1. Introduction

About 70 million people worldwide suffer from epilepsy, a neurological disease that negatively affects patients’ quality of life and that of their families. Antiepileptic drugs (AEDs), despite many scientists’ and the medical community’s efforts, are still the basic tool in the treatment of patients with epilepsy, a disease of unknown etiology. To date, about thirty first–third generation AEDs are used in various types of epilepsy. The new AED cenobamate (CNB) was approved by the US FDA in 2019 to reduce uncontrolled partial-onset seizures in adults ^[1][2][1,2]. AEDs possess a long tradition in the treatment of different modes of epilepsy and over the decades many analytical methods have been developed for the therapeutic monitoring of AEDs. More improved liquid chromatography (LC) methods have also appeared in the last years, based mainly on sensitive mass spectrum detection. Generally, the elaborated methods are focused on determining many AEDs and their metabolites in one analytical run. The main problem of recent years has been developing procedures for sample micronization and automatic preparation of clinical material for quantitative determination. Microextraction techniques like solid-phase microextraction (SPME), single-drop microextraction (SDME), or dispersive liquid-liquid microextraction (DLLME) were elaborated for the isolation of many medicines, including AEDs, from biological matrices.

Therapeutic Drug Monitoring (TDM) of AEDs involves measuring the drug concentrations in blood, serum, or plasma to improve the efficacy of applied treatment according to known rule if you can’t measure it, you can’t improve it. TDM was introduced in the late 1960s to minimize the toxicity effect caused by aminoglycosides. The main aim of TDM is to optimize drug exposure to minimize toxicities and maximize efficacy. The therapeutic tool is particularly essential when a drug that has a narrow therapeutic window and is characterized by a significant correlation between drug concentration and its toxicity, and when the clinical results depend on the drug level—total and/or free in the blood, not on the dose taken [3]. AEDs are monitored mainly due to the high inter-individual variability resulting from non-linear pharmacokinetics and the narrow therapeutic scope. New reference ranges were established for total and free concentrations of AED’s [4]. Essentials are interactions of AEDs in the pharmacokinetic phase. Carbamazepine (CBZ), oxcarbazepine (OXC), vigabatrin (VGB) are indicated for epileptic patients with focal seizures. Clobazam (CLB) for myoclonic episodes and valproic acid (VPA) are dedicated to generalized epilepsy. Levetiracetam (LEV), lamotrigine (LTG), rufinamide (RFM), topiramate (TPM), and zonisamide (ZSM) with a broad spectrum of action are used for most types of seizures. They require monitoring also due to the rapid absorption process, short half-life or changing with dose (nonlinearity) and significant changes in concentrations during the day or the degree of binding with blood proteins [5].

Recommendations for TDM in the treatment of epilepsy are the following ^[5][6][7][5,6,7]:

• small differences between the toxic effect of the drug and the symptoms of the disease
• doses of the drug that do not relieve symptoms
• patient belongs to one of the groups at increased risk: Elderly, children, pregnant, patients with renal and hepatic dysfunction
• required changes in the dosage of a given drug when the patient is in weak condition
• the use of polytherapy of AED’s drugs
• a drug with dose-dependent—non-linear pharmacokinetics.

Currently, a great benefit for patients using various AEDs is adjusting the dosage to their individual needs and monitoring the sufficient concentrations. To correctly interpret the results obtained during TDM, pharmacokinetics knowledge on absorption, distribution, biotransformation, drug excretion and an understanding of the physiological, pathophysiological and environmental factors affecting the therapy’s success are necessary ^[6][8][6,8].

Initially in TDM colorimetric methods were used, and later more modern techniques were implemented, such as immunochemistry in the early 1980s, or LC. In the 2000s, the LC-MS method allowed for the quantification of free drug concentrations in biological matrices, mainly in plasma and interstitial fluids, with very sensitive detection [9]. The control of drug concentrations has provided a new perspective in the treatment in line with pharmacokinetic principles and clinical observations [7].

The matrix for TDM is often plasma or serum, rarely urine through, and increasingly unstimulated saliva, DBS or dried saliva spot (DSS). For a reliable test, the concentration of free drug in the blood has to be measured that reflects effective levels in the brain, heart tissue. In most clinical trials, the ratio between free and bounded fractions is constant at the equilibrium state. Therefore, a measurement of the total concentration (protein-bounded and free form) is acceptable in drugs with low protein binding. Such analysis is more comfortable, cheaper and requires less time expenditure ^[5][10][5,10].

Sample preparation is a crucial step during the whole process of the quantitative determination of the clinical samples. Simple protein precipitation or different extraction procedures before injection into LC of monitored analytes is equally important as a separation technique before detection. The most useful method for extraction includes liquid-solid, liquid-gas or liquid-liquid extraction, and mechanical separation. It is also essential to precisely determine the degree of reliability of a given method using statistical parameters known as the validation process. Validation determines the suitability of analytical methods based on the evaluation of validation parameters, like selectivity, recovery, calibration curve, determination of the linearity range, sensitivity, accuracy, precision, limit of quantification (LOQ), and limit of detection (LOD). Designed for the TDM bioanalytical method has to fulfill validation criteria established mainly by the European Medicines Agency (EMA) [11] or the Food Drug Administration (FDA) [12].

2. Bioanalytical Methods

Methods that meet the validation criteria, including high sensitivity and selectivity, are the basis for TDM, and finally, for effective and safe pharmacotherapy. This statement also applies to AEDs, for which monitoring of drug concentrations becomes the rule nowadays, not only in terms of total concentration (VPA, PHT, CBZ, GBP, LTG, lithium, TPM, LEV) but also of unbound fractions, especially for those drugs with a high (90%) or even higher protein binding rate. Free concentrations of many AEDs including PHT, CBZ, and VPA are in some laboratories routinely determined. An increase in TDM is expected for the most recent AEDs for which the therapeutic concentration ranges have been established. This concerns BRV (0.2–2 mg/L), PER (0.1–1 mg/L), STM (2–10 mg/L) and STP (4–22 mg/L) [4]. For other AEDs: ESL, LTG, OXC, PGB, TPM, and VPA, the reference ranges have been updated or harmonized. It can be expected that free fraction will be determined for other AEDs: CLB, and its active metabolite N-CLB, PER, TGB, retigabine, STP, medicines with high protein binding ^[13][203].

There is still a need to develop new, rapid methods that meet the validation criteria. This trend has been observed in the last few years in the bioanalysis of AEDs, where LC-MS/MS is a dominated technique; however, it is costly. Data presented in the review show that near 50% of the LC methods applied the high-resolution detection despite the technical problems with the stability of MS/MS response. The sensitivity connected with the high-resolution LC-MS/MS causes that deuterated analogues of the AEDs are used as effective internal standards even with very small differences in molecular mass as well as retention time ^{[14][15][16][17][18][19][20][21][22][23][24][25][26][27][28]}[21,22,37,40,42,65,84,119,146,147,149,151,157,176,177]. There is still an increase in the full automatization of analytical determination of AEDs in order to decrease hands-on time as well as consumption of organic solvents, protect the natural environment and improve the analytical process in terms of accuracy, precision, repeatability, and sensitivity. To realize the purpose also different microextraction techniques for AEDs analysis were developed, including ultrasound-assisted emulsification microextraction ^[29][47], MEPS ^[30][31][32][66,75,167], microextraction combined with micro-derivatization-increased detection (MDID) ^[17][40], DLLME ^[33][34][59,133]. Automatization was applied for the DBS extraction of VPA, PHB, PHT, CBZ and its active 10,11 epoxide ^[14][21]. The effective online SPE coupled with an analytical column of LC-HRMS system was applied for quantification of the first generation of AEDs: PHB, PHT and CBZ and its active metabolite in clinical samples ^[35][23]. The interesting idea seems to be online extraction using RACNTs for analysis of CBZ, PHB and PRM, previously used only for extraction cadmium and lead [36].

However, traditional protocols with steps involving sample preparation using traditional LLE and newer SALLE ^[37][50], as well as protein precipitation, are frequently applied in the bioanalysis of AEDs, more often than advanced micro-extraction techniques.

Among the presented LC methods, the columns with different length, dimensions and particle sizes based on popular lipophilic chain C-18 are very often applied. Generally, in LC-MS/MS and UHPLC-MS/MS, the columns are characterized by dimensions of 2–3 mm, and smaller than 5 µm particle size (1.7–3.5 µm, Table S2). The other types of LC columns, like those containing biphenyl phase ^[38][88], chiral column Chiracel with normal phase ^[39][76], or columns with cyano stationary phase (polar phase) ^[40][121] were rarely used in quantification of AEDs. Less popular methods like MEKC gained attention for TDM of PIR ^[41][92], capillary electrophoresis for TPM and PRM ^[42][43][122,184]. Worth mentioning is that HPTLC, not so useful in TDM was also developed for determination of OXC and ESL in biological samples ^[37][50]. The GC-MS technique loses its importance compared to LC-MS/MS, although it has been developed in recent years to determine TPR ^[44][120], GBP ^[24][149], LCM ^[45][28][175,177].

Moreover, alternative specimen sampling is proposed employing non-invasive and patient-friendly techniques, including DBS or collection of saliva. These alternative specimens, appropriate for TDM, are especially valuable in specific clinical situations involving pediatric patients or critically ill patients.