SciELO - Scientific Electronic Library Online

Home Pagelista alfabética de revistas  

Servicios Personalizados




Links relacionados


Journal of the Chilean Chemical Society

versión On-line ISSN 0717-9707

J. Chil. Chem. Soc. v.48 n.3 Concepción sep. 2003 

J. Chil. Chem. Soc., 48, N 3 (2003) ISSN 0717-9324




1 Departamento de Farmacia, Facultad de Farmacia, Universidad de Concepción,
casilla 237, Concepción, Chile. E-mail:
2 Departamento de Bromatología, Nutrición y Dietética, Facultad de Farmacia,
Universidad de Concepción, casilla 237, Concepción, Chile.


An instrumental planar chromatography method for quantitative analysis of carbamazepine in saliva was developed. The linearity, precision, detection and quantification limits, accuracy and selectivity of the method were validated. The method was linear between 0.50 and 15.00 ng/spot, with a regression coefficient of 0.999. The intra-assay variation (repeatability) was between 5.1% and 7.4%, and the inter-assay (reproducibility) was between 5.6% and 7.4%. The detection limit was 0.18 ng, and the quantification limit was 0.54 ng. The method proved accurate, with a recovery percentage of 109.8%, and it was selective for the active principle tested. In conclusion, is a good method for the quantitative determination of carbamazepine in saliva.

Key Words: carbamazepine, HPTLC (high performance thin layer chromatography or instrumental thin layer chromatography or instrumental planar chromatography), quantitative analysis, saliva, analytical parameters, drugs, epilepsy.


Epilepsy is a chronic disorder of the central nervous system characterized by recurrent, unpredictable and self-limited seizures. Periodically, neurons depolarize abnormally causing paroxysmal alteration in motor, sensory, or behavioral activity (1-3).

Carbamazepine is one of the most widely used anticonvulsive drugs in the treatment of epilepsy (2-4), especially for the control of partial and tonicoclonic seizures (2,4,5).

This drug has a low solubility and its metabolism is self-induced or induced by other drugs as well. This produces a poor relationship between dosage and effect. Therefore, it is fundamental to monitor plasma concentrations of carbamazepine in order to obtain safe, non-toxic and effective therapeutic regimens (2,6,7,8). Therapeutic concentrations are reported to be 3 to 14 ug/mL (2,7,8) although considerable variation occurs. Side effects related to the CNS are frequent at concentrations above 9 ug/mL (2).

This study deals with the use of saliva as a biologic fluid in the analysis of carbamazepine, using instrumental planar chromatography (HPTLC) as the analytical technique.

Saliva is a fluid that can be easily obtained with the aid of non-invasive methods, an aspect of crucial importance in the studies of children, elderly patients and pregnant women, as well as in all situations where samples need to be obtained without exposing a patient to discomfort, skin irritation and risk of infection. The use of saliva instead of blood is a help for the adherence of the patients to their controls.

Furthermore, the rationale for measuring drug concentration in saliva is that the pharmacological action depends on the unbound fraction of the drug in the plasma; this fraction is usually excreted by the salivary gland into the saliva (8-10).

Carbamazepine is non-ionized at the pH range of saliva, an important factor in the analysis of this fluid (8). Also, this drug is widely distributed throughout the body in areas such as the brain, duodenal fluids, bile and saliva (7). The reported saliva concentrations are between 1.2 and 3.5 ng/uL (11), and the plasmatic concentrations are correlated with those in saliva (7).

HPTLC is a technique carried out within a short period of time, requires few mobile phases and allows for the analysis of a large number of samples simultaneously. The ability of HPTLC to analyse many samples in parallel has the advantage over other techniques in that the separation of 10 or 20 samples only takes the same amount of time as the separation of 1 sample.

Amounts on the order of nanograms (UV) and smaller than picograms (fluorescence) can be detected (12).



In this study, USP standards of carbamazepine, carbamazepine-epoxide and carbamazepine-dihydroxide were used. All of the reagents were of pro-analysis quality (Merck).

Carbamazepine solutions were prepared in ethanol (13). Saliva of volunteer donors was used for the recovery study.


Chromatography was carried out on silica gel F254 HPTLC plates, previously washed in methanol and activated at 130 °C during 20 minutes.

Sample application was carried out on 3 mm bands using a semiautomatic Linomat IV device.

For the chromatographic development, ethylacetate/toluene/methanol (5/4/1, v/v/v) was used. The length of development was 5cm ocurring in a period of 10 minutes. Horizontal development chambers were used.

After development, carbamazepine fluorescence activation was obtained by immersion of the plates in 60% HClO4 in ethanol/water (1/1, v/v) and then placing the plates on a heater at 120 °C for 7 minutes (14).

Densitometry readings were carried out using a Scanner Camag III spectrodensitometer assisted by a computer equipped with CATS version 4.05, and a mercury lamp was used as the radiation source. Determinations were performed at a wavelength of 366 nm.


The parameters validated were linearity, detection limit, quantification limit, precision, accuracy and selectivity.

Validation was performed in compliance with international standards (13,15,16), using adequate statistic estimates (relative standard deviation, Student t, variance analysis) (15-16).


To establish the linearity order of the method, a calibration curve was prepared by applying the standard concentrations between 0.50 and 15.00 ng for spot. Each solution was applied in triplicate.


In order to determine detection and quantification limits, analyte concentrations in the lower part of the linear range of the calibration curve were used. Concentrations solutions of 0.50, 1.00 and 2.00 ng/uL were prepared, each of them being applied 3 times (1 uL every time).

The amount of analyte by spot versus average response (peak area) was graphed and the equation for this curve was determined, thereby obtaining an estimate of the target response: ybl (15). The ybl value corresponds to the intersection of the curve. Subsequently, a second curve was graphed showing the amount of analyte by spot versus standard deviation of the responses. From the equation of this curve, we obtained an estimate of the standard deviation for target: sbl (15), which corresponds to the intersection of this curve. Detection and quantification limits were calculated by means of the equations (15): detection limit = (ybl + 3sbl)/b; quantification limit = (ybl + 10sbl)/b, where "b" corresponds to the slope obtained in the linearity study of the method.


Carbamazepine solutions of 1.00, 5.00 and 15.00 ng/uL were used in this study. These cover the linear range determined (applying 1 uL).

To assess the repeatability of the analytical method, five solutions of every concentration were prepared, each of them being applied one time (13,15) Repeatability of the system, however, was done using one solution of every concentration and each was spotted in triplicate (15). All applications were made on the same plate, by the same analyst, on the same instrument and on the same day.

To assess reproducibility, one solution of every concentration was applied five times during five days, using a different plate each day.


Known amounts of carbamazepine was added to drug-free saliva in order to obtain solutions of 0.50, 2.00, 5.00 and 15.00 ng/uL.

Accuracy was evaluated by means of recovery assays. The solutions were centrifuged at 5000 rpm and the supernatant was used. The recovery of the active principle was determined by comparison of a prepared concentration and an obtained concentration. Each solution was applied three times in order to obtain the mean recovery. The obtained concentration was calculated from the equation of concentration versus peak areas graph.


To assess selectivity of the method, a carbamazepine solution (CBZ) plus carbamazepine-10,11-epoxide and carbamazepine- 10, 11- dihydroxide was prepared. The latter two are the main metabolites of carbamazepine and can therefore be found in saliva samples. Carbamazepine epoxide is as active as the parent compound, and its concentration in plasma and brain may reach 50% of those of CBZ (2). The solution was prepared at a concentration of 0.20 ug/uL of CBZ, 0.10 ug/uL of CBZ-10,11- epoxide, and 0.30 ug/uL of CBZ-10,11-dihidroxyde.


Instrumental planar chromatography method for analysis of carbamazepine in saliva proved to be linear between 0.50 and 15.00 ng/spot. This range include to the expected carbamazepine concentrations in saliva (1.2 - 3.5 ng/uL) (11). The calibration curve for carbamazepine had a regression coefficient of 0.999, (Figure 1) indicating that a linear relationship exists between the concentration of carbamazepine and peak area over the range investigated.

Fig.1. Calibration curve for the determination of carbamazepine by the proposed HPTLC method.

The sensitivity of this method is adequate for the detection and quantification of carbamazepine in saliva, since the detection limit was 0.18 ng and the quantification limit was 0.54 ng. Figure 2 shows the curve for the amount of analyte by spot versus average response (peak area) in this study. The regression coefficient of this curve was 0.999.

Fig.2. Quantity of carbamazepine by spot versus means of peak areas for each concentration. Curve for determination of detection and quantification limits. (ybl = 21.44)

Precision analysis studies showed an intra-assay variation between 5.1% and 7.4% and an inter-assay variation between 5.6% and 7.4%. These values are adequate for the analysis of drugs in biological fluids (15).

The average recovery percentage (n = 10) was 109.8. The student's test allowed us to conclude that there were no significant differences between average recuperation and 100 (t observed = - 1.53 and t tabulated = 1.83). The degrees of freedom were 9 and a = 0.05.

The method allows separation of carbamazepine from its main metabolites with a resolution of 2.5 for the peaks of carbamazepine-dihydroxide and carbamazepine-epoxide, and 1.5 for the peaks of carbamazepine-epoxide and carbamazepine. (Figure 3).

Fig.3. Selectivity of the method: peak n1 = carbamazepine-dihydroxide; peak n2 = carbamazepine-epoxide; peak n 3 = carbamazepìne.

The fact that no extraction is required gives advantages that need not be emphasized. The very small amounts of sample required for the assay make the method ideally suitable for the study of carbamazepine levels in body fluids that usually can be obtained only in small amounts (e.g., saliva).

Although we made an exhaustive search at libraries and the internet, we were unable to find other studies on carbamazepine similar to ours, however we found some investigations done on drug determination in saliva involving other drugs or other analytical methods (17-25).


The method proved to be linear, precise and accurate within the range of concentrations tested. Accuracy was optimal and remained close to 100%. The method showed high sensitivity, within the ng order, and was selective for the active principle studied, allowing for good separation of its metabolites. Therefore, we propose this method for the quantitative determination of carbamazepine in saliva.


[1] W. Garnett, Pharmacotherapy. 20, 191S-199S (2000)         [ Links ]

[2] J.Hardman, L. Lee. Goodman/Gilman, Las bases farmacológicas de la terapéutica, Mc Graw Hill Interamericana, México (1996)         [ Links ]

[3] J.Gorrel, U.S.Pharmacist. 75-84 (2002)         [ Links ]

[4] J. Delgado,W. Remers. Wilson/Gisvolds, Textbook of organic medicinal and pharmaceutical chemistry, Lippincott-Reaven, Philadelphia (1998)         [ Links ]

[5] N.Ryan et al, J.Am.Acad.Child Adolesc. Psychiatry. 38, 529-536 (1999)         [ Links ]

[6] C. Page et al. Farmacología Integrada, Harcout, Madrid (1999)         [ Links ]

[7] Drug Evaluations Annual 1995, American Medical Association, USA (1995)         [ Links ]

[8] S. Pichini et al, Clin. Pharmacokinet. 30, 211-228 (1996)         [ Links ]

[9] D.Malamud, BMJ. 305, 207 (1992)         [ Links ]

[10] M. Lifhitz, Z. Ben-Zui, R. Gorodischer, Ther. Drug Monit. 12, 334-338 (1990)         [ Links ]

[11] G.W. Rylance and T.A. Moreland, Archives of Disease in Chilhood.56, 637-640 (1981)         [ Links ]

[12] I. Wilson, Ther. Drug Monit. 18, 482-492 (1996)         [ Links ]

[13] The United States Pharmacopeia / The National Formulary (USP 25/ NF 20), United States Pharmacopeial Convection Inc, USA (2002)         [ Links ]

[14] J.Chromatography. 93, 238-242 (1974)         [ Links ]

[15] O. Quattrochi, S. Abelaira, R. Laba. Introducción a la HPLC. Aplicación y Práctica, Ed. Artes Gráficas Farro S.A, Buenos Aires (1992)         [ Links ]

[16] W. Daniel. Bioestadística, Base para el análisis de las ciencias de la salud, Uteha Noriega Editores, México (1996)         [ Links ]

[17] P.I.Tomlin, I.McKinlay and I.Smith, Development Medicine/Child Neurology. 28 (6), 713-718 (1986)         [ Links ]

[18] S.Pichini, I.Altieri, P.Zuccaro and R.Pacifici, Clin.Pharmacokinetic. 30 (3), 211-228 (1996)         [ Links ]

[19] N.A.Klitgaard and O.Kristensen, Eur.J.Cli.Pharmacol. 31, 91-94 (1986)         [ Links ]

[20] T.Ohkubo, M.Suno, M.Kudo, T.Uno and K.Sugawara, Ther.Drug Monit. 18(5), 598-603 (1996)         [ Links ]

[21] J.H.H.Thijssen, C.C.Gispen-De Wied, G.M.van Heeswijk and W.Veeman, Clinical Chemistry. 42(8), 1238-1242 (1996)         [ Links ]

[22] J.H.Rojas, I.Riberos, M.Martínez and C.Troncoso, Libro de Resúmenes IX Congreso Latinoamericano de Cromatografía y Técnicas Afines, 79-80 (2002)         [ Links ]

[23] M.Lifshitz, Z.Ben-Zvi and R.Gorodischer, Ther.Drug Monit. 12, 334-338 (1990)         [ Links ]

[24] H.P.Kruger, E.Schulz and H.Magerl, 13th Internacional Confer ence on Alcohol, Drugs and Traffic Safety. 1, 62 (1995)         [ Links ]

[25] G.A.Starmer, D.J.Mascord, B.Tattam and R.Zeleny, Report CR141, Federal Office of Road Safety (1994)         [ Links ]


The authors thank the Direction of Investigation at the University of Concepción, through Project DIUC N° 201.074.026-1.0.

Creative Commons License Todo el contenido de esta revista, excepto dónde está identificado, está bajo una Licencia Creative Commons