MS Reanalysis

Article pubs.acs.org/ac

Practical Approaches to Incurred Sample LC-MS/MS Reanalysis: Confirming Unexpected Results Wenkui Li,* Hui Lin, Jimmy Flarakos, and Francis L. S. Tse Department of Drug Metabolism and Pharmacokinetics, Novartis Institutes for BioMedical Research, One Health Plaza, East Hanover, New Jersey 07936, United States ABSTRACT: Incurred sample reanalysis (ISR) is an important step in assuring the quality of an LC-MS/MS bioanalytical assay and the integrity of bioanalysis conduct. A conventional ISR involves analysis of at least 20 samples taken from an in vivo study a second time using the method that was described in prestudy validation and employed in generating the initial study sample results. However, this practice is sometimes inadequate to confirm bioanalytical results that are unexpected. The present report discusses several additional exploratory activities that were performed to confirm the unexpected plasma concentration−time results of NVP-1, an investigational drug candidate, observed in the plasma samples collected from patients in a phase II trial. These approaches include (1) LC-MS/MS reanalysis of the study samples after multiple freeze/thaw cycles followed by a short-term benchtop storage, (2) evaluation of additional MS/MS transitions in LC-MS/MS, (3) employment of a different sample preparation procedure in LC-MS/MS, and (4) study sample dilution using plasma samples from healthy volunteers. These procedures are practical and can be readily implemented in the confirmatory LC-MS/MS bioanalysis of other small molecules.

I

community on what confirmatory experiments should be conducted and to what extent. Compound NVP-1 (C18H17ClN4O2S, structure not shown) is an aldosterase inhibitor currently under clinical investigation for the treatment of hypertension. The compound has a molecular weight of 388 Da and is highly protein bound (∼96%) with pKa values of ∼3.5 and 9.6. An in vitro study using human liver microsomes and hepatocytes and an in vivo study in rats suggested that N-demethylation, N-glucuronidation, and cysteine conjugation are the main pathways of metabolism for the compound. The human plasma LC-MS/MS assay was developed and validated according to the current health authority guidances and industry practice to support a first-in-human (FIH) study in healthy volunteers. In the first-in-patient (FIP) study, hypertensive patients were instructed to take 10 mg once-a-day (QD) or 30 mg twice-a-day (BID) of the test article orally. Surprisingly, the measured analyte concentrations in the trough (or predose) plasma samples collected on the scheduled study days 7, 14, 21, 28, and 29 from the first group of patients taking 30 mg BID of the test article were 300% to 1000% (Figure 1) higher than the highest values predicted on the basis of the FIH study results. The apparent higher-than-expected concentrations have exceeded the predetermined safety limit, leading to the termination of the trial. Although the cause for these unexpectedly high patient plasma concentrations remains unknown, several possibilities were worthy of investigation. These possibilities can be related, but not limited to, (1) the prediction model, (2)

ncurred sample reanalysis is an important step in the current practice of LC-MS/MS bioanalysis. At least 20 samples taken from a typical in vivo study are assayed a second time, using the method that was described in prestudy validation and used in generating the initial study sample results.1−4 The purpose of conducting incurred sample reanalysis is to assess the reproducibility of an assay method and the integrity of the bioanalytical process and the associated bioanalytical results.1−4 Several recent publications have demonstrated that incurred sample reanalysis can reveal various methodological and/or operational issues that may lead to bioanalytical errors.5−10 These issues include, but are not limited to, instability of the analyte(s) of interest,5 suboptimal sample preparation procedures,5−10 population specific matrix effect and/or interference,5,11 inadvertent sample switching,6 inadequate chromatographic selectivity,10,11 and sample nonhomogeneity.5,6,9 Analyte concentrations observed in study samples collected from patients sometimes do not correlate well with those obtained from healthy volunteers. These unexpected analyte concentrations may be the result of the differences in absorption, distribution, metabolism, and excretion of a drug candidate(s) between the study patient population and healthy volunteers. Alternatively, unexpected analyte concentrations could also be a result of noncompliance by patients in study drug administration. Nevertheless, when unexpected analyte concentrations are observed, the validity of the bioanalytical assay and/or the conduct of study sample analysis are often the first question that comes to mind. In order to confirm the unexpected results, the bioanalytical laboratory often needs to conduct experiments well beyond the conventional incurred sample reanalysis. Currently, little information is available to guide the bioanalytical © 2013 American Chemical Society

Received: November 25, 2012 Accepted: January 13, 2013 Published: January 30, 2013 2405

dx.doi.org/10.1021/ac3034135 | Anal. Chem. 2013, 85, 2405−2413

Analytical Chemistry

Article

Figure 1. Representative unexpected trough NVP-1 plasma concentrations observed from the first group of patients on study days 7, 14, 21, 28, and 29 vs predicted plasma trough concentrations (Cmin and Cmax) based on the FIH study outcomes.

with a dwell time of 100 ms for each analyte and ISTD mass transition. The mass spectrometer was operated at unit mass resolution (half-height peak width, 0.7 Da) for both the first quadrupole and the third quadrupole. Preparation of Matrix Calibration and QC Samples. Two primary NVP-1 stock solutions were prepared in methanol, each at a concentration of 1.00 mg/mL, in 20 mL vials. The LCMS/MS responses of stock solutions, from the two weighings, were required to be within 5%. An ISTD working solution containing 50.0 ng/mL of [M + 8]NVP-1 was prepared from the ISTD stock solution (1.00 mg/mL) with 50% methanol (v/v). The calibration standards were prepared by serial dilution of the stock solution with blank human plasma. Eight nonzero calibration standards were prepared at concentrations of 0.100, 0.200, 0.800, 2.00, 8.00, 20.0, 80.0, and 100 ng/mL. Pooled quality control (QC) samples were prepared by serial dilution of a separate stock solution using human plasma. Six QC sample concentrations at 0.100 (LLOQ), 0.300, 0.900, 9.00, 75.0, and 500 (dilution QC) ng/mL were prepared. Sample Preparation via Liquid−Liquid Extraction. A 100 μL of blank plasma (for matrix blanks and zero samples), calibration standard, QC sample, and study sample was added to the appropriate well of a 96-well assay plate. A 25 μL aliquot of the internal standard working solution (50.0 ng/mL of [M + 8]NVP-1 in 50% aqueous methanol, v/v) was added to each well except to the matrix blanks, where a 25 μL aliquot of 50% aqueous methanol (v/v) was added. A 100 μL aliquot of a sodium carbonate (100 mM) solution in water was added to each individual well, and the plate was vortex-mixed for about 0.5 min on a pulse-vortex mixer with a motor speed setting of ∼65 units. A 500 μL aliquot of methyl tert-butyl ether (MTBE) was added to each well, and the plate was covered and vortex-mixed for approximately 10 min using the same vortex-mixing setting as above. The sample plate was centrifuged at approximately 4000g for ∼10 min at 4 °C. The supernatant (400 μL) from each well was transferred via a TomTec Quadra 96 system (Hamden, CT, USA) to the corresponding well in a 1 mL, 96-well plate. This was followed by evaporation of the supernatants to dryness under a stream of nitrogen at 45 °C. The sample residues were reconstituted with a 100 μL volume of reconstitution solution (50% aqueous methanol, v/v). After a brief vortex-mixing and centrifugation, a 10 μL volume of the reconstituted sample extracts was injected onto the LC-MS/MS system. Sample Preparation via Protein Precipitation. A 100 μL of blank plasma (for matrix blanks and zero samples), calibration

the bioanalytical method and bioanalysis conduct, and (3) patient noncompliance in drug administration. The current work highlights some practical approaches that a bioanalytical laboratory can take in performing confirmatory LC-MS/MS bioanalysis of study samples with unexpected analyte concentrations.

■

EXPERIMENTAL SECTION Materials. NVP-1 and the stable labeled internal standard ([M + 8]NVP-1, ISTD, 13C2C162H6H11ClN4O2S, structure not shown) were synthesized at Novartis Pharmaceuticals Corporation (East Hanover, NJ, USA). Control human plasma with K2EDTA as the anticoagulant was obtained from Bioreclamation (Westbury, NY, USA). HPLC grade methanol, acetonitrile, laboratory grade methyl tert-butyl ether (MTBE), and formic acid (85%) were obtained from Fisher Scientific (Fair Lawn, NJ, USA). Deionized water was generated by an ELGA ultrapure water system (ELGA, Oxford, UK). Chromatography. An integrated Shimadzu liquid chromatography system consisting of a CBM-20A controller, DGU-20A multichannel mobile phase degasser, CTO-20A column heater, two LC-20AD pumps, and SIL-20ACHT autosampler (Shimadzu, Columbia, MD, USA) and an ACE3 C18 (50 × 4.6 mm, 3 μm particle size, MAC-MOD Analytical, Chadds Ford, PA, USA) HPLC column were used for the chromatographic separation of NVP-1 and the ISTD from the matrix components. Mobile phase A was water (containing 0.5% formic acid by volume) and mobile phase B was acetonitrile (containing 0.5% formic by volume). Isocratic chromatographic elution using 50% mobile phase B was employed from 0 to 2.0 min at a flow rate of 1.0 mL/min. This was followed by a column wash-out using 95% mobile phase B at a flow rate of 1.5 mL/min. Column re-equilibration was made using 50% mobile phase B at a flow rate of 1.0 mL/min. The injection cycle time was approximately 4.2 min. MS/MS Detection. A Sciex API5000 triple quadrupole mass spectrometer (AB Sciex, Concord, Ontario, Canada) with a TurboIonSpray (TIS) interface was operated in the positive ionization mode for the selected reaction monitoring (SRM) LCMS/MS analyses. The optimized instrumental conditions were as follows: TIS source temperature, 550 °C; TIS voltage, 5000 V; curtain gas, 20 units; nebulizing gas (GS1), 70 units; TIS (GS2) gas, 80 units; CID (collision-induced-dissociation) gas, 5 units; collision energy, 37 eV for both the analyte and the ISTD. The following precursor → product ion transitions were used for SRM: analyte, m/z 389.1 → 280.0; ISTD, m/z 397.2 → 284.2, 2406



Article

Figure 2. Representative product ion mass spectra of NVP-1(A) and its [M + 2] isotope (B).

(benchtop stability, room temperature stability, long-term storage stability, and autosmpler stability, etc.). The details of method development and validation are not listed here. Conventional Incurred Sample Reanalysis of “First-inPatient” Study Samples. Following the initial analysis of the plasma samples collected from patients, a conventional incurred sample reanalysis was conducted for 24 selected samples according to the current industry practice. The obtained concentrations with bias (%) values within ±20% of the initial results [(repeat result − initial result)/initial result × 100%] were considered acceptable. Assessment of Incurred Sample Short-Term Stability for NVP-1 in Human Plasma. A set of 21 randomly selected human plasma samples from the FIH study were subjected to two additional freeze/thaw cycles, followed by storage at room temperature (∼22 °C) on the laboratory bench for about 20 h. The samples were then reanalyzed along with a set of calibration standards and regular QCs. The obtained results with bias (%) values within ±20% of the initial results were considered acceptable. Confirmatory Analysis of Incurred Samples Using Three Additional SRM Channels in LC-MS/MS. The product ion mass spectrum of NVP-1 was re-examined. Three additional SRM channels (m/z 389.1 → 295.0, 391.1 → 282.0, 391.1 → 297.0, Figure 2) along with the original m/z 389.1 → 280.0

standard, QC sample, and study sample was added to the appropriate well of a 96-well assay plate. A 25 μL aliquot of the internal standard working solution (50.0 ng/mL of [M + 8]NVP1 in 50% aqueous methanol, v/v) was added to each well except to the matrix blanks, where a 25 μL aliquot of 50% aqueous methanol (v/v) was added. A 500 μL aliquot of the protein precipitation solution (acetonitrile/methanol/formic acid, 95/ 5/0.1, v/v/v) was added to each well, and the plate was covered and vortex-mixed for approximately 10 min on a pulse-vortex mixer with a motor speed setting of ∼65 units. The sample plate was centrifuged at approximately 4000g for ∼10 min at 4 °C. The supernatant (400 μL) from each well was transferred via a TomTec Quadra 96 system (Hamden, CT, USA) to the corresponding well in a 1 mL, 96-well plate, followed by evaporation of the supernatants to dryness under a stream of nitrogen at 45 °C. The sample residues were reconstituted with 100 μL volume of reconstitution solution (50% aqueous methanol, v/v). After brief vortex-mixing and centrifugation, a 10 μL volume of the reconstituted sample extracts was injected onto the LC-MS/MS system. Prestudy Validation Using LLE. Prestudy validation was conducted using the standards and QCs prepared as above. The validation activities cover the evaluations on selectivity, specificity, linearity, intraday and interday assay precision and accuracy, matrix effect, extraction recovery, and stability 2407



Article

transition were included in the confirmatory LC-MS/MS reanalysis of the study samples. In this experiment, 24 plasma samples from healthy volunteers and 24 plasma samples from patients were selected and assayed using the same sample preparation procedure, i.e., LLE, as employed in the FIH/FIP sample analysis. The obtained results from LC-MS/MS via each of the four SRM channels were compared to those obtained initially via the single SRM channel. The bias (%) values within ±20% of the initial results were considered acceptable. Confirmatory Analysis of Incurred Samples via a Different Sample Preparation Procedure. Protein precipitation, instead of liquid−liquid extraction, was employed to provide additional confirmation. In this experiment, 24 plasma samples from healthy volunteers and 24 samples from patients were selected and assayed using the protein precipitation method. The obtained results with bias (%) values within ±20% compared to the initial results were considered acceptable. Confirmatory Analysis of Incurred Samples via Dilution with Samples from a Different Study. Seventeen (17) FIP plasma samples were each diluted with 17 FIH plasma samples using a one-to-ten ratio. This was followed by LC-MS/ MS analysis of the diluted samples along with the calibration standards and regular QCs. The obtained concentration results with bias (%) values within ±20% of the theoretical results [concentration of FIH sample × dilution factor + concentration of FIP sample × dilution factor] were considered acceptable.

Table 2. Result Summary of Incurred Sample Short-Term (3 F/T Cycles Followed by 20 h at Room Temperature) Stability Assessment for NVP-1 in the Plasma Samples Collected from Healthy Volunteers

■

RESULTS AND DISCUSSION When the unexpected analyte plasma concentrations (Figure 1) were observed in the initial LC-MS/MS analysis, a conventional incurred sample reanalysis was performed using the verified calibration standards (0.100 to 100 ng/mL) and QCs [0.300,

original concentration (ng/mL)

repeat concentration (ng/mL)

difference (%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

672 770 767 326 715 360 461 322 407 702 1.55 2.70 3.16 6.35 4.80 55.8 55.5 42.5 76.4 36.4 179 105 224 307

709 778 780 346 759 362 501 343 425 715 1.67 2.73 3.24 6.47 4.95 56.8 57.2 44.0 78.3 35.9 189 111 237 324

5.5 1.0 1.7 6.1 6.2 0.6 8.7 6.5 4.4 1.9 7.7 1.1 2.5 1.9 3.1 1.8 3.1 3.5 2.5 −1.4 5.6 5.7 5.8 5.5

original concentration (ng/mL)


difference (%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

4.02 2.60 2.58 2.45 7.73 1.59 1.81 2.00 2.79 2.92 2.88 20.2 16.8 10.4 11.1 34.2 16.1 8.28 7.03 2.86 5.36

4.22 2.44 2.42 2.35 7.71 1.58 1.80 2.36 2.82 2.85 2.74 18.9 16.1 9.76 10.8 34.3 15.3 7.84 6.62 2.67 5.00

5.0 −6.2 −6.2 −4.3 −0.3 −0.6 −0.4 17.9 1.1 −2.4 −4.9 −6.2 −3.9 −6.2 −2.8 0.3 −5.0 −5.3 −5.9 −6.5 −6.7

0.900, 9.00, 75.0, and 500 ng/mL (dilution QC)] for 24 clinical plasma samples. The same sample preparation procedure and LC-MS/MS conditions as for the initial analysis were employed in the incurred sample reanalysis (ISR) evaluation. As shown in Table 1, the difference (%) between the repeat and initial results was within ±10% for all 24 incurred samples. However, despite the excellent ISR results, the unexpectedly high analyte concentrations still presented bioanalytical and clinical challenges. From a bioanalysis perspective, as the method development and validation was performed using the metabolitefree matrix collected from healthy volunteers, several possibilities needed to be considered: (1) possible instability of incurred samples during freeze/thaw cycles and benchtop storage in the process of sample analysis, (2) possible interference in MS/MS detection by unknown endogenous components in the patient plasma samples, (3) possible lack of specificity in the sample preparation procedure, and (4) possible patient plasma-specific matrix effect that might have resulted in “signal enhancement” for the analyte of interest in the initial LC-MS/MS analysis. In the current confirmatory work, specific experiments were designed and implemented to address the above concerns. Assessment of Incurred Sample Short-Term Stability for NVP-1 in Human Plasma. In prestudy validation, stability of NVP-1 in metabolite-free human plasma QC samples was demonstrated after three freeze/thaw cycles, for at least 24 h at room temperature and at least 119 days following storage at ≤−60 °C. However, it is well-known that metabolite-free QC samples might not always reflect the stability of the analyte(s) of interest in incurred samples.7,12 As a considerable number of phase II conjugates were identified in the rat ADME study using 14 C-labeled material and in vitro studies using human liver microsomes and hepatocytes as well, a short-term stability evaluation was conducted for the human plasma samples collected from the FIH study. This evaluation was to confirm

Table 1. Result Summary of Conventional Incurred Sample Reanalysis (ISR) of the Plasma Samples Collected from Patients sample index

sample index

2408



Article

Figure 3. Representative LC-MS/MS chromatograms with four MS/MS transition channels (A: m/z 389.1 → 280.0; B: m/z 389.1 → 295.0; C: m/z 391.1 → 282.0; D: m/z 391.1 → 297.0) for the analyte (left side) and one MS/MS transition (m/z 397.2 → 284.2) for ISTD (right side).

whether the unknown phase II conjugates were stable in the incurred samples during the bioanalytical process that covers the initial analysis, repeat analysis, and ISR (three freeze/thaw cycles) and possible unexpected long-term storage at room temperature (∼22 °C) on the laboratory bench. As shown in Table 2, the measured analyte concentrations from all 21 selected human plasma samples were within the acceptance window when compared to those obtained from the initial analysis. These results confirmed that the unknown phase II conjugates of NVP1 are stable in human plasma samples under the conditions described above and that similar analytical results can be obtained for the compound in the incurred samples after 3 freeze/thaw cycles followed by extended storage at room temperature for 20 h. Confirmatory Analysis of Incurred Samples with Addition of Three SRM Channels in LC-MS/MS. NVP-1

contains a chlorine in the molecule and, as shown in Figure 2, both the protonated molecule (m/z 389) and protonated isotopic molecule (m/z 391) produced abundant product ions, respectively, at m/z 280 and 295 (Figure 2A) and m/z 282 and 297 (Figure 2B). Considering the three stage MS/charge (m/z) specific separation, i.e., Q1 parent ion selection and Q2 CID and Q3 product ion selection, in the tandem mass spectrometric (MS/MS) detection, an MS related deficiency in quantitative LC-MS/MS bioanalysis would be unlikely. However, predose plasma samples were not collected from the patients in this study. Therefore, the presence or absence of a possible endogenous component cannot be directly confirmed, which may share the same LC retention time and MS/MS transition (i.e., m/z 389.1 → 280.0) as NVP-1. To address this concern, an investigation was carried out using 24 FIH samples and 24 FIP samples together with verified calibration standards and QCs. Three 2409



Article

Table 3. Result Summary of Reanalysis of the Plasma Samples Collected from Healthy Volunteers Using the Four SRM Channel LC-MS/MS Method original conc. (ng/mL)a

a

sample index 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24


difference (%) compared to the original concentration

m/z 389.1/280.0

m/z 389.1/280.0

m/z 389.1/295.2

m/z 391.1/282.0

m/z 391.1/297.0

m/z 389.1/280.0

m/z 389.1/295.2

m/z 391.1/282.0

m/z 391.1/297.0

42.5 45.6 51.8 57.1 40.7 44.0 18.3 14.5 13.6 11.6 7.35 2.90 16.6 44.7 45.7 41.2 43.9 33.4 22.0 16.3 7.69 8.23 4.62 2.61

44.4 48.7 48.8 57.5 41.6 43.9 19.3 15.5 13.6 12.0 7.45 2.90 16.2 45.8 45.8 42.7 45.2 34.2 24.1 18.1 7.94 8.47 4.92 2.67

43.5 48.2 49.0 57.5 41.4 43.5 19.3 15.5 13.5 11.8 7.22 2.88 16.0 46.0 45.3 41.9 44.0 34.2 23.7 18.0 7.96 8.44 4.88 2.67

43.3 49.2 50.0 58.2 41.6 43.8 19.6 15.3 13.5 12.0 7.24 2.85 16.3 46.5 46.8 42.8 45.9 34.2 24.1 18.2 7.87 8.46 4.92 2.60

42.4 47.3 48.5 56.2 40.1 43.3 18.8 15.2 13.1 11.8 7.22 2.82 15.6 45.0 44.4 41.3 43.8 32.9 23.1 17.7 7.81 8.25 4.78 2.63

4.5 6.8 −5.8 0.7 2.2 −0.2 5.5 6.9 0.0 3.4 1.4 0.0 −2.4 2.5 0.2 3.6 3.0 2.4 9.5 11.0 3.3 2.9 6.5 2.3

2.4 5.7 −5.4 0.7 1.7 −1.1 5.5 6.9 −0.7 1.7 −1.8 −0.7 −3.6 2.9 −0.9 1.7 0.2 2.4 7.7 10.4 3.5 2.6 5.6 2.3

1.9 7.9 −3.5 1.9 2.2 −0.5 7.1 5.5 −0.7 3.4 −1.5 −1.7 −1.8 4.0 2.4 3.9 4.6 2.4 9.5 11.7 2.3 2.8 6.5 −0.4

−0.2 3.7 −6.4 −1.6 −1.5 −1.6 2.7 4.8 −3.7 1.7 −1.8 −2.8 −6.0 0.7 −2.8 0.2 −0.2 −1.5 5.0 8.6 1.6 0.2 3.5 0.8

Measured 8 months earlier.

additional SRM channels, i.e., m/z 389.1 → 295.0, 391.1 → 282.0, and 391.1 → 297.0, as well as the original channel of m/z 389.1 → 280.0 were incorporated in the LC-MS/MS of both the study samples and calibration standards and QCs. Any systematic difference (outside of ±20%) in the measured analyte concentrations between any of the above three additional SRM channels and the original one (i.e., m/z 389.1 → 280.0) would suggest possible deficiency in MS/MS detection. The representative LC-MS/MS chromatograms showing the four channel MS/MS transitions for the study samples are given in Figure 3. Apparently, the LC-MS/MS peak response ratio for the analyte between the two MS/MS channels for each precursor ion was consistent with the relative abundance of respective product ions in the MS2 mass spectra for the analyte (Figure 2A and Figure 2B). Tables 3 and 4 summarize the analyte concentrations from the reanalysis of the 24 FIH samples and 24 FIP samples using the four channels in LC-MS/MS. None of the measured NVP-1 concentrations using any SRM channel were outside the ±20% window compared to the results obtained using any other MS/MS channels or the results from the initial analysis. These results further demonstrated that the LC-MS/MS assay possessed the appropriate selectivity and specificity to distinguish the molecule of interest from the possible endogenous compound(s). Confirmatory Analysis of Incurred Samples via a Different Sample Preparation Procedure. For the current assay, the matrix pH of the study samples was adjusted via the addition of 100 μL of 100 mM sodium carbonate solution (pH ∼11) prior to liquid−liquid extraction using MTBE. Although the soundness of incurred sample reproducibility, incurred

sample short-term stability, and LC-MS/MS methodology has been confirmed as above, one question that remains to be answered is whether the unknown phase II conjugates in the study samples are stable under the above sample processing conditions prior to LC-MS/MS. However, a conventional ISR might not be always adequate to answer this question. This is because that, as long as incurred sample short-term (F/T and RT) stability is demonstrated, the results from a bench-error-free ISR run should be similar (with ±20%) to the initial results as the same sample processing procedure is employed in both runs. Considering that sample preparation can have an impact on the unknown phase II metabolites which are not to be measured,7,10−12 liquid−liquid extraction under the basic condition (pH ∼11) as above was revisited as part of the present confirmation work. In this evaluation, 24 FIH samples and 24 FIP samples were reanalyzed via both the liquid−liquid extraction and protein precipitation. Acetonitrile/methanol/ formic acid (95/5/0.1, v/v/v) was used as the protein precipitation solution. Under this mild acidic condition (pH ∼3), the unknown phase II conjugates that might not be stable under basic conditions (pH ∼11) during LLE are expected to be intact with no back-conversion to the parent. Table 5 summarizes the bias (%) values for the obtained results for the 24 FIH and 24 FIP samples via protein precipitation vs liquid−liquid extraction. All measured values are within ±20% of each other, confirming the absence of any speculated deficiency in the sample preparation procedure. Confirmatory Analysis of Incurred Samples via Dilution with Samples from a Different Study. The total measurement error in the results obtained from study sample 2410



Article

Table 4. Result Summary of Reanalysis of the Plasma Samples Collected from Patients Using the Four SRM Channel LC-MS/MS Method original conc. (ng/mL) sample index 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24


difference (%) compared to original concentration

m/z 389.1/280.0

m/z 389.1/280.0

m/z 389.1/295.2

m/z 391.1/282.0

m/z 391.1/297.0

m/z 389.1/280.0

m/z 389.1/295.2

m/z 391.1/282.0

m/z 391.1/297.0

0.846 1.42 2.00 135 63.5 207 65.5 22.0 15.8 38.5 1.99 5.34 213 309 332 342 314 260 225 19.9 11.0 11.9 19.6 35.8

0.825 1.45 1.95 130 63.0 207 64.4 22.5 15.6 39.1 1.96 5.46 214 298 324 333 306 257 222 19.5 11.0 12.1 19.6 35.9

0.832 1.44 1.98 130 62.2 204 63.8 22.4 15.6 38.9 1.94 5.37 212 300 317 333 303 255 221 19.2 11.0 11.9 19.2 35.8

0.832 1.41 1.99 129 63.3 203 64.4 22.5 15.5 39.2 1.88 5.43 217 302 325 335 309 255 223 19.5 11.0 12.0 19.3 35.9

0.830 1.45 2.03 129 62.4 199 62.5 21.8 15.6 38.2 1.96 5.35 210 293 316 329 297 248 215 19.1 10.7 11.9 18.8 35.2

−2.5 2.1 −2.5 −3.7 −0.8 0.0 −1.7 2.3 −1.3 1.6 −1.5 2.2 0.5 −3.6 −2.4 −2.6 −2.5 −1.2 −1.3 −2.0 0.0 1.7 0.0 0.3

−1.7 1.4 −1.0 −3.7 −2.0 −1.4 −2.6 1.8 −1.3 1.0 −2.5 0.6 −0.5 −2.9 −4.5 −2.6 −3.5 −1.9 −1.8 −3.5 0.0 0.0 −2.0 0.0

−1.7 −0.7 −0.5 −4.4 −0.3 −1.9 −1.7 2.3 −1.9 1.8 −5.5 1.7 1.9 −2.3 −2.1 −2.0 −1.6 −1.9 −0.9 −2.0 0.0 0.8 −1.5 0.3

−1.9 2.1 1.5 −4.4 −1.7 −3.9 −4.6 −0.9 −1.3 −0.8 −1.5 0.2 −1.4 −5.2 −4.8 −3.8 −5.4 −4.6 −4.4 −4.0 −2.7 0.0 −4.1 −1.7

Table 5. Result Summary of Reanalysis of the Plasma Samples Collected from Both Healthy Volunteers and Patients via Liquid− liquid Extraction vs Protein Precipitation first-in-patient samples

first-in-healthy volunteer samples

sample index

LLE conc. (ng/mL)

PPT conc. (ng/mL)

difference (%)

sample index

LLE conc. (ng/mL)a

PPT conc. (ng/mL)

difference (%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

0.846 1.42 2.00 135 63.5 207 65.5 22.0 15.8 38.5 1.99 5.34 213 309 332 342 314 260 225 19.9 11.0

0.804 1.45 1.94 136 62.2 204 67.8 21.9 16.4 38.9 1.90 5.20 214 312 333 330 317 258 228 19.1 11.0

−5.0 2.1 −3.0 0.7 −2.0 −1.4 3.5 −0.5 3.8 1.0 −4.5 −2.6 0.5 1.0 0.3 −3.5 1.0 −0.8 1.3 −4.0 0.0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

14.0 22.8 20.6 17.3 27.0 43.6 31.9 27.8 15.8 10.7 4.66 5.38 5.97 7.20 60.1 53.7 42.8 34.4 35.5 15.0 11.3

14.2 23.1 22.3 17.1 29.8 46.5 34.9 27.9 15.9 11.0 4.53 5.54 6.16 7.38 65.7 54.7 44.9 35.4 35.7 14.9 11.8

1.4 1.3 8.3 −1.2 10.4 6.7 9.4 0.4 0.6 2.8 −2.8 3.0 3.2 2.5 9.3 1.9 4.9 2.9 0.6 −0.7 4.4

a


that are linked to the measured values. For bioanalysis of any analyte(s) with proven freeze/thaw stability, room temperature stability, and long-term storage stability in both QC samples and

analysis is related to the closeness of the values measured and the true values, although the latter might never be known. This closeness is affected by both the systematic and random errors 2411



Article

Table 6. Result Summary of Incurred Sample Accuracy Assessment via Dilution of Patient Plasma Samples Using the Plasma Samples Collected from Healthy Volunteers

a

index of patient plasma sample

original conc. (ng/mL)

theoretical conc. (ng/mL) after a 10-fold dilution (1/10)

index of healthy volunteer plasma sample

original conc. (ng/mL)a

theoretical conc. (ng/mL) after mixing (9/10)

theoretical conc. (ng/mL) of the mixture

measured conc. (ng/mL) of the mixture

bias (%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

672 770 767 326 715 360 461 322 407 702 42.5 76.4 36.4 179 105 224 307

67.2 77.0 76.7 32.6 71.5 36.0 46.1 32.2 40.7 70.2 4.25 7.64 3.64 17.9 10.5 22.4 30.7

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

2.61 4.62 2.90 45.6 8.23 57.1 45.7 44.0 16.6 7.69 42.5 16.3 51.8 18.3 11.6 22.0 33.4

2.35 4.16 2.61 41.0 7.41 51.4 41.1 39.6 14.9 6.92 38.3 14.7 46.6 16.5 10.4 19.8 30.1

69.5 81.2 79.3 73.6 78.9 87.4 87.2 71.8 55.6 77.1 42.5 22.3 50.3 34.4 20.9 42.2 60.8

76.4 83.5 84.4 79.0 84.5 89.7 91.6 72.0 57.0 76.4 44.1 24.9 49.6 35.5 21.8 43.7 62.0

9.9 2.9 6.4 7.3 7.1 2.6 5.0 0.3 2.4 −0.9 3.8 11.6 −1.3 3.3 4.1 3.6 2.0


initially observed from both the patient and healthy volunteer samples after correction of a dilution factor. Again, the same ±20% acceptance rule was applied to determine whether a systematic error related to the analysis of a patient sample might have contributed to the unexpected analyte concentrations. Seventeen (17) FIP plasma samples were diluted 10-fold with 17 FIH samples, followed by sample extraction and analysis via the same procedure used in the initial analysis. As shown in Table 6, the bias values (%) of all samples are within the ±20% acceptance window. Therefore, it can be concluded that there was no systematic error(s) in the measurement of the analye concentrations in the collected patient samples.

incurred samples, although conventional ISR provides information about the random errors (e.g., contamination in sample processing, carryover in LC-MS/MS, inadvertent sample switching, or sample in-homogeneity, etc) in generating the bioanalytical results, it does not always provide information about the systematic errors that may have affected the accuracy of the obtained initial concentrations. Those errors can be due to (1) study sample specific matrix effect, (2) impact of coadministered drugs and their metabolites, or (3) different extraction recovery for the patient samples from the samples collected from healthy volunteers. In general, the matrix employed for the preparation of calibration standards and QCs is different from the incurred samples and therefore, to some extent, the STD/QC samples may not adequately mimic the study samples.2 Recently, incurred sample accuracy was introduced to the bioanalysis community for assessing systematic bioanalytical errors, including constant (absolute) or proportional (relative) error.13 In this assessment, a fixed volume of standard or QC (X0) sample is added to a group of incurred samples for reanalysis. The measured analyte concentrations are subtracted by the added analyte concentration(s) with the remaining values (similar to ISR) to be compared with the initial values of the incurred samples. Any difference in the obtained values within ±20% of the “theoretical” ones would suggest an absence of systematic error. However, the calculated recoveries (%) were largely affected by random experimental errors at relatively high concentrations with respect to low spiked concentrations and vice versa. Therefore, careful selection of the concentration to be used for the standard addition experiment is necessary.13 This process can be very timeconsuming. In the current confirmatory reanalysis, a workable approach to identify the possible systematic error is to reanalyze a group of patient samples with arbitrary initial concentrations (system 1) after a random one-to-one dilution using a group of samples previously analyzed from healthy volunteers (system 2) with arbitrary analyte concentrations as well. The final measured analyte concentrations in the diluted patient plasma samples were compared against the sum of the analyte concentrations

■

CONCLUSIONS In the current work, several approaches were explored to confirm the presence of unexpected NVP-1 concentrations in the plasma samples collected from a group of patients enrolled in a phase II clinical study. Serving as extensions to the conventional incurred sample reanalysis, these approaches included reanalysis of the study samples of interest after multiple freeze/thaw cycles followed by a short benchtop storage, inclusion of additional MS/MS transition channels in the LC-MS/MS analysis, incorporation of a different sample preparation, and dilution of study sample using samples from a different study. By employing those approaches, the presence of unexpectedly high NVP-1 concentrations was confirmed. Although the exact cause of the unexpected results still remains unknown, the present investigation further confirmed the robustness of the LC-MS/MS bioanalytical method and validity of the conduct of bioanalysis of the study samples. The procedures described herein are practical and can be readily implemented in confirmatory LC-MS/MS bioanalysis of other small molecule drugs.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel: 1-862-778-4255. Fax: 1973-781-7579. 2412



Article

Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS We would like to extend our gratitude to Harold T. Smith for his review of the manuscript. REFERENCES

(1) Fast, D. M.; Kelley, M.; Viswanathan, C. T.; O’Shaughnessy, J.; King, S. P.; Chaudhary, A.; Weiner, R.; DeStefano, A. J.; Tang, D. AAPS J. 2009, 11 (2), 238−241. (2) Booth, B. Bioanalysis 2011, 3 (9), 927−928. (3) Viswanathan, C. T.; Bansal, S.; Booth, B.; DeStefano, A. J.; Rose, M. J.; Sailstad, J.; Shah, V. P.; Skelly, J. P.; Swann, P. G.; Weiner, R. AAPS J. 2007, 9 (1), E30−E42. (4) European Medicines Agency; Committee for Medicinal Products for Human Use (CHMP); http://www.ema.europa.eu/docs/en_GB/ document_library/Scientific_guideline/2011/08/WC500109686.pdf. (5) Meng, M.; Reuschel, S.; Bennett, P. Bioanalysis 2011, 3 (4), 449− 465. (6) Yadav, M.; Shrivastav, P. S. Bioanalysis 2011, 3 (9), 1007−1024. (7) Rocci, M. L., Jr; Collins, E.; Wagner-Caruso, K. E.; Gibbs, A. D.; Fellows, D. G. Bioanalysis 2011, 3 (9), 993−1000. (8) Tan, A.; Gagnon-Carignan, S.; Lachance, S.; Boudreau, N.; Lévesque, A.; Massé, R. Bioanalysis 2011, 3 (9), 1031−1038. (9) Fu, Y.; Li, W.; Smith, H. T.; Tse, F. L. Bioanalysis 2011, 3 (9), 967− 972. (10) Dicaire, C.; Bérubé, E. R.; Dumont, I.; Furtado, M.; Garofolo, F. Bioanalysis 2011, 3 (9), 973−982. (11) Côté, C.; Lahaie, M.; Latour, S.; Bergeron, M.; Dicaire, C.; Savoie, N.; Furtado, M.; Garofolo, F. Bioanalysis 2011, 3 (9), 935−938. (12) Li, W.; Zhang, J.; Tse, F. L. S. Biomed. Chromatogr. 2011, 25 (1− 2), 258−277. (13) de Boer, T.; Wieling, J. Bioanalysis 2011, 3 (9), 983−992.

2413


MS Reanalysis

Recommend Documents