Figure 3A contains a sample chart list for the processed samples, including the calibrators 0 – 7 ("Kalibrator 0" – "Kalibrator 7"), quality controls, and patient sera, that are indicated with the injection number (#); the sample identification text (Sample Text); the measured concentration in mg/L (Conc.); the sample type that is either a blank, standard, quality control, or patient sample (Type); the nominal concentration of the calibrators in mg/L (Std. Conc); the analytical retention time (RT); the response that is the ratio of the peak area of the analyte/peak area IS (Response); the deviation from the nominal concentration value (%Dev); the vial position (Vial); and the acquisition time (Acq.Time). The key parameter used for the quantification is the Response, gradually increasing with the analyte concentration, due to the constant amount of added isotope-labeled internal standard.
Figure 3B shows the calibration curve. In regression, the coefficient of determination r2 should be > 0.995. The following calibration model is used for all analytes described in this method: curve type = linear; origin = included; weighting = 1/x; axis transformation = none. In the given example, the calibration curve and quality controls fulfill all quality criteria: r2 > 0.995 for the calibration curve and the deviation of the calibrators (including the LLOQ) and the QC samples is within ± 15% of the nominal value.
The measured parent-to-daughter ion transitions (MRM) are given in Figure 3C, showing four peaks at the same retention time: the two upper peaks depict two transitions that are measured for the analyte of interest, the lower two peaks represent the transitions for the corresponding isotope-labeled internal standard. For the quality assessment, the analyte peaks in the respective retention time windows are visually checked and manually reintegrated at the baseline, when necessary.
The minimally inhibitory concentration (MIC) is the central component of the antimicrobial TDM, defining the pharmacokinetic exposure that is required to achieve a target pharmacokinetic/pharmacodynamic (PK/KD) ratio13,17. Accordingly, the target antibiotic TDM concentration levels are expressed in relation to the MIC of the causative pathogen. Given that the action of beta-lactam antibiotics is time-dependent, their efficacy is maximized through the achievement of the therapeutic concentrations that exceed the MIC 4x -5x (fT > 4-5x MIC).When facing unknown infectious pathogens, the target trough concentration range of free (protein-unbound) piperacillin is, therefore, 64 mg/L, corresponding to approximately 90 mg/L total piperacillin18.
The first patient (sample #11) has a satisfactory high serum trough level of 83.4 mg/L piperacillin that is also sufficient for problem pathogens, such as Pseudomonas aeruginosa. The second patient (sample #12) has a concentration of approximately 0.2 mg/L, which is below the lowest calibrator (LLOQ). Perhaps the patient has recovered, and the administration of piperacillin was discontinued. The result "< 0.5 mg/L" is, therefore, reported in the hospital information system. The third patient (sample #13) has a low piperacillin trough concentration of only 5.3 mg/L that is not sufficient for the clear majority of pathogens. For effective antimicrobial chemotherapy, the dosage should be increased by the physician.
Figure 3: Exemplary quality assessment and quantification for the analyte piperacillin. These panels represent the mass spectrometry data analysis. (A) This panel shows the sample list, including the calibrators (Standard, samples #1 – #8), quality controls (QC, samples #9 and #10), and patient sera (samples #11 – #13). Calibrator 0 refers to the blank without analyte, but with the addition of an internal standard. 9951 represents QC B, 9953 represents QC D. (B) This panel shows the calibration curve for piperacillin. The percentage deviations from the nominal calibrator concentrations are given in the upper graph (y-axis: residual), the lower graph depicts the linear calibration range. (C) This panel shows the multiple reaction time monitoring (MRM) for piperacillin and the corresponding internal standard piperacillin-D5 for patient serum sample #12. Two parent-to-daughter ion transitions are presented with their retention time and respective signal intensities. Please click here to view a larger version of this figure.
Supplemental File. Please click here to download this file.
In this manuscript, we report the protocol for a simple and robust tandem mass spectrometry-based method for the quantification of frequently used antibiotics in ICU19, namely cefepime, meropenem, ciprofloxacin, moxifloxacin, linezolid, and piperacillin14. A spreadsheet accompanies the manuscript for the preparation of antibiotic stock solutions, calibrators, and quality controls, taking into account the purity of the antibiotics and the molecular weight of their counterions. Given that the concentrations of the antibiotics are rather high, their quantification should be no particular challenge from an analytical perspective. Accordingly, we are confident that this protocol is applicable to various MS instrumental platforms. For a method transfer, users are encouraged to quantify the extra-column volume and hold-up volume of their chromatographic system and to adapt the gradient start time accordingly16. During the method set-up, the system should also be evaluated for carry-over and, if necessary, a blank sample must be injected after the highest calibrator and patient samples with high antibiotic concentrations. Users must also consider the possibility of detector saturation that occurs when too many ions enter a tandem mass spectrometer. Relevant detector saturation can be eliminated with smaller injection volumes, a higher analyte dilution during the sample cleanup, and/or a detuning of a target analyte (e.g., downgrading the optimal voltage settings).
Contrary to other methods, the calibration range allows both a quantification of concentrations close to the MIC of susceptible pathogens, as well as peak concentrations (cmax) that are obtained with a bolus administration. The highest Cmax-values for adults are reported in the corresponding professional information sheets on the FDA drug safety database as follows: 163.9 mg/L for cefepime20, 112 mg/L for meropenem21, 4.6 mg/L for ciprofloxacin22, 4.1 mg/L for moxifloxacin23, 21.2 mg/L for linezolid24, and 298 mg/L for piperacillin25. Antibiotic concentration monitoring in the patient's blood circulation allows a dose adjustment to the susceptibility of the involved pathogens, but the pharmacokinetic area under the curve can also be obtained through multiple blood sampling with the given protocol.
Many antibiotics (especially beta-lactam meropenem) are chemically unstable once dissolved. The most critical step in this protocol is, therefore, the preparation of the stock solutions, calibrators, and quality controls under cold conditions26,27. In that respect, it is also essential to freeze patient samples as quickly as possible. Although serum storage at -80 °C is recommended26, our stability experiments show that samples can also be stored up to 3 days at -20 °C without any significant decrease of antibiotics concentrations (even at the trough levels).
We recommend performing a system suitability test before each HPLC-MS/MS analysis of patient samples (e.g., with calibrator 3). Generally, a system suitability test is used to verify the repeatability of the LC-MS/MS system and to see if it is also adequate for the analysis to be done. Thus, for instance, decreasing signal intensities are caused by a contamination of the MS sweep cone, which, then, requires its cleaning with an organic solvent. To keep the MS source clean, a divert valve can be introduced after the chromatography column, directing "analyte-free" portions of the mobile phase to the waste before they reach the mass spectrometer. On the other hand, an overall increase of the pressure can indicate column clogging over time. To increase the column longevity usage of a cost-effective precolumn filter is recommendable. If the pressure still continues to be a problem, a flow rate of 0.4 mL/min can also be used with the chromatographic gradient in this protocol.
A minor limitation of this technique is that it requires three separate manual steps for sample clean-up, resulting in a total turnaround time of approximately 30 min. Adding the isotope-labeled internal standards to the precipitation agent may save some processing time. However, this should only be done for high sample throughput rates and with the precipitation agent being stored in the cold (e.g., at -20 °C), as the internal standards also degrade in vitro at elevated temperatures.
The described protocol has been developed for sample processing in standard 1.5 mL polypropylene tubes. Should a higher throughput rate be required for antibiotic TDM, the procedure can be upgraded to the multi-well plate format using adequate centrifuge inserts or filter plates with a vacuum manifold.
The authors have nothing to disclose.
The authors thank Dr. Schütze for his help with establishing the presented method and Dr. Zoller for the valuable input regarding the proper calibration range. The authors also acknowledge the technical staff of the mass spectrometry facility.
cefepime hydrochloride | Sigma-Aldrich | 1097636 | USP Reference Standard |
meropenem trihydrate | Sigma-Aldrich | Y0001252 | EP Reference Standard |
ciprofloxacin | Sigma-Aldrich | 17850 | |
moxifloxacin hydrochloride | Sigma-Aldrich | SML1581 | |
linezolid | Toronto Research Chemicals | L466500 | |
piperacillin sodium salt | Sigma-Aldrich | 93129 | |
cefepime-13C12D3 sulfate | Alsachim | C1297 | Isotope labelled internal standard for cefepime |
meropenem-D6 | Toronto Research Chemicals | M225617 | Isotope labelled internal standard for meropenem |
ciprofloxacin-D8 | Toronto Research Chemicals | C482501 | Isotope labelled internal standard for ciprofloxacin |
moxifloxacin-13C1D3 hydrochloride | Toronto Research Chemicals | M745003 | Isotope labelled internal standard for moxifloxacin |
linezolid-D3 | Toronto Research Chemicals | L466502 | Isotope labelled internal standard for linezolid |
piperacillin-D5 | Toronto Research Chemicals | P479952 | Isotope labelled internal standard for piperacillin |
methanol | JT Baker | 8402 | |
HPLC grade water | JT Baker | 4218 | |
formic acid | Biosolve | 6914132 | |
acetic acid | Biosolve | 1070501 | |
ammonium formate | Sigma-Aldrich | 70221-25G-F | |
tert-Butyl methyl ether | Merck | 101845 | |
Fortis 3 μm C8 100 * 2.1 mm | Fortis | F08-020503 | |
Ti-PEEK-encased Prifilter (2 μm) | Chromsystems | 15011 | |
2795 Alliance HPLC system | Waters | 176000491 | |
Quattro micro API Tandem Quadrupole System | Waters | 720000338 | |
QuanLynx 4.1 software | Waters | / | Data evaluation software provided by the mass spectrometer manufacturer |