Valproic acid determination by liquid chromatography coupled to mass spectrometry (LC–MS/MS) in whole blood for forensic purposes

Abstract Valproic acid (VPA) is a well‐known drug prescribed as anti‐epileptic. It has a narrow therapeutic range and shows great individual differences in pharmacodynamics and pharmacokinetics. Consequently, the therapeutical drug monitoring (TDM) in patient's plasma is of crucial importance. Liquid chromatography coupled to mass spectrometry (LC–MS/MS) has gained importance in TDM applications for its features of sensitivity, selectivity and rapidity. However, in case of VPA, the LC–MS/MS selectivity could be hampered by the lack of a sufficient number of multiple reaction monitoring (MRM) transitions describing the molecule. In fact, the product ion scan of deprotonated molecules of VPA does not produce any ion and thus most LC–MS/MS methods are based on the detection of the unique MRM transition m/z 143➔143. In this way, the advantages of selectivity in LC–MS cannot be effectively exploited. In the present method, stable analyte adducts were exploited for the determination of VPA in blood. An Acquity HSS C18 column and mobile phases consisting of 5‐mM ammonium formate and acetonitrile both added 0.1% formic acid were used. Source worked in negative acquisition mode and parameters were optimized to increase the adduct (m/z 189) and dimer (m/z 287) stability, and their fragmentation were used to increase the selectivity of MRM detection. The method has been validated according to the toxicological forensic guidelines and successfully applied to 10 real blood samples. Finally, the present method showed suitable for the rapid LC–MS/MS detection of VPA in whole blood, demonstrating the possibility to increase specificity by exploiting stable in‐source adducts. This should be considered of utmost importance in the case of forensic applications.


| INTRODUCTION
Valproic acid (2-propylpentanoic acid or VPA) is a short-chain fatty acid derivative prescribed as anti-epileptic, mood stabilizer, and for migraine prophylaxis. 1 Moreover, VPA is used in the treatment of alcohol and substance withdrawal and dependence. 2 It is commonly administered as sodium salt, sodium valproate, or a mixture of its acidic form and salt, showing high oral bioavailability and high binding to plasma proteins, with a plasma/whole blood ratio around 1.8. 3 VPA has a narrow therapeutic range (50-100 μg/ml) and shows great individual differences in pharmacodynamics and pharmacokinetics. Consequently, the therapeutical drug monitoring (TDM) in patient's plasma is of crucial importance. From the analytical point of view, many methods for VPA determination in plasma exist in literature, and beside immunometric methods commonly used in clinical settings, high-performance liquid chromatography (HPLC-UV) 4 and gas chromatography-mass spectrometry (GC-MS) 5 have been gaining importance in recent years. Moreover, liquid chromatography coupled to mass spectrometry (LC-MS/ MS) [6][7][8] has gained importance in TDM method development for its features of sensitivity, selectivity, and rapidity. However, in case of VPA determination, the LC-MS/MS selectivity could be hampered by the lack of a sufficient number of multiple reaction monitoring (MRM) transitions describing the molecule. In fact, the product ion scan of deprotonated molecules of VPA does not produce any ion, and thus most LC-MS/MS methods are based on the detection of the unique MRM transition m/z 143à143. 6,9,10 In this way, the advantages of MRM detection cannot be effectively exploited. To overcome this limitation, adducts formation between VPA and mobile phase components as acetate and acetic acid have been described. 10,11 Adduct formation is based on complicated equilibria reactions at the source level, and these adducts, if noted, are usually not considered because of low reproducibility. However, some authors 12   Stock solutions of VPA were prepared in methanol. Internal standard working solution was prepared at the final concentration of 20 μg/ml in methanol. All solutions were stored at À20 C and left at room temperature 2 h for equilibration prior use.

| LC-MS conditions
The LC-MS system consisted of an UHPLC Water Acquity coupled to a Xevo TDQ triple quadrupole mass spectrometer (Waters, Milford, MA), equipped with an Acquity UPLC HSS C18 column  Table 1.

| Method validation
The method was validated following the forensic toxicology guidelines by considering selectivity, sensitivity, linearity accuracy, precision, matrix effect, and stability. 15 Selectivity was assessed by analyzing eight drug-free blood samples, six without IS and two with IS. Matrix-matched calibration curve was set in the range of 5-800 μg/ml (5-10-50-100-200-400-600-800 μg/ml). Three QCs at 25-75-500 μg/ml were also prepared in whole blood. Accuracy and precision were evaluated by analyzing QC samples in five replicates on four non-consecutive days. Accuracy and precision were obtained as bias (%) and relative standard deviation (RSD %).
Sensitivity, expressed as limit of detection (LOD) and limit of quantification (LOQ), was determined by a signal-to-noise ratio of 3 and 10, respectively, and experimentally verified by spiking the calculated amount. To obtain an average evaluation of the matrix effect over the entire quantification range, expressed as average bias %, the slopes of the calibration curves prepared in water and blood (n. 10 samples) were used. 16,17 Stability was assessed on QCs after 24 h storage at À20 C and calculating percent deviation on freshly prepared QCs.   Figure 2).
As can be noted, a discrepancy from therapeutic plasma/serum concentrations was observed, even in consideration of the plasma/blood ratio. These data were already observed in VPA analysis on postmortem blood samples, and stability of VPA in stored blood was considered the major issue. 19 To overcome VPA stability issue, the alternative way of collection by means of dried blood spots had been investigated by some authors with promising results. 9,20,21 Unfortunately, this modality of collection could not be exploited in our cases, but it represents an interesting further application of the present method. Moreover, in the future, the advantage of working with mobile phases commonly used for drug analysis in positive ion mode could productively be exploited by integrating the analysis by adding more anti-epileptic drugs to the same run.