Abstract
Objectives
Urine sample manipulation including substitution, dilution, and chemical adulteration is a continuing challenge for workplace drug testing, abstinence control, and doping control laboratories. The simultaneous detection of sample manipulation and prohibited drugs within one single analytical measurement would be highly advantageous. Machine learning algorithms are able to learn from existing datasets and predict outcomes of new data, which are unknown to the model.
Methods
Authentic human urine samples were treated with pyridinium chlorochromate, potassium nitrite, hydrogen peroxide, iodine, sodium hypochlorite, and water as control. In total, 702 samples, measured with liquid chromatography coupled to quadrupole time-of-flight mass spectrometry, were used. After retention time alignment within Progenesis QI, an artificial neural network was trained with 500 samples, each featuring 33,448 values. The feature importance was analyzed with the local interpretable model-agnostic explanations approach.
Results
Following 10-fold cross-validation, the mean sensitivity, specificity, positive predictive value, and negative predictive value was 88.9, 92.0, 91.9, and 89.2%, respectively. A diverse test set (n=202) containing treated and untreated urine samples could be correctly classified with an accuracy of 95.4%. In addition, 14 important features and four potential biomarkers were extracted.
Conclusions
With interpretable retention time aligned liquid chromatography high-resolution mass spectrometry data, a reliable machine learning model could be established that rapidly uncovers chemical urine manipulation. The incorporation of our model into routine clinical or forensic analysis allows simultaneous LC-MS analysis and sample integrity testing in one run, thus revolutionizing this field of drug testing.
Funding source: Emma Louise Kessler Foundation
Acknowledgments
The authors express their gratitude to Emma Louise Kessler, MD for her generous legacy she donated to the Institute of Forensic Medicine at the University of Zurich, Switzerland for research purposes.
Research funding: The Emma Louise Kessler Foundation directly funded Akos Dobay.
Author contributions: All authors have accepted responsibility for the entire content of this manuscript and approved its submission.
Competing interests: Authors state no conflict of interest.
Ethical approval: According to Swissethics (Humanforschungsgesetz), no further ethical approval from the cantonal ethics commission was necessary as the research was not aiming to investigate diseases or functions of the human body. A statement of the Cantonal Ethics Board of the Canton of Zurich (document BASEC-No. Req-2017-00946 and E-14/2009) was obtained.
Code availability:https://github.com/gbril/urine-adulteration.
References
1. Gmeiner, G, Geisendorfer, T. Urine manipulation with liquid soap: a case report. Drug Test Anal 2020;12:575–8. https://doi.org/10.1002/dta.2780.Search in Google Scholar PubMed
2. Lin, SY, Lee, HH, Lee, JF, Chen, BH. Urine specimen validity test for drug abuse testing in workplace and court settings. J Food Drug Anal 2018;26:380–4. https://doi.org/10.1016/j.jfda.2017.01.001.Search in Google Scholar PubMed
3. Fu, S. How do people try to beat drugs test? Effects of synthetic urine, substituted urine, diluted urine, and in vitro urinary adulterants on drugs of abuse testing. In: Dasgupta, A, editor. Critical issues in alcohol and drugs of abuse testing. Academic Press; 2019:359–89 pp.10.1016/B978-0-12-815607-0.00026-5Search in Google Scholar
4. Jaffee, WB, Trucco, E, Levy, S, Weiss, RD. Is this urine really negative? A systematic review of tampering methods in urine drug screening and testing. J Subst Abuse Treat 2007;33:33–42. https://doi.org/10.1016/j.jsat.2006.11.008.Search in Google Scholar PubMed
5. Fu, S. Adulterants in urine drug testing. Adv Clin Chem 2016;76:123–63. https://doi.org/10.1016/bs.acc.2016.05.003.Search in Google Scholar PubMed
6. Cody JT, Valtier, S. Effects of Stealth adulterant on immunoassay testing for drugs of abuse. J Anal Toxicol 2001;25:466–70. https://doi.org/10.1093/jat/25.6.466.Search in Google Scholar PubMed
7. Urry, FM, Komaromy-Hiller, G, Staley, B, Crockett, DK, Kushnir, M, Nelson, G, et al.. Nitrite adulteration of workplace urine drug-testing specimens. I. Sources and associated concentrations of nitrite in urine and distinction between natural sources and adulteration. J Anal Toxicol 1998;22:89–95. https://doi.org/10.1093/jat/22.2.89.Search in Google Scholar PubMed
8. Dasgupta, A, Wahed, A, Wells, A. Rapid spot tests for detecting the presence of adulterants in urine specimens submitted for drug testing. Am J Clin Pathol 2002;117:325–9. https://doi.org/10.1309/9q2g-6cth-xt16-hccc.Search in Google Scholar
9. Wu, AH, Bristol, B, Sexton, K, Cassella-McLane, G, Holtman, V, Hill, DW. Adulteration of urine by “urine Luck”. Clin Chem 1999;45:1051–7. https://doi.org/10.1093/clinchem/45.7.1051.Search in Google Scholar
10. Scholer, A. The effect of urine manipulation on substance abuse testing. ToxChem und KrimTech 2004;71:127–37.Search in Google Scholar
11. Huppertz, B, Bartling, C, Baum, K. Adulteration of urine samples, discovery and mitigation. J Appl Life Sci 2018;16:1–8. https://doi.org/10.9734/jalsi/2018/40427.Search in Google Scholar
12. Pham, AQN, Kelly, T, Fu, S. Urine adulteration: can bleach be used to mask MDMA use? Anal Methods 2013;5:3948–55. https://doi.org/10.1039/c3ay40543e.Search in Google Scholar
13. Dasgupta, A. The effects of adulterants and selected ingested compounds on drugs-of-abuse testing in urine. Am J Clin Pathol 2007;128:491–503. https://doi.org/10.1309/fqy06f8xktqpm149.Search in Google Scholar PubMed
14. Olivieri, B, Marić, M, Bridge, C. Determining the effects of adulterants on drug detection via enzyme-linked immunosorbent assay and adulterant tests strips. Drug Test Anal 2018;10:1383–93. https://doi.org/10.1002/dta.2404.Search in Google Scholar PubMed
15. Raouf, M, Bettinger, JJ, Fudin, J. A practical guide to urine drug monitoring. Fed Pract 2018;35:38–44.Search in Google Scholar
16. Dasgupta, A, Chughtai, O, Hannah, C, Davis, B, Wells, A. Comparison of spot tests with AdultaCheck 6 and Intect 7 urine test strips for detecting the presence of adulterants in urine specimens. Clin Chim Acta 2004;348:19–25. https://doi.org/10.1016/j.cccn.2004.01.003.Search in Google Scholar PubMed
17. Paul, BD, Jacobs, A. Spectrophotometric detection of iodide and chromic (III) in urine after oxidation to iodine and chromate (VI). J Anal Toxicol 2005;29:658–63. https://doi.org/10.1093/jat/29.7.658.Search in Google Scholar PubMed
18. Singh, J, Elberling, JA, Hemphill, DG, Holmstrom, J. The measurement of nitrite in adulterated urine samples by high-performance ion chromatography. J Anal Toxicol 1999;23:137–40. https://doi.org/10.1093/jat/23.3.137.Search in Google Scholar PubMed
19. Kinkennon, AE, Black, DL, Robert, TA, Stout, PR. Analysis of nitrite in adulterated urine samples by capillary electrophoresis. J Forensic Sci 2004;49:1094–100. https://doi.org/10.1520/jfs2003361.Search in Google Scholar
20. Steuer, AE, Arnold, K, Schneider, TD, Poetzsch, M, Kraemer, T. A new metabolomics-based strategy for identification of endogenous markers of urine adulteration attempts exemplified for potassium nitrite. Anal Bioanal Chem 2017;409:6235–44. https://doi.org/10.1007/s00216-017-0567-4.Search in Google Scholar PubMed
21. Steuer, AE, Arnold, K, Kamber, D, Kraemer, T. Suitability evaluation of new endogenous biomarkers for the identification of nitrite-based urine adulteration in mass spectrometry methods. Drug Test Anal 2019;11:230–9. https://doi.org/10.1002/dta.2481.Search in Google Scholar PubMed
22. Steuer, AE, Kamber, D, Kraemer, T. Evaluation of endogenous urinary biomarkers for indirect detection of urine adulteration attempts by five different chemical adulterants in mass spectrometry methods. Drug Test Anal 2019;11:638–48. https://doi.org/10.1002/dta.2539.Search in Google Scholar PubMed
23. Emmert-Streib, F, Yang, Z, Feng, H, Tripathi, S, Dehmer, M. An introductory review of deep learning for prediction models with big data. Front Artif 2020;3. https://doi.org/10.3389/frai.2020.00004.Search in Google Scholar PubMed PubMed Central
24. Nazarenko, DV, Kharyuk, PV, Oseledets, IV, Rodin, IA, Shpigun, OA. Machine learning for LC–MS medicinal plants identification. Chemometr Intell Lab 2016;156:174–80. https://doi.org/10.1016/j.chemolab.2016.06.003.Search in Google Scholar
25. Wilkes, EH, Rumsby, G, Woodward, GM. Using machine learning to aid the interpretation of urine steroid profiles. Clin Chem 2018;64:1586–95. https://doi.org/10.1373/clinchem.2018.292201.Search in Google Scholar PubMed
26. Streun, GL, Elmiger, MP, Dobay, A, Ebert, L, Kraemer, T. A machine learning approach for handling big data produced by high resolution mass spectrometry after data independent acquisition of small molecules - proof of concept study using an artificial neural network for sample classification. Drug Test Anal 2020;12:836–45. https://doi.org/10.1002/dta.2775.Search in Google Scholar PubMed
27. Boxler, MI, Schneider, TD, Kraemer, T, Steuer, AE. Analytical considerations for (un)-targeted metabolomic studies with special focus on forensic applications. Drug Test Anal 2019;11:678–96. https://doi.org/10.1002/dta.2540.Search in Google Scholar PubMed
28. Chong, J, Soufan, O, Li, C, Caraus, I, Li, S, Bourque, G, et al.. MetaboAnalyst 4.0: towards more transparent and integrative metabolomics analysis. Nucleic Acids Res 2018;46:486–94. https://doi.org/10.1093/nar/gky310.Search in Google Scholar PubMed PubMed Central
29. Chollet, F, Keras. GitHub: GitHub repository; 2015.Search in Google Scholar
30. Dancho, M. Deep learning with keras to predict customer churn. RStudio AI Blog; 2018.Search in Google Scholar
31. Kuhn, M. Building predictive models in R using the caret package. J Stat Softw 2008;28:1–26. https://doi.org/10.18637/jss.v028.i05.Search in Google Scholar
32. Abadi, M, Barham, P, Chen, J, Chen, Z, Davis, A, Dean, J, et al.. TensorFlow: large-scale machine learning on heterogeneous systems; 2015.Search in Google Scholar
33. Ribeiro, MT, Singh, S, Guestrin, C. “Why should I trust you?”: explaining the predictions of any classifier. In: Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining. San Francisco, CA, USA: Association for Computing Machinery; 2016:1135–44 pp.10.1145/2939672.2939778Search in Google Scholar
34. Smith, CA, Want, EJ, O’Maille, G, Abagyan, R, Siuzdak, G. XCMS: processing mass spectrometry data for metabolite profiling using nonlinear peak alignment, matching, and identification. Anal Chem 2006;78:779–87. https://doi.org/10.1021/ac051437y.Search in Google Scholar PubMed
35. Tsugawa, H, Cajka, T, Kind, T, Ma, Y, Higgins, B, Ikeda, K, et al.. MS-DIAL: data-independent MS/MS deconvolution for comprehensive metabolome analysis. Nat Methods 2015;12:523–6. https://doi.org/10.1038/nmeth.3393.Search in Google Scholar PubMed PubMed Central
Supplementary Material
The online version of this article offers supplementary material (https://doi.org/10.1515/cclm-2021-0010).
© 2021 Walter de Gruyter GmbH, Berlin/Boston
