Lacosamide and its deuterated form (lacosamide-d3) were purchased from Toronto Research Chemicals (Toronto, ON, Canada). Next, HPLC-grade methanol was obtained from FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan). An HPLC-grade aqueous ammonium acetate solution was purchased from Nacalai Tesque (Kyoto, Japan). Pooled breast milk was obtained from LEE BioSolutions (Maryland Heights, MO, USA) for method validation. Pooled plasma samples from healthy human donors were obtained from Cosmo Bio (Tokyo, Japan). To investigate the matrix effect, breast milk samples from six donors were obtained from LEE BioSolutions. Individual human plasma samples from six healthy female donors were obtained from Cosmo Bio.

Standard and internal standard (IS) stock solutions were prepared using methanol. Concentrations of the standard stock solutions in methanol were 40, 80, 200, 400, 800, 2,000, 4,000, and 8,000 ng/mL. These stock solutions were stored at − 80 °C. Calibration samples were prepared immediately before analysis by appropriately diluting the stock solutions in blank breast milk or plasma. Specifically, the calibration samples were made by spiking 200 μL of blank pooled breastmilk (LEE BioSolutions) or plasma (Cosmo Bio) with 2.5 μL of each stock solution. Fifty microliters of the calibration sample was used for the procedures described in Sample preparation. Calibration curves of eight concentrations (0.5, 1.0, 2.5, 5.0, 10, 25, 50, and 100 ng/mL) were obtained by plotting the peak area ratio (lacosamide/IS) against the theoretical concentration and were fitted using least-squares regression with 1/x weighting.

To 50 μL of breast milk or plasma samples, 200 μL of methanol containing the IS (1 ng/mL) was added and mixed by vortexing. The mixture was centrifuged at 13,000 × g for 15 min at 4 °C. Subsequently, 100 μL of the supernatant was carefully collected and filtered through a DISMIC-13HP filter (0.2 μm, ADVANTEC, Tokyo, Japan). Next, 2 μL of the sample was injected into the UPLC/MS/MS.

Ultra-performance liquid chromatography (ACQUITY UPLC H-Class, Waters) and an ACQUITY UPLC HSS T3 column (2.0 × 50 mm, 1.8 µm, Waters) were used for chromatographic separation. An in-line column filter kit (Waters) was used to protect the column. The column temperature was set at 40 °C. The mobile phase was an isocratic flow of methanol/10mM ammonium acetate (30:70, v/v) at a flow rate of 0.4 mL/min. The total runtime was 3.5 min. Positive-ion electrospray tandem mass spectrometric analysis was performed using a Xevo TQ-S (Waters) with multiple reaction monitoring (MRM). Transitions m/z 251.2 → 108.0 for lacosamide and m/z 254.0 → 108.0 were monitored for quantification. As a qualifier ion, m/z 251.2 → 91.0 was also monitored. The dwell time for each ion was 200 ms/ion. The cone was set at 20V, and the collision energies were 6, 6, and 18V for the lacosamide, IS, and lacosamide qualifier, respectively. Data were analyzed using TargetLynx software.

Method validation was performed according to the Food and Drug Administration (Guidance for Industry: Bioanalytical Method Validation 2018) and European Medicines Agency (Guideline on Bioanalytical Method Validation 2011) guidelines.

Intra-day precision and accuracy were investigated by analyzing six samples at four concentrations (0.5, 1.25, 12.5, and 80 ng/mL) on the same day. Inter-day precision and accuracy were assessed by analyzing the samples at four concentrations (0.5, 1.25, 12.5, and 80 ng/mL) on six days. The samples were processed as described in Sample preparation and measured. The relative error (R.E.) was calculated as [(measured concentration − nominal concentration)/nominal concentration] × 100 (%). The precision was defined as the relative standard deviation (R.S.D.). The lower limit of quantification (LLOQ) was determined using precision and accuracy data and a signal-to-noise (S/N) ratio > 10. The accuracy and precision should be <  ± 15%, except for LLOQ, for which the acceptable limit is <  ± 20%.

To investigate selectivity, six different lots of matrix (human breast milk or plasma) with or without the LLOQ (0.5 ng/mL) of lacosamide were prepared and analyzed. The interference peak detected at the retention time of lacosamide should be < 20% of the LLOQ and 5% of IS. Carry-over was investigated by injecting a blank matrix (human breast milk or plasma) after injecting the highest concentration of the sample (100 ng/mL). The area responses of the blank matrices were compared with those of the LLOQ. Further, the peak area of the blank samples after injecting the highest concentration should not be over 20% of the peak area at LLOQ and 5% of the peak area of the IS.

To assess matrix effects, the accuracy and precision of multiple lots from different donors were determined. Six lots of breast milk or plasma were spiked with lacosamide at three concentrations (1.25, 12.5, and 80 ng/mL). The samples were processed as described in Sample preparation and measured. The concentration of lacosamide in each matrix was quantified using a calibration curve prepared from the data of pooled breast milk or plasma. The precision should not be greater than 15%.

To assess dilution integrity, lacosamide in breast milk or plasma at a concentration of 1 μg/mL was diluted 100-fold with blank breast milk or plasma. Furthermore, lacosamide in breast milk or plasma at a concentration of 10 μg/mL was diluted 200-fold with blank breast milk or plasma. Six samples were measured, and the accuracy and precision were determined. The precision and accuracy should not be over 15%.

Lacosamide stability (1.25 and 80 ng/mL) in breast milk and plasma was investigated. The lacosamide remaining 24 h after being stored in breast milk or plasma at 4 °C was quantified. Furthermore, the lacosamide quantity 1 month after being stored in breast milk or plasma at − 30 °C was measured. Freeze–thaw stability was investigated by three freeze–thaw cycles (− 30 °C and room temperature). Six replicates were prepared and analyzed.

This study was approved by the Ethics Committee of the Hokkaido University Hospital (020–0139). The volunteer received a detailed explanation of the study and freely gave her consent to participate. The volunteer was a woman admitted to the Obstetrics Department of Hokkaido University Hospital. She had focal onset epilepsy and had been prescribed levetiracetam and lamotrigine. Because her seizures were could not be controlled by these drugs, lacosamide was introduced several years ago. After the introduction of lacosamide, her seizure frequency was reduced. Because lacosamide could not completely control her seizure, clobazam was additionally prescribed. During the peripartum period, she was administered only lacosamide. The patient delivered at term without complications. This was her second pregnancy. She participated in the study three days after delivery. Although non-compliance is suspected during pregnancy, she certainly took lacosamide at least after delivery. She was orally administrated lacosamide twice a day (100 mg × 2). Before the day she participated in the research, the dosage of lacosamide was increased from 50 mg × 2. Her body weight was 55.72 kg. Plasma samples were obtained 75 min before and 60 min after the oral administration of lacosamide. Breast milk samples were collected 75 min before and after administration (60, 115, 300, and 500 min). The samples were stored at − 30 °C until measurement. The samples were prepared as described in Sample preparation and quantified using UPLC/MS/MS. As the lacosamide concentrations exceeded the calibration range, the samples were diluted 100-fold with drug-free breast milk or plasma.

M/P ratios were expressed as the concentration in breast milk / in plasma. RID was estimated as: RID (%) = (Cmilk × Vmilk / Dmaternal) × 100%,

where Cmilk　is the lacosamide concentration in breast milk; Vmilk is the daily volume of milk intake by infants; and Dmaternal is the maternal body weight-adjusted lacosamide daily dose in the mother (mg/kg/day).

Cmilk is the average concentration of lacosamide in breast milk (mg/mL) calculated from the area under the curve (AUC0–12). AUC0–12 was estimated using the linear trapezoidal method. Since we did not obtain the time point 12 h after the last lacosamide administration, the data were compensated by the value at the trough. In the study, Vmilk was used average breast milk intake (150 mL/kg/day) for calculation.

In this study, the protein binding of lacosamide in plasma was assessed by quantifying ultrafiltrates using Centrifree® Ultrafiltration Devices (Merck Millipore, Tulagreen, Ireland). The protein binding ratio was expressed as follows: [(total lacosamide concentration – free lacosamide concentration)/total lacosamide concentration] × 100 (%).