An Improved GC Method for Rapid Analysis of Valproic Acid in Human Plasma Without Derivatization

1. Introduction

     Epilepsy is the most common serious neurological disorder affecting people of all ages. Statistics show that epilepsy affects 1% of the world population [1]. Valproic acid (2-propylpentanoic acid, VPA) has been widely used in the control of a variety of seizure types as a major antiepileptic drug [2, 3]. Determination of antiepileptic drugs in body fluids on a routine basis is important to clarify and control their therapeutic and toxic effects and to assess the patient's compliance to therapy [4-8].

     Reported methods for analysis of VPA in serum and plasma include immunological techniques, chromatographic approaches [9-14] and techniques based upon capillary elec-trophoresis (CE) [15]. Most of the chromatographic methods require long sample preparation steps and prior derivatization [16]. Chromatography of the analytes without prior derivatization would significantly simplify the method and thus shorten the analysis time.

     In the present study, we describe a simple and rapid method with improved sensitivity and reproducibility for sample preparation and GC analysis of VPA using a Stabilwax^®-DA column with satisfactory performance.

2. Materials and methods

2.1. Chemicals

     Valproic acid was obtained from Sigma (St. Louis, MO, USA). Perchloric acid (70%), chloroform and octanoic acid (analytical grade), methanol (HPLC grade) were purchased from Merck (Darmstadt, Germany). Laboratory prepared distilled water was used throughout the experiments. All reagents were used without further purification.

2.2. Apparatus

     The gas chromatograph was a Hewlett-Packard (HP) model 6890 equipped with a flame-ionization detector (FID). For instrumental control, data collection and processing GC- Chemstation software was employed. The column was a Stabilwax^®-DA, 15 m×0.25 mm I.D.×0.25 m film thickness (RESTEK, Buckinghamshire,UK).

Table 1. Valproic acid in plasma, within-day assay precision, accuracy.

 ^aAverage recovery calculated with 5 samples.

^b%RE=[(mean value/theoretical value)-1]´100

The deviation of the mean from the nominal value serves as the measures of accuracy.

2.3. Standard solutions

      Stock standard solutions (5 mg/ml) of VPA and (1 mg/ml) octanoic acid (internal standard; IS) were prepared in methanol and stored at -20 °C. Working standard solutions of VPA (1-500 μg/ml) were prepared from the stock standard solution in methanol and then aliquots of each of these solutions were added to the separated micro tubes .The spiked plasma samples were prepared freshly every day by adding 500 μl blank plasma to these tubes after solvent evaporation under nitrogen stream at the room temperature.

2.4. Processing of plasma samples

     After addition of 20 μl of IS (1 mg/ml) and vortex for 1 min., 100 μl of perchloric acid (12%, w/w) and 250 μl of chloroform were added to each sample. The tubes were shaken for 5 min. vigorously and then centrifuged at 12000 rpm for 5 min. Then the top aqueous phase was removed and discarded. The chloroform layer was transferred to another small tapered vial (ca. 200 μl was recovered). After evaporation of solvent under nitrogen stream and reconstitution with 50 μl of methanol, 3 μl of the sample was injected to GC inlet.

Figure 1. Chromatograms of  A) plasma (blank), B) plasma spiked with 40 mg/ml octanoic acid, C) plasma spiked with 150 ng/ml VPA, and D) plasma spiked with 20 mg/ml VPA.

2.5. Gas chromatography conditions

     Nitrogen was used as carrier gas at a flow rate of 3 ml/min. in a constant flow mode. The make up gas was also nitrogen. Splitless injection was used. The initial oven temperature was 100 °C (held for 2 min.) followed by an increase to 115 °C at a rate of 5 °C/min. (0 min. hold time) and at 0.6 °C/min. increase to 117 °C (1.5 min. hold time) and finally at 30 °C/min. to 220 °C (5 min. hold time). Optimum detector and injector temperature was 300 and 250 °C, respectively.

Table 2. Valproic acid in plasma, between-day assay precision, accuracy. 

^aAverage recovery calculated with 5 samples.

^b%RE=[(mean value/theoretical value)-1]´100

The deviation of the mean from the nominal value serves as the measures of accuracy.

3. Results

3.1. Stability and recovery

     Stability of stock standard solutions at +4 °C has been verified during a month. Stability of 10 μg/ml spiked plasma at -20 °C has been verified for VPA and octanoic acid during the sample storage period (two weeks). The extraction efficiency (recovery) was determined for both VPA and octanoic acid by comparing peak areas from drug-free plasma spiked with known amounts of these compounds (in the range of concentrations of the calibration curves) and standard solutions in methanol injected directly into the GC inlet (Figure 1). Each sample was determined in triplicate. The results are presented in Tables 1 and 2. The extraction efficiencies for VPA and IS were 70-85%.

3.2. Linearity, precision and accuracy of the method

     The method within and between-day precision in human plasma were assessed by performing five determinations per concentration of spiked plasma with VPA (at six levels) on the same and different days (n=5). The results are presented in Tables 1 and 2, respectively. Precision and accuracy were characterized by the RSD% and deviation from the nominal concentration, respectively. Linearity of the calibration curve was investigated in the 0.45-100 μg/ml concentration range, using spiked plasma samples. The calibration equation was calculated from the peak area values.

4. Discussion

     So far, various GC methods have been used for determination of VPA in plasma, most of these methods have benefited from derivatization techniques to improve chro-matographic characteristics and LOD of VPA but because of its multistep procedure and nature of chemical reactions, derivatization is not a good choice for rapid and robust quantitative analysis which are essential characteristics in TDM and clinical toxicological analysis. The aim of this study was to develop a selective and sensitive quantitative analytical method for rapid determination of VPA in human plasma. In order to determine the trace level of drugs, it is advantageous to eliminate any possible interference from samples. Using perchloric acid with chloroform to precipitate the plasma proteins and to extract the analyte in one step shorten the analysis time. By using a high polarity solvent such as methanol for reconstitution after previous non-polar solvent, we could gain a cleaner and more concentrate extract at the same time. After analysis of more than three hundred samples, there was no need to clean the sample inlet, and there was no significant increase in instrument noise. The chromatograms obtained after extraction of plasma blank, plasma spiked with 150 ng/ml, 20 μg/ml of VPA are shown in Figure 1. The chromatograms indicate that even in low levels, the target compound is separated from the interferences of the biological extract. The calibration curve for detection of VPA was obtained by performing a linear regression analysis on spiked plasma samples using the ratio of VPA to IS signal area. Good linearity was obtained with correlation coefficient of r > 0.99. According to calibration calculations, the equation of the curve fitted to the calibration points were: Y= 0.026X-0.015 (r² = 0.998). The detection limit of the method for VPA ranged from 120 to 180 ng/ml which was estimated at a signal-to-noise (S/N) ratio of 3. Compared with the previous published methods, the LOD of this method is better. The limit of quantification (LOQ) estimated by measuring the response of the calibration curve samples between run and defined as the concentration which yields an S/N equal to 10, was determined to be 450 ng/ml. The reproducibility of this method was represented by the percentage of the relative standard deviation (RSD) at each fortification level for VPA. The results show that the precision of the method was within 10% which is very satisfactory. The within-day reproducibility ranged from 3.14 to 8.70% and between-day reproducibility ranged from 4.11 to 9.50%.

5. Conclusion

     Although there are many GC methods in the literature for VPA analysis in human plasma, a few of them could achieve sensitivity and selectivity similar to this method but with more complicated sample preparation. By using these optimized sample preparation, column length and GC temperature programming, good resolution and sensitive determination was achieved in human plasma. We developed and validated a rapid and precise method with sufficient selectivity and sensitivity for determination of VPA in human plasma. The assay is based on fast sampling procedure which include protein precipitation with perchloric acid and liquid/liquid extraction with chloroform in one step, then reconstitution with methanol and finally direct injection to GC column without derivatization. The described method demonstrated to be suitable and sufficiently robust for TDM, toxicological and pharmaco-kinetic studies.