An isotope dilution-liquid chromatography-tandem mass spectrometry (ID-LC-MS/MS)-based candidate reference measurement procedure (RMP) for the quantification of gabapentin in human serum and plasma

Chemicals and reagents

LC-MS grade methanol (CAS 67-56-1) and formic acid (CAS 64-18-6) were purchased from Biosolve (Valkenswaard, The Netherlands). Dimethyl sulfoxide (DMSO; CAS 67-68-5, ACS reagent, ≥99.99%), ammonium acetate (CAS 631-61-8, LC-MS grade), isopropanol (CAS 67-63-0, high-performance liquid chromatography [HPLC] grade) and D₂O (CAS 7789-20-0) were purchased from Sigma-Aldrich (Taufkirchen, Germany).

qNMR ISTD DSS-d6 (TraceSure, Lot Nr. APL6177) was purchased from WAKO FUJIFILM (Neuss, Germany). 5 mm 8-inch NMR tubes were purchased from Sigma-Aldrich and/or Eurisotop GmbH (Hadfield, United Kingdom).

Native human serum was obtained from Merck (Darmstadt, Germany), TDM-free human serum (multi-individual pooled, surrogate matrix) was obtained from Roche Diagnostics GmbH (Mannheim, Germany), native plasma matrix (lithium-heparin [Li-Hep], dipotassium [K2]-ethylene diamine tetraacetic acid [EDTA], and tripotassium [K3]-EDTA) was obtained from anonymized patient samples, and water was purified using a Millipore Milli-Q 3 UV system from Merck. The patient samples were anonymized residual samples, and pools prepared from them, in accordance with the Declaration of Helsinki.

The standard of gabapentin (CAS 60142-96-3, Cat. No. PHR1049, Lot LRAB7794), gabapentin-lactam (CAS 64744-50-9, Cat. No. Y0001281), was purchased from Sigma-Aldrich (Taufkirchen, Germany), and the deuterated ISTD [²H₆]-gabapentin HCl (CAS 1432061-73-8, Art. No. C467, Lot JA-ALS-14-094-P1), was purchased from Alsachim (Illkirch Graffenstaden, France).

General requirements for laboratory equipment

All equipment used was calibrated and certified by the manufacturer. The minimum sample weight for the microbalance used (XPR2, Mettler Toledo, Columbus, OH, USA) was determined according to the United States Pharmacopeial Convention (USP) guidelines (USP Chapters 41 and 1251). Direct displacement pipettes were used to measure organic solvents and serum. For the preparation of stock and spike solutions, volumetric glassware (Class A volumetric flasks) that fulfilled the requirements of ISO 1042 and USP was used.

qNMR determination of the purity of the standard material

qNMR measurements were performed on a JEOL 600 MHz NMR with a He-cooled Cryoprobe. Single-pulse-¹H{¹³C}NMR (Supplemental Figure 1) was utilized for the quantitation (NH₂CH₂; δ= 2.39 ppm, 2H). Additional details about NMR acquisition and processing parameters can be found in Supplemental Material 2.

Preparation of calibrators and quality control (QC) samples

The preparation of calibrator and QC levels, as well as the estimation of calibrator level uncertainties (type B uncertainty), was executed based on Taibon et al. [37]. A detailed description of the methods used to calculate calibrator level uncertainty can be found in Supplemental Material 3.

In brief, two individual calibrator stock solutions were prepared and further used for the preparation of working and spike solutions. For each stock solution, 50 mg of gabapentin was weighed in tin boats on a microbalance (XPR2, Mettler Toledo) and dissolved in 5 mL 50% methanol in Milli-Q water (v/v) using a volumetric flask to achieve concentrations of 10 mg/mL. The exact concentration of the stock solutions was calculated based on the purity of the reference material (99.6% ± 0.3%, determined by qNMR) and the exact amount weighed.

These primary stock solutions were further diluted with 50% methanol to obtain working solutions with a concentrations of 0.5 mg/mL and 1 mg/mL. Stock and working solutions were used (separately) for further preparation of the eight calibrator spike solutions in 50% methanol (v/v). Final matrix-based calibrators, uniformly distributed from 0.300–48.0 μg/mL (see Figure 1), were prepared with a 1 + 99 dilution (v/v) in human serum matrix.

Figure 1:

Schematic overview of gabapentin calibrator and control levels chosen to allow optimal coverage of measurement and therapeutic reference range. Cal, calibrator; QC, quality control.

Four levels of matrix-based QC samples (QC levels) were prepared in the same way as described for the calibrator levels using a third independent stock solution and resulting spike solutions. The concentrations for the QC levels were set at four critical control points: above the limit of quantification (0.500 μg/mL), below (1.50 μg/mL) and within (6.00 μg/mL) the therapeutic reference range, and at the laboratory alert level (25.0 μg/mL) [12] (Figure 1). All samples (spiked and native material) were stored at −20 °C for a maximum of 27 days and all liquids were allowed to reach room temperature prior to use.

ISTD solution

For the preparation of the ISTD stock solution [²H₆]-gabapentin HCl was dissolved in the appropriate amount of 50% DMSO in Milli-Q water (v/v) to obtain a 1,000 μg/mL ISTD stock solution and stored at −20 °C until further use. The ISTD working solution was freshly prepared by a twofold dilution of ISTD stock solution: 88 µL 50% DMSO in Milli-Q water (v/v) was mixed with 12 µL of ISTD stock solution, followed by the addition of 3,900 µL Milli-Q water, to achieve a final concentration of 3 μg/mL.

Sample preparation

Native human serum, TDM-free serum (surrogate matrix) and plasma (Li-Hep, K2-EDTA, and K3-EDTA) were used as sample matrix. 100 µL of the ISTD working solution was pipetted into a 2 mL tube (Eppendorf Safe-Lock Tubes, Eppendorf, Hamburg, Germany) and 50 µL of the sample specimen (native sample/calibrator/QC) was added. To ensure an ISTD equilibration, the sample was mixed briefly on a vortex and then incubated for 15 min (500 rpm, 37 °C) on a thermomixer (Eppendorf). Proteins were precipitated by adding 1,000 µL precipitation solution (75% methanol in Milli-Q water [v/v]), followed by an incubation step on a thermomixer for 10 min (2,000 rpm, 23 °C) and subsequently centrifuged for 10 min at 4 °C and 20,000 rcf. 10 µL of the supernatant was then diluted 1 + 99 (v/v) using mobile phase A in an HPLC vial (Wicom, Heppenheim, Germany).

Liquid chromatograph-mass spectrometry (LC-MS)

Chromatographic separation was performed by an Agilent 1290 Infinity II LC system (Santa Clara, California, USA) equipped with a binary pump, a vacuum degasser, an autosampler at 7 °C, and a column compartment tempered to 40 °C.

An Agilent Zorbax Eclipse XDB-C8 column (100 × 3 mm, 3.5 µm) was used for chromatography as it showed the best peak shapes compared with other reversed-phase columns. The mobile phases consisted of 2 mM ammonium acetate in Milli-Q water with 0.1% formic acid (A) and methanol/2 mM ammonium acetate in Milli-Q-water 95 + 5 (v/v) with 0.1% formic acid (B).

All measurements were performed at a flowrate of 0.6 mL/min using the following gradient: t=0.0 min, 100% A; t=1.0 min, 100% A; t=4.5 min, 50% A; t=4.6 min, 0% A; t=7 min, 0% A; t=7.1 min, 100% A; t=8.5 min, 100% A. The injection volume was set to 5 μL. A divert valve, switching the eluent flow before 0.8 min and after 5.5 min to the waste, was used to reduce contamination of the mass spectrometer.

Gabapentin was detected in multiple reaction monitoring (MRM) mode using an AB Sciex Triple Quad 6,500+ mass spectrometer (Framingham, Massachusetts, USA), with a Turbo V ion source operating in positive electrospray ionization mode. An ion spray voltage of 5000 V and a temperature of 500 °C were applied. Curtain gas, collision gas, ion gas source 1, and ion gas source 2 were set at 35 psi, 10 psi, 50 psi, and 50 psi, respectively. For all mass transitions, a declustering potential of 57 V and a dwell time of 50 ms were applied. The quantifier transition is the basis of the quantitative method and is associated with a corresponding transition of the ISTD. The additional qualifier allows checking for possible interferences in clinical samples by applying the “branching ratio concept”. Table 1 shows an overview of the selected reaction monitoring transitions as well as the remaining compound-dependent MS settings.

System suitability test (SST)

An SST was performed prior to each analysis to check the sensitivity of the system, chromatographic performance, and possible carryover effects. Concentration levels of the SST1 and SST2 corresponded to the analyte concentration within the processed calibrator level 1 and 8, respectively.

If the signal-to-noise ratio of the quantifier was ≥10 for SST1 and the retention time for SST1 and SST2 was within 3.6 min (±0.5 min), the SST was considered passed and chromatographic performance was assured. For the investigation of potential carryover effects, the highly concentrated sample SST2 was injected, followed by the injection of two solvent blanks. The analyte peak area at the expected retention time observed in the first blank after the injection of the SST2 sample was required to be ≤20% of the analyte peak area of SST1 (alternatively, calibrator 1) to pass.

Calibration and structure of analytical series

The system was calibrated using the calibrators prepared as described in the ‘Preparation of calibrators and quality control (QC) samples’ section. Calibrator levels were measured in increasing concentration at the beginning and at the end of the analytical series (bracketing mode) using both calibrations to generate the final calibration function. The final calibration function was obtained by linear regression of the area ratios of the analyte and ISTD (y) against the analyte concentration (x), resulting in the function, y = ax + b. The sequence setup is described in Supplemental Material 1.

Data processing

Data evaluation was performed using the Analyst software, Version 1.6.3, employing the IntelliQuant algorithm. Gabapentin and its ISTD showed a retention time of 3.6 min and were integrated within a 30 s window. For the peak integration, a smoothing and peak splitting factor of 3 were used. The noise percentage was set to 90% with a base sub window of 30 s.

Method validation

Assay validation and determination of measurement uncertainty were performed according to the Clinical and Laboratory Standard Institute Guidelines C62A “Liquid Chromatography-Mass Spectrometry Methods” [38], the International Conference on Harmonization guidance document “Harmonised Tripartite Guideline Validation of Analytical Procedures: Text and Methodology Q2 (R1)” [39] and the Guide to the expression of uncertainty in measurement (GUM) [40]. Further details are published in Taibon et al. [37].

Traceability to SI units

The use of qNMR-characterized reference standards for calibration of the RMP leads to direct traceability to SI units. Since there are currently no primary reference materials for gabapentin available, traceability was established by using qNMR ISTDs that are directly traceable to primary standards. Therefore, six individual experiments (Supplemental Figure 2 in Supplemental Material 2), involving six individual weightings of the analyte and DSS-d6 were performed yielding a final content value of 99.6 ± 0.3% (coverage factor k=1).

Selectivity/specificity

All elements of the novel method were carefully evaluated during the design phase of the project, and a prototype method was evaluated prior to method validation. The optimization of the LC-MS/MS setup included mobile phase composition and gradient, stationary phase selection, and ion source optimization during the setup of the selected reaction monitoring (SRM) experiment.

Baseline separation of gabapentin and the known interfering compound gabapentin-lactam (see Figure 2) was ensured by the developed gradient combined with the reversed-phase column (Agilent Zorbax Eclipse XDB-C8). The chromatographic resolution (R) was ≥8.5 for all tested matrices, calculated as follows: R=2 * (t_R2–t_R1)/(w_b1+w_b2), where t_R1 and t_R2 was the retention time of peaks 1 and 2, respectively, and w_b was the peak width at baseline.

Figure 2:

Gabapentin LC-MS/MS derived analytical readouts. Analytes on the left-hand side, gabapentin ISTD always on the right-hand side. (A) Chromatogram of gabapentin (black) and gabapentin-lactam (grey) both with a concentration of 4.00 μg/mL in native serum; (B) lowest gabapentin calibrator peak with a concentration of 0.300 μg/mL in native serum; (C) patient sample pool of five individual patients with a concentration of 9.20 μg/mL. ISTD, internal standard.

The selectivity was further determined by analyzing sample pools of analyte-free native human serum, surrogate serum, and human plasma (Li-Hep). No signals were detected at the expected retention time. Furthermore, no residual unlabeled analyte was observed in the ISTD [²H₆]-gabapentin.

Matrix effects

To reduce the occurrence of matrix-dependent effects caused by salts, proteins, and phospholipids sample preparation optimization included fluid handling, selection of optimal pipettes, protein precipitation including equilibration times as well as the dilution into the linear range of the MS detector.

The post-column infusion experiment showed no matrix component-mediated effect on the ionization field of gabapentin at the expected retention time for serum and plasma matrix (K2-EDTA plasma, K3-EDTA plasma, and Li-Hep plasma).

For further evaluation of matrix effects, slopes and coefficients of determination of calibrations in different matrices (native serum, surrogate serum, and Li-Hep plasma) were compared against calibrations in neat solution. Slopes (n=6 sample preparations) were found to be 0.253 (95% confidence interval [CI] 0.251 to 0.256) for the native serum matrix, 0.251 (95% CI 0.248 to 0.255) for the neat solution, 0.253 (95% CI 0.251 to 0.256) for the surrogate serum matrix and 0.255 (95% CI 0.251 to 0.258) for the plasma matrix. The CIs of the slopes are overlapping, which leads to the assumption that they are not significantly different from each other, supporting the absence of matrix effects. Moreover, mean r² values were ≥0.999 independent of the matrix used for calibration.

Matrix samples (n=6 sample preparations), when calibrated against neat, showed a relative bias for all matrices and levels ranging from −1.2 to 4.2% with CVs of less than 4.4%.

Additionally, matrix effects were evaluated based on Matuszewski et al. [42]. Analyte peak areas, ISTD peak areas, and area ratios from spiked matrix samples were compared with spiked neat samples. Matrix effect values of >100% indicate an ionization enhancement and values <100% indicate an ionization suppression. No matrix effect was observed, with values ranging from 99 to 104% for the analyte and from 103 to 105% for the ISTD in all tested matrices (Table 2). The area ratios were between 96 and 101%. Thus, the compensating effect of the present labelled ISTD is confirmed.

Linearity

Linearity of the method was demonstrated by analyzing calibration curves covering an extended measuring range by ±20% (n=6, sample preparations). Figure 3 shows that residuals were randomly distributed in a linear regression model, thus it was chosen for assay calibration. Correlation coefficients were r ≥0.999 for all individual calibrations. Based on these results, samples 1 to 11 were evaluated and showed a linear dependence with a correlation coefficient of 1.00. The relative deviation ranged from −0.6 to 3.1% and CV was determined to be ≤2.2%.

Figure 3:

Residuals plot of the calibration model evaluation. A linear model was chosen and six individual sample set preparations were performed to evaluate repeated model performance.

Lower limit of measuring interval (LLMI)

The LLMI was determined using spiked samples with the concentration of the lowest calibrator level. Relative bias showed a deviation of −2.0% while the CV was found to be 2.6% at a concentration of 0.301 μg/mL.

Precision and accuracy

Precision and accuracy were determined within a multi-day validation experiment using spiked serum and plasma samples, as well as native patient samples.

Bias was evaluated using three preparations by different operators (n=6). The relative mean bias ranged from −0.8–1.0% for all levels and matrices within the measuring range. Dilution integrity was proven for highly concentrated samples above the measuring range with a bias ranging from −1.3–2.0% (Table 3). The corresponding confidence intervals showed that the results were distributed around the nominal concentration.

Intermediate precision, including variances in between-day -calibration, -preparation and -injection, was found to be less than 2.3% (Table 4), independent of the matrix. Repeatability CV ranged from 0.9–2.0% across all concentration levels (Table 4).

Sample stability

Processed samples were found to be stable at 7 °C for 6 days with recoveries ranging from 99 to 108%. Stability of spiked frozen serum calibrator and QC materials stored at −20 °C was examined after 28 days. Freshly prepared calibrators and QC samples were used to evaluate the stored samples. These were found to be stable for 27 days with recoveries ranging from 96 to 104% compared to the original value.

Equivalence of results between independent laboratories

From the 122 native anonymized patient samples used in the method comparison study, 19 were not considered, as 18 samples were below the limit of quantitation, and one sample was detected as an outlier by the application of the generalized extreme studentized deviate test [44]. Passing-Bablok regression analysis showed very good agreement between the two laboratories and resulted in a regression equation with a slope of 1.00 (95% CI 0.98 to 1.03) and an intercept of −0.05 (95% CI −0.14 to 0.03). Pearson’s correlation coefficient was found to be ≥0.996 (Figure 4).

Figure 4:

Results from the patient sample-based gabapentin method comparison study performed between two independent laboratories. (A) Passing–Bablok regression plot including the pearson regression analysis for the method comparison study of the RMP (n=103 patients) between the independent laboratories (Laboratory 1: Risch site, and Laboratory 2: Roche site). The regression analysis resulted in a regression equation y=1.00x−0.05 with a 95% confidence interval (CI) from 0.98 to 1.03 for the slope and from −0.14 to 0.03 for the intercept. The pearson correlation coefficient was 0.996. (B) Bland–Altman plot for the method comparison study of the RMP (n=103 patients) between two independent laboratories (Laboratory 1: Risch site, and Laboratory 2: Roche site). Relative measurement difference, based on the mean of the individual values, are plotted against the mean of the individual measurement values (“relative Bland-Altman plot”). The inter-laboratory measurement bias was −1.5% (95% CI −2.6% to −0.4%) and the ±1.96 SD interval of the relative differences ranged from −12.5 to 9.6% (95% CI 3.8%).

Bland-Altman analysis also showed good agreement between the two laboratories (Figure 4). The data scatter of the relative differences between the laboratories ranged from −12.5 to 9.6 (95% CI 3.8%). The result bias in the patient cohort was −1.5% and statistically significantly different from zero (95% CI interval of the bias range: −0.4% to −2.6%). It is however in the range to be expected from the uncertainty associated with the production of independent calibrator samples. Both data scatter and data bias indicate that the proposed gabapentin RPM is transferable between independent laboratories.

Precision performance evaluated within the second laboratory showed CVs for repeatability and mean precision of ≤2.0 and ≤4.9%, respectively. The variation between-calibration and between-day is higher within laboratory 2 (n=36 measurements per level), resulting in a higher intermediate precision that of laboratory 1 (n=60 measurements per level).

Uncertainty of results

Total measurement uncertainties for single measurements were found to be ≤2.5% (k=1), independent of the concentration level and the nature of the sample (Table 5). To obtain an expanded measurement uncertainty, the derived total uncertainty was multiplied by a coverage factor of k=2, which corresponds to an approximate confidence level of 95% assuming a normal distribution. Determined uncertainties are expressed as CVs for better interpretability.