Dicorcia 2012
Dicorcia 2012
Dicorcia 2012
Journal of Chromatography B
journal homepage: www.elsevier.com/locate/chromb
Short communication
a r t i c l e i n f o a b s t r a c t
Article history: A simple procedure for the quantitative determination in hair samples of 13 common drugs of abuse or
Received 15 September 2011 metabolites (morphine, 6-acetylmorphine, codeine, amphetamine, methamphetamine, 3,4-methylene-
Accepted 1 May 2012 dioxyamphetamine, 3,4-methylenedioxymethamphetamine, 3,4-methylenedioxyethylamphetamine,
Available online 5 May 2012
benzoylecgonine, cocaine, buprenorphine, methadone and 9 -tetrahydrocannabinol) has been devel-
oped and fully validated. The analytes were extracted from the matrix by a simple overnight incubation
Keywords:
with methanol at 55 ◦ C. An aliquot of the extract was directly injected into an ultra-high performance liq-
Hair
uid chromatography system equipped with Waters Acquity UHPLC BEH C18 column (100 mm × 2.1 mm,
Ultra-high performance liquid
chromatography
1.7 m). The mobile phase eluted with a linear gradient (water/formic acid 5 mM:acetonitrile; v:v) from
Multidrugs 98:2 to 0:100 in 4.5 min, followed by isocratic elution at 100% B for 1.0 min. The flow rate was 0.6 mL/min
Validation and the total run time was 8.0 min including re-equilibration at the initial conditions. The compounds
were revealed by a triple quadrupole mass spectrometer operating in the selected reaction monitoring
mode. The absence of matrix interferents, together with excellent repeatability of both retention times
and relative abundances of diagnostic transitions, allowed the correct identification of all analytes tested.
The method proved linear in the interval from the limit of quantification to 5.0 ng/mg (1.0 ng/mg for 9 -
tetrahydrocannabinol) with correlation coefficient values ranging from 0.9970 to 0.9997. Quantitation
limits were below the cut-off values recommended by the Society of Hair Testing and ranged from 0.02
to 0.08 ng/mg. Application of the present UHPLC–MS/MS procedure and instrumentation to hair analysis
allows high sample-throughput, together with excellent sensitivity and selectivity, in workplace drug-
screening controls and forensic investigations. These qualities, combined with minimal sample workup,
make the cost of this screening affordable for most private and public administrations.
© 2012 Elsevier B.V. All rights reserved.
1570-0232/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jchromb.2012.05.003
D. Di Corcia et al. / J. Chromatogr. B 899 (2012) 154–159 155
Most of the cited methods were implemented for the determina- was 0.6 mL/min and the total run time was 8.0 min including re-
tion of a few analytes, as is evident from the synoptic table reported equilibration at the initial conditions. The mass analyzers were
in the cited review [14]. In order to perform wide range screening operated in the selected reaction monitoring (SRM) mode. In order
of abused drugs, more than one procedures are likely to be utilized, to establish appropriate SRM conditions, optimization of the mass
with direct impact on efficiency and costs. In particular, exten- spectrometer was conducted by direct infusion of the analytes into
sive application of hair analysis in workplace drug testing is often the electrospray ionization capillary and the declustering poten-
prevented by its cost, crucially high for public administrations. tial (DP) was adjusted to maximize the intensity of the protonated
Our goal was to develop and validate a sensitive multi-class molecular species. The collision offset (CE) voltage values were
and multiresidual screening method for drugs of abuse or metabo- selected so as to preserve approximately 10% of each precursor ion.
lites in hair samples using a dedicated UHPLC–MS/MS protocol. In Nitrogen was employed as the collision gas (5 × 10−3 Pa). The ESI
comparison with the previously reported procedures, the present source was held at 550 ◦ C. Precursor ions and the corresponding
method used a simple sample extraction and direct injection into product ion SRM transitions employed for all analytes and internal
the UHPLC–MS system, avoiding both solid-phase and liquid–liquid standards are presented in Table S1, supplementary data.
extraction. Furthermore, the utilization of recent UHPLC–MS/MS
technology allowed a drastic reduction of the analysis time, with-
2.4. Method validation
out resolution loss. The method proved simple, accurate, rapid and
highly sensitive, allowing the simultaneous detection of all the
The method was validated for linearity, selectivity, quantita-
most common drugs, including THC, and high sample throughput,
tion limits (LOD and LOQ), intra/inter-day precision, and accuracy
resulting in significantly reduced costs of analysis.
[22,23]. Carry-over and matrix effect phenomena were also evalu-
ated.
2. Experimental
Table 1
Calibration levels, R2 values for calibration curve and matrix effect; LODs and LOQs values of the 13 analytes investigated.
Compound Internal standard Calibration levels (ng/mg) Regression coefficient (R2 ) Matrix effecta LOD (ng/mg) LOQb (ng/mg)
2.4.5. Precision and accuracy decision and 2006 SOFT/AAFS guidelines criteria [22]. Moreover,
For all analytes, intra-day precision (expressed as percent vari- the intra-day precision values for retention times measured at 0.2,
ation coefficient, CV%) and accuracy (expressed as bias %) were 1.0 and 5.0 ng/mg concentration (0.04, 0.2 and 1.0 ng/mg for THC)
evaluated at LOQ and at different concentration levels. Ten repli- were below 0.5%, confirming that retention times are repeatable
cates of blank hair samples spiked with the standard solutions at and not affected by the analytes concentration.
the final concentration of LOQs, 0.2, 1.0 and 5.0 ng/mg (0.04, 0.2
and 1.0 ng/mg for 9 -THC) and analyzed by the described method. 3.1.3. Linearity, limit of detection and limit of quantitation
Inter-day precision and accuracy were evaluated by preparing and The SRM protocol described (see Table S1) was used to build
analyzing for three consecutive days one set of ten hair samples the calibration plots for all thirteen analytes. Adequate linearity
spiked with target compounds at LOQ and at the final concentration was observed for all compounds. Table 1 reports the resulting R2
of 1.0 ng/mg (0.2 ng/mg for 9 -THC). Standard criteria designated values, ranging from 0.9970 and 0.9997 and indicating good fit and
satisfactory assay precision when CV% values were below 25% for linearity of the calibration curves.
lower concentrations and below 15% for upper concentrations. Table 1 also reports LOD and LOQ values, calculated from the
Satisfactory accuracy was achieved when the experimentally deter- analysis of multiple blank samples. Detection of analytes at LOD
mined concentrations lied within ±25% from the expected values. levels was confirmed experimentally (see Fig. 1). LOD values ranged
Bias % (%B) was estimated as the percent difference between the between 0.006 ng/mg and 0.027 ng/mg while LOQ values were
average value of a set of measurements (X) and the “true value” (T) estimated between 0.02 ng/mg and 0.08 ng/mg. When compared
following the formula %B = (100/T) × (X − T). The parameters most to existing procedures using similar techniques [14], the present
commonly changing in everyday toxicological analysis, namely method provided equal or lower LODs and LOQs. This demonstrates
sample volume, reagent batch and operator, were deliberately var- that the inclusion of a large set of analytes within the screening
ied to test if satisfactory accuracy was maintained. protocol and the simplified sample treatment did not affect signifi-
cantly the method sensitivity, while the newest UHPLC technology
2.4.6. Carry-over coupled with the last generation of mass spectrometers guaran-
The background chromatographic profiles for each analyte were tees short analysis time with concurrent improvement of analytical
monitored during the analysis of blank hair sample injected for five performances.
times after the chromatographic run of a spiked blank hair sample
containing all the analytes at 5.0 ng/mg concentration. To assure 3.1.4. Matrix effect
the absence of carry-over, the signal to noise ratio (S/N) for each The variability among different hair samples was acceptable
transition had to be lower than 3. (CV% <25%, as shown in Table 1), in consequence we pooled together
the five sources of hair to perform validation experiments such
3. Results and discussion as precision and accuracy. The effect of real hair matrix did not
appear significant for most of the analytes tested (see Table 1).
3.1. Method validation A moderate ion suppression was observed for morphine, 6-MAM,
benzoylecgonine and cocaine, while the matrix influence was
3.1.1. Selectivity minor for methadone, all the amphetamines and buprenorphine.
The SRM chromatograms obtained from five blank hair sam- Only codeine and 9 -THC underwent a considerable ion suppres-
ples presented no peaks arising from endogenous interferences (i.e. sion from keratinic matrix (values <−10%). To compensate as much
S/N < 3) at the expected retention time for all analytes. This demon- as possible the matrix effects present in real hair samples analy-
strated that the method is selective for the tested compounds and sis, all calibration and validation tests were conducted on a pool of
free from positive interference from hair components and column human hair samples, spiked with standard analytes solutions. The
bleeding. good linearity observed in the calibration plots demonstrated how-
ever that the observed matrix effect is proportionally constant, i.e.
3.1.2. Identification criteria does not depend on the analytes’ concentrations.
The SRM transitions selected for each analyte provided at least 4
identification points, while the substantial stability of their relative 3.1.5. Precision and accuracy
abundances proved compliant to the unambiguous identification of Intra- and inter-day data on precision and accuracy are reported
all analytes included in the assay, in agreement with CE/2002/657 in Table 2. The results show satisfactory intra-day repeatability, as
D. Di Corcia et al. / J. Chromatogr. B 899 (2012) 154–159 157
the percent variation coefficient (CV%) is lower than 15% for all the highest concentration level the upper and lower limits were +0.5%
spiked analytes at LOQ, low, medium and high concentration. In (benzoylecgonine) and −13.8% (MDMA). On the whole, all the
particular, intra-day precision exhibits CV% values ranging between experimental bias values were largely below the acceptable limit of
0.7% and 14.9%. ±25% at all concentrations. Only for methamphetamine, the intra-
The intra-day accuracy results proved better than acceptable in day accuracy calculated at LOQ showed a significant overestimation
most cases, particularly at the lowest concentration tested, where (+47.5%).
percent bias values ranged from −4.0% (MDMA) to +10.0% (codeine At the intermediate concentration level (1.0 ng/mg for all ana-
and THC). This upshot is particularly significant, as the 0.2 ng/mg lytes, 0.2 ng/mg for 9 -THC), also the inter-day precision proved
level correspond to the generally accepted cut-off for many ana- satisfactory, as the CV% values ranged from 8.1% for codeine to
lytes, such as opiates, methadone and amphetamines. Indeed, 22.5% for MDEA, likewise the intra-day accuracy, ranging from
high accuracy at cut-off levels is an essential pre-requisite in a −8.4% to +19.6. As for intra-day results, also inter-day preci-
screening method. At intermediate spiking concentrations the bias sion data calculated at LOQ were satisfactory for all compounds
ranged from −5.4% (morphine) to +18.0% (MDEA), whereas at the while methamphetamine, MDMA and methadone were largely
158 D. Di Corcia et al. / J. Chromatogr. B 899 (2012) 154–159
Table 2
Intra/inter-day precision (CV%) and accuracy (bias %) for each analyte tested.
Morphine 8.1 3.9 4.4 3.5 −18.5 +6.0 −5.4 −13.3 8.8 9.9 −24.3 +8.4
Codeine 6.4 3.2 2.6 3.9 −8.0 +10.0 +3.0 −13.2 7.4 8.1 −22.0 −8.4
Amphetamine 5.8 6.3 5.2 9.7 +10.0 +4.0 −2.2 −12.8 5.4 12.0 +5.6 +18.4
6-MAM 14.9 5.0 6.2 2.4 −14.0 +9.0 +7.0 −10.8 17.4 11.0 −14.8 +16.4
MDA 8.1 7.1 5.3 5.9 +3.5 −1.0 −1.0 −6.7 9.1 11.6 −3.0 +4.8
Methamphetamine 13.0 2.6 4.7 5.6 +47.5 +7.0 +10.2 −12.5 21.5 15.5 +52.8 +12.8
MDMA 2.7 5.7 0.7 5.6 +22.0 −4.0 +14.8 −13.8 25.3 16.9 +31.3 +6.0
MDEA 2.6 10.9 8.8 4.4 +14.0 −1.0 +18.0 −11.0 5.5 22.5 +16.0 +14.8
Benzoylecgonine 1.4 3.9 9.0 4.6 −18.0 +6.0 +8.0 +0.5 4.9 17.8 −22.7 +18.0
Cocaine 3.0 9.9 4.7 4.5 +15.0 +9.0 +14.8 −6.5 5.1 12.0 +10.7 −6.4
Buprenorphine 6.7 2.7 2.7 6.9 −3.0 +3.0 −4.2 −9.8 4.3 22.4 −2.2 +9.2
Methadone 7.7 7.5 4.8 5.9 +18.5 +9.0 +9.4 −2.2 10.3 9.4 +26.3 +19.6
THC 12.4 12.4 4.1 5.9 +10.0 +10.0 +1.0 −3.0 20.0 10.8 +14.0 +6.0
a
Low concentration: 0.2 ng/mg (0.04 ng/mg for THC).
b
Medium concentration: 1.0 ng/mg (0.2 ng/mg for THC).
c
High concentration: 5.0 ng/mg (1.0 ng/mg for THC).
performances are high and relatively uniform for all the studied [5] P. Kintz, M. Villain, V. Cirimele, Ther. Drug Monit. 28 (2006) 442.
analytes, so that the present protocol may find easy application in [6] P. Kintz, Anal. Bioanal. Chem. 388 (2007) 1467.
[7] J. Lozano, O. García-Algar, O. Vall, R. de la Torre, G. Scaravelli, S. Pichini, Ther.
routine analysis for toxicological investigations. Drug Monit. 29 (2007) 711.
[8] M. Barroso, M. Dias, D.N. Vieira, M. López-Rivadulla, J.A. Queiroz, Anal. Biochem.
Acknowledgments Chem. 396 (2010) 3059.
[9] E. Gerace, A. Salomone, S. Pellegrino, M. Vincenti, Forensic Sci. Int. 215 (2012)
121.
We gratefully thank Ms. Tonia Lombardo for preparing the sam- [10] A. Salomone, E. Gerace, D. Di Corcia, G. Martra, M. Petrarulo, M. Vincenti, Int. J.
ples. The generous financial contribution from the Compagnia di Legal Med. 126 (2012) 451.
[11] R. Kronstrand, I. Nystrom, J. Strandberg, H. Druid, Forensic Sci. Int. 145 (2004)
San Paolo (Turin, Italy) for acquiring the analytical instrumenta- 183.
tion utilized in the present work is gratefully acknowledged (Grant [12] A. Pelander, J. Ristimaa, I. Rasanen, E. Vuori, I. Ojanpera, Ther. Drug Monit. 30
411/PV-2009.1993). We are also indebted to the Regione Piemonte (2008) 717.
[13] M.K. Klose Nielsen, S.S. Johansen, P.W. Dalsgaard, K. Linnet, Forensic Sci. Int.
for its continuous financial support of our toxicology laboratory.
196 (2010) 85.
[14] M. Wada, R. Ikeda, N. Kuroda, K. Nakashima, Anal. Bioanal. Chem. 397 (2010)
1039.
Appendix A. Supplementary data [15] Y.H. Wu, K.L. Lin, S.C. Chen, Y.Z. Chang, J. Chromatogr. B 870 (2008) 192.
[16] O. Quintela, E. Lendoiro, A. Cruz, A. de Castro, A. Quevedo, C. Jurado, M. López-
Supplementary data associated with this article can be found, Rivadulla, Anal. Bioanal. Chem. 396 (2010) 1703.
[17] F. Bucelli, A. Fratini, P. Bavazzano, N. Comodo, J. Chromatogr. B 877 (2009) 3931.
in the online version, at http://dx.doi.org/10.1016/j.jchromb.
[18] J.C. Dominguez-Romero, J.F. García-Reyes, A. Molina-Díaz, J. Chromatogr. B 879
2012.05.003. (2011) 2034.
[19] S. Vogliardi, D. Favretto, M. Tucci, G. Stocchero, S.D. Ferrara, Anal. Bioanal. Chem.
400 (2011) 51.
References [20] A.U. Jackson, J.F. Garcia-Reyes, J.D. Harper, J.S. Wiley, A. Molina-Díaz, Z. Ouyang,
R.G. Cooks, Analyst 135 (2010) 927.
[1] P. Kintz, Analytical and Practical Aspects of Drug Testing in Hair, CRC Press, [21] S. Vogliardi, D. Favretto, G. Frison, S. Maietti, G. Viel, R. Seraglia, P. Traldi, S.D.
Boca Raton, FL, USA, 2007. Ferrara, Anal. Bioanal. Chem. 396 (2010) 2435.
[2] F. Pragst, M.A. Balikova, Clin. Chim. Acta 370 (2006) 17. [22] SOFT, AAFS, Forensic Toxicology Laboratory Guidelines, SOFT and AAFS
[3] E.J. Cone, Forensic Sci. Int. 121 (2001) 7. (01.11.06).
[4] Y.H. Caplan, B.A. Goldberger, J. Anal. Toxicol. 25 (2001) 396. [23] F.T. Peters, O.H. Drummer, F. Musshoff, Forensic Sci. Int. 165 (2007) 216.