Accepted Manuscript Title: Determination of low levels of benzodiazepines and their metabolites in urine by hollow-fiber liquid-phase microextraction (LPME) and gas chromatography-mass spectrometry (GC-MS) Author: Andr´e Valle de Bairros Rafael Menck de Almeida Lorena Pantale˜ao Thiago Barcellos Sidnei Moura e Silva Mauricio Yonamine PII: DOI: Reference:

S1570-0232(14)00681-3 http://dx.doi.org/doi:10.1016/j.jchromb.2014.10.040 CHROMB 19189

To appear in:

Journal of Chromatography B

Received date: Revised date: Accepted date:

15-5-2014 24-10-2014 29-10-2014

Please cite this article as: A.V. Bairros, R.M. Almeida, L. Pantale˜ao, T. Barcellos, S.M. Silva, M. Yonamine, Determination of low levels of benzodiazepines and their metabolites in urine by hollow-fiber liquid-phase microextraction (LPME) and gas chromatography-mass spectrometry (GC-MS), Journal of Chromatography B (2014), http://dx.doi.org/10.1016/j.jchromb.2014.10.040 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 proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Determination of low levels of benzodiazepines and their metabolites in urine by hollow-fiber liquid-phase microextraction (LPME) and gas chromatography-mass spectrometry (GC-MS) André Valle de Bairrosab*, Rafael Menck de Almeidaa, Lorena Pantaleãoa, Thiago Barcellosc,

Faculty of Pharmaceutical Sciences, University of São Paulo, Brazil

b

Center of Chemical, Pharmaceutical and Food Sciences, Federal University of Pelotas, Brazil

Institute of Biotechnology, University of Caxias do Sul, Brazil

us

c

cr

a

ip t

Sidnei Moura e Silvac and Mauricio Yonaminea

an

*Corresponding author. Postal address: Av. Professor Lineu Prestes, 580, bloco 13B CEP: 05508-900, São Paulo, SP, Brazil. Fax: +55 11 3031 9055.

M

E-mail address: [email protected]; [email protected]

ed

ABSTRACT

In this study, it is shown a method for the determination of benzodiazepines and their main metabolites in urine samples by hollow-fiber liquid-phase microextraction (LPME) in the

pt

three-phase mode. Initially, the hydrolysis step was performed using 100 µL of sodium acetate 2.0 mol/L buffer solution (pH 4.5), 25 µL of β-glucuronidase enzyme and incubation

Ac ce

for 90 min at 55°C. In parallel with hydrolysis, the LPME fiber (9 cm) was prepared. Its pores were filled with a mixture of dihexyl ether: 1-nonanol (9:1). Afterwards, a solution of 3.0 mol/L of HCl was introduced into the lumen of the fiber (acceptor phase). After hydrolysis, the fiber was submersed in the alkalinized urine (pH 10) containing 10% NaCl. Samples were then submitted to orbital shaking (2400 rpm) for 90 min. The acceptor phase was later withdrawn from the fiber, dried and the residue derivatized with trifluoroacetic anhydride (TFAA)

for

10

minutes

at

butyldimethylsilyltrifluoroacetamide

60°C

with

containing

further 1%

addition

of

N-methyl-N-tert-

tert-butyldimethylchlorosilane

(MTBSTFA) for 45 minutes at 90°C followed by determination by gas chromatography–mass spectrometry (GC–MS). The calibration curves obtained showed linearity over the specified range, with a similar sensitivity to traditional techniques and a higher detection capability compared to most of the miniaturized methods described in the literature. The method has

Page 1 of 32

been developed and successfully validated and applied to urine samples from real cases of benzodiazepines intake. Keywords: Benzodiazepines; Gas chromatography; Mass spectrometry; Urine; Liquid phase

Ac ce

pt

ed

M

an

us

cr

ip t

microextraction; Double derivatization.

Page 2 of 32

1. Introduction Benzodiazepines are central nervous system (CNS) depressant drugs often prescribed for the treatment of insomnia, anxiety, and epilepsy. However, this class of drugs can promote both tolerance and dependence and may also have an impact on the performance of given situations such as driving a vehicle or the ability to remain focused at work [1,2]. Despite the

ip t

distribution control of benzodiazepines in many countries, this class of drugs has been used in suicide attempts [3,4]. Because of its pharmacological effects, high commercialization and relative ease in obtaining such substances, they are also used in many countries in drug-

cr

facilitated crimes (DFC). Diazepam, clonazepam and flunitrazepam were the most frequently

us

benzodiazepines found in biological samples of DFC victims [5-8]. In these cases, benzodiazepines and their metabolites are the target analytes according to the Society of Forensic Toxicologists (SOFT) and United National Office on Drugs and Crime (UNODC)

an

[9,10]. One of the analytical features required by these organizations is the sensitivity to determine very low concentrations of benzodiazepines and its metabolites in biological fluids

M

when DFC is suspected.

A number of studies have been reported in the literature on the determination of

ed

benzodiazepines and their metabolites in biological fluids. Liquid and gas chromatography (LC and GC, respectively) are the main techniques to determinate these molecules, while liquid-liquid extraction (LLE) and solid phase extraction (SPE) are the most frequently used

pt

sample preparation techniques [11-15]. However, LLE and SPE require costly consumables such as solvents and cartridges. Having this in mind, miniaturized methods such as hollow-

Ac ce

fiber liquid-phase microextraction (LPME) could be an interesting alternative for the extraction of benzodiazepines. However, this technique has been scarcely explored and is still seen as a challenge [16-18].

Hollow-fiber liquid-phase microextraction (LPME) is a relatively new technique and can

be performed in a two-phase or three-phase mode. In a two-phase extraction system, the hydrophobic solvent is immobilized as a thin supported liquid membrane (SLM) into the pores of a porous hollow fiber. The lumen of the fiber is also filled with hydrophobic solvent (acceptor phase) and the system is placed in contact with the sample (donor phase). The analytes are extracted from the sample (aqueous phase), through the SLM (organic phase) and finally into the acceptor phase. In the case of a three-phase extraction system, the analytes must be ionized with an aqueous solution (acceptor phase) inside the lumen of the hollow

Page 3 of 32

fiber. Consequently, the analytes will get trapped in the acceptor phase. Due to the high sample-to-acceptor volume ratio, very high enrichments can be obtained by using LPME, especially in the three-phase mode. This technique provides a better recovery when compared with the two-phase described previously. It has also an easy drying process and subsequent derivatization of the analytes plus a chemical reaction that increases the selectivity and stability of the compounds. Therefore, an excellent clean-up has been reported from complex

ip t

biological matrices such as urine samples due to the size of the pores as they provide microfiltration of macromolecules [19-22].

cr

Urine samples continue to be widely used as a biological matrix for the analysis of

us

psychoactive substances in forensic cases because of the large volume of sample available for analysis, the relative simplicity of sample preparation and also its broad drug detection window compared to blood [9,10,23]. During the metabolism process of benzodiazepines up

an

to their excretion, these molecules undergo reactions of phase I followed by a second stage where there is an addition of a glucuronide group. The glucuronidation binds hydroxy-

M

benzodiazepines, transforming these compounds into more polar products and thus excreted by this route. Therefore, it is necessary to perform an hydrolysis step for drug extraction procedure for GC analysis [12,24]. In this case, enzymatic hydrolysis is preferable compared

ed

to chemical hydrolysis, avoiding benzodiazepines degradation which can generate benzophenones [12,24,25]. Although enzymatic hydrolysis allows free drug/metabolite in the

pt

urine sample, benzodiazepines metabolites present in this biological matrix are more polar than the parent drug, requiring a derivatization step when determined by GC-MS [26,27].

Ac ce

Silylation, acylation and alkylation derivatizing agents are the most commonly used in

the determinations of benzodiazepines and its metabolites by GC-MS [11,13,27,28]. In some circumstances, double derivatization (two steps) is used to reach the analytical purpose [13,29]. According to other studies, this procedure is the key to improve GC-MS selectivity and sensitivity over single-step procedures studies. This strategy is useful to avoid artifact formation, tailing peaks and to produce a better repeatability without the need for different procedures in sample preparation. Furthermore, there is an increase of sensitivity and a wider range of evaluated analytes [13,29-31]. The aim of the present study was to develop a sensible method for the determination of 11 benzodiazepines and their main metabolites (medazepam, chlordiazepoxide, diazepam, nordiazepam,

oxazepam,

lorazepam,

nitrazepam,

flunitrazepam,

clonazepam,

7-

Page 4 of 32

aminoflunitrazepam and 7-aminoclonazepam) in urine samples using hollow-fiber liquid phase microextraction (LPME) in three-phase mode and double derivatization for further detection by gas chromatography-mass spectrometry (GC–MS). The method has been fully validated and it was successfully applied in urine samples from real cases involved with

ip t

benzodiazepine exposure.

2. Experimental

cr

2.1. Chemicals

nitrazepam, 7-aminoflunitrazepam, (1.0

mg/mL)

and

lorazepam, clonazepam and 7-aminoclonazepam

internal

standards

diazepam-D5,

oxazepam-D5,

7-

an

solutions

us

Medazepam, diazepam, nordiazepam, flunitrazepam, chlordiazepoxide, oxazepam,

aminoflunitrazepam-D7, clonazepam-D4 and 7-aminoclonazepam-D4 were purchase from Cerilliant Analytical Reference Standards® (Round Rock, TX, USA). Dihexyl ether, 1-

M

nonanol, undecane, decanol, 1-octanol, xilol, β-glucuronidase from Helix pomatia type 2 in aqueous solution 100.000 units/mL, N-methyl-N-tert-butyldimethylsilyltrifluoroacetamide

ed

with 1% tert-butyldimethylchlorosilane (MTBSTFA), trifluoroacetic anhydride (TFAA) and ethyl acetate were purchased from Sigma-Aldrich® (MO, USA), while sodium hydroxide,

2.2. Instrumentation

pt

sodium acetate and hydrochloric acid were purchased from Merck® (Darmstadt, Germany).

Ac ce

Hollow-fiber Q3/2 Accurel KM polypropylene (600 µm i.d., 200 µm wall thickness and

0.2 µm pore size) was purchased from Membrana® (Wuppertal, Germany). Gel-loading pippete tips Round CC 4853 (0.5 mm; 1-200 µL) were purchased from Costar® (Corning, NY, USA). Extraction was performed using a multi-tube vortexer® model VWR VX-2500 (Thorofare, NJ, USA). The analyses was performed using an Agilent 6850 Network GC System gas chromatograph coupled with an Agilent® 5975 Series quadrupole mass seletive detector (MSD) (Wilmington, DE, USA). Samples were injected into the GC-MS by means of an autosampler (Agilent 7693). Injections were made using splitless mode (2 min and afterwards split vent was turned on in a ratio of 1:50). Chromatographic separation was achieved on a HP-5MS fused-silica capillary column (30 m x 0.25 mm x 0.1 µm film thickness) using helium as the carrier gas with 1.0 mL/min at a constant flow rate mode. The column oven temperature program was as follows: first held at 150°C (hold 1 min), then

Page 5 of 32

programed at 30°C/min to 220°C (hold 1 min); 20°C/min at 300°C (hold 3 min). The total analytical time was 11.33 min. Injection port and transfer line were set at 260°C and 280°C respectively. The MS was operated by electron ionization (70 eV) in selected ion monitoring (SIM) mode. 2.3. Optimization of the method

ip t

The study of method optimization was performed taking into consideration the choice of supported liquid membrane, pH of donor phase, influence of acceptor phase, time of

cr

extraction, shaking speed, salt addition on the extraction yield and two steps derivatization strategy. Fortified urine samples at a concentration of 50 ng/mL for each analyte were

us

submitted to the previous described method. The efficiency of extraction was evaluated by the absolute area produced by each analyte in all tested conditions. The following parameters

an

were studied: supported liquid membrane (dihexyl ether, xilol, 1-nonanol, decanol, 1-octanol and different proportions of the mixture between dihexyl ether and 1-nonanol); pH of donor phase (pH 8, 9, 10, 11, 12 and 13); acceptor phase (0.05, 0.1, 1.0, 3.0 and 5.0 mol/L HCl);

M

time of extraction (15, 30, 60, 90 and 120 min) and agitation (1200, 1560, 1800, 2040 and 2400 rpm). The salting out effect was also tested by adding 0, 5, 10 and 20% of NaCl (m/v) in

ed

the sample before extraction. All remaining parameters were fixed at a certain rate while changing a given variable of those described in section 2.5 (Sample preparation). For the study of double derivatization process, it was verified the influence of TFAA and MTBSTFA

pt

at different temperatures (TFAA - 60, 70 and 80°C; MTBSTFA – 70, 80 and 90°C) and incubation times (TFAA – 10, 20 and 30 minutes; MTBSTFA – 15, 30 and 45 minutes) for all

Ac ce

analytes.

2.4. Urine samples

Drug-free urine samples were obtained from volunteers and used for the validation of the

analytical method. The urine samples collected from real cases were obtained from the Vila Serena Treatment Center (São Paulo, Brazil) and stored at –20°C until further analysis. 2.5. Sample preparation An aliquot of 2.0 mL of urine was transferred into a 4 mL glass tube, followed by the addition of 200 ng of each internal standard (diazepam-D5, oxazepam-D5, 7aminoflunitrazepam-D7, clonazepam-D4 and 7-aminoclonazepam-D4). Initially, hydrolysis step was performed according to Meatherall [32] with minor adjustments. First, it was added

Page 6 of 32

100 µL of sodium acetate 2 mol/L buffer solution (pH 4.5), 25 µL of β-glucuronidase enzyme and incubated for 90 min at 55°C. During this process, the LPME fiber was prepared in parallel (9 cm), by immersion into the supported liquid membrane for about 30 seconds to impregnate the pores with organic solvent and subsequently sonicated in distilled water for 10 seconds using ultrasonic bath to remove the excess of solvent. The fiber lumen was filled with 75 µL of HCl aqueous solution (3 mol/L) using a gel-loading pipette tip where the first two

ip t

drops coming out of the hollow-fiber have been dismissed. Both ends of the fiber were sealed by mechanical pressure with a plier. After the hydrolysis step, the sample was alkalinized

cr

with 10% NaOH (m/v) aqueous solution (2 mol/L) until reaching a measure of pH 10. The previously treated fiber was then introduced in a U-shape into the sample solution. During

us

extraction, the assembly was submitted to shaking at 2400 rpm for 90 minutes using a multitube vortex. After extraction, the acceptor phase was withdrawn from the fiber and dried

an

under nitrogen stream at 40ºC. The first step derivatization was performed using 50 µL of TFAA at 60ºC for 10 min (after drying at 40°C under N2 stream), while the second derivatization was achieved by using 35 µL of MTBSTFA at 90ºC for 45 min in the same

M

vial. After cooling down, an aliquot of 2.0 µL was taken from this solution and injected into the GC–MS system.

ed

2.6. Validation of the method

Method validation was carried out by establishing several parameters as the limit of

pt

detection (LoD) and quantitation (LoQ), specificity/selectivity, linearity, intra and inter-assay accuracy, precision and recovery after optimization. All validation procedures were

Ac ce

performed in six replicates for each parameter evaluated. Further explanation on method validation can be seen below. 2.6.1. Limit of detection (LoD) and limit of quantification (LoQ) First of all, the limit of blank (LoB) was established to ensure a reliable LoD. The LoB is

the highest apparent concentration expected to be found when 60 replicates of 20 different blank samples containing no analytes are tested. The data is expressed in the following equation: LoB = mean

blank

+ 1.645 (standard deviation

blank).

LoD can be defined as the

lowest concentration of analyte in a sample which can be distinguished from the LoB with a significant reliability and at which detection is feasible. The procedure to determine LoD was performed in a similar way to that of LoB, by using low concentrations of the analyte spiked into negative urine samples and using the equation: LoD = LoB + 1.645 (standard deviation

Page 7 of 32

low concentration sample).

When a value of 1.645 is used as a standard deviation, no more than 5% of

the values should be less than the LoB. If the observed LoD sample values meet this criterion, then LoD is considered established or verified. The LoQ can be defined as the lowest concentration of a sample that can still be quantified with acceptable precision and accuracy. The acceptance criteria used were values of < 20% (RSD) for precision and accuracy between

ip t

80-120% [33,34]. 2.6.2. Selectivity/specificity

cr

Selectivity/specificity was evaluated by assessing ten different drug-free urine samples and 16 potential interfering drug compounds. Urine samples were extracted and analyzed

us

according to the previously described procedure for assessment of endogenous substances. Peaks at the retention time of interest were compared to those from urine samples spiked with

an

the analytes at the LOQ. The method was also evaluated for potential interfering substances through the analysis of blank urine samples spiked with 1000 ng/mL of amphetamine, methamphetamine,

3,4-methylenedioxy-N-methylamphetamine,

11-nor-9-carboxy-

M

tetrahydrocannabinol, benzoylecgonine, norketamine, morphine, codeine, acetylsalicylic acid, phenobarbital, nicotine, furosemide, clomipramine, lidocaine, naphazoline and caffeine.

ed

Acceptance criteria for this assay was based on the absence of interfering substances at the retention times of the analyte(s) of interest and their respective IS [34,35].

pt

2.6.3. Linearity

Linearity was established by analyzing urine samples containing all the analytes of

Ac ce

interest in the following concentrations: LoQ, 50, 100, 150, 200 and 250 ng/mL. Six replicates were analyzed at each concentration level. 2.6.4. Precision and accuracy Precision and accuracy studies were performed by analyzing urine samples containing

three different known concentrations of benzodiazepines and their metabolites during three consecutive days. The analyzes were performed in six replicates for each day. The lowest concentrations obtained were of 1.5 ng/mL (nordiazepam); 7.5 ng/mL (medazepam); 15 ng/mL (diazepam, chlordiazepoxide, oxazepam, nitrazepam and 7-aminoclonazepam); 30 ng/mL

(clonazepam

and

7-aminoflunitrazepam)

and

90

ng/mL

(lorazepam

and

flunitrazepam). Additional concentrations of 130 and 235 ng/mL were also measured for all analytes. Precision, defined as the relative standard deviation (RSD), was determined by intra-

Page 8 of 32

and inter-day replications. Experimental concentrations were obtained using the standard calibration curve. Accuracy was expressed as a percentage of the known concentration, i.e., the acquired mean concentration/nominal concentration × 100. 2.6.5. Recovery Recovery studies were performed by preparing two sets of urine samples with the same

ip t

concentrations for this study. One set of samples (set A), consisting of three concentrations for each analyte was analyzed in six replicates for each concentration according to the method

cr

described in Section 2.5. The lowest acquired concentrations were 1.5 ng/mL (nordiazepam); 7.5 ng/mL (medazepam); 15 ng/mL (diazepam, chlordiazepoxide, oxazepam, nitrazepam and

us

7-aminoclonazepam); 30 ng/mL (clonazepam and 7-aminoflunitrazepam) and 90 ng/mL (lorazepam and flunitrazepam). Further concentrations of 130 and 235 ng/mL were also

an

studied for all analytes.

The second set (set B) also comprised of samples in six replicates for each

M

analyte/concentration as previously described for set A. However, in set B, the analytes were spiked into the samples immediately after the LPME extraction procedure. Absolute recovery was evaluated by comparison of the mean response obtained for both set A (processed) and B

pt

2.6.6. Dilution integrity

ed

(unprocessed). The unprocessed response represented 100% recovery.

Dilution integrity is a parameter that allows evaluation of samples with analyte levels

Ac ce

above the calibration curve. When dilution integrity is required, it should be demonstrated by means of spiking the matrix with analytes above the highest concentration of the curve and then diluting this sample with a blank matrix (at least five determinations per dilution factor). Accuracy and precision should be within the set criteria, i.e. within ±15% [36]. 2.6.7. Stability

Stability of the analytes in urine samples were evaluated at the concentration of 100 ng/mL in three different conditions: autosampler (21°C/12 and 24h), 4°C/24h and -20°C/24h. Stability tests were performed in triplicate. 2.7. Application to real cases

Page 9 of 32

The analytical method was applied in real cases that involved benzodiazepine intake. The study protocol was previously reviewed and approved by the Faculty of Pharmaceutical Sciences Ethics Committee, University of Sao Paulo, Brazil (Ethics Protocol Approval n° CEP 98156). The quantification was based on the ratios of the ion peak areas of the compounds to the IS ion peak areas. The calibration curves were used to determine the

ip t

concentrations of benzodiazepines and their metabolites.

cr

3. Results and Discussion

us

3.1. Derivatization strategy

TFAA and MTBSTFA were used for the derivatization of benzodiazepines and their

an

metabolites. MTBSTFA promotes tert-butyldimethylsilyl (t-BDMS) derivatives of the benzodiazepines tested, except for 7-aminoflunitrazepam (diazepam, medazepam and flunitrazepam are molecules that do not suffer derivatization). Consequently, it was not

M

possible to detect the levels indicated by SOFT and UNODC (5 ng/mL) [9,10] of this analyte in a pool of benzodiazepines. Hence, double derivatization was performed to increase the

ed

capacity of detection for 7-aminoflunitrazepam in a benzodiazepines mixture. First, an acylation reaction using TFAA for all benzodiazepines studied found its

pt

optimum conditions at 60°C for 10 minutes. Only 7-aminoflunitrazepam and 7aminoclonazepam were able to react with TFAA under such conditions. Aromatic primary

Ac ce

amine group present on the 7-aminoflunitrazepam and 7-aminoclonazepam attacks the carbonyl group of the acylating reagent by an addition-elimination pathway, producing the monotrifluoroacetylated products in the conditions tested. It is possible that the formation of an amide group after the reaction with TFAA molecule conceives nitrogen less nucleophilic due to the deactivation caused by the monotrifluoroacetylated group. This condition allows only one molecule of TFAA to be successful in this reaction, as it was shown by Lin & Lua [37]. Indeed, high reactivity of TFAA allows a decrease in the time of reaction. Therefore this inexpensive reagent which is also stable at room temperature is widely used in toxicology laboratories [26,28]. All these factors represent its use as an advantage in comparison with others acylating reagents. According to Elian [29], Gunnar and co-authors [11], MTBSTFA showed greater sensitivity and stability of carried-out product when compared with other silylating agents.

Page 10 of 32

Because of these characteristics, MTBSTFA was used in the second step. The best conditions observed in this study were 90°C for 45 minutes for all benzodiazepines studied. Silicon atom is an element that has affinity for electronegative atoms and can explain why this reaction does not occur in the primary amine [38]. It is possible that monotrifluoroacetylated products from the first reaction with TFAA express a carbonyl group that suffers a strong inductive effect arising from the CF3 group. Thus, the oxygen atom forms a carbonyl group which binds

ip t

to the silicon atom, forming a complex in a first step reaction and then an electronic rearrangement occurs and the silicon atom is attacked by the nitrogen from the amide group,

cr

similarly to a mechanism of bimolecular nucleophilic substitution (SN2). This condition allows the silylation reaction to occur in the absence of a base as catalyst due to the nature of

us

the leaving group in the MTBSTFA as described in the literature [11,26,39].

In order to confirm the double derivatization of 7-aminoflunitrazepam and 7-

an

aminoclonazepam, a mass spectrum and its respective fragmentations were obtained using a GC-MS with different ionizations energies (5, 15 and 70eV). As shown in the Figure 1, by

M

decreasing the collision energy, there is an increase in the molecular ion signal, confirming the conversion of 7-aminoclonazepam [609 m/z and isotopic abundance (100.0%), 611 m/z (32.0%), 610 m/z (31.4%) and 612 m/z (10.0%)] and 7-aminoflunitrazepam [493 m/z and abundance

(100.0%),

494

ed

isotopic

m/z

(26.0%),

495

m/z

(3.3%)]

in

trifluoroacetamides/silylated derivatives.

pt

This technique of two steps derivatization proposed in the current study allowed the achievement of very specific mass spectra for both 7-aminoflunitrazepam and 7-

Ac ce

aminoclonazepam, enhancing selectivity and sensitivity of aminobenzodiazepines, which are considered as the most problematical metabolites of this class of drugs [40-42]. This is relevant for identification purposes from unknown samples such as cases of chemical submission. Other benzodiazepines studied were not negatively affected by double derivatization. A given set of ions were chosen for SIM analysis and its respective retention times, this can be seen in Table 1. 3.2. Sample preparation The majority of methods developed to detect benzodiazepines and their respective metabolites in urine samples comprise of conventional liquid-liquid extraction (LLE) or solid phase extraction (SPE) as sample preparation techniques previous to chromatographic analysis. However, these techniques require relatively large volumes of organic solvents that

Page 11 of 32

can be toxic to the analyst and hazardous to the environment. For this reason, miniaturized techniques have adopted a procedure of using little or no organic solvent in analytical procedures [16-18,43-45]. Solid phase microextraction (SPME), dispersive liquid-liquid microextraction (DLLME) and microextraction in packed sorbents (MEPS) have been related to the measurement of benzodiazepines and their metabolites in urine samples (Table 2). However, these techniques

ip t

comprise of some disadvantages such as high cost (SPME and MEPS), fragility of fiber (SPME), low absolute recovery values (1-10%) (SPME), possible carry-over effect (SPME

cr

and MEPS), traditional problems found in SPE protocols (MEPS), wide use of halogenated solvents (DLLME), proper formation of cloudy solution (DLMME), unsuitable interaction

us

between complex matrices and solvents, and the requirement for additional steps in the preparation of biological samples (DLMME) [20,45,56].

an

In this context, LPME showed to be an interesting alternative for the microextraction of drugs and metabolites in biological samples. This technique provides high recovery values, excellent clean-up of endogenous compounds and suitable application in different matrices.

M

Furthermore, it eliminates the use of toxic solvents, fragility of the material and carry-over effect since the hollow fibers can be discarded after each extraction due to their low cost. In

ed

summary, LPME combines extraction, concentration and sample clean-up in one step, especially LPME three-phase [19-22]. Only two recent publications considered the use of LPME for the analysis of benzodiazepines in urine based in a two-phase system [16,17]. Cui

pt

and co-authors [16] did not validate their method for flunitrazepam while Ugland and coauthors [17] did not apply their method for the determination of diazepam and nordiazepam in

Ac ce

real cases. Therefore, both studies [16,17] analyzed simply 1 or 2 analytes, while the procedure purposed in the current work is validated and it is also able to measure 11 analytes with good sensitivity, including aminobenzodiazepines (7-aminoflunitrazepam and 7aminoclonazepam). Indeed, this is the first report of aminobenzodiazepines by LPME found in scientific literature.

Depending on the configuration used, there is a limitation related to the number of samples that can be processed in the same batch. Here we used an U-shape configuration without any supporters connecting the hollow-fiber ends. Due to the simplicity of these LPME units, many samples can be processed at the same time, providing a high sample throughput, similar to others studies [57,58]. This is an advantage when compared to other miniaturized techniques such as SPME. The major disadvantage is the lack of automation of the process since it is a relatively new technique.

Page 12 of 32

3.3. LPME optimization Benzodiazepines are considered weak basic drugs, therefore urine samples were adjusted from pH 8 to 13 using an aqueous solution of NaOH (2 mol/L). Donor phase at pH 10 was chosen due to its high uniformity among all benzodiazepines analyzed. Medazepam was not affected significantly by pH changes. When using a donor phase at pH 10, diazepam,

ip t

nordiazepam, oxazepam, chlordiazepoxide, lorazepam, flunitrazepam, nitrazepam and clonazepam showed a lower LPME extraction, while 7-aminoflunitrazepam and 7-

cr

aminoclonazepam had an increase of the signal. The behavior of the analyzed molecules can be explained by the lower pKa values of most benzodiazepines (1.30-4.60) which are similar

us

to those reported by Ugland and co-authors [17]. In sample solutions with higher pH values (10-13), there is a favored no-ionization of the NH2 group and such conditions promote the

an

ionization of anionic groups present in the remaining benzodiazepines studied. When it comes to the supported liquid membrane, 7-aminoclonazepam and 7-

M

aminoflunitrazepam were efficiently extracted using 1-nonanol while dihexyl ether showed better results for other analytes. Once it was concluded the impossibility of extracting all benzodiazepines and its metabolites in the concentrations required by UNODC and SOFT

ed

[9,10] by using just one organic solvent, the mixture of dihexyl ether and 1-nonanol was evaluated in the following ratios: 9:1; 8:2; 7:3; 6:4; 5:5. The solvent combinations are

pt

traditionally described in the literature for LLE in an attempt to improve their extraction process [12,32,59] and Ugland and co-authors [17] described a similar procedure for LPME.

Ac ce

In our experiment, an optimum result was achieved by using 9 parts of dihexyl ether and 1 part of 1-nonanol (9:1). This condition allowed the acquirement of low levels for the majority of the analytes by LPME in the three-phase mode. In the Figure 2, it is shown the relative extraction efficiencies using different combination of dihexyl ether and 1-nonanol. For better visualization of results, the highest average absolute area obtained with the experiments was considered as 100%.

Taking into consideration the acceptor phase, LPME three-phase mode is considered a greater technique compared to LPME two-phase [19-22]. There is one study using LPME three-phase mode for the determination of alprazolam, diazepam, nitrazepam and nordiazepam [18]. According to Ugland and co-authors [18], the acceptor phase should comprise of a pH value of at least 3.3 units below the pKa of the benzodiazepines to promote ionization, hence the binding of the molecule. Among the range of acceptor phase

Page 13 of 32

concentrations studied, 3.0 mol/L of HCl presented better results, possibly due to the majority of benzodiazepines having a low pKa. However, an acceptor phase with a concentration above 3 mol/L of HCl produced a dramatic reduction of the peak area for 7aminoflunitrazepam, 7-aminoclonazepam and clonazepam. It may be due to the high concentrations of acidic solution causing the degradation of these molecules. However it was not observed any significant differences for others analytes. Thus, 3.0 mol/L of HCl was

ip t

chosen as the concentration for the acceptor phase.

The equilibrium of analytes between donor phase and acceptor phase can be established

cr

more rapidly with the agitation of the LPME system [20]. The effect of agitation velocity is

us

essential for the extraction of benzodiazepines in a LPME three-phase mode as it was seen an increase of the analyte peak areas at 2400 rpm. It was also observed that, 7-aminoclonazepam, 7-aminoflunitrazepam, nitrazepam, lorazepam, medazepam and oxazepam were highly

an

influenced by the strength of shaking. Subsequently, an agitation rate of 2400 rpm was chosen for further studies.

M

Both shaking speed and time are some essential parameters to promote an extraction by LPME three-phase mode [20]. The maximum peak area for the analytes was obtained with 90

ed

minutes of agitation, except for flunitrazepam and clonazepam. Indeed, these benzodiazepines showed a better performance when extracted after 15 minutes. After 120 minutes, all analytes showed a decrease in the response. Hence, an extraction time of 90 minutes was selected as

pt

ideal, especially for 7-aminoflunitrazepam and 7-aminoclonazepam. As LPME is considered a miniaturization of conventional LLE, salting out effect was evaluated by increasing the

Ac ce

concentration of NaCl from 0 to 20% (w/v) in sample solution. The concentration of salt at 10% (w/v) of showed a higher peak area and better chromatograms while the absence of salt was better for both medazepam and nordiazepam. The maximum concentration evaluated (20%) is only suitable for aminobenzodiazepines. 3.4. Validation of the method

Method validation for benzodiazepines and its metabolites was performed after optimization of the extraction of these analytes and was carried out by establishing specificity/selective, linearity, recovery, intra and inter-day precision, accuracy, limit of detection (LoD), limit of quantification (LoQ) and dilution integrity. The internal standards (IS) and its respective analytes were as follows: diazepam-D5, medazepam, diazepam, nordiazepam and chlordiazepoxide; clonazepam-D4, flunitrazepam, clonazepam and

Page 14 of 32

nitrazepam;

oxazepam-D5,

oxazepam

and

lorazepam;

7-aminoflunitrazepam-D7,

7-

aminoflunitrazepam; 7-aminoclonazepam-D4 and 7-aminoclonazepam. No interfering peaks due to endogenous or exogenous substances were observed at the retention time of the compounds of interest. The validation parameters of the method are summarized in the Table 3.

ip t

This is a sensitive method that determines 11 benzodiazepines and its metabolites demonstrating that this new technique has similar results to traditional techniques [13,15,60] and miniaturized methods (Table 2) for extracting various substances of this class of drugs.

cr

The validated methodology includes 8 compounds (medazepam, diazepam, nordiazepam,

us

chlordiazepoxide, oxazepam, clonazepam, nitrazepam and 7-aminoclonazepam) with LoD values between 0.1-2.5 ng/mL, below the threshold designated by UNODC and SOFT [9,10]. The compound 7-aminoflunitrazepam is detected within the value set (5 ng/mL), while

an

lorazepam and flunitrazepam are identified above the given concentration by these guidelines (15 ng/mL). LoQ values ranged from 0.5 to 30 ng/mL demonstrating an advantage of this

M

method in evaluating this class of drugs in view of the physicochemical differences of benzodiazepines and their metabolites. Recovery values varied among the analytes (3.30 to 92.77% for lorazepam and medazepam, respectively) which may have been caused by

ed

different chemical(s) group(s) among the benzodiazepines evaluated, promoting distinct ionization capacity and partition coefficient of these analytes.

pt

In the calibration curve range (from the LoQ of each analyte to 250 ng/mL) the phenomenon of heteroscedasticity was observed (evaluated through the distribution F).

Ac ce

Ordinary least-squares linear regression methods could result in large errors in the calculation of drugs in concentrations especially in the smaller range of values. By using weighted least squares linear regression of the sum of percentage relative error (% RE) over the whole range, it indicated "goodness of fit". The effectiveness of the weighting factors evaluated were 1/x; 1/x1/2; 1/y1/2; 1/y; 1/y2 and 1/x2 [61]. Heteroscedasticity was observed in 7 analytes (medazepam, diazepam, nordiazepam, chlordiazepoxide, oxazepam, clonazepam and nitrazepam) using 1/x1/2 as weighted squares linear regression while 1/x was used for flunitrazepam. The weighted least squares linear regression equations and coefficients of correlation were: Y= 0.0252X+0.00706, r2= 0.9970 (medazepam); Y= 0.00855X+0.00023, r2 = 0.9973 (diazepam); Y= 0.03018X+0.010868, r2 = 0.9973 (nordiazepam); Y= 0.001093X0.00201, r2 = 0.9907 (chlordiazepoxide); Y= 0.00967X-0.00176, r2 = 0.9980 (oxazepam); Y= 0.00874X+0.008079, r2 = 0.9966 (clonazepam); Y= 0.07866X-0.34848, r2 = 0.9915

Page 15 of 32

(nitrazepam) and Y= 0.02578X+0.004905, r2= 0.9916 (flunitrazepam). Lorazepam, 7aminoclonazepam and 7-aminoflunitrazepam were considered homoscedastic, therefore it was not needed weighted squares linear regression. The respective calibration curves and coefficients of correlations were: Y= 0.0001X-0.008, r2= 0.999 (lorazepam); Y= 0.009X-0.023, r2 = 0.999 (7-aminoclonazepam); Y= 0.014X+0.012, r2 = 0.998 (7-aminoflunitrazepam).

ip t

The values of precision for these analytes at three different concentration levels were less than 20% at the lowest concentration level and 15% for the remaining concentration levels. Intra and inter-day precision were considered suitable for all tested substances (RSD% < 11.5

cr

and 19.4, respectively). Simply 7-aminoflunitrazepam showed values slightly higher than

us

indicated by the UNODC guideline (RSD% < 15.6-20.5) in the inter-day precision. Accuracy values of the benzodiazepines evaluated at three different concentration levels ranged from

an

88.4 to 119.6%.

Due to the possibility of obtaining urine samples with concentrations above calibration curve, the study of dilution integrity was performed to ensure the method is both precise and

M

accurate after the procedure. Dilution integrity was performed by spiking the blank urine sample with 1300 ng/mL and diluting it using the same matrix, 10 times (six replicates per

ed

dilution factor). Accuracy of these dilutions for benzodiazepines and its metabolites ranged from 89.1-113.2%, while the precision was above 12.3%. These results are in concordance

pt

with the criteria established by EMEA (RSD + 15%) [36]. In Table 4 it can be seen the results of the stability test for urine samples submitted to the

Ac ce

proposed method in different conditions: autosampler (21°C/12 and 24h), 4°C/24h and 20°C/24h. The studies showed that analytes were stable for all tested conditions (differences of 80-120% compared to freshly prepared samples).

4. Application of the method in real samples Four volunteers (real cases 1-4) reported the use of benzodiazepines (diazepam, lorazepam and clonazepam) before sample collection and one unknown urine sample (real case 5) was analyzed according to proposed methodology. Samples from volunteers were collected in the morning (8-9 hs). Real cases 1 and 2 had taken 5 mg of diazepam during the night before sample collection while real cases 3 and 4 ingested 2 and 0.5 mg of lorazepam and clonazepam, respectively. No relevant information about drug and/or concentration intake

Page 16 of 32

was obtained for real sample 5. Figure 3 shows the chromatograms obtained with the practical use of this method for the analyses of urine samples (blank samples spiked with analytes and internal standards, blank samples spiked only with internal standards and samples deriving from real cases). Nordiazepam was identified only in real case 2 (69.1 ng/mL) while oxazepam was

ip t

detected in real cases 1 (109.6 ng/mL) and 2 (307.9 ng/mL). Oxazepam is the last metabolite excreted with relative abundance in urine for many benzodiazepines, including diazepam [62,63]. In the real case 1, oxazepam concentration is below the cut off levels used in some

cr

immunoassays screening tests (200 ng/mL), which could lead to false negative results and

us

consequently erroneous toxicological reports.

The analysis of urine sample from volunteer number 3 showed the presence of lorazepam

an

at a concentration 151.1 ng/mL of. In spite of the low lorazepam recovery (3.3-4.2%) observed during validation of the method, it was possible to detect this substance in relatively low levels. In real cases 4 and 5, clonazepam was not detected and only 7-aminoclonazepam

M

was determinated in both cases. Volunteer number 4 admitted oral intake of clonazepam (0.5 mg) and it was found 23.5 ng/mL of 7-aminoclonazepam in the urine sample. This data

ed

suggests that small therapeutic doses promote low levels of metabolite excretion, considering that the biological matrix analyzed was the first collection after drug intake.

pt

The analysis of urine sample from real case number 5 revealed the presence of 7aminoclonazepam at a concentration of 224 ng/mL, indicating that this analytical

Ac ce

methodology is reliable and double derivatization (TFAA and MTBSTFA) allowed an unequivocally determination of 7-aminoclonazepam, the major metabolite of clonazepam, showing uniform peak shapes and no artifact formation from urine sample without the knowledge of drug intake as shown in the Figure 3. These characteristics of the proposed methodology are essential to reveal DFC situations. The developed method is also currently used in our laboratory for the routine analysis in

workplace drug testing programs. Five samples previously submitted to immunoassay screening confirmed the presence of oxazepam and nordiazepam.

5. Conclusions

Page 17 of 32

A sensitive and reliable LPME/GC-MS method using double derivatization for the determination of benzodiazepines and its metabolites in human urine samples was developed and fully validated. The methodology suggested offers a wider range for this class of drugs which can be measured by a miniaturized technique in urine matrix. To our knowledge, this is the first full validation of these molecules for LPME with an application of real cases. Compared to other methods, it revealed some practical advantages such as a simple device,

ip t

with a relative low cost and also sensitivity. This procedure can be applied when there is the

cr

need for analysis of benzodiazepines and its metabolites in urine samples.

us

Acknowledgments

Financial supports from Fundação de Amparo à Pesquisa do Estado de São Paulo

an

(FAPESP), Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) are gratefully

M

acknowledged. The authors have declared no conflicts of interest. We also thank to Dr. Claudio Martins Pereira de Pereira for his suggestions on given issues of the manuscript and

References

pt

revision of the paper.

ed

Ana Miguel Fonseca Pego, from University of Glasgow, Scotland, UK, for the English

Ac ce

[1] E. Gallardo, M. Barroso, J.A. Queiroz, Drug Test. Anal. 1, (2009) 109. [2] S.A. Legrand, P. Silverans, P. de Paepe, W. Buylaert, A.G. Verstraete. Traffic Inj. Prev. 14 (2013) 461.

[3] J. Kim, S. Lee, S. In, H. Choi, H. Chung, J. Chromatogr. B. 879 (2011) 878. [4] H.L. Shih, M.C. Lin, C.C. Lin, H.C. Hsu, H.L. Lee, C.H. Chi, F.C. Sung, Y.J. Chang, C.H. Kao, Psychopharmacology 229 (2013) 665. [5] M. Scott-Ham, F.C. Burton, J. Clin. Forensic Med. 12 (2005) 175. [6] B. Marc. Ther. Drug Monit., 30 (2008) 218.

Page 18 of 32

[7] S. Djezzar, F. Questel, E. Burin, S. Dally, Chemical submission: results of 4-year French inquiry. Int. J. Legal Med. 123 (2009) 213. [8] I.J. Bosman, M. Verschraagen, K.J. Lusthof, J. Forensic Sci. 56 (2011) 1562. [9] UNODC, United Nations Office on Drugs and Crime, Guidelines for the forensic analysis

2011.

Available

ip t

of drug facilitating sexual assault and other criminal acts 2011. United Nations, New York, from:

http://www.unodc.org/documents/scientific/forensic_analys_of_drugs_facilitating_sexual_ass

cr

ault_and_other_criminal_acts.pdf. Accessed in 15/01/ 2014.

us

[10] SOFT, Society of Forensic Toxicologists, Recommended minimum performance limits for common DFSA drugs and metabolites in urine samples. United States of America, Mesa, 2014.

http://soft-tox.org/sites/default/files/SOFT%20DFC%20Rec%20Det%20Limits%201-

an

2014.pdf. Accessed in 26/01/2014.

M

[11] T. Gunnar, K. Ariniemi, P. Lillsunde, J. Chromatogr. B, 818 (2005) 175. [12] S. Fu, J. Lewis, H. Wang, J. Keegan, M. Dawson, J. Anal Toxicol. 34 (2010) 243.

ed

[13] I.I. Papoutsis, S.A. Athanaselis, P.D Nikolaou, C.M. Pistos, C.A Spiliopoulou, C.P. Maravelias, J. Pharm. Biomed. Anal. 52 (2010) 609.

pt

[14] M. Nakamura, Biomed. Chromatogr. 25 (2011) 1283.

Ac ce

[15] K. Arnhard, R. Schmid, U. Kobold, R. Thiele, Anal. Bioanal. Chem. 403 (2012) 755. [16] S. Cui, S. Tan, G. Ouyang, J. Pawliszyn, J. Chromatogr. A 1216 (2009) 2241. [17] H.G. Ugland, M. Krogh, K.E. Rasmussen, J. Chromatogr. B, 749 (2000) 85. [18] H.G. Ugland, M. Krogh, L. Reubsaet, J. Chromatogr. B 798 (2003) 127. [19] J. Lee, H.K. Lee, K.E. Rasmussen, S. Pedersen-Bjergaard, Anal. Chim. Acta 624 (2008) 253. [20] S. Pedersen-Bjergaard, K.E. Rasmussen, J. Chromatogr. A. 1184 (2008) 132. [21] R.A. Menck, D.S.D. Lima, S.C. Seulin, V. Leyton, C.A. Pasqualucci, D.R. Muñoz, M.D. Osselton, M. Yonamine, J. Sep. Sci. 35 (2012) 3361.

Page 19 of 32

[22] L. N. Pantaleao, B.A.B. Paranhos, M. Yonamine, J. Chromatogr. A. 7 (2012) 1. [23] R.J. Dinis-Oliveira, T. Magalhães, Toxicol. Mech. Methods 23 (2013) 471. [24] H. Maurer, K. Pfleger, J. Chromatogr. A 422 (1987) 85. [25] M. Japp, K. Garthwaite, A.V. Geeson, M.D. Osselton. J. Chromatogr. A 439 (1988) 317. [26] J. Segura, R. Ventura, C. Jurado, J. Chromatogr. B 713 (1998) 61.

ip t

[27] M.N. Udin, V.F. Samanidou, I.N. Papadoyannis, J. Chromatogr. Sci. 51 (2013) 587. [28] M. Terada, S. Masui, T. Hayashi, R. Watanabe, H. Inoue, M. Iino, M. Nakatome, R.

us

[29] A.A. Elian, Forensic Sci. Int. 101 (1999) 107.

cr

Matoba, T. Shinozuka, T. Murai, E. Tanaka, K. Honda, Leg. Med. 5(2003) S96.

an

[30] J.Y. Kim, M.K. In, J.H. Kim, Rapid Commun. Mass Spectrom. 20 (2006) 3159. [31] G.A. Cavalcanti, B.C. Garrido, F.D. Leal, M.C. Padilha, X. de la Torre, H.M. Pereira,

M

F.R. de Aquino Neto, Rapid Commun. Mass Spectrom. 24 (2010) 3297. [32] R. Meatherall, J. Anal. Toxicol. 18 (1994) 382.

ed

[33] D.A. Armbruster, T. Pry, Clin. Biochem. Rev. 29 (2008) S49. [34] F.T. Peters, O.H. Drummer, F. Musshoff, Forensic Sci. Int. 165 (2007) 216.

pt

[35] UNODC, United Nations Office on Drugs and Crime, Guidelines for the Validation of Analytical Methodology and Calibration of Equipment used for Testing of Illicit Drugs in

Ac ce

Seized Materials and Biological Specimens. United Nations, New York, 2009. Available from: http://www.unodc.org/documents/scientific/validation_E.pdf. Accessed in 15/01/ 2014. [36] EMEA, European Medicine Agency, Guideline on bioanalytical method validation. United

Kingdom,

London,

2011.

http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2011/08/WC5 00109686.pdf. Accessed in 15/01/2014. [37] H.R. Lin, A.C. Lua, Rapid Commun. Mass Spectrom. 20 (2006) 1724. [38] J. Clayden, N. Greeves, S. Warren, Sulfur, silicon and phosphorus in organic chemistry, in: J. Clayden, N. Greeves, S. Warren (Eds.), Organic Chemistry, 2° edition, Oxford University Press, São Paulo, 2012, p.656-693.

Page 20 of 32

[39] B. Khakimov, M.S. Motawia, S. Bak, S.B. Engelsen, Anal. Bioanal. Chem. 405 (2013) 9193. [40] I. Deinl, G. Mahr, L. Von Meyer, J. Anal. Toxicol. 22 (1998) 197. [41] M.D. Robertson, O.H. Drummer, J. Forensic Sci. 43 (1998) 5.

ip t

[42] M. Chèze, M. Villain, G. Pépin, Forensic Sci. Int. 145 (2004) 123. [43] V. Samanidou, L.Kovatsi, D. Fragou, K. Rentifis. Bioanalysis 3 (2011) 2019.

cr

[44] M. Barroso, I. Moreno, B. Fonseca, J.A. Queiroz, E. Gallardo. Bioanalysis 4 (2012)

us

1805.

[45] G. Alves, M. Rodrigues, A. Fortuna, A. Falcão, J. Queiroz. Bioanalysis 5 (2013) 1409.

an

[46] S. Rani, A.K. Malik. J. Sep. Sci. 35 (2012) 2970.

M

[47] H.L. Lord, M. Rajabi, S. Safari, J. Pawliszyn. J. Pharm. Biomed. Anal. 40 (2006) 769. [48] A. Aresta, L. Monaci, C.G. Zambonin. J. Pham. Biomed. Anal. 28 (2002) 965.

ed

[49] W.M. Mullett, J. Pawliszyn. Anal. Chem. 74 (2002) 1081.

Toxicol. 23 (1999) 54

pt

[50] F. Guan, H. Seno, A. Ishii, K. Watanabe, T. Kumazawa, H. Hattori, O. Suzuki. J. Anal.

[51] K.J. Reubsaet, H.R. Norli, P. Hemmersbach, K.E. Rasmussen. J. Pharm. Biomed. Anal.

Ac ce

18 (1998) 667.

[52] H. Yuan, Z. Mester, H. Lord, J. Pawliszyn. J. Anal. Toxicol. 24 (2000) 718. [53] S. Khodadoust, M. Ghaedi. J. Sep. Sci. 36 (2013) 1734. [54] M.B. Melwanki, W.S. Chen, H.Y. Bai, T.Y. Lin, M.R. Fuh. Talanta 78 (2009) 618. [55] M. Ghobadi, Y. Yamini, B. Ebrahimpour. J. Sep. Sci. 37 (2014) 287. [56] F. Rezaei, Y. Yamini, M. Moradi, B. Daraei. Anal. Chim. Acta 4 (2013) 135. [57] L. do Nascimento Pantaleao, B.A. Bismara Paranhos, M. Yonamine, J. chromatogr. A 7 (2012) 1. [58] A.V. Bairros, R. Lanaro, R.M. Almeida, M. Yonamine. Forensic Sci. Int. 243 (2014) 47.

Page 21 of 32

[59] S.L. Chou, M.H. Yang, Y.C. Ling, Y.S. Giang, J. Chromatogr. B 799 (2004) 37. [60] M.A. Elsohly, W. Gul, K.M. Elsohly, B. Avula, I.A. Khan, J. Anal. Toxicol. 30 (2006) 524. [61] A.M. Almeida, M.M. Castel-Branco, A.C. Falcão, J. Chromatogr. B 774, (2002) 215.

ip t

[62] S. Kinani, S. Bouchonnet, N. Milan, I. Ricordel, J. Chromatogr. A 1141 (2007) 131.

cr

[63] R. Lennestål, H.A. Lakso, M. Nilsson, T. Mjörndal, J. Anal. Toxicol. 32 (2008) 402.

us

Legends

Figure 1- Mass spectra of aminobenzodiazepines after double derivatization. A) 70 eV; B) 15eV; C) 5ev. NHCZ, 7-aminoclonazepam; NHFZ, 7-aminoflunitrazepam.

an

Figure 2- Bars graph (triplicate for each sample) showing extraction efficiencies of the supported liquid membrane combination using dihexyl ether: 1-nonanol in different ratios

M

(9:1; 8:2; 7:3; 6:4; 5:5) to analyze benzodiazepines and its metabolites. MZ, medazepam; DZ, diazepam; ND, nordiazepam; FZ, flunitrazepam; CDZ, chlordiazepoxide; OXZ, oxazepam; NZ, nitrazepam; NHFZ, aminoflunitrazepam; LZ, lorazepam; CZ, clonazepam; NHCZ,

ed

aminoclonazepam. Data are presented as means + standard deviations. The highest average absolute area obtained with the experiments was considered as 100%.

pt

Figure 3- Selected ion monitoring (SIM) chromatograms obtained by the LPME and GC-MS analysis of urine sample. A) Blank sample spiked with 100 ng/mL of all analytes without

Ac ce

internal standards; B) Blank sample spiked with 100 ng/mL of internal standards; C) Real sample (case 2) containing 69.1 ng/mL of ND and 307.9 ng/mL of OXZ; D) Real sample (case 5) containing 224 ng/mL of NHCZ. MZ, medazepam; DZ, diazepam; ND, nordiazepam; FZ, flunitrazepam; CDZ, chlordiazepoxide; OXZ, oxazepam; NZ, nitrazepam; NHFZ, 7aminoflunitrazepam; LZ, lorazepam; CZ, clonazepam; NHCZ, 7-aminoclonazepam.

Page 22 of 32

*Highlights (for review)

Highlights

Determination of low levels of benzodiazepines and their metabolites in urine by hollow-fiber liquid-phase microextraction (LPME) using gas chromatography-mass spectrometry (GC-MS)

ip t

André Valle de Bairrosab*, Rafael Menck de Almeidaa, Lorena Pantaleãoa, Thiago Barcellosc,

Faculty of Pharmaceutical Sciences, University of São Paulo, Brazil

c

Center of Chemical, Pharmaceutical and Food Sciences, Federal University of Pelotas, Brazil

us

b

Institute of Biotechnology, University of Caxias do Sul, Brazil

an

a

cr

Sidnei Moura e Silvac and Mauricio Yonaminea

(LPME)

for

Ac

ce pt

ed

M

 Development of hollow-fiber liquid-phase microextraction benzodiazepines in urine samples  Low-cost method  Wide range of analytes with different characteristics  Highly sensitive  New derivatization procedure for GC-MS analysis  Applicable in the fields of clinical and forensic toxicology

Page 23 of 32

Ac

ce

pt

ed

M

an

us

cr

i

Figure1

Page 24 of 32

Ac

ce

pt

ed

M

an

us

cr

i

Figure2

Page 25 of 32

Ac ce p

te

d

M

an

us

cr

ip t

Figure3

Page 26 of 32

Table 1

Table 1. Qualifier and quantifier ions of the benzodiazepines and internal standards (m/z) obtained from the GC-MS analysis after derivatization with TFAA and MTBSTFA. Qualifier ion

Quantifier ion

Retention time (min)

Medazepam

244, 270

242

5.64

Diazepam

258, 284

256

6.58

Nordiazepam

329, 384

327

Flunitrazepam

266, 286

312

Chlordiazepoxide

282, 358

356

Oxazepam

459, 514

457

7.84

Nitrazepam

292, 394

338

7.93

7-aminoflunitrazepam

280, 493

436

8.09

Lorazepam

493, 515

491

8.40

Clonazepam

326, 374

372

8.44

368, 554

552

8.80

ip t

Analyte

6.81

7.66

Diazepam-D5

ed

M

an

us

cr

7.38

289, 287

261

6.55

Oxazepam-D5

464, 519

462

7.81

7-aminoflunitrazepam-D7

287, 500

443

8.05

Clonazepam-D4

330, 378

376

8.42

7-aminoclonazepam-D4

372, 558

556

8.78

Ac

ce pt

7-aminoclonazepam

GC-MS, gas chromatography-mass spectrometry; TFAA, trifluoroacetic anhydride; N-methyl-N-tertbutyldimethylsilyltrifluoroacetamide with 1% tert-butyldimethylchlorosilane (MTBSTFA).

Page 27 of 32

cr

ip t

Table 2

Analytes

Extraction

CDZ, CZ, DZ, FZ, LZ, MZ, ND, NHCZ,

LPME

Detection

LoD (ng/mL)

Reference

GC-MS

0.1-15

Purposed methodology

FZ

LPME

GC-MS/MS

0.001

16

DZ and ND

LPME

GC-NPD

5.7-31

17

M

an

NHFZ, NZ and OXZ

us

Table 2. Summary of miniaturized techniques for the determination of BZD and its metabolites in urine samples.

MEPS

GC-MS

0.1

46

SPME

LC-MS/MS

0.02

47

SPME

LC-UV

5

48

SPME

LC-UV

46-600

49

BZ, CZ, DZ, FZ, HZ, NZ, NTZ

SPME

GC-ECD

2-20

50

AZ, DZ, FZ, ND, PZ and OXZ

SPME

GC-NPD

2.85-148

51

DZ, ND, TZ, OXZ, NHFZ, NHDFZ and CZ

SPME

LC-MS

0.02-2.0

52

CDZ

DLLME

LC-UV

2.4

53

NHFZ

DLLME

LC-MS/MS

0.025

54

DZ, MIZ and AZ

SPE-DLLME

GC-FID

0.1-0.2

55

AZ, NZ, CZ, MIZ and DZ

VMME

LC-UV

2.0

56

AZ

DLZ

Ac c

ep te

DZ, ND, OXZ, TZ and CZ

d

NHFZ

AZ, alprazolam; CDZ, chlordiazepoxide; CZ, clonazepam; DLZ, delorazepam; DZ, diazepam; FZ, flunitrazepam; HZ, haloxazolam; LZ, lorazepam; MIZ, midazolam; ND, nordiazepam; NHDFZ, 7-aminodesmethylflunitrazepam; NHFZ, 7-aminoflunitrazepam; NZ, nitrazepam; NTZ, nimetazepam; OXZ, oxazepam; PZ, Prazepam; TZ, temazepam. DLLME, dispersive liquid-liquid microextraction; GC-FID, gas chromatography with flame ionization detector; GC-ECD, gas chromatography-electron capture detector; GC-MS, gas chromatography-mass spectrometry; GC-NPD, gas chromatography with nitrogenous-phosphorus detector; LPME, hollow fiber-liquid phase

Page 28 of 32

ip t cr

us

microextraction; VMME, hollow fiber-vesicular mediated microextraction; LC-MS, liquid chromatography-mass spectromeytry; LC-UV, liquid chromatography-ultraviolet;

Ac c

ep te

d

M

an

MEPS, microextraction by packed sorbents; SPE, solid phase extraction; SPME, solid phase microextration.

Page 29 of 32

cr

ip t

Table 3

point). MZ

DZ

ND

CDZ

C1

79.5

75.8

75.1

39.9

C2

79.2

75.6

81.0

C3

92.7

84.7

85.2

LoD (ng/mL)

0.5

1.0

LoQ (ng/mL)

2.5

5.0

LZ

FZ

CZ

NZ

7-NHFZ

7-NHCZ

20.9

3.3

22.2

21.5

47.2

40.4

34.3

35.1

25.1

3.5

23.4

22.5

51.2

43.8

37.2

32.6

28.3

4.2

25.0

25.8

54.7

45.7

39.7

1.0

2.5

15

15

5.0

2.5

5.0

1.0

0.5

5.0

5.0

30

30

10

5.0

10

5.0

8.0

3.8

10.2

3.4

3.5

6.2

10.1

4.1

9.4

3.7

4.9

9.4

2.0

5.3

11.5

2.4

8.9

11.0

2.2

5.4

6.5

5.6

9.8

1.8

5.5

10.1

1.8

9.7

10.5

1.7

9.2

6.5

3.3

C1

8.0

2.7

7.8

19.4

3.0

7.6

9.6

4.1

6.8

20.5

5.2

C2

7.1

5.0

4.0

14.8

4.9

7.6

11.3

2.3

5.1

15.6

3.9

C2 C3

d 0.1

ep te

Ac c

C1

M

Recovery (%)

Intra-day precision (RSD%)

OXZ

an

Performance Parameters

us

Table 3. Validation parameters of the develped method for the determination of benzodiazepines and its metabolites in urine samples (six replicates for each

Inter-day precision (RSD%)

Page 30 of 32

ip t 5.3

11.2

C1

119.6

100.2

119.4

103.2

111.1

C2

97.9

102.5

99.2

88.0

C3

96.8

105.0

99.8

10 times

11.9

4.4

7.4

15.9

6.2

102.7

117.1

91.6

103.8

103.5

93.0

97.3

102.9

105.0

88.4

95.8

106.3

100.9

94.3

103.2

105.6

113.6

91.5

92.9

107.5

104.0

d 8.3

8.7

9.4

12.3

7.6

6.6

10.5

6.3

5.8

9.8

4.2

102.2

111.4

113.2

105.1

99.7

89.1

97.6

102.8

103.5

97.0

Ac c

Accuracy (%)

ep te

Dilution integrity

10 times

10.7

an

Accuracy (%)

Precision (RSD%)

5.8

cr

5.5

us

14.1

M

C3

111.6

C1, 1.5 ng/mL (ND); 7.5 ng/mL (MZ); 15 ng/mL (DZ, CDZ, OXZ, NZ and NHCZ); 30 ng/mL (CZ and NHFZ); 90 ng/mL (LZ and FZ); C2, 130 ng/mL; C3, 235 ng/mL; LoD, limit of detection; LoQ, limit of quantification; RSD%, relative standard deviation; 10 times, 1300 ng/mL; MZ, medazepam; DZ, diazepam. ND, nordiazepam; CDZ, chlordiazepoxide; OXZ, oxazepam; LZ, lorazepam; FZ, flunitrazepam; CZ, clonazepam; NZ, nitrazepam; NHFZ, 7-aminoflunitrazepam; NHCZ, 7-aminoclonazepam.

2

Page 31 of 32

cr

ip t

Table 4

us

Table 4. Stability results of urine samples submitted to purpose method in different conditions: autosampler (12 and 24h), 4°C/24h and -

CZ

NHCZ

FZ

NHFZ

Autosampler (21°C/12h)

99.8

95.5

92.0

98.0

Autosampler (21°C/24h)

91.2

86.5

88.2

4°C/24h

89.0

90.1

-20°C/24h

92.3

DZ

ND

OXZ

LZ

CDZ

NZ

MZ

94.8

94.1

105.7

109.8

106.6

93.0

96.0

82.9

97.5

96.1

95.7

93.7

113.7

85.6

98.9

89.0

91.7

102.6

88.2

89.9

102.3

87.5

97.9

85.3

96.1

92.1

89.7

90.4

92.4

88.8

95.9

an

Sample stability

M

20°C/24h at 100 ng/mL. Results are present in percentage compared to freshly prepared samples.

d

92.0

91.0

ep te

88.4

MZ, medazepam; DZ, diazepam. ND, nordiazepam; CDZ, chlordiazepoxide; OXZ, oxazepam; LZ, lorazepam; FZ, flunitrazepam; CZ, clo nazepam; NZ, nitrazepam; NHFZ,

Ac c

7-aminoflunitrazepam; NHCZ, 7-aminoclonazepam

Page 32 of 32