Gas chromatographic determination of cocaine and its metabolites in blood and urine from cocaine users in northwestern Spain

COCAINE USERS IN NW SPAIN 283

Copyright © 2004 John Wiley & Sons, Ltd. J. Appl. Toxicol. 24, 283–287 (2004)

JOURNAL OF APPLIED TOXICOLOGYJ. Appl. Toxicol. 24, 283–287 (2004)Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jat.994

Received 6 April 2004Revised 6 May 2004

Copyright © 2004 John Wiley & Sons, Ltd. Accepted 7 May 2004

* Correspondence to: P. Fernández, Institute of Legal Medicine, ForensicToxicology Service, Faculty of Medicine, C/San Francisco, s/n. 15782,Santiago de Compostela, Spain.E-mail: [email protected]

Gas Chromatographic Determination ofCocaine and its Metabolites in Bloodand Urine from Cocaine Users inNorthwestern Spain

P. Fernández,* L. Buján, A. M. Bermejo and M. J. TaberneroInstitute of Legal Medicine, Forensic Toxicology Service, Faculty of Medicine, University of Santiago de Compostela,Spain

Key words: GC/FID; solid-phase extraction; derivatization; intoxications; cocaine; benzoylecgonine; ecgonine methyl ester;urine; blood.

A gas chromatographic method with flame ionization detection (GC/FID) was developed for the deter-mination of cocaine and its metabolites in blood and urine samples from cocaine users in NorthwesternSpain. After a solid-phase extraction with Bond Elut Certify cartridges and a derivatization with bis-trimethylsilyltrifluoroacetamide–trimethylchlorosilane (1%), calibration curves were constructed over 0.4–4 µµµµµg ml−−−−−1 for urine and 0.1–2 µµµµµg ml−−−−−1 for blood, using proadifen as the reference compound. The average extractionrecoveries were 75% for urine and 78% for blood. The limits of detection and quantitation were 0.071 and0.24 µµµµµg ml−−−−−1, respectively. Coefficients of variation were <<<<<10% and accuracy was within ±±±±±12%. The averageblood concentrations of cocaine, benzoylecgonine and ecgonine methyl ester in 42 living patients were 0.22,1.43 and 0.16 µµµµµg ml−−−−−1, respectively. Urine samples were collected from individuals in the criminal justice system(70 cases), from drug abusers admitted to emergency rooms (36 cases) and from patients under detoxificationtreatment (36 cases). The second group exhibited the highest average concentrations (e.g. 0.97 µµµµµg ml−−−−−1 forcocaine, 5.23 µµµµµg ml−−−−−1 for benzoylecgonine and 0.39 µµµµµg ml−−−−−1 for ecgonine methyl ester). Sixty-five fatal intoxicationsdue to cocaine alone or in combination with other drugs were studied, and average blood levels were found tobe higher in the deaths related to cocaine alone (e.g. 0.40 µµµµµg ml−−−−−1 for cocaine, 2.38 µµµµµg ml−−−−−1 for benzoylecgonineand 0.38 µµµµµg ml−−−−−1 for ecgonine methyl ester). Copyright © 2004 John Wiley & Sons, Ltd.

0.5 µg ml−1 are potentially fatal. However, death and toxic-ity can occur following ingestion of trivial amounts of thedrug and be associated with very low levels in plasma. Infact, there is no upper limit for lethal blood concentra-tions, nor is there an absolutely safe lower limit (Smart &Anglin, 1986; Karch, 2002). Even though generally only10–20% of the dose is released unaltered in urine, cocainecan be found at relatively high concentrations in urine fromcrack users, and chronic users have been reported to havepersistent levels of cocaine in urine (Schramm et al., 1993).

In this work, gas chromatography with flame ionizationdetection (GC/FID) was developed to determine cocaineand its metabolites in blood and urine samples from 184living patients and 65 individuals who died from an over-dose of cocaine, whether alone or in combination withother drugs.

EXPERIMENTAL

Reagents and standards

All solvents used were purchased from Merck (Darmstadt,Germany). Proadifen, cocaine, BEG, EME, N,O-bis-trimethylsilyltrifluoroacetamide (BSTFA) and trimethyl-chlorosilane (TMCS) were supplied by Sigma-Aldrich(Madrid, Spain). Working solutions were 2, 5, 10, 20 and

INTRODUCTION

Drug testing for cocaine use is done in many situations,including drug treatment, employment screening andinvestigations of the criminal justice system (Preston et al.,1999). Although benzoylecgonine (BEG) is normally usedto provide evidence for cocaine consumption, the detec-tion of ecgonine methyl ester (EME) and cocaine (COCA)can provide valuable, significant additional forensic infor-mation (De Giovanni & Rossi, 1994; Peterson et al., 1995;Kidwell et al., 1997).

Once absorbed, cocaine undergoes enzymatic hydrolysisto BEG and EME. Because EME retains the methyl esterfunction, it is liable — like cocaine — to chemical hydroly-sis, so it is rapidly removed from blood in vivo (Pan &Hedaya, 1997). As a result, blood concentrations ofcocaine and/or EME higher than that of BEG can berelated to recent use of the drug. The interpretation ofblood levels also depends on how the cocaine is adminis-tered (Isenschmid et al., 1992). Thus, Corburt and Koves(1994) believe that blood concentrations of cocaine above

284 P. FERNÁNDEZ ET AL.


50 µg ml−1 for each drug and 20 µg ml−1 of proadifen(reference compound) in methanol. Urine and blood solu-tions for calibration were prepared by adding appropriateamounts of the working solutions to drug-free samples.The concentrations were 0.4, 0.8, 1.6, 3.2 and 4 µg ml−1 inurine and 0.1, 0.2, 0.5, 1 and 2 µg ml−1 in blood.

The samples from patients studied were frozen until theiranalyses in order to avoid the potential decay of cocaine(Hippensteil & Gerson, 1994).

Chromatographic conditions

All samples were analysed on a Hewlett–Packard 6890gas chromatograph furnished with an HP-1 cross-linkedmethylsiloxane column (12 m × 0.2 mm × 0.33 µm). Theinjector was operated in the split mode at 250 °C, using asplit ratio of 22.5:1 and helium as the carrier gas. The oventemperature was raised from 130 to 280 °C in three ramps;thus, the initial temperature was held for 3 min, thenincreased at 20 °C min−1 to 200 °C (held for 1 min), thenraised at 5 °C min−1 to 250 °C (held for 1 min) and finallyincreased at 20 °C min−1 to 280 °C (held for 3 min). Theflame ionization detector was kept at 290 °C.

Procedure

The samples (1 ml) were subjected to a solid-phase extrac-tion with Bond Elut Certify columns (Fernández et al.,1996), using proadifen as the reference compound. Thedry residue was reconstituted in 40 µl of derivatizing agent,BSTFA containing 1% TMCS, shaken and maintained inan oven at 100 °C for 25 min to form trimethylsilylderiva-tives of high stability. Immediately afterwards, 2 µl wereinjected into the chromatograph.

Once the chromatographic conditions were optimized,calibration curves were constructed in urine and bloodusing the ranges 0.4–4 and 0.1–2 µg ml−1, respectively. Theconcentrations of the initial solutions are modified afterthe extraction, because the post-extraction volume (40 µl)is 25 times less than the sample volume (1000 µL). Thus,the final solutions were 25 times more concentrated thanthe initial samples.

The method was applied to 142 urine and 42 bloodsamples from 184 living cocaine users distributed into threegroups, namely: individuals in the criminal justice systemfor the purpose of confirming their drug addict condition(Group 1); drug abusers admitted to emergency roomsbecause of a neurological or psychiatric complication orsome traumatism resulting from a traffic accident (Group2); and patients under detoxification treatment to controltheir abstinence (Group 3).

The method also was applied to 39 urine and 62 bloodsamples from 65 fatally poisoned individuals classified intotwo groups, namely: 25 deaths due to cocaine alone, whereonly cocaine or cocaine + benzodiazepines and/or canna-bis were detected — the latter at too low concentrations tocause death (Group A); and 40 deaths related to cocaine+ opiates, alcohol and/or methadone (Group B).

Samples were frozen at −20 °C and preserved with 1%sodium fluoride prior to testing. The high concentrationsfound in some samples (mainly cocaine and BEG in urineand BEG in blood) required multiple dilutions in order toaccommodate them within the range of the calibrationgraphs. Thus, real concentrations were affected by dilu-tion prior to extraction

Creal = Ccal × Vpostext/Vinit

where Creal is the real concentration, Ccal is that calculatedfrom the calibration graph, Vpostext is the post-extractionvolume with which dry extracts were reconstituted (40 µlof the derivatizing mixture) and Vinit is the initial urine orblood volume.

RESULTS AND DISCUSSION

The retention times were used to identify EME (5.1 min),cocaine (10.1 min), BEG (11.1 min) and proadifen(11.8 min). Calibration curves were obtained by least-squares linear regression analysis, resulting in the follow-ing equations

COCA in urine y = 0.0238x + 0.0437 R = 0.9987COCA in blood y = 0.0244x + 0.0083 R = 0.9966BEG in urine y = 0.0277x + 0.0256 R = 0.9989BEG in blood y = 0.0327x + 0.0058 R = 0.9971EME in urine y = 0.0158x + 0.0282 R = 0.9980EME in blood y = 0.0173x + 0.0036 R = 0.9955

There was a good linearity at the intervals studied for eachdrug, according to the criteria of Bressolle et al. (1996).

The recoveries were assessed at two levels of concen-tration, with ten replicates of each level. The values oscil-lated between 68% and 83% in urine or between 72% and84% in blood.

The sensitivity was studied by means of limits of detec-tion (LOD) and quantitation (LOQ) (Armbruster et al.,1994). The LODs ranged from 0.044 to 0.070 µg ml−1, andthe LOQs from 0.15 to 0.23 µg ml−1.

The precision and accuracy were calculated, with goodresults (Bressolle et al., 1996). The precision, expressed asthe coefficient of variation for each level of concentration,was <10%. The accuracy, expressed as the relative errorobtained from the theoretical concentration of the sam-ples tested, was in the range ±12%.

Tables 1 and 2 show the concentration ranges of cocaineand its metabolites in 142 urine and 42 blood samples from184 living cocaine users, and the number of positive casesfor each compound. Complementary information such asthe presence of other drugs is also included. Males (148cases) prevailed over females (36 cases), which was tobe expected because most of the drug addicts in Galicia(Northwestern Spain) at the time of this research weremales (Navarro and Sánchez, 2002).

With respect to the concentration distribution of eachanalyte, 78% of the cases in Group 1 exhibited urine con-centrations of EME and COCA below the respectiveaverage values; also, 80% of the cases in Group 2 hadBEG and COCA levels lower than the average values inurine. On the other hand, 28 of the 36 individuals in Group3 had urine concentrations of EME, COCA and BEGbelow the respective averages.

A comparison of the results for the three groups revealsthat the urine levels of the three compounds were highestin the drug abusers admitted to emergency rooms (Group2). The ratios obtained from the average values confirmthat BEG concentrations were consistently several timeshigher than the COCA and EME concentrations: thus, theBEG/COCA ratio was 10.3 for Group 1, 5.4 for Group 2



and cannabis. One individual was found to have taken fivedifferent drugs.

The three compounds were found to occur simultane-ously in 57 urine samples (15 from Group 1, 25 from Group2 and 17 from Group 3). Fifteen urine samples (12 fromGroup 1 and 3 from Group 3) contained BEG + EME,whereas 44 urine samples (22 from Group 1, 10 from Group2 and 12 from Group 3) contained BEG + COCA. Nosample was found to contain COCA + EME only becauseBEG was detected in all studied individuals; this confirmsthat the presence of BEG is the best marker for previoususe of cocaine (Logan and Peterson, 1994). Twenty-sixurine samples were found to contain BEG alone; accord-ing to some authors, BEG remains in urine for longer thanEME and COCA do, so the originating episode may haveoccurred several days before the sample was collected(Ramcharitar et al., 1995).

Regarding the 42 blood samples studied, all from drugabusers admitted to emergency rooms, 23 (55%) had EMEconcentrations below the average value. This was also thecase with COCA in 31 samples (74%) and with BEG in28 samples (67%). The three compounds were foundsimultaneously in 31 individuals, BEG + EME in 1 andBEG + COCA in 10. As with the 142 urine samplesstudied, BEG was detected in all blood samples andalways at higher concentrations than COCA and EME;thus, the BEG/COCA ratio was 6.5 and the BEG/EMEratio was 8.9. In those cases where cocaine was the soledrug involved (29%), these ratios increased to 12.9 and42.3, respectively.

It has been reported that not only the route of cocaineadministration but also the concurrent use of additionaldrugs can have a significant impact on the pharmacolo-gical outcome (Oyler et al., 1996).

Tables 3 and 4 list the number of positive cases for eachcompound and the concentration ranges of cocaine and

Table 1—Results for the urine samples from 142 living patients

Group 1a Group 2b Group 3c

Gender Female 12 cases 7 cases 8 casesMale 58 cases 29 cases 28 cases

Positive cases EME 27 cases 25 cases 20 casesCOCA 37 cases 35 cases 29 casesBEG 70 cases 36 cases 36 cases

Average levels EME 0.05 ± 0.09 0.39 ± 0.57 0.21 ± 0.62(Ranges) (0–0.51 ) (0–2.71) (0–3.64)(µµµµµg ml−−−−−1) COCA 0.06 ± 0.15 0.97 ± 2.02 0.17 ± 0.28

(0–0.95) (0–10.17) (0–1.20)BEG 0.62 ± 0.38 5.23 ± 7.08 2.46 ± 4.54

(0.03–1.7) (0.4–32.47) (0.07–26.01)

Relationships BEG > COCA 68 cases 33 cases 35 casesbetween BEG > EME 68 cases 34 cases 36 casescompounds COCA > EME 28 cases 18 cases 19 cases

Other drugs Opiates 63 cases 30 cases 31 casesMethadone 5 cases 4 cases 8 casesBenzodiazepines 7 cases 1 case 4 casesCannabis 2 cases 1 case 1 caseNone 6 cases 4 cases 3 cases

a Group 1: individuals in the criminal justice system.b

Group 2: drug abusers admitted to emergency rooms.c Group 3: patients under detoxification treatment.

Table 2—Results for the blood samples from 42 living drugabusers admitted to emergency rooms

Gender Female 9 casesMale 33 cases

Positive cases EME 32 casesCOCA 41 casesBEG 42 cases

Average levels EME 0.16 ± 0.34(Ranges) (0–1.46)(µµµµµg ml−−−−−1) COCA 0.22 ± 0.28

(0–1.14)BEG 1.43 ± 1.62

(0.11–7.60)

Relationships BEG > COCA 41 casesbetween BEG > EME 41 casescompounds COCA > EME 31 cases

Other drugs Opiates 23 casesAlcohol 12 casesBenzodiazepines 3 casesCannabis 2 casesNone 10 cases

and 14.5 for Group 3; and the BEG/EME ratio was 12.4for Group 1, 13.4 for Group 2 and 11.7 for Group 3. Byexception, in six urine samples the BEG levels were lowerthan the COCA levels. This may have been the result of ashorter distance in time from the last dose of cocaine(Ramcharitar et al., 1995). In the cases where cocaine wasthe sole drug involved, the BEG/COCA and BEG/EMEratios were threefold and twofold higher, respectively.Our series exhibited polydrug consumption in many indi-viduals (88%), with a prevalence of the simultaneoususe of heroin followed by methadone, benzodiazepines

286 P. FERNÁNDEZ ET AL.


Table 3—Results for 25 deaths due to cocaine alone (Group A)

Blood Urine(24 samples) (18 samples)

Positive cases EME 17 cases 17 casesCOCA 23 cases 18 casesBEG 24 cases 18 cases

Average levels EME 0.38 ± 0.78 1.19 ± 1.64(ranges) (0–2.12) (0–5.32)(µµµµµg ml−−−−−1) COCA 0.40 ± 0.43 3.98 ± 6.83

(0–1.15) (0.11–18.38)BEG 2.38 ± 1.27 13.32 ± 11.51

(0.47–4.47) (3.00–35.50)

Relationships BEG > COCA 24 cases 18 casesbetween BEG > EME 24 cases 18 casescompounds COCA > EME 23 cases 18 cases

Table 4—Results for 40 deaths due to cocaine + opiates and/orother drugs (Group B)

Blood Urine(38 samples) (21 samples)

Positive cases EME 17 cases 7 casesCOCA 29 cases 20 casesBEG 38 cases 21 cases

Average levels EME 0.10 ± 0.13 0.88 ± 1.59(ranges) (0–0.38) (0–4.40)(µµµµµg ml−−−−−1) COCA 0.17 ± 0.33 3.88 ± 7.30

(0–1.73) (0–33.24)BEG 1.20 ± 1.05 10.85 ± 12.18

(0.04–4.26) (0.27–42.58)

Relationships BEG > COCA 38 cases 19 casesbetween BEG > EME 38 cases 21 casescompounds COCA > EME 29 cases 18 cases

COCA, which in turn was usually higher than that of EME,both in urine and in blood. These two cases correspondedto quick deaths in which the cocaine probably had no timeto metabolize; opiates and alcohol were also involved andthe latter is known to inhibit the hepatic metabolism ofcocaine and increase its plasmatic levels, which are raisedabove those of BEG as a result (Farré et al., 1993).

Most of the individuals exhibited higher COCA, BEGand EME levels in urine than in blood, with a urine/bloodaverage concentration ratio of 10 for COCA, 6 for BEGand 3 for EME in Group A and 23 for COCA, 9 for BEGand 9 for EME in Group B. The only exceptions were twocases corresponding to violent, quick deaths (suffocationby submersion) associated with high doses of cocaine,heroin and alcohol.

The blood concentrations obtained were higher in GroupA, where deaths were related to cocaine alone. The valuesin our series are consistent with those found previouslyby some authors following intranasal and/or intravenousadministration (Wetli and Fishbain, 1985) but lower thanthose obtained after swallowing a package containingcocaine (Fineschi et al., 2002).

We detected BEG in all urine post-mortem samples, atlevels higher than those of EME. This can be ascribed tococaine being released to a greater extent as BEG than asEME in urine and to the elimination half-life of BEGbeing longer than that of EME.

Correct interpretation of the blood and urine post-mortem levels of cocaine found in overdose cases requiresthe individual study of each one and the knowledge ofadditional data such as dose, administration route, timefrom administration to death and storage conditions ofthe samples. One must also consider differences amongindividuals regarding the presence of other substances andthe prior history of consumption, tolerance and metabolicalterations caused by changes in blood pH, among others.

its metabolites in urine and blood samples from 65 fatalintoxications, distributed into two groups, namely: 25deaths due to cocaine alone (Group A) and 40 deathsrelated to cocaine + opiates and/or other drugs (Group B).Males (85.5%) clearly prevailed over females (14.5%),which is consistent with the results of other authors whohave established that the probability of death for men ishigher than that for women because the former take drugsmore frequently (De la Fuente et al., 1995; Sánchez et al.,1995). In our series, the average age was 29.2 ± 5.5 years(range 20–44 years).

Group A exhibited low levels of benzodiazepines inseven cases and low levels of cannabis in four cases; theseindividuals presented increased blood and urine con-centrations of the drug and its metabolites. In Group B,opiates (38 cases) were found to prevail among the con-comitants, followed by alcohol (17 cases), methadone(8 cases), benzodiazepines (5 cases) and cannabis (4 cases).

Fifty-three individuals were drug addicts by parenteralroute and used the intravenous route (results taken from58 cases). HIV tests were positive in 59% of the 46 casesstudied. In the 56 cases where the origin of the corpse wasknown, the individual’s home and vicinity were the mostfrequent death sites (43 cases), followed by hospital(9 cases) and hotel rooms (4 cases). We can thus concludethat the average profile of the dead individual in our serieswas a 29-year-old male, who was a habitual consumer ofcocaine and heroin by the intravenous route, HIV-positiveand died in his own home or nearby.

The ratios obtained from the average urine concentra-tions were 3.4 and 3.9 for BEG/COCA in Groups A andB, respectively, and 11.2 and 17.1 for BEG/EME in GroupsA and B, respectively. On the other hand, the ratiosobtained from the average blood concentrations were 6.0(Group A) and 7.1 (Group B) for BEG/COCA and 6.3(Group A) and 12.0 (Group B) for BEG/EME. Except fortwo cases, the concentration of BEG exceeded that of

REFERENCES

Armbruster DA, Tillman MD, Hubbs LM. 1994. Limit ofdetection (LOD)/limit of quantitation (LOQ): comparisonof the empirical and the statistical methods exemplifiedwith GC–MS assays of abused drugs. Clin. Chem. 40: 1233–1238.

Bressolle F, Bromet-Petit M, Audran M. 1996. Validation of liquidchromatographic and gas chromatographic methods. Appli-cations to pharmacokinetics. J. Chromatogr. B 686: 3–10.

Corburt MR, Koves EM. 1994. Gas chromatography/mass spectro-metry for the determination of cocaine and benzoylecgonine



over a wide concentration range (<0.005–5 mg/dl) in post-mortem blood. J. Forens. Sci. 39: 136–149.

De Giovanni N, Rossi SS. 1994. Simultaneous detection ofcocaine and heroin metabolites in urine by solid-phaseextraction and gas chromatography/mass spectrometry. J.Chromatogr. B 658: 69–73.

De la Fuente L, Barrio G, Vicente J, Bravo MJ, Santacreu J. 1995.The impact of drug-related deaths on mortality amongyoung adults in Madrid. Am. J. Public Health 85: 102–105.

Farré M, De la Torre M, Llorente M, Lamas X, Ugena B, SeguraJ, Cami J. 1993. Alcohol and cocaine interactions in humans.J. Pharmacol. Exp. Ther. 266: 1364–1373.

Fernández P, Lafuente N, Bermejo AM, López-Rivadulla M,Cruz A. 1996. HPLC determination of cocaine and benzoylec-gonine in plasma and urine from drug abusers. J. Anal.Toxicol. 20: 224–228.

Fineschi V, Centini F, Monciotti F, Turillazzi E. 2002. The cocaine‘body-stuffer’ syndrome: a fatal case. Forens. Sci. Int. 126:7–10.

Hippensteil MJ, Gerson B. 1994. Optimization of storage con-ditions for cocaine and benzoylecgonine in urine. J. Anal.Toxicol. 18: 104–109.

Isenschmid DS, Fischman MW, Foltin RW, Caplan YH. 1992.Concentration of cocaine and metabolites in plasma ofhumans following intravenous administration and smokingof cocaine. J. Anal. Toxicol. 16: 311–314.

Karch SB. 2002. Karch’s Pathology of Drug Abuse. CRC Press:Boca Raton, FL.

Kidwell DA, Blanco MA, Smith FP. 1997. Cocaine detection in auniversity population by hair analysis and skin swab testing.Forens. Sci. Int. 84: 75–86.

Logan BK, Peterson KL. 1994. The origin and significance ofecgonine methyl ester in blood samples. J. Anal. Toxicol. 18:124–125.

Navarro J, Sánchez L. 2002. Observatorio de Galicia sobreDrogas. Informe General 2002. Xunta de Galicia: Santiago deCompostela.

Oyler J, Darwin WD, Preston KL, Suess P, Cone EJ. 1996. Cocainedisposition in meconium from newborns of cocaine abusingmothers and urine of adult drug users. J. Anal. Toxicol. 20:453–462.

Pan WJ, Hedaya MA. 1997. Sensitive and specific high-performance liquid chromatographic assay with ultravioletdetection for the determination of cocaine and its metabo-lites in rat plasma. J. Chromatogr. 703: 129–138.

Peterson KL, Logan BK, Christian GD. 1995. Detection of cocaineand its polar transformation products and metabolites inhuman urine. Forens. Sci. Int. 73: 183–196.

Preston KL, Huestis MA, Wong CJ, Umbricht A, Goldberger BA,Cone EJ. 1999. Monitoring cocaine use in substance-abuse-treatment patients by sweat and urine testing. J. Anal.Toxicol. 23: 313–322.

Ramcharitar V, Levine B, Smialek JE. 1995. Benzoylecgonine andecgonine methyl ester concentrations in urine specimens.J. Forens. Sci. 40: 99–101.

Sánchez J, Rodríguez B, De la Fuente L, Barrio G, Vicente J,Roca J, Royuela L. 1995. Opiates or cocaine: mortality fromacute reactions in six major Spanish cities. State InformationSystem on Drug Abuse (SEIT) Working Group. J. Epidemiol.Commun. Health 49: 54–60.

Schramm W, Craig PA, Smith RH, Berger GE. 1993. Cocaine andbenzoylecgonine in saliva, serum and urine. Clin. Chem. 39:481–487.

Smart R, Anglin R. 1986. Do we know the lethal dose of cocaine?J. Forens. Sci. 32: 303–312.

Wetli C, Fishbain D. 1985. Cocaine-induced psychosis andsudden death in recreational cocaine users. J. Forens. Sci.30: 873–888.

Documents

Gas chromatographic determination of cocaine and its metabolites in blood and urine from cocaine users in northwestern Spain