Determination of acetaminophen and its main metabolites in urine by capillary electrophoresis hyphenated to mass spectrometry

Talanta 205 (2019) 120108

Contents lists available at ScienceDirect

Talanta journal homepage: www.elsevier.com/locate/talanta

Determination of acetaminophen and its main metabolites in urine by capillary electrophoresis hyphenated to mass spectrometry

T

Marie Lecoeura, Guy Rabenirinaa, Nadège Schifanoa, Pascal Odoua,b, Sabine Ethgena,c, Gilles Lebuffea,c, Catherine Foulona,* a

Univ. Lille, EA 7365 - GRITA - Groupe de Recherche sur les Formes Injectables et Technologies Associées, F-59000, Lille, France CHU Lille, Institut de Pharmacie, F-59000, Lille, France c CHU Lille, Pôle d’Anesthésie-Réanimation, F-59000, Lille, France b

ARTICLE INFO

ABSTRACT

Keywords: Capillary electrophoresis (CE) Electrospray ionization interface (ESI) Mass spectrometry (MS) Acetaminophen (APAP) Metabolites Urinary samples

In this study, a capillary electrophoresis-tandem mass spectrometry method combining efficient separation and sensitive detection has been developed and validated, for the first time, to quantify acetaminophen and five of its metabolites in urine samples. Optimization of the method has led us to perform detection in positive ESI mode using MeOH-ammonium hydroxide (0.1%) (50:50, v/v) as sheath liquid. Moreover, optimal separation has been obtained in less than 9 min after anodic injection, using an ammonium acetate solution (40 mM, pH 10) as BGE. It was shown that the dilution solvent and the dilution factor to use for sample preparation are critical parameters to avoid peak splitting, to gain in sensitivity and then to obtain an effective analysis method. While a 200fold factor dilution was shown to be suitable for quantitation of acetaminophen, acetaminophen mercapturate, acetaminophen sulfate and acetaminophen glucuronide, a 20-fold dilution was finally selected for methoxyacetaminophen and 3-methylthioacetaminophen analysis, thus requiring two successive analyses to be carried out in order to quantify all metabolites. Hyphenation of CE with MS/MS versus UV permits to improve LOQ (10–20-fold factor with respect to previous works for acetaminophen, acetaminophen sulfate and acetaminophen glucuronide). Moreover, use of CE versus HPLC, permits to quantify two additional metabolites, i.e. 3-methylthio-acetaminophen and methoxy-acetaminophen. The method has been validated using the accuracy profile approach with a total error (accuracy) included in the ± 20% range. Thereby, the method allows the quantitation of acetaminophen and acetaminophen mercapturate in the range (0.1–1 mg mL-1), and of acetaminophen sulfate, methoxy-acetaminophen, acetaminophen glutathione and 3-methylthio-acetaminophen in the ranges (0.5–5 mg mL-1), (0.025–0.4 mg mL-1), (9.22–30 mg mL-1) and (0.073–0.4 mg mL-1), respectively. The method was finally applied to the analysis of urine samples of eighteen patients belonging to three different inclusion groups of the ongoing clinical trial, demonstrating that the method is suitable to highlight different metabolic profiles. This work will be subsequently extended to the analysis two hundred and seventy urine samples from patients included in a clinical trial dedicated to the study of acetaminophen metabolism changes after hepatic resection.

1. Introduction The management of pain is a determining factor for recovery after major abdominal liver surgery. Multimodal analgesia, combining local anaesthetic techniques with systemic analgesics, that exhibit different action mechanisms, is the reference strategy for post-operative pain treatment [1,2]. Acetaminophen (APAP) is almost systematically used. Although this drug is very safe in therapeutic doses (4 g per day) for

most patients, it sometimes exhibits hepatic toxicity even in these therapeutic doses, especially among population with similar type risk factors such as those encountered in cases of hepatic surgery, acute or chronic hepatopathies, chronic alcohol consumption or fasting and malnutrition [3]. These last years, hepathic surgery has experienced significant growth with an enlargement of therapeutic indications resulting from an improvement of surgery techniques and anaesthesic management.

Abbreviations: Acetaminophen, (APAP); acetaminophen sulfate, (APAP-S); acetaminophen glucuronide, (APAP-G); 3-methoxy-acetaminophen, (M-APAP); Acetaminophen glutathione, (APAP-GT); acetaminophen mercapturate, (APAP-M); 3-methylthio-acetaminophen, (MT-APAP) * Corresponding author. Univ Lille, EA 7365, UFR Pharmacie, 59000, Lille, France. E-mail address: [email protected] (C. Foulon). https://doi.org/10.1016/j.talanta.2019.07.003 Received 25 March 2019; Received in revised form 27 June 2019; Accepted 1 July 2019 Available online 02 July 2019 0039-9140/ © 2019 Elsevier B.V. All rights reserved.

Talanta 205 (2019) 120108

M. Lecoeur, et al.

Fig. 1. Schema of acetaminophen metabolism.

Unfortunately, in this context, acetaminophen, could induce an overdose, as hepatic function can be altered by reduction of hepatocyte volume and oxidative stress [4]. The result is a decrease in the liver glutathione stocks that may induce a change in the acetaminophen metabolism [5]. Indeed, despite the major routes involve the conjugation to endogeneous glucuronide and sulfate (Fig. 1) to form inactive metabolites (APAP-G and APAP-S, respectively) [6], a small portion is oxidized by cytochrome P-450 enzymes, to form the reactive intermediate N-acetylbenzoquinone imine (NAPQI) which is preferentially conjugated with glutathione (APAP-GT) and then excreted in the urine as cysteine (APAP-C) and mercapturic acid conjugates (APAP-M) or finally as 3-methylthio-acetaminophen (MT-APAP). After major liver resection, the depletion of glutathione may lead to NAPQI accumulation and to its covalent binding to vital cell constituents, resulting in a liver necrosis [7,8]. A few studies have been dedicated to the evaluation of

acetaminophen metabolism changes after hepatic resection [2,9–11]. Some of them led to contradictory outcomes. Furthermore, recommendations focusing on acetaminophen intake after hepatic resection are conflicting, as some of them discouraging the use of acetaminophen [12,13], while others consider its use safe except for major resection [14]. Hence, a clinical trial including ninety patients, has been initiated at Lille University hospital (France), in order to assess the metabolism of acetaminophen administered post-operatively in major hepatic surgery (resection of three or more hepatic segments) with respect to a less extensive hepatic resection and to a hepatic re-operation. In addition to the assessment of the acetaminophen plasma levels for five days in post-operative liver surgery, this study includes the evaluation of the distribution of acetaminophen and its metabolites in urine. This requires a sensitive and robust method to quantify acetaminophen and its main metabolites in urine. Several methods using 2

Talanta 205 (2019) 120108

M. Lecoeur, et al.

HPLC were previously developed for their analysis in biological samples. Most of them employ UV detection for quantitation of APAP, APAP-S and APAP-G [15–17] and two more analytes, i.e. APAP-M and APAP-C in urine [18,19]. Separations were performed in 6–45 min and limits of quantitation, were ranged between 2 and 50 mg L−1, depending on the analyte and the study. HPLC methods using mass spectrometry (MS) detection were also developed for quantitation of acetaminophen and its metabolites (from one to five) in plasma and/or urine [9,20–23]. The greater the number of metabolites analyzed is, the longer the analysis duration is (between 6.5 and 26 min for one to 5 metabolites, respectively). Moreover, among UHPLC-MS/MS recently developed [24,25], only the method suggested by Flint et al. [25] makes it possible to analyze a large number of metabolites while having a reduced duration of analysis, i.e. five metabolites in 11 min. Moreover, to our knowledge, only two research teams have explored the analysis of acetaminophen and its metabolites by capillary electrophoresis. Hence, Heitmeier et al. [26] have demonstrated the performances of capillary electrophoresis for the analysis of acetaminophen and four of its metabolites (APAP-G, APAP-S, APAP-C and APAP-M) in urine and serum samples, with an efficient separation in less than 10 min. In this study, MS detection was only used for metabolites identification, while quantitation was performed by UV detection unfortunately with poor limits of quantitation in urine (between 7.5 and 19 mg L−1). Ullsten et al. [27] use CE-MS/MS and multivariate data analysis for metabolic profiling of urine using acetaminophen as a model. As a result, our goal was to develop and validate a CE-MS/MS method combining efficient separation and sensitive detection for quantitation of acetaminophen and six metabolites in urine samples using a simple dilution of the urine samples as this is the main used approach. This method was finally applied to quantify acetaminophen and its metabolites, in urine of patients included in the clinical trial.

were received from the Centre de Ressources Biologiques de Lille (CRB/ CIC1403). Each patient received APAP by intravenous route at a dose of 1 g, 30 min before the end of the surgery then 1 g every 6 h, for five days. Three urine samples were collected for five days (24 h, 72 h and 100 h after the first APAP administration) and were stored in polypropylene tubes at −20 °C. 2.3. Background electrolyte and sample preparation 2.3.1. Background electrolyte Separations were carried out in a pH 10 buffer solution prepared from a 40 mM ammonium acetate solution whose pH has been adjusted by the addition of a 1 M ammonium hydroxide solution, unless otherwise specified. Before use, this background electrolyte (BGE) was filtered through a 0.45 μm cellulose syringe filter (Alltech, Templemars, France). 2.3.2. Standard samples Stock solutions of APAP, APAP-M, M-APAP, MT-APAP, APAP-GT, [2H3]-APAP, [2H5]-APAP-M, [2H3]-MT-APAP and [2H3]-APAP-G (500 mg L−1) were prepared in MeOH, while stock solutions of APAP-S and APAP-G (500 mg L−1) were prepared in deionized water. These solutions were aliquoted and stored at −20 °C. Each day, an aliquot of each corresponding solution was thawed at room temperature, then diluted in the BGE in order to obtain working solutions of: APAP, APAPM, M-APAP or MT-APAP at 12.5 mg L−1; [2H3]-APAP, [2H5]-APAP-M or [2H3]-MT-APAP at 25 mg L−1; APAP-S or APAP-GT at 50 mg L−1; [2H3]-APAP-G at 100 mg L−1 and APAP-G at 250 mg L−1, respectively. These solutions were used to prepare two distinct series of calibration and validation standards. The first series was prepared from blank urine after a 200-fold factor dilution in the BGE and contained APAP, APAPM, APAP-S and APAP-G, as their internal standards, i.e. [2H3]-APAP (0.5 mg L−1), [2H5]-APAP-M (for APAP-M and APAP-S; 0.2 mg L−1) and [2H3]-APAP-G (10 mg L−1), respectively. The second series was prepared from blank urine after a 20-fold factor dilution and contained MAPAP and MT-APAP and chosen internal standard [2H3]-MT-APAP (0.1 mg L−1). Concentration range of the calibration standards and of the validation standards are summarized in Table 1.

2. Materials and methods 2.1. Chemicals Acetaminophen (APAP), acetaminophen sulfate (APAP-S), acetaminophen glucuronide (APAP-G), 3-methoxy-acetaminophen (MAPAP), [2H3]-acetaminophen glucuronide ([2H3]-APAP-G) were purchased from Santa Cruz Biotechnologies (Heidelberg, Germany). Acetaminophen glutathione (APAP-GT), acetaminophen mercapturate (APAP-M), 3-methylthio-acetaminophen (MT-APAP), [2H3]-acetaminophen ([2H3]-APAP), [2H5]-acetaminophen mercapturate ([2H5]APAP-M), [2H3]-3-methylthio-acetaminophen ([2H3]-MT-APAP) were supplied from Toronto Research Chemicals (Toronto, Canada). All solvents and reagents were of analytical grade. Ammonium acetate, ammonium hydroxide (28%) and methanol were purchased from SigmaAldrich (Saint-Quentin Fallavier, France). Ethanol, formic acid and sodium hydroxide were supplied from VWR (Fontenay-sous-Bois, France). All background electrolytes and sample solutions were prepared using purified water produced by a MilliQ purification system (Millipore, Guyancourt, France).

2.3.3. Urine samples Just before assay, patients’ urine samples stored at −20 °C were thawed at room temperature, centrifuged at 14,000 rpm for 10 min to obtain the corresponding supernatant. Two distinct sample types were then prepared in duplicate in the BGE: 200-fold diluted samples spiked with [2H3]-APAP, [2H5]-APAP-M and [2H3]-APAP-G for the assay concerning APAP, APAP-M, APAP-S and APAP-G and 20-fold diluted samples spiked with [2H3]-MT-APAP to analyse M-APAP and MT-APAP. 2.4. Instrumentation and operating conditions CE-MS/MS experiments were performed using a HP3DCE apparatus hyphenated with an Agilent Series 6420 triple quadrupole mass spectrometer via a commercial sheath liquid interface equipped with a stainless-steel needle (Agilent Technologies, Les Ulis, France). The CE apparatus was equipped with a diode-array detector, an autosampler and a power supply able to deliver a voltage up to 30 kV. The cassette temperature was set at 25 °C. Data were collected and analyzed using the Masshunter software. Separations were carried out in a bare fused-silica capillary (80 cm length, 50 μm i. d., 375 μm o. d.) obtained from Polymicro (CM Scientific, Silsden, UK). Prior to use, new capillaries were activated using the following three-step sequence: 1 M NaOH, 0.1 M NaOH and H2O, for 10 min each, by applying a pressure of 950 mbar. When not in use, capillary was rinsed by the procedure described previously and air-dried. Before each analysis, capillaries were successively flushed with 0.1 M NH4OH solution (5 min, 950 mbar), water (5 min, 950 mbar) and finally with the pH 10 BGE (5 min,

2.2. Urine samples collection Blank urine samples of two healthy volunteers who did not receive acetaminophen within the previous fifteen days were collected in the morning fasting. They were pulled together and centrifuged at 14,000 rpm for 10 min. The resulting supernatant was aliquoted in polypropylene tubes before being stored at −20 °C for a maximum of six months until use. Eighteen frozen urine samples of patients uniformly distributed in three inclusion groups (G1: major hepatectomy (≥3 liver segments); G2: minor hepatectomy (< 3 liver segments); G3: liver re-operation) 3

Talanta 205 (2019) 120108

M. Lecoeur, et al.

Table 1 Validation results of the CE-MS/MS method: response function, trueness and precision for the analysis of acetaminophen (APAP), acetaminophen sulfate (APAP-S), acetaminophen glucuronide (APAP-G), acetaminophen mercapturate (APAP-M), 3-methylthio-acetaminophen (MT-APAP) and methoxy-acetaminophen (M-APAP). Validation parameters

APAP

APAP-S

APAP-G

APAP-M

MT-APAP

M-APAP

0.1–1 Y = 1.23X-0.027

0.5–5 Y = 0.391X-0.332

2–30 Y = 1.09X+0.016

0.1–1.0 Y = 1.493X+0.064

0.025-0.4 Y = 0.491x+0.019

0.025-0.4 Y = 0.561X+0.236

1.029 −0.032 0.994

1.100 −0.594 0.991

1.003 −0.070 0.946

1.018 0.054 0.992

1.076 −0.043 0.980

0.948 0.068 0.959

Response functiona Conc. range f(x) Linearity Slope Y intercept r2 Trueness

Ccalculated (mg.L−1)

Relative bias (%)

Ccalculated (mg.L−1)

Relative bias (%)

Ccalculated (mg.L−1)

Relative bias (%)

Ccalculated (mg.L−1)

Relative bias (%)

Ccalculated (mg.L−1)

Relative bias (%)

Ccalculated (mg.L−1)

Relative bias (%)

VS1 VS2 VS3 VS4 Precision

0.10 0.24 0.42 1.02 Rep. (% RSD)

1.18 −3.68 −5.80 2.19 Inter. prec. (% RSD)

0.55 1.50 2.46 5.46 Rep. (% RSD)

9.69 −0.06 −1.46 9.22 Inter. prec. (% RSD)

1.60 8.80 19.60 29.40 Rep. (% RSD)

−19.54 −9.97 −2.16 −2.01 Inter. prec. (% RSD)

0.11 0.26 0.48 1.02 Rep. (% RSD)

6.19 5.87 6.85 2.46 Inter. prec. (%RSD)

0.03 0.10 0.20 0.43 Rep. (% RSD)

18.6 2.21 −0.45 7.93 Inter. prec. (% RSD)

0.02 0.11 0.20 0.38 Rep. (% RSD)

0.44 5.45 0.46 −4.18 Inter. prec. (% RSD)

VS1 VS2 VS3 VS4

2.19 4.01 6.00 2.18

7.83 4.01 6.67 4.26

9.93 6.12 7.59 3.11

12.04 6.12 7.59 3.90

18.34 9.64 11.36 13.10

30.15 9.64 11.36 16.82

10.54 5.95 2.08 2.40

12.07 10.10 8.42 4.79

4.90 8.86 9.09 7.85

21.13 12.36 9.09 9.08

15.04 16.39 17.73 11.65

16.80 16.39 17.73 11.65

Conc. range: concentration range. Rep.: repeatability. Inter. prec.: intermediate precision. b Concentrations of validation standards were the following: APAP (0.1, 0.25, 0.45 and 1.0 mg L−1); APAP-S (0.5, 1.5, 2.5 and 5.0 mg L−1); APAP-G (2.0, 10.0, 20.0 and 30.0 mg L−1); APAP-M (0.1, 0.25, 0.45 and 1.0 mg L−1); MT-APAP (0.025, 0.10, 0.20 and 0.40 mg L−1); M-APAP (0.025, 0.10,0.20 and 0.40 mg L−1). a Calibration curves, expressing the peak area ratio (analyte/EI) versus concentration ratio (analyte/EI) were obtained from six calibration standard by using least square weighted (1/X) linear regression. Table 2 MS/MS detection parameters. Analyte

M (g/mol)

Acceleration voltage (V)

Collision energy (eV)

Transition m/z

APAP APAP-S APAP-M APAP-G M-APAP MT-APAP APAP-GT [2H3]-APAP [2H5]-APAP-M [2H3]-MT-APAP [2H3]-APAP

151.1 231.0 312.1 327.1 181.1 197.0 456.1 154.1 317.1 200.1 330.1

110 130 105 135 110 110 135 110 105 110 135

14 11 13 12 17 16 12 14 13 16 12

152.0 → 110.0 231.8 → 152.0 312.9 → 207.9 349.9 → 174.0 182.0 → 108.0 198.0 → 108.1 456.9 → 327.9 155.0 → 111.1 317.9 → 211.9 201.0 → 159.0 352.9 → 176.9

950 mbar). Separations were carried out in the pH 10 BGE at 25 kV and 25 °C, after anodic hydrodynamic injection (100 mbar, 20 s) of the samples. The triple quadrupole mass spectrometer (Agilent Technologies) equipped with an orthogonal ESI source was used in the positive ionization mode, unless otherwise specified. Nitrogen was used as nebulising (NG) and as a drying gas (DG). In the optimized conditions, the nebulizer gas pressure was set at 10 psi, and the drying gas flow rate and temperature were set at 10 L min−1 and 250 °C, respectively. The ESI voltage was set at + 4000 V in positive mode and −4500 V in negative mode. The optimized coaxial sheath liquid consisting of MeOH–H20 (50:50, v/v) contained 0.1% NH4OH. It was delivered at a flowrate of 4 μL min−1 by a 1100 series isocratic pump (Agilent) equipped with a splitter (1:100). MS detection was carried out

in multiple reaction monitoring (MRM) mode. The acceleration voltages, collision energies and MRM transition selected for MS acquisition are summarized in Table 2. Peak width and dwell time were set at 0.12 min and 50 ms, respectively. 2.5. Validation of the method The validation of the method was performed according the validation guidelines proposed by the French Society of Pharmaceutical Sciences and Technologies – SFSTP [28–30]. Two distinct validations were performed: the first was dedicated to APAP, APAP-S, APAP-M and APAP-G analysis (200-fold dilution factor of the urines); the second, to M-APAP and MT-APAP (20-fold dilution factor). Hence, in each case, 4

Talanta 205 (2019) 120108

M. Lecoeur, et al.

Fig. 2. Effect of the sheath liquid composition on the MS sensitivity of acetaminophen and its metabolites. Each analyte (30 mg L−1) prepared in the sheath liquid was individually infused at a flow rate of 4 μL/min. S1: MeOH-formic acid (0.1%) (50:50, v/v); S2: MeOH-ammonium acetate (20 mM) (50:50, v/v); S3: MeOH-ammonium hydroxide (0.1%) (50:50, v/v). APAP: acetaminophen; M-APAP: 3-methoxy acetaminophen; MT-APAP: 3-methyl-thio-acetaminophen; APAP-S: acetaminophen sulfate; APAP-G: acetaminophen glucuronide; APAP-GT: acetaminophen gluthatione; APAP-M: acetaminophen mercapturate. Results are presented in relative scale where [M+H]+ of M-APAP obtained in S3 corresponds to 100%.

Finally, specificity, response function, linearity, trueness and precision (repeatability and intermediate precision) were studied. The acceptance criteria for precision and trueness were fixed in accordance with the requirements of the EMA [31] which recommends relative standard deviation (RSD) and relative bias under 15% for the VS samples (except 20% for the LLOQ). Lastly, accuracy profiles were assessed using NeoLiCy ® software (version:1.8.2.2) taking into account the ± 20% acceptance limits at a risk of 5%. They were used to determine the lower limit of quantitation (LLOQ). In addition, carryover was estimated by injecting blank samples after a mixture of the analytes (concentration equal to upper limit of quantification) and internal standards. Stability of the stock solutions and working solutions of the analytes and internal standards was assessed by comparison of freshly prepared solutions to solutions that have been stored at −20 °C for approximately three and six months. The stability of the analytes in urine was determined from two validation standards (lowest (VS2) and highest concentrations (VS4); n = 3 per concentration) in different storage conditions: at room temperature (20 ± 5 °C) for 6 h, 12 h and 24 h (sample processing and autosampler temperature), after three freezing-thawing cycles, at freezing temperature (−20 °C, temperature corresponding to the long term storage).

Fig. 3. Influence of the BGE pH (ammonium acetate 40 mM adjusted to suitable pH by addition of 1 M NH40H) on the migration times. Separation conditions: 25 kV; 25 °C; anodic injection.

the validation of the method was carried out on three consecutive days to estimate the prediction errors. Each day, six calibration standards (CS) and four validation standards (VS; each prepared in triplicate), a blank urine sample (CAL00), obtained by suitable dilution of the blank urine (200 or 20-fold dilution factor) and the same diluted blank urine spiked with internal standard (CAL0) were prepared and analysed.

3. Results and discussion The aim of this work was to develop and validate a sensitive CE-ESI/ MS-MS method to evaluate the distribution of acetaminophen and its 5

Talanta 205 (2019) 120108

M. Lecoeur, et al.

Fig. 4. Influence of the sampling solvent on the acetaminophen peak shape. Analysis of a 10-fold diluted blank urine spiked with analytes and their internal standard at 1 mg L−1, in (a) 4 mM ammonium acetate solution (10-fold diluted BGE, pH 10) and (b) 40 mM ammonium acetate solution (BGE, pH 10). Analysis conditions: pH 10 BGE (40 mM ammonium acetate adjusted to pH 10 by addition of 1 M NH4OH); anodic injection (100 mbar, 10 s); 25 kV; 25 °C.

urinary metabolites for patients who have undergone a major hepatic resection with respect to a less extensive one and to a hepatic re-operation.

improved at apparent pH 10.1, which is in accordance with an increased proportion of anionic species [M − H]- formed from acidic compounds, in alkaline conditions. In positive mode, a distinct behaviour was emphasized for monoacidic compounds versus polyacidic ones. For APAP, M-APAP and MT-APAP (monoacidic compounds), the MS response was clearly improved at apparent pH 10.1, while it slightly decreases for APAP-GT and APAP-M. For APAP-G and APAP-S, no substantial difference was observed by changing the sheath liquid composition. Moreover, except for APAP-M, the sensitivity was much higher in positive mode using pH 10.1 sheath liquid. This justifies the choice to operate in positive mode using MeOHammonium hydroxide (0.1%) (50:50, v/v) (apparent pH 10.1) as sheath liquid, in accordance with the works of Park et al. [9], contrarily to Heitmeier et al.[26].

3.1. Method development 3.1.1. Optimization of MS detection conditions 3.1.1.1. Sheath liquid composition. The composition and the flow rate of the sheath liquid play a major role when using a CE/ESI-MS hyphenated via a sheath liquid interface. Sheath liquid is usually composed of hydro-organic medium implemented with a volatile salt. Alcoholic solvents (MeOH, EtOH) are preferred to ACN due to their lower boiling point in favour of the desolvatation process. Since APAP and its metabolites possess a wide range of physico-chemical properties (namely one or several ionization functions), three sheath liquids were initially investigated to promote analyte ionization: MeOHformic acid (0.1%) (50:50, v/v) (apparent pH 3.1), MeOH-ammonium acetate (20 mM) (50:50, v/v) (apparent pH 7.1) and MeOH-ammonium hydroxide (0.1%) (50:50, v/v) (apparent pH 10.1). Each analyte was individually infused into the ESI-MS apparatus at a flowrate of 4 μL min−1. Fig. 2 exhibits the MS sensitivity of all the ions ([M − H]- and [M+H]+), determined in negative and positive ESI mode, respectively. In negative mode, the MS response was clearly

3.1.1.2. MS parameters. The selection of MRM transition is a key to obtain a reliable and sensitive method. Hence, for each analyte and selected internal standards, the acceleration voltage and the collision energy were optimized in the product ion mode, by infusion into the ESI-MS apparatus. Optimal operational parameters and MRM transitions from the precursor ion to the dominant product ion selected are summarized in Table 2. 6

Talanta 205 (2019) 120108

M. Lecoeur, et al.

Fig. 5. Specificity: Electropherograms obtained in optimized analytical conditions for (1) a blank urine (CAL00), (2) a calibration standard at the LLOQ and (3) a blank urine spiked with internal standards (CAL0). For APAP-G, APAP-M, APAP-S and APAP: 200-fold factor dilution of the blank urine; for APAP-GT, MT-APAP and APAP-M: 20-fold factor dilution of the blank urine. Analyte concentration: APAP-G (10 mg L−1), APAP-M (0.1 mg L−1), APAP-S (0.5 mg L−1), MT-APAP (0.1 mg L−1), M-APAP (0.025 mg L−1) or APAP (0.1 mg L−1) (LLOQ) and (3) a blank Separation conditions: pH 10 BGE (40 mM ammonium acetate adjsuted to pH 10 by addition of 1 M NH4OH); anodic injection (100 mbar, 20 s); 25 kV; 25 °C. * in-source fragmentation of APAP-G.

3.1.2. Optimization of the CE separation conditions 3.1.2.1. BGE pH and concentration. According to the acidic properties of acetaminophen and of its metabolites, it is obvious that the simplest separation strategy is to perform analysis at an alkaline pH, so that analytes are ionized, i.e. under anionic forms. While the non-volatile borax buffer is often used in CE-UV, when CE-MS is used, it must be replaced by a volatile buffer, such as the ammonium buffer. Heitmeier et al. [26] have preconized the use of pH 9.8 ammonium acetate buffer and hydrodynamic anodic injection. Separation experiments were then performed in order to determine the optimal pH and the concentration of ammonium acetate buffer to use as the BGE. In each case, a 10-fold diluted BGE solution spiked with acetaminophen and its metabolites at 3 mg L−1 was analysed after hydrodynamic anodic injection (100 mbar, 20 s) applying a 25 kV voltage. Separations were first carried out in buffer solutions with pH ranging from 8 to 10.5, i.e. around the pKa of phenolic group of

acetaminophen, in order to obtain both a strong EOF and the dissociation of the weakly acidic analytes. These buffers were prepared from a 40 mM ammonium acetate solution whose pH has been adjusted by the addition of a 1 M ammonium hydroxide solution. As shown in Fig. 3, when pH increases a marked decrease in the migration time is observed for APAP-S and APAP-G (30% of decrease) although ionization does not change in the pH range studied. This can only result from an increase in the EOF velocity despite no change in the ionic strength takes place. This phenomenon has already been observed by Grundmann et al. [32] during a CE-MS analysis of hyaluronan oligosaccharides (20% decrease in migration times between pH 8 and 9). These molecules with carboxylic acid groups whose pKa vary between 3 and 4 therefore exist in 100% anionic form in the corresponding pH range. This phenomenon was unfortunately not explained in this study. According to the origin of the EOF, it can be assumed that the addition of increasing amounts of ammonia to reach concentration between 2.5 7

Talanta 205 (2019) 120108

M. Lecoeur, et al.

Fig. 6. Accuracy profiles of CE-MS method for the quantitation of (1) APAP, (2) APAP-S, (3) APAP-G, (4) APAP-M, (5) MT-APAP and (6) M-APAP. The dashed lines represent the acceptance limits at 20%, the red plain line corresponds to the bias and the tolerance interval of the bias for a risk of 5% was materialized by dotted blue line. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

and 800 mM corresponding to the desired pH, although it does not modify the ionic strength (I = 40 mM), would result in a change in the ratio of the dielectric constant to the viscosity of the separating electrolyte as when using an organic modifier. According to Duong et al. [33], the dielectric constant of aqueous ammonia solution decreases slightly as the ammonia concentration increases (decrease of about 5%, for the concentration range studied). However, Fernandes et al. demonstrated that above 5.7 mM, the viscosity of an ammonia solution is inversely proportional to the concentration [34]. Therefore, it seems that the change in the buffer viscosity prevails on those of dielectric constant, leading to an increase in EOF. For APAP, M-APAP, MT-APAP, APAP-M and APAP-GT, which ionization degree increase along with pH, the increase in the EOF velocity is probably compensated by an increase in the counterflow

electrophoretic mobility of the anions, resulting in an insignificant change in their migration time. To obtain the best separation, a pH 10 BGE was in the end chosen. Without co-migration of the analyte, ionic suppression is also avoided. In a second step, the influence of ammonium acetate concentration (pH 10) was studied in the range 20–60 mM. While an increase in the concentration from 20 to 30 mM led to a mean increase in the migration times of 16%, no additional increase was observed for higher concentrations and separation was unchanged. A concentration of 40 mM was finally chosen as a compromise between a high buffer capacity and a low current (25 μA) flowing the CE system. In these conditions, migration times of all analytes were between 5.9 and 9.4 min, APAP-GT exhibiting the lowest apparent mobility as shown in Fig. 3. 8

Talanta 205 (2019) 120108

M. Lecoeur, et al.

Fig. 7. Electropherograms of an urine sample collected 72 h after the first APAP administration, for a patient belonging to the G2 inclusion group of the clinical trial. For APAP-G, APAP-M, APAP-S and APAP: 200-fold factor dilution; for APAP-GT, MT-APAP and APAP-M: 20-fold factor dilution. Separation conditions: pH 10 BGE (40 mM ammonium acetate + 1 M NH4OH); anodic injection (100 mbar, 20 s); 25 kV; 25 °C. * in-source fragmentation of APAP-G; + non identified compounds.

3.1.3. Optimization of the sample preparation method 3.1.3.1. Dilution solvent. Despite several authors have demonstrated that the direct injection of urine samples in CE-MS is possible for separation and detection of APAP and its metabolites [26,27], no data relative to their quantitation are available. In order to avoid ion suppression due to the presence of high concentrations of salt in urine samples, blank urine samples spiked with analytes and their internal standards at 1 mg L−1 were analysed after a 10-fold dilution in 4 mM ammonium acetate solution (10-fold diluted BGE, pH 10) to sharpen peak by stacking phenomenon. Unfortunately, this sample preparation led to a splitting of the APAP peak (Fig. 4a). This phenomenon was already observed by Heitmeier et al. [26] when APAP was dissolved in water while borax BGE (50 mM, pH 9.4) was used for the separation. To avoid this splitting due to the existence of zone boundaries, blank urine samples spiked with analytes and their internal standards was then performed after a 10-fold dilution directly in the BGE. Efficacy and peak symmetry were improved (3-fold factor) and APAP peak splitting has disappeared (Fig. 4). In this case, the higher concentration of the BGE probably permits to set the ionic strength of the sample, whatever the initial salt concentration of the urine sample. A direct dilution of the urine samples in the BGE was finally selected.

the dynamic range. Additionally, the highest the dilution factor is, the smallest the quantity of urine species introduced in the MS-MS system is. In these conditions, the fouling of the system is limited, as well as the ion suppression occurring during the analysis. In order to select the dilution factor to use, urine samples of three different patients per inclusion group, collected 24 h, 72 h and 100 h after the first APAP administration, were analysed after 200-fold, 100fold, 50-fold, 20-fold and 10-fold dilutions in the BGE and internal standard addition. According to quantitation limits and dynamic ranges established for all analytes during preliminary experiments, a 200-fold dilution factor was finally selected for the analysis of APAP, APAP-M, APAP-S and APAP-G. A 20-fold dilution factor was chosen for M-APAP and MT-APAP. As APAP-GT was not observed in any of the 10-fold diluted patient urine samples analysed (estimated LOD = 0.1 mg L−1), it was consequently discarded of the study. 3.2. Method validation Firstly, specificity of the method with respect to APAP, APAP-S, APAP-M, APAP-G, M-APAP and MT-APAP was assessed by comparing electropherograms obtained for blank urine samples diluted in the BGE (CAL00), a blank urine sample spiked with internal standards (CAL0) and a blank urine sample spiked with the analytes at the previously estimated LLOQ and their respective internal standards at concentrations used for samples analysis (Fig. 5). Absence of interfering components demonstrated the specificity of the method. It is important to

3.1.3.2. Dilution factor. Ideally, the dilution factor to apply must be the same for all analytes and for all patients. It must be such as analyte concentration is higher than the limit of quantitation and included in 9

Talanta 205 (2019) 120108

M. Lecoeur, et al.

Fig. 8. Concentration of acetaminophen (APAP), acetaminophen mercapturate (APAP-M), acetaminophen sulfate (APAP-S) and acetaminophen glucuronide (APAPG) in urine collected 72 h after the first APAP administration, for patients in G1, G2 or G3 inclusion groups of the clinical trial.

underline that electropherogram recorded with the specific MRM transition of APAP (152 - > 110) exhibits two peaks (tM = 6.4 and 6.8 min). The first one corresponds to native APAP, while the second one with the same migration time as APAP-G is related to its ion-source fragmentation. Calibration curves, expressing the peak area ratio (analyte/IS) versus concentration ratio (analyte/EI), were established from six calibration standards either in 200-fold diluted blank urine samples for APAP, APAP-S, APAP-M and APAP-G or in 20-fold diluted blank urine samples for M-APAP and MT-APAP. They were used to assess the response functions using least square weighted (1/X) linear regression. Obtained results are summarized in Table 1. The back-calculated concentrations of the calibration standards were within ± 15% of the nominal values, even for the LLOQ, in accordance with EMA acceptance criteria. The validity of the model was attested by evaluation of the linearity of the method: whatever the analyte, the introduced concentrations and the back-calculated one can be expressed by linear models (R2 > 0.946) with slopes close to the unit and y-intercept that can be considered to be equal to zero (Student test, with α = 5%). Trueness was evaluated by comparing nominal and back-calculated concentrations and was expressed as relative bias. Repeatability and intermediate precision were expressed as the relative standard deviation (RSD), at each concentration level. A shown in Table 1, relative bias varied between ± 10% except for the validation standards exhibiting the lowest concentration in the case of APAP-G and MT-APAP, with relative bias within ± 20% of the nominal values. These results

are in accordance with EMA. For APAP, APAP-S, APAP-G, APAP-M and MT-APAP, repeatability, i.e. intra-day precision, ranged from 2.19% to 18.34% for the LLOQ and from 2.08% to 10.10% for others validation standards. Concerning the intermediate precision (inter-day precision), RSD values were less than 12% except for APAP-G and MT-APAP at the LLOQ, with values higher than the acceptance criteria of ± 20%. For M-APAP, both repeatability and intermediate precision were outside the ± 15% interval for low and medium validation standards (VS2 and VS1). These trueness and precision results were used to calculate the upper and the lower confidence limits for validation standards required to establish the accuracy profiles (Fig. 6). Whatever the concentration level, the tolerance interval (β = 5%) is included in the acceptance limits set at ± 20% for APAP, APAP-S, APAP-M and M-APAP. It is worth noting that for M-APAP, while intermediate precision was too high for VS2 and VS3 to fulfil EMA criteria, bias was quite low to obtain a total error included in the acceptance limits, guaranteeing that the error will be less than 20% in 95% of cases. The lower limit of quantitation (LLOQ) is equal to 0.1 mg L−1 for APAP and APAP-M, to 0.5 mg L−1 for APAP-S and to 0.025 mg L−1 for M-APAP. For APAP-G and MT-APAP, the accuracy profiles exhibit a tolerance interval excluded from the acceptance limits for the lowest concentrations. Hence, for each of these analytes, the LLOQ can be calculated by taking intersection point between the acceptability limit and the corresponding tolerance interval; they are equal to 9.22 and 0.073 mg L−1 for APAP-G and MT-APAP, respectively. 10

Talanta 205 (2019) 120108

M. Lecoeur, et al.

Carryover was estimated by injecting blank samples after a mixture of the analytes (concentration equal to upper limit of quantification) and internal standards. The rinsing protocol consisting in the following multistep sequence: 0.1 M NH4OH solution (5 min, 950 mbar), water (5 min, 950 mbar) and finally pH 10 BGE (5 min, 950 mbar), was suitable since the peaks areas obtained for the blank sample following the high concentration standard were less than 10% of the LLOQ for the analytes and less than 5% for the internal standard. This result is in accordance with EMA criteria. Analyte stock solutions appeared stable for six months of storage at −20 °C as assessed by a deviation from the initial concentration within ± 15%. APAP and its metabolites were shown to be adequately stable in urine for 24 h at room temperature (deviation within ± 22%), at freezing temperature (−20 °C) for 6 months (deviation within ± 17%) and after three freeze-thaw cycles (deviation within ± 16%).

approved by the Ethics Committee (CPP N°2016-002632-32) and was authorized by the competent authorities (ANSM N°1060661A-42; France). References [1] H. Kehlet, H. Holte, Effect of postoperative analgesia on surgical outcome, J. Anaest. 87 (2001) 62–72 https://doi.org/10.1093/bja/87.1.62. [2] O. Mimoz, P. Incagnoli, C. Josse, M.C. Gilllon, L. Kuhlman, A. Mirard, H. Soileux, D. Fletcher, Analgesic efficacy and safety of nefopam vs. Propacetamol following hepatic resection, Anaesthesia 56 (2001) 520–525 https://doi.org/10.1046/j.13652044.2001.01980.x. [3] M. Seifari, A. Iten, A. Hadengue, Acetominophen: hepatotoxicity at therapeutic dose and risk factors, Rev. Med. Suisse 3 (2007) 2345–2349. [4] M. Van de Poll, C. Dejong, M. Fischer, A. Bats, G. Koek, Decreased hepotosplanchnic antioxidant uptake during hepatic ischaemia/reperfusion in patients undergoing liver resection, Clin. Sci. 114 (2008) 553–560 https://doi.org/10.1042/ CS20070317. [5] M. Kretzschmar, A. Krüger, W. Schirrmeister, Hepatic ischemia-reperfusion syndrome after partial liver resection: hepatic venous oxygen saturation, enzyme pattern, reduced and oxidized glutathione, procacitonin and interleukine-6, Exp. Toxicol. Pathol. 54 (2003) 423–431 https://doi.org/10.1078/0940-2993-00291. [6] A. Bertolini, A. Ferrari, A. Ottani, S. Guerzoni, R. Tacchi, S. Leone, Paracetamol: new vistas of an old drug, CNS Drug Rev. 12 (2006) 250–275 https://doi.org/10. 1111/j.1527-3458.2006.00250.x. [7] M. Black, Acetominophen hepatotoxicity, Ann. Rev. med. 35 (1984) 577–593 https://doi.org/10.1146/annurev.me.35.020184.003045. [8] M. McGill, H. Jaeschke, Metabolism and disposition of acetaminophen: recent advances in relation to hepatotoxicity and diagnosis, Pharma Res. 30 (2013) 2174–2187 https://doi.org/10.1007/s11095-013-1007-6. [9] J.M. Park, Y.S. Lin, J.C. Calamia, K.E. Thummel, J.T. Slattery, T.F. Kalhorn, R.L. Carithers, A.E. Levy, C.L. Marsh, M.F. Hebert, Transiently altered acetaminophen metabolism after liver transplantation, Clin. Pharmacol. Ther. 73 (2003) 545–553 https://doi.org/10.1016/S0009-9236(03)00062-6. [10] M. Galinski, B. Delhotal-Landes, D.J. Lockey, J. Rouaud, S. Bah, A.E. Bossard, F. Lapostolle, M. Chauvin, F. Adnet, Reduction of paracetamol metabolism after hepatic resection, Pharmacology 77 (2006) 161–165 https://doi.org/10.1159/ 000094459. [11] M. Galinski, S.X. Racine, A.E. Bossard, M. Fleyfel, L. Hamza, N. Bouchemal, F. Adnet, L. Le Moyec, Detection and Follow-up, after partial liver resection, of the urinary paracetamol metabolites by proton NMR spectroscopy, Pharmacology 93 (2014) 18–23 https://doi.org/10.1159/000357095. [12] A. Hartog, G. Mills, Anaesthesia for hepatic surgery, Cont. Educ. Anaesth. Crit. Care Pain 9 (2009) 1–5 https://doi.org/10.1093/bjaceaccp/mkn050. [13] C. Vezinet, D. Eyraud, E. Savier, Anesthésie pour chirurgie hépatique, Prat. Anesthésie Réanimation 13 (2009) 418–428 https://doi.org/10.1016/j.pratan. 2009.11.001. [14] R. Stümpfle, A. Riga, R. Deshpande, S. Singh Mudan, R. RaoBaikady, Anaesthesia for metastatic liver resection surgery, Curr. Anaesth. Crit. Care 20 (2009) 3–7 https://doi.org/10.1016/j.cacc.2008.10.009. [15] A. Di Giralamo, W.M. O'Neill, I.W. Wainer, A validated method for the determination of paracetamol ans its glucuronide and sulfate metabolites in the urine of HIV +/AIDS patients using wavelength-switching UV detection, J. Pharm. Biomed. Anal. 17 (1998) 1191–1197 https://doi.org/10.1016/1.cacc.2008.10.009. [16] L.S. Jensen, J. Valentine, R.W. Milne, A.M. Evans, The quantification of paracetamol, paracetamol glucuronide and paracetamol sulfate in plasma and urine using single high-performance liquid chromatography assay, J. Pharm. Biomed. Anal. 34 (2004) 585–593 https://doi.org/10.1016/S0731-7085(03)00573-9. [17] I. Baranowska, A. Wilczek, Simultaneous RP-HPLS determination of sotalol, metoprolol, a-Hydroxymetoprolol, paracetamol and its glucuronide and sulfates metabolites in human urine, Anal. Sci. 25 (2009) 769–772 https://doi.org/10.2116/ analsci.25.769. [18] D. Howie, P.I. Adriaenssens, L.F. Prescott, Paracetamol metabolism following overdosage: application of high performance liquid chromatography, J. Pharm. Pharmacol. 29 (1977) 235–237 https://doi.org/10.1111/j.2042-7158.1977. tb11295.x. [19] D. Reith, N.J. Medlicott, R. Kumara De Silva, L. Yang, J. Hickling, M. Zacharias, Simultaneous modelling of the Michaelis-Menten kinetics of paracetamol sulphation and glucuronidation, Clin. Exp. Pharmacol. Physiol. 36 (2009) 35–42 https:// doi.org/10.1111/j.1440-1681.2008.05029.x. [20] A.K. Hewavitharana, S. Lee, P.A. Dawson, D. Markovich, P.N. Shaw, Development of an HPLC-MS/MS method for the selective determination of paracetamol metabolites in mouse urine, Anal. Biochem. 374 (2008) 106–111 https://doi.org/10. 1016/j.ab.2007.11.011. [21] Q. Tan, R. Zhu, H. Li, F. Wang, M. Yan, L. Dai, Simultaneous quantitative determination of paracetamol and its glucuronide conjugate in human plasma and urine by liquid chromatography coupled to electrospray tandem mass spectrometry : application to a pharmacokinetic study, J. Chromatogr. B 893–894 (2012) 162–167 https://doi.org/10.1016/j.jchromb.2012.02.027. [22] S. Cook, A. King, J. Anker, D. Wilkins, Simultaneous quantification of acetaminophen and five acetaminophen metabolites in human plasma and urine by high-performance liquid chromatography-electrospray ionization –tandem mass spectrometry: method validation and application to a neonatal pharmacokinetic study, J. Chromatogr. B 1007 (2015) 30–42 https://doi.org/10.1016/j.jchromb.

3.3. Application of the CE-MS/MS method to the quantitation of acetaminophen and its metabolites in patient's urine The analysis method was finally used for quantitation of APAP and its metabolites in the urine of eighteen patients belonging respectively to the three inclusion groups of the ongoing clinical trial (G1: major hepatic surgery; G2: minor hepatic resection; G3: hepatic re-operation; six patients per group). For example, Fig. 7 shows the electropherograms obtained for one of these samples. Whatever the patient's inclusion group, both concentrations of M-APAP and MT-APAP were between 0.30 and 0.50 mg L−1 and between 0.12 and 1.46 mg L−1, respectively, i.e. between LOD and LLOQ determined previously for these analytes, when 20-fold dilution factor is taken into account. For APAP, APAP-M, APAP-S and APAP-G, the analysis method was successfully applied to assess their concentrations. For example, Fig. 8 shows box plots established from concentrations measured for APAP, APAP-M and, APAP-S and APAP-G. The differences in concentrations observed according to the membership of the different inclusion groups, show that the method developed will be sufficiently discriminating to obtain relevant information on metabolism after liver resection. Indeed, for these patients, as illustrated in Fig. 8, it seems that APAP concentrations were quite similar for all groups, while APAP-M, APAP-S and APAP-G concentrations seem to be of different magnitude when a hepatic resection was performed instead of a minor resection or a resurgery. This will be confirmed using all the data that will be obtained after inclusion of all patients in the clinical trial. 4. Conclusion A CE-MS/MS method was for the first time developed and validated for the quantitation of APAP and five of its metabolites in urine samples. It was shown that the dilution solvent and dilution factor to use for sample preparation are critical parameters to avoid peak splitting, to gain in sensitivity and then to obtain an effective analysis method. Hyphenation of CE with MS/MS versus UV permits to improve LOQ (10–20-fold factor with respect to previous works for APAP, APAP-S and APAP-G). Use of CE versus HPLC, permits to quantify two additional metabolites, i.e. MT-APAP and M-APAP. This method was successfully applied to the analysis of eighteen patients’ urines belonging to three inclusion groups of a clinical trial. From these preliminary experiments, it seems that differences in the distribution of APAP and its urinary metabolites versus inclusion group occur. The CE-MS/MS method will be further use for the analysis of urines samples of the ninety patients included in the clinical trial and collected for five days. Acknowledgments The study was supported by the University Hospital Center (CHU) of Lille (France). As required by the french law, the study protocol was 11

Talanta 205 (2019) 120108

M. Lecoeur, et al. 2015.10.013. [23] A.D.B. Vliegenthart, R.A. Kimmitt, J.H. Seymour, N.Z. Homer, J.I. Clarke, J.I. Eddleston, A. Gray, D.M. Wood, P.I. Dargan, J.G. Cooper, D.J. Antoine, D.J. Webb, S.C. Lewis, D.N. Bateman, J.W. Dear, Circulating acetaminophen metabolites are toxicokinetics biomarkers of acute liver injury, Clin. Pharamacol. Ther. 101 (2017) 531–540 https://doi.org/10.1002/cpt.541. [24] D. Tonoli, E. Varesio, G. Hopfgartener, Quantification of acetaminophen and two of its metabolites in human plasma by ultra-high performance liquid chromatographylow and high resolution tandem mass spectrometry, J. Chromatogr. B 904 (2012) 42–50 https://doi.org/10.1016/j.jchromb.2012.07.009. [25] R. Flint, P. Mian, B. Van der Nagel, N. Slijkhuis, B. Koch, Quantification of acetominiphen ant its metabolites in plasma using UHPLC-MS: doors open to therapeutic drug monitoring in special patient populations, Ther. Drug Monit. 39 (2017) 164–171 https://doi.org/10.1097/FTD.0000000000000379. [26] S. Heitmeier, G. Blaschke, Direct determination of paractetamol and its metabolites in urine and serum by capillary electrophoresis with ultraviolet and mass spectrometry detection, J. Chromatogr. B 721 (1999) 93–108 https://doi.org/10.1016/ S0378-4347(98)00415-0. [27] S. Ullsten, R. Danielsson, D. Bäckerström, P. Sjöberg, J. Bergquist, Urine profiling using capillary electrophoresis-mass spectrometry and multivariate data analysis, J. Chromatogr. A 1117 (2006) 87–93 https://doi.org/10.1016/j.chroma.2006.03.048. [28] P. Hubert, J.J. Nguyen-Hu, B. Boulanger, E. Chapuzet, P. Chiap, N. Cohen, P.A. Compagnon, W. Dewé, M. Feinberg, M. Lallier, M. Laurentie, N. Mercier, G. Muzard, C. Nivet, L. Valat, Harmonization of strategies for the validation of quantitative analytical procedures: a SFSTP proposal—part I, J. Pharm. Biomed.

Anal. 36 (2004) 579–586 https://doi.org/10.1016/j.jpba.2004.07.027. [29] P. Hubert, J.J. Nguyen-Hu, B. Boulanger, E. Chapuzet, P. Chiap, N. Cohen, P.A. Compagnon, W. Dewé, M. Feinberg, M. Lallier, M. Laurentie, N. Mercier, G. Muzard, C. Nivet, L. Valat, E. Rozet, Harmonization of strategies for the validation of quantitative analytical procedures: a SFSTP proposal – Part II, J. Pharm. Biomed. Anal. 45 (2007) 70–81 https://doi.org/10.1016/j.jpba.2007.06.013. [30] P. Hubert, J.J. Nguyen-Hu, B. Boulanger, E. Chapuzet, N. Cohen, P.A. Compagnon, W. Dewé, M. Feinberg, M. Laurentie, N. Mercier, G. Muzard, L. Valat, E. Rozet, Harmonization of strategies for the validation of quantitative analytical procedures: a SFSTP proposal–Part III, J. Pharm. Biomed. Anal. 45 (2007) 82–96 https://doi. org/10.1016/j.jpba.2007.06.032. [31] Guideline on Bioanalytical Method Validation, European Medicines Agency, Committe for Medicinal Products for Human Use, (2011) United Kingdom, London https://www.ema.europa.eu/documents/scientific-guideline/guidelinebioanalytical-method-validation_en.pdf , Accessed date: 22 January 2019. [32] M. Grundmann, M. Rothenhöfer, G. Bernhardt, A. Buschauer, F.M. Matysik, Fats counter-electroosmotic capillary electrophoresis-time-of-flight mass spectrometry of hyaluronan oligosaccharides, Anal. Bioanal. Chem. 402 (2012) 2617–2623 https://doi.org/10.1007/s00216-011-5254-2. [33] D. Duong, The Complex Dielectric Properties of Aqueous Ammonia from 2 GHz – 8.5 GHz in support of the Nasa Juna Mission, Masters’thesis, Georgia Institute of Technology (2011). [34] B.L. Fernandes, J.R. Figueiredo, Modeling viscosity of liquid ammonia-water mixture, Revi. Ciênc. Tecnol. 15 (2012) 51–57.

12