ANNALS OF CLINICAL AND LABORATORY SCIENCE, Vol. 25, No. 4 Copyright © 1995, Institute for Clinical Science, Inc. Mechanism of Interferences for Gas Chromatography/Mass Spectrometry Analysis of Urine for Drugs of Abuse*f ALAN H. B. WU, Ph.D. Toxicology Laboratory, Hartford Hospital, Hartford, CT 06102 ABSTRACT Although gas chromatography/mass spectrometry (GS/MS) is recognized as the definitive procedure for confirming positive immunoassay screening results of urine for drugs of abuse, targeted GC/MS analysis does have limitations. False negative results can occur when interfering drugs are present at high relative concentrations. If an interfering drug competes with the targeted drug for the derivatization reagent, low results are pro­ duced. If the interfering drug chromatographically co-elutes with the target drug, the ionization efficiency of the target compound by electron impact (El) ionization may be affected. False positive results can also occur through a number of mechanisms. Two substances with the same mass spectrum require gas chromatographic conditions that enable adequate separation of the compounds prior to MS analysis. In the case of optical isomers, special columns or derivatives must be used for identification and quantification. The widespread use of selected ion monitoring may further limit GC/MS assays. Drugs that produce similar high molecular-weight mass fragment ions could potentially interfere if they have similar GC retention times and if inappropriate ions are selected for monitoring. The conversion of one drug to another by the GC/MS instrument itself is a particularly insidious problem. False positive and negative results have serious forensic consequences and must be recognized and avoided. In contrast, the consumption of poppy seeds or meats from livestock given drugs such as methenolone can produce unexpected true positive results for opiates and anabolic steroids, respectively. Introduction for drugs of abuse, most laboratories use a combination of screening using a non­ Although there are many methods isotopic immunoassay and confirmation available today for the analysis of urine by gas chromatography/mass spectrome­ try. Many consider GC/MS as the “gold * Presented in part at the 1994 Spring Meeting of standard” for confirmation analysis. The the Association of Clinical Scientists, Charleston, SC. Substance Abuse and Mental Health Ser­ t Send reprint requests to: Alan H. B. Wu, Ph.D., Toxicology Laboratory, 80 Seymour Street, P.O. vices Administration (SAMHSA) man­ Box 5037, Hartford, CT 06102-5037. dates the use of targeted GC/MS analysis 319 0091-7370/95/0700-0319 $01.80 © Institute for Clinical Science, Inc. 320 wu for confirming positive results of urine Methodology collected for workplace testing of federal em ployees . 1,2 A “ targeted analysis” is F a l s e N e g a t i v e R e s u l t s directed towards determining the absence or quantitative presence of only Competition for derivatization rea­ a few drugs. This is in contrast to an gents. An important step for most GC/MS “unknown analysis” where the objective assays is derivatization of the target drug. is to identify the presence of as many Derivatization functions to increase the drugs or substances as possible. Most volatility of the substance, improve the laboratories use selected ion monitoring efficiency of chromatographic elution, (SIM) for targeted analysis because this and increase the molecular weight of improves the analytical sensitivity of the mass fragments so that they are more dis­ assay.3 Analysis, by SIM usually involves tinct from low molecular weight “back­ monitoring the intensity of three princi­ ground” mass fragments. Derivatization pal ions of a mass spectrum and comput­ reagents react with specific functional ing ion ratios of responses between ions. groups on the parent molecule including Full-scan mass spectral analysis coupled amides, amines, acids, and hydroxides.4 with a computerized library and search Accurate GC/MS analysis requires routines are necessary for conducting derivatization reactions that are stoichio- unknown analyses. metrically complete, as underivatized Data from GC/MS analysis are consid­ drugs elute at different times and would ered the most defensible in legal pro­ not be detected by analyses targeted ceedings. Some attorneys and lay indi­ towards the derivative. viduals might consider such analyses to As an example of drug interference be beyond challenge. However, GC/MS caused by competition, Brunk described is like any other analytical technique, the GC/MS analysis of the marijuana and there are important limitations. In metabolite 11-nor-A9 -THC-9-carboxylic this paper, mechanisms for false negative acid (9-carboxy THC [tetrahydrocan- and positive results of GC/MS analysis nabinols] ) . 5 As shown in figure 1, this for drugs of abuse are reviewed from assay involves the formation of the published reports. dimethyl derivative with dimethylsulfoxide c h 3 c « 3 (CH 3 )2 CHCH 2 ^ > -C H C O O H (CH 3 )2 CHCH 2 CHCOOCH 3 COOH delta-9-THC F ig u r e 1. The presence of ibuprofen can compete with delta-9-THC (tetrahydrocannabinols) for the formation of methyl derivatives. MECHANISM OF INTERFERENCES FOR GC/MS OF URINE FOR DRUGS OF ABUSE 321 (DMSO) and tétraméthylammonium excess to the other, saturation of the ion­ hydroxide (TMAH). The presence of ization source can occur, resulting in a high concentration of ibuprofen inter­ decrease in the ionization efficiency of fered with the derivatization of 9-carboxy both compounds (figure 2B). Under these THC by competing with the derivatiza­ conditions, the concentration of the tar­ tion reagent. This problem can be elimi­ get drug may be dramatically decreased nated by using enough derivatizating by the presence of the interferant. The reagent so that it does not get depleted gas chromatograph normally separates when other potentially interferring drugs drugs from potential interferants so that are present. both are not simultaneously present Interference of the ionization source in the ionization chamber of the mass by coelution. A critical tenant for the spectrometer. However, if pre-analytical success of GC/MS analysis is that the separation does not occur, a substance interaction of a substance with the instru­ that coelutes with the target compound mentation is independent to the pres­ may produce an interference. ence of another substance.6 In figure 2A Specific examples of these interactions is depicted the production of positive have been reported .6 ,7 In one report, the ions and mass fragments from the bom­ detection of cocaine metabolite ben- bardment of electrons for two different zoylecgonine (BE, target compound) was drugs present in equal concentrations. interfered by the coelution of the antifun- The presence of drug A does not interfere gal drug fluconazole (interferant), pres­ with the ability of the electron stream to ent at much higher concentrations rela­ ionize drug B. However, if the concentra­ tive to BE .6 This situation was detected tion of one of the drugs is present in great because a positive immunoassay screen A. B. to MS F igu re 2. A. Two substances (A and B) present in equal and nonsaturating concentrations within the electron impact chamber of a mass spectrometer independently produce parent ions (Al, Bl) and fragments (A2, B2). B. Substance B is present at saturating concentrations relative to the target substance A. Under these conditions, some of the target drug and interferant are unable to be ionized by the electron beam or are unable to reach the filtering segment of the mass spectrometer. 322 wu for BE was unexpectedly negative by Once the presence of a coeluting inter­ GC/MS. In the second example (figure 3), fering compound has been detected, cor­ high concentrations of a methadone rection of false negative results can be metabolite interfered with detection and made by removal of the interférant itself quantitation of BE .7 As shown in figure 3, prior to GC/MS analysis. In the example 2-ethylidene-l,5-dimethyl-3,3-diphenyl- presented in figure 3, EDDP was pyrrolidine (EDDP) coelutes with the removed by thin-layer chromatography cocaine metabolite, resulting in a very resulting in full recovery of BE from the small GC peak at m/z 300, corresponding sam ple.7 Another mechanisms to remove to the major fragment of the pentafluoro- interférants is to alter the gas chromato­ propionic (PFP) derivative of BE. graphic analysis conditions (e.g., chang­ The presence of a coeluting band is ing columns or temperature program) so very obvious when full scan MS analysis that the interférant no longer coelutes is used. In figure 3 is clearly illustrated with the target compound. This may not the presence of a large potentially inter­ always be possible, as interférant peaks fering chromatographic peak. When can be very broad. Moving the interférant selected ion monitoring mass spectro- may affect detection of other important metric techniques are used, however, substances such as drug metabolites. detection of an interférant is not apparent unless one (or more) of the ions selected for the target drug is coincidentally one F a l s e P o s i t i v e R e s u l t s of the major fragments of the interférant. This highlights the importance of using Substances with identical mass spec­ coeluting deuterated internal standards tra. Qualitative identification by mass (IS) . 8 The recovery of a coeluting internal spectrometry is predicated on the standard can be used to evaluate the assumption that no two compounds will quality of the GC/MS analysis, as the produce the identical mass spectrum. presence of a coeluting interférant will Although this may be theoretically true, affect the ionization efficiency of both the in actual practice, many pairs of com­ IS and target drug. Since a constant pounds have mass spectra that are essen­ amount of the IS is added to all samples, tially identical (i.e., within the limits of a low IS recovery might be caused by current instrumentation).