Anal. Chem. 2004, 76, 4672-4675
Extraction of Thiamylal in Serum Using Hydrophilic Acetonitrile with Subzero-Temperature and Salting-Out Methods Manabu Yoshida,*,† Atsushi Akane,† Mayumi Nishikawa,‡ Toshimitsu Watabiki,† and Hitoshi Tsuchihashi‡
Department of Legal Medicine, Kansai Medical University, Moriguchi, 570-8506, Japan, and Forensic Science Laboratory, Osaka Prefectural Police H.Q., Osaka, 541-0053, Japan
For high-performance liquid chromatography (HPLC) of thiamylal, one of the barbiturates, the drug in serum samples was extracted by two alternative liquid-liquid extraction techniques using hydrophilic acetonitrile as a solvent and subzero-temperature and salting-out methods. Acetonitrile was mixed with the sample, separated by cooling at -20 °C or addition of sodium chloride, and injected directly into the HPLC apparatus. In both the methods, thiamylal was extracted effectively in the acetonitrile phase and pH adjustment of the sample was not required. The salting-out extraction method is rapid and would be suitable for quantitation of drugs in many samples. To avoid coextraction of added salt, the subzerotemperature extraction method was applied to identification of thiamylal by gas chromatography/mass spectrometryandliquidchromatography-tandemmassspectrometry. Drugs in biological materials are generally extracted and purified prior to assay by high-performance liquid chromatography (HPLC), gas chromatography, and mass spectrometry. In conventional liquid-liquid extraction methods,1-4 hydrophobic organic solvents are used to extract drugs from aqueous materials. For HPLC, however, the organic solvents should be vaporized, and dried drugs should be dissolved in hydrophilic organic solvents such as acetonitrile and methanol. In our previous report,5 a novel subzero-temperature liquid-liquid extraction method was applied to the assay of benzodiazepines by HPLC, in which acetonitrile was mixed with the aqueous samples, separated from them by cooling at -20 °C, and injected directly into the HPLC apparatus. In this study, an alternative salting-out liquid-liquid extraction technique was applied to HPLC analysis of thiamylal, * To whom correspondence should be addressed. Fax: +81-6-6992-6582. E-mail:
[email protected]. † Kansai Medical University. ‡ Osaka Prefectural Police H.Q. (1) Smith, R. H.; MacDonald, J. A.; Thompson, D. S.; Flacke, W. E. Clin. Chem. 1977, 23, 1306-1309. (2) Kudo, K.; Nagata, T.; Kimura, K.; Uehori, R.; Noda, M. Forensic Sci. Int. 1988, 37, 193-200. (3) Hosotsubo, H.; Takeda, K.; Hosotsubo, K.; Yoshiya, I. J. Chromatogr. 1989, 487, 204-209. (4) Costantino, A. G.; Caplan, Y. H.; Levine, B. S.; Dixon, A. M.; Smialek, J. E. J. Forensic Sci. 1990, 35, 89-96. (5) Yoshida, M.; Akane, A. Anal. Chem. 1999, 71, 1918-1921.
4672 Analytical Chemistry, Vol. 76, No. 16, August 15, 2004
in which acetonitrile was separated by adding sodium chloride,6 as by well as the subzero-temperature liquid-liquid extraction method. EXPERIMENTAL SECTION Subzero-Temperature Liquid-Liquid Extraction. Acetonitrile (0.5 mL, Wako, Osaka, Japan) was mixed well with 0.5 mL of each sample (human serum or an aqueous solution containing 5 µg/mL thiamylal sodium salt (Kyorin Pharmaceutical, Tokyo, Japan) in a 1.5-mL polypropylene microtube. Following centrifugation at 1200-3000g for 5 min, and refrigeration at -20 °C for 30 min, 10 µL of the separated acetonitrile upper phase was injected into a high-performance liquid chromatograph. Salting-Out Liquid-Liquid Extraction. Acetonitrile (0.5 mL) was mixed well with 0.5 mL of each sample and 0.2 g of sodium chloride (Wako) in a microtube for 30 s using a vortex mixer. Following centrifugation at 1200-3000g for 5 min, 10 µL of the separated acetonitrile upper phase was injected into the chromatograph. HPLC. For quantitation of thiamylal, the chromatograph system was composed of a LC-6A pump (Shimadzu, Kyoto, Japan), a UV-visible spectrophotometric detector (SPD-6AV, Shimadzu) set at 287 nm, a Rheodyne model 7125 injector (Cotati, CA) fitted with a 100-µL loop or a Shimadzu SIL-7A autoinjector, and a Develosil C8-UG-5 column (4.6 mm i.d. × 150 mm, Nomura Chemical, Seto, Japan). The mobile phase was 0.05% acetic acid/ 50% acetonitrile, flowing at 0.7 mL/min. For identification of thiamylal, the chromatograph was equipped with a LC-10AS pump (Shimadzu) and a photodiode array detector (SPD-M10Avp, Shimadzu), instead of the LC-6A pump and the SPD-6AV detector. Experiments. First, the effect on the separation efficiency of nondiluted acetonitrile was studied. Acetonitrile was mixed with 5 µg/mL aqueous thiamylal solution and separated by the saltingout liquid-liquid extraction method. Following the injection of 5, 10, 20, 50, or 100 µL of the acetonitrile phase into the chromatograph, the separation efficiency was investigated by the appearance of the thiamylal peak in the chromatogram and its number of theoretical plates. The number was calculated as 16 (tR/w)2, where tR is the retention time and w is the peak width measured in the same units as tR. (6) Tsuchihashi, H.; Katagi, M.; Nishikawa, M.; Tatsuno, M. J. Anal. Toxicol. 1998, 22, 383-388. 10.1021/ac040065a CCC: $27.50
© 2004 American Chemical Society Published on Web 06/30/2004
Second, thiamylal concentrations in the acetonitrile and aqueous phases were determined by HPLC. Third, the effect on the extraction of the pH of the aqueous phase was investigated. An aqueous solution (0.5 mL) containing 5 µg/mL thiamylal was mixed with 50 µL of Milli-Q ultrapure water, concentrated acetic acid, 10% tartaric acid, 20% sodium carbonate solution, or concentrated ammonia. The pH of each sample was determined with a pH meter (Twin pH, Horiba, Kyoto, Japan). Following the separation, the thiamylal concentration in each phase was determined as well. Fourth, calibration curves for the drug extracted by the saltingout method were obtained with aqueous solutions and human serum containing 0.02-40 µg/mL thiamylal. Case Study. Thiamylal in an anesthetized patient’s serum was extracted by the subzero-temperature method for identification and by the salting-out method for quantitation. The drug in 10 µL of the acetonitrile phase was identified and quantitated by HPLC as described above. The drug was also identified by gas chromatography/mass spectrometry (GC/MS) and HPLC-tandem mass spectrometry (HPLC-MS-MS). GC/MS assay was performed as follows. The separated acetonitrile phase (0.5 mL) was dried with a Himac CE1 centrifugal evaporator (Hitachi, Hitachinaka, Japan), and dissolved in 10 µL of acetonitrile befor being injected into the QP2000A gas chromatograph/mass spectrometer (Shimadzu) equipped with a CBP1M25 column (25 m × 0.2 mm i.d., 0.25-µm film thickness, Shimadzu). The GC/MS conditions were as follows: carrier gas, helium flow 0.56 mL/min; column temperature, 60 °C, maintained for 5 min, to 290 °C programmed at 20 °C/min; injection temperature, 280 °C; injection method, splitless method for 30 s from the injection, and then split (1/85); interface temperature, 290 °C; mode, electron impact ionization; ion source temperature, 280 °C; electron energy, 70 eV; total emission current, 400 mA; scanning mass range, 35-540 m/z, and scanning interval, 2.0 s. HPLC-MS-MS assays were performed as follows. The dried and dissolved acetonitrile phase was injected into a HewlettPackard (Agilent) HP1100 Series pump coupled to a Micromass Quattro LC (Micromass Co, Manchester, U.K.) using an electrospray interface. The tandem mass spectrometer was operated in the positive ion mode. The separation column was Develosil C8UG-5. The mobile phase was 0.1% acetic acid/50% acetonitrile, flowing at 0.1 mL/min (10:1 ratio split of 1.0 mL/min). The sample inlet utilized a heated nebulizer set at 185 °C, the sample cone voltage was set to 3.5 kV, the collision energy was set to 12 eV; the collision gas consisted of argon with the gas cell set at 1.9 × 10-3 mbar. A m/z 255 ion was used as the precursor ion. RESULTS Table 1 shows the relationship between the injected volume of acetonitrile and the number of theoretical plates (N). Greater volume gave lower N and the broader peaks resulting in shortened retention times. Since retention times for 10-µL injection and 5-µL injection were the same, and since 5 µL of acetonitrile contains less thiamylal than 10 µL, the injection amount was set at 10 µL. The chromatograms obtained by HPLC assay of acetonitrile and aqueous phases separated by the subzero-temperature and salting-out methods are shown in Figure 1. The retention time of thiamylal was 8.4 min. From the aqueous solution containing 5 µg/mL thiamylal, 99.2 and 99.7% of the added amount of the drug
Table 1. Effect of Injection Amount on the Column Separation of Thiamylal by HPLCa injection amt (µL)
no. of theoretical plates
retention time (min)
5 10 20 50 100
11082 ( 25.8 (0.233) 9763 ( 63.1 (0.647) 6612 ( 137.(2.08) 1944 ( 52.3 (2.69) 5.33 ( 2.08 (39.6)
8.62 ( 0.251 (2.91) 8.57 ( 0.229 (2.67) 8.49 ( 0.168 (1.98) 8.38 ( 0.133 (1.58) 7.82 ( 0.0147 (0.192)
a Mean ( standard deviation for three determinations. Coefficients of variance (%) in parentheses.
Figure 1. High-performance liquid chromatograms of acetonitrile and aqueous phase. Left: Salting-out extraction method (5 µg/mL thiamylal/aqueous solution), Right: Subzero-temperature extraction method (5 µg/mL thiamylal/serum), AC, acetonitrile phase; AQ, aqueous phase.
were recovered into the acetonitrile phase by the methods (Table 2). The coefficients of variance suggested high reproducibility. When acetonitrile and sodium chloride were added to human serum containing thiamylal, the separated aqueous phase became emulsive. When the mixture of acetonitrile and serum was refrigerated, the aqueous phase in which most of serum contents were detected (Figure 1) was not emulsive. No peak was detected at the retention time of 8.3 min in samples containing no thiamylal. Peak areas of thiamylal obtained by the injection of 10 µL of acetonitrile phase separated from the aqueous sample by the subzero-temperature and salting-out methods were 1.9- and 1.4fold greater, respectively, than that obtained by the direct injection of 10 µL of the aqueous sample into the chromatograph before the addition of acetonitrile. Thus, the target drug was concentrated by the extraction methods. Peak areas of thiamylal extracted by the subzero-temperature method from the aqueous and serum samples were 1.43- and 1.35-fold greater, respectively, than those by the salting-out method (Table 2), suggesting that the former method concentrated thiamylal more than the other. The effect of the pH of samples on the extraction is shown in Table 3. Recovery rates of thiamylal by both methods from acid and neutral solutions were high (>99.0%), but acetonitrile could not be separated from concentrated acetic acid by the subzeroAnalytical Chemistry, Vol. 76, No. 16, August 15, 2004
4673
Table 2. Extraction Rates and the Reproducibility of Thiamylal Analysis by the Two Extraction Methodsa aqueous solution acetonitrile phase
serum aqueous phase
peak area extraction rate (%)b coefficient of variance (%)
Subzero-Temperature Method 214426 ( 3444 1713 ( 305 99.2 ( 1.59 0.793 (0.141 1.61 17.8
peak area extraction rate (%)b coefficient of variance (%)
150208 ( 3055 99.7 ( 2.03 2.03
Salting-Out Method 429 ( 146 0.285 ( 0.0967 34.0
acetonitrile phase
aqueous phase
175285 ( 5242 98.5 ( 2.95 2.99
2689 ( 564 1.51 ( 0.317 21.0
130052 ( 2340
ndc
1.80
a Each value of peak area and extraction rate represents mean ( standard deviation for seven determinations. b Sum of peak areas of thimylal in acetonitrile and aqueous phases as 100%. c Not detected because of emulsion.
Table 3. Effect of pH of the Aqueous Solution on the Extraction of Thiamylal extraction rate of thiamylal (%)a aqueous phase
a
additive
pH
ultrapure water acetic acid tartaric acid sodium carbonate ammonia
7.0 1.9 1.9 10.8 11.0
ultrapure water acetic acid tartaric acid sodium carbonate ammonia
7.0 1.9 1.9 10.8 11.0
acetonitrile phase
peak 1 (7.8 min)
Subzero-Temperature Method 99.0 -b (phases were not separated) 99.2 96.8 3.16 81.0 15.1 Salting-Out Method 99.4 99.4 99.9 97.4 2.64 97.1 2.77
peak 2 (8.3 min)
sum of peak areas
0.984
247 144
0.835 nd 3.93
256 178 174 445 175 772
0.634 0.595 0.113 0.183
154 949 134 610 144 799 138 013 146 116
Sum of areas of the peaks as 100%. Each value represents the mean of two determinations. b -, no peak was detected.
temperature method. The extraction rates from alkaline solutions were slightly lower, especially when an ammonia solution was treated by the subzero-temperature liquid-liquid extraction method. Furthermore, thiamylal extracted from aqueous ammonia by the two methods was detected as twin peaks (peaks 1 and 2 in Table 3 and Figure 2). The retention time of thiamylal extracted from acid and neutral solutions was the same as that of peak 2, whereas thiamylal extracted from alkaline sodium carbonate solution had the same retention time as peak 1. Peaks 1 and 2 detected in the ammonia phase were analyzed by the photodiode array detector. The peaks showed the same absorption spectra as that of standard thiamylal (Figure 3). The HPLC assay of thiamylal purified by the salting-out liquidliquid extraction from the aqueous solution and human serum containing the drug gave linear calibration curves with the correlation coefficients of 0.999 99 and 0.999 82, respectively. The detection limit was 20 ng/mL (S/N ) 9.8) in each specimen. In the case study, a peak at a retention time of 15.3 min was detected by GC/MS, and the mass and HPLC-photodiode array spectra of the peak were identical to those of standard thiamylal. The concentration of thiamylal was 23.3 µg/mL. Trace amounts of secobarbital, a metabolite of thiamylal, were also detected by GC/MS. A peak at a retention time of 6.4 min was also detected by HPLC-MS-MS. In the total ion chromatogram, a peak of thiamylal and peaks of impurities were detected. However, the 4674 Analytical Chemistry, Vol. 76, No. 16, August 15, 2004
Figure 2. Twin peaks detected in high-performance liquid chromatograms. Aqueous solution containing 5 µg/mL thiamylal mixed with concentrated ammonia was extracted using the salting-out extraction method. AC, acetonitrile phase; AQ, aqueous phase.
tandem mass spectrum obtained with the use of the m/z 255 ion as the precursor ion showed m/z 255 ([MH]+) and 184 ions clearly (Figure 4), which was identical to that of standard thiamylal.
Figure 3. Absorption spectra of peaks 1 and 2 (Figure 2) using the photodiode array detector. Solid line, peak 1; broken line, peak 2.
Figure 4. Product ion spectrum of thiamylal by Micromass Quattro (HPLC-MS-MS) (precursor ion, m/z 255).
DISCUSSION Subzero-temperature liquid-liquid extraction is a novel method we had established to extract benzodiazepines from aqueous samples using acetonitrile.5 This extraction method is simple, but it is time-consuming to refrigerate the samples. As an alternative method, we had first applied the salting-out liquid-liquid extraction technique to analyze metabolites of O-ethyl S-2-diisopropylaminoethyl methylphosphonothiate, known as a chemical weapon VX.6 From the serum sample, 2-(diisopropylaminoethyl)methyl sulfide, a metabolite of VX, was purified by the conventional liquid-liquid extraction method using dichloromethane. Then, from the aqueous phase, the other metabolite ethyl methylphosphonic acid was extracted by the salting-out method. Here we applied the salting-out method as the subzero-temperature method to direct extraction of thiamylal from serum as well. Thiamylal is one of the barbiturates, an ultra-short-acting anaesthetic. It is necessary to monitor its blood concentration because this drug has side effects such as respiratory and circulatory suppression. Applicability of the two alternative extraction methods was investigated for the HPLC assay of thiamylal.
Using each of the extraction methods, most of the thiamylal in serum could be extracted in the separated acetonitrile phase, and serum components remained in the aqueous phase. Addition of acetonitrile to deproteinize the samples results in dilution of them.3 However, the subzero-temperature and salting-out extractions concentrated the target drug as described above. The extraction efficiency in conventional liquid-liquid extraction is known to be affected by pH of the samples.1,7 The saltingout method was less affected by the pH as shown in Table 3. However, the subzero-temperature method was affected by addition of acetic acid or ammonia. Acetic acid inhibited the separation of the phases while ammonia lowered the extraction of thiamylal. Alkaline samples also affect the retention time of thiamylal in the aqueous phase separated by both methods. Sodium carbonate shortened the retention time from 8.3 to 7.8 min. Addition of ammonia resulted in the detection of two peaks with retention times of 7.9 and 8.3 min. The area of the peak at 7.9 min was greater than that of the other. The cause of this phenomenon is unclear. However, in our previous study in which benzodiazepines were extracted by the subzero-temperature method for HPLC assay, ammonia added to the specimens corroded the HPLC column (Capcell Pak C18). The column used in this study was different, but it might be affected by alkaline solutions as well. These results indicated that pH adjustment of neutral serum samples is unnecessary for both extraction methods. In this study, two alternative extraction methods were investigated for analysis of thiamylal. These methods could be used to extract benzodiazepines as in our previous study5 and would be applicable to many other drugs. The salting-out liquid-liquid extraction technique was less time-consuming than the subzerotemperature method, so that the former method is suitable for analyzing many aqueous samples for quantitation of drugs. However, the acetonitrile phase separated by the salting-out method might contain salts as well as water, which would be harmful to mass spectrometers. For identification by GC/MS or LC-MS(-MS), drugs should be extracted by the subzerotemperature extraction rather than the salting-out method. In the case of thiamylal analysis, the drug in serum could be identified by HPLC (photodiode array analysis), GC/MS, and HPLC-MSMS following the subzero-temperature extraction and could be quantitated by HPLC following the salting-out extraction. If concentrations of drugs in the acetonitrile phase are too small to be analyzed, the injection amount of the phase should not be increased, but the phase should be concentrated, because increase of the injection amount results in decrease of number of theoretical plates.
Received for review April 8, 2004. Accepted May 6, 2004. (7) Stockham, T. L.; McGee, M. P.; Stajic, M. J. Anal. Toxicol. 1991, 15, 155156.
AC040065A
Analytical Chemistry, Vol. 76, No. 16, August 15, 2004
4675
