- Email: [email protected]
Automated fast procedure for the simultaneous extraction of hair sample performed with an automated workstation
Forensic Science International 218 (2012) 15–19
Contents lists available at SciVerse ScienceDirect
Forensic Science International journal homepage: www.elsevier.com/locate/forsciint
Automated fast procedure for the simultaneous extraction of hair sample performed with an automated workstation§ I. Angeli *, M. Minoli, A. Ravelli, F. Gigli, F. Lodi Section of Forensic Toxicology, Department of Human Physiology, University of Milan, Italy
A R T I C L E I N F O
A B S T R A C T
Article history: Received 1 April 2011 Accepted 1 July 2011 Available online 27 October 2011
Introduction: Hair testing has a leading role in toxicology practice and even more in those aspects tightly linked to the assessment of psychoactive drug use and abuse in social life. Aim: The objective of the present study was to develop and validate an automated SPE samplepreparation step, suited for GC/MS confirmation analysis of basic drugs in hair drug control. The method was studied and optimized for quantitative determination and in a second time it was extended to real hair samples. The purpose of method validation was to ensure good reliability, reproducibility and quickness. Methods: Janus Automated Workstation (PerkinElmer) was employed to perform SPE hair extraction, using 96-well plate SPEC MP1 acquired from Varian (Agilent Technologies). After derivatization of dried extracts, screening confirmations were performed using gas chromatography (GC) followed by mass spectrometry (MS). GC/MS data were validated following standard guidelines, but our attention was focused on three headings: samples cross-contamination, ‘‘memory effect’’ and extraction recovery. Results: Validation requests were fully accomplished and we always obtained best results with the automated procedure. For instance, analytes mean recovery was between 70 and 90% and data analysis proved that no contamination between samples occurred. Conclusions: The automated workstation has shown good reliability (cross contamination and ‘‘memory effect’’ were tested and excluded), effectiveness (no false negative was detected), solvent saving (500 mL/ sample vs traditionally LLE 4 mL/sample) and quickness (50 min for 96 tests cycle). ß 2011 Elsevier Ireland Ltd. All rights reserved.
Keywords: Drugs Hair Automated SPE
1. Introduction Hair testing currently has a leading rule in toxicology practice and in the last years more procedures were discussed and validated for the assessment of psychoactive drug use and abuse in social life [1–5]. The promulgation of new laws on workers’ health and the assessment of drug abuse in workplace have led to a significant increase in laboratory workload. Moreover, Ser.T (public detoxification centers) and driving license committees have applied new drugs cut-off. For these reasons, it is necessary to consider the possibility of changing the method in use in our laboratory, using automated SPE sample preparation instead of LLE. Drugs of abuse determination in hair and SPE procedures are well documented; the SPE extract purity and the advantages of automating sample preparation guarantee high throughput, improved reproducibility and robustness [6–10]. Janus Varispan 8 channels Automated
§ This paper is part of the special issue entitled: Selected papers from the Chamonix 2011 Society of Hair Testing Meeting, Guest-edited by Pascal Kintz. * Corresponding author. E-mail address: [email protected] (I. Angeli).
0379-0738/$ – see front matter ß 2011 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2011.10.008
Workstation (Perkin Elmer) met many of our goals and was employed to perform hair extraction, using the SPEC MP1 96well plate acquired from Varian (Agilent Technologies). The method applied was studied and optimized for quantitative determination of basic drugs in hair matrix, so as to ensure reproducibility, increase analytical sensibility and shorten sample preparation. During method development, our attention was focused on three headings: samples cross-contamination, ‘‘memory effect’’ and extraction recovery. 2. Materials and methods 2.1. Working solutions and chemicals All the working solutions used in the present study were hand-prepared from following certified standard solutions acquired from Cerilliant (LCG Standard, Milan, Italy). Cocaine and 6-monoacetylmorphine were 0.1 mg/mL in acetonitrile. Benzoylecgonine, morphine, codeine, methadone, amphetamine, methamphetamine, MDA and MDMA, were 0.1 mg/mL in methanol. Deuterated Internal and External Standards (IS and ES) were morphine-d3, benzoylecgonine-d3, amphetamine-d5 and MDMA-d5, 0.1 mg/mL in methanol and cocaine-d3 0.1 mg/mL in acetonitrile. All these solutions were stored at 20 8C. Derivatizing agents used were pentafluoropropionic anhydride (PFPA) and hexafluoroisopropanol (HFIP), acquired from Sigma Aldrich S.r.l. (Milan, Italy).
16
I. Angeli et al. / Forensic Science International 218 (2012) 15–19
Solvents used were trichloromethane, n-heptane, methylene chloride, methanol, isopropanol and ethyl acetate (Panreac Quimica Sa, Barcelona, Spain). Hydrochloric acid and ammonia hydroxide were purchased from JT Baker (Deventer, Netherlands). 2.2. Instrumentation – JANUS liquid handling workstation/SPE cartridge – SPEC MP 1 The automated workstation for liquid handling chosen was JANUS Integrator with Varispan arm 8 channels needles VersaTip Plus with Gripper Arm, equipped with 500 mL syringes and Vacuum System Option. The JANUS liquid handler is equipped with VersaTip technology that allows to equip each needle with a disposables tip or to use it as a fixed tip. The JANUS dispensing technology is based on liquid displacement. This means that all aspiration and distribution steps are done by moving internal system liquid columns. WinPREP-JANUS software coordinates all liquid handling steps, vacuum steps and gripper movements used in the automated sample preparation. To obtain best results, instrument settings (volumes, vacuum grade and arm movements) can be modified as preferred. The 96-well plate SPEC MP1 was acquired from Varian (Agilent Technologies) and was used to perform sample extraction. The SPEC MP1 technology, that allows charging the acidic hydrolyzed hair solution directly, was chosen so as to avoid pH neutralizing step. 2.3. Samples preparation Blank hair were washed for 2 min with 2 mL of methylene chloride (3 cycles). Dried hair were cut into small pieces. One hundred samples were constituted by weighting 50 mg of hair in glass test tubes. Samples were then hydrolyzed overnight at 45 8C in 2 mL of 0.1 M HCl. Acidic aqueous solution was then transferred into 12 mm 75 mm test tubes and placed in the Janus holder plate for the execution of the automated and programmed extraction steps. 2.4. Calibration points Automated protocol validation and analysis were performed on the construction of 24 calibration curves. Calibration levels were 0–0.05–0.1–0.2–0.5–1–5 ng/mg and IS levels were 100 ng for cocaine-d3, benzoylecgonine-d3, MDMA-d5, amphetamine-d5 and 50 ng for morphine-d3. 2.5. Automated SPE procedure For the validation procedure Janus system was programmed to extract samples through different steps. Standards, IS and samples were loaded completely in automation. Elution was accomplished by passing and collecting 700 mL of methylene chloride:isopropanol:ammonia hydroxide (77:20:3, v/v). Disposable tips steps (1000 mL or 200 mL volumes) were executed after tips conditioning (five cycles), so as to avoid solution dropping out of well container, well plate or test tubes. 2.6. Derivatization Samples were dried under a steam of N2 and derivatized with PFPA (50 mL)–HFIP (30 mL) for 20 min at 70 8C in sealed vials. PFPA–HFP in excess was eliminated under a steam of nitrogen. Dried extract was dissolved in 50 mL of ethyl acetate.
In order to ensure a linear correlation between peak area and analyte concentration, squared correlation coefficient (r2) was calculated for each curve. 2.8.2. Sensitivity/specificity Sensibility was expressed in terms of LOD and LOQ. They were calculated by injecting decreasing concentrations of the drug spiked in drug-free hair, until a response, that met peak shape, resolution, acceptable retention time and qualifier ion ratio within 20% of the average of six calibrators, was obtained. The LOQ is the lowest measurement that met LOD criteria and had a %RSD lower or equal to 20% in six replicates. In order to detect any possible interferences from hair matrix, cartridge and/ or solvents, blank (no analytes or IS) samples were extracted with the automated method, derivatized and analyzed in GC/MS. 2.8.3. Precision and accuracy Intra-assay accuracy and precision were assessed over the linear range. Six blank hair samples were spiked with low, medium, and high analytes concentration (0.1– 1–5 ng/mg) and extracted according to the automated method. Accuracy was calculated as percentage deviation of the mean measured concentration of six extracts from the theoretical spiked concentration. The intra-assay precision was expressed as a percent relative standard deviation (%RSD) calculated using the six individual values. Interday-assay for accuracy and precision was assessed with the same method and was applied to 24 replicates of three levels of calibration curve, extracted in three different days through the complete automated procedure. 2.8.4. SPE vs LLE recovery The two preparative techniques were compared in order to evaluate the difference in extraction recovery efficiency. With the purpose of eliminating the influence of the derivatization step in the recovery evaluation, ES were added after extraction. The extraction recovery was calculated dividing the mean peak-area ratio of target ion/internal standard for the samples in which the standard solution was added before extraction, with the mean peak-area ratio of target ion/internal standard for the samples in which the standard solution was added after extraction. 2.8.5. Cross contamination and memory effect Cross contamination was investigated at two different levels. To investigate upper cross contamination, which may occur during dispensation processes, 18 blank samples were spiked with high levels of analytes (1 mg/mL) and were extracted as follows. A 96-well plate was prepared so as to ensure that the 18 high positive samples were surrounded by negative samples. In each well internal standards were added and all samples were analyzed in GC/MS. To avoid lower cross contamination that may occur during elution, we used the Clean Thru Tip (CTT) technology, which eliminates sample carry over from the vacuum manifold. This technology provides SPE columns with a disposable tip that attaches to the end of each column, eliminates any contact between the sample, wash solvent and the extraction apparatus. To simplify extraction method, lower cross contamination was tested even without the CTT devices. To ensure the absence of instrument ‘‘analyte memory’’, directly prior to and after extraction of the previous 18 highly positive samples, a plate with 96 blanks was extracted with the same automated method and analyzed in GC/MS. Blank samples were compared and quantified with the calibration curves built for the method validation to ensure the absence of each drug in negative samples.
2.7. Gas chromatography–mass spectrometry method Analysis was performed using a GC/MS. The instrument configuration used consisted of an Agilent 6890N Gas Chromatograph coupled with an Agilent 5973 Mass Selective Detector. GC column was CP-SIL 8–15 m (ID 0.25 mm to 0.25 mm film thickness), carrier gas was helium (1 mL/min). Injection was executed in splitless mode, using a sample volume of 1 mL. Injector temperature was 250 8C. Oven initial temperature was 70 8C. First ramp was 20 8C/min up to 200 8C, second ramp was 6 8C/min up to 220 8C, and third ramp was 30 8C/min up to 290 8C (hold for 2 min). Transfer line temperature was 280 8C. Ions were monitored as follows (underscored ion was chosen as target for quantitative determination): amphetamine (190–118–91 m/z), amphetamine-d5 (194–123–92 m/z), methamphetamine (204–160–91 m/z), MDA (325–135–162 m/ z), MDMA (204–162–135 m/z), MDMA-d5 (208–163–136 m/z), benzoylecgonine (439–318), benzoylecgonine-d3 (442–321), cocaine (303–272–82), cocaine-d3 (306–275), morphine (414–577 m/z), morphine-d3 (417–580 m/z), codeine (445– 282 m/z), 6-monoacetylmorphine (414–473 m/z), and methadone (294–72 m/z). 2.8. Method validation The automated method was validated in combination with GC/MS analysis. Data analysis was executed checking standard validation parameters. 2.8.1. Linearity Calibration curves were built using a linear regression (y = mx + b) between the interested range of 0.05–5 ng/mg.
2.8.6. Real samples analysis Finally, this automated procedure was applied for 30 real cases and two saliva quality controls (sent by ISS, Roma). All these samples were processed with our LLE and with the automated SPE. We divided the acidic hair solution and the prebuffered saliva (pH 4 – 1:1) in two aliquots of 1 mL each for hair solution and of 200 mL each for saliva. The first part was neutralized and then extracted at pH 9 with 4 mL of a mixture of trichloromethane:n-heptane:isopropanol (50:33:17, v/v). The second part was processed with the SPE automated method. Both types of extract were then dried, derivatized and analyzed in GC/MS.
3. Results and discussion 3.1. Method validation Tables 1 and 2 summarize all the results obtained in the validation procedure. The r2 values ranged from 0.996 to 0.999 (Table 1). This correlation demonstrates the capacity of the instrument to obtain fine calibration curves, minimizing the variability introduced by the operator. Sensitivity parameters were satisfying for many analytes (LOQs ranged from 0.038 to 0.046 ng/mg). However amphetamines and MDMA calculated LOQs impose to build calibration curve with
I. Angeli et al. / Forensic Science International 218 (2012) 15–19 Table 1 Results of linearity test.
Cocaine Benzoylecgonine Morphine Codeine 6-MAM Methadone Amphetamine Methamphetamine MDMA MDA
Target ion (m/z)
Equation
r2
303 439 414 445 414 294 190 204 204 162
y = 0.02196x 0.00195 y = 0.05025x + 0.00611 y = 0.01221x 0.003336 y = 0.007694x 0.00343 y = 0.01553x 0.05677 y = 0.002155x 0.00147 y = 0.010220x 0.0191 y = 0.000958x 0.00404 y = 0.000828x 0.003429 y = 0.004011x + 0.005012
0.999 0.999 0.999 0.996 0.998 0.998 0.998 0.997 0.999 0.998
higher starting point (0.1 ng/mg instead of 0.05 ng/mg) and suggest the necessity to use at least 20 mg of hair to reach the workplace cutoff value of 0.2 ng/mg. Furthermore, regarding to specificity, endogenous interferents that may be present from the matrix or from the components of the SPE column, from solvents or buffers did not produce any peaks that interfere with our analysis. The %RSDs for precision were always between 13.4% and 5.5%. Best accuracy values were obtained for amphetamine ( 1.2%) and for benzoylecgonine (+4.2%). 3.2. Cross contamination and memory effect After the execution of the previous described procedure with samples spiked with high levels of analytes (1 mg/mL), they were analyzed in GC/MS and often resulted totally negative. However in some cases, cross contamination occurred. In these conditions, analytes (always cocaine and benzoylecgonine) were present, but their concentration was always under the lower calibration point of 0.05 ng/mg. In real hair, concentration of 20 ng/mg are rare but possible, for instance in the profile of a cocaine abuser. In the face of
17
this sporadic event, it is necessary to consider the decisional cutoff for positivity or always keep in mind that some rare contamination between samples may occur. We suggest performing hair analysis always coupled with a semi-quantitative screening procedure, so as to dilute very high positive samples before SPE. 3.3. LLE vs SPE recovery As can be seen in Table 3, we obtained evidences of best recovery extraction with SPE rather than LLE. 3.4. Real samples analysis The quantitative results obtained were in good agreement with those obtained by the conventional method and never differ than 10% RSD. Moreover no false positive or negative was detected. Despite those obtained with the SPE procedure, in two cases with low cocaine positivity, the determination of benzoylecgonine in the liquid/liquid extract was negative, due to the bad recovery rate of the substance with this technique. As to saliva quality controls, our results perfectly match both the target compounds and their concentrations. In the first sample both cocaine (19.7 ng/mL) and benzoylecgonine (18.1 ng/mL) were identified, whereas the second one was negative. In Fig. 1 it is shown the overlaid chromatograms of target ions between LLE and SPE for cocaine (COC), morphine (MOR) and benzoylecgonine (BEG) in a real testing hair (COC: 1.83 ng/ mg – MOR: 0.88 ng/mg – BEG: 0.64 ng/mg). 3.5. Extract purity Samples were extracted both with SPE and LLE method, so as to verify extract purity in GC/MS. For instance, in Fig. 2 it is represented one of the real sample (MDMA – 1.74 ng/mg), in which
Table 2 Results of validation parameters. Method performance parameters determined with spiked hair samples by GC/MS Cocaine
Benzoylecgonine
303 439 Target ion (m/z) Retention time (min) 11.4 9.6 LOD (ng/mg) 0.012 0.014 LOQ (ng/mg) 0.038 0.042 Intra-day reproducibility (%) (3 different levels) Low level (0.1 ng/mg) 7.5 8.6 Medium level (1 ng/mg) 6.4 5.5 High level (5 ng/mg) 6.9 7.9 Inter-day reproducibility (%) (3 different levels) Low level (0.1 ng/mg) 8.1 9.0 Medium level (1 ng/mg) 7.1 6.0 High level (5 ng/mg) 8.0 8.5 Accuracy (BIAS %) (3 different levels) Low level (0.1 ng/mg) 5.8 6.5 Medium level (1 ng/mg) 10.2 4.2 High level (5 ng/mg) 9.3 6.7
Morphine
Codeine
6-MAM
Methadone
Amphetamine
Methamphetamine
MDMA
MDA
414 11.8 0.016 0.046
445 12.2 0.016 0.045
414 12.6 0.015 0.043
294 10.8 0.016 0.042
190 5.3 0.0350 0.057
204 6.1 0.037 0.056
204 8.0 0.035 0.096
162 7.4 0.031 0.083
8.5 5.7 8.1
10.8 7.4 9.3
8.4 6.0 9.5
8.9 7.3 8.1
10.7 7.6 9.3
12.1 8.4 10.5
11.3 8.6 10.1
11.5 9.7 10.6
9.0 6.5 8.4
11.4 7.9 10.1
8.9 7.0 10.6
10.4 9.1 9.9
12.7 8.9 11.4
13.4 9.5 11.0
13.5 9.8 10.9
12.5 10.5 11.5
8.5 5.4 19.1
8.6 6.1 9.1
9.1 5.4 8.5
6.2 8.5 4.5
8.5 7.7 1.2
9.8 9.1 2.6
9.6 8.5 5.5
9.8 8.7 3.1
Table 3 Recovery percentages (SPE and LLE).
SPE Low level (0.1 ng/mg) Medium level (1 ng/mg) High level (5 ng/mg) LLE Low level (0.1 ng/mg) Medium level (1 ng/mg) High level (5 ng/mg)
Cocaine
Benzoylecgonine
Morphine
Codeine
6-MAM
Methadone
Amphetamine
Methamphetamine
MDMA
MDA
74 83 80
95 101 102
73 93 76
84 83 89
69 70 73
75 90 94
81 86 74
61 61 60
73 76 83
70 92 90
61 79 46
24 32 21
66 72 78
41 52 48
43 47 47
44 53 53
38 65 64
24 36 19
33 59 58
38 62 63
18
I. Angeli et al. / Forensic Science International 218 (2012) 15–19
Fig. 1. Overlaid chromatograms of cocaine, morphine and benzoylecgonine, respectively in a real case.
Figs. 2–4. Examples of extract purity.
the ion 204 m/z (MDMA) is overlaid. In Figs. 3 and 4 it is shown the TIC of a spiked sample positive for all analytes obtained in SPE and LLE, respectively. GC/MS analysis has demonstrated the great extract purity obtained with the SPE performed with the SPEC MP1 cartridge.
4. Conclusions The usefulness of automated procedures has often been discussed in many trained works. The advantages of SPE automation are well-known and include high throughput, improved precision and robustness. The JANUS Integrator with Varispan arm 8 channels needles – VersaTipPlus, Gripper Arm & Vacuum Filtration Option was used to perform a SPE extraction from human hair to determine cocaine, morphine, amphetamines, MDMA, methadone and their most representative metabolites by GC/MS. In addition, accurate and steady liquid handling ensures good reproducibility. The capacity of the instrument to quickly build a calibration curve makes the validation of every batch results very simple. The automated workstation has shown good reliability (cross contamination and ‘‘memory effect’’ were tested and with some accuracy excluded), effectiveness (no false positive was detected), solvent saving (500 mL/sample vs traditionally LLE 4 mL/sample) and quickness (50 min for 96 tests cycle). The introduction of a 96-well plate simplifies the procedure. The SPEC MP1 resulted suitable for the automation and it lets the
use of the acidic hair extract and the occurrence of the cartridge dry up. So the method has fully satisfied required validation parameters such as extract purity, demonstrated by the reached sensibility, and mean recovery rates (70–90% for all the substances of interest). Finally we have compared different analysis of real hair samples, after ELISA screening for amphetamines, cocaine, opiates and methadone, whose preparation was achieved with our routine LLE or with the developed and validated method proposed. We always obtained best sample purification and an increased analyte recovery (this aspect is particularly enhanced with benzoylecgonine). The procedure was also useful to analyze saliva samples with a simple method and optimal result. References [1] R.L. DuPont, W.A. Baumgartner, Drug testing by urine and hair analysis: complementary features and scientific issues, Forensic Sci. Int. 70 (1995) 63–76. [2] P. Kintz, P. Mangin, Simultaneous determination of opiates, cocaine and major metabolites of cocaine in human hair by gas chromatography/mass spectrometry (GC/MS), Forensic Sci. Int. 73 (1995) 93–100. [3] M. Barroso, M. Dias, D.N. Vieira, J.A. Queiroz, M. Lopez-Rivadulla, Development and validation of an analytical method for the simultaneous determination of cocaine and its main metabolite, benzoylecgonine, in human hair by gas chromatography/mass spectrometry, Rapid Commun. Mass Spectrom. 22 (2008) 3320–3326. [4] F. Musshoff, B. Madea, New trends in hair analysis and scientific demands on validation and technical notes, Forensic Sci. Int. 165 (2007) 204–215. [5] M.J. Telepehak, T.F. August, G. Chaney. Forensic and Clinical Applications of Solid Phase Extraction, Humana Press, 2004.
I. Angeli et al. / Forensic Science International 218 (2012) 15–19 [6] P. Zimmerli, Ch. Staub, Analyse de drogues dans les cheveux par extraction automatisee en phase solide, Rev. Fr. Labor. 282 (1996) 51–55. [7] Ch. Girod, F. De Dominicis, R. Giovannini, Ch. Staub, Analyse de drogues dans les cheveux par extraction automatise’e en phase solide: validation et utilisation dans la routine, Toxicorama 1 (1999) 46–50. [8] C. Girod, C. Staub, Analysis of drugs of abuse in hair by automated solid-phase extraction, GC/EI/MS and GC ion trap/CI/MS, Forensic Sci. Int. 107 (2000) 261–271.
19
[9] T. Sibum, M. Kohal, E. Koster, K. Resch, V. Kaever, Direct and fast determination of antiretroviral drugs by automated online solid phase extraction-liquid chromatography–tandem mass spectrometry in human plasma, Clin. Chem. Lab. Med. 44 (2006) 299–305. [10] F.X. Diamond, W.E. Vickery, J. de Kanel, Extraction of benzoylecgonine (cocaine metabolite) and opiates (codeine and morphine) from urine samples using the Zymark RapidTraceTM, J. Anal. Toxicol. 20 (1996) 587–591.
Contents lists available at SciVerse ScienceDirect
Forensic Science International journal homepage: www.elsevier.com/locate/forsciint
Automated fast procedure for the simultaneous extraction of hair sample performed with an automated workstation§ I. Angeli *, M. Minoli, A. Ravelli, F. Gigli, F. Lodi Section of Forensic Toxicology, Department of Human Physiology, University of Milan, Italy
A R T I C L E I N F O
A B S T R A C T
Article history: Received 1 April 2011 Accepted 1 July 2011 Available online 27 October 2011
Introduction: Hair testing has a leading role in toxicology practice and even more in those aspects tightly linked to the assessment of psychoactive drug use and abuse in social life. Aim: The objective of the present study was to develop and validate an automated SPE samplepreparation step, suited for GC/MS confirmation analysis of basic drugs in hair drug control. The method was studied and optimized for quantitative determination and in a second time it was extended to real hair samples. The purpose of method validation was to ensure good reliability, reproducibility and quickness. Methods: Janus Automated Workstation (PerkinElmer) was employed to perform SPE hair extraction, using 96-well plate SPEC MP1 acquired from Varian (Agilent Technologies). After derivatization of dried extracts, screening confirmations were performed using gas chromatography (GC) followed by mass spectrometry (MS). GC/MS data were validated following standard guidelines, but our attention was focused on three headings: samples cross-contamination, ‘‘memory effect’’ and extraction recovery. Results: Validation requests were fully accomplished and we always obtained best results with the automated procedure. For instance, analytes mean recovery was between 70 and 90% and data analysis proved that no contamination between samples occurred. Conclusions: The automated workstation has shown good reliability (cross contamination and ‘‘memory effect’’ were tested and excluded), effectiveness (no false negative was detected), solvent saving (500 mL/ sample vs traditionally LLE 4 mL/sample) and quickness (50 min for 96 tests cycle). ß 2011 Elsevier Ireland Ltd. All rights reserved.
Keywords: Drugs Hair Automated SPE
1. Introduction Hair testing currently has a leading rule in toxicology practice and in the last years more procedures were discussed and validated for the assessment of psychoactive drug use and abuse in social life [1–5]. The promulgation of new laws on workers’ health and the assessment of drug abuse in workplace have led to a significant increase in laboratory workload. Moreover, Ser.T (public detoxification centers) and driving license committees have applied new drugs cut-off. For these reasons, it is necessary to consider the possibility of changing the method in use in our laboratory, using automated SPE sample preparation instead of LLE. Drugs of abuse determination in hair and SPE procedures are well documented; the SPE extract purity and the advantages of automating sample preparation guarantee high throughput, improved reproducibility and robustness [6–10]. Janus Varispan 8 channels Automated
§ This paper is part of the special issue entitled: Selected papers from the Chamonix 2011 Society of Hair Testing Meeting, Guest-edited by Pascal Kintz. * Corresponding author. E-mail address: [email protected] (I. Angeli).
0379-0738/$ – see front matter ß 2011 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2011.10.008
Workstation (Perkin Elmer) met many of our goals and was employed to perform hair extraction, using the SPEC MP1 96well plate acquired from Varian (Agilent Technologies). The method applied was studied and optimized for quantitative determination of basic drugs in hair matrix, so as to ensure reproducibility, increase analytical sensibility and shorten sample preparation. During method development, our attention was focused on three headings: samples cross-contamination, ‘‘memory effect’’ and extraction recovery. 2. Materials and methods 2.1. Working solutions and chemicals All the working solutions used in the present study were hand-prepared from following certified standard solutions acquired from Cerilliant (LCG Standard, Milan, Italy). Cocaine and 6-monoacetylmorphine were 0.1 mg/mL in acetonitrile. Benzoylecgonine, morphine, codeine, methadone, amphetamine, methamphetamine, MDA and MDMA, were 0.1 mg/mL in methanol. Deuterated Internal and External Standards (IS and ES) were morphine-d3, benzoylecgonine-d3, amphetamine-d5 and MDMA-d5, 0.1 mg/mL in methanol and cocaine-d3 0.1 mg/mL in acetonitrile. All these solutions were stored at 20 8C. Derivatizing agents used were pentafluoropropionic anhydride (PFPA) and hexafluoroisopropanol (HFIP), acquired from Sigma Aldrich S.r.l. (Milan, Italy).
16
I. Angeli et al. / Forensic Science International 218 (2012) 15–19
Solvents used were trichloromethane, n-heptane, methylene chloride, methanol, isopropanol and ethyl acetate (Panreac Quimica Sa, Barcelona, Spain). Hydrochloric acid and ammonia hydroxide were purchased from JT Baker (Deventer, Netherlands). 2.2. Instrumentation – JANUS liquid handling workstation/SPE cartridge – SPEC MP 1 The automated workstation for liquid handling chosen was JANUS Integrator with Varispan arm 8 channels needles VersaTip Plus with Gripper Arm, equipped with 500 mL syringes and Vacuum System Option. The JANUS liquid handler is equipped with VersaTip technology that allows to equip each needle with a disposables tip or to use it as a fixed tip. The JANUS dispensing technology is based on liquid displacement. This means that all aspiration and distribution steps are done by moving internal system liquid columns. WinPREP-JANUS software coordinates all liquid handling steps, vacuum steps and gripper movements used in the automated sample preparation. To obtain best results, instrument settings (volumes, vacuum grade and arm movements) can be modified as preferred. The 96-well plate SPEC MP1 was acquired from Varian (Agilent Technologies) and was used to perform sample extraction. The SPEC MP1 technology, that allows charging the acidic hydrolyzed hair solution directly, was chosen so as to avoid pH neutralizing step. 2.3. Samples preparation Blank hair were washed for 2 min with 2 mL of methylene chloride (3 cycles). Dried hair were cut into small pieces. One hundred samples were constituted by weighting 50 mg of hair in glass test tubes. Samples were then hydrolyzed overnight at 45 8C in 2 mL of 0.1 M HCl. Acidic aqueous solution was then transferred into 12 mm 75 mm test tubes and placed in the Janus holder plate for the execution of the automated and programmed extraction steps. 2.4. Calibration points Automated protocol validation and analysis were performed on the construction of 24 calibration curves. Calibration levels were 0–0.05–0.1–0.2–0.5–1–5 ng/mg and IS levels were 100 ng for cocaine-d3, benzoylecgonine-d3, MDMA-d5, amphetamine-d5 and 50 ng for morphine-d3. 2.5. Automated SPE procedure For the validation procedure Janus system was programmed to extract samples through different steps. Standards, IS and samples were loaded completely in automation. Elution was accomplished by passing and collecting 700 mL of methylene chloride:isopropanol:ammonia hydroxide (77:20:3, v/v). Disposable tips steps (1000 mL or 200 mL volumes) were executed after tips conditioning (five cycles), so as to avoid solution dropping out of well container, well plate or test tubes. 2.6. Derivatization Samples were dried under a steam of N2 and derivatized with PFPA (50 mL)–HFIP (30 mL) for 20 min at 70 8C in sealed vials. PFPA–HFP in excess was eliminated under a steam of nitrogen. Dried extract was dissolved in 50 mL of ethyl acetate.
In order to ensure a linear correlation between peak area and analyte concentration, squared correlation coefficient (r2) was calculated for each curve. 2.8.2. Sensitivity/specificity Sensibility was expressed in terms of LOD and LOQ. They were calculated by injecting decreasing concentrations of the drug spiked in drug-free hair, until a response, that met peak shape, resolution, acceptable retention time and qualifier ion ratio within 20% of the average of six calibrators, was obtained. The LOQ is the lowest measurement that met LOD criteria and had a %RSD lower or equal to 20% in six replicates. In order to detect any possible interferences from hair matrix, cartridge and/ or solvents, blank (no analytes or IS) samples were extracted with the automated method, derivatized and analyzed in GC/MS. 2.8.3. Precision and accuracy Intra-assay accuracy and precision were assessed over the linear range. Six blank hair samples were spiked with low, medium, and high analytes concentration (0.1– 1–5 ng/mg) and extracted according to the automated method. Accuracy was calculated as percentage deviation of the mean measured concentration of six extracts from the theoretical spiked concentration. The intra-assay precision was expressed as a percent relative standard deviation (%RSD) calculated using the six individual values. Interday-assay for accuracy and precision was assessed with the same method and was applied to 24 replicates of three levels of calibration curve, extracted in three different days through the complete automated procedure. 2.8.4. SPE vs LLE recovery The two preparative techniques were compared in order to evaluate the difference in extraction recovery efficiency. With the purpose of eliminating the influence of the derivatization step in the recovery evaluation, ES were added after extraction. The extraction recovery was calculated dividing the mean peak-area ratio of target ion/internal standard for the samples in which the standard solution was added before extraction, with the mean peak-area ratio of target ion/internal standard for the samples in which the standard solution was added after extraction. 2.8.5. Cross contamination and memory effect Cross contamination was investigated at two different levels. To investigate upper cross contamination, which may occur during dispensation processes, 18 blank samples were spiked with high levels of analytes (1 mg/mL) and were extracted as follows. A 96-well plate was prepared so as to ensure that the 18 high positive samples were surrounded by negative samples. In each well internal standards were added and all samples were analyzed in GC/MS. To avoid lower cross contamination that may occur during elution, we used the Clean Thru Tip (CTT) technology, which eliminates sample carry over from the vacuum manifold. This technology provides SPE columns with a disposable tip that attaches to the end of each column, eliminates any contact between the sample, wash solvent and the extraction apparatus. To simplify extraction method, lower cross contamination was tested even without the CTT devices. To ensure the absence of instrument ‘‘analyte memory’’, directly prior to and after extraction of the previous 18 highly positive samples, a plate with 96 blanks was extracted with the same automated method and analyzed in GC/MS. Blank samples were compared and quantified with the calibration curves built for the method validation to ensure the absence of each drug in negative samples.
2.7. Gas chromatography–mass spectrometry method Analysis was performed using a GC/MS. The instrument configuration used consisted of an Agilent 6890N Gas Chromatograph coupled with an Agilent 5973 Mass Selective Detector. GC column was CP-SIL 8–15 m (ID 0.25 mm to 0.25 mm film thickness), carrier gas was helium (1 mL/min). Injection was executed in splitless mode, using a sample volume of 1 mL. Injector temperature was 250 8C. Oven initial temperature was 70 8C. First ramp was 20 8C/min up to 200 8C, second ramp was 6 8C/min up to 220 8C, and third ramp was 30 8C/min up to 290 8C (hold for 2 min). Transfer line temperature was 280 8C. Ions were monitored as follows (underscored ion was chosen as target for quantitative determination): amphetamine (190–118–91 m/z), amphetamine-d5 (194–123–92 m/z), methamphetamine (204–160–91 m/z), MDA (325–135–162 m/ z), MDMA (204–162–135 m/z), MDMA-d5 (208–163–136 m/z), benzoylecgonine (439–318), benzoylecgonine-d3 (442–321), cocaine (303–272–82), cocaine-d3 (306–275), morphine (414–577 m/z), morphine-d3 (417–580 m/z), codeine (445– 282 m/z), 6-monoacetylmorphine (414–473 m/z), and methadone (294–72 m/z). 2.8. Method validation The automated method was validated in combination with GC/MS analysis. Data analysis was executed checking standard validation parameters. 2.8.1. Linearity Calibration curves were built using a linear regression (y = mx + b) between the interested range of 0.05–5 ng/mg.
2.8.6. Real samples analysis Finally, this automated procedure was applied for 30 real cases and two saliva quality controls (sent by ISS, Roma). All these samples were processed with our LLE and with the automated SPE. We divided the acidic hair solution and the prebuffered saliva (pH 4 – 1:1) in two aliquots of 1 mL each for hair solution and of 200 mL each for saliva. The first part was neutralized and then extracted at pH 9 with 4 mL of a mixture of trichloromethane:n-heptane:isopropanol (50:33:17, v/v). The second part was processed with the SPE automated method. Both types of extract were then dried, derivatized and analyzed in GC/MS.
3. Results and discussion 3.1. Method validation Tables 1 and 2 summarize all the results obtained in the validation procedure. The r2 values ranged from 0.996 to 0.999 (Table 1). This correlation demonstrates the capacity of the instrument to obtain fine calibration curves, minimizing the variability introduced by the operator. Sensitivity parameters were satisfying for many analytes (LOQs ranged from 0.038 to 0.046 ng/mg). However amphetamines and MDMA calculated LOQs impose to build calibration curve with
I. Angeli et al. / Forensic Science International 218 (2012) 15–19 Table 1 Results of linearity test.
Cocaine Benzoylecgonine Morphine Codeine 6-MAM Methadone Amphetamine Methamphetamine MDMA MDA
Target ion (m/z)
Equation
r2
303 439 414 445 414 294 190 204 204 162
y = 0.02196x 0.00195 y = 0.05025x + 0.00611 y = 0.01221x 0.003336 y = 0.007694x 0.00343 y = 0.01553x 0.05677 y = 0.002155x 0.00147 y = 0.010220x 0.0191 y = 0.000958x 0.00404 y = 0.000828x 0.003429 y = 0.004011x + 0.005012
0.999 0.999 0.999 0.996 0.998 0.998 0.998 0.997 0.999 0.998
higher starting point (0.1 ng/mg instead of 0.05 ng/mg) and suggest the necessity to use at least 20 mg of hair to reach the workplace cutoff value of 0.2 ng/mg. Furthermore, regarding to specificity, endogenous interferents that may be present from the matrix or from the components of the SPE column, from solvents or buffers did not produce any peaks that interfere with our analysis. The %RSDs for precision were always between 13.4% and 5.5%. Best accuracy values were obtained for amphetamine ( 1.2%) and for benzoylecgonine (+4.2%). 3.2. Cross contamination and memory effect After the execution of the previous described procedure with samples spiked with high levels of analytes (1 mg/mL), they were analyzed in GC/MS and often resulted totally negative. However in some cases, cross contamination occurred. In these conditions, analytes (always cocaine and benzoylecgonine) were present, but their concentration was always under the lower calibration point of 0.05 ng/mg. In real hair, concentration of 20 ng/mg are rare but possible, for instance in the profile of a cocaine abuser. In the face of
17
this sporadic event, it is necessary to consider the decisional cutoff for positivity or always keep in mind that some rare contamination between samples may occur. We suggest performing hair analysis always coupled with a semi-quantitative screening procedure, so as to dilute very high positive samples before SPE. 3.3. LLE vs SPE recovery As can be seen in Table 3, we obtained evidences of best recovery extraction with SPE rather than LLE. 3.4. Real samples analysis The quantitative results obtained were in good agreement with those obtained by the conventional method and never differ than 10% RSD. Moreover no false positive or negative was detected. Despite those obtained with the SPE procedure, in two cases with low cocaine positivity, the determination of benzoylecgonine in the liquid/liquid extract was negative, due to the bad recovery rate of the substance with this technique. As to saliva quality controls, our results perfectly match both the target compounds and their concentrations. In the first sample both cocaine (19.7 ng/mL) and benzoylecgonine (18.1 ng/mL) were identified, whereas the second one was negative. In Fig. 1 it is shown the overlaid chromatograms of target ions between LLE and SPE for cocaine (COC), morphine (MOR) and benzoylecgonine (BEG) in a real testing hair (COC: 1.83 ng/ mg – MOR: 0.88 ng/mg – BEG: 0.64 ng/mg). 3.5. Extract purity Samples were extracted both with SPE and LLE method, so as to verify extract purity in GC/MS. For instance, in Fig. 2 it is represented one of the real sample (MDMA – 1.74 ng/mg), in which
Table 2 Results of validation parameters. Method performance parameters determined with spiked hair samples by GC/MS Cocaine
Benzoylecgonine
303 439 Target ion (m/z) Retention time (min) 11.4 9.6 LOD (ng/mg) 0.012 0.014 LOQ (ng/mg) 0.038 0.042 Intra-day reproducibility (%) (3 different levels) Low level (0.1 ng/mg) 7.5 8.6 Medium level (1 ng/mg) 6.4 5.5 High level (5 ng/mg) 6.9 7.9 Inter-day reproducibility (%) (3 different levels) Low level (0.1 ng/mg) 8.1 9.0 Medium level (1 ng/mg) 7.1 6.0 High level (5 ng/mg) 8.0 8.5 Accuracy (BIAS %) (3 different levels) Low level (0.1 ng/mg) 5.8 6.5 Medium level (1 ng/mg) 10.2 4.2 High level (5 ng/mg) 9.3 6.7
Morphine
Codeine
6-MAM
Methadone
Amphetamine
Methamphetamine
MDMA
MDA
414 11.8 0.016 0.046
445 12.2 0.016 0.045
414 12.6 0.015 0.043
294 10.8 0.016 0.042
190 5.3 0.0350 0.057
204 6.1 0.037 0.056
204 8.0 0.035 0.096
162 7.4 0.031 0.083
8.5 5.7 8.1
10.8 7.4 9.3
8.4 6.0 9.5
8.9 7.3 8.1
10.7 7.6 9.3
12.1 8.4 10.5
11.3 8.6 10.1
11.5 9.7 10.6
9.0 6.5 8.4
11.4 7.9 10.1
8.9 7.0 10.6
10.4 9.1 9.9
12.7 8.9 11.4
13.4 9.5 11.0
13.5 9.8 10.9
12.5 10.5 11.5
8.5 5.4 19.1
8.6 6.1 9.1
9.1 5.4 8.5
6.2 8.5 4.5
8.5 7.7 1.2
9.8 9.1 2.6
9.6 8.5 5.5
9.8 8.7 3.1
Table 3 Recovery percentages (SPE and LLE).
SPE Low level (0.1 ng/mg) Medium level (1 ng/mg) High level (5 ng/mg) LLE Low level (0.1 ng/mg) Medium level (1 ng/mg) High level (5 ng/mg)
Cocaine
Benzoylecgonine
Morphine
Codeine
6-MAM
Methadone
Amphetamine
Methamphetamine
MDMA
MDA
74 83 80
95 101 102
73 93 76
84 83 89
69 70 73
75 90 94
81 86 74
61 61 60
73 76 83
70 92 90
61 79 46
24 32 21
66 72 78
41 52 48
43 47 47
44 53 53
38 65 64
24 36 19
33 59 58
38 62 63
18
I. Angeli et al. / Forensic Science International 218 (2012) 15–19
Fig. 1. Overlaid chromatograms of cocaine, morphine and benzoylecgonine, respectively in a real case.
Figs. 2–4. Examples of extract purity.
the ion 204 m/z (MDMA) is overlaid. In Figs. 3 and 4 it is shown the TIC of a spiked sample positive for all analytes obtained in SPE and LLE, respectively. GC/MS analysis has demonstrated the great extract purity obtained with the SPE performed with the SPEC MP1 cartridge.
4. Conclusions The usefulness of automated procedures has often been discussed in many trained works. The advantages of SPE automation are well-known and include high throughput, improved precision and robustness. The JANUS Integrator with Varispan arm 8 channels needles – VersaTipPlus, Gripper Arm & Vacuum Filtration Option was used to perform a SPE extraction from human hair to determine cocaine, morphine, amphetamines, MDMA, methadone and their most representative metabolites by GC/MS. In addition, accurate and steady liquid handling ensures good reproducibility. The capacity of the instrument to quickly build a calibration curve makes the validation of every batch results very simple. The automated workstation has shown good reliability (cross contamination and ‘‘memory effect’’ were tested and with some accuracy excluded), effectiveness (no false positive was detected), solvent saving (500 mL/sample vs traditionally LLE 4 mL/sample) and quickness (50 min for 96 tests cycle). The introduction of a 96-well plate simplifies the procedure. The SPEC MP1 resulted suitable for the automation and it lets the
use of the acidic hair extract and the occurrence of the cartridge dry up. So the method has fully satisfied required validation parameters such as extract purity, demonstrated by the reached sensibility, and mean recovery rates (70–90% for all the substances of interest). Finally we have compared different analysis of real hair samples, after ELISA screening for amphetamines, cocaine, opiates and methadone, whose preparation was achieved with our routine LLE or with the developed and validated method proposed. We always obtained best sample purification and an increased analyte recovery (this aspect is particularly enhanced with benzoylecgonine). The procedure was also useful to analyze saliva samples with a simple method and optimal result. References [1] R.L. DuPont, W.A. Baumgartner, Drug testing by urine and hair analysis: complementary features and scientific issues, Forensic Sci. Int. 70 (1995) 63–76. [2] P. Kintz, P. Mangin, Simultaneous determination of opiates, cocaine and major metabolites of cocaine in human hair by gas chromatography/mass spectrometry (GC/MS), Forensic Sci. Int. 73 (1995) 93–100. [3] M. Barroso, M. Dias, D.N. Vieira, J.A. Queiroz, M. Lopez-Rivadulla, Development and validation of an analytical method for the simultaneous determination of cocaine and its main metabolite, benzoylecgonine, in human hair by gas chromatography/mass spectrometry, Rapid Commun. Mass Spectrom. 22 (2008) 3320–3326. [4] F. Musshoff, B. Madea, New trends in hair analysis and scientific demands on validation and technical notes, Forensic Sci. Int. 165 (2007) 204–215. [5] M.J. Telepehak, T.F. August, G. Chaney. Forensic and Clinical Applications of Solid Phase Extraction, Humana Press, 2004.
I. Angeli et al. / Forensic Science International 218 (2012) 15–19 [6] P. Zimmerli, Ch. Staub, Analyse de drogues dans les cheveux par extraction automatisee en phase solide, Rev. Fr. Labor. 282 (1996) 51–55. [7] Ch. Girod, F. De Dominicis, R. Giovannini, Ch. Staub, Analyse de drogues dans les cheveux par extraction automatise’e en phase solide: validation et utilisation dans la routine, Toxicorama 1 (1999) 46–50. [8] C. Girod, C. Staub, Analysis of drugs of abuse in hair by automated solid-phase extraction, GC/EI/MS and GC ion trap/CI/MS, Forensic Sci. Int. 107 (2000) 261–271.
19
[9] T. Sibum, M. Kohal, E. Koster, K. Resch, V. Kaever, Direct and fast determination of antiretroviral drugs by automated online solid phase extraction-liquid chromatography–tandem mass spectrometry in human plasma, Clin. Chem. Lab. Med. 44 (2006) 299–305. [10] F.X. Diamond, W.E. Vickery, J. de Kanel, Extraction of benzoylecgonine (cocaine metabolite) and opiates (codeine and morphine) from urine samples using the Zymark RapidTraceTM, J. Anal. Toxicol. 20 (1996) 587–591.