1 Introduction Since the end of the 1970s, toxicological hair analysis has
increasingly been used to investigate drug abuse histories
In the determination of forensic drugs in biological sam- and is now widely accepted as a tool to monitor past,
CE and CEC
ples a major analytical challenge is the limit of detection chronic exposure to drugs [7, 8]. Unfortunately, drug and
and quantitation [1] because of the low concentration lev- drug metabolite concentration in the hair matrix is low (in
els, which may be of analytical interest. Moreover, addi- the order of fractions of nanogram per milligram of hair)
tional problems come from the high and variable content and, for aesthetic reasons, the amount of hair sample
of proteins, lipids and ions which may interfere with drug which can be collected is limited. In hair analysis, CE has
determination [2, 3]. Therefore, the analysis of biological become a valuable tool because its separation principles
fluids generally requires preliminary sample clean-up and are completely different from those of gas and liquid chro-
preconcentration steps [4, 5]. Capillary electrophoresis matography, which are at present the most used tech-
(CE) is a relatively new analytical technique, which has niques. Furthermore, CE analysis is characterized by
been applied to many aspects of drug analysis [6]. As is minimal sample consumption, which is important in order
well known, although very high mass sensitivity detection to preserve hair sample for further confirmation analyses.
is easily obtained, CE, particularly when applied with UV In our experience [9], capillary zone electrophoresis
absorbance detection, suffers from limited concentration (CZE), carried out in borate buffer at pH 9.2, allowed mor-
sensitivity due to the short optical pathlength inherent in phine and cocaine detection in hair extracts, but the assay
the in-capillary detection. sensitivity was moderate, i.e., insufficient for use in real
cases of chronic but moderate drug abuse.
Correspondence: Dr. Giulia Manetto, Department of Public
Medicine and Health, Section of Forensic Medicine, Policlinico
Enhanced limits of detection in CZE have been reported
Borgo Roma, 37134 Verona, Italy
E-mail: giumnt@borgoroma.univr.it by using on-column concentration techniques like field-
Fax: +39-045-505-259 amplified sample stacking (FASS), which is based on a
mismatch between the electric conductivity of the sample
Abbreviations: FASS, field-amplified sample stacking; IS, inter-
and that of the running buffer [1013]. If the sample con-
nal standard; 6-MAM, 6-monoacetylmorphine ductivity is lower than that of the background electrolyte
(BGE), the higher electric field and consequently the high- with a diode array UV detector. Uncoated fused-silica
er migration velocities of the analytes result in an injected capillaries (47 cm total length, 40 cm effective length,
sample plug in comparison with the separation buffer, and 75 mm ID) from Composite Metal Services (The Chase,
this produces a sharpening of the analyte zone at the Hallow, Worcestershire, UK) were used throughout.
boundary with the BGE. A preliminary separation using a
sample stacking method coupled with CZE separation in
plain aqueous medium resulted in about 100-fold in- 2.3 Electrophoretic conditions
creased sensitivity, thus allowing UV spectral identifica- The electrophoretic separations were carried out using a
tion and quantitation of abused drugs in hair samples running buffer composed of 0.1 M sodium phosphate
down to low ng/mg concentration levels [14]. The present adjusted to pH 2.5. Binary running buffers were com-
study describes a further optimization of a head-column posed of aqueous phosphate solution (pH 2.5) supple-
FASS approach in a binary buffer system, aimed at the mented with ethylene glycol at percentages ranging be-
high efficiency separation and high sensitivity detection of tween 2560% v/v. A constant voltage of 20 kV was
a mixture of commonly available opiates (i.e., methadone, applied with the anode at the injection end. UV absor-
morphine, codeine, 6-monoacetylmorphine (6-MAM ace- bance detection took place at the fixed wavelength of 214
tylcodeine, heroin). Also, a preliminary application of the nm (filter UV detector) or by recording the full spectrum
optimized CZE conditions to hair analysis is reported as between 190 and 400 nm with the diode array detector.
well as the use of diode array UV detection in order to The temperature of the capillary was kept constant at
improve analytical selectivity and the identification power 25oC. Conditioning of fresh capillaries was effected by
of CZE. rinsing with 1 M NaOH (4 min), 0.1 M NaOH (4 min), H2O
expressed as number of theoretical plates, was calcu- constant of binary media is proportional to the volume
lated as N=16 (t/W)2, where t is the migration time and W fraction of the components [17] and thus could be calcu-
is the peak width at peak base. The length of the water lated from pure solvents (78.54 and 37.7 for water and
plug injected by pressure application was calculated ethylene glycol, respectively) according to the following
according to the following equation: L=Pr2t /8hLt, where P equation: e=78.5 (1-x)+37.7 x, where x represents the vol-
is the pressure difference between the capillary ends, r is ume fraction of ethylene glycol.
the capillary radius, h is the viscosity of the buffer, and Lt
is the total length of the capillary. The viscosity of buffers
and solutions was measured by the CE apparatus [16,
3 Results and discussion
17] according to the equation h=hO t /tO, where hO indi-
3.1 Optimization of CZE conditions
cates the tabulated viscosity of a reference liquid (e.g.,
water) and t and tO the measured times to reach the It is well known that, because of the limited volume of the
detector under constant pressure for the studied liquid capillary, separation in CE is highly affected by the length
and for the reference liquid, respectively. The dielectric of the injection plug. FASS provides a way to concentrate
Electrophoresis 2000, 21, 28912898 FASS in toxicological hair analysis 2895
(0.03 ng/mg), solely to confirm heroin use or to deny Table 2. RSDs of absolute and relative migration times
external contamination of hair. The hair blank showed no and peak area of pure standards (intraday:
significant peaks that could interfere with the investigated n=10; day-to-day: n=5)
opioids in the time window from 34 to 40 min. On the RSD% RSD%
other hand, the spiked sample shows well-separated and Compound Migration time Peak area
symmetric peaks of codeine, morphine, ethylmorphine Intraday Day-to-day Intraday Day-to-day
(IS), 6-MAM, acetylcodeine, and heroin, which are clearly
Methadone 0.76 - 14.12 -
distinguishable from the background. The FASS increase
Methadone/IS 0.05 - 3.04 -
in the injected amount of sample allowed the recording of
Codeine 0.79 6.44 13.1 26.98
UV spectra from the peaks, which are overimposable onto Codeine/IS 0.03 0.19 1.15 3.96
the spectra from standards. Figure 5 shows the electro- Morphine 0.80 6.47 11.36 31.02
pherograms of a hair sample from a real user of heroin Morphine/IS 0.02 0.12 0.82 3.98
(containing 0.22 ng/mg of morphine according to a vali- 6-MAM 0.84 6.54 11.22 33.97
dated HPLC method [15]) after direct liquid-liquid extrac- 6-MAM/IS 0.06 0.08 5.71 4.41
tion. On the basis of migration times and UV spectra,
codeine and morphine were identified, at levels of 0.07
and 0.20 ng/mg, respectively. Figure 6 shows the UV on average. The theoretical sensitivity, calculated on
spectra of morphine in the hair of a heroin user superim- the basis of the results from spiked samples, in the
posed with the UV spectra of the pure standard. The back hair extract was 100 pg/mL for codeine, 75 pg/mL for
extraction into 1 mM H3PO4 of the Toxi-Lab organic morphine and 6-MAM, 150 pg/mL for ethylmorphine,
extract gave a slightly cleaner electropherogram, with 750 pg/mL for acetylcodeine and heroin (with a signal-to-
similar quantitative results for codeine and morphine. noise ratio of 3). In terms of drug concentration in hair the
LODs correspond to 20 pg/mg for codeine, 15 pg/mg for
3.4 Validation morphine and 6-MAM, 30 pg/mg for ethylmorphine, and
150 pg/mg for acetylcodeine and heroin, which perfectly
Under optimized conditions, the absolute and relative satisfied the cutoff limits applied in our forensic labora-
within-day relative standard deviations (RSD%) for migra- tory.
tion times and peak areas were determined for ten repli-
cate samples of pure standards at concentrations of
4 Concluding remarks
10 ng/mL. The day-to-day precision was determined over
a period of five days with three replicate samples at con- In recent years, hair analysis for drugs has gained adva-
centrations of 0.1 mg/mL. The typical reproducibility data nces in the forensic toxicology environment, as a tool for
obtained are presented in Table 2. As expected, absolute investigating past, chronic exposure to illicit drugs, as
peak area reproducibility with FASS was poor: day-to-day shown by an increasing body of literature [7, 8]. To this
RSD% values were determined to be between 26.98 and purpose, CE has several unique advantages over current
33.97%, but the area normalization on the basis of the IS techniques (GC, GC-MS,, HPLC) nowadays applied,
(ethylmorphine) allowed us to limit RSDs to values such as small sample size, minimal sample preparation,
4.41%. Absolute intraday RSD values were found to be speed of analysis, and low cost of capillaries and chemi-
in the range between 11.2214.12%, but with internal cals, but its sensitivity in terms of concentration was diffi-
standardization RSD were better than 5.71%. The intra- cult to deal with, considering the minute amounts of drugs
day RSDs of absolute migration times were always <1%, present in hair samples.
while the day-to-day RSDs were <6.54%; for relative
migration times, RSDs were <0.06% and 0.2%, respec- In the present paper, our attention was focused on the
tively. optimization of sample enrichment offered by a FASS
before the separation in a binary CZE buffer, providing
With pure standard solutions, using hydrodynamic injec- excellent resolution power, separation efficiency and ana-
tion (0.5 psi for 10 s; no FASS) the limit of detection lytical precision as the sensitivity was more than enough
(LOD) at 214 nm wavelength (S/N = 3) was 0.15 mg/mL to determine some of the major opioids in the hair sam-
for methadone, 0.12 mg/mL for codeine, 0.10 mg/mL for ples. Furthermore, the substantial increase in sensitivity
morphine, and 0.19 mg/mL for 6-MAM. Under FASS- achieved by sample stacking allowed the recording of UV
optimized conditions the LOD values achieved were: 60 spectra from the peaks of drugs in hair even at the cutoff
pg/mL for methadone, 50 pg/mL for codeine, 40 pg/mL for levels adopted (100 pg/mg) and consequently the spec-
morphine, and 90 pg/mL for 6-MAM. Thus an approxi- tral confirmation of results based on migration times. In
mately 2500-fold increase in sensitivity can be calculated conclusion, CZE coupled to head-column FASS once
2898 G. Manetto et al. Electrophoresis 2000, 21, 28912898
again proved to offer a sensitive, precise, easy, low-cost [8] Tagliaro, F., Chiarotti, M., Deyl, Z., Forens. Sci. Int. 1993,
and rugged method for toxicological hair analysis, to be 63, 1315.
used as a complementary technique for confirmation of [9] Tagliaro, F., Poiesi, C., Aiello, R., Dorizzi, R., Ghielmi, S.,
Marigo, M., J. Chromatogr. 1993, 638, 303309.
results obtained with traditional methodologies.
[10] Thormann, W., Zhang, C.-X., Anal. Chem. 1996, 68, 2523
2532.
This work was co-funded by research grants awarded by
[11] Thormann, W., Zhang, C.-X., Anal. Chem. 1998, 70,
Regione Veneto (# 765/01/97), CARIVERONA and
540548.
M.U.R.S.T. #9906404127.
[12] Burg, D. S., Chien, R.-L., Anal. Chem. 1991, 63,
20422047.
Received March 27, 2000
[13] McGrath, G., Smyth, W. F., J. Chromatogr. B 1996, 681,
125131.
5 References [14] Tagliaro, F., Manetto, G., Crivellente, F., Scarcella, D.,
Marigo, M., Forens. Sci. Int. 1998, 92, 201211.
[1] Taylor, R. B., Toasaksiri, S., Reid, G., Electrophoresis
1998, 19, 27912797. [15] Tagliaro, F., De Battisti, Z., Lubli, G., Neri, C., Manetto, G.,
Marigo, M., Forens. Sci. Int. 1997, 84, 129135.
[2] Garcia, L. L., Shihabi, Z. K., J. Chromatogr. 1993, 652,
465469. [16] Weast, R. C., Astle, M. J., Beyer, W. H. (Eds.), Handbook of
Chemistry and Physics, CRC Press, Boca Raton, FL 1978,
[3] Shihabi, Z. K., J. Chromatogr. 1993, 652, 471475.
C294.
[4] Lloyd, D. L., J. Chromatogr. A 1996, 735, 2942.
[17] Thormann, W., Zhang, C.-X., J. Chromatogr. A 1997, 764,
[5] Boone, C. M., Waterval, J. C. M., Lingeman, H., Ensing, K., 157168.
Underberg, W. J. M., J. Pharmaceut. Biomed. Anal. 1999,
[18] Grossman, P. D., Colburn, J. C. (Eds.), Capillary Electro-
20, 831863.
phoresis: Theory and Practice, Academic Press, San Diego,
[6] Thormann, W., Caslavska, J., Electrophoresis 1998, 19, CA 1992, pp. 111132.
26912694.
[7] Kinz, P. (Ed.), Drug Testing in Hair, CRC Press, Boca
Raton, FL 1996, pp. 1293.