Journal of Analytical Toxicology, Vol. 21, January/February 1997
Improved CEDIA| BenzodiazepineAssay Eliminates Sertraline Crossreactivity Robert t. Fitzgerald* and David A. Herold VeteransAffairs Medical Center, San Diego and Department of Pathology, University of California, 5an Diego, California
] Abstract ] Initial experiments demonstrated that the original CEDIA (cloned enzyme donor immunoassay)benzodiazepine assaycrossreacted with sertraline and sertraline metabolites. In responseto this phenomenon, Boehringer Mannheim Corporation developed an improved CEDIA benzodiazepine assay in order to eliminate sertraline crossreactivity.The improved CEDIA assaywas evaluated againstthe original CEDIA product, EMIT II (enzyme multiplied immunoassay technique) benzodiazepine assay,and electron capture negative chemical ionization (ECNCI) gas chromatography--massspectrometry (GC-MS). Five hundred and lhirty-one urine drug screenswere tested by the immunoassays. Sensitivity and specificity of these immunoassaysfor the 5-aryl-7chloro-l,4-benzodiazepine compoundswere 92 and 98%, respectively, for the improved CEDIA assay;92 and 93%, respectively, for the current CEDIA assay;and 87 and 98%, respectively, for EMIT II. The improved CEDIA assayperformed almost identically to the EMIT II assay,both of which had a significant advantage over the original CEDIA product, which was subject to crossreactivity because of sertraline metabolites. The cc-hydroxyketone metabolites of sertraline are identified in human urine specimensfor the first time using ECNCI GC-MS.
we were able to detect several characteristic peaks in chromatograms from many of the false-positive specimens. The mass spectra of these peaks contained base peaks consistent with a compound containing two chlorine atoms. This prompted us to review the patients' medical records to determine if the patients had prescriptions that could explain the laboratory findings. Of the 23 false-positivespecimens, 16 were obtained from patients with prescriptions for one of the newer antidepressants, Zoloft| (sertraline) (1). Based on these observations and other evaluations, the manufacturer modified the package insert to indicate that sertraline metabolites gave positive results on the CEDIAassay for benzodiazepines and developed an improved assay that would not crossreact with sertraline (3). In this report, we evaluate the improved CEDIAassay and compare it with the original CEDIA assay, EMIT II (enzyme multiplied immunoassay technique), and ECNCI GC-MS on 535 random drug screens submitted to this laboratory. In addition, we report the identification of a-hydroxy ketone metabolites of sertraline in human urine.
Experimental Introduction Boehringer Mannheim Corporation (IndianapoJis, IN) introduced CEDIA(cloned enzyme donor imrnunoassay) DAU, a line of homogeneous drugs-of-abuse testing reagents that are based on the use of recombinant enzyme fragments. When we evaluated the original CEDIA drugs-of-abuse screening assay for benzodiazepines, a substantial number of samples (23 of 503) that could not be confirmed to contain benzodiazepines using electron capture negative chemical ionization (ECNCI) gas chromatography-mass spectrometry (GC-MS) screened positive (1,2). Because the GC-MS procedure used in our laboratory uses full scan mode and not selected ion monitoring, * Addresscorrespondenceto RobertL. Fitzgerald,Ph.D., VA Medical Center-113,3350 La Jolla village Drive, San Diego, CA 92161, e-mail [email protected]
The CED1Aand EMITII assays were performed according to the package inserts supplied by the manufacturer. We obtained sertraline and N-desmethylsertraline from Pfizer (New York, NY). Any specimens that were positive by either EMITII or the improved CEDIA assay were analyzed by ECNCI GC-MS as described previously (2). The GC-MS assay quantitates nordiazepam, oxazepam, temazepam, lorazepam, N-l-hydroxyethylflurazepam, a-hydroxyalprazolam, and a-hydroxytriazolarn using full scan analysis. All samples were collected at the Veterans Affairs Medical Center in San Diego, California, with our Institutional Review Board's approval for the use of human subjects. Samples were kept frozen (-20~ from time of sample collection until time of analysis. For purposes of comparing the various immunoassays, each kit was analyzed according to the manufacturer's cutoff of
Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission.
Journal of Analytical Toxicology, Vol. 21, January/February 1997
300 ng/mL using oxazepam as the cutoff calibrator. A true positive was defined as any sample that gave an immunoassay positive result and contained benzodiazepines by ECNCIGC--MS. Afalse positivewas any sample giving an immunoassay positive result but not containing benzodiazepinesby ECNCIGC-MS.A true negative was any sample negative by immunoassay that contained total benzodiazepinesless than the stated cutoff of the immunoassay as quantitated by ECNCIGC-MS.A false negative was any immunoassay negative result for a sample that contained total benzodiazepines, as determined by ECNCIGC-MS, greater than the immunoassay cutoff.
tained only lorazepam tested negative by EMITII and the original CEDIAassay. The improved CEDIAassay was negative for all of the lorazepam-only specimens, with the exception of the sample that contained 619 ng/mL of lorazepam.
Discussion Sertraline is a serutonin reuptake blocker that was approved by the Food and Drug Administration in February 1992 for treatment of depression. Since sertraline's introduction, it has been well-receivedby the medical community as approximately 400,000 prescriptions were written in the first quarter of 1992. It is structurally unrelated to other antidepressants but does share limited similarity to the benzodiazepines (Figure 3). It is supplied as 50- and 100-rag tablets, with daily doses ranging from 50 to 200 rag. Peak plasma concentrations are reached
Table I shows the false positives, false negatives, true positives, true negatives, sensitivity, specificity, and predictive values of the immunoassays for the 5-aryl-7-chloro-l,4-benzodiazepine compounds using GC-MS as the reference Table I. Resultsof Immunoassayfor Detecting Presence method. The improved CEDIAassay performed almost identiof Benzodiazepinesusing GC-MS as the Reference rally to the EMITII assay, and both showed substantially fewer Method false positives than the original CEDIAassay. Improved Original To confirm that sertraline was the cause of most false posiCEDIA CEDIA EMIT II tives in the original CEDIAassay, urine specimens from ten patients with prescriptions for Zoloft were tested by all three False positives 9 38 6 assays. All ten specimens were positive by the original CEDIA False negatives 3 3 5 True positives 36 38 35 assay and negative by EMIT II and by the improved CEDIA True negatives 487 456 489 assay. Specimens from patients with sertraline prescriptions had Sensitivity (%) 92 93 88 two characteristic peaks in the total ion chrumatograms shown Specificity (%) 98 92 99 in Figure 1. Peaks I and 2 of this chromatogram had similar Pos Pred Value (%) 80 50 85 mass spectra, which are shown in Figure 2. The only difference Neg Pred Value (%) 99 99 99 noticed in the ECNCIspectra of the two peaks was the ratio of the molecular ion relative to the base peak. Peak 1 had a molecular ion intensity of less than 10%,whereas peak 2 had a molecular ion of about 35% of the base peak. Peaks 1 and 2 had retention times of 1.07 and 1.17, respectively,relative to oxazepam-d5,which eluted at 4:08 rain. The other peaks are the other benzodiazepine internal standards. Peak 3 is thought to be a sertraline metabolite that could not be identified. Forty-three specimens submitted for ~c GC-MS analysis confirmed positive for benzodiazepines. Most of these specimens (iV = 27) contained oxazepam and related metabolites (nordiazepam and temazepam) at total benzodiazepine concentrations ranging from 34 to 4207 ng/mL. Twelve specimens had lorazepam (range, 29-719 4:05 4:30 3:16 3:41 4:54 5: ng/mL) combined with oxazepam and Time (min) related metabolites(range, 78-1340 ng/mL). Figure 1, Total ion chromatogram for a patient taking sertraline. Peaks 1 and 2 are sertraline metaboThe other positive specimens contained lites with retention times of 1.07 and 1.17 min relative to oxazepam-ds ($) which eluted at 4:08 rain. lorazepam as the only benzodiazepine idenPeak3 is an unidentified compound thought to be a sertraline metabolite. Peakslabeled with an asterisk tiffed at concentrations of 62, 412, 553, and are the other benzodiazepine internal standards. 619 ng/mL. The four specimens that con=
Iournal of Analytical Toxicology,Vol. 21, lanua~/February 1997
and long half-life of metabolites limited the usefulness of the original CEDIAassay. We obtained sertraline and N-desmethylsertraline from the manufacturer, Pfizer, to identify the compounds responsible for the observed crossreactivity in the original CEDIAassay. Drug-free urine samples 288 100% spiked with these compounds crossreacted with the CEDIAassay for benzodiazepines, causing a positive result at 500 ng/mL. However, after extraction, derivatization, and CC-MS analysis of these standards, it was clear that the compounds present in the patients' urine were not the parent or demethylated analogue. This result was not totally unexpected because it had been previously reported that these compoundswere 378 present in verysmall mounts in urine specimens (5). The false-positive urine specimens were analyzed by ECNCI GC-MS for the major ions of sertraline and desmethylsertraline. 245 II 270 ~ alS 344 3~2 . . , I. sso Neither of these compounds as detected. A review of the literature showed that in rats m/z and dogs, after demethylation, sertraline is Figure 2. Electron capture negative chemical ionization spectra consistentwith the trimethylsilylat~l further oxidatively deaminated to a ketone hydroxy ketone metabolite of sertraline (MW= 378). This compound exists as a pair of diastereomers and eventually to a stereoisomeric pair of (peaks 1 and 2 in Figure 1). keto-alcohols (Figure 3) that are subsequently glucuronidated (6). The negative ion mass spectra of the two peaks in patient NH2 specimens were similar; both had base peaks of m/z 288 with other diagnostic ions at m/z 378 and 252. These spectra are consistent with the hydroxy ketone metabolite of sertraline which exists as a pair of diastereomers with molecular weights of 378 O atomic mass units. The other two ion fragCI ments are consistent with the loss of trimethylsilyl alcohol followed by a further N-Desmethylsertraline loss of HCI. These losses have been observed in our previous studies of benzodiazepines (2). Although we have not conclusively identified which metabolite of sertraline is CI NHCH3 responsible for the crossreactivity, the fact that sertraline and the demethylated metab~-Hydroxyketone ias~ereomers olite do crossreact with the original CEDIA benzodiazepine assay, combined with the patient histories and GC-MS data, is strong evidence that a metabolite of sertraline O caused the false-positive results. ~ CI Because benzodiazepines are excreted as glucuronides, which are alpha to a carSertraline Ketone bonyl group, it is understandable that an metabolite immunoassay designed to detect benzodiazepines could have crossreacted with sertraline metabolites, which are excreted as Cl alpha hydroxy ketone glucuronides. Apparently, these functional groups were an Figure 3. Partial metabolic route of sertraline, important part of the antigenic determinant
4--8 h postadministration, followed by first order elimination with a plasma half-life of 26-32 h. N-Desmethylsertraline, its primary metabolite in plasma, is 8-10 times less potent and has a longer elimination half-life of 71-200 h (4). Widespread use
Journal of Analytical Toxicology,Vol. 21, January/February1997
of the antibody used in the initial CEDIAassay. According to 8oehringer Mannheim, the antibody and the antigen were changed in the current CEDIAassay to eliminate the sertraline crossreactivity. We recently compared several different benzodiazepine immunoassay screens (Roche ONLINE,Abbott TDx, Syva EMIT,and Diagnostic Products Corporation and Biosite TRIAGE)and did not find that any of these assays were subject to false-positive results due to sertraline metabolites (7), suggesting that the antibodies employed have different antigenic determinant than the initial CEDIAassay. In addition, we tested the Roche ONLINE, the improved CEDIAassay, and EMIT II with concentrations of sertraline and desmethyl sertraline up to 1000 ng/mLand did not get a positive response. None of the sertraline patients that were positive on the original CEDIAassay were positive by Roche ONLINE,the improved CEDIAassay, or EMIT II. When screening urine for drugs, which sometimes occurs in postmortem forensic toxicology, it is necessary to analyze for the keto-alcohol forms of sertraline to ensure that use of this compound will be reliably detected. The combination of glucuronidase hydrolysis and derivatization will also enable the analyst to detect the polar metabolites of a variety of other drugs as well. Neither EMIT II nor the CEDIAassays were able to consistently detect lorazeparn in patient specimens. The improved CEDIA assay did detect Iorazepam in one specimen at 619 ng/mL, which appeared to be near the detection limit for this compound because it was negative for a specimen that contained 553 ng/mL. The negative predictive value of all three immunoassays for the 5-aryl-7-chlom-l,4-benzodiazepine compounds was very good, although these assays did not detect iorazepam reliably. The positive predictive values of the improved CEDIAassay (80%) and EMITII (85%) were better than that of the original CEDIAassay (50%), but all assays were subject to at least six false positives. The false positives were falsely elevated by all three immunoassays, suggesting that these samples contained a benzodiazepine not detected by the GC-MS method (one possibility is flunitrazepam or its metabolites, which have not been evaluated by our GC-MS method) or some crossreacting
substance such as oxaprozin (8). The improved CEDIA assay has clearly eliminated the problem of sertraline crossreactivity, which will substantially enhance the clinical utility of this product.
Acknowledgment The authors would like to thank Kenneth Schlussel and Susan Priem for technical assistance. Boehringer Mannheim is gratefully acknowledged for supplying reagents and financial assistance.
References 1. R.L. Fitzgerald, S. Priem, T. Leech, K. Schlussel, and D.A. Herold. Comparison of CEDIA and ONLINE drugs of abuse assays. Clin. Chem. 41($6): 5125 (1995). 2. R.L. Fitzgerald, D.A. Rexin, and D.A. Herold. Benzodiazepine analysis by negative chemical ionization gas chromatography/mass spectrometry. J. Anal Toxicol. 17:342-47 (1993). 3. N.F. Belier, H.A. Scholtz, M.L. Opel, E. Padilla, F.R. Galloway, and P. Khanna. Improved CEDIA| DAU benzodiazepine assay for urine drug testing. J. Anal ToxicoL 20:76 (1996). 4. S.J.Warrington. Clinical implications of the pharmacology of sertraline. Int. Clin. Psychopharmacol. 2:11-21 (t991). 5. B. Levine, A.J. Jenkins, and i.E. Smialek. Distribution of sertraline in postmortem cases. J. Forensic Sci. 18:272-74 (1994). 6. L.M. Tremaine, W.M. Welch, and R.A. Ronfeld. Metabolism and disposition of the S-hydroxytryptamine uptake blocker sertraline in the rat and dog. Drug Metab. Dispos. 17" 542-50 (1989). 7. R.L. Fitzgerald, D. Rexin, and D.A. Herold. Detecting benzodiazepines: Immunoassays compared with negative chemical ionization gas chromatography/mass spectrometry. Clin. Chem. 40: 373-80 (1994). 8. P.D. Camara, L. Audette, K. Velletri, P. Breitenbecher, M. Rosner, and W.C. Griffiths. False positive immunoassay results for urine benzodiazepine in patients receiving oxaprozin (Daypro). Clin. Chem. 41:115-16 (1995). Manuscript received February 12, 1996; revision received July 8, 1996.