ANALYST MAY 1986 VOL. 111 525 An Iodine-I25 Radioimmunoassay for the Direct Detection of Benzodiazepines in Blood and Urine Colin P. Goddard A. Howard Stead Peter A. Mason Brian Law and Anthony C. Moffat" Central Research Establishment Home Office Forensic Science Service Aldermaston Reading Berkshire RG74PN UK and Margaret McBrien and Simon Cosby Northern Ireland Forensic Science Laboratory 180 Newtownbreda Road Belfast BT8 4QR UK A radioimmunoassay (RIA) for the direct detection of benzodiazepines in blood and urine is described. It is based on a commercially available antiserum and an easily synthesised radio-iodinated derivative of clonazepam that allows the use of relatively simple gamma-counting procedures. The assay can detect low therapeutic levels of all of the benzodiazepines currently available in the UK in 50-$ samples of blood and urine (1-50 ng ml-I depending on the drug); no prior sample preparation is required.It is inexpensive rapid, simple to perform and is broadly specific for the benzodiazepine class of drugs. The assay offers a most suitable means of screening large numbers of samples of forensic interest for the presence of the benzodi azepi nes. Keywords Radioimmunoassay; benzodiazepine detection; blood; urine The benzodiazepines are a group of chemically related drugs used mainly as hypnotics and sedatives. Forensic interest in these compounds is due primarily to their frequent occurrence in drugsldriving cases,1-3 i.e. those cases where a drug is suspected of having made a significant contribution to the impairment of driving ability.A large number of drugddriving cases are seen by forensic science laboratories each year each case requiring an analysis for unknown drugs and metabolites in very small volumes of blood. With the benzodiazepines in particular the analyst is faced with further problems arising from the fact that there is a steadily increasing number of different compounds available (20 in the UK alone4) and because commercial companies are producing more potent drugs all the time the newer com-pounds are often present at very low concentrations. A rapid and simple screening assay for benzodiazepines that is sensitive and broadly specific to this group of drugs is therefore desirable. Methods currently used to detect benzodiazepines include high-performance liquid chromatography (HPLC) ,5,6 gas -liquid chromatography with electron-capture detection (GLC - ECD)7-9 and radioimmunoassays (RIAp14 and radiorecep-tor assays (RRA)15,16 using tritiated labels.Use of HPLC and GLC for routine screening is relatively time consuming. In comparison immunoassay techniques offer many advantages as methods for the efficient screening of large numbers of biological samples for the presence of chemically related drugs. Unfortunately of those RIAs available for the detection of benzodiazepines most are specific to particular drugs and assays of this type are in general of limited use for drug screening. Radioreceptor assays and screening RIAs capable of detecting a broad range of benzodiazepines are convenient to use but published methods are based on tritiated labels and suffer the disadvantages associated with liquid scintillation counting viz.the need for long counting times the need for prior sample extraction the use of expensive scintillant and problems associated with quenching when coloured biological samples are assayed. Two enzyme immunoassay kits one for benzodiazepines in serum and one for benzodiazepine metab-* Present address Home Office Forensic Science Laboratory, Hinchingbrooke Park Huntingdon PE18 8NP. Crown Copyright 1986. olites in urine are commercially available [Syva (UK), Maidenhead Berkshire] but neither has the sensitivity required to detect low therapeutic levels of all available benzodiazepines. In addition neither can cope with the haemolysed blood samples often encountered in forensic cases.The screening assay described in this paper is based on an iodine-125 labelled derivative of clonazeparn. It is sensitive, broadly specific and allows the direct analysis of forensic samples. Experimental Materials and Equipment All chemicals unless specified otherwise were obtained from BDH Chemicals Poole Dorset . Phosphate buffer (0.1 M pH 7.4) containing 0.2% mlV of bovine y-globulin (Cohn fraction I1 from Sigma Chemical, Poole Dorset) and 0.01% mlV of sodium azide was used throughout the assay. The antisera were obtained from Emit TOX serum benzodiazepine and Emit DAU benzodiazepine metabolite kits [Syva (UK)]. The Emit TOX antiserum was diluted 1 + 299 and the Emit DAU antiserum 1 + 999 with assay buffer immediately before each run of the assay.These dilutions give approximately 50% binding with the amount of radiolabel added to each assay tube. The maximum binding of label to each antiserum was approximately 75-80%. The radiolabelled benzodiazepine derivative [7-1251]-iodoclonazepam (specific activity 5.6 TBq mmol-1; 12.7 MBq pg-1) was prepared from the 7-amino derivative of clonazepam via a diazonium intermediate as previously described17 and stored in methanol at 4 "C. It was diluted with assay buffer to give approximately 10000 counts min-l (550 Bq; 0.099 ng) per 100 p1. Standard solutions of diazepam were prepared in synthetic urine at concentrations of 0 1 2 3 5 10 and 20 ng ml-1. The synthetic urine was prepared by adding 22 g of urea 1.8 g of NaH2P04.2H20 1.1 g of Na2HP04 8.25 g of NaCl 5.2 g of KCl 1.5 g of creatinine and 0.1 g of NaN3 to distilled water and making the total volume up to 1 1 (modified from reference 18).Swine serum was obtained from Flow Laboratories Ltd., Irvine Ayrshire 526 ANALYST MAY 1986 VOL. 111 synthetic urine samples containing 1.5 and 5 ng ml-1 of diazepam. The effects of various urine and blood preservatives on the assay were studied by adding the following preservative tablets (1) to known blank urine samples (2.5 ml) phenyl-mercury(I1) nitrate and sodium fluoride (50 + 100 mg), sodium fluoride and sodium sulphate (300 + 300 mg) and sodium azide (50 mg); (2) to known blank blood samples (1 ml) sodium fluoride and potassium oxalate (37.5 + 18.7 mg) and sodium nitrite (25 mg).The final concentrations of the added preservatives were approximately 5 (urine) and 2.5 (blood) times greater than those recommended for forensic use. The assay was used to measure serum levels of diazepam following oral ingestion of single therapeutic doses of the drug (10 mg) by two volunteers. Blood samples were taken before drug ingestion and then at intervals over a period of 36 h afterwards. The Northern Ireland Forensic Science Laboratory (NIFSL) conducted a trial to compare the results obtained using the described RIA with those obtained by conventional methods. The NIFSL routinely uses GLC with ECD GLC with NPD and HPLC methods for the detection and quanti-fication of benzodiazepines in case samples. The trial involved using chromatographic techniques in parallel with the RIA method on each of 80 samples (including clinical driving and post-mortem cases) where benzodiazepines were suspected.The levels of both parent drug and any major active metabolites were obtained by the chromatographic methods, and these were compared with the total benzodiazepine level as measured by RIA. Polyethylene glycol (PEG M 8000) was obtained from Sigma Chemical and was used to prepare a 27.5% mlV solution in distilled water (550 g of PEG plus 1525 ml of water). Disposable polypropylene microcentrifuge tubes (plastic point) were obtained from Alpha Laboratories Eastleigh, Hampshire. Gamma-counting was performed using an NE 1600 counter (Nuclear Enterprises Beenham Berkshire) which had an efficiency of approximately 50% for iodine-125.Method The dilution for the two commercially available antibodies was determined by plotting dilution curves for each against the purified [7-12~I]iodoclonazepam and measuring the dilu-tion needed to bind approximately 50% of the total radio-activity. Cross-reactivities of a small number of benzodiazepines (chlordiazepoxide diazepam flurazepam lorazepam nitra-zepam oxazepam temazepam and triazolam) were obtained using both the TOX and the DAU antisera at the optimum dilution by comparing the calibration graph of each with that of diazepam. The results indicated that the TOX antiserum was more suitable for a general screening assay owing to a narrower range of cross-reactivities; all further work was performed using only this antiserum.The cross-reactivities of all the benzodiazepines currently available in the UK, together with some of their major active and inactive metabolites were then obtained. The concentration of each drug required to give a 50% depression of binding was measured and expressed relative to the value for diazepam. The extent of cross-reaction of several commonly encoun-tered but structurally unrelated drugs was also measured and compared with the cross-reactivity of diazepam. The depres-sion of binding was measured using solutions containing 100 pg ml-1 of the drug unless stated otherwise. The drugs included acetylsalicylic acid (400 pg ml-I) amitriptyline, amphetamine bromocryptine butriptyline caffeine, codeine ephedrine glutethimide imipramine isoprenaline, lysergic acid diethylamide methadone methylphenobar-bitone mianserin morphine nicotine paracetamol (250 pg ml-I) phenytoin and protriptyline.As diazepam is one of the strongest cross-reacting benzo-diazepines and is also one of the most commonly prescribed, it was used to prepare solutions for the standard curve. Assays were performed in duplicate using the following conditions. Into each microcentrifuge tube were pipetted 100 p1 each of sample radiolabel and antiserum. The 100 p1 of sample consisted of (for urine samples and standards) urine or standard (50 pl) and swine serum (50 pl) or (for blood samples) blood (50 p1) together with synthetic urine (50 pl). This procedure ensured that each tube contained approxi-mately the same amounts of salt protein and water.The tubes were then capped vortexed and incubated at room temperature for at least 1 h. Equilibration was attained after 60 min and was stable for up to 24 h after that time. PEG solution (500 pl) was then added and the tubes were re-capped vortexed thoroughly and then centrifuged at 9000 g for 3.5 min. The supernatant was removed by aspiration and the tubes containing the residual bound fraction were each counted for 60 s in the gamma counter. Whole blood samples were assayed neat and after dilution by factors of 10 and 100 with swine serum. Urine samples were assayed neat and after dilution by factors of 10 100 and 1000 with synthetic urine. Cut-off levels for the assay were determined by the analysis of 100 unpreserved blank urine and blood samples obtained from volunteers not taking any drugs.The conditions of the blood samples varied from fresh unhaemolysed to haemo-lysed/putrefied. Intra- and inter-assay coefficients of variation for the assay were determined by repeated analyses of Results and Discussion The optimum antibody dilutions (ie. those giving 50% binding of the added radiolabel) were measured as 1 + 299 for the Emit TOX antiserum and 1 + 999 for the Emit DAU and these dilutions were used in all further work. Table 1. Cross-reactivities of eight common benzodiazepines relative to diazepam. Determined at 50% depression of binding (equivalent to 2.5 ng ml-1 of diazepam) Relative reactivity Benzodiazepine Emit TOX Emit DAU Diazepam .. . . . . . 1 .O 1 .o Triazolam . . . . . . . 1.4 3.3 Flurazepam . . . . . . 2.8 2.5 Temazepam . . . . . . 4.8 12 Nitrazepam . . . . . . 6.4 8.8 Oxazepam . . . . . . . . 11 38 Lorazepam . . . . . . 36 107 Chlordiazepoxide . . . . 52 115 0 10 20 Diazepam concentrationhg ml-1 Fig. 1. Typical calibration graph for diazepam obtained using radioirnmunoassa ANALYST MAY 1986 VOL. 111 527 Table 2. Cross-reactivities of benzodiazepines and their metabolites with the Emit TOX antiserum Concentration reauired Therapeutic Compound Prazepam . . . . . . . . Diazepam . . . . . . Pinazepam . . . . . . Alprazolam . . . . . . Triazolam . . . . . . N-Desmethyldiazepam . Flurazepam . . . . . . Midazolam . . . . . . Temazepam . . . . . . Tetrazepam .. . . . . Clorazepate . . . . . . Nitrazepam . . . . . . Clobazam . . . . . . Flunitrazepam . . . . . . Medazepam . . . . . . Oxazepam . . . . . . Lormetazepam . . . . . . Ketazolam . . . . . . Loprazolam . . . . . . Clonazepam . . . . . . Lorazepam . . . . . . Desalkylflurazepam . . . . Bromazepam . . . . . . Chlordiazepoxide . . . . Desmethylclobazam . . . . Demoxepam . . . . . . Desmethylchlordiazepoxide Prazepam benzophenone . . Diazepam benzophenone . . 7-Acetylaminonitrazepam . . Clozapine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . * .. . . . . . . . . . . . . . . . . . . . . . . . . . . . Relative activity* 0.60 1 .oo 1.20 1.20 1.40 2.40 2.80 3.20 4.80 5.6 6.0 6.4 10.8 10.8 10.8 11.2 11.2 14.8 21.6 32 36 44 52 52 52 >loo > 100 >loo > 100 >>loo >>loo for 50% depresiion of bindinghg ml-1.5 2.5 3.0 3 .O 3.5 6.0 7.0 8.0 12 14 15 16 27 27 27 28 28 37 54 80 90 110 130 130 130 >250 >250 >250 >>250 >250 ~ 2 5 0 rangel ng ml-lf 8-40 50-2000 --5-25 100-1 500 1-5 -360-850 300-1500 20-150 300-900 10-160 500-2000 -5-20 ---10-100 50-240 30-90 80-150 1000-3000 2500-3500 ------* Determined at 50% depression of binding (equivalent to 2.5 ng ml-I of diazepam).t Approximate range of blood concentrations at the steady state wherever possible; data from references 19 and 20. 100 8 & 2 50 0 I I I 1 102 104 +/ Drug concentrationhg ml-1 Fig. 2. Full range of cross-reactivities for benzodiazepines. B = bound activity; B = bound activity when no unlabelled drug is present. A Prazepam B bromazepam The relative cross-reactivities of eight common benzodiaz-epines were measured for both the Emit TOX and the Emit DAU antisera. The results (Table 1) indicate that the two antisera are very similar but that the spread of cross-reactivity values is less with the Emit TOX antiserum. The most important feature of any screening assay is that the sensitivity towards the least cross-reacting drugs is sufficient to detect them at the required levels.The Emit TOX antiserum is more sensitive towards drugs of lower cross-reactivity and conse-quently it was selected for the development of the general screening assay; this antiserum was used in all further work. A typical assay calibration graph for diazepam is shown in Fig. 1. The relative cross-reactivities of all available benzodiaz-epines together with a number of their active and inactive metabolites were measured using the Emit TOX antiserum. The results are given in Table 2. The full range of cross-reactivities of the 20 benzodiazepines available in the UK,4 from prazepam (the strongest) to bromazepam (the weakest), is shown in Fig. 2.The range of cross-reactivities demon-strated by these results shows that all of the parent benzo-diazepines bind relatively strongly to the antiserum their desalkyl metabolites less so and compounds in which the diazepine ring structure is broken (the substituted benzo-phenones) exhibit greatly decreased binding. Further the range suggests that the assay is sufficiently sensitive to detect all available benzodiazepines. Even low-dose drugs such as alprazolam prazepam and triazolam should be easily detected at therapeutic levels because their cross-reactivities are approximately the same as that of diazepam. Ironically, because of its poor cross-reactivity one of the drugs that may be most difficult to detect at low therapeutic levels is clonazepam the drug from which the radiolabel is derived.Other drugs that might prove difficult to detect in similar circumstances include flunitrazepam and flurazepam. However even with these compounds therapeutic levels will usually be above the assay cut-off values and therefore will be detectable. Of the 20 common non-benzodiazepine drugs that were tested for their cross-reactivity in the assay only morphine at a concentration of 100 pg ml-1 caused a depression in binding greater than the positivehegative cut-off value of the assay (see below). This concentration is much higher than normal fatal levels of the drug. The analysis of 100 unpreserved blank blood samples gave a mean level of background cross-reactivity of 0.28 k 0.18 (S.D.) ng ml-l. Similar analyses of 100 blank urine samples gave a mean value of -0.02 k 0.16 ng ml-1.The mean plu 528 ANALYST MAY 1986 VOL. 111 Table 3. Comparison of RIA and GLC/HPLC analyses of blood and urine samples from clinical driving and post-mortem cases involving benzodiazepines GLC or HPLC resulthg ml-1 Drug Diazepam . . . . . . . . . . . . . Lorazeparn . . . . . . . . . . . . . . Flurazepam . . . . . . . . . . . . . . . . Temazepa-m . . . . . . . . . . . . . . . . Lormetazepam . . . . . . . . . . . . . . . . Chlordiazepoxide . . . . . . . . . . . . . . Bromazepam . . . . . . . . . . . . . . . . Desmethyldiazepam/flurazepam . . . . . . . . Diazepam/desmethyldiazepam/ temazepam . . . . . . . . . . . . . . . . Sample [blood (b) , urine (u)] b b b b b b b b b b b b b b b U U b b b b b b b b b b Parent 30 70 200 420 260 460 460 1250 1000 70 1820 130 160 320 100 1400 450 310 750 340 2180 40 240 2600 400/1000 -300/440/3 10 Met a bolit e * --340 130 90 670 ------+ 390 Diazepam equivalent (total) 1-30 70 142 254 420 298 460 739 1250 1000 70 1820 3.6 4.4 8.9 2.8 38.9 >161 120 156 71 454 3.6 4.6 50 524 548 RIA result/ ng ml-* 98 147 210 250 270 350 570 630 820 850 1000 1230 2.3 8.0 10.4 23 41 1630 2300 120 240 520 3.8 27 32 430 540 * Metabolite levels were not determined unless stated otherwise; - no or only trace amounts of parent or metabolite detected; + metabolite t Corrected for cross-reaction of parent drug and/or its metabolite.detected but not quantified. three standard deviations is 0.82 ng ml-1 for blood and 0.46 ng ml-1 for urine. A single positivehegative cut-off level of 1.0 ng ml-1 was set for both blood and urine samples for simplicity ensuring a >99.7%0 probability of obtaining a true positive result. Coefficients of variation for 1.5 and 5 ng ml-1 of diazepam in synthetic urine were 6.7 and 7.9% intra-assay (n = 20) and 6.4 and 11.4% inter-assay (n = 20) respectively. Of the blood and urine preservatives tested the only one to cause a measurable depression in binding of the radiolabel to the antiserum was the sodium fluoride - sodium sulphate (300 + 300 mg) preservative when added to urine at concentrations five times the normal.However as urinary benzodiazepine concentrations are generally much higher than those seen in blood and as the measured background levels in urine caused by this preservative are only equivalent to approximately 24 ng ml-1 of diazepam it is unlikely that the interference will create any serious analytical problem. Despite this urinary levels of less than 30 ng ml-1 should be treated with caution unless a complementary blood level is available. The results of RIA analysis of blood samples obtained from two volunteers given a single oral dose of diazepam are shown in Fig. 3; peak levels of diazepam in both subjects are over 300 ng ml-1. Most points shown in Fig. 3 were obtained following dilution of samples by a factor of 10 or 100 (shortly after ingestion) demonstrating that the assay is capable of measur-ing therapeutic levels of diazepam in blood with at least an order of magnitude to spare.Because of this assay results must be interpreted with great care as a positive result might easily be obtained in a sample taken long after any pharmaco-logical effects have ceased. In practice it is recommended that the assay is performed both on neat samples of blood and after dilution of samples by factors of 10 and 100 and the higher drug levels found in urine may require sample dilution by factors of 10 100 and 1000 in order to obtain results that can be measured accurately from the calibration graph. Of the 80 samples tested by the NIFSL 55 were positive for benzodiazepines by RIA; 19 gave results (mean 28 ng ml-1 diazepam equivalents; range 1.3-180 ng ml-1) that could not be confirmed independently by GLC or HPLC.The remain-ing 36 samples were positive both by GLC or HPLC and by RIA. Nine of these 36 results were rejected for the compari-son study either owing to lack of data on metabolite concentrations or because the parent drug was only detected in trace amounts by GC or HPLC. Table 3 summarises the data obtained from the remaining 27 cases. The range of drugs identified in the case samples is fairly typical of that occurring in routine toxicology. Parent drug and/or metabolite levels as measured by GLC or HPLC are presented. These results ar ANALYST MAY 1986 VOL. 111 529 1.5 3 4.5 Time after ingestiodh Fig.3. single 10-mg oral dose of the drug Serum concentrations of diazepam in two volunteers given a also expressed as total diazepam equivalents calculated from the cross-reactivities of the respective compounds to allow direct comparison with the RIA results. A statistical comparison of the RIA and GLC or HPLC methods was performed using diazepam and lorazepam as examples. Diazepam provides an example typical of those benzodiazepines that undergo complex metabolism giving rise to one or more active metabolites; its major metabolite in blood is desmethyldiazepam. Lorazepam provides an example of a benzodiazepine that is not extensively metabolised and for which active metabolites have not been identified. Good correlations exist between results from the two methods if the parent drug and its major active metabolite in blood are measured as with diazepam ( r = 0 .7 6 ; ~ <0.005 y = 0.506~ + 259 n = 12). Similarly where the drug has no active metabolites and only the parent compound is measured as with lorazepam there is again a relatively high degree of correlation ( r = 0.85 p C0.02 y = 0.0849~ + 6.99 n = 5). In both instances drug levels measured by RIA are generally higher than those obtained by the chromatographic methods. This is not unexpected as it is likely that the RIA will detect those chemically related metabolites of the benzodiazepines that the more specific GLC and HPLC techniques will not. The only results to show large differences between the two methods were observed with flurazepam.RIA results were much higher than chromatographic results even after allow-ing for the major metabolite desalkylflurazepam. It is possible that other metabolites that cross-react well are present in sufficient amounts to raise the level obtained by RIA. Metabolites such as desethylflurazepam didesethylfluraze-pam and N-1-hydroxyethylflurazepam were not measured by chromatography but should cross-react and may be present in significant concentrations. The combination of the Emit TOX antiserum and the [7-12~I]iodoclonazepam label provides an assay that is specific for the benzodiazepines as a group making it ideal as a general screen for this class of drugs. The examination of case samples has shown that by comparison with the less sensitive chromatographic techniques used the RIA gives no false negative results.There is overall good agreement between the results obtained by RIA and those obtained using these chromatographic methods. However the broad range of cross-reactivities and the interaction of active metabolites with the assay means that quantification based on RIA is of little value unless the identity of the drug is known; even low levels detected in screening may represent significant amounts of drugs such as lorazepam and bromazepam. A positive result can only be a semi-quantitative indication of the presence of a drug and even if a single compound is known to be present an accurate quantification of concentrations should not be attempted by this method unless a calibration graph for the specific drug is prepared.All positive results should be confirmed and the drug identified by more specific procedures such as GLC HPLC or mass spectrometry. Conclusions In conclusion the assay is easy to perform quick to set up and run inexpensive and very reliable. Intra- and inter-assay results confirm its reproducibility. The sensitivity and group specificity of the assay allow low therapeutic levels of all of the benzodiazepines currently available in the UK to be detected in very small volumes (50 p1) of sample without the need for any prior preparation procedures. The antiserum used is commercially available. The assay has been used successfully for testing for the presence of benzodiazepines in a large number of cases submitted for forensic analysis and should prove extremely useful to both the clinical and the forensic toxicologist.1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. References Garriott J. C. DiMaio V. J. M. Zumwelt R. E. andPetty, C. S. J. Forensic Sci. 1977 22 383. Robinson T. A. J. Forensic Sci. SOC. 1979 19 237. Taylor J. F. in Goldberg L. Editor “Alcohol Drugs and Traffic Safety,” Almqvist and Wiksell Stockholm 1981, p. 478. “MIMS Monthly Index of Medical Specialities,” Haymarket Press London August 1984. Osselton M. D. Hammond M. D. and Twitchett P. J., J. Pharm. Pharmacol. 1977 29 460. Mehta A. C. Talanta 1984 31 1. Greenblatt D. J. Divoll M. Moschitto L. J. and Shader, R. I. J. Chromatogr. 1981 225 202. Zlatkis A. and Poole C. F. Editors “Electron Capture, Theory and Practice in Chromatography,” Elsevier Amster-dam 1981 p. 306. Douse J. M. F. J. Chromatogr. 1984 301 137. Gelbke H. P. Schlicht H. J. and Schmidt G. Arch. Toxicol. 1977 38 295. Rutterford M. G. and Smith R. N. J. Pharm. Pharmacol., 1980 32 449. Dixon R. in Langone J. J. and Vunakis H. V. Editors, “Methods in Enzymology Volume 84 Immunochemical Tech-niques Part D Selected Immunoassays,” Academic Press, New York and London 1982 p. 490. Aderjan R. and Schmidt G. 2. Rechtsmed. 1979 83 191. Robinson K. Rutterford M. G. and Smith R. N. J. Pharm. Pharmacol. 1980 32,773. Jochemsen R. Horbach G. J. M. J. and Breimer D. D., Res. Commun. Chem. Pathol. Pharmacol. 1982 35 259. Lund J. Scand. J. Clin. Lab. Invest. 1981 41 275. Goddard C. P. Law B. Mason P. A. and Stead A. H., J. Labelled Cpds. Radiopharm. in the press. Rodgers R. Crowl C. P. Eimstad W. M. Hu M. W. Kam, J. K. Ronald R. C. Rowley G. L. and Ullman E. F. Clin. Chem. 1978,24 95. Schutz H. “Benzodiazepines A Handbook,” Springer-Verlag Berlin 1982. Stead A. H. and Moffat A. C. Human Toxicol. 1983 3, 437. Paper A51393 Received October 31st 1985 Accepted December 9th I98