Rapid and Effective Determination of Ethyl Glucuronide in Hair by Micro Extraction by Packed Sorbent (MEPS) and LC‐MS/MS

ABSTRACT

Ethyl glucuronide (EtG) in hair is a reliable biomarker of alcohol consumption habits. Due to its small concentration incorporated into hair, analytical methods sensitive enough to reliably quantify EtG in this matrix are required. Sample preparation is critical in hair analysis, especially for EtG, for which extraction efficiency and matrix effect can strongly influence the results; furthermore, miniaturized methods are sought, to reduce solvent use and times of sample preparation. A micro extraction by packed sorbent (MEPS) procedure coupled to a high‐performance liquid chromatography–tandem mass spectrometry method was developed and validated for quantitation of EtG in human hair samples. Fifty milligrams of hair samples were cut into snippets and extracted in water. The cleanup of the extract was carried out by using a MEPS syringe packed with anion exchange sorbent (SAX); all parameters for conditioning, washing, loading and eluting steps were optimized and the eluted aqueous volume was directly injected in the LC–MS/MS system operating in the negative ionization mode. The method was fully validated assessing LOD, LOQ, calibration curve, repeatability, accuracy, matrix effect and carryover. The method was subsequently applied to QCs and authentic hair samples. The developed MEPS method is quick and effective, with low solvent purchase and discard costs, allowing the differentiation between social drinkers and chronic excessive alcohol consumers, according to the cut‐offs established by the Society of Hair Testing (SoHT).

The development and validation of a MEPS method for the purification of hair extract for the EtG determination and quantification in the keratin matrix is the main focus of this study. The method is fast and effective and, as a miniaturised technique, requires only small amounts of solvent.

1. Introduction

The determination of alcohol abuse is of importance for both clinical and forensic toxicology. Ethyl glucuronide (EtG), a non‐oxidative minor metabolite of ethanol, has been used as a specific biomarker of alcohol consumption. EtG is formed in the body only as a direct product of alcohol consumption, not influenced by other bodily parameters, and is therefore currently considered more reliable, sensitive and specific than the indirect markers [1, 2, 3, 4]. The detection of EtG in hair is useful for retrospective monitoring of alcohol consumption over time, as it has been shown to be stably incorporated into hair after repeated alcohol consumption. Some publications have investigated the correlation between the concentration of EtG in the hair and the ingested dose of alcohol, resulting in a good correlation [4, 5, 6, 7]. Furthermore, its levels in hair, unlike other drugs of abuse in the keratin matrix, were proven not to be influenced by hair colour [8, 9]. The 2019 Society of Hair Testing Consensus established two cut‐off levels for EtG determination in the proximal hair segment from 3 to 6 cm in length, the first at 5 pg/mg, to distinguish between abstinence and repeated alcohol consumption, and the second at 30 pg/mg to determine chronic excessive consumption [10]. Due to the low concentrations of EtG in hair and the relative cut‐off limits, the fact that it is a polar and relatively low mass compound in a complex matrix such as hair, EtG determination is analytically challenging and requires sensitive techniques for both sample preparation and analysis. Therefore, GC‐MS/MS after derivatization [4, 11, 12, 13] and LC‐MS/MS [3] have been mainly used for EtG determination. Hair extracts are often purified by SPE [11, 12, 13] or analysed directly without further pretreatment [4].

Micro extraction by packed sorbent (MEPS) is a miniaturized solid‐phase extraction in which the sorbent material is placed in the barrel of a syringe. Compared to traditional SPE, where the solution flows in one direction only, MEPS allows the solution to pass through the stationary phase multiple times, increasing extraction efficiency. In addition, MEPS has low solvent consumption and can easily be connected with both GC and LC without modification. Since its first application in the early 2000s, MEPS was used to analyse various biological samples including blood or plasma [14, 15], urine [16, 17, 18], oral fluid [19, 20, 21] and hair extracts [22, 23], for the determination of analytes of interest in forensic toxicology.

The study was aimed at the development of a method using MEPS for the clean‐up and enrichment of EtG in hair extracts. To the best of our knowledge, this is the first application of MEPS for the analysis of EtG in hair.

2. Materials and Methods

2.1. Standards and Reagents

Water, formic acid and methanol were purchased from Sigma‐Aldrich (Milan, Italy); ammonium formate was from Agilent (Agilent Technologies, Santa Clara, CA, USA). All reagents and solvents were of LC‐MS grade.

Methanolic solutions (1 mg/mL) of EtG and EtG‐D5 were obtained from Chebios (Rome, Italy). Standard compounds were stored according to supplier recommendations.

Working solutions of EtG and EtG‐D5 were prepared at a concentration of 0.1 μg/mL. Working solutions were stored at −20 C until use.

2.2. Sample Preparation

2.2.1. Hair Extraction Procedure

Hair samples were collected by cutting a strand in the posterior vertex zone with scissors, as close as possible to the scalp.

The samples were decontaminated with a single 1 min wash with 1‐mL dichloromethane with a vortex mixer. Then, the hair samples were air‐dried. Samples were subsequently cut into small pieces with scissors. Fifty milligrams of sample was added with 450 μL of water and 10 μL of the deuterated internal standard (0.1 μg/mL, final concentration 20 pg/mg) and left at 40°C for overnight incubation, followed by sonication for 2 h.

2.2.2. MEPS Extraction

The hair extract was purified by using MEPS syringe (250 μL) packed with SAX sorbent from SGE Italy (Milan, Italy). The sorbent was initially conditioned by flushing three times methanol and three times water, as suggested by the supplier. The MEPS procedure used for sample clean‐up was optimized, resulting in the following final conditions: For the loading step, the ‘draw‐eject’ mode was used; thus, 100 μL of the hair extract was passed through the MEPS dispensing and aspirating 12 times. After the washing step with 100 μL of water, analytes elution was achieved by 50‐μL 1% HCOOH. This volume was aspirated and ejected from the syringe five times, resulting in five load/release cycles with the same solvent. The eluate was collected in a vial, and 10 μL was directly injected in the UPLC‐MS/MS instrument.

2.3. UPLC‐MS/MS Equipment and Condition

A Waters Acquity UPLC system (Waters, Milford, MA, USA) was used for separation, equipped with a Raptor EtG/EtS column (2.1 mm ID × 100 mm, particle size 2.7 μm, from Restek, Milan, Italy) maintained at 40°C. The LC conditions were as follows: gradient elution with water containing 0.01% (v/v) formic acid as mobile phase A and methanol with 0.01% formic acid as mobile phase B. The flow rate was 500 μL/min, and the gradient was programmed as follows: linear increase from 5% to 35% of phase B in 2.5 min, in 0.01 min phase B percentage decreased again to 5%, this condition is held from 2.51 to 4.00 min.

A Xevo TQ‐S micro tandem‐quadrupole mass spectrometer (Waters, Milford, MA, USA) equipped with a Z‐spray electrospray interface was used. Negative electrospray ionization (ESI) was performed in the multiple reaction monitoring (MRM) mode. The capillary voltage was set to −2.5 kV. The source block temperature was 150°C, and the desolvation gas (nitrogen) was heated to 650°C and delivered at a flow rate of 1000 L/h. The following MRM transitions were monitored: for EtG m/z 220.97 ➔ 74.79, 84.80 and 112.77 using for collision energy 14, 18 and 14 V, respectively, and for EtG‐D5 225.97 ➔ 74.79 and 84.80, using for collision energy 14 and 18 V, respectively. The cone voltage was set to 42 V for all the transitions followed. MRM transition ions, both precursors and products, cone voltages and collision energies were optimized by using Waters IntelliStart. System operation and data acquisition were controlled using Mass Lynx 4.2 software (Waters, Milford, MA, USA). All data were processed with the Target Lynx quantification program (Waters, Milford, MA, USA).

2.4. Method Validation

Method validation considered the following parameters: specificity, the limit of detection (LOD), the limit of quantification (LOQ), linear correlation of the calibration curve, intraday precision and accuracy, recovery, matrix effect and carry‐over.

Hair samples collected from teetotallers and children, previously tested negative for EtG, were used for method validation.

The LOD was calculated at a concentration value that gave a signal‐to‐noise ratio S/N > 3 for three EtG MRM transitions. This parameter was estimated from five replicate analyses of spiked hair samples extracted and then purified with MEPS at decreasing concentrations.

The LOQ was determined as the concentration of the spiked sample extracted and treated with MEPS that gave a signal‐to‐noise ratio equal to or greater than 10 and a CV% not exceeding 20%.

The concentration–response relationship was evaluated using the least squares regression method. A calibration curve was built by adding the appropriate amount of EtG to the blank hair extracted by the above procedure to obtain five calibrators, with triplicate analyses for each level. The concentration levels prepared were 5, 10, 20, 30 and 50 pg/mg.

Intraday precision and accuracy were assessed by preparing QC samples in triplicate at four different concentration levels, 5, 10, 30 and 50 pg/mg. Intraday precision was expressed as CV%. Accuracy was calculated as the percentage deviation from the nominal concentration normalised to the nominal concentration.

Recovery and matrix effect were evaluated by preparing three sets of QC samples, at the four concentrations already used to assess intraday precision and accuracy, each level prepared in triplicate. The first set consisted of blank hair samples fortified before the MEPS procedure, the second set consisted of samples fortified after the MEPS extraction and the last set consisted of neat solutions containing the appropriate amount of EtG without any extraction.

Recovery was then calculated as the ratio of the average absolute EtG areas for the first set to the average absolute areas for EtG in the second set. The matrix effect was the average of the absolute areas of the second set to the average of the absolute areas of the third set. Process efficiency for MEPS extraction can be calculated both by multiplying the value obtained for recovery and that for matrix effect by dividing the average of the absolute areas for the first set by the average of the absolute areas for the third set.

Carry‐over was evaluated by extracting with MEPS a blank sample after the extraction of the highest calibrator, 50 pg/mg. Further experiments were performed by preparing hair spiked with 50 pg/mg EtG, extracting and treating with MEPS and analysing a second recovery volume obtained by eluting with 50‐μL 1% HCOOH with five load/release cycles after the first elution step. Both the experiments were performed in triplicate.

2.5. Authentic Cases and Proficiency Test

The method was applied to 95 authentic samples from forensic caseworks and also to samples from the SoHT Proficiency Test. Authentic samples were collected according to SoHT recommendations [10], cut as close as the scalp as possible and with length between 3 and 6 cm.

3. Results

3.1. MEPS Optimization

For MEPS extraction, the loading, washing and elution steps were optimized. Each of the experiments was prepared in triplicate. The experiments were performed on the extracts obtained by the above procedure applied to hair samples spiked at 30 pg/mg.

For the loading phase, different conditions were tested, including ‘draw‐eject’ and ‘extraction‐discard’ modes. The former consists of repeated cycles of aspiration and injection in the same vial, whereas for the latter, each aliquot of the sample, after being aspired, is subsequently discarded in waste. The loading conditions tested were the following: whole sample aspired and ejected for 1, 3 or 5 cycles; extraction‐discard mode with 3 aliquots of 150 μL; draw‐eject mode with aliquots of 100 μL for 3, 6, 9, 12 or 15 cycles. The different loading procedures were evaluated by completing the MEPS extraction and comparing the absolute areas in the eluate, 3 replicates for each experiment, with that of a sample prepared at the same concentration in the solvent. The presence of the analyte in the residue remaining after loading was also evaluated. The best results were obtained for the draw‐eject procedure with 12 and 15 cycles in terms of both highest areas in the eluates and lowest residues in loading samples. Because the results were similar, the load with 12 cycles was chosen. Figure 1 shows the different load experiments results.

FIGURE 1.

Percentages of the ratios between the mean absolute areas obtained for EtG after different loading conditions with respect to the mean absolute area of a neat standard prepared at the same concentration in solvent. The conditions tested were the following: whole sample aspirated and ejected for 1, 3 or 5 cycles; extraction‐discard mode (E‐D) with 3 aliquots of 150 μL; draw‐eject mode (D‐E) with aliquots of 100 μL for 3, 6, 9, 12 or 15 cycles. Washing step was made with 250 μL water for 1 aspiring/ejecting cycle and recovery with 250 μL of methanol 2% formic acid for 5 aspiring/ejecting cycles.

Once the most suitable conditions for sample loading had been defined, the eluting step was evaluated. Different solvents (water, methanol, water/methanol 1:1 v/v) and formic acid percentages (0.1, 1 and 2) were evaluated first. Also, volumes (50, 100, 150 and 250 μL) and number of cycles (1, 3, 5 and 7) were evaluated. All the experiments were repeated in triplicate. The recoveries were similar with all the solvents tested, although better chromatography results, in terms of peak shape, were achieved by using water. Considering the formic acid, 0.1% reduced the recovery from the MEPS, whereas 1% and 2% gave similar results. By reducing the volume of the solvent for the elution step, the recoveries slightly decreased, whereas absolute areas increased, due to the higher enrichment factor. Table 1 shows the differences in terms of absolute EtG area and in terms of percentage for the ratios between the absolute EtG area in the experiment samples versus neat standard at the same concentration by varying the final volume used for recovery. The higher the number of the aspiring/ejecting cycles, the higher the recoveries, with similar values for 5 and 7 cycles. Evaluating all the parameters investigated, the best conditions were obtained by using 50 μL of water 1% HCOOH for aspired and ejected 5 times.

TABLE 1.

Mean absolute area for EtG and ratio between mean absolute area for EtG versus mean absolute area for EtG in solvent samples as the volume used for recovery varies.

Volume used for recovery	Mean area	Mean area ratio %
250 μL	204 ± 18	44% ± 4%
150 μL	273 ± 48	43% ± 8%
100 μL	275 ± 93	20% ± 7%
50 μL	550 ± 90	32% ± 5%

Note: Solvent used for recovery was water 1% HCOOH. Number of aspiring/ejecting cycles was 7.

In the end, the washing step was evaluated. The washing solutions, tested in triplicate, were 250, 150 or 100 μL of water for 1 cycle, 100 μL of water for 3 cycles, 250‐μL 0.01% HCOOH for 1 cycle and 250‐μL 0.1% NH₃ for 1 cycle. Results were evaluated in terms of absolute areas and signal‐to‐noise. Except for washing with ammonia, which suppressed the signal, all experiments gave similar results. Consequently, the condition chosen was the simplest, one wash with 100 μL of water.

The same packed syringe was effectively used for 90 extractions of authentic hair sample extract with no loss of efficiency or carry‐over effects.

3.2. Validation Results

LOD and LOQ were experimentally established at 3 and 5 pg/mg, respectively.

The calibrators fitted with a linear curve that gave a correlation coefficient of 0.991.

Intraday precision and accuracy were always below 15%. Recovery values ranged from 33% to 39% and matrix effect from 65% to 102%, whereas the MEPS process efficiency was from 23% to 34%. Intraday precision, accuracy, recovery, matrix effect and process efficiency values are shown in Table 2.

TABLE 2.

Percentage precision, expressed as percentage coefficient of variation (%CV), accuracy, expressed as percentage error (%E), recovery (R%), matrix effect (ME%) and process efficiency (%PE) at four concentration levels, 5, 10, 30 and 50 pg/mg.

	%CV	%E	%R	%ME	%PE
5 pg/mg	6	−1	39	69	27
10 pg/mg	12	−1	36	65	23
30 pg/mg	9	−3	33	102	34
50 pg/mg	8	+4	35	92	32

No carry‐over was observed, even after the extraction of the highest calibrator. The signal obtained from this sample was indistinguishable from that obtained from a blank sample.

3.3. Authentic Case and Proficiency Test Results

To evaluate the applicability of the method, 95 hair samples from forensic toxicology cases were prepared according to the developed MEPS method for EtG analysis. Eighty‐five samples resulted negative, meaning below the LOQ, whereas five sample had EtG concentrations between 5 and 30 pg/mg, and 15 samples had concentrations above 30 pg/mg.

The method was also applied to proficiency test samples provided from the SoHT. Sample A and C results were consistent with the target values with good z‐scores, whereas sample B calculated concentration exceeded the evaluation limit given by the SoHT.

Table 3 summarizes the results of the SoHT proficiency tests together with the target values, evaluation limits and z‐scores provided by the SoHT.

TABLE 3.

SoHT proficiency test results for the samples analysed (A, B and C), compared to target values, evaluation limits and z‐scores provided by the SoHT.

Sample	MEPS‐LC–MS/MS result (pg/mg)	Target value (pg/mg)	Evaluation limit (pg/mg)	z‐score
A	16.0	13.75	5.35–22.15	0.53
B	25.0	12.51	4.75–20.27	3.21
C	47.0	58.57	29.83–87.31	−0.80

4. Discussion

The determination of EtG in hair samples can be challenging, due to its low accumulation in keratin matrix and to its low molecular weight, affecting its chromatographic behaviour. Therefore, sample preparation is highly recommended to achieve adequate sensitivity for the analysis whilst reducing the matrix effect that can affect its determination. Additional issues are related to cutting or pulverising the matrix for extraction. The SoHT recommended powdering hair for EtG analysis. In fact, powered hair extraction has been widely proven to be more effective than cutting [13, 24, 25]. On the other hand, some authors stated that a longer incubation time can result in higher and in some cases even similar EtG extraction percentages from cut and powdered hair [25, 26, 27]. The incubation procedure proposed in this study for cut hair include overnight incubation (at list 18 h) and a subsequent 2 h extraction under sonication. In addition, our main focus was to develop a technique able to enhance the sensitivity for EtG determination, consequently efficiency for its extraction from hair matrix was not evaluated.

MEPS, like traditional SPE, is a simple and rapid technique, highly efficient in extracting and purifying analytes from complex matrices. Being a microextraction technique, it has the advantage of working with low solvents and sample volumes. A further advantage is that MEPS can be used repeatedly, on authentic biological samples, without loss of extraction efficiency, depending on the specimen extracted. In case of hair aqueous extraction medium, a quite clean specimen, it can be re‐used up to a hundred times.

The method was validated showing good results for the parameters evaluated. In particular, accuracy (%E) and repeatability (%CV) are very good, always below 12%. Linearity range (R ² 0.991) includes the cut‐offs proposed by the SoHT for differentiation of social drinkers/repeated alcohol consumption and of excessive alcohol consumption.

MEPS reduces the matrix effect, which is a critical parameter for EtG analysis due to its low retention time during the chromatographic run and the consequent potential for co‐elution with the matrix components.

The recoveries were between 33% and 39%, but the high enrichment factor for the extraction, which is 9, together with the reduction of the matrix effect, led to good sensitivity and adequate limits of determination and quantification. Low recovery rates, sometimes even lower than the one obtained in this study, were reported also in previous publications in which MEPS was applied for the extraction of other classes of analytes [15, 17, 18, 20, 22]. No carry‐over was detected when analysing a blank after the highest calibration point. Anyway, repeating the activation and conditioning step for each extraction avoids the possibility of any presence of the analyte in the dead volume of the syringe for the subsequent analysis, guaranteeing the reusability of MEPS without carry‐over phenomena.

The method was applied to authentic hair samples from forensic caseworks. In 10 samples, EtG was below the LOQ (5 pg/mg), which is consistent with alcohol abstinence, according to the SoHT cut‐off. In one case the concentration was 25 pg/mg, indicating repeated/social alcohol consumption in the period corresponding to the length of the segments analysed, and in another one the concentration was 36 pg/mg, which suggests chronic excessive alcohol consumption. Figure 2 shows the chromatograms obtained for an authentic hair sample. The method was applied also to three proficiency test samples. The evaluation of the concentration obtained for two of them, in term of comparison with target values and z‐scores, was good. The results for the third one was out of the evaluation limit given by the SoHT, but the calculated value fell anyway within the repeated/social alcohol consumption range.

FIGURE 2.

Ionic chromatograms of EtG and its internal standard, EtG‐D5, in an authentic hair sample (calculated concentration 25 pg/mg).

5. Conclusions

A simple, fast and reliable MEPS‐UPLC‐MS/MS method has been developed and validated for the determination of EtG in hair samples. The combination of a microextraction technique for sample pretreatment and tandem mass spectrometry resulted in a sensitive EtG determination. In addition, the method was able to cope with low solvent volumes. The linearity range, LOD and LOQ were adequate for forensic purposes and allowed the differentiation between social drinkers and chronic excessive drinkers, according to the cut‐off established by the SoHT. In addition, the method validation showed good reproducibility and accuracy, with satisfactory recovery and matrix effect reduction.

The method has been successfully applied both to authentic hair from forensic caseworks and to proficiency test samples.

To the best of our knowledge, this is the first application of MEPS for EtG analysis.

Conflicts of Interest

The authors declare no conflicts of interest.

Odoardi S., Mestria S., Valentini V., Biosa G., and Rossi S., “Rapid and Effective Determination of Ethyl Glucuronide in Hair by Micro Extraction by Packed Sorbent (MEPS) and LC‐MS/MS,” Drug Testing and Analysis 17, no. 7 (2025): 1160–1166, 10.1002/dta.3824.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.