# Detection of phenobarbital adulteration in dietary supplements: simultaneous analysis of 16 sedative-hypnotics and sleep-inducers by ultra-high-performance liquid chromatography with UV detection (UPLC-UV) and quadruple Orbitrap mass spectrometry (Q-Orbitrap-MS)

• Lee, Ji Hyun (Division of Advanced Analysis, National Institute of Food and Drug Safety Evaluation, Ministry of Food & Drug Safety, Osong Health Technology Administration Complex) ;
• Choi, Ji Yeon (Division of Advanced Analysis, National Institute of Food and Drug Safety Evaluation, Ministry of Food & Drug Safety, Osong Health Technology Administration Complex) ;
• Park, Hanna (Division of Advanced Analysis, National Institute of Food and Drug Safety Evaluation, Ministry of Food & Drug Safety, Osong Health Technology Administration Complex) ;
• Min, Ah Young (Division of Advanced Analysis, National Institute of Food and Drug Safety Evaluation, Ministry of Food & Drug Safety, Osong Health Technology Administration Complex) ;
• Kim, Nam Sook (Division of Advanced Analysis, National Institute of Food and Drug Safety Evaluation, Ministry of Food & Drug Safety, Osong Health Technology Administration Complex) ;
• Park, Seong Soo (Division of Advanced Analysis, National Institute of Food and Drug Safety Evaluation, Ministry of Food & Drug Safety, Osong Health Technology Administration Complex) ;
• Park, Sung-Kwan (Division of Advanced Analysis, National Institute of Food and Drug Safety Evaluation, Ministry of Food & Drug Safety, Osong Health Technology Administration Complex) ;
• Kang, Ho-il (Division of Advanced Analysis, National Institute of Food and Drug Safety Evaluation, Ministry of Food & Drug Safety, Osong Health Technology Administration Complex)
• Accepted : 2019.01.15
• Published : 2019.02.25
• 447 35

#### Abstract

The safety of food is occasionally questionable, as there have been some reports of products contaminated with illegal adulterants. In this study, the presence of 16 sedative-hypnotics and sleep inducers in dietary supplements was determined by ultra-high-performance liquid chromatography with UV detection (UPLC-UV) and quadruple Orbitrap mass spectrometry (Q-Orbitrap-MS). The UPLC method was validated, providing a linearity (R2) of more than 0.999, and LODs and LOQs that ranged from 0.2 to 0.5 and 0.6 to $1.5{\mu}g\;mL^{-1}$, respectively. The repeatabilities were 0.2-8.4 % (intra-day) and 0.3-4.5 % (inter-day), and the accuracies were 89.0-117.0 % (intra-day) and 87.8-111.9 % (inter-day). The mean recoveries of the spiked samples ranged from 98.7 to 107.3 %. The relative standard deviation (%RSD) of the stability was less than 2.4 %. Using the developed method, one sedative-hypnotic compound, phenobarbital, was detected in one of the nineteen samples tested. In addition, the major characteristic fragment ions of each target compound were confirmed using Q-Orbitrap-MS for higher accuracy. Monitoring the presence of these 16 sedative-hypnotics and sleep inducers in dietary supplements should be pursued in the interest of human health, and the results of this study confirmed that the developed method has value for this application.

# 1. Introduction

Over the past few years, illegal adulteration with sedative-hypnotic and sleep-inducer compounds suchas diazepam, chlordiazepoxide, nitrazepam, clonazepamand estazolam has been routinely detected in dietary supplements without labelling.1-4 Sedative-hypnotic medications represent a variety of chemically distinct groups of compounds, including barbiturates, benzo-diazepines, and selective non-benzodiazepine hypnotics.

These medications shared a general sedative profile of clinical effects, but they differ in other pharmacological properties, including their therapeutic profile, safety, and potential to produce adverse behavioral effects, such as abuse or dependence. Benzodiazepines and selective hypnotics have a more favorable efficacy and safety profile, and their safety profile has led to adecrease in the prescription and clinical issues associated with the use of older agents, such as barbiturates.5-6

However, taking sedative-hypnotics and sleepinducers can induce drowsiness, dizziness, weakness, respiratory depression and other side-effects, whilelong-term use may lead to tolerance, dependence and addiction. These side effects have been used for the purposes of criminal activity, such as robberies and rape. But some manufacturers adulterate theirmerchandise with sedative-hypnotics and sleep inducersin order to achieve high profits and deceive consumersinto thinking that they are experiencing the claimed health benefits by enhancing short-term effects.7 Therefore, there is an urgent need to develop analytical methods with high sensitivity and high selectivity toscreen for the presence of sedative-hypnotic and sleep inducer compounds in dietary supplements.

The analytical methods hitherto used for the detection of sedative-hypnotics and sleep inducers in dietary supplements include HPLC,4,13 GC-MS3,8,14 and LC-MS.1,9,15 However, the simultaneous analysis of morethan 16 compounds in dietary supplements using UPLC has not yet been reported.

Therefore, we conducted qualitative analyses with two ‘state-of-the-art’ instruments. First, qualitative analysis of 16 sedative-hypnotics and sleep inducers was conducted via UPLC. LC with UV detection has advantages including its low cost, and straight forwardapplication and interpretation. The equipment is of ten standard in medicine control laboratories and readily available in most laboratories.10 UPLC buildsupon the well-established principles of LC, but usessub-2-μm porous particles. These particles operate atelevated mobile phase linear velocities to producerapid separation with increased resolution. Theseattractive features prompted us to develop a rapid, sensitive and specific UPLC assay method for thesimultaneous determination of multiple sedative-hypnotics and sleep inducers as possible adulterants in dietary supplements. Second, we wanted to usequadruple Orbitrap mass spectrometry (Q-Orbitrap-MS) to identify the detected substances once again toperform an accurate and in-depth analysis. The major characteristic fragment ions were confirmed using Q-Orbitrap-MS for higher accuracy. Furthermore, after confirming the presence of the sedative-hypnotics and sleep inducer compounds, quantitative analysis was conducted using UPLC.

In this study, we developed and validated a UPLC-UV method for the identification and quantification of 16 sedative-hypnotics and sleep inducer compoundsin illegally adulterated dietary supplements. In addition, most of the other studies have analysed single or afew compounds using UPLC. The advantages of ourmethod is the first to carry out simultaneous analysis of more than 16 sedative-hypnotics and sleep inducercompounds using UPLC, which is the most commonly used for analysis. The results of this study are expected to be useful in confirming the safety andeffectiveness of sedative-hypnotics and sleep inducercompounds.

# 2. Materials and Methods

## 2.1. Standards and reagents

Alprazolam, Clemastine, Clonazepam, Diphen-hydramine, Estazolam, Flunitrazepam, Flurazepam, Hexobarbital, Lorazepam, Mequitazine, Midazolam, Pentobarbital, Phenobarbital, Temazepam, Triazolamand Zolpidem were obtained from the following companies: U.S. Pharmacopeia (Rockville, MD, USA), Lipomed AG (Arlesheim, Switzerland), Merck(Darmstadt, Germany), Cayman (Ann Arbor, MI, USA) and TRC (Toronto, Canada). Sodium phosphate and phosphoric acid were purchased from Sigma-Aldrich (St. Louis, MO, USA). High-purity de ionised water was prepared using a Milli-Q purification system (Millipore, Bedford, MA, USA). HPLC-grade methanoland acetonitrile (ACN) were obtained from Burdic and Jackson (Muskegon, MI, USA). Stock solutions (1000 μg mL−1) were prepared by dissolving each compound in methanol. Each stock solution was stored at 4 °C, and working mixtures were prepared daily by diluting the stock solutions in methanol to a concentration appropriate for the calibration curves and method validation.

## 2.2. UPLC conditions

A Waters Acquity UPLC system (Milford, MA, USA) equipped with binary pumps, a sample manager, a column oven and a photodiode array (PDA) detectorwas used for all analyses. Chromatographic separation was performed using an Acquity UPLC HSS T3 column (2.1 × 150 mm, 1.8 μm; Waters, Milford, MA, USA), and the elution peaks were detected by UV absorption at 210 nm. The column temperature was maintained at 40°C, and the flow rate was 0.2 mL min− 1, with an injection volume of 1 μL. The mobile phase consisted of (A) 0.5 mM sodium phosphate indeionised water (DW) with 0.1 % phosphoric acid, and (B) 95 % ACN. The gradient program proceeded as follows: 0 min, 32 % B; 2 min, 32 % B; 15 min, 50 % B; 15.1 min, 100 % B; 17 min, 100 % B; 17.1 min, 32 % B; 20 min, 32 % B.

## 2.3. Sample preparation

Dietary supplements were purchased from various markets. Notably, these products were imported to South Korea or sold via online sites or offline stores. A total of 19 samples were obtained and analysed. The samples were homogenised with a blender. Each sample (1 g) was placed into a 50 mL volumetric flask and dissolved in 100 % methanol. The obtained mixture was extracted in an ultrasonic bath for 30 min and then further methanol was added to the 50 mL in volumetric flask after cooling. The extract was subsequently filtered through a 0.22 μm PTFE filter (Millipore) and injected into the UPLC and Q-Orbitrap-MS systems.

## 2.4. Method validation

A series of analyses to evaluate the specificity, linearity, stability, repeatability, limits of detection (LOD) and quantification (LOQ) were conducted tovalidate the performance of our method. Standardsolutions containing the 16 sedative-hypnotic and sleep-inducer were prepared and diluted with methanolto appropriate concentrations for the construction of calibration curves. Calibration curves were developed by plotting the peak areas versus the corresponding concentrations of each analytes. The repeatability of the method was evaluated by analysing the 16 standard compounds. The %RSD of the peak areas was used to evaluate the repeatability of the method. The LOD and LOQ for each analyte were determined at a signal-to-noise ratio (S/N) of approximately 3 and 10, respectively. A recovery test was used to evaluate the accuracy of this method by adding the corresponding standard compounds at low (near the LOQ), medium (~10 fold above the LOQ) and high (~50 fold above the LOQ) concentrations to each dietary supplement, which had been analysed previously. The mixtures were extracted and analysed using the a forementioned method in triplicate.

## 2.5. Mass spectrometry conditions

The experiment was performed with a Q-Exactive Orbitrap mass spectrometer (Thermo Scientific, San Jose, CA) coupled to a Thermo Dionex UltiMate 3000 LC. Target compounds were separated with a BEH-C18 (100 × 2.1 mm i.d., 1.7 μm) column maintained at 40°C. The mobile phases consisted of 0.1% formic acid in both DW (A) and acetonitrile (B). The injection volume was 1 µL and the flow rate was 0.25 mLmin−1. Elution was conducted with the following gradient profile: 0.0-2.0 min (A: 80 %, B: 20 %), 2.0-7.0 min (A: 80-0 %, B: 20-100 %), 7.0-11.0 min (A: 0 %, B: 100 %), 11.0-11.1 min (A: 0-80 %, B: 100-20 %), and 11.1-13 min (A: 80 %, B:20 %).

Full MS/ddMS2 (data-dependent MS2) was imple-mented as the mass analysis mode. The mass calibration was performed according to the manufacturer's specifications. Data were obtained using X calibur3.0 software. The mass conditions were as follows: HESI ion source; positive ion mode, except forphenobarbital and pentobarbital; spray voltages of 3.5 kV(+) and 3.0 kV(−); capillary temperature of 320 ℃; sheath gas flow at 42 arbitrary units; auxiliary gas flow at 10 arbitrary units; probe heater temperatureset to 350 °C(+) and 300 °C(−); S-lens RF level of 50; resolution of 70,000 (full scan) and 17,500 (MS/MS); automatic gain control (AGC) target of 3 × 106 (full scan) and 1 × 105 (MS/MS); scan range from m/z 50 to 1000; maximum infusion time (IT) of 100 ms (full scan) and 50 ms (MS/MS); single microscancount; loop count of 5; MSX count of 1; Top N of 5;isolation window of 4 m/z; underfill ratio of 1.0 %;intensity threshold of 2.0 × 104; isotope exclusion; dynamic exclusion of 10.0 s.

# 3. Results

## 3.1. Optimization of sample preparation

In order to develop the optimal sample preparation conditions, the extraction parameters, solvent proportion, sample weight and sonication time were considered. The recovery under each set of conditions was determined as the fortified standardarea, where the 16 sedative-hypnotic and sleep-inducers were fortified with a precisely known amount prior to extraction.

The recovery efficiency was greater with methanolextraction than with ethanol. The use of 100 % methanol led to a better recovery than either 70 % or 50 % methanol (v/v). The sonication time was varied from 10 to 60 min; sonication of most compounds for 30 min led to superior recovery. The sample preparation optimization results are shown in Table 1.

Table 1. Recovery of each compound under different extraction conditions

## 3.2. Optimization of UPLC conditions

To achieve a successful separation of the 16 sedative-hypnotics and sleep-inducers, we evaluated several mobile phases and buffers, including sodiumphosphate, phosphoric acid and potassium phosphate. In phosphoric acid and potassium phosphate buffers, we observed that an excellent resolution was achievable in 0.5 mM sodium phosphate in deionized water (DW) with 0.1 % phosphoric acid.

We also evaluated several stationary phase columns, and we found that the Acuity UPLC HSS T3 columngave better separation, peak shapes, and resolution compared with either a BEH C8 (1.7 μm, 2.1 × 150mm; Waters, Milford, MA, USA) or an HSS C18 column (1.8 μm, 2.1 × 100 mm; Waters, Milford, MA, USA); the latter two both showed broad peaks with poor analyte resolution. Therefore, we chose the HSS T3 column for further analysis. The UPLC chromatogram and UV spectra of a standard mixture of the 16 sedative-hypnotics and sleep-inducers areshown in Fig. 1.

Fig. 1. LC chromatogram of the 16 sleep inducers analysed in this study and the individual PDA spectra of each compound.

## 3.3. Method validation

The developed method was validated in accordance with the guidelines established by the AOAC, CODEX, FDA and at the ICH. The performance of the method was evaluated by estimation of the specificity, LOD, LOQ, linearity, recovery, and repeatability.

### 3.3.1. Specificity

The specificity of the method was guaranteed by comparing the retention times (RTs) of the samples with those of reference materials in blank samples. Fig. 1 shows the individual chromatograms of the 16 sedative-hypnotics and sleep-inducers, which do notexhibit significant matrix interferences at theirrespective RTs. The developed UPLC method iscapable of separating all analytes under the givengradient conditions within 20 min.

### 3.3.2. Linearity, LOD, and LOQ

The calibration plots based on linear regression analysis revealed good linear relationships between the response and six different concentrations between 0.6-30 μg mL−1, based on the LOQ of the 16 sedative-hypnotics and sleep-inducers. Acceptable linearity with R2 values between 0.9990 and 1.0000 (Table 2) is obtained.

The LODs and LOQs for all analytes were defined at a S/N of 3:1 and 10:1, respectively. The LOD and LOQ values are presented in Table 2. The LODs and LOQs of 16 sleep-inducing compounds range from 0.20 to 0.50 and 0.60 to 1.50 μg mL−1, respectively.

Table 2. Summary of the calibration curves, limits of detection (LOD), and limits of quantification (LOQ) for 16 sleep inducers by UPLC

### 3.3.3. Repeatability and accuracy

The intra-day and inter-day repeatabilities wereassessed using the relative standard deviation (RSD, %) at low, medium and high concentrations. Theaccuracy of the method was determined as therecovery (%) of the theoretical concentration of the target compounds. Intra-day assays were carried outin triplicate using samples of low, medium and high concentrations on the same day, and inter-day assays were carried out in triplicate using samples of low, medium and high concentrations on three separatedays. The intra- and inter-day repeatability values are 0.2-8.4 % and 0.3-4.5 %, respectively, and the intra-and inter-day accuracy values are 89.0-117.0 % and 87.8-111.9 %, respectively (Table 3).

Table 3. Intra-and inter-day variation in the three concentrations of the 16 sleep inducers using UPLC

### 3.3.4. Recovery

The recovery (%) of each compound was calculated by comparing it to the response for the true concentration of the liquid and solid reference standards. Each sample was analysed at the same concentration threetimes. As shown in Table 4, the mean recoveries of the solid and liquid samples are 99.3-105.0 % and 98.7-107.0 %, respectively. The %RSD values are less than 2.4 %, which is within the acceptable limit (15 %; UNODC 2009). In solid and liquid samples, therecoveries tend not to be affected by matrix effects.

Table 4. The recovery efficiency of the 16 sleep inducers from dietary supplement samples using UPLC

### 3.3.5. Stability

The stability of each compound was measured several times over 48 h, after incubation in the sample solution. The stability was assessed using three different concentrations of samples stored atroom temperature for 6 h, and samples in autosamplervials stored at 4 °C for 48 h. All stability samples were analysed in triplicate.

The %RSD values were taken as indicators of the stability of the compounds analysed by the analytical method. The %RSD values are < 8.4 % (Table 5). The results indicate that the prepared sample solutions are sufficiently stable at room temperature for 6 h. Also, the sample solutions could be stored at 4 ℃ for up to 48 h, and no stability-related problems would be expected during routine analysis of the 16 sedative-hypnotics and sleep-inducers.

Table 5. Stability of the 16 sleep inducer compounds over 48 h analysed with UPLC.

### 3.3.6. Confirmation by Q-Orbitrap

For the 16 sedative-hypnotics and sleep-inducers, protonated [M+H]+and deprotonated [M-H] molecules were observed in Q-Orbitrap in positive and negative ion mode. The exact masses and RTs of the 16 sedative-hypnotics and sleep-inducers were obtained from a database. It is necessary to identify the majorcharacteristic fragment ions present in the mass spectra, if these spectra are to be used for qualitative analysis. Towards this end, reference MS spectra of the target compounds were studied and the fragmentation pathways leading to the formation of the majorcharacteristic fragment ions were investigated. The Q-Orbitrap method was demonstrated as suitable fordetecting the target compounds, evaluated using a combination of RT, mass accuracy, and fragmentation.

### 3.3.7. Application in real samples

19 samples were collected from online and offlinemarkets; their adulteration with 16 sedative-hypnotics and sleep-inducers were evaluated using the UPLC-UV and Q-Orbitrap methods. First, qualitative analysis of the sleep-inducing compounds in the food samples was conducted via UPLC-UV. After qualitative analysis for screening of the presence of the sleep inducers infood samples by UPLC-UV, confirming the result of qualitative analysis by Q-Orbitrap. The risk of a false-positive was minimised by applying several criteria, such as RT, mass accuracy and MS2 productions employed to identify the sedative-hypnotics and sleep inducer compounds present in dietary supplements. Fig. 2 show the representative XIC's of a phenobarbital standard and phenobarbital in apositive sample. The MS spectra of the samples that contained one sleep-inducing compound, phenobarbital, and phenobarbital standards are shown in Fig. 3.

Phenobarbital was detected as an illegal adulterant. The RT, accurate mass and MS2 product ions confirmed the presence of phenobarbital in the positive samples. After confirming the presence of the sleep inducing compounds in food samples, quantitative analysis was conducted using UPLC-UV. It is important tonote that the positive samples were illegally adulterated with phenobarbital at high levels (24.45 mg g−1) in this study. Therefore, for safety, these screening results should be publicised. Furthermore, these results suggest that this sort of health and food productmonitoring should be continued in order to safeguard human health.

Fig. 2. XIC of a phenobarbital standard (a) and 500-fold diluted positive sample (b).

Fig. 3. Mass spectra of a phenobarbital standard (a) and a sample that contains phenobarbital (b).

# 4. Discussion

Previously, most of the other studies have analysed single or a few sedative-hypnotics and sleep-inducing compounds using UPLC; these methods cansimultaneously determine of sedative-hypnotics and sleep-inducers in illegally adulterated dietary supple-ments. In this method, the successful separation produced by UPLC-UV, in addition, the majorcharacteristic fragment ions were confirmed using Q-Orbitrap-MS for higher accuracy. The UPLC-UV separation was achieved on a HSS T3 column by using mobile phase of 0.5 mM sodium phosphate indeionized water (DW) with 0.1 % phosphoric acidand acetonitrile of a gradient elution mode. The optimized method was validated for specificity, linearity, LOD, LOQ, repeatability, accuracy, recovery and stability according to ICH guideline. The developed method was successfully applied to determine sedative-hypnotics and sleep-inducing compounds in dietary supplements without any interference. The results demonstrated that the values were within the acceptablerange.

# 5. Conclusions

We attempted to detect the presence of 16 sedative-hypnotics and sleep inducers in dietary supplements advertised to improve sleep functions, by developing and fully validating a sensitive, accurate, and selective UPLC-UV method. Nineteen representative samples were screened by UPLC-UV and Q-Orbitrap. Only one sample was tested positive for illegal adulteration with a sleep-inducing compound, phenobarbital, athigh levels. The novel UPLC-UV method has been proven to be a very promising and powerful method for routine screening of illegal adulterated sedative-hypnotics and sleep inducers in dietary supplements, ensuring food safety and public human health.

## Funding

This work was supported by the Ministry of Food and Drug Safety [grant number 15181MFDS521].

#### Acknowledgement

Supported by : Ministry of Food and Drug Safety

#### References

1. Q. Gao, B. Che, Z. Zhang, H. Dai, and H. Guo, J. Chin. Pharm. Univ., 39, 142-146 (2008).
2. B. Zhu, Z. Zhao, Z. Wang, Q. Xue, and L. Zhao, Chin. J. Pharm. Anal., 30, 745-751 (2010).
3. J. Liu, L. Xiao, X. Guan, Y. Xiong, T. Wang, and J. Li, Chin. J. Pharm. Anal., 31, 376-379 (2011).
4. P. Jin, Z. Wang, W. Wang, D. Zou, J. Ma, and X. Hu, Chin Pharm. J., 47, 462-466 (2012).
5. L. J. Greenfield, S. H. Zaman, M. L. Sutherland, S. C. Lummis, M. I. Niemeyer, E. A. Barnard, and R. L. Macdonald, Neuropharmacology, 42, 502-521 (2002). https://doi.org/10.1016/S0028-3908(01)00196-4
6. C. Gottesmann, Neuroscience, 111, 231-239 (2002). https://doi.org/10.1016/S0306-4522(02)00034-9
7. J. Wang, B. Chen, and S. Yao, Food addit. Contam., 25, 822-830 (2008). https://doi.org/10.1080/02652030801946553
8. Z. Wang, J. Zhang, and Y. Zhang, J. Chin. Mass Spec. Soc., 30, 282-286 (2009).
9. A. Yan, X. Li, C. Xi, L. Zhang, S. Xia, G. Wang, B. Tang, and Z. Mu, Chinese J. Anal. Chem., 41, 509-516 (2013). https://doi.org/10.3724/SP.J.1096.2013.20694
10. E. Deconinck, P. Y. Sacre, P. Courselle, and J. O. De Beer, J. Chromatogr. Sci., 51, 791-806 (2013). https://doi.org/10.1093/chromsci/bmt006
11. AOAC Official Methods of Analysis. Internationalconference on harmonization of harmonized tripartiteguideline, validation of analytical procedures: text andmethodology.
12. AOAC International, Rockville, MD (2005).
13. ICH International conference on Harmonization. Harmonized tripartite guideline, validation of analyticalprocedures: text and methodology Q2
14. (R1). European Medicines Agency (EUMA), London, UK .
15. A, El-Gindy, Il Farmaco, 60, 745 (2005). https://doi.org/10.1016/j.farmac.2005.06.002
16. R. A. Menck, D. S. De Lima, S. C. Seulin, V. Leyton, C. A. Pasqualucci, D. R. Munoz, M. D. Osseltore, M. Yonamine, J. Sep. Sci., 35, 3361-3368 (2012). https://doi.org/10.1002/jssc.201200444
17. X. Zhang, Z. Lin, J. Li, Z. Huang, Y. Rao, H. Liang, J. Yan, and F. Zheng. Drug Testing and Analysis, 9, 588-595 (2017). https://doi.org/10.1002/dta.2002