Theron J. Hamilton 1 *, Harry Z. Qui 1, Katherine V. R. Dozier 1 and Zachary J. Fuller 2. Testing Laboratory, Fort Meade, MD, USA

Transcription

1 Journal of Analytical Toxicology 2014;38: doi: /jat/bku069 Advance Access publication June 20, 2014 Technical Note Phentermine Interference and High L-Methamphetamine Concentration Problems in GC EI-MS SIM Analyses of R-(2)-a-Methoxy-a-(Trifluoromethyl)Phenylacetyl Chloride-Derivatized Amphetamines and Methamphetamines Theron J. Hamilton 1 *, Harry Z. Qui 1, Katherine V. R. Dozier 1 and Zachary J. Fuller 2 1 Department of Chemistry, United States Naval Academy, Annapolis, MD, USA, and 2 United States Army Forensic Toxicology Drug Testing Laboratory, Fort Meade, MD, USA *Author to whom correspondence should be addressed. theron@usna.edu In order to achieve chromatographic separation, urine samples shown to be initially positive for amphetamines and methamphetamines in US Department of Defense immunoassays are derivatized with R-(2)-a-methoxy-a-(trifluoromethyl)phenylacetyl chloride (R-(2)-MTPA) prior to gas chromatography electron impact-mass spectrometry (GC EI-MS) analysis. Phentermine, a member of the phenethylamine class of drugs and a common appetite suppressant, interferes with GC EI-MS assays of R-(2)-MTPA-derivatized D-amphetamine, degrading the chromatography of the internal standard and analyte ions and skewing concentration calculations. Additionally, when specimens with high concentrations of L-methamphetamine are derivatized with R-(2)-MTPA, signal peaks have the potential to be misidentified by integration software as D-methamphetamine. We have found that replacing R-(2) MTPA with (S)- (1)-a-methoxy-a-(trifluoromethyl)phenylacetyl chloride reduces phentermine interference problems related to internal standard chromatography, reduces the possibility of concentrated L-methamphetamine peaks being misidentified by integration software, improves resolution of D-methamphetamine in the presence of high L-methamphetamine concentrations, and is a cost-neutral change that can be applied to current amphetamines GC EI-MS methods without the need for method modification. Introduction Amphetamine and methamphetamine are routinely included in Department of Defense (DoD) drug screening panels. In the DoD regimen, urine specimens are screened with three amphetamine-sensitive commercial immunological assays before confirmation of drug presence is ultimately made through gas chromatography electron impact-mass spectrometry (GC EI-MS) analysis using selected ion monitoring (SIM). Since 2004, the standard confirmational GC EI-MS SIM and sample preparation methods for amphetamine and methamphetamine used by the DoD have closely followed the recommendations published by Paul et al. (1), which suggest the use of R-(2)- a-methoxy-a-(trifluoromethyl)phenylacetyl chloride (R-(2)- MTPA) as a chiral derivatizing agent for distinguishing enantiomeric pairs of amphetamines and methamphetamines. The use of MTPA allows for the complete chromatographic separation of the S-(þ) or D andther-(2) or L molecular stereoisomers The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of the Department of the Navy, Department of the Army and Department of Defense nor the US Government. of amphetamine and methamphetamine. However, the US Army Forensic Toxicology Drug Testing Laboratory (FTDTL), at Fort Meade, MD, USA, has recently noticed several issues related to the interference of GC EI-MS assays by phentermine (C 10 H 15 N) an FDA approved appetite suppressant for the treatment of obesity and a member of the phenethylamine class of drugs. Phentermine is not eliminated by specimen pretreatment with iodate in a way similar to phenylpropanolamine and ephedrine. Attempts to modify the current DoD GC EI-MS method in order to achieve chromatographic separation between the interferent and analyte ions were unproductive; moreover, a prevalence study showed that thousands of tested military specimens per year could be contaminated with phentermine at the Ft. Meade FTDTL alone. This interference issue translates into a relatively significant operational and economic burden that is likely to also affect other drug testing laboratories. In particular, we have observed that the effect of phentermine interference on the established DoD GC EI-MS amphetamine method (no phentermine interference was observed with methamphetamines) can be divided into two cases. Case 1: GC EI-MS analyses showing the complete absence of amphetamines in samples that initially screened positive in three separate amphetamines-sensitive immunological assays. The problem in this case is troublesome for two reasons. First, it highlights a limitation of this application of the immunoassay screening reactions which cross-react with phentermine a well-known problem that has proven difficult to circumvent (2, 3). Secondly, although there is an absence of target analyte, phentermine complicates GC EI-MS analysis by interfering with the internal standard quantification and qualifier ions. The effect is a shift of the internal standard ion peaks to retention times outside of the defined retention time integration windows and possibly unacceptable chromatography. In response, standard operating procedure calls for manual integration of the internal standard ion peaks and subsequent re-extractions of the original specimen. According to the US Army FTDTL drug testing protocol, even negative results with internal standard failures will not be released without the lab first attempting to obtain acceptable internal standard results through follow-up assays. In other DoD drug testing laboratories, acceptable internal standard results with analyte-negative samples are defined as those exhibiting retention times and mass ion ratios that, minimally, allow for the identification of the relevant internal standard ions. However, even with this less strict criterion, rework is still likely since most phentermine-contaminated samples will have skewed retention times. For drug testing laboratories experiencing a high prevalence of phentermine-contaminated specimens, the Published by Oxford University Press This work is written by (a) US Government employee(s) and is in the public domain in the US.

2 economic and man-hour costs of additional extractions, new GC EI-MS confirmation assays and multiple data reviews as a result of internal standard failures can become significant. Case 2: GC EI-MS analyses showing the presence of D- and L-amphetamine in the sample, but failing acceptability criteria because of poor chromatographic resolution of the 162 m/z and 118 m/z analyte qualifier ions due to interference from co-eluting phentermine fragments. Instances of this problem in practice are less impactful than those in Case 1 since the occurrence frequency is low, but nonetheless, we noticed some interesting consequences. For example, the effect of phentermine fragments co-eluting with the internal standard target and qualifier ions is quantifiable. Consider the following experiment with control samples. In general, final analyte concentrations are given by C ¼ C 0 ða dt=a dis Þ ða ct =A cis Þ where C 0 is the concentration cutoff, A dt is the drug target ion area, A dis is the drug internal standard area, A ct is the calibrator target ion area and A cis is the calibrator internal standard ion area. Regardless of whether or not the qualifier ions resolve, in these control experiments, the internal standard quantification (264 m/z) and the drug quantification (262 m/z) ion peaks are consistently within defined retention time integration windows with acceptable chromatographic resolution. Our data show that, when compared with the known control concentration, experimental samples containing phentermine will have calculated analyte concentrations up to 15% below the known value, implying that latent co-eluting phentermine fragments are artificially increasing internal standard peak area values (data not shown). This result is consistent with the equation above, which predicts that larger internal standard areas will concurrently reduce final analyte concentrations, C. Considering this fact, we can envision at least some scenarios where the potential to release artificially low quantification results exists when phentermine concentrations are not high enough to degrade chromatography enough to fail acceptability standards. A more serious problem unrelated to phentermine interference is revealed when analyzing urine with high L- methamphetamine concentrations. It is possible that GC EI-MS assays of specimens overloaded with L-methamphetamine isomers will show L-methamphetamine peak migration to later retention times. Achieving urinary L-methamphetamine concentrations, high enough to induce this effect is possible through the use of over-the-counter products. For example, published studies have reported urinary concentrations of L-methamphetamine as high as 6,000 ng/ml following the use of a Vicks w inhaler (4). A chromatographically well resolved, but retention time-shifted L-methamphetamine peak could cause integration software to misidentify it as D-methamphetamine when derivatized with R-(2)-MTPA. If also coupled with proper mass ion ratios and chromatographically acceptable internal standard ion peaks, recognizing the error could be particularly difficult for data review officials. This fact and the potential for the artificially lowconcentration values discussed above are particularly disconcerting in an environment where accurate quantitation is critical to both the success of the DoD s drug-use deterrence efforts and to those individuals whose careers hinge on the fairness and accuracy of the results. Here, we report a simple and cost-neutral improvement to reduce the problems discussed above based on the use of a derivatizing agent with chirality opposite to the one currently used in DoD amphetamine protocols, S-(þ)-a-methoxy-a-trifluoromethylphenylacetyl (S-(þ)-MTPA). Substitution of R-(2)-MTPA with S-(þ)-MTPA reduces phentermine interference by changing the elution order of the D- andl-enantiomers. This change has the effect of shifting target D-amphetamine ion away from the retention times of the phentermine fragments, removing the internal standard interference issues and the potential for calculating artificially low-concentration values in the process. In this application, the use of S-(þ)-MTPA could significantly reduce the amount of rework triggered by phentermine-related GC EI-MS amphetamine confirmation assay failures in DoD and other drug testing laboratories. Most importantly, the possibility of erroneous peak integration with R-(2)-MTPAderivatized specimens containing high concentrations of L-methamphetamine isomers is avoided as a result of shifted retention times. Finally, it should be noted that the substitution of R-(2)-MTPA for S-(þ)-MTPA can be made without any modification to the current DoD amphetamine GC EI-MS analysis method. Experimental The extraction and derivatization procedures as well as the GC EI-MS SIM parameters used in all experiments were identical to those described in Paul et al., with the following exceptions: (i) the sample preparation procedures employed in the study of phentermine prevalence did not include a derivatization step and (ii) a calibrator cutoff concentration of 100 ng/ml for DoD amphetamines assays was used to quantitate isomer concentrations rather than a 500 ng/ml calibrator in all cases. S-(þ)-MTPA was purchased from Sigma Aldrich Corporation (St. Louis, MO, USA). Service member urine specimens analyzed in the effort to estimate the prevalence of phentermine had been held for further investigation after screening positive in amphetamine immunoassays, but confirming negative with GC EI-MS. In order to verify the presence of phentermine, aliquots of the original specimens were extracted, purified without derivatization and subjected to a total (raw signal) mass scan on an Agilent w 6890 GC EI-MS. Mass spectra were extracted from all recorded chromatographic peaks and compared against reference mass spectra from a commercially available National Institutes of Standards and Technology library. Validation of the use of S-(þ)-MTPA as a replacement derivatizer for amphetamine quantitation was conducted with a series of side-by-side control experiments. Each contained two groups of a racemic amphetamine mixture at concentrations of 20, 50, 75, 100 (reporting cutoff concentration), 150, 500 and 1,000 ng/ml in a certified drug-free urine matrix. Each amphetamine concentration contained 50,000, 25,000, 12,500 or 0ng/mL of phentermine. One group was derivatized with R-(2)-MTPA and the other with S-(þ)-MTPA. In order to demonstrate how the use of R-(2)-MTPA could increase the risk of peak misidentification in L-methamphetamine-overloaded specimens, a 98.7% chirally pure Cerilliant w L-amphetamine/methamphetamine isomer control set was made with analyte concentrations increasing in Interference and Concentration Problems in Amphetamines GC MS Assays 457

3 1,000 ng/ml increments from 1,000 to 10,000 ng/ml in certified urine. Following extraction, the samples were derivatized with R-(2)-MTPA and subjected to GC EI-MS analysis. In a separate experiment designed to demonstrate improved D-methamphetamine ion resolution in samples with high L-methamphetamine ion concentrations, duplicate 10-sample control sets were prepared. Each sample had a constant 100 ng/ml D-amphetamine concentration with L-methamphetamine concentrations increasing in 1,000 ng/ml increments from 1,000 to 10,000 ng/ml in certified urine. Following extraction, one set was derivatized with R-(2)-MTPA and the other with S-(þ)- MTPA followed by analysis with GC EI-MS. Results The number of phentermine-contaminated specimens triggering rework at FTDTL, Ft. Meade, was found to be significant. Based on the acceptability criterion that only spectra with a minimum match of 70% to the reference library spectrum were considered reliable, 68% of the 279 service member specimens that had screened initially positive for amphetamines, but confirmed amphetamine negative during a 2-week period in July of 2013 were shown to contain phentermine (data not shown). Assuming that the submission frequency of phentermine-contaminated specimens remains constant and that the number can be extrapolated to.4,000 specimens per year with the potential to interfere with the testing process. Substitution of S-(þ)-MTPA for R-(2)-MTPA can help reduce the need for confirmation rework in some of those cases and for routine amphetamines GC EI-MS assays the two derivatizers were shown to be equivalent. Concentration calculations of phentermine-free amphetamine controls used in the experiments showed no significant difference between those derivatized with R-(2)-MTPA and those derivatized with S-(þ)-MTPA. Additionally, with both derivatizers, chromatography results for target and qualifier analyte ions were similar with or without phentermine and at all phentermine to amphetamine concentration ratios. Unfortunately, these similarities also include phentermine-induced resolution problems with the m/z analyte qualifier ion that were present when using the R-(2)-MTPA derivatizer and were not improved with the use of S-(þ)-MTPA. Efforts to correct the issue by replacing the m/z with the m/z ion yielded similarly poor results even after various reductions in the He flow rate and a series of temperature ramp changes between 15 and 458C/s. However, derivatization with S-(þ)-MTPA eliminated all phentermine-induced problems affecting internal standard ions. Figure 1 illustrates this effect with the results of the control experiment using a 1:2,500 amphetamine to phentermine concentration (ng/ml) ratio the lowest studied. Figure 1a and b Figure 1. The results of a control experiment with a 20 50,000 amphetamines to phentermine concentration ratio (ng/ml) showing internal standard ion resolution when derivatized with either R-(2)-MTPA or S-(þ)-MTPA. (a) The D-methamphetamine/L-methamphetamine m/z internal standard quantification ions with R(2)-MTPA. The retention time of D-methamphetamine internal standard is 5.21 s. The peaks are unresolved due to phentermine interference. (b) The D-methamphetamine/L-methamphetamine m/z 98.0 internal standard qualification ions with R-(2)-MTPA. The peaks are unresolved due to phentermine interference. (c) and (d) The result of an identical control experiment with S-(þ)-MTPA used in place of R-(2)-MTPA. The retention times of the m/z and m/z 98.0 D-methamphetamine internal standard ions are now 4.99 and 4.98 s, respectively. The effect of phentermine interference on the D-methamphetamine internal standard ions has been removed. 458 Hamilton et al.

4 Figure 2. Data from a GC EI-MS assay of a control sample with 98.7% pure L-methamphetamine at 8,000 ng/ml. The data show how specimens with high concentrations of L-methamphetamine may be misidentified by integration software when L-methamphetamine peaks migrate into retention time windows defined for D-methamphetamine following derivatization with R-(2)-MTPA. All chromatographic and mass ion ratio criteria are met. (a) m/z target analyte ion. (b) m/z qualifier analyte ion. (c) m/z qualifier ion. (d) m/z D-methamphetamine internal standard quantification ion (the integrated peak). (e) m/z 98.0 D-methamphetamine internal standard qualifier ion (integrated peak). shows the chromatographic failure of the D-methamphetamine internal standard quantification and qualifier ions as result of phentermine interference. Figure 1C and D shows the result of the same experiment, but with S-(þ)-MTPA used as the derivatizing agent in place of R-(2)-MTPA. In this case, although chromatography for the analyte ions failed acceptability criteria, the effect of phentermine interference on the D-methamphetamine internal standard ions has been completely removed and the results of such an assay in Army labs could be released as negative without the need for rework. As Figure 2 shows, specimens with high concentrations of L-methamphetamine could lead to peak misidentification by integration software when L-methamphetamine peaks migrate into retention time windows defined for D-methamphetamine when derivatized with R-(2)-MTPA. Figure 2d and e shows the D-methamphetamine/L-methamphetamine internal standard quantification and qualifier ion windows the D- methamphetamine peak is properly integrated and meets all US Army FTDTL chromatographic and mass ion ratio acceptability criteria. Figure 2a c shows the drug target and two qualifier ion windows, respectively, with properly integrated and chromatographically acceptable peaks. However, as described above, the only analyte present in this control sample is 98.7% pure L-methamphetamine at 8,000 ng/ml. In fact, in this experiment with R-(2)-MTPA-derivatized analyte, all L-methamphetamine concentrations between 8,000 and 10,000 ng/ml produced results that were misidentified as D-methamphetamine by the integration software due to retention time shifts concentrations achievable with the use of a typical Vicks w inhaler (4). The data plotted in Figure 3 tracks the shift in retention times of the L-methamphetamine (O) signal peak as a function of L-methamphetamine concentrations. Clearly, when such samples are instead derivatized with S-(þ)-MTPA, L-methamphetamine peaks will elute at later retention times, sufficiently far from the defined D-methamphetamine integration window that any potential for peak misidentification will be effectively avoided. The second dataset ( ) in Figure 3 shows that there is also a concurrent increase in the retention time of the D-methamphetamine internal standard ion peak; however, in these experiments the change was not enough for the peak to fall outside of the defined integration window (see Figure 2). Furthermore, the ratio of the rates of retention time shift for L-methamphetamine and the D-methamphetamine internal standard with respect to L-methamphetamine concentration, dt L drug =dt D IS ; shows that the L-methamphetamine peak shifts 1.33 times faster than the D-methamphetamine internal standard peak. Finally, Figure 4 shows the advantage S-(þ)-MTPA has over R-(2)-MTPA as a derivatizing agent in terms of D-methamphetamine resolution when concentrations of L-methamphetamine Interference and Concentration Problems in Amphetamines GC MS Assays 459

5 Figure 3. L-Methamphetamine (O) and D-methamphetamine ( ) internal standard retention times as a function of L-methamphetamine drug concentration from a control experiment using 98.7% pure L-methamphetamine. Concentrations at and.8,000 ng/ml caused L-methamphetamine (O) signal peaks to migrate into the retention time integration windows of D-methamphetamine analyte and qualifier ions. D-Methamphetamine ( ) internal standard peaks also migrate as a function of L-methamphetamine drug concentration, but at rate that is 33% slower. Figure 4. Choosing S-(þ)-MTPA over R-(2)-MTPA as a derivatizing agent for methamphetamines improves D-methamphetamine peak resolution when concentrations of L-methamphetamine are high relative to the D isomer. (a) Target and qualifier ions of R-(2)-MTPA-derivatized D-methamphetamine at a concentration of 100 ng/ml and L-methamphetamine at a concentration 10,000 ng/ml. None of the three ions meet DoD resolution acceptability criteria. (b) The same experiment, but with S-(þ)-MTPA used as the derivatizing agent. In this case, all three ions are now resolved and all mass ion ratios are in acceptable ranges. 460 Hamilton et al.

6 are high. In this experiment, D-amphetamine concentrations were held constant at 100 ng/ml while L-amphetamine concentrations were increased in increments of 1,000 ng/ml. Analysis showed samples derivatized with R-(2)-MTPA failed resolution criteria with L-methamphetamine concentrations at and.2,000 ng/ml, but analysis of all samples derivatized with S-(þ)-MTPA displayed well-resolved D-methamphetamine ion peaks even at the highest L-methamphetamine concentrations. Figure 4a shows the R-(2)-MTPA-derivatized target and qualifier ions with 100 ng/ml D-methamphetamine and 10,000 ng/ml L-methamphetamine. Figure 4b shows the same experiment, but with S-(þ)-MTPA used as the derivatizing agent all DoD-defined chromatographic and mass ion ratio criteria are met for analyte and internal standard ions. resolution of D-methamphetamine peaks in the presence of high L-methamphetamine concentrations. Acknowledgments The authors thank William Bronner, PhD for suggesting the use of (S)-(þ)-MTPA in place of (R)-(2)-MTPA as a means of improving this method; Ms Paula Underwood for preparing all experimental samples and controls; and Ms. Michelle Serafin and Mr James Callies, FTCB for their thoughtful critiques of this report. The authors are especially grateful to Major Matthew Moser, PhD, and Captain Robert Nadeau, PhD of the United States Army Forensic Toxicology Drug Testing Laboratory, Fort Meade, for their generous support. Conclusion Substitution of R-(2)-MTPA with S-(þ)-MTPA as an amphetamines derivatizing agent should aid drug testing laboratories which are experiencing GC EI-MS assay failures as a result of phentermine interference with internal standards. In particular, the use of S-(þ)-MTPA should allow data review officials to release negative D-amphetamine results without subjecting the original specimen to re-extractions and GC EI-MS analyses. Most importantly, S-(þ)-MTPA would ensure the avoidance of peak misidentification in cases of high L-methamphetamine concentrations. Both problems have the potential to undermine the accuracy of reported results. Reducing these uncertainties with a simple, cost-neutral change in derivatizing agent is easy, requiring no modification to existing amphetamines GC EI-MS methods and has the added benefit of enhancing the References 1. Paul, B.D., Jemionek, J., Lesser, D., Jacobs, A., Searles, D.A. (2004) Enantiomeric separation and quantitation of (þ/-)-amphetamine, (þ/-)-methamphetamine, (þ/-)-mda, (þ/-)-mdma, and (þ/-)-mdea in urine specimens by GC-EI-MS after derivatization with (R)-(-)- or (S)-(þ)-alpha-methoxy-alpha-(trifluoromethy)phenylacetyl chloride (MTPA). Journal of Analytical Toxicology, 28, Stout, P.R., Klette, K.L., Horn, C.K. (2004) Evaluation of ephedrine, pseudoephedrine and phenylpropanolamine concentrations in human urine samples and a comparison of the specificity of DRI amphetamines and Abuscreen online (KIMS) amphetamines screening immunoassays. Journal of Forensic Sciences, 49, Herring, C., Muzyk, A.J., Johnston, C. (2011) Interferences with urine drug screens. Journal of Pharmacy Practice, 24, Fitzgerald, R.L., Ramos, J.M., Jr, Bogema, S.C., Poklis, A. (1988) Resolution of methamphetamine stereoisomers in urine drug testing: urinary excretion of R(2)-methamphetamine following use of nasal inhalers. Journal of Analytical Toxicology, 12, Interference and Concentration Problems in Amphetamines GC MS Assays 461