MSACL 2023 Abstract

Using LC-MS/MS LDT to Determine Fentanyl Prevalence and Evaluate the FEN2 Immunoassay’s Real-World Clinical Performance in Tertiary Care Settings Menlyadiev M.R., Suhandynata R.T., Lund K., Kelner M.J., Alabi A., Fitzgerald R.L. University of California, San Diego, CA Marlen Menlyadiev, Ph.D. (Presenter) UCSD Health >> POSTER (PDF)

Presenter Bio: Marlen Menlyadiev was trained as PhD analytical chemist specializing in the areas of differential mobility spectrometry and mass spectrometry. He then continued his training as Caltech postdoctoral scholar at JPL working on analysis of trace organics for space applications. After several years in industry where he developed methods and instrumentation for narcotics detection, Dr. Menlyadiev shifted his focus to something he has always been passionate about: applying mass spectrometry to the field of medicine. This led him to the Brigham and Women's Hospital where he worked on the metabolomic and lipidomic biomarker discovery. Currently, Marlen is a second year clinical chemistry fellow at UCSD working with Dr. Robert Fitzgerald. His projects include clinical implementation of fentanyl immunoassays, development of LC-MS/MS methods for measurement of cannabinoids in support of clinical research projects (marijuana and driving, CBD in treatment of autism) and use of UPLC-QTOF for identification of unknowns in clinical toxicology cases. He feels fortunate to have joined the UCSD program which is particularly strong in clinical mass spectrometry. In his free time, Marlen likes reading, hiking, and spending time with his family and friends.

Introduction
LC-MS/MS laboratory-developed tests (LDTs) are the mainstay of modern clinical toxicology testing. While widely used to support the development of FDA-cleared drug immunoassays, their significance in the clinical implementation and evaluation of such assays is less recognized. In this work we report on the use of our LC-MS/MS opiates method for determining the prevalence of fentanyl in urine drug screen (UDS) samples and assessing the real-world clinical performance of the Roche FEN2 fentanyl assay in routine clinical use. This work expands on our previous report of the FEN2 assay implementation.

Methods
Excess specimens from a total of 250 consecutive random UDS clinical samples were collected between 05/04/22 and 05/17/22 under UCSD IRB protocol 181656. These specimens were first screened using the DRI assay (Thermo Fisher Scientific) followed by the analysis by the FEN2 assay (Roche Diagnostics). Each specimen in the study was sent to the clinical toxicology laboratory for quantitative analysis by LC-MS/MS method for fentanyl and norfentanyl. Fentanyl prevalence in the tested population was calculated by examining the extracted ion chromatograms from analyzed samples for fentanyl and norfentanyl peaks (retention times, quantifier-to-qualifier ion, and signal-to-noise (S/N) ratios, etc.). For calculation of the clinical sensitivity and specificity of the FEN2 (and the DRI) immunoassay, sample were classified as true positive if they contained at least 2 ng/mL of fentanyl and/or norfentanyl. The same cutoffs were used when querying EHR to evaluate real-world clinical performance (screening and confirmation positivity rates, rates of false positives and negatives) of the FEN2 and the DRI immunoassays. False positive and negative rates in queried EHR cohorts were determined from samples that screened positive by immunoassay and did not confirm by LC/MS/MS or, conversely, from samples that screened negative, but had fentanyl and/or norfentanyl by LC-MS/MS.

Results
Thirty-eight of 250 study samples were found to contain fentanyl and 49 samples - norfentanyl at ≥ 2ng/mL concentration. Fifty-one samples contained fentanyl, norfentanyl or both analytes at ≥ 2ng/mL. The median fentanyl and norfentanyl concentrations in these 51 samples were 5 and 15.5 ng/mL, respectively, with corresponding inter-quartile ranges (IQRs) of 43 and 85 ng/mL. In addition, in 6 samples from the 250-sample study pool, fentanyl and/or norfentanyl were detected (signal-to-noise ratio >3, acceptable quantifier-to-qualify ion ratios, etc.) but not quantified. These findings corresponded to 22.8% prevalence of fentanyl in our study population.
Of the 51 LC-MS/MS true positive samples in the study, 31 and 50 were classified correctly by the DRI and the FEN2, respectively. Both assays classified 198 of the 199 LC-MS/MS-confirmed true negatives as negative. The clinical sensitivity and specificity calculated from these data were 61% and 99.5% for the DRI and 98% and 99.5% for the FEN2. The real-world screening positivity rate with the DRI and the FEN2 assays during 1 month testing period was, respectively, 13.3% and 17.3%, with the corresponding LC-MS/MS confirmation rate for immunoassay-positive samples of 88.8% and 96.8%. Higher immunoassay positivity rate for FEN2 was likely due to its ability to detect norfentanyl. The false positive rates for the DRI and the FEN2 in queried EHR cohorts (1000+ entries) were, 11.2% and 3.2%, respectively, while false-negativity rates (using smaller subset of total immunoassay screens) were 22% and 5.5% for the DRI and the FEN2 assays.

Conclusion
The use of LC-MS/MS LDT enabled estimation of the prevalence of the fentanyl in the study population (urban tertiary care hospital) and evaluation of the clinical performance of the FEN2 and DRI assay in both the study population and during 1 month of routine clinical use. Our results demonstrate the FEN2 assay has greater clinical sensitivity and is less prone to false positive results as compared with the DRI assay. These findings support the implementation of the FEN2 in a routine clinical practice and underline the broader role of mass spectrometry-based LDTs in clinical toxicology testing.