Journal of Chromatography B, 990 (2015) 164–168
Contents lists available at ScienceDirect
Journal of Chromatography B journal homepage: www.elsevier.com/locate/chromb
A validated method for the quantification of fosfomycin in human plasma by liquid chromatography–tandem mass spectrometry Jens Martens-Lobenhoffer ∗ , Stefanie M. Bode-Böger Institute of Clinical Pharmacology, Otto-von-Guericke University, Magdeburg, Germany
a r t i c l e
i n f o
Article history: Received 21 January 2015 Accepted 29 March 2015 Available online 3 April 2015 Keywords: Fosfomycin Human plasma HILIC LC–MS/MS
a b s t r a c t Fosfomycin is a small, hydrophilic antibiotic drug with activity against Gram-positive as well as Gramnegative pathogens. It is in increasing use in intensive care units as a last line antibiotic since it shares no cross-resistance with other antibiotics. It is not metabolized and plasma levels are dependent on renal excretion rate and renal replacement therapy such as hemofiltration or hemodialysis. Measurement of fosfomycin plasma concentrations is therefore highly desirable in order to optimize dosing. We have developed a method for the quantification of fosfomycin in human plasma using HILIC chromatography on a silica stationary phase and tandem mass spectrometric detection. Sample preparation consisted only of protein precipitation without derivatization. Propylphosphonic acid was used as internal standard. Two calibration ranges from 15 to 150 g/ml and 100 to 750 g/ml were necessary to cover the whole range of plasma concentrations expected from intensive care patients. Intraday precision ranged from 4.0% to 6.4%, depending on the concentration level, with accuracies ranging from −1.1% to 11.5%. The corresponding interday precisions and accuracies were 2.0–11.0% and 0.6–7.8%, respectively. Fosfomycin was stable in human plasma under all storing conditions relevant for clinical samples. First experiences with this method in clinical routine use confirmed the applicability and ruggedness of the analytical procedure. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Fosfomycin is a broad-spectrum antibiotic drug with activity against Gram-negative as well as Gram-positive bacteria. It is a small and highly hydrophilic phosphonic acid derivative (see Fig. 1), which is not significantly metabolized and excreted mostly unchanged via the kidneys [1]. It is structurally unrelated to other antibiotic substances and shares no cross-resistance with them. Owing to the growing incidence of infections by multidrug resistant bacteria strains, therapy with fosfomycin is applied in increasing rates in intensive care units as a last line antibiotic [2]. Daily doses of up to 24 g fosfomycin divided into 3–4 intravenous applications per day were administered to counteract sensitive pathogens with MIC ≤ 32 g/ml [3,4], leading to very high maximal plasma concentrations of more than 300 g/ml. However, studies evaluating the effectiveness and safety of fosfomycin in intensive care settings remain scarce [2]. Patients in critical conditions may
∗ Corresponding author at: Otto-von-Guericke Universität, Insitut für Klinische Pharmakologie, Leipziger Str. 44, 39120 Magdeburg, Germany. Tel.: +49 391 6713067; fax: +49 391 6713062. E-mail address: [email protected] (J. Martens-Lobenhoffer). http://dx.doi.org/10.1016/j.jchromb.2015.03.029 1570-0232/© 2015 Elsevier B.V. All rights reserved.
exhibit rapidly changing renal excretion rates, making predictions about the pharmacokinetic parameters of fosfomycin very difficult. Furthermore, these patients often undergo renal replacement therapy by hemofiltration or hemodialysis, making a proper dosing estimation even more difficult [5,6]. Monitoring of the plasma concentrations of fosfomycin in intensive care unit patients can be helpful in adjusting individual dosage regimes and may increase therapeutic efficiency and prevent induction of drug resistance. Thus, a fast and rugged method for the determination of fosfomycin in human plasma optimized for this patient group is highly desirable. The extremely acidic and polar nature of fosfomycin and the lack of chromophores in the molecule make its analytical separation and detection quite difficult. Because of these properties, gas chromatographic separation is only possible after adequate derivatization, as it was described previously for conventional [7,8] or mass spectrometric detection [9]. On the other hand, capillary electrophoresis was utilized to analyze fosfomycin in its underivatized state [10], but the unselective UV-detection led to difficult to interpret chromatograms. Applying LC–MS/MS, fosfomycin can be analyzed unambiguously in its underivatized state [11,12], but the chromatographic retention was marginal despite the highly aqueous mobile phases and the cyano- or polar-modified stationary phases. A retention on a reversed phase stationary phase was
J. Martens-Lobenhoffer, S.M. Bode-Böger / J. Chromatogr. B 990 (2015) 164–168
O H3C H
P O
OH
O
OH
P
H
Fosfomycin
H3C
165
2.4. Sample preparation
OH OH
Propylphosphonic acid (I.S.)
Fig. 1. Molecular structures of fosfomycin and the I.S. propylphosphonic acid.
only possible after derivatization [13]. Yet, derivatization reactions make an analytical method tedious, time consuming and potentially error prone and are therefore less appropriate for the sample preparation of urgent clinical samples. Recently, a LC–MS/MS method applying the HILIC separation technique on a zwitterionic (sulfobetaine) modified column for the retention of underivatized fosfomycin was reported [14]. However, this method experienced some problems in terms of distorted peak shapes with long tailing and instable retention times. Thus, the aim of this study was to develop an easy and rugged LC–MS/MS method providing reasonable chromatographic retention without derivatization using a HILIC separation on a bare silica column for the precise and accurate detection of fosfomycin in plasma samples of intensive care unit patients. 2. Materials and methods 2.1. Chemicals Fosfomycin was purchased as its disodium salt (purity > 99%) from Sigma–Aldrich (Munich, Germany). The internal standard (I.S.) propylphosphonic acid (purity 99.6%) was also obtained from Sigma–Aldrich. All other chemicals were of analytical grade or better. Pooled drug free human plasma was obtained from the blood bank of the Otto-von-Guericke University (Magdeburg, Germany). 2.2. Instrumentation The HPLC part of the analytical apparatus consisted of an Agilent 1100 system (Waldbronn, Germany) comprising a binary pump, an autosampler and a thermostatted column compartment. The chromatographic separation took place on an AtlantisTM HILIC silica column with 5 m particle size and with the dimensions 150 mm × 2.1 mm (Waters, Eschborn, Germany), protected by a SecurityGuard system (Phenomenex, Aschaffenburg, Germany) equipped with a 4 mm × 2 mm silica filter insert. The analytes were detected by a Thermo Fisher Scientific (Waltham, MA, USA) TSQ Discovery Max triple quadrupole mass spectrometer, equipped with an electrospray ionization (ESI) ion source. System control and data handling were carried out by the Thermo Electron Xcalibur software, version 1.2. 2.3. Calibration and quality control samples For both fosfomycin and the I.S. propylphosphonic acid, 10 mg/ml stock solutions in water were prepared. From the fosfomycin stock solution, after appropriate dilution, calibration samples were prepared in blank human plasma. The calibration range was divided into 2 sub-ranges containing each 5 calibration levels, one covering the concentrations from 15 to 150 g/ml, the other from 100 to 750 g/ml. Batches of quality control (QC) samples were prepared at 15 g/ml (low), 150 g/ml (medium) and 750 g/ml (high) from blank human plasma. The QC samples were stored at −20 ◦ C until analysis.
Unknown human plasma samples underwent 2 distinct sample preparation procedures, differentiated by the amount of I.S. added. In both cases, to 10 l plasma 10 l of I.S. solution (20 g/ml in water for the low calibration range or 200 g/ml in water for the high calibration range) was added. Subsequently, 20 l of buffer solution (2% ammonium formate and 1% formic acid in water) and 360 l acetonitrile were added. Precipitated proteins were separated by centrifugation at 10,000 × g for 5 min and about 100 l of the clear supernatant was transferred to autosampler vials with microliter inserts. 2.5. Chromatographic separation From the prepared samples, 10 l were injected into the HPLCsystem in case of the low range calibration and 1 l in case of the high range calibration. In order to clean the autosampler tubing and switching valve from residual sample content, 40 l of a purging mixture consisting in equal parts of mobile phase A and B (see below) was injected after a runtime of 8 min. On the analytical column, fosfomycin was chromatographically separated by the HILIC methodology. Mobile phase A consisted of 0.1% ammonium formate plus 0.05% formic acid in water and mobile phase B was prepared from 90% acetonitrile containing 0.1% ammonium formate plus 0.05% formic acid and 10% mobile phase A. The stationary phase was pure silica (see Section 2.2). The mobile phase composition started at 100% B, evolved linearly to 45% A and 55% B in 4 min and was held constant until 8.5 min. The flow rate was 0.35 ml/min and the column temperature was set to 35 ◦ C. A post run equilibration time of 4 min was necessary. Under the described conditions, fosfomycin eluted at 7.2 min and the I.S. slightly earlier at 7.1 min. 2.6. Mass spectrometric analysis The mobile phase flow from the HPLC-system was directed without splitting into the ESI ion source of the mass spectrometer. The ion source was operated in the negative mode, applying an ionization voltage of −3.5 kV and a capillary temperature of 350 ◦ C, with the sheath gas and auxiliary gas (both nitrogen) set to 20 and 15 units, respectively. Under these conditions, the quasimolecular [M−H]−1 ions with the mass-charge ratios m/z = 137 for fosfomycin and m/z = 123 for the I.S. were produced. These ions underwent collision induced fragmentation using argon as collision gas at a pressure of 1.5 mTorr. For fosfomycin, the fragment ions m/z = 137 → 79 at 20 eV fragmentation energy and m/z = 137 → 63 at 14 eV fragmentation energy were produced. The former fragment ion was recorded for quantification and the latter one for qualification. For the I.S., only the fragment ion m/z = 123 → 79 was produced at 22 eV fragmentation energy. 2.7. Validation The extraction yield was tested by comparing spiked and extracted plasma samples with samples spiked after the sample extraction (n = 5). Ion suppression was examined by comparing the results of spiked and extracted plasma samples with spiked and extracted aqueous samples (n = 5). The intra- and inter-day precision and accuracy of the method was tested using the QC-samples. The QC-low and QC-high samples were used in the low and the high calibration ranges, respectively, whereas the QC-medium samples applied to both calibration ranges. For the evaluation of the stability of fosfomycin in prepared samples, QC-samples were injected in their freshly prepared state and again after standing at room temperature for 24 h in the autosampler tray. Furthermore, the
166
J. Martens-Lobenhoffer, S.M. Bode-Böger / J. Chromatogr. B 990 (2015) 164–168
stability of fosfomycin in human plasma was assessed by preparing QC-samples after standing for 24 h at room temperature and comparing them with freshly thawed ones. The long term stability was tested by calculating trends of the QC-results over a time range of 5 months. The matrix independency of the method was tested by measuring 6 individual human plasma samples spiked with fosfomycin at two concentration levels (50 and 150 g/ml) and inspecting the results for their accuracy. 3. Results and discussion 3.1. Sample preparation The sample preparation consisted only of the addition of the I.S. to the samples, buffering and protein precipitation with acetonitrile. In this way, a solution free of proteins with a composition comparable to the HPLC mobile phase at starting conditions was obtained. Therefore, no peak broadening occurred due to injection of a sample with higher elution strength or different buffer content than the mobile phase. The extraction yield was 94.7% ± 5.87%, while no significant difference from a theoretical 100% extraction yield (p = 0.152) was observed. This result was expected since fosfomycin exhibits negligible protein binding [15] and coprecipitation with the plasma proteins was not observed. Applying this simple extraction procedure, the relatively “dirty” prepared samples caused an ion suppression of 23.22% ± 4.76% at the retention time of the analytes. 3.2. Chromatography and mass spectrometry Being of very polar and hydrophilic nature, fosfomycin and its I.S. seemed to be ideally suited for hydrophilic interaction liquid chromatography (HILIC). However, in a recently published HILIC method applying a zwitterionic stationary phase [14], the fosfomycin peak appeared substantially deformed showing a more than 1 min long tailing pointing to poorly controlled secondary interactions with the stationary phase. Furthermore, when using a different matrix (urine extract instead of plasma extract) the retention times of fosfomycin and its I.S. were reduced to near the solvent front and the peaks broadened considerably, irrespective of unchanged chromatographic conditions. This erratic behavior illustrated the difficulties in developing a stable and rugged HILIC chromatographic separation. In the here described method, we used a HILIC gradient elution on a bare silica column to retain the underivatized analytes with a mobile phase containing high amounts of acetonitrile and elute them with a higher aqueous content of the mobile phase. In order to obtain tailing-free peaks, an adjusted buffering of the mobile phase was necessary to keep the diprotic organo-phosphonic acid analytes in a defined state of ionization in the mobile phase throughout the complete gradient elution. As can be seen in Fig. 2, under the described conditions nearly tailing free peaks of fosfomycin and its I.S. with stable retention times were observed. Unfortunately, the well-known property of phosphonic acids to form coordination bonds to iron atoms led to the tendency of fosfomycin and its I.S. to stick to the steel surfaces of the autosampler tubing. In order to prevent carry over effects, the autosampler had to be purged after each injection by injecting a cleaning solution consisting of equal amounts of mobile phase A and B. Applying the described conditions, fosfomycin and its I.S. eluted as sharp and symmetric peaks at 7.2 and 7.1 min, respectively. The low relative chromatographic separation of the two analytes referred to their physicochemical similarity and was not regarded as a drawback but as a requirement of an I.S. to be as similar to the analyte as possible. Separation of the two analytes was achieved by mass spectrometric means. The low surface tension
and good evaporability of the HILIC mobile phase led to favorable ionization properties in the ESI ion source of the mass spectrometer. Since the analytes were strong acids, the ionization took place in the negative mode of the ESI source, leading to single charged [M−H]−1 quasimolecular ions. No adduct ion formation or in-source fragmentation was observed under the described conditions. In the MS2 -mass spectrum, the most significant fragment ions produced by both analytes was m/z 79, referring to the abstraction of the organic side chain from the phosphonic acid moiety. In the case of fosfomycin, a less intense fragment ion m/z = 63 was observed, most probably related to a fragmentation where additionally an oxygen atom from the phosphonic acid moiety was transferred onto the organic side chain during the brakeage of the phosphorous–carbon bond. This fragment ion was recorded as qualifier ion to support the unambiguous identification of fosfomycin in the chromatograms. No such fragment ion was observed in the case of the I.S. In Fig. 2, typical chromatograms obtained with the here described method are depicted. As can be seen, no peaks of endogenous substances and no crosstalk between fosfomycin and the I.S. were observed. 3.3. Calibration The full calibration range of the method covered 15–750 g/ml in order to monitor plasma samples with Cmin as well as Cmax concentrations anticipated in patients from intensive care units. As it turns out during the method development, this large range produced a non-linear calibration function using a single I.S. concentration. Probably, this non-linearity originated from ionization competition between fosfomycin and the I.S. Therefore, the calibration range was cut into 2 slightly overlapping sub-ranges, a low calibration range from 15 to 150 g/ml and a high calibration range from 100 to 750 g/ml. In the high calibration range, a tenfold higher I.S. concentration and a tenfold lower injection volume in comparison to the low calibration range was applied. Using these sub-ranges, both calibration functions were linear with the parameters intercept = −0.09765 ± 0.03715, slope = 0.01853 ± 0.00074 and a correlation coefficient R = 0.9976 for the low calibration range and intercept = −0.00807 ± 0.02024, slope = 0.00160 ± 0.00007 and R = 0.9969 for the high calibration range. The drawback of the split calibration range was the necessity of preparing and measuring unknown samples in both calibration ranges. The lower limit of quantification was defined as the lower end of the low calibration range, i.e. 15 g/ml. The lower limit of detection was 5 g/ml at a signal-to-noise ratio of 3. 3.4. Validation Due to the dual calibration ranges, the monitoring of the precision and accuracy of the method required three QC-levels, the QC-low samples with a concentration at the lower end of the low calibration range, the QC-medium at the upper end of the low calibration range as well as near the lower end of the high calibration range and the QC-high sample at the upper end of the high calibration range. Using these three QC-sample concentrations, the intra-day and the inter-day precisions and accuracies were assessed. The results are summarized in Table 1. As can be seen, the widely accepted requirements issued in the “Guidance for industry: Bioanalytical Method Validation” by the Federal Food and Drug Administration (FDA, USA) were met. The stability of fosfomycin in plasma samples was also assessed using the QC-samples. The comparison of freshly prepared plasma samples, prepared samples injected again after 24 h on the autosampler tray, plasma samples prepared after 24 h at room temperature and plasma samples stored at −20 ◦ C over a time range of 5 month are summarized in Table 2. The only significant difference showed up in the QC-low samples, where after 24 h in
J. Martens-Lobenhoffer, S.M. Bode-Böger / J. Chromatogr. B 990 (2015) 164–168 40000
40000
A)
35000 30000
20000 15000
30000
20000 15000
5000
5000
Intensity
10000
Fosfomycin m/z = 137 → 79
30000 25000
35000
Fosfomycin m/z = 137 → 79
30000 25000
20000
20000
15000
15000
10000
10000
5000
5000
0
RT: 7.15
Propylphosphonic acid (I.S.) m/z = 123 → 79
25000
10000
35000
B)
35000 RT: 7.14
Propylphosphonic acid (I.S.) m/z = 123 → 79
25000
Intensity
167
RT: 7.27
0 3
4
6
5
7
8
3
9
4
5
6
Time (min)
8
9
8
9
70000
70000
C)
60000
D)
60000
RT: 7.10
50000
50000 RT: 7.08
40000
40000
Propylphosphonic acid (I.S.) m/z = 123 → 79
30000 20000
20000
10000
RT: 7.20
60000
40000
10000
60000
Fosfomycin m/z = 137 → 79
50000
Propylphosphonic acid (I.S.) m/z = 123 → 79
30000
Intensity
Intensity
7
Time (min)
Fosfomycin m/z = 137 → 79
50000 40000
30000
30000
20000
20000
10000
RT: 7.22
10000 RT: 4.52 0
0 3
4
5
6
7
8
9
4
3
5
6
7
Time (min)
Time (min)
Fig. 2. Representative LC–MS/MS chromatograms: panel (A) plasma blank in the low calibration range; panel (B) quality control level low (15 g/ml) in the low calibration range; panel (C) quality control level high (750 g/ml) in the high calibration range; panel (D) patient sample (461 g/ml) in the high calibration range.
Table 1 Precision and accuracy. Sample
QC-low QC-medium QC-medium QC-high a
Calibration range
Low Low High High
RSD: relative standard deviation.
Concentration (g/ml)
15 150 150 750
Intra-day (n = 10)
Inter-day (n = 5) a
Mean (g/ml)
RSD (%)
Accuracy (%)
Mean (g/ml)
RSD (%)
Accuracy (%)
16.6 167.2 148.4 748.3
6.0 6.4 5.3 4.0
10.9 11.5 −1.1 −0.2
15.5 150.9 161.7 796.9
2.0 11.0 5.2 5.7
3.3 0.6 7.8 6.3
168
J. Martens-Lobenhoffer, S.M. Bode-Böger / J. Chromatogr. B 990 (2015) 164–168
Table 2 Stability of fosfomycin. Sample
QC-low QC-medium QC-medium QC-high a b
Calibration range
Low Low High High
Concentration (g/ml)
15 150 150 750
Fresh plasma sample
24 h in autosampler at 23 ◦ C
Plasma 24 h at 23 ◦ C
Plasma at −20 ◦ C
Mean ± SDa (g/ml)
Mean ± SD (g/ml)
Mean ± SD (g/ml)
Difference (%)
Mean ± SD (g/ml)
−11.3 −1.5 −3.0 1.0
14.1 147.5 141.2 701.4
14.8 141.6 139.0 710.7
± ± ± ±
0.7 7.9 13.1 25.5
14.1 140.4 141.8 691.8
± ± ± ±
0.4 5.9 13.7 25.7
Difference (%) −5.2 −0.8 2.0 −2.7
13.2 139.5 134.9 717.8
± ± ± ±
0.5 13.2 17.0 37.2
± ± ± ±
1.0 15.2 11.3 77.7
ANOVAb
Difference (%) −4.5 −4.0 1.6 −1.3
p-value 0.030 0.531 0.799 0.903
SD: standard deviation. ANOVA: analysis of variance.
plasma at room temperature a decay of 11.26% was observed. In all other QC-sample measurements the differences between the storage conditions were insignificant. Furthermore, in the QC-samples stored at −20 ◦ C over a time range of 5 months, no significant trend toward higher or lower concentration results during this time period was observed. Thus, sufficient stability of fosfomycin in human plasma in all relevant storing conditions was confirmed. The matrix independency of the method was investigated by analyzing 6 different human plasma samples spiked with 50 g/ml for the low calibration range and 300 g/ml for the high calibration range. The resulting measurements were 53.2 ± 1.7 g/ml and 305.9 ± 17.9 g/ml. None of the plasma samples showed a deviation > 10% from the expected result. Thus, the method can be considered as matrix independent with regard to individual human plasma samples. 3.5. Clinical application of the method First experiences with this method were gathered with samples from patients routinely treated with intravenous fosfomycin to counteract severe infections. No interferences from endogenous substances or from the various co-administrated drugs in these samples were observed. All concentrations found in these samples fitted in the calibrated and validated concentration ranges. However, it was impossible to estimate beforehand in which of the two calibration ranges the concentration of unknown samples would fit. Therefore, unknown samples always had to be prepared and analyzed in both calibration ranges to receive timely and correct results. 4. Conclusion The here described method for the quantification of fosfomycin in human plasma was easy, rugged and precise. These characteristics were achieved by HILIC chromatography with tandem mass spectrometric detection and an I.S. resembling closely the physico-chemical properties of the analyte. The calibration range was optimized for the requirements of patients receiving high intravenous doses of fosfomycin. These features made this method superior to earlier methods which rely on tedious, time consuming and error prone derivatization of the analytes or based on unspecific detection systems like UV-absorption resulting in ambiguous selectivity and quantification. Methods using LC–MS/MS are inherently of much better selectivity, but it proved difficult to retain underivatized fosfomycin on reversed phase stationary phases. On the other hand, a recent approach to separate fosfomycin applying the HILIC methodology showed difficulties with peak shape and retention time stability. Our method applied a HILIC gradient elution and overcame in this way such problems. Early results obtained with this method have proven its applicability in clinical practice. It is therefore a valuable tool for the therapeutic drug monitoring (TDM) of patients receiving fosfomycin intravenously for the treatment of severe infections.
Acknowledgement The development and validation of the method was financially supported by InfectoPharm Arzneimittel und Consilium GmbH, Heppenheim, Germany. References [1] S.S. Patel, J.A. Balfour, H.M. Bryson, Fosfomycin tromethamine. A review of its antibacterial activity, pharmacokinetic properties and therapeutic efficacy as a single-dose oral treatment for acute uncomplicated lower urinary tract infections, Drugs 53 (1997) 637–656. [2] S. Parker, J. Lipman, D. Koulenti, G. Dimopoulos, J.A. Roberts, What is the relevance of fosfomycin pharmacokinetics in the treatment of serious infections in critically ill patients? A systematic review, Int. J. Antimicrob. Agents 42 (2013) 289–293. [3] M.E. Falagas, K.P. Giannopoulou, G.N. Kokolakis, P.I. Rafailidis, Fosfomycin: use beyond urinary tract and gastrointestinal infections, Clin. Infect. Dis. 46 (2008) 1069–1077. [4] K. Pontikis, I. Karaiskos, S. Bastani, G. Dimopoulos, M. Kalogirou, M. Katsiari, A. Oikonomou, G. Poulakou, E. Roilides, H. Giamarellou, Outcomes of critically ill intensive care unit patients treated with fosfomycin for infections due to pandrug-resistant and extensively drug-resistant carbapenemaseproducing Gram-negative bacteria, Int. J. Antimicrob. Agents 43 (2014) 52–59. [5] J.L. Bouchet, C. Quentin, H. Albin, G. Vincon, J. Guillon, P. Martin-Dupont, Pharmacokinetics of fosfomycin in hemodialyzed patients, Clin. Nephrol. 23 (1985) 218–221. [6] R. Gattringer, B. Meyer, G. Heinz, C. Guttmann, M. Zeitlinger, C. Joukhadar, P. Dittrich, F. Thalhammer, Single-dose pharmacokinetics of fosfomycin during continuous venovenous haemofiltration, J. Antimicrob. Chemother. 58 (2006) 367–371. [7] M. Dios-Vieitez, M. Goni, M. Renedo, D. Fos, Determination of fosfomycin in human urine by capillary gas chromatography: application to clinical pharmacokinetic studies, Chromatographia 43 (1996) 293–295. [8] A. Loste, E. Hernandez, M. Bregante, M. Garcia, C. Solans, Development and validation of a gas chromatographic method for analysis of fosfomycin in chicken muscle samples, Chromatographia 56 (2002) 181–184. [9] A. Longo, M. Di Toro, E. Pagani, A. Carenzi, Simple selected ion monitoring method for determination of fosfomycin in blood and urine, J. Chromatogr. 224 (1981) 257–264. [10] D. Leveque, C. Gallion, E. Tarral, H. Monteil, F. Jehl, Determination of fosfomycin in biological fluids by capillary electrophoresis, J. Chromatogr. B 655 (1994) 320–324. [11] L. Li, X. Chen, X. Dai, H. Chen, D. Zhong, Rapid and selective liquid chromatographic/tandem mass spectrometric method for the determination of fosfomycin in human plasma, J. Chromatogr. B 856 (2007) 171–177. [12] W. Poeppl, T. Lingscheid, D. Bernitzky, O. Donath, G. Reznicek, M. Zeitlinger, H. Burgmann, Assessing pharmacokinetics of different doses of fosfomycin in laboratory rats enables adequate exposure for pharmacodynamic models, Pharmacology 93 (2014) 65–68. [13] T.A. Papakondyli, A.M. Gremilogianni, N.C. Megoulas, M.A. Koupparis, A novel derivatization method for the determination of Fosfomycin in human plasma by liquid chromatography coupled with atmospheric pressure chemical ionization mass spectrometric detection via phase transfer catalyzed derivatization, J. Chromatogr. A 1332 (2014) 1–7. [14] S.L. Parker, J. Lipman, J.A. Roberts, S.C. Wallis, A simple LC–MS/MS method using HILIC chromatography for the determination of fosfomycin in plasma and urine: application to a pilot pharmacokinetic study in humans, J. Pharm. Biomed. Anal. 105 (2015) 39–45. [15] N. Roussos, D.E. Karageorgopoulos, G. Samonis, M.E. Falagas, Clinical significance of the pharmacokinetic and pharmacodynamic characteristics of fosfomycin for the treatment of patients with systemic infections, Int. J. Antimicrob. Agents 34 (2009) 506–515.
Contents lists available at ScienceDirect
Journal of Chromatography B journal homepage: www.elsevier.com/locate/chromb
A validated method for the quantification of fosfomycin in human plasma by liquid chromatography–tandem mass spectrometry Jens Martens-Lobenhoffer ∗ , Stefanie M. Bode-Böger Institute of Clinical Pharmacology, Otto-von-Guericke University, Magdeburg, Germany
a r t i c l e
i n f o
Article history: Received 21 January 2015 Accepted 29 March 2015 Available online 3 April 2015 Keywords: Fosfomycin Human plasma HILIC LC–MS/MS
a b s t r a c t Fosfomycin is a small, hydrophilic antibiotic drug with activity against Gram-positive as well as Gramnegative pathogens. It is in increasing use in intensive care units as a last line antibiotic since it shares no cross-resistance with other antibiotics. It is not metabolized and plasma levels are dependent on renal excretion rate and renal replacement therapy such as hemofiltration or hemodialysis. Measurement of fosfomycin plasma concentrations is therefore highly desirable in order to optimize dosing. We have developed a method for the quantification of fosfomycin in human plasma using HILIC chromatography on a silica stationary phase and tandem mass spectrometric detection. Sample preparation consisted only of protein precipitation without derivatization. Propylphosphonic acid was used as internal standard. Two calibration ranges from 15 to 150 g/ml and 100 to 750 g/ml were necessary to cover the whole range of plasma concentrations expected from intensive care patients. Intraday precision ranged from 4.0% to 6.4%, depending on the concentration level, with accuracies ranging from −1.1% to 11.5%. The corresponding interday precisions and accuracies were 2.0–11.0% and 0.6–7.8%, respectively. Fosfomycin was stable in human plasma under all storing conditions relevant for clinical samples. First experiences with this method in clinical routine use confirmed the applicability and ruggedness of the analytical procedure. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Fosfomycin is a broad-spectrum antibiotic drug with activity against Gram-negative as well as Gram-positive bacteria. It is a small and highly hydrophilic phosphonic acid derivative (see Fig. 1), which is not significantly metabolized and excreted mostly unchanged via the kidneys [1]. It is structurally unrelated to other antibiotic substances and shares no cross-resistance with them. Owing to the growing incidence of infections by multidrug resistant bacteria strains, therapy with fosfomycin is applied in increasing rates in intensive care units as a last line antibiotic [2]. Daily doses of up to 24 g fosfomycin divided into 3–4 intravenous applications per day were administered to counteract sensitive pathogens with MIC ≤ 32 g/ml [3,4], leading to very high maximal plasma concentrations of more than 300 g/ml. However, studies evaluating the effectiveness and safety of fosfomycin in intensive care settings remain scarce [2]. Patients in critical conditions may
∗ Corresponding author at: Otto-von-Guericke Universität, Insitut für Klinische Pharmakologie, Leipziger Str. 44, 39120 Magdeburg, Germany. Tel.: +49 391 6713067; fax: +49 391 6713062. E-mail address: [email protected] (J. Martens-Lobenhoffer). http://dx.doi.org/10.1016/j.jchromb.2015.03.029 1570-0232/© 2015 Elsevier B.V. All rights reserved.
exhibit rapidly changing renal excretion rates, making predictions about the pharmacokinetic parameters of fosfomycin very difficult. Furthermore, these patients often undergo renal replacement therapy by hemofiltration or hemodialysis, making a proper dosing estimation even more difficult [5,6]. Monitoring of the plasma concentrations of fosfomycin in intensive care unit patients can be helpful in adjusting individual dosage regimes and may increase therapeutic efficiency and prevent induction of drug resistance. Thus, a fast and rugged method for the determination of fosfomycin in human plasma optimized for this patient group is highly desirable. The extremely acidic and polar nature of fosfomycin and the lack of chromophores in the molecule make its analytical separation and detection quite difficult. Because of these properties, gas chromatographic separation is only possible after adequate derivatization, as it was described previously for conventional [7,8] or mass spectrometric detection [9]. On the other hand, capillary electrophoresis was utilized to analyze fosfomycin in its underivatized state [10], but the unselective UV-detection led to difficult to interpret chromatograms. Applying LC–MS/MS, fosfomycin can be analyzed unambiguously in its underivatized state [11,12], but the chromatographic retention was marginal despite the highly aqueous mobile phases and the cyano- or polar-modified stationary phases. A retention on a reversed phase stationary phase was
J. Martens-Lobenhoffer, S.M. Bode-Böger / J. Chromatogr. B 990 (2015) 164–168
O H3C H
P O
OH
O
OH
P
H
Fosfomycin
H3C
165
2.4. Sample preparation
OH OH
Propylphosphonic acid (I.S.)
Fig. 1. Molecular structures of fosfomycin and the I.S. propylphosphonic acid.
only possible after derivatization [13]. Yet, derivatization reactions make an analytical method tedious, time consuming and potentially error prone and are therefore less appropriate for the sample preparation of urgent clinical samples. Recently, a LC–MS/MS method applying the HILIC separation technique on a zwitterionic (sulfobetaine) modified column for the retention of underivatized fosfomycin was reported [14]. However, this method experienced some problems in terms of distorted peak shapes with long tailing and instable retention times. Thus, the aim of this study was to develop an easy and rugged LC–MS/MS method providing reasonable chromatographic retention without derivatization using a HILIC separation on a bare silica column for the precise and accurate detection of fosfomycin in plasma samples of intensive care unit patients. 2. Materials and methods 2.1. Chemicals Fosfomycin was purchased as its disodium salt (purity > 99%) from Sigma–Aldrich (Munich, Germany). The internal standard (I.S.) propylphosphonic acid (purity 99.6%) was also obtained from Sigma–Aldrich. All other chemicals were of analytical grade or better. Pooled drug free human plasma was obtained from the blood bank of the Otto-von-Guericke University (Magdeburg, Germany). 2.2. Instrumentation The HPLC part of the analytical apparatus consisted of an Agilent 1100 system (Waldbronn, Germany) comprising a binary pump, an autosampler and a thermostatted column compartment. The chromatographic separation took place on an AtlantisTM HILIC silica column with 5 m particle size and with the dimensions 150 mm × 2.1 mm (Waters, Eschborn, Germany), protected by a SecurityGuard system (Phenomenex, Aschaffenburg, Germany) equipped with a 4 mm × 2 mm silica filter insert. The analytes were detected by a Thermo Fisher Scientific (Waltham, MA, USA) TSQ Discovery Max triple quadrupole mass spectrometer, equipped with an electrospray ionization (ESI) ion source. System control and data handling were carried out by the Thermo Electron Xcalibur software, version 1.2. 2.3. Calibration and quality control samples For both fosfomycin and the I.S. propylphosphonic acid, 10 mg/ml stock solutions in water were prepared. From the fosfomycin stock solution, after appropriate dilution, calibration samples were prepared in blank human plasma. The calibration range was divided into 2 sub-ranges containing each 5 calibration levels, one covering the concentrations from 15 to 150 g/ml, the other from 100 to 750 g/ml. Batches of quality control (QC) samples were prepared at 15 g/ml (low), 150 g/ml (medium) and 750 g/ml (high) from blank human plasma. The QC samples were stored at −20 ◦ C until analysis.
Unknown human plasma samples underwent 2 distinct sample preparation procedures, differentiated by the amount of I.S. added. In both cases, to 10 l plasma 10 l of I.S. solution (20 g/ml in water for the low calibration range or 200 g/ml in water for the high calibration range) was added. Subsequently, 20 l of buffer solution (2% ammonium formate and 1% formic acid in water) and 360 l acetonitrile were added. Precipitated proteins were separated by centrifugation at 10,000 × g for 5 min and about 100 l of the clear supernatant was transferred to autosampler vials with microliter inserts. 2.5. Chromatographic separation From the prepared samples, 10 l were injected into the HPLCsystem in case of the low range calibration and 1 l in case of the high range calibration. In order to clean the autosampler tubing and switching valve from residual sample content, 40 l of a purging mixture consisting in equal parts of mobile phase A and B (see below) was injected after a runtime of 8 min. On the analytical column, fosfomycin was chromatographically separated by the HILIC methodology. Mobile phase A consisted of 0.1% ammonium formate plus 0.05% formic acid in water and mobile phase B was prepared from 90% acetonitrile containing 0.1% ammonium formate plus 0.05% formic acid and 10% mobile phase A. The stationary phase was pure silica (see Section 2.2). The mobile phase composition started at 100% B, evolved linearly to 45% A and 55% B in 4 min and was held constant until 8.5 min. The flow rate was 0.35 ml/min and the column temperature was set to 35 ◦ C. A post run equilibration time of 4 min was necessary. Under the described conditions, fosfomycin eluted at 7.2 min and the I.S. slightly earlier at 7.1 min. 2.6. Mass spectrometric analysis The mobile phase flow from the HPLC-system was directed without splitting into the ESI ion source of the mass spectrometer. The ion source was operated in the negative mode, applying an ionization voltage of −3.5 kV and a capillary temperature of 350 ◦ C, with the sheath gas and auxiliary gas (both nitrogen) set to 20 and 15 units, respectively. Under these conditions, the quasimolecular [M−H]−1 ions with the mass-charge ratios m/z = 137 for fosfomycin and m/z = 123 for the I.S. were produced. These ions underwent collision induced fragmentation using argon as collision gas at a pressure of 1.5 mTorr. For fosfomycin, the fragment ions m/z = 137 → 79 at 20 eV fragmentation energy and m/z = 137 → 63 at 14 eV fragmentation energy were produced. The former fragment ion was recorded for quantification and the latter one for qualification. For the I.S., only the fragment ion m/z = 123 → 79 was produced at 22 eV fragmentation energy. 2.7. Validation The extraction yield was tested by comparing spiked and extracted plasma samples with samples spiked after the sample extraction (n = 5). Ion suppression was examined by comparing the results of spiked and extracted plasma samples with spiked and extracted aqueous samples (n = 5). The intra- and inter-day precision and accuracy of the method was tested using the QC-samples. The QC-low and QC-high samples were used in the low and the high calibration ranges, respectively, whereas the QC-medium samples applied to both calibration ranges. For the evaluation of the stability of fosfomycin in prepared samples, QC-samples were injected in their freshly prepared state and again after standing at room temperature for 24 h in the autosampler tray. Furthermore, the
166
J. Martens-Lobenhoffer, S.M. Bode-Böger / J. Chromatogr. B 990 (2015) 164–168
stability of fosfomycin in human plasma was assessed by preparing QC-samples after standing for 24 h at room temperature and comparing them with freshly thawed ones. The long term stability was tested by calculating trends of the QC-results over a time range of 5 months. The matrix independency of the method was tested by measuring 6 individual human plasma samples spiked with fosfomycin at two concentration levels (50 and 150 g/ml) and inspecting the results for their accuracy. 3. Results and discussion 3.1. Sample preparation The sample preparation consisted only of the addition of the I.S. to the samples, buffering and protein precipitation with acetonitrile. In this way, a solution free of proteins with a composition comparable to the HPLC mobile phase at starting conditions was obtained. Therefore, no peak broadening occurred due to injection of a sample with higher elution strength or different buffer content than the mobile phase. The extraction yield was 94.7% ± 5.87%, while no significant difference from a theoretical 100% extraction yield (p = 0.152) was observed. This result was expected since fosfomycin exhibits negligible protein binding [15] and coprecipitation with the plasma proteins was not observed. Applying this simple extraction procedure, the relatively “dirty” prepared samples caused an ion suppression of 23.22% ± 4.76% at the retention time of the analytes. 3.2. Chromatography and mass spectrometry Being of very polar and hydrophilic nature, fosfomycin and its I.S. seemed to be ideally suited for hydrophilic interaction liquid chromatography (HILIC). However, in a recently published HILIC method applying a zwitterionic stationary phase [14], the fosfomycin peak appeared substantially deformed showing a more than 1 min long tailing pointing to poorly controlled secondary interactions with the stationary phase. Furthermore, when using a different matrix (urine extract instead of plasma extract) the retention times of fosfomycin and its I.S. were reduced to near the solvent front and the peaks broadened considerably, irrespective of unchanged chromatographic conditions. This erratic behavior illustrated the difficulties in developing a stable and rugged HILIC chromatographic separation. In the here described method, we used a HILIC gradient elution on a bare silica column to retain the underivatized analytes with a mobile phase containing high amounts of acetonitrile and elute them with a higher aqueous content of the mobile phase. In order to obtain tailing-free peaks, an adjusted buffering of the mobile phase was necessary to keep the diprotic organo-phosphonic acid analytes in a defined state of ionization in the mobile phase throughout the complete gradient elution. As can be seen in Fig. 2, under the described conditions nearly tailing free peaks of fosfomycin and its I.S. with stable retention times were observed. Unfortunately, the well-known property of phosphonic acids to form coordination bonds to iron atoms led to the tendency of fosfomycin and its I.S. to stick to the steel surfaces of the autosampler tubing. In order to prevent carry over effects, the autosampler had to be purged after each injection by injecting a cleaning solution consisting of equal amounts of mobile phase A and B. Applying the described conditions, fosfomycin and its I.S. eluted as sharp and symmetric peaks at 7.2 and 7.1 min, respectively. The low relative chromatographic separation of the two analytes referred to their physicochemical similarity and was not regarded as a drawback but as a requirement of an I.S. to be as similar to the analyte as possible. Separation of the two analytes was achieved by mass spectrometric means. The low surface tension
and good evaporability of the HILIC mobile phase led to favorable ionization properties in the ESI ion source of the mass spectrometer. Since the analytes were strong acids, the ionization took place in the negative mode of the ESI source, leading to single charged [M−H]−1 quasimolecular ions. No adduct ion formation or in-source fragmentation was observed under the described conditions. In the MS2 -mass spectrum, the most significant fragment ions produced by both analytes was m/z 79, referring to the abstraction of the organic side chain from the phosphonic acid moiety. In the case of fosfomycin, a less intense fragment ion m/z = 63 was observed, most probably related to a fragmentation where additionally an oxygen atom from the phosphonic acid moiety was transferred onto the organic side chain during the brakeage of the phosphorous–carbon bond. This fragment ion was recorded as qualifier ion to support the unambiguous identification of fosfomycin in the chromatograms. No such fragment ion was observed in the case of the I.S. In Fig. 2, typical chromatograms obtained with the here described method are depicted. As can be seen, no peaks of endogenous substances and no crosstalk between fosfomycin and the I.S. were observed. 3.3. Calibration The full calibration range of the method covered 15–750 g/ml in order to monitor plasma samples with Cmin as well as Cmax concentrations anticipated in patients from intensive care units. As it turns out during the method development, this large range produced a non-linear calibration function using a single I.S. concentration. Probably, this non-linearity originated from ionization competition between fosfomycin and the I.S. Therefore, the calibration range was cut into 2 slightly overlapping sub-ranges, a low calibration range from 15 to 150 g/ml and a high calibration range from 100 to 750 g/ml. In the high calibration range, a tenfold higher I.S. concentration and a tenfold lower injection volume in comparison to the low calibration range was applied. Using these sub-ranges, both calibration functions were linear with the parameters intercept = −0.09765 ± 0.03715, slope = 0.01853 ± 0.00074 and a correlation coefficient R = 0.9976 for the low calibration range and intercept = −0.00807 ± 0.02024, slope = 0.00160 ± 0.00007 and R = 0.9969 for the high calibration range. The drawback of the split calibration range was the necessity of preparing and measuring unknown samples in both calibration ranges. The lower limit of quantification was defined as the lower end of the low calibration range, i.e. 15 g/ml. The lower limit of detection was 5 g/ml at a signal-to-noise ratio of 3. 3.4. Validation Due to the dual calibration ranges, the monitoring of the precision and accuracy of the method required three QC-levels, the QC-low samples with a concentration at the lower end of the low calibration range, the QC-medium at the upper end of the low calibration range as well as near the lower end of the high calibration range and the QC-high sample at the upper end of the high calibration range. Using these three QC-sample concentrations, the intra-day and the inter-day precisions and accuracies were assessed. The results are summarized in Table 1. As can be seen, the widely accepted requirements issued in the “Guidance for industry: Bioanalytical Method Validation” by the Federal Food and Drug Administration (FDA, USA) were met. The stability of fosfomycin in plasma samples was also assessed using the QC-samples. The comparison of freshly prepared plasma samples, prepared samples injected again after 24 h on the autosampler tray, plasma samples prepared after 24 h at room temperature and plasma samples stored at −20 ◦ C over a time range of 5 month are summarized in Table 2. The only significant difference showed up in the QC-low samples, where after 24 h in
J. Martens-Lobenhoffer, S.M. Bode-Böger / J. Chromatogr. B 990 (2015) 164–168 40000
40000
A)
35000 30000
20000 15000
30000
20000 15000
5000
5000
Intensity
10000
Fosfomycin m/z = 137 → 79
30000 25000
35000
Fosfomycin m/z = 137 → 79
30000 25000
20000
20000
15000
15000
10000
10000
5000
5000
0
RT: 7.15
Propylphosphonic acid (I.S.) m/z = 123 → 79
25000
10000
35000
B)
35000 RT: 7.14
Propylphosphonic acid (I.S.) m/z = 123 → 79
25000
Intensity
167
RT: 7.27
0 3
4
6
5
7
8
3
9
4
5
6
Time (min)
8
9
8
9
70000
70000
C)
60000
D)
60000
RT: 7.10
50000
50000 RT: 7.08
40000
40000
Propylphosphonic acid (I.S.) m/z = 123 → 79
30000 20000
20000
10000
RT: 7.20
60000
40000
10000
60000
Fosfomycin m/z = 137 → 79
50000
Propylphosphonic acid (I.S.) m/z = 123 → 79
30000
Intensity
Intensity
7
Time (min)
Fosfomycin m/z = 137 → 79
50000 40000
30000
30000
20000
20000
10000
RT: 7.22
10000 RT: 4.52 0
0 3
4
5
6
7
8
9
4
3
5
6
7
Time (min)
Time (min)
Fig. 2. Representative LC–MS/MS chromatograms: panel (A) plasma blank in the low calibration range; panel (B) quality control level low (15 g/ml) in the low calibration range; panel (C) quality control level high (750 g/ml) in the high calibration range; panel (D) patient sample (461 g/ml) in the high calibration range.
Table 1 Precision and accuracy. Sample
QC-low QC-medium QC-medium QC-high a
Calibration range
Low Low High High
RSD: relative standard deviation.
Concentration (g/ml)
15 150 150 750
Intra-day (n = 10)
Inter-day (n = 5) a
Mean (g/ml)
RSD (%)
Accuracy (%)
Mean (g/ml)
RSD (%)
Accuracy (%)
16.6 167.2 148.4 748.3
6.0 6.4 5.3 4.0
10.9 11.5 −1.1 −0.2
15.5 150.9 161.7 796.9
2.0 11.0 5.2 5.7
3.3 0.6 7.8 6.3
168
J. Martens-Lobenhoffer, S.M. Bode-Böger / J. Chromatogr. B 990 (2015) 164–168
Table 2 Stability of fosfomycin. Sample
QC-low QC-medium QC-medium QC-high a b
Calibration range
Low Low High High
Concentration (g/ml)
15 150 150 750
Fresh plasma sample
24 h in autosampler at 23 ◦ C
Plasma 24 h at 23 ◦ C
Plasma at −20 ◦ C
Mean ± SDa (g/ml)
Mean ± SD (g/ml)
Mean ± SD (g/ml)
Difference (%)
Mean ± SD (g/ml)
−11.3 −1.5 −3.0 1.0
14.1 147.5 141.2 701.4
14.8 141.6 139.0 710.7
± ± ± ±
0.7 7.9 13.1 25.5
14.1 140.4 141.8 691.8
± ± ± ±
0.4 5.9 13.7 25.7
Difference (%) −5.2 −0.8 2.0 −2.7
13.2 139.5 134.9 717.8
± ± ± ±
0.5 13.2 17.0 37.2
± ± ± ±
1.0 15.2 11.3 77.7
ANOVAb
Difference (%) −4.5 −4.0 1.6 −1.3
p-value 0.030 0.531 0.799 0.903
SD: standard deviation. ANOVA: analysis of variance.
plasma at room temperature a decay of 11.26% was observed. In all other QC-sample measurements the differences between the storage conditions were insignificant. Furthermore, in the QC-samples stored at −20 ◦ C over a time range of 5 months, no significant trend toward higher or lower concentration results during this time period was observed. Thus, sufficient stability of fosfomycin in human plasma in all relevant storing conditions was confirmed. The matrix independency of the method was investigated by analyzing 6 different human plasma samples spiked with 50 g/ml for the low calibration range and 300 g/ml for the high calibration range. The resulting measurements were 53.2 ± 1.7 g/ml and 305.9 ± 17.9 g/ml. None of the plasma samples showed a deviation > 10% from the expected result. Thus, the method can be considered as matrix independent with regard to individual human plasma samples. 3.5. Clinical application of the method First experiences with this method were gathered with samples from patients routinely treated with intravenous fosfomycin to counteract severe infections. No interferences from endogenous substances or from the various co-administrated drugs in these samples were observed. All concentrations found in these samples fitted in the calibrated and validated concentration ranges. However, it was impossible to estimate beforehand in which of the two calibration ranges the concentration of unknown samples would fit. Therefore, unknown samples always had to be prepared and analyzed in both calibration ranges to receive timely and correct results. 4. Conclusion The here described method for the quantification of fosfomycin in human plasma was easy, rugged and precise. These characteristics were achieved by HILIC chromatography with tandem mass spectrometric detection and an I.S. resembling closely the physico-chemical properties of the analyte. The calibration range was optimized for the requirements of patients receiving high intravenous doses of fosfomycin. These features made this method superior to earlier methods which rely on tedious, time consuming and error prone derivatization of the analytes or based on unspecific detection systems like UV-absorption resulting in ambiguous selectivity and quantification. Methods using LC–MS/MS are inherently of much better selectivity, but it proved difficult to retain underivatized fosfomycin on reversed phase stationary phases. On the other hand, a recent approach to separate fosfomycin applying the HILIC methodology showed difficulties with peak shape and retention time stability. Our method applied a HILIC gradient elution and overcame in this way such problems. Early results obtained with this method have proven its applicability in clinical practice. It is therefore a valuable tool for the therapeutic drug monitoring (TDM) of patients receiving fosfomycin intravenously for the treatment of severe infections.
Acknowledgement The development and validation of the method was financially supported by InfectoPharm Arzneimittel und Consilium GmbH, Heppenheim, Germany. References [1] S.S. Patel, J.A. Balfour, H.M. Bryson, Fosfomycin tromethamine. A review of its antibacterial activity, pharmacokinetic properties and therapeutic efficacy as a single-dose oral treatment for acute uncomplicated lower urinary tract infections, Drugs 53 (1997) 637–656. [2] S. Parker, J. Lipman, D. Koulenti, G. Dimopoulos, J.A. Roberts, What is the relevance of fosfomycin pharmacokinetics in the treatment of serious infections in critically ill patients? A systematic review, Int. J. Antimicrob. Agents 42 (2013) 289–293. [3] M.E. Falagas, K.P. Giannopoulou, G.N. Kokolakis, P.I. Rafailidis, Fosfomycin: use beyond urinary tract and gastrointestinal infections, Clin. Infect. Dis. 46 (2008) 1069–1077. [4] K. Pontikis, I. Karaiskos, S. Bastani, G. Dimopoulos, M. Kalogirou, M. Katsiari, A. Oikonomou, G. Poulakou, E. Roilides, H. Giamarellou, Outcomes of critically ill intensive care unit patients treated with fosfomycin for infections due to pandrug-resistant and extensively drug-resistant carbapenemaseproducing Gram-negative bacteria, Int. J. Antimicrob. Agents 43 (2014) 52–59. [5] J.L. Bouchet, C. Quentin, H. Albin, G. Vincon, J. Guillon, P. Martin-Dupont, Pharmacokinetics of fosfomycin in hemodialyzed patients, Clin. Nephrol. 23 (1985) 218–221. [6] R. Gattringer, B. Meyer, G. Heinz, C. Guttmann, M. Zeitlinger, C. Joukhadar, P. Dittrich, F. Thalhammer, Single-dose pharmacokinetics of fosfomycin during continuous venovenous haemofiltration, J. Antimicrob. Chemother. 58 (2006) 367–371. [7] M. Dios-Vieitez, M. Goni, M. Renedo, D. Fos, Determination of fosfomycin in human urine by capillary gas chromatography: application to clinical pharmacokinetic studies, Chromatographia 43 (1996) 293–295. [8] A. Loste, E. Hernandez, M. Bregante, M. Garcia, C. Solans, Development and validation of a gas chromatographic method for analysis of fosfomycin in chicken muscle samples, Chromatographia 56 (2002) 181–184. [9] A. Longo, M. Di Toro, E. Pagani, A. Carenzi, Simple selected ion monitoring method for determination of fosfomycin in blood and urine, J. Chromatogr. 224 (1981) 257–264. [10] D. Leveque, C. Gallion, E. Tarral, H. Monteil, F. Jehl, Determination of fosfomycin in biological fluids by capillary electrophoresis, J. Chromatogr. B 655 (1994) 320–324. [11] L. Li, X. Chen, X. Dai, H. Chen, D. Zhong, Rapid and selective liquid chromatographic/tandem mass spectrometric method for the determination of fosfomycin in human plasma, J. Chromatogr. B 856 (2007) 171–177. [12] W. Poeppl, T. Lingscheid, D. Bernitzky, O. Donath, G. Reznicek, M. Zeitlinger, H. Burgmann, Assessing pharmacokinetics of different doses of fosfomycin in laboratory rats enables adequate exposure for pharmacodynamic models, Pharmacology 93 (2014) 65–68. [13] T.A. Papakondyli, A.M. Gremilogianni, N.C. Megoulas, M.A. Koupparis, A novel derivatization method for the determination of Fosfomycin in human plasma by liquid chromatography coupled with atmospheric pressure chemical ionization mass spectrometric detection via phase transfer catalyzed derivatization, J. Chromatogr. A 1332 (2014) 1–7. [14] S.L. Parker, J. Lipman, J.A. Roberts, S.C. Wallis, A simple LC–MS/MS method using HILIC chromatography for the determination of fosfomycin in plasma and urine: application to a pilot pharmacokinetic study in humans, J. Pharm. Biomed. Anal. 105 (2015) 39–45. [15] N. Roussos, D.E. Karageorgopoulos, G. Samonis, M.E. Falagas, Clinical significance of the pharmacokinetic and pharmacodynamic characteristics of fosfomycin for the treatment of patients with systemic infections, Int. J. Antimicrob. Agents 34 (2009) 506–515.
