Artículo de Investigación

Validation of an analytical methodology by gas chromatography for the assay of barbiturates in blood samples

Validación de una metodología analítica para la determinación de barbitúricos en muestras de sangre por cromatografía gaseosa con detector de ionización de llama

Catalina Daza¹, Yady Cabrera¹, Jorge A. Martínez² y Jaime H. Rojas¹

¹ Universidad Nacional de Colombia, Departamento de Farmacia, A. A. 14490, Bogotá, D. C., Colombia. Correo electrónico: jhrojasb@unal.edu.co
² Instituto Nacional de Medicina Legal y Ciencias Forenses. Correo electrónico: jamartinezra@unal.edu.co

Recibido para evaluación: marzo 16 de 2007 Aceptado para publicación: diciembre 17 de 2007

SUMMARY

An analytical methodology was developed for the quantitative assay of phenobarbital, pentobarbital and butobarbital in blood samples, using aprobarbital as an internal standard. After sample cleaning by liquid-liquid extraction, the separation was achieved by liquid gas chromatography using a capillary column in pulsed splitless mode, temperature ramp, and helium as carrier gas and flame ionization detector. The methodology was selective, linear, precise and accurate. The linearity was adequate, between 20 and 100 mcg/mL, for all of three barbiturates studied. The developed methodology was applied in two forensic cases presented in the Toxicology lab of the National Institute of Legal Medicine and Forensic Sciences in Bogotá.

RESUMEN

Una metodología analítica fue desarrollada y validada para la determinación de butabarbital, pentobarbital y fenobarbital en muestras de sangre, utilizando aprobarbital como estándar interno. Para la identificación de los diferentes fármacos se llevó a cabo una extracción líquido-líquido con posterior separación cromatográfica empleando una columna capilar en modo de inyección pulsado sin división conrampa de temperatura, helio como gas de arrastre y detector de ionización de llama. La metodología validada fue selectiva, lineal, precisa y exacta. La linealidad fue evaluada para concentraciones entre 20 y 100 mcg/mL para cada uno de los tres barbitúricos. La metodología desarrollada se aplicó en dos casos forenses presentados en el Laboratorio de Toxicología del Instituto Nacional de Medicina Legal y Ciencias Forenses en Bogotá.

Palabras clave: Barbitúricos, cromatografía gaseosa, muestras forenses, detector de ionización de llama, validación.

INTRODUCTION

Numerous cases associated with intoxications from a great variety of substances arrive at the National Institute of Legal Medicine and Forensic Sciences (NILMFC). Some of those substances are benzodiazepines, illegal recreational drugs and barbiturates. Such intoxications are related to criminal acts such as theft and sexual crimes. This is the reason why the identification and quantification of these substances is important.

In the field of forensics, several preliminary or screening tests are used for identification of these substances. The most commonly used are the inmunoenzimatic methods, which in spite of their high sensitivity can show a lack of selectivity and can therefore produce false positives. For this reason, and to confirm any preliminary positive result, chromatographic techniques are used.

In the determination of barbiturates, liquid chromatography (1-4) and gas chromatography (5-9) techniques are used. Wand et al. reported the use of micellar capillary electrophoresis for the analysis of two barbiturates, barbital and phenobarbital (10). In this research, the results of development and validation of a gas chromatographic method for the determination of butabarbital, pentobarbital and phenobarbital in blood samples are presented. Aprobarbital was used as internal standard. For the validation of methodology we used FDA's parameters and guidelines (11) as well as some recommendations made by other authors (12, 13).

EXPERIMENTAL PHASE OF DEVELOPMENT

Instrumentation and materials

Reagents and standards

Methanol, n-hexane, dichloromethane, n-butyl chloride, chloroform, monosodium phosphate, disodium phosphate, acetic acid, sodium hydroxide (all analytical grade). Deionized water was used throughout all experiments. The standards butabarbital, phenobarbital, pentobarbital, aprobarbital, paracetamol, nicotinamide, caffeine, ibuprofen, and aspirin were purchased from Sigma Aldrich (USA).

Specimens

Blank preserved blood fluid, which was used for the extraction procedures, development of the methodology and validation stage, was obtained from Cruz Roja Colombiana.

Optimization of chromatographic parameters

For this development stage, two analytical procedures reported by Anderson y Fuller (6) and Ojanpera (7) for the identification and determination of acid and neutral drugs were revised and evaluated. The chromatographic conditions and parameters used in this work are shown in Table 1.

After the corresponding trials, Method 1 was selected for the subsequent validation stage considering resolution of the different substances peaks signs and the internal standard, as well as the reproducibility of selection criteria. The internal standard, aprobarbital, was used in concentration of 60 mg/mL. Table 2 shows the results for the GC system suitability parameters related to the first method.

The obtained values for capacity factor, number plates, resolution and tailing factor for all drugs were according to pharmacopoeia requirements (14).

Sample extraction and purification standardization

In accordance with the reviewed literature, there are two keys for the best adequacy and extraction of barbiturates in biological fluids, namely, pH of the buffer and organic solvent used (5-9). For this purpose, a 22 factorial design was used to determine the best combination of these variables. The independent variable was the amount of drug extracted and the working pH values were 4.5 and 5.6. For the solvent categorical levels, butyl chloride and heptane-dichloromethane mixture (50 -50) were used. The factorial designs are shown in Tables 3and Table 4

The computed recovery differences between treatments were statistically significant (Scheffe's test, p = 0,05). As a result of these statistical analyses, the combination of a pH 4.5 buffer and the 50/50 heptane-dichloromethane mixture was selected as the best treatment.

In summary, the extraction procedure established was: to 2 mL of the blood sample 0,1 mL of the internal standard solution, 3 mL of the pH 4,5 buffer and 10 mL of extraction solvent were added. The mix was placed in the rotomixer for 15 minutes, centrifuged at 2000 rpm for 20 minutes and the organic phase removed afterwards by evaporation at 50 °C under nitrogen atmosphere. The residue was dissolved in 100 mL of methanol and 2 mL injected into GC/FID system.

Sample stability

Ina accordance with the FDA guidelines (11) it is necessary to know the stability of the substances under study in the different biological matrixes selected and under different conditions. This stability depends on the nature of the substance, the type of matrix, the storage conditions, and also on the container of the matrix. The stability of substances in the selected matrix was experimentally evaluated under the freezing cycles (at -20°C) between 12 and 24 hours and thawing it at room temperature, which was repeated three times. The short-term stability was analyzed, using a 4-hour period at room temperature; the storage temperature of -20°C. The long term stability for freezing was carried out at -20°C during the period between the preparation of the first sample and the date of the last analysis, as well as in the auto-sampler tray.

The chromatogram of the blank blood sample obtained after storage at -20°C during 2 months, showed change when compared to the chromatogram obtained at the beginning of the study (Figure 1).

Table 5 shows the results of the study expressed as a percentage of the initial concentration.

Validation stage

Selectivity. Figure 1shows no signals of endogen interferences that may come from blank blood samples.

Likewise, and according to the FDA's guidelines (11), effectiveness of the chromatographic system was verified against the use of some commonly used drugs, such as ibuprofen, nicotine, caffeine and aspirin, without finding any interferences or overlapping with the chromatographic signs produced by the barbiturates under study. Retention times for such substances were 5,1, 8,3, 11,5 and 14,6 minutes respectively.

Calibration standard curve. In order to determine the proportionality of the answers as a function of the concentration for each one of the drugs, standard solutions for the system linearity and spiked blank blood samples for the method linearity were studied. Five and three concentration levels for the system and method were used respectively. Figure 2 shows the regression curve for phenobarbital, system and method.

The intercept and slope as well as the experimental t values for student test, system and method, are shown in Table 6

For the system, F's experimental values show significant regressions for the three drugs and non-significant linearity deviations. Similar results were found for the method.

Minimum detectable and quantifiable concentration. For estimate the detection limit (LOD) and the quantification limit (LOQ) the intercept and the slope of the calibration curve were used (11). The detection and quantification limits found are shown in Table 8.

Precision. To evaluate the repeatability and intermediate precision of the method, three concentration levels of each drug were used: 20, 60 y 100 mcg/mL. Intermediate precision was evaluated between days and between analysts. Six and three determinations were carried out for repeatability and intermediate precision, respectively. Table 9 reports the coefficients of variation (RSD) for repeatability, as well as the coefficients for each one of two analysts during the three days of assay. The global values found in all the cases were lower than 15%, values established as the maximum permissible value for bio-analytical methodologies (11).

In accordance with the results for each one of the three barbiturates, experimental F-values are lower than tabulated F-values, showing that there are not significant differences between analysts, days and determinations.

Accuracy. This parameter was evaluated through the recovery percentage, replacing the Y data or the average answer values in the system regression equation obtained for each drug. The value of the experimental concentration was compared to the real concentration added to the matrix, obtaining the percentage of recovery for each drug and for each concentration level (Table 11).

For each one of the barbiturates the recovery percentage is around 80%, which is a value higher than the minimum recommended by the la FDA (11) for bio-analytical methodologies (60%).

Robustness. The variables selected for this study were modified in order to observe the influence of these changes in the answer. In Table 12, the robustness parameters or variables studied are reported.

]]>

For the three barbiturates, according to Youden's method reported by Quattrochi and Laba (15) the critical factors that have to be carefully controlled during the analysis are: equipment, column, liner, material and redisolution time of the dry extract.

APPLICATION. To determine the usefulness of the validated methodology, we carried out a preliminary analysis of ten forensic samples that could contain some of the barbiturates under study. Of them, two showed positive results for phenobarbital with concentrations of 21 and 15,7 mcg/mL respectively. One of them showed toxic concentration (see Table 13). Phenobarbital identity was also confirmed with a second validated HPLC method. Figure 3 shows the chromatogram obtained with one of the two forensic samples.

CONCLUSIONS

An analytical method was develop and validated for the identification and quantification of butabarbital, pentobarbital and phenobarbital in blood samples by gas chromatography with flame ionization detector and pulsed splitless. The validated bioanalytical methodology is specific, linear, precise and accurate between 20 and 100 mcg/mL for phenobarbital, pentobarbital and butabarbital. The stability of the drugs used was rigorously evaluated, mainly for samples stored at room temperature. The validated methodology, as it was demonstrated, proved useful in establishing the presence of studied barbiturates and for their quantitative determination in blood forensic samples.

ACKNOWLEDGES

BIBLIOGRAPHY

1. Pihlainen K., Sippola E. y Kostiainen R., Rapid identification and quantification of compounds with forensic interest using phase liquid chromatography-ion trap mass spectrometry and library searching, Journal of Chromatography A, 995, 93-102 (2003).        [ Links ]

2. Miyaguchi, H., Kuwayama K., Tsujikawa T., Kanamon T, Iwata Y. y Kishi, T., A method for screening for various sedative-hypnotics in serum by liquid chromatography/single quadruple mass spectrometry, Forensic Science International, 157, 57-70 (2006).        [ Links ]

3. Capella-Pieró M., Gil-Agusti M., Martinavarro-Domínguez A. y Esteve-Romero J., Determination in serum of some barbiturates using micellar liquid chromatography with direct injection, Analytical Biochemistry, 309, 261-268 (2002).        [ Links ]

4. Esteve-Romero J., Carda-Broch S., Gil-Agusti M. Capella-Pieró M. y Bose, D. Micellar liquid chromatography for the determination of drugs materials in pharmaceutical preparations and biological samples, TrAc Trends in Analytical Chemistry, 24, 75-91 (2005).        [ Links ]

5. Zhao H., Wang L., Qiu Y., Zhou Z., Li X. y Zhong W., Simultaneous determination of three residual barbiturates in pork using accelerated solvent extraction and gas chromatography-mass spectrometry, Journal of Chromatography B., 840, 139-145 (2006).        [ Links ]

6. Anderson W.H. y Fuller D.C., A simplified procedure for the isolation, characterization and identification of weak acid and neutral drugs from whole blood, Journal of Analytical Toxicology, 11, Sep-Oct (1987).        [ Links ]

7. Ojanpera, I., Automated quantitative screening for acidic and neutral drugs in whole blood by dual-column capillary gas chromatography, Journal of Analytical Toxicology, Jul- Aug, 14(4), 2048 (1997).        [ Links ]

8. Whitter P. y Cary, P., A rapid method for the identification of acidic, neutral and basic drugs in postmortem liver specimens by toxi-lab, Journal of Analytical Toxicology, Mar-Apr., 10, 68-71 (1986).        [ Links ]

9. Perchalki, R., Rapid, Simultaneous GLC determination of phenobarbital, primidone and diphenylhidantoin, Journal of Pharmaceutical Sciences, 62, 1735-1736 (1973).        [ Links ]

10. Wang Q-L., Fan L-Y., Zhang W. y Cao C-X., Sensitive analysis of two barbiturates in human urine by capillary electrophoresis with sample stacking induced by moving reaction boundary, Analytica Chimica Acta, 580, 200-205 (2006).         [ Links ]

11. Guidance for Industry, Bioanalytical Methods Validation, Guidance, Food and Drug Administration, Centre for Drug Evaluation and Research CDER, May 2001.        [ Links ]

12. Singtoroj. T, Tarning J. A new approach to evaluate regression models during validation of bioanalytical assays, Journal of Pharmaceutical and Biomedical Analysis, 41, 219-227 (2006).        [ Links ]

13. Boulanger B. Chiap. P. An analysis of the SFSTP guide on validation of chromatographic bioanalytical methods: progresses and limitations, Journal of Pharmaceutical and Biomedical Analysis, 32, 753-765 (2003).        [ Links ]

14. The United States Pharmacopoeia, USP 29, The Pharmacopoeia Convention Inc., 2006, page 588.        [ Links ]

15. O.A. Quattrochi, R.F. Laba y S.I. Belaira, "Introducción a la Cromatografía Líquida de Alta Eficiencia, Aplicación y Práctica", Artes Gráficas Farro S.A., B.A., Argentina, 1992.        [ Links ]

16. Anthony C. Moffat, M. David Osselton, Clarke's Analysis of drugs and Poisons, The Pharmaceutical Press, 3a Ed., 2004.        [ Links ] ]]> 1 2003 995 93-102 2 2006 157 57-70 3 2002 309 261-268 4 2005 24 75-91 5 2006 840 139-145 6 1987 11 7 1997 14 ^s4 ^s4 4 2048 8 1986 10 68-71 9 1973 62 1735-1736 10 2006 580 200-205 11 Guidance for Industry 2001 12 2006 41 219-227 13 2003 32 753-765 14 2006 588 15 1992 16 2004 3