Therapeutic Drug Monitoring of 6 New-Generation Antiseizure Medications Using a Mass Spectrometry Method: Analysis of 2-Year Experience in a Large Cohort of Korean Epilepsy Patients

Lacosamide, perampanel, pregabalin, gabapentin, vigabatrin, and rufinamide calibrators (3PLUS1 Multilevel Plasma Calibrator Set Antiepileptic Drugs/EXTENDED), and quality control (QC; MassCheck Antiepileptic Drugs/EXTENDED Plasma Controls) materials were purchased from Chromsystems (Munich, Germany). ClinCal Serum Calibrator for Antiepileptics 5 (level 0–3) was purchased from RECIPE Chemicals + Instruments GmbH (Munich, Germany). For internal standards (ISs), lacosamide-d3, perampanel-d5, gabapentin-d6, vigabatrin-d3, and rufinamide-d2 were purchased from Toronto Research Chemicals, and pregabalin-d4 was purchased from TLC Pharmaceutical Standards. LC-MS/MS-grade acetonitrile, distilled water, and methanol were purchased from Burdick and Jackson Brand (Muskegon, Michigan). Formic acid and ammonium acetate were purchased from Sigma-Aldrich (St Louis, Missouri).

Five levels of calibrators were prepared by dissolving the calibrators (3PLUS1 Multilevel Plasma Calibrator Set Antiepileptic Drugs/EXTENDED) into drug-free blank serum included in the Chromsystems calibrator set, achieving final concentrations as shown in Supplemental Table 1 (see the supplemental digital content, containing 3 tables, at https://meridian.allenpress.com/aplm in the January 2025 table of contents). QC samples were prepared at low and high concentrations according to the manufacturer’s guidance. The concentrations of IS solutions were prepared at 0.1 µg/mL for gabapentin, lacosamide, rufinamide, and vigabatrin, 0.2 μg/mL for pregabalin, and 1.0 ng/mL for perampanel in methanol solution (methanol to acetonitrile 1:1).

A 10-μL patient sample, calibrators, and quality control materials were protein precipitated by treatment with methanol and acetonitrile, and the mixture was added to 200 μL of IS working solution. The mixture was then vortex mixed for 5 seconds and centrifuged at 15 000 rpm for 10 minutes, and the final protein-precipitated sample was analyzed by LC-MS/MS with an injection volume of 1 μL.

LC-MS/MS analysis was performed on an Agilent 6460 triple quad mass spectrometer with an Agilent 1260 high-performance liquid chromatography system (Agilent Technologies, Santa Clara, California). Chromatographic separation was performed on an Agilent Poroshell 120 EC-C18 column (3.0 × 50 mm, 2.7 µm) using distilled water containing 0.1% formic acid and 2 mM ammonium acetate (mobile A) and acetonitrile containing 0.1% formic acid (mobile B) as mobile phases. The flow rate was 0.5 mL/min.

Quantitative analysis was performed in multiple reaction monitoring mode with a jet stream electrospray ionization source operating in positive-ion detection mode (quantifiers: lacosamide, m/z 251.1→108.1; lacosamide-d3, m/z 254.2→108.1; perampanel, m/z 350.1→247.1; perampanel-d5, 355.2→248.1; pregabalin, m/z 160.1→142.1; pregabalin-d4, 164.2→146.1; gabapentin, m/z 172.1→154.1; gabapentin-d6, 178.2→160.1; vigabatrin, m/z 130.1→71.1; vigabatrin-d3, 133.1→74.1; rufinamide, m/z 239.0→127.0; and rufinamide-d2, 241.2→129.0; qualifiers are summarized in Supplemental Table 2). Other settings for the MS were as follows: gas flow 12 L/min at 350°C, capillary voltage 3500 V, nebulizer pressure 413.7 kPa, and sheath gas flow 12 L/min at 350°C. Figure 1 shows a typical chromatogram. Quantification was performed using the ratio of the integrated peak area of analytes to that of IS using MassHunter Workstation software (version B.06, Agilent Technologies).

Since 2019, we have participated in external proficiency testing of the LGC proficiency testing programs for all 6 ASMs, which included up to 46 institutes, and CAP programs for lacosamide, pregabalin, gabapentin, and rufinamide, which included up to 26 institutes.

Samples referred for serum concentration measurement of lacosamide, perampanel, pregabalin, vigabatrin, gabapentin, and rufinamide were retrospectively identified at Samsung Medical Center (Seoul, Korea) between August 2019 and June 2021. Clinical information of age, sex, body weight, underlying diseases, blood sampling time, drug regimen, times of start and discontinuation of drug, and concomitant antiepileptic drugs were collected from patient medical records. Among the concentrations, only steady-state concentrations were included by selecting those measured after 5 half-lives of each drug. Blood samples were collected 10 to 12 hours after the last dosing. All variables except sex, age, and type of epilepsy were analyzed regarding each concentration as an independent case. The concentration-to-dose (CD) ratio was calculated by dividing the serum concentration (µg/mL) by the dose adjusted for body weight (mg/kg). Lastly, use of concomitant ASMs was categorized as follows^19–22 :

1. Enzyme inducers (phenytoin, carbamazepine, phenobarbital, or primidone), even in combination with other ASMs mentioned below.

2. Non–enzyme inducers, also including weak enzyme inducers/inhibitors (oxcarbazepine, perampanel, topiramate, lamotrigine, levetiracetam, zonisamide, clobazam, clonazepam, or valproic acid) without the use of any ASM categorized in 1.

The Samsung Medical Center Institutional Review Board (SMC 2020-05-182) waived the need for informed consent for this study.

After checking for normality, data showing a nonnormal distribution were reported as the median and range, and data showing a normal distribution were reported as mean with SD. Considering the nonparametric characteristics of the data, Spearman rank correlation was used for correlation analysis, and Wilcoxon rank sum test was used for comparison of CD ratio between the 2 antiseizure comedication subgroups. Linear regression was used for regression analysis in Figure 2. A P value <.05 was considered statistically significant. Statistical analyses were performed using Medcalc version 77.5.1.0 (MedCalc Software Ltd, Ostend, Belgium), IBM SPSS Statistics version 25 (IBM, Armonk, New York), and R 4.0.3 (R Foundation for Statistical Computing, Vienna, Austria).

As shown in Figure 1, the elution times for lacosamide, perampanel, pregabalin, gabapentin, vigabatrin, and rufinamide were less than 4 minutes, and no interfering peaks were observed. The total analytic run time was 5 minutes per sample, and the sample preparation time was less than 10 minutes for a maximum batch size validated up to 100 injections. Peak area ratios of analyte and IS were used for quantification.

The validation results are summarized in Table 1. The method exhibited a good linearity for all analytes. The deviation from the assigned value for each of the 5 calibration standards of each metabolite was within 15% (±20% for LLOQ) and the coefficient of determination (R²) was higher than 0.995 for all 6 ASMs. LOD and LLOQ ranges were 0.001 to 0.03 µg/mL and 0.01 to 0.8 µg/mL, respectively. Intraday and interday imprecision values were less than 5.0% coefficient of variation. The accuracy using QC materials ranged from 95.7% to 105.0%, and the accuracy using calibrators and proficiency test materials ranged from 85.2% to 111.4%. Results of external proficiency testing provided by CAP and LGC were all acceptable (representative results summarized in Supplemental Table 3). No interfering signals or false-positive results were observed during analysis of the 10 selectivity test samples. No interference from commonly prescribed antidepressants or neuroleptics was observed. Moreover, no interfering peaks have been observed in any patient samples to date. No carryover effect was observed. Extraction recovery ranged from 86.0% to 100.1% for analytes and from 85.9% to 98.8% for ISs. Matrix effect ranged from 87.3% to 113.5% for analytes and from 81.7% to 109.8% for ISs.

Patient characteristics are presented in Table 2. A total of 458 serum samples were obtained from 363 Korean epilepsy patients for TDM of 6 ASMs. Among the 6 ASMs, lacosamide (239 samples from 208 patients) and perampanel (139 from 131 patients) were the 2 most monitored drugs, whereas vigabatrin (25 from 7 patients) and gabapentin (8 from 8 patients) were the 2 least monitored. Median age was 35 years (range, 3–80 years), with pediatric (age <18 years) and elderly patients (age ≥65 years) accounting for 3% (n = 15) and 5% (n = 21), respectively. There was no case of a pregnant patient. A total of 88% of the patients were receiving more than 1 ASM (403 of 458), and a median of 2 ASMs were used in cases of polypharmacy (range, 2–11). Thirty-two patients used more than 1 of the 6 analyzed ASMs, 69% (22 of 32) of whom were using lacosamide and perampanel concomitantly. Among the patients on ASM polypharmacy, 29% (118 of 403) were using concomitant CYP enzyme inducers. The most commonly used concomitant ASMs were levetiracetam, valproic acid, and clobazam. Serum drug concentration and CD ratio were highly variable in all 6 drugs. When evaluated according to the therapeutic range suggested in the literature,^4,23,24  only 38% (175 of 458) of the concentrations were in the therapeutic range, whereas 53% (244 of 458) were subtherapeutic.

Among patients taking lacosamide and perampanel, 16% (38 of 239) and 14% (20 of 139) were taking concomitant CYP enzyme-inducing ASMs, respectively. In patients using lacosamide, levetiracetam (41%; 98 of 239) and valproic acid (34%; 81 of 239) were the 2 most commonly used concomitant ASMs. In patients using perampanel, valproic acid (40%; 55 of 139) and oxcarbazepine (40%; 29 of 139) were the most commonly used concomitant ASMs. Figure 2, A and B, shows the correlation between weight-adjusted daily dose and serum concentration according to the use of concomitant enzyme inducers in patients using lacosamide and perampanel, respectively. Both serum lacosamide and perampanel concentration increased with dose. However, the correlation was weak in perampanel (correlation coefficient r = 0.36, P < .001) compared with that in lacosamide (r = 0.732, P < .001). Although the slope was similar between cases with and without enzyme inducers in lacosamide, it was significantly greater in cases without enzyme inducers in perampanel. The median CD ratio was significantly lower in patients using concomitant enzyme inducers than in patients on monotherapy in both patients using lacosamide and perampanel, with greater difference in patients using perampanel (monotherapy versus with enzyme inducers, 1.64 versus 0.96 [µg/mL]/[mg/kg], P < .001 in lacosamide, Figure 2, C; 7.09 versus 1.85 [µg/mL]/[mg/kg], P < .001 in perampanel, Figure 2, D).

In this study, we developed and validated an LC-MS/MS method for simultaneous measurement of lacosamide, perampanel, pregabalin, vigabatrin, gabapentin, and rufinamide and successfully applied it to routine clinical practice. In a clinical perspective, our study opens important opportunities for the improvement of the management of patients using these ASMs by allowing a more rational approach to dose adjustment.

Currently, LC-MS/MS is the most preferred method for measurement of ASMs because of its high sensitivity and specificity with reduced sample volume and short run time.¹⁶  Previously, 2 UPLC-MS/MS methods for simultaneous measurement of multiple ASMs have been described.¹⁷  One method simultaneously measured 22 ASMs; however, matrix effect could not be compensated because ISs were not used. The other method simultaneously measured levetiracetam, lamotrigine, lacosamide, primidone, topiramate, and zonisamide with stable isotope-labeled ISs. Although levetiracetam, lamotrigine, topiramate, and zonisamide are analyzed using LC-MS/MS in many laboratories, those that measure lacosamide, perampanel, pregabalin, gabapentin, vigabatrin, and rufinamide are limited.^13,14  Methods that individually measure these drugs have been described in previous literature.^25–29  However, as polypharmacy is common in ASMs, simultaneous measurement is clinically more efficient. Moreover, our method used high-performance liquid chromatography instead of UPLC while demonstrating similar analytic run time (4.4 minutes) and injection volume (1 µL).¹⁷ 

With the TDM data obtained using this method, we analyzed the pharmacokinetic characteristics of the 6 ASMs in the Korean population. The weight-adjusted daily doses were smaller than those reported in previous studies for all 6 antiepileptic drugs.^{23,24,30–33}  In particular, in patients using pregabalin, vigabatrin, or gabapentin, the weight-adjusted daily dose was nearly half of those in previous reports from Spain and Germany (mean weight-adjusted daily dose from the present versus previous study; vigabatrin 36.8 versus 60.8 mg/kg/day, gabapentin 14.5 versus 35.9 mg/kg/day, and rufinamide 13.8 versus 27.2 mg/kg/day).^30,32,33  In the Spanish study evaluating gabapentin, which provided information regarding participant diagnosis, the dose was twice that of the present study despite similar diagnostic characteristics of the participants (75% and 83% had a diagnosis of focal epilepsy in the present and Spanish study, respectively).³²  Despite the low doses of pregabalin, vigabatrin, and gabapentin, the mean pregabalin and vigabatrin concentrations were not lower in our study than in previous studies (mean serum concentrations of pregabalin and vigabatrin in the present versus previous studies; 2.84 versus 2.1 µg/mL and 6.35 versus 5.3 µg/mL, respectively), with most of the concentrations within the reference range (56%, 96%, and 75% of pregabalin, vigabatrin, and gabapentin concentrations, respectively).^30,33 

The serum concentrations of lacosamide and perampanel were significantly decreased by the use of concomitant enzyme inducers, in agreement with previous findings.^23,24,34  The effect of an enzyme inducer was greater with perampanel than with lacosamide. Lacosamide is metabolized mainly by CYP2C19, CYP2C9, and CYP3A4, and also by CYP-independent mechanisms, whereas perampanel is extensively metabolized in the liver, primarily by CYP3A4.^35,36  Because lacosamide is metabolized by multiple pathways, the influence of increased activity of a specific CYP enzyme likely will be smaller compared with that of perampanel in which the influence of increased CYP3A4 activity would be substantial. Therefore, in patients using perampanel with concomitant enzyme inducers, the serum concentration should be closely monitored when changing the dose.

This study has several limitations. First, because of the retrospective nature of the data, we had limitations in controlling potential confounders, such as sampling time. For instance, although clinicians and patients were informed to collect samples at a drug-fasting state to obtain trough concentration, non–drug-fasting concentrations could not be completely excluded. Second, because of the small number of patients included in the TDM data of vigabatrin and gabapentin, bias cannot be excluded. Lastly, although we observed extensive pharmacokinetic variability that supports the use of TDM in these 6 ASMs, further studies that evaluate the relationship between drug concentration and clinical response would be needed to establish the clinical effectiveness of TDM in these ASMs.

We successfully developed a simple and rapid LC-MS/MS method for the simultaneous measurement of 6 new-generation ASMs, and successfully applied it to clinical practice. The described method was suitable for routine clinical TDM because it required simple sample preparation and allowed rapid and accurate analysis. In our study, most of the patients were using multiple ASMs in combination, and the drug concentration and CD ratio were highly variable among individuals in all 6 ASMs. Hopefully, we expect this method and our experience will address the needs of laboratories and clinicians seeking to introduce mass spectrometry–based TDM of new-generation ASMs.