Screening of Western Drug Adulterants in Complementary Health Products by High Performance Liquid Chromatography/Diode Array Detection/Mass Spectrometry

Gas chromatography/mass spectrometry (GC/MS) and high performance liquid chromatography/diode array detection (HPLC/DAD) are the techniques used for screening most of the common western drug adulterants in complementary health products. However, co-elution with poor baseline separation of peaks is commonly observed with complex sample matrices, thus posing a challenge for identifying any suspected drugs. Hence, we see the need to develop a high performance liquid chromatography/diode array detection/mass spectrometry (HPLC/DAD/MS) screening method based on the detection of the molecular ion, ultraviolet (UV) spectra and the retention time of each drug. This method was successfully established for 213 drugs, together with their respective limit of detection (LOD). An In-house UV library and In-house MS library, both comprising more than 200 drug substances, were also successfully built and used in the HPLC/DAD/MS screening of western drug adulterants in complementary health products.

1. Introduction

In Singapore, complementary health products generally include the Chinese proprietary medicines, traditional medicines and health supplements. These products are widely used to maintain or improve health, with public beliefs that the complementary health products are safe and have no side effects [1]. One of the greatest safety concerns of complementary health products is adulteration with undeclared pharmaceutical drugs and their analogues, in illicit attempts to evade detection [2,3,4]. It represents a problem in product quality and is one of the major causes for adverse events [5].
The more commonly used techniques for screening western drug adulterants in complementary health products are gas chromatography/mass spectrometry (GC/MS) and high performance liquid chromatography/diode array detection (HPLC/DAD) [6,7]. However, HPLC/DAD screening method commonly encountered issue of co-elution with poor baseline separation of peaks especially for complex sample matrices, thus posing a challenge for identifying any suspected drugs.
Due to its desirable sensitivity and selectivity, liquid chromatography/mass spectrometry (LC/MS) method has been applied for screening purposes too in health supplements [8,9,10]. As health supplements have complicated matrices, LC/MS would be a more suitable screening technique since it is highly selective and is capable of providing additional information about the molecular mass of the drugs. HPLC method coupled with diode array detection and electrospray ionisation mass spectrometry (HPLC/DAD/MS) could potentially eliminate false negative results and enhance the screening capability. A few reports focused on screening of a limited number of western drug adulterants by HPLC/DAD/MS [9,11]. There has not been a report on simultaneous screening of a wide range of western drug adulterants in complementary health products by HPLC/DAD/MS so far. Thus, a specific method that can simultaneously screen a wide range of western drug adulterants in complementary health products needed to be developed.
In this study, a HPLC/DAD/MS method was developed as a screening method for a wide range of western drug adulterants in complementary health products based on the detection of the molecular ion, ultraviolet (UV) spectra and the retention time of each drug in a single run. An in-house UV library and in-house MS library were also successfully built and used in the screening of western drug adulterants in complementary health products.

2. Experimental

2.1 Materials
The majority of the standard drugs (150 drugs) used were purchased from United States Pharmacopoeia (Rockville, MD, USA), British Pharmacopoeia (Teddington, Middlesex, UK) or the European Pharmacopoeia (Strasbourg, France). The remaining 63 secondary standard drugs were acquired from TLC Pharmaceutical Standards Ltd. (Aurora, Ontario, Canada) or Sigma-Aldrich (St. Louis, MO, USA). All 213 standards were prepared in methanol to achieve a concentration of 1 mg/mL. They were then further diluted to 0.1 mg/mL using methanol for injection into the HPLC/DAD/MS system.

2.2 HPLC/DAD/MS screening
An Agilent 1260 series Infinity II HPLC chromatograph with diode-array detector and coupled with single quadrupole mass spectrometer detector (MSD) 6135B with electrospray ionisation (ESI) (Waldbronn, Germany) was used for the analysis. The column used was a Thermo Scientific Accucore C18, 150 mm x 2.1 mm, 2.6 µm (PA, USA). The UV spectra from 200 to 400 nm were recorded online during the chromatographic run. The mass spectra were acquired at a mass range from m/z 100 to 1000. The method conditions are listed in Table 1.

2.3 Sample preparation
The ten complementary health products selected for the study were in the form of capsules, granules, pills, powder and liquid. Capsules were opened and the contents were used for analysis. Granules and pills were ground before use. Powder and liquid samples were directly used for analysis. All the samples were thoroughly homogenised before the test. About 1 g of the homogenised sample, pre-spiked with selected western drug adulterants, was transferred to a test-tube and 10 mL of methanol was added. The mixture was sonicated in an ultrasonic bath for 15 minutes and filtered using a 0.45 µm PTFE membrane filter, which had been validated and showed no adsorption, for
HPLC/DAD/MS analysis.

3. Results and Discussion

3.1 Development of UV and MS library
The UV spectra of the 213 standard drugs were obtained using DAD and compiled as an UV library using the OpenLAB CDS ChemStation Data Analysis software (Agilent Technologies). The screening of the drugs in complementary health products was carried out by library search using the same software. The UV spectra corresponding to the peaks in the unknown sample were compared with those in the library. Library matches of UV spectra were automatically calculated for each peak and a score of 1000 represented a perfect match, while a score of below 900 represented a poor match [12]. A peak identification result with a library match above 950 could be considered as identification with good certainty. For the LC/DAD library search, the retention time window was set at ± 20% and the match threshold was set at 900. For those peaks with close retention times (± 1 min) to those drug substances suggested by the In-house library and library match ≥ 900, the UV spectra of the peaks would be matched with those of the reference drug substances from the library to identify the presence of any adulterants.
The MS spectra of the 213 standard drugs were obtained using the single quadrupole mass spectrometer. The mass-to-charge ratio (m/z) of the molecular ion and retention time of the 213 drugs were compiled as a MS library using Microsoft Access. The screening of the drugs in Complementary Health Products was carried out by library search using our In-house developed program ‘MS Library Search Report Generator’ with retention time window set at ± 1 min and ion mass window set at ± 0.2 Da. The m/z and retention time corresponding to the peaks in the unknown sample were compared with those in the library. Library matches were automatically calculated for each peak and a score of 100 represented a perfect match. For those peaks with close retention times (± 1 min) and m/z within ± 0.2 Da to those reference drug substances suggested by the In-house LCMS library, they would be listed as suspect adulterants.
Table 2 shows the drug name, retention time, molecular ion and limit of detection (LOD) of the 213 drugs in the UV and MS library.
Co-elution of two or more compounds remains one of the major causes of errors in the HPLC/DAD screening method. Thus, the development of the in-house UV library and in-house MS library would be very useful in the HPLC/DAD/MS screening of western drug adulterants in Complementary Health Products since MS is a much more selective technique than DAD.

3.2 Screening of western drug adulterants in Complementary Health Products
To validate the HPLC/DAD/MS screening system, two positive samples (each containing one drug) and eight negative complementary health product sample matrices were selected. A total of fourteen drugs were spiked into the eight negative complementary health product sample matrices as unknowns. Table 3 showed the list of drugs that were detected in the positive samples (Product 1 and 2), and the list of drugs with their spiked concentrations in the negative samples (Product 3-5, 7-8 and 10). There were two negative samples without any drugs being spiked in (Product 6 and 9). All the ten samples were then analysed using the HPLC/DAD/MS screening system and the results were as shown in Table 3. The LC/DAD library search reports were able to successfully detect and identify fourteen out of the sixteen drugs with a library match score of more than 990, indicating a good match of UV spectra with those drugs in the UV library. Two of the drugs, namely Piroxicam in Product 4 and Hydroxyhomosildenafil in Product 7, were not detected by the LC/DAD library search report due to co-elution with other peaks at the same retention time. However, when MS technique was used in the screening, Piroxicam in Product 4 and Hydroxyhomosildenafil in Product 7 were successfully detected and identified in the LC/MS library search reports with a perfect score of 100.
For illustration, two drugs namely, Desmethylsildenafil and Hydroxyhomosildenafil with close retention times at about 10.8 min as shown in Figure 1A and identical UV spectra as shown in Figure 1B, were spiked at 10 mg/L into Product 7. The LC/DAD library search report was not able to identify Hydroxyhomosildenafil as shown in Figure 1D. The presence of m/z 461.1 in Figure 1C, represents the molecular ion [M + H]+ of Desmethylsildenafil, while
the ion at m/z 505.1 in Figure 1C corresponds to the molecular ion [M + H]+ of Hydroxyhomosildenafil. Since both molecular ions at m/z 461.1 and 501.1 were observed in the MS spectra, this indicated that both drugs were present in Product 7. The LC/MS library search report was able to correctly identify Desmethylsildenafil and Hydroxyhomosildenafil with a good match quality as shown in Figure 1E, even though these two drugs co-eluted at the same retention time with identical UV spectra. The HPLC/DAD/MS screening method proved to be successful in the identification of all sixteen drugs in the ten samples during
the study.
The limit of detection (LOD) was established as the lowest concentration at which a signal-to-noise ratio (S/N) of at least 3 was obtained for UV detection. The limit of detection (LOD) of 213 drugs were determined to be between 5 and 100 mg/L as shown in Table 2. 121 drugs had the lowest LOD of 5 mg/L. Hydralazine and Hyoscine N-Butylbromide had the highest LOD of 100 mg/L. Most of the LOD of the 213 drugs (92%) fall within the range of 5 – 20 mg/L, which is considered a reasonable limit for adulterant screening.
Due to the complexity of the matrices in complementary health products, and the co-elution of matrix and drug compounds, the UV spectra of the drugs are not be able to match the UV spectra of the drugs in the UV screening library, resulting in false negative results. It is very important therefore to use a highly selective LC/MS screening technique based on the detection of the molecular ion to complement the DAD technique, and hence avoiding false negative results.
A good data processing software programme is also required for the screening of western drug adulterants in complementary health products to conduct auto-library searches against the in-house UV and MS screening libraries. The OpenLAB CDS ChemStation Data Analysis software can do auto-library search against the UV library with retention time included in the matching, however such auto-library search function is not available for the MS library. It would be ideal to have a powerful data processing software suite that could generate the UV and MS library search results together in the same report, which could greatly ease data processing and interpretation.   

4. Conclusion

The newly developed HPLC/DAD/MS screening method was successfully established for 213 drugs, together with their respective limit of detection (LOD) in complementary health product matrices. The study demonstrated the application of an in-house UV and MS library for effective screening and identification of 213 drugs without the need of standards for each analysis.

5. References

1.    S.R.B.M. Eussen, H. Verhagen, O.H. Klungel, J. Garssen, H.van Loveren, H J.van Kranen, C.J.M. Rompelberg, Eur. J. Pharmacol. 668 (2011) S2.
2.    E. Ernst, J. Intern. Med. 252 (2002) 107.
3.    T. Rocha, J.S. Amaral, M.B.P.P. Oliveira, Compr. Rev. Food Sci. Food Saf. 15 (2016) 43.
4.    J. Calahan, D. Howard, A.J. Almalki, M.P. Gupta, A.I. Calderon, Planta Med. 82 (2016) 505.
5.    P. A. Chyka, Am. J. Med., 2000, 109, 122–130.
6.    12. S.H. Kim, J. Lee, T. Yoon, J. Choi, D. Choi, D. Kim, S.W. Kwon, Biomed. Chromatogr. 23 (2009) 1259.
7.    S.Y. Liu, S.O. Woo, H.L. Koh, J. Pharm. Biomed. Anal. 24 (2001) 983.
8.    J. Wang, B. Chen, S. Yao, Food Addit. Contam. A 25 (2008) 822.
9.    X. Zhu, S. Xiao, B.O. Chen, F. Zhang, S. Yao, Z. Wan, D. Yang, H. Han, J Chromat. A 1066 (2005) 89.
10. S.R. Gratz, C.L. Flurer, K.A. Wolnik, J Pharm Biomed Anal. 36 (2004) 525.
11. Z. Huang, S. Xiao, D. Luo, B. Chen, S. Yao, J Chromat Sci, 46 (2008) 707.
12. L. Huber, S. George, Diode Array Detection in HPLC, Marcel Dekker, New York, 1993.