MSACL 2018 US Abstract

Topic: Small Molecules / Tox

Confirming 23-Benzodiazepines in Urine for Pain Management Using LC-MS/MS – a Challenging Assay Development

Richard Lahr (Presenter)
Development Technologist, Mayo Clinic

Bio: 2008 - Bachelor's of Arts in Chemistry, with a minor in Mathematics from St. John's University, Collegeville, MN 2008-present: Toxicology and Drug Monitoring Lab, Mayo Clinic, Rochester, MN Current Title: Development Technologist

Authorship: Richard G. Lahr, Loralie J. Langman, Paul J. Jannetto
Mayo Clinic, Rochester, MN

Short Abstract

During the routine validation of a benzodiazepine LC-MS/MS method, it was noted that lorazepam, triazolam, and alpha-hydroxytriazolam showed a quadratic shift/bias in the calibration curve particularly at the high end. The ultimate cause of this bias was determined to be due to the natural presence of chlorine (Cl) in these benzodiazepines where the mass of the heavy isoforms of Cl was the same as the respective deuterium labeled internal standards for these compounds. One solution to this bias was to identify and use the Cl heavy isoforms of the respective labeled internal standards.

Long Abstract


Benzodiazepines represent a large family of medications used to treat a wide range of disorders from anxiety to seizures and they are also used in pain management. With a high risk for abuse/diversion, professional practice guidelines recommend compliance monitoring for these medications using urine drug tests[1,2]. Utilizing a definitive assay can give clinicians a better indication of compliance to controlled substances. This new quantitative confirmation assay provides a more comprehensive and sensitive alternative to traditional GC/MS assays for twenty-three benzodiazepines and metabolites (aprazolam, alpha-hydroxyalprazolam, chlordiazepoxide, clonazepam, 7-amino clonazepam, diazepam, nordiazepam, oxazepam, temazepam, flunitrazepam , 7-amino flunitrazepam, flurazepam, hydroxyethylflurazepam, lorazepam, midazolam, alpha-hydroxymidazolam, triazolam, alpha-hydroxytriazolam, zolpidem, zolpidem phenyl-4-carboxylic acid, clobazam, norclobazam, and prazepam).


A protein crash/dilute-and-shoot method was developed in which urine samples (100 µL) were first hydrolyzed (enzymatically) for 10 minutes. The subsequent hydrolyzed sample was crashed with acetonitrile and a portion of the supernatant was diluted with 20 mM ammonium formate/0.2% formic acid in clinical laboratory reagent water (CLRW). The resulting diluted sample was analyzed by LC-MS/MS using an ABSciex™ 6500 QQQ with heated electrospray ionization in positive mode. A Restek™ Ultra Biphenyl (5 µm 50 x 2.1 mm) column achieved complete chromatographic separation of isobaric compounds using a linear gradient with a total run time of 7.5 minutes, a flow rate of 0.5 mL/min, and the ability to multiplex samples. Mobile phase A contained 20mM ammonium formate with 0.2% formic acid in CLRW, while mobile phase B contained 0.1% formic acid in acetonitrile. Compounds were identified by retention time, relative retention time to a deuterated internal standard, and Q1/Q3 ion pair ratios.


Initially, an unusual calibration bias was observed for three of the analytes (lorazepam, triazolam and alpha-hydroxytriazolam), most notably at the upper limit of quantitation. On the surface, the error appeared to associate with the saturation of the detector or, possibly, incorrect settings within the quantitation software. When this data was plotted using the analyte-internal standard area ratio against the targeted concentrations the curve was quadratic. However, when the analyte peak area was plotted against the targeted concentrations, the plot was linear and had an r2 of greater than 0.990. In contrast, a plot of the IS showed that an increase over the analytical range. As part of the investigation samples were spiked and measured with only the non-labeled analytes (without internal standard), and a separate set of samples were spiked and measured with only the deuterated labeled internal standards (did not contain any non-labeled analytes). As expected, the samples spiked with only the internal standards showed consistent response across the analytical measuring range while the analyte spiked samples (without internal standard) showed a linear increase in the internal standard response.

The common thread between all three of the compounds showing the calibration bias was that they contained two chlorines – in fact 19 of the 23 benzodiazepines compounds contain at least one chlorine atom (flunitrazepam, 7-amino flunitrazepam, zolpidem, and zolpidem-phenyl-4-carboxylic acid being the exceptions). Chlorine (Cl) has two stable, commonly occurring masses – Cl25 and Cl27 – with the heavy form being ~25% naturally present[3]. The final piece of evidence is that all three compounds utilize a D4 internal standard. To academically confirm the presence of naturally abundant heavy compounds, the instrument was re-optimized to look for one- and two- heavy mass Cl on lorazepam, triazolam, and alpha-hydroxytriazolam. The experiment yielded a linear increase in the abundance of heavy isoforms of the compounds as the concentration increased and matched the predicted polynomial distribution.

While the results further proves the natural abundance of heavier forms of Cl, it didn’t directly solve the overall problem of how to monitor the compounds of interest that have an isotopic form at the same mass as its respective internal standard. Changing the isotopic labels on lorazepam, for example, to a [13C]2-D4, was one solution, but no alternatively labeled internal standard was available for triazolam and alpha-hydroxytriazolam. Therefore, in the absence of another isotopically labeled internal standard, we chose to monitor the heavy chlorine isoforms of the deuterated internal standards thus eliminating the interference. Following the re-optimization of the internal standards, all of the compounds were linear over the desired analytic measurement range.

Conclusions & Discussion

The presence of a stable, heavy isotope in a compound that matched the mass of the deuterated-labeled internal standard presented a unique challenge to the development of this assay. However, the laboratory was able to successfully validate a highly specific and sensitive LC-MS/MS confirmation assay using LC-MS/MS system for benzodiazepines in urine.

References & Acknowledgements:

1.Charney DS, Mihic SJ, Harris RA: Hypnotics and Sedatives. In The Pharmacological Basis of Therapeutics. Chapter 16. Volume 11. Eduted by LL Brunton, JS Lazo, KL Parker. Goodman & Gilman's, McGraw-Hill Companies, Inc, 2006. Available from URL:

2.Levine B: Central Nervous System Depressants. In Principles of Forensic Toxicology. Volume 2. Edited by B Levine B. Washington DC: AACC Press, 2003:157-72.

3.De Laeter, et al., ATOMIC WEIGHTS OF THE ELEMENTS: REVIEW 2000 (IUPAC Technical Report), Pure Appl. Chem., Vol. 75, No. 6, pp. 745–746, 2003.

Financial Disclosure

SalaryyesMayo Clinic, Rochester, MN
Board Memberno

IP Royalty: no

Planning to mention or discuss specific products or technology of the company(ies) listed above: