Accepted Manuscript

Title: A simple LC-MS/MS method using HILICchromatography for the determination of fosfomycin inplasma and urine: application to a pilot pharmacokinetic studyin humans

Author: Suzanne L. Parker Jeffrey Lipman Jason A. RobertsSteven C. Wallis

PII: S0731-7085(14)00578-0DOI: http://dx.doi.org/doi:10.1016/j.jpba.2014.11.042Reference: PBA 9829

To appear in: Journal of Pharmaceutical and Biomedical Analysis

Received date: 17-10-2014Revised date: 21-11-2014Accepted date: 23-11-2014

Please cite this article as: S.L. Parker, J. Lipman, J.A. Roberts, S.C. Wallis, Asimple LC-MS/MS method using HILIC chromatography for the determinationof fosfomycin in plasma and urine: application to a pilot pharmacokineticstudy in humans., Journal of Pharmaceutical and Biomedical Analysis (2014),http://dx.doi.org/10.1016/j.jpba.2014.11.042

This is a PDF file of an unedited manuscript that has been accepted for publication.As a service to our customers we are providing this early version of the manuscript.The manuscript will undergo copyediting, typesetting, and review of the resulting proofbefore it is published in its final form. Please note that during the production processerrors may be discovered which could affect the content, and all legal disclaimers thatapply to the journal pertain.

Page 1 of 23

Accep

ted

Man

uscr

ipt

ICU patients

plasma and urine

pharmacokinetic profile

HILIC and LC-MS/MS bioanalysis

Fosfomycin

*Graphical Abstract

Page 2 of 23

Accep

ted

Man

uscr

ipt

1

1

A simple LC-MS/MS method using HILIC chromatography for the 2

determination of fosfomycin in plasma and urine: application to a pilot 3

pharmacokinetic study in humans. 4

Suzanne L. Parker1,a, Jeffrey Lipmana,b, Jason A. Robertsa,b , Steven C. Wallisa 5

a Burns, Trauma and Critical Care Research Centre, The University of Queensland, 6

Brisbane, Australia 7

b Royal Brisbane and Women’s Hospital, Brisbane, Australia 8

1Author for correspondence: Tel: +6173346 5104 9

E-mail: suzanne.parker@uq.edu.au 10

Present address: School of Medicine, The University of Queensland, Herston, 11

Queensland, AUSTRALIA 12

13

14

Keywords: 15

Fosfomycin; Antibiotic; LC-MS/MS; Pharmacokinetic; HILIC 16

17

18

Page 3 of 23

Accep

ted

Man

uscr

ipt

2

19

Abstract 20

A high performance liquid chromatography - tandem mass spectrometry (LC-MS/MS) 21

method, using hydrophilic interaction liquid chromatography (HILIC) 22

chromatography for the analysis of fosfomycin in human plasma and urine, has been 23

developed and validated. The plasma method uses a simple protein precipitation 24

using a low volume sample (10 μL) and is suitable for the concentration range of 1 to 25

2000 μg/mL. The urine method involves a simple dilution of 10 μL of sample and is 26

suitable for a concentration range of 0.1 to 10 mg/mL. The plasma and urine results, 27

reported respectively, are for recovery (68, 72%), inter-assay precision (≤9.1%, 28

≤8.1%) and accuracy (range -7.2 to 3.3%, -1.9 to 1.6%), LLOQ precision (4.7%, 3.1%) 29

and accuracy (1.7% and 1.2%), and includes investigations into the linearity, stability 30

and matrix effects. The method was used in a pilot pharmacokinetic study of a 31

critically ill patient receiving IV fosfomycin, which measured a maximum and 32

minimum plasma concentration of 222 μg/mL and 172 μg/mL, respectively, after the 33

initial dose, and a maximum and minimum plasma concentration of 868 μg/mL and 34

591 μg/mL, respectively, after the fifth dose. The urine concentration was 2.03 35

mg/mL after the initial dose and 0.29 mg/mL after the fifth dose. 36

37

HIGHLIGHTS 38

A simple and robust LC-MS/MS method for the quantification of fosfomycin 39

in human plasma and urine has been developed. 40

The method uses HILIC chromatography that supports a simple treatment of 41

fosfomycin from biological fluids. 42

The developed LC–MS/MS method has been validated according to published 43

US FDA guidelines and shows excellent performance. 44

Results of a critically ill patient from a pilot pharmacokinetic study with 45

receiving IV fosfomycin are included. 46

This is the first method published that is suitable for the quantification of 47

fosfomycin in both plasma and urine. 48

Page 4 of 23

Accep

ted

Man

uscr

ipt

4

50

1. Introduction 51

Fosfomycin (FOM) is a broad-spectrum antibiotic that is generating substantial 52

interest as an intravenous or enteral therapy for the treatment of multi-drug 53

resistant (MDR) infections [1, 2]. Antibiotic resistance is a significant and immediate 54

global health concern and an increasing prevalence of MDR bacteria is steadily 55

decreasing the number of usable antibiotics available [3]. In this context, fosfomycin 56

represents an important treatment option. 57

FOM was first discovered in Spain in 1969 and is used as treatment therapy for 58

uncomplicated urinary tract infections [4] and often combined with beta-lactams or 59

aminoglycosides for a synergistic effect against Pseudomonas aeruginosa [5, 6]. FOM 60

is structurally unrelated to other antibiotics: it is a small (138 Da), highly hydrophilic 61

phosphonic acid and, with negligible protein binding [7], exhibits excellent 62

penetration into tissue [8]. FOM has a unique mechanism of action by way of its 63

ability to inhibit the synthesis of peptidoglycan found in the inner cell wall of Gram-64

negative and Gram-positive bacteria [9]. These characteristics support the 65

effectiveness of FOM for the treatment of MDR pathogens and it has been used 66

extensively as a last-line antibiotic for treatment of critically ill patients [2]. 67

Investigating the pharmacokinetic (PK)/pharmacodynamic (PD) characteristics of 68

FOM may enable development of robust dosing strategies that maximize the PD 69

response to treatment while minimizing the potential future development of 70

antibiotic resistance. A reliable method of quantification of FOM in plasma is needed 71

to define the pharmacokinetic profile of the drug and urine data can provide 72

valuable information on elimination rates. The information obtained can be used to 73

characterize the PK/PD relationship. 74

There are several analytical techniques available for the determination of FOM in 75

biological fluids, using gas chromatography [10-14], liquid chromatography (LC) - 76

spectrophotometric detection [15], LC - photometric detection [16], capillary zone 77

electrophoresis [17, 18], and, more recently, with derivatization and LC - 78

Page 5 of 23

Accep

ted

Man

uscr

ipt

5

atmospheric pressure chemical ionization mass spectrometry [19] and LC – tandem 79

mass spectrometry (MS/MS) [20-22]. 80

A method suitable for use in a pharmacokinetic study with patients receiving 81

multiple intravenous doses, up to around 12-24 g/day (as are now being used 82

clinically [2]) would require a lower limit of quantification (LLOQ) of around 1 μg/mL 83

for plasma and 0.1 mg/mL for urine. However, patients with renal insufficiency or 84

with altered pharmacokinetics – as is commonly seen in the critically ill – receiving 85

multiple doses of antibiotics can exhibit very high concentrations in their plasma. 86

This has, therefore, led to an unusually extended concentration range (from 1 to 87

2000 μg/mL for plasma) being described here, which is supported by the data from 88

the pilot study. While there are many methods currently available for the analysis of 89

FOM in biological fluids, and the method by Li offers a rapid and sensitive alternative 90

for plasma [20], we are unaware of any methods suitable for the analysis of FOM in 91

both plasma and urine that offer the concentration range to meet these clinical 92

specifications at this time. 93

This paper describes an analytical technique using hydrophilic interaction 94

chromatography (HILIC) – tandem mass spectrometry that offers a simple and 95

reliable determination of FOM in plasma and urine, with a quick and reproducible 96

sample preparation. 97

2. Experimental 98

2.1. Chemicals and materials 99

Fosfomycin (FOM), ethylphosphonic acid (EPA, internal standard) and acetonitrile 100

(HPLC gradient-grade solvent) were purchased from Sigma-Aldrich and ammonium 101

acetate was obtained from Ajax Univar. Ultra-pure water was obtained using a 102

Permutit system. Drug-free human plasma was obtained from the Australian Red 103

Cross Blood Service and drug-free urine was obtained from healthy volunteers. 104

2.2. Instrumentation and conditions 105

The LC-MS/MS used is a Shimadzu Nexera UHPLC equipped with a triple quadrupole 106

8030+ Shimadzu mass spectrometer (MS) detector. An electro-spray ionization (ESI) 107

Page 6 of 23

Accep

ted

Man

uscr

ipt

6

source interface operating in negative-ion mode was used for the multiple reaction 108

monitoring (MRM) LC-MS/MS analysis. MS conditions for FOM and the internal 109

standard (IS; EPA) are reported in Table 1. The interface settings consisted of the 110

nebulizing gas flow of 3 L/min, a de-solvation line temperature of 250°C, heat block 111

temperature of 400°C and a drying gas flow of 15 L/min. 112

The compounds were separated on a Merck SeQuant zic-HILIC, 2.1 x 50 mm, 5.0 μm 113

analytical column (operated at 24° C) protected by a 20 mm SeQuant zic-HILIC 114

guard cartridge using an isocratic mobile phase containing acetonitrile with 2 mM 115

ammonium acetate, pH 4.8 (85/15 v/v) at a flow rate of 0.3 mL/min. The injection 116

volumes used were 0.1 μL for the plasma assay and 0.5 μL for the urine assay. The 117

retention time for both FOM and EPA was 2.5 min. 118

2.3. Stock and standard solution preparation 119

2.3.1. Standards for plasma analysis 120

Aqueous stock solutions for plasma standard preparation (at 1, 2 and 10 mg/mL) 121

were stored at -80°C. On the day of assay these were diluted with drug free plasma 122

to yield calibration standards from 1 to 2000 µg/mL that was processed alongside 123

the clinical samples. 124

2.3.2. Standards for urine analysis 125

Aqueous stock solutions for urine standards of FOM (2, 5, 10, and 50 mg/mL) were 126

stored at -20°C. On the day of assay these were diluted with drug free urine to 127

prepare calibration standards from 0.1 to 10 mg/mL that were processed alongside 128

the clinical samples. 129

2.3.3. Internal standard solution 130

Ethylphosphonic acid (EPA) in acetonitrile was used as internal standard for the 131

plasma assay (10 µg/mL), and an aqueous EPA solution was used as internal standard 132

for the urine assay (1 mg/mL). Solutions were stored at -20°C. 133

Page 7 of 23

Accep

ted

Man

uscr

ipt

7

2.3.4. Quality Controls 134

Quality controls were prepared by spiking drug free plasma with FOM to 135

concentrations of 3, 800 and 1600 µg/mL, and stored at -80°C until assay. On the day 136

of assay an additional QC at 80 µg/mL was prepared by diluting with blank plasma 137

the QC at 800 µg/mL. The four sets of QCs were assayed alongside clinical samples. 138

Quality controls for urine analysis were prepared at FOM concentrations of 0.3, 2 139

and 8 mg/mL. The urine QCS were stored at -20°C until assayed alongside clinical 140

samples. 141

2.4. Analytical procedure 142

2.4.1. Clinical plasma sample preparation 143

Clinical samples were prepared by combining 10 μL of clinical sample, 10 μL of 144

water, and 90 μL of drug-free blank plasma. Internal standard (300 μL, 10 μg/mL of 145

EPA in acetonitrile) was then added and the sample vortexed and then centrifuged 146

(for 5 min at 14,000 rpm) to remove precipitated proteins. A volume of 0.1 μL was 147

injected onto the LC-MS/MS system. 148

2.4.2. Clinical urine sample preparation 149

All urine samples were filtered using a 0.22 µm filter prior to use. Clinical samples 150

were prepared by combining 10 μL of sample with 10 μL water, followed by internal 151

standard (20 μL, 1 mg/mL of EPA). The sample was then diluted with 200 μL of 152

mobile phase and 0.5 μL was injected into the LC-MS/MS system. 153

2.4.3. Data analysis 154

For both plasma and urine the concentration of each clinical sample was obtained 155

using the data from the calibration curve prepared (in either plasma or urine) within 156

the batch using a quadratic regression with peak-area ratio (drug/internal standard 157

area responses) against concentration (x), with a 1/x2 weighting as the mathematical 158

basis of the quantification. 159

Page 8 of 23

Accep

ted

Man

uscr

ipt

8

2.5. Validation 160

The validation was performed in accordance with the guidelines provided by the US 161

Food and Drug Administration (USFDA) and met the criteria required to demonstrate 162

the method is suitable for intended purpose [23]. The validation for both plasma and 163

urine was assessed for matrix effects, lower limit of quantification (LLOQ), linearity, 164

inter-day precision and accuracy, freeze-thaw stability of quality control samples and 165

the stability of standard solutions. 166

2.5.1. Limit of quantification 167

The lower limits of quantification for FOM were evaluated by analysis of replicate 168

standards, for both plasma and for urine samples. 169

2.5.2. Linearity 170

To investigate linearity, calibration curves were prepared using the corresponding 171

concentration ranges suitable for each matrix. 172

2.5.3. Inter-day precision and accuracy 173

Precision and accuracy for FOM throughout the calibration range of both plasma and 174

urine was evaluated by the analysis of QC samples at four different concentrations 175

for plasma and three different concentrations for urine with the QC concentrations 176

determined against freshly prepared standard curves. In addition to precision and 177

accuracy data obtained from QC samples, an incurred sample reanalysis was 178

performed. 179

Acceptance criteria were applied according to the US Food and Drug Administration 180

guidelines [23]; with acceptance criteria on an incurred sample reanalysis applied 181

according to the European Medicines Agency guidelines [24]. 182

2.5.4. Matrix effects 183

Plasma matrix effects were evaluated to identify any suppression or enhancement of 184

signal from an interfering substance around the retention time of FOM and EPA by 185

using the matrix factor test. Five blank plasma samples were assayed at spiked low 186

and high concentration levels and with internal standard, and the area results 187

Page 9 of 23

Accep

ted

Man

uscr

ipt

9

compared to those produced following the same extraction procedure using water 188

instead of plasma. The precision of the matrix factor (normalized against the internal 189

standard) was used to determine if any concentration level demonstrated a trend of 190

variation from the expected result. 191

Five urine blanks were assayed at a spiked low and high concentration level and the 192

precision and accuracy calculated, with respect to the nominal concentrations, to 193

determine if any concentration level demonstrated a trend of variation from the 194

expected result. 195

2.5.5. Recovery 196

The recovery of FOM and EPA was evaluated by comparing the peak area for plasma 197

or urine samples spiked prior to protein precipitation (for plasma) or dilution (for 198

urine) with samples spiked after protein precipitation or dilution. Care was taken to 199

ensure the injection matrix was identical in comparable samples. 200

2.5.6. Stability 201

Stability of FOM in plasma and urine was assessed across three freeze-thaw cycles 202

(from -80°C to ambient temperature) using three replicates of the QC samples at 203

low, medium and high concentrations. Stability of stock solutions was assessed 204

comparing aqueous solutions stored at both -20°C and -80°C to freshly prepared 205

solutions. 206

2.6. Pharmacokinetic application 207

The method was developed for the analysis of plasma and urine samples from a 208

pharmacokinetic study with critically ill patients receiving an intravenous dose of 6 g 209

of FOM every 6 hours with expected peak plasma concentrations of around 200 210

µg/mL, an expected plasma half-life of 2 h, and urinary concentrations of around 5 211

mg/mL [25]. 212

One critically ill patient was administered an intravenous dose of 6 g fosfomycin 213

disodium. Blood samples (3 mL) were taken prior to dosing (0 h) and 0.5, 0.75, 1, 1.5, 214

2, 4, and 6 h post administration using heparinized vacuum tubes (Greiner Bio-One, 215

Vacuette® LiHep) on the first day of FOM administration and after receiving the fifth 216

Page 10 of 23

Accep

ted

Man

uscr

ipt

10

FOM dose. Blood samples were centrifuged at 3000 rpm for 10 min to obtain plasma 217

samples. Plasma samples were transferred into 2 mL polypropylene tubes, capped 218

and stored at -80°C until analysis. 219

Similarly, a urine sample was collected 6 h post administration of the FOM dose. The 220

urine was transferred into a urine specimen vial, capped and stored at -80°C until 221

analysis. 222

This procedure was conducted in accordance with the principles laid down by the 223

ICH guidelines for Good Clinical Practice and approved by the University of 224

Queensland Medical Research Review Committee (clearance # 2012000870) and the 225

Epistimoniko Symvouleio (Scientific Committee) of Attikon University Hospital 226

(approval MEΘ-84/13-3-12). 227

3. Results and discussion 228

3.1. Chromatography 229

This method has been established using HILIC technology which offers excellent 230

selectivity for polar hydrophilic compounds like FOM, and the use of a high organic-231

solvent content in the mobile phase leads to a rapid evaporation of the solvent 232

during electrospray ionization [26], endowing the method with a simple 233

compatibility with mass spectrometry. Additionally, the use of HILIC technology in 234

this method obviates the requirement for further modification of the sample, after a 235

solvent-based protein precipitation, to closely correspond with the organic content 236

of the mobile phase for improved peak shape. This simplicity of extraction leads to 237

low detection levels when using low sample volumes. 238

Retention of the analyte by the stationary phase is caused by hydrophilic partitioning 239

within an aqueous-enriched liquid layer and/or with the positive or negative charges 240

on the HILIC stationary phase [27]; the balance of the partition is provided by the 241

aqueous content and pH of the mobile phase. Therefore, the aqueous layer is critical 242

to the efficiency of HILIC separation. Published HILIC methodology often 243

recommends mobile phases consisting of high ionic strength (from 5 – 20 mM, with 244

upper limits of 200-300 mM) [28] but this presented difficulties in development. The 245

Page 11 of 23

Accep

ted

Man

uscr

ipt

11

use of mobile phases with high ionic strength of buffers, low and variable aqueous 246

content, combined with the high pressure applied in an HPLC system led to blocking 247

of the column and the capillary in the ionization source on the mass spectrometer. 248

This isocratic method uses a low ionic strength of 2 mM ammonium acetate buffer in 249

85% acetonitrile which has been reliable and provided consistent results with 250

minimal loss to chromatographic shape and reproducibility. Regenerating the 251

column after every 300 – 400 injections of samples, particularly the urine samples, 252

and keeping buffer concentration to as low an ionic strength as possible, was 253

advantageous to long term use. Using a guard column extended the column life but 254

as the separation is highly dependent on the salt concentration and its impact on the 255

stationary phases aqueous-rich layer, the regeneration of the column was critical to 256

maintaining the quality of chromatography. 257

Another aspect of the method development that was found to affect both the 258

chromatographic retention and peak shape was the injection volume and the 259

composition of the sample being injected. Despite the low injection volumes being 260

used, 0.1 μL for plasma and 0.5 μL for urine, we conclude that the changes in the 261

quality of the chromatography observed were due to the sensitivity of the 262

interaction of FOM with the stationary phase from changes in ionic strength and pH 263

of the buffer [26, 27]. For the development of a bio-analytical assay using HILIC 264

chromatography, consideration of the organic to aqueous ratio, concentration of 265

salts, and finally, the presence of acids or base, is required. A dilution with plasma 266

was used in the plasma method as the ratio of acetonitrile to aqueous concentration 267

had an impact on the quality of the chromatographic peak shape and retention; the 268

concentrations found in clinical samples from the pilot study allowed this dilution. 269

The urine method included a dilution of sample that improved the chromatographic 270

separation by either reducing the presence of endogenous salts in the sample or 271

controlling the pH. 272

Buszewski [26] and Alpert describe the mechanism of HILIC separation as being 273

based on an interplay of a partitioning equilibrium in the aqueous layer (based on 274

the hydrophilicity of the analyte), weak electrostatic mechanisms, and dipole-dipole 275

interactions (including hydrogen-bonding) [29] the impact of each parameter on the 276

Page 12 of 23

Accep

ted

Man

uscr

ipt

12

selectivity and reproducibility of chromatography requires a more sophisticated 277

management than in general reversed-phase chromatography, but which once 278

overcome can lead to a highly stable and robust method. 279

3.2. Validation 280

The LLOQ was determined as 1 μg/mL for plasma and 0.1 mg/mL urine with 281

precision calculated as 4.7 and 3.1%, respectively, and accuracy calculated as 1.7 and 282

1.2%, respectively (Table 2). The signal to noise ratio of the lowest standard in the 283

calibration curve was 23.2 for plasma and 178 for urine and this data, combined with 284

the excellent precision and accuracy obtained at 1 μg/mL for plasma and 0.1 mg/mL 285

for urine, suggests substantial scope for achieving a lower LLOQ for both matrices 286

(see Figure 2 and 3 for a representative chromatogram of the LLOQ standard 287

extracted from plasma and urine, respectively). The lower limit of detection (LLOD) is 288

defined as being reliably distinguished from the background noise and calculated as 289

≥ three-times the noise of the blank plasma sample. From the validation the LLOD 290

was determined as being approximately 0.01 μg/mL for plasma and <0.01 mg/mL for 291

urine. 292

A regression model with a weighted (1/x2) quadratic curve provided the lowest 293

distribution of error across the substantial concentration range (1 to 2000 μg/mL for 294

plasma and 0.1 to 10 mg/mL for urine). The results of the linearity study are 295

described in Table 3. 296

Precision and accuracy of the QC samples are shown in Table 4 for both plasma and 297

urine. All results were within the acceptance criteria of ± 15% of the nominal 298

concentration, indeed the results of all plasma QCs samples were within 9.1% and 299

urine within 4.2%. An incurred sample reanalysis was performed on a subset of 300

clinical samples and the results meet the current guidelines [23, 24] with >67% of 301

repeated results being within 30% of the mean. Indeed, 100% of the repeated results 302

were within 11%. 303

No signal suppression/enhancement was evident for either FOM or the internal 304

standard from the matrix study performed. The matrix effect evaluation is reported 305

in Table 5. 306

Page 13 of 23

Accep

ted

Man

uscr

ipt

13

Despite using a very simple protein-precipitation for the extraction of FOM from 307

plasma the recovery was somewhat low at 68%. However, this extraction recovery is 308

not atypical for a drug with a highly hydrophilic nature due to preferential aqueous 309

partitioning. As was seen from the LLOQ testing, the variability was reliable 310

(precision 6.1%) and sensitivity (LLOD ca. 0.01 μg/mL) easily achievable. The recovery 311

of the internal standard EPA was good at 72% when tested at the undiluted 312

concentration of 10 μg/mL. The urine preparation was a simple dilution with internal 313

standard and as such provided recoveries of 98% when tested at 0.4 mg/mL. The 314

recovery results are reported in Table 5. 315

Stock solution stability for FOM was tested for aqueous solutions stored for over 16 316

months at -80°C and for over 11 months at -20°C and it was found to be stable. FOM 317

was also found to be stable in plasma and urine across three freeze-thaw cycles 318

when stored at -80°C and thawed at ambient temperature in water (see Table 5). 319

Overall, the validation of this method was highly successful for both plasma and 320

urine with the method showing an excellent degree of reproducibility and accuracy, 321

and is suitable for use in the analysis of patient samples in a pharmacokinetic study. 322

This technique offers a simple and robust method for the analysis of FOM in both 323

plasma and urine in patient samples. Other quantitative methods have been 324

described for the determination of FOM in serum or plasma [10, 12, 14, 15, 20, 22] 325

and urine [11]. However, these methods often require a significant amount of time 326

in sample preparation or technique, and none offer a chromatographic system 327

suitable for a pharmacokinetic study of FOM in both plasma and urine. 328

3.3. Pharmacokinetic analysis 329

The plasma concentration-time profile obtained in the pilot pharmacokinetic study is 330

shown in Figure 4. The peak plasma concentration in this patient after receiving the 331

initial dose was 222 µg/mL, and the trough concentration was 141 µg/mL. Increased 332

plasma concentrations were observed after receiving the fifth dose of FOM, with the 333

peak plasma concentration recorded as 868 µg/mL, and the trough concentration 334

was 592 µg/mL. The urinary concentration determined from a 6 h sample taken 335

Page 14 of 23

Accep

ted

Man

uscr

ipt

14

post-dose and was 2.03 mg/mL after the initial dose and 0.29 mg/mL after the fifth 336

dose of FOM. 337

4. Conclusion 338

The developed analytical method is a sensitive, simple and robust tool to analyse 339

FOM in plasma and urine of patients. With the increasing prevalence of MDR 340

organisms and the reduced effectiveness of currently available antibiotics this 341

method allows the opportunity to study the disposition of FOM, particularly in at-risk 342

patient groups. This research may improve dosing strategies which could minimize 343

the risk of increasing resistance and bring an effective antibiotic back into the hands 344

of treating physicians. 345

346

Page 15 of 23

Accep

ted

Man

uscr

ipt

15

347

Table 1. MS conditions for FOM and EPA 348

MS FOM EPA

Product Ion 137.1 (M.H-) 109.1 (M.H-)

Daughter Ion 78.9 78.9

Dwell Time (ms) 100 100

Q1 (V) +14 +10

Q3 (V) +28 +13

Collision Energy (V) +25 +20

349

Table 2: Lower limit of quantification 350

Matrix Mean Precision (%) Accuracy (%)

Plasma (n = 13) 1.02 µg/mL ±4.7 +1.7

Urine (n = 7) 0.10 mg/mL ±3.1 +1.2

351

Table 3: Linearity analysis 352

Matrix Calibration Range Correlation Coefficient

(Mean)

Maximum deviation*

(%)

Plasma (n=12) 1 to 2000 µg/mL 0.9963 -14.0

Urine (n=5) 0.1 to 10 mg/mL 0.9959 +10.5

* Reported maximum deviation from nominal (%) across all standard curves and all concentration levels. 353

Page 16 of 23

Accep

ted

Man

uscr

ipt

16

354

Table 4: Inter-assay Precision and Accuracy 355

Matrix Concentration Replicates Mean Accuracy (%) Precision (%)

Pla

sma

3 μg /mL 22 2.93 -2.4 9.1

80 μg /mL 27 78.0 -3.1 7.8

800 μg/mL 22 826 3.3 4.3

1600 μg/mL 23 1486 -7.2 5.9

Uri

ne

0.3 mg /mL 9 0.30 0.0 4.2

2 mg /mL 9 2.0 1.6 8.1

8 mg/mL 9 7.9 -1.9 2.3

356

357

Table 5: Matrix, recovery and freeze-thaw stability studies 358

359

Study Matrix Concentration Mean Accuracy (%) Precision (%)

Mat

rix

Plasma 3 μg /mL 1.05a 8.8

800 μg /mL 0.98a 3.9

Urine 0.2 mg/mL 0.189 4.6 7.2

5 mg/mL 5.36 -5.4 7.1

Rec

ove

ry

Plasma 80 μg /mL 68 7.7

Urine 10 μg/mL 98 4.2

Stab

ility

(fr

eeze

-th

aw)

Plasma 0.3 μg /mL 0.31 3.2 4.0

5 μg /mL 4.8 -3.1 5.4

80 μg /mL 86 7.2 5.8

Urine 0.3 mg /mL 0.30 0.9 2.7

2.1 mg /mL 2.3 8.2 3.1

7.8 mg /mL 8.2 4.4 1.0 a matrix factor: calculated as a ratio of peak area of FOM in the presence of matrix to the peak area in 360

the absence of matrix (normalized using the internal standard). 361

362

363

Page 17 of 23

Accep

ted

Man

uscr

ipt

17

364

References 365

1. M.E. Falagas, A.C. Kastoris, A.M. Kapaskelis, D.E. Karageorgopoulos, 366 Fosfomycin for the treatment of multidrug-resistant, including extended-367 spectrum beta-lactamase producing, Enterobacteriaceae infections: a 368 systematic review. Lancet Infect. Dis. 10 (2010) 43-50. 369

2. A. Dinh, J. Salomon, J.P. Bru, L. Bernard, Fosfomycin: Efficacy against 370 infections caused by multidrug-resistant bacteria. Scand. J. Infect. Dis. 44 371 (2012) 182-189. 372

3. S. Parker, J. Lipman, D. Koulenti, G. Dimopoulos, J.A. Roberts, What is the 373 relevance of fosfomycin pharmacokinetics in the treatment of serious 374 infections in critically ill patients? A systematic review. Int. J. Antimicrob. 375 Agents 42 (2013) 289-293. 376

4. R. Raz, Fosfomycin: an old-new antibiotic. Clinical Microbiology and Infection 377 18 (2012) 4-7. 378

5. K. Takahashi, H. Kanno, Synergistic activities of combinations of beta-lactams, 379 fosfomycin, and tobramycin against Pseudomonas-aeruginosa. Antimicrobial 380 Agents and Chemotherapy 26 (1984) 789-791. 381

6. F. Tessier, C. Quentin, In vitro activity of fosfomycin combined with 382 ceftazidime, imipenem, amikacin, and ciprofloxacin against Pseudomonas 383 aeruginosa. Eur. J. Clin. Microbiol. Infect. Dis. 16 (1997) 159-162. 384

7. N. Roussos, D.E. Karageorgopoulos, G. Samonis, M.E. Falagas, Clinical 385 significance of the pharmacokinetic and pharmacodynamic characteristics of 386 fosfomycin for the treatment of patients with systemic infections. Int. J. 387 Antimicrob. Agents 34 (2009) 506-515. 388

8. M. Frossard, C. Joukhadar, B.M. Erovic, P. Dittrich, P.E. Mrass, M. Van Houte, 389 H. Burgmann, A. Georgopoulos, M. Müller, Distribution and antimicrobial 390 activity of fosfomycin in the interstitial fluid of human soft tissues. 391 Antimicrobial Agents and Chemotherapy 44 (2000) 2728-2732. 392

9. S.S. Patel, J.A. Balfour, H.M. Bryson, Fosfomycin tromethamine - A review of 393 its antibacterial activity, pharmacokinetic properties and therapeutic efficacy 394 as a single-dose oral treatment for acute uncomplicated lower urinary tract 395 infections. Drugs 53 (1997) 637-656. 396

10. M.C. Dessalles, J. Levieux, M. Souleau, G. Mahuzier, Determination of 397 fosfomycin by gas-chromatographic method in biological fluids. Pathologie 398 Biologie 35 (1987) 200-204. 399

11. M.C. DiosVieitez, M.M. Goni, M.J. Renedo, D. Fos, Determination of 400 fosfomycin in human urine by capillary gas chromatography: Application to 401 clinical pharmacokinetic studies. Chromatographia 43 (1996) 293-295. 402

Page 18 of 23

Accep

ted

Man

uscr

ipt

18

12. E. Hernandez, A. Loste, M.A. Bregante, M.A. Garcia, C. Solans, Determination 403 of fosfomycin in chicken plasma samples by gas chromatography: Application 404 to pharmacokinetic studies. Chromatographia 54 (2001) 365-368. 405

13. A. Loste, E. Hernandez, M.A. Bregante, M.A. Garcia, C. Solans, Development 406 and validation of a gas chromatographic method for analysis of fosfomycin in 407 chicken muscle samples. Chromatographia 56 (2002) 181-184. 408

14. G.K. Webster, R.G. Bell, Gas chromatographic analysis of fosfomycin in 409 plasma for pharmacokinetic analysis. Journal of AOAC International 82 (1999) 410 620-624. 411

15. G.A. Pianetti, L.M.M. Decampos, P. Chaminade, A. Baillet, D. Baylocqferrier, 412 G. Mahuzier, Application of ion chromatography with indirect 413 spetcrophotometric detection to the sensitive determination of 414 alkylphosphonic acids and fosfomycin. Analytica Chimica Acta 284 (1993) 415 291-299. 416

16. Y.L. Hu, Y.Q. Feng, Q.H. Zhang, S.L. Da, Determination of fosfomycin by 417 indirect spectrophotometric method. Talanta 49 (1999) 47-52. 418

17. A. Baillet, G.A. Pianetti, M. Taverna, G. Mahuzier, D. Baylocqferrier, 419 Fosfomycin determination in serum by capillary zone electrophoresis with 420 indirect ultraviolet detection. J. Chromatogr.-Biomed. Appl. 616 (1993)311-421 316. 422

18. M. Petsch, B.X. Mayer-Helm, R. Sauermann, C. Joukhadar, E. Kenndler, 423 Capillary electrophoresis analysis of fosfomycin in biological fluids for clinical 424 pharmacokinetic studies. Electrophoresis 25 (2004) 2292-2298. 425

19. T.A. Papakondyli, A.M. Gremilogianni, N.C. Megoulas, M.A. Koupparis, A 426 novel derivatization method for the determination of Fosfomycin in human 427 plasma by liquid chromatography coupled with atmospheric pressure 428 chemical ionization mass spectrometric detection via phase transfer 429 catalyzed derivatization. J. Chromatogr. A 1332 (2014) 1-7. 430

20. L. Li, X. Chen, X. Dai, H. Chen, D. Zhong, Rapid and selective liquid 431 chromatographic/tandem mass spectrometric method for the determination 432 of fosfomycin in human plasma. Journal of Chromatography B-Analytical 433 Technologies in the Biomedical and Life Sciences 856 (2007) 171-177. 434

21. W. Poeppl, T. Lingscheid, D. Bernitzky, O. Donath, G. Reznicek, M. Zeitlinger, 435 H. Burgmann, Assessing Pharmacokinetics of Different Doses of Fosfomycin 436 in Laboratory Rats Enables Adequate Exposure for Pharmacodynamic Models. 437 Pharmacology 93 (2014) 65-68. 438

22. S. Dieguez, A. Soraci, O. Tapia, R. Carciochi, D. Pérez, R. Harkes, O Romano, 439 Determination of antibiotic fosfomycin in chicken serum by liquid 440 chromatography-tandem mass spectrometry. Journal of Liquid 441 Chromatography & Related Technologies 34 (2011) 116-128. 442

23. Centre for Drug Evaluation and Research: Food and Drug Administration: 443 Guidance for Industry: Bioanalytical Method Validation. (May 2001). 444

Page 19 of 23

Accep

ted

Man

uscr

ipt

19

24. European Medicines Agency : Guideline on bioanalytical method validation. 445 (EMEA/CHMP/EWP/192217/2009), (21 July 2011). 446

25. W.M.M. Kirby: Pharmacokinetics of Fosfomycin. Chemotherapy 23 (1977) 447 141-151. 448

26. B. Buszewski, S. Noga, Hydrophilic interaction liquid chromatography (HILIC)-449 a powerful separation technique. Analytical and Bioanalytical Chemistry 402 450 (2012) 231-247. 451

27. B. Dejaegher, D. Mangelings, Y.V. Vander Heyden, Method development for 452 HILIC assays. Journal of Separation Science 31 (2008) 1438-1448. 453

28. Merck: ZIC®-HILIC HPLC Column: General Instructions for Care and Use. ZA-2-454 11-02 (2002). 455

29. A.J. Alpert, Hydrophilic-Interaction Chromatography for the separation of 456 peptides, nucleic-acids and other polar compounds. J. Chromatogr. 499 457 (1990) 177-196. 458

30. Chemistry RSO: ChemSpider. (15 September 2014), CSID 394204 and 177126 459 (2014). 460

31. Drugbank: DrugBank: Fosfomycin (DB00828). 2013 (26 March 2013). 461

462

463

Page 20 of 23

Accep

ted

Man

uscr

ipt

Figure 1: Structure of fosfomycin (FOM, left) and the internal standard,

ethylphosphonic acid (EPA, right),[29,30].

Figure 1

Page 21 of 23

Accep

ted

Man

uscr

ipt

Figure 2: Chromatograms of a blank sample (top) and the LLOQ (1 µg/mL) plasma

standard (FOM, centre; EPA, bottom).

1.00 1.25 1.50 1.75 2.00 2.25 2.50 2.75 3.00 3.25

0

10

20

30

40137.10>78.90(-)

1.00 1.25 1.50 1.75 2.00 2.25 2.50 2.75 3.00 3.25

0

500

1000

1500

2000

137.10>78.90(-)

FFM

1.25 1.50 1.75 2.00 2.25 2.50 2.75 3.00 3.25

0

100000

200000

300000

400000

500000

109.10>78.90(-)

EPA

Time (min)

Intensity

(cps)

Figure 2

Page 22 of 23

Accep

ted

Man

uscr

ipt

Figure 3: Chromatograms of blank sample (top) and the LLOQ (0.1 mg/mL) urine

standard (FOM, centre; EPA, bottom)

0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 2.75

0

25

50

75

100137.10>78.90(-)

0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 2.75

0

1000

2000

3000

4000

5000

6000

137.10>78.90(-)

FFM

0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 2.75

0

25000

50000

75000

100000109.10>78.90(-)

EPA

Time (min)

Intensity

(cps)

Figure 3

Page 23 of 23

Accep

ted

Man

uscr

ipt

Figure 4: Plasma concentration – time profiles of FOM in a critically ill patient

receiving a 6 g FOM IV dose every 6 hours, for the first and fifth doses.

0

200

400

600

800

1000

0 1 2 3 4

FOM plasma concentration

(μg/mL )

Time (h)

Dose 1

Dose 5

Figure 4