Urine methamphetamine-to-amphetamine ratio by LC-MS/MS to differentiate methamphetamine use from pharmaceutical impurity in patients prescribed amphetamine

Abstract

Introduction: Patients compliant with prescribed amphetamine (AMPH) should not have detectable methamphetamine (METH) in their urine; detectable METH typically indicates illicit use. However, we have identified patients with results suggestive of METH as an impurity in prescribed AMPH.

Objectives: Derive a METH:AMPH ratio cut-off from a training set of patients compliant with AMPH prescriptions to differentiate METH as an impurity from illicit use.

Methods: Retrospective review of AMPH and METH-positive cases by liquid chromatography-tandem mass spectrometry (LC-MS/MS) at Brigham and Women's Hospital (BWH) and Luxor Scientific. Correlated results with clinical and medication history and compliance with prescribed medications.

Results: The median ± interquartile range (IQR) METH:AMPH ratio for the Adderall training sets was 0.43 ± 0.31 % and 0.05 ± 0.040 %, with a maximum ratio of 1.125 % and 0.125 % at BWH and Luxor, respectively. The median ± IQR METH:AMPH ratio for the Luxor d-AMPH training set was 0.039 ± 0.028 %, with a maximum ratio of 0.09 %; not statistically different from the Adderall training set. Assessment of the BWH test set where METH < AMPH (n = 22) revealed that METH was likely due to an impurity (n = 10), distant METH mis/use (n = 11), or requiring further analysis (n = 1). METH was also detected by LC-MS/MS in a commercial AMPH calibrator and in Adderall XR.

Discussion: METH may represent an impurity in the AMPH formulation. Laboratories are encouraged to define a METH:AMPH ratio below which an impurity is the likely explanation for METH and/or to increase the METH positivity cut-off to 50 or 100 ng/mL to reduce potential false-accusations of illicit METH use.

Keywords: Amphetamine; Compliance monitoring; Liquid chromatography-tandem mass spectrometry; Methamphetamine; Pharmaceutical impurity; Toxicology.

Fig. 1

Flow diagram of the retrospective study performed at BWH and Luxor. LC-MS/MS toxicology results with detectable AMPH (≥5 ng/mL) were reviewed for inclusion in the Adderall training sets (BWH and Luxor), the d-AMPH training set (Luxor), or compiled into the test sets (BWH and Luxor), which were further classified into Groups A-C based on the METH:AMPH ratio. Group A had METH:AMPH ratio below the maximum ratio observed in the Adderall training set (excluding outliers). Group B included samples with METH:AMPH greater than that of the Adderall training set but where METH < AMPH. Group C included samples where METH:AMPH ratio ≥ 100 %.

Fig. 2

METH:AMPH for the Adderall training set and test set at BWH. The median, 25th percentile, 75th percentile, minimum, and maximum METH:AMPH ratios, shown as percentages, are presented for the Adderall training set (n = 16) and test set (n = 41). Red dashed lines indicate Group A (≤0.125, corresponding to the maximum cut-off of the training set) (n = 10, 24 %), test set Group B (>0.125; METH < AMPH) (n = 12, 30 %), and Group C (METH > AMPH) (n = 19, 46 %). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 3

METH:AMPH for the Adderall training set and test set at METH:AMPH Luxor. The median, 25th percentile, 75th percentile, minimum, and maximum METH:AMPH ratios shown as percentages, are presented for the Adderall training set (n = 38) and test set (n = 566). Red dashed lines indicate Group A (≤0.125, corresponding to the maximum cut-off of the training set) (n = 108, 19 %), test set Group B (>0.125; METH < AMPH) (n = 95, 17 %), and Group C (METH > AMPH) (n = 363, 64 %). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 4

Comparison of METH:AMPH in Adderall and d-AMPH (Vyvanse, Procentra) Formulations. The median, 25th percentile, 75th percentile, minimum, and maximum METH:AMPH ratios are presented for the Luxor Adderall training set (n = 38) and d-AMPH training set (n = 11). The p-value between groups was 0.8574, indicating no statistically significant difference.