- Email: [email protected]
Investigation of transition ion ratio variation for liquid chromatography-tandem mass spectrometry: A case study approach
Clinica Chimica Acta 486 (2018) 205–208
Contents lists available at ScienceDirect
Clinica Chimica Acta journal homepage: www.elsevier.com/locate/cca
Investigation of transition ion ratio variation for liquid chromatographytandem mass spectrometry: A case study approach☆
T
⁎,2
Dustin R. Bunch1, Adam J. McShane, Sihe Wang
Department of Laboratory Medicine, Cleveland Clinic, Cleveland, OH, United States of America
A R T I C LE I N FO
A B S T R A C T
Keywords: Transition ion ratio Liquid chromatography Mass spectrometry
Background: A transition ion ratio (TIR) is the ratio of one fragment over another from the same precursor and is frequently monitored in liquid chromatography-tandem mass spectrometry (LC-MS/MS) assays for analyte identification. The Clinical and Laboratory Standards Institute (CLSI) C50-A guidelines give a static percent allowable TIR deviation based on the TIR level. Anecdotally, we observed failures of these rules for some of our LC-MS/MS assays. We determined what parameters may affect TIRs in a clinical setting and whether TIR variations may be analyte, matrix, instrument service, and/or concentration dependent. Methods: Data was collected from the validation and selected periods after implementation for urine benzodiazepines (7 analytes) and plasma azole antifungals (6 analytes). TIRs for the calibrators and quality control materials on a Thermo TSQ™ Quantum Ultra from July 2016 to February 2017 for benzodiazepines in urine and Thermo TSQ™ Vantage from May 2016 to Oct 2016 for azoles in serum were monitored. Results: The statistically significant day-to-day TIR shift ranged from 5.7 to 27.0% of the days studies for benzodiazepines and from 5.6 to 27.8% of the days studied for azoles excluding shifts caused by instrument services. Instrument service had significant impact on all benzodiazepines except oxazepam with p-values ranging from 1.79 × 10−6 to 1.53 × 10−39 and 4 of the 6 azoles (fluconazole, isavuconazole, voriconazole, and itraconazole) with (p from 7.89 × 10−3 to 1.98 × 10−12). Lorazepam, α-hydroxyalprazolam, and hydroxyitraconazole showed significant concentration dependent TIR variations. Conclusions: TIR variations may be affected by instrument services, and can be concentration and analyte dependent. Instead of using a static percent deviation rule, establishment of TIR variation criteria for each analyte during test development and validation may provide a more useful tool for analyte identification.
1. Introduction Mass spectrometry (MS) is increasingly used in the clinical laboratory [1–6]. The most common type of MS in the clinical laboratory for targeted quantitation is triple quadrupole MS operating in selected reaction monitoring (SRM) mode. One precursor-to-product ion is termed a transition or product ion. Transition ions are thought to be a physical property of a compound of interest (COI). If one maintains the same MS settings, a MS should produce consistent ion fragmentation. By measuring multiple transition ions, transition ion ratios (TIR) can be
produced that reflect the COI. If used correctly, monitoring the TIR improves the specificity of a liquid chromatography/tandem mass spectrometry (LC- MS/MS) assay. These TIRs allow mass spectroscopists to determine if a peak is truly a COI or an analytical interferent [7]. However, for isobaric compounds with very similar structures these TIRs may not provide sufficient specificity. TIR guidelines in clinical laboratories are available through Clinical and Laboratory Standards Institute (CLSI) C50-A guidelines [8]. These guidelines give a static percent allowable TIR deviation based on the TIR level [8]. As an example, a TIR of 10–20% has a static allowable
Abbreviations: UAMCLZ, 7-aminoclonazepam; UOHALP, α-hydroxyalprazolam; UOHTR, α-hydroxytriazolam; AMU, atomic mass unit; CLSI, Clinical and Laboratory Standards Institute; COI, compound of interest; FLU, fluconazole; OHITRA, hydroxyitraconazole; ISA, isavuconazole; ITRA, itraconazole; ULORZP, lorazepam; UNDIAZ, nordiazepam; UOXAZP, oxazepam; POSA, posaconazole; SAD, static allowable deviation; SRM, selected reaction monitoring; UTEMZ, temazepam; TIR, transition ion ratio; VOR, voriconazole ☆ A portion this study was presented at the 69th AACC Annual Meeting in 2017. ⁎ Corresponding author at: LL3-140, 9500 Euclid Ave, Cleveland, OH 44195, United States of America. E-mail addresses: [email protected], [email protected] (S. Wang). 1 Current address: Department of Laboratory Medicine, Yale-New Haven Hospital, New Haven, CT. 2 Current address: Pathology and Laboratory Medicine, Akron Children's Hospital, Akron, OH. https://doi.org/10.1016/j.cca.2018.08.009 Received 4 May 2018; Received in revised form 12 June 2018; Accepted 7 August 2018 Available online 09 August 2018 0009-8981/ © 2018 Elsevier B.V. All rights reserved.
Clinica Chimica Acta 486 (2018) 205–208
D.R. Bunch et al.
Fig. 1. Benzodiazepine Normalized TIRs. Average daily TIR were calculated then a min-max normalization was performed. Min-max normalization sets the min to zero and the max to 1. Each COI is stacked on the y-axis and the x-axis is day number. Statistical significance was identified using two tailed student's t-test. Dashed line indicates instrument service performed to the instrument. * = p-value < 0.05; ** = p-value < 0.01.
Fig. 2. Azole Normalized TIRs. Average daily TIR were calculated then a min-max normalization was performed. Statistical significance was identified using two tailed student's t-test. Dashed line indicates instrument service performed to the instrument. * = p-value < 0.05; ** = p-value < 0.01.
spiked human urine with known levels (calibrators and quality controls [QC]) of 7-aminoclonazepam (UAMCLZ; precursor ion: 286 m/z→ ;product ion 1: 121 m/z; product ion 2: 222 m/z), lorazepam (ULORZP; 321→229;194), nordiazepam (UNDIAZ; 271→140;208), α-hydroxyalprazolam (UOHALP; 325→279;243), α-hydroxytriazolam (UOHTRI; 359→176;242), oxazepam (UOXAZP; 287→241;104), and temazepam (UTEMZ; 301→177;255) and was analyzed by LC-MS/MS with patient and validation samples. Briefly, internal standard solution and urine sample were mixed and enzymatically hydrolyzed at 60 °C with red abalone β-glucuronidase. Then the mixture was diluted with methanol and water prior to direct injection. Calibrator levels were 50, 100, 200, 500, 1000, 5000 ng/ml and QC values were 200 and 1000 ng/ ml. Data analyzed contained TIRs for the calibrators and QCs on a Thermo TSQ™ Quantum Ultra from July 2016 to February 2017. Azole antifungal analysis [10], the second data set, measured fluconazole
deviation (SAD) of 30%. There could be circumstances where these SADs are too large for some COIs and too small for others; hence, the SAD may lead to improper interpretation of the COI or interference. These levels are in place under the assumption that on a specific instrument and for a certain COI these levels do not change. Previously, a study showed that TIRs vary based on ion intensity which was used as a proxy for concentration [9]. Anecdotally, we observed failures of these rules for some of our LC-MS/MS assays and began an investigation into the factors that have impacts on TIR variations.
2. Methods TIRs were collected from assay validation and selected periods after assay implementation. Data sets were non-contiguous, single day preparations. A benzodiazepine data set that contained multiple lots of 206
Clinica Chimica Acta 486 (2018) 205–208
D.R. Bunch et al.
confirm between-day shifts with 2 significant value cut-offs considered, p-value of < 0.05 and < 0.01 (Figs. 1 and 2). The number of statistically significant TIR shift days excluding the service events identified below was normalized to the total number of days for comparison. Percent significant days, calculated based on p < 0.05, of TIR shifts ranged from 5.7 to 27.0% of the time for benzodiazepines and from 5.6 to 27.8% of the time for azoles. If one uses 20% as an acceptable cutoff for selecting stable TIRs, 6 of 7 benzodiazepines, while only 3 of the 6 azoles would be acceptable. However, if CLSI C50-A guidelines were used 4 of 7 benzodiazepines and 5 of 6 azoles would be acceptable. As shown in Table 1, if service events were not excluded there were more unacceptable TIR variations for benzodiazepines. For this analysis, the TIR was not reset to the daily mean calibrator TIR, which has the potential to reduce the number of significant TIR shift days. Based on this data, the CLSI C50-A TIR guidelines did not completely reflect the analyte TIR variations experienced during the routine clinical analysis. The second data analysis identified instrument service events as a cause of TIR shifts in min-max normalized datasets. Development and validation of these two assays spanned a significant time period during which preventative maintenance and other instrument service events occurred. Some of the compounds had shifts that were statistically relevant when single factor ANOVA was applied. Major benzodiazepine service events occurred between days 7 and 8 (Fig. 1). All benzodiazepines tested, except UOXAZP (p-value = 0.96), were shifted with pvalues ranging from 1.79 × 10−6 to 1.53 × 10−39. For the azole compounds, the instrument service event was between days 11 and 12 (Fig. 2) and 4 of the 6 compounds (FLU, ISA, VOR, and ITRA) had a statistically significant shift with p-values ranging from 7.89 × 10−3 to 1.98 × 10−12. This information may prompt the laboratory to check the TIR after an instrument service event and adjust accordingly. The final analysis was conducted to determine if TIR variation changes were concentration dependent. ULORZP, UOHALP, and OHITRA had increasing TIR variation changes (> 20%) as concentrations decreased (Figs. 3–4). The reason for the TIR variation change is not fully understood; but possible reasons include recalibration of the ion optics and pressure gauge in the collision cell of the MS. Another possibility is the absolute number of ions that arrive in the collision cells affects the TIR efficiency which could also explain the concentration-dependent affects. Typically, mass spectrometers receiving preventive maintenance or service are recalibrated. The recalibration may affect the ion optics allowing more ions to be transmitted. In addition, source parameters can affect absolute ion number through ion formation efficiency. These data indicate that TIRs may significantly change during normal operations, instrument service events, and/or variable analyte
Table 1 Initial TIR levels and CLSI C50A recommended maximum tolerances for all analytes. Analyte
UOXZAP ULORZP UNDIAZ UTEMZ UOHALP UOHTRI UAMCLZ POSA VOR FLU ISA ITRA OHITRA
Initial day TIR
5.7 12.8 12.9 13.2 14.9 18.7 26.4 31.4 76.3 78.7 83.8 87.2 114.3
Measured TIR
CLSI recommended
Min
Max
Min
Max
1.8 8.2 0.9 10.2 9.0 12.1 22.1 30.0 72.0 67.1 72.2 76.5 50.7
20.7 44.3 67.7 18.5 96.4 116.4 28.8 34.7 107.4 87.5 92.0 102.6 152.8
2.9 9.0 9.0 9.2 10.4 13.1 19.8 23.6 61.0 63.0 67.0 69.8 91.4
8.6 16.6 16.8 17.2 19.4 24.3 33.0 39.3 91.6 94.4 100.6 104.6 137.2
(FLU; 307→238;220), itraconazole (ITRA; 705→392;432), hydroxyitraconazole (OHITRA; 721→408;430), isavuconazole (ISA; 438→ 224;127), posaconazole (POSA; 701→683;614), and voriconazole (VOR; 350→127;281). The azole data set used a single lot of spiked human serum with calibrators at 0.3, 0.6, 1.3, 2.5, 5.0, and 10.0 μg/ml for all except FLU which had calibrators at 0.5, 0.9, 1.9, 3.8, 7.5, 15.0, and 30.0 μg/ml. The azole QC concentrations were prepared in pooled patient serum at 0.7 and 5.8 μg/ml for all azoles except FLU, which were at 1.9 and 15.0 μg/ml. TIRs for the calibrators and QCs analyzed on a Thermo TSQ™ Vantage from May 2016 to Oct 2016 were obtained. All TIRs were extracted and compiled using a Perl script with statistics and graphs produced using Microsoft Office Excel and GraphPad Prism7. Three different data analyses were performed. First, the between-day variation for the ion ratios was investigated. All daily calibrators and QCs were compared to the next sequential day's calibrators and QCs using a two-tailed t-test. Secondly, ion ratio pre- and postinstrument service were compared using single factor ANOVA with a pvalue < 0.05 being considered significant. Finally, concentration dependent TIR variation change was investigated. TIR variation change that was > 20% at specific analyte concentrations was considered significant. 3. Results and discussion When performing an LC-MS/MS assay, the TIR day-to-day variation can have a significant impact on operation. The initial analysis was to
Fig. 3. Concentration Dependent Average Benzodiazepine TIRs. Concentration averaged TIR were calculated. Error bars are at the min and max TIR mean change for the specific concentration. 207
Clinica Chimica Acta 486 (2018) 205–208
D.R. Bunch et al.
dependent. During routine operation, means of TIR at each concentration level may be recalculated from the calibrators in the same batch. Operators should be aware of the limitations of TIR and utilize this tool properly. References [1] V.M. Carvalho, The coming of age of liquid chromatography coupled to tandem mass spectrometry in the endocrinology laboratory, J. Chromatogr. B Anal. Technol. Biomed. Life Sci. 883–884 (2012) 50–58. [2] H. Ketha, S. Kaur, S.K. Grebe, R.J. Singh, Clinical applications of LC-MS sex steroid assays: evolution of methodologies in the 21st century, Curr. Opin. Endocrinol. Diabetes Obes. 21 (3) (2014) 217–226. [3] G. la Marca, Mass spectrometry in clinical chemistry: the case of newborn screening, J. Pharm. Biomed. Anal. 101 (2014) 174–182. [4] K.S. Leung, B.M. Fong, LC-MS/MS in the routine clinical laboratory: has its time come? Anal. Bioanal. Chem. 406 (9–10) (2014) 2289–2301. [5] J.M. van den Ouweland, I.P. Kema, The role of liquid chromatography-tandem mass spectrometry in the clinical laboratory, J. Chromatogr. B Anal. Technol. Biomed. Life Sci. 883–884 (2012) 18–32. [6] A.H. Wu, D. French, Implementation of liquid chromatography/mass spectrometry into the clinical laboratory, Clin. Chim. Acta 420 (2013) 4–10. [7] C. Heideloff, A.J. McShane, D.R. Bunch, K. Lembright, S. Lawson, S. Wang, Monitoring two transitions by LC-MS/MS may not be sufficient to positively identify benzoylecgonine in patient urine samples, Clin. Chim. Acta 456 (2016) 67–68. [8] C.L.S.I. (CLSI), Mass Spectrometry in the Clinical Laboratory: General Principles and Guidance; Approved Guideline C50-A, Wayne, PA, USA, (2007). [9] H.G. Mol, P. Zomer, M. Garcia Lopez, R.J. Fussell, J. Scholten, A. de Kok, A. Wolheim, M. Anastassiades, A. Lozano, A. Fernandez Alba, Identification in residue analysis based on liquid chromatography with tandem mass spectrometry: experimental evidence to update performance criteria, Anal. Chim. Acta 873 (2015) 1–13. [10] A.J. McShane, S. Wang, Development and validation of a liquid chromatographytandem mass spectrometry assay for the simultaneous quantitation of 5 azole antifungals and 1 active metabolite, Clin. Chim. Acta 474 (2017) 8–13.
Fig. 4. Concentration Dependent Average Azole TIRs. Concentration averaged TIR were calculated for each analyte. Error bars are at the min and max TIR mean change for the specific concentration.
concentrations. During LC-MS/MS method development and validation, the TIRs should be monitored and acceptance criteria should be established for the assay. However, this approach cannot address other factors (e.g., instrument service) that may change the TIRs. Operationally, TIR mean shifts can be partially addressed by adjustment of the analyte TIR means based on the calibrators performed in the same batch. More research is warranted to study the mechanism of changing TIRs in LC-MS/MS assays affected by different factors. At this stage, we recommend that TIRs should be established for each analyte during assay development and validation, which can be concentration
208
Contents lists available at ScienceDirect
Clinica Chimica Acta journal homepage: www.elsevier.com/locate/cca
Investigation of transition ion ratio variation for liquid chromatographytandem mass spectrometry: A case study approach☆
T
⁎,2
Dustin R. Bunch1, Adam J. McShane, Sihe Wang
Department of Laboratory Medicine, Cleveland Clinic, Cleveland, OH, United States of America
A R T I C LE I N FO
A B S T R A C T
Keywords: Transition ion ratio Liquid chromatography Mass spectrometry
Background: A transition ion ratio (TIR) is the ratio of one fragment over another from the same precursor and is frequently monitored in liquid chromatography-tandem mass spectrometry (LC-MS/MS) assays for analyte identification. The Clinical and Laboratory Standards Institute (CLSI) C50-A guidelines give a static percent allowable TIR deviation based on the TIR level. Anecdotally, we observed failures of these rules for some of our LC-MS/MS assays. We determined what parameters may affect TIRs in a clinical setting and whether TIR variations may be analyte, matrix, instrument service, and/or concentration dependent. Methods: Data was collected from the validation and selected periods after implementation for urine benzodiazepines (7 analytes) and plasma azole antifungals (6 analytes). TIRs for the calibrators and quality control materials on a Thermo TSQ™ Quantum Ultra from July 2016 to February 2017 for benzodiazepines in urine and Thermo TSQ™ Vantage from May 2016 to Oct 2016 for azoles in serum were monitored. Results: The statistically significant day-to-day TIR shift ranged from 5.7 to 27.0% of the days studies for benzodiazepines and from 5.6 to 27.8% of the days studied for azoles excluding shifts caused by instrument services. Instrument service had significant impact on all benzodiazepines except oxazepam with p-values ranging from 1.79 × 10−6 to 1.53 × 10−39 and 4 of the 6 azoles (fluconazole, isavuconazole, voriconazole, and itraconazole) with (p from 7.89 × 10−3 to 1.98 × 10−12). Lorazepam, α-hydroxyalprazolam, and hydroxyitraconazole showed significant concentration dependent TIR variations. Conclusions: TIR variations may be affected by instrument services, and can be concentration and analyte dependent. Instead of using a static percent deviation rule, establishment of TIR variation criteria for each analyte during test development and validation may provide a more useful tool for analyte identification.
1. Introduction Mass spectrometry (MS) is increasingly used in the clinical laboratory [1–6]. The most common type of MS in the clinical laboratory for targeted quantitation is triple quadrupole MS operating in selected reaction monitoring (SRM) mode. One precursor-to-product ion is termed a transition or product ion. Transition ions are thought to be a physical property of a compound of interest (COI). If one maintains the same MS settings, a MS should produce consistent ion fragmentation. By measuring multiple transition ions, transition ion ratios (TIR) can be
produced that reflect the COI. If used correctly, monitoring the TIR improves the specificity of a liquid chromatography/tandem mass spectrometry (LC- MS/MS) assay. These TIRs allow mass spectroscopists to determine if a peak is truly a COI or an analytical interferent [7]. However, for isobaric compounds with very similar structures these TIRs may not provide sufficient specificity. TIR guidelines in clinical laboratories are available through Clinical and Laboratory Standards Institute (CLSI) C50-A guidelines [8]. These guidelines give a static percent allowable TIR deviation based on the TIR level [8]. As an example, a TIR of 10–20% has a static allowable
Abbreviations: UAMCLZ, 7-aminoclonazepam; UOHALP, α-hydroxyalprazolam; UOHTR, α-hydroxytriazolam; AMU, atomic mass unit; CLSI, Clinical and Laboratory Standards Institute; COI, compound of interest; FLU, fluconazole; OHITRA, hydroxyitraconazole; ISA, isavuconazole; ITRA, itraconazole; ULORZP, lorazepam; UNDIAZ, nordiazepam; UOXAZP, oxazepam; POSA, posaconazole; SAD, static allowable deviation; SRM, selected reaction monitoring; UTEMZ, temazepam; TIR, transition ion ratio; VOR, voriconazole ☆ A portion this study was presented at the 69th AACC Annual Meeting in 2017. ⁎ Corresponding author at: LL3-140, 9500 Euclid Ave, Cleveland, OH 44195, United States of America. E-mail addresses: [email protected], [email protected] (S. Wang). 1 Current address: Department of Laboratory Medicine, Yale-New Haven Hospital, New Haven, CT. 2 Current address: Pathology and Laboratory Medicine, Akron Children's Hospital, Akron, OH. https://doi.org/10.1016/j.cca.2018.08.009 Received 4 May 2018; Received in revised form 12 June 2018; Accepted 7 August 2018 Available online 09 August 2018 0009-8981/ © 2018 Elsevier B.V. All rights reserved.
Clinica Chimica Acta 486 (2018) 205–208
D.R. Bunch et al.
Fig. 1. Benzodiazepine Normalized TIRs. Average daily TIR were calculated then a min-max normalization was performed. Min-max normalization sets the min to zero and the max to 1. Each COI is stacked on the y-axis and the x-axis is day number. Statistical significance was identified using two tailed student's t-test. Dashed line indicates instrument service performed to the instrument. * = p-value < 0.05; ** = p-value < 0.01.
Fig. 2. Azole Normalized TIRs. Average daily TIR were calculated then a min-max normalization was performed. Statistical significance was identified using two tailed student's t-test. Dashed line indicates instrument service performed to the instrument. * = p-value < 0.05; ** = p-value < 0.01.
spiked human urine with known levels (calibrators and quality controls [QC]) of 7-aminoclonazepam (UAMCLZ; precursor ion: 286 m/z→ ;product ion 1: 121 m/z; product ion 2: 222 m/z), lorazepam (ULORZP; 321→229;194), nordiazepam (UNDIAZ; 271→140;208), α-hydroxyalprazolam (UOHALP; 325→279;243), α-hydroxytriazolam (UOHTRI; 359→176;242), oxazepam (UOXAZP; 287→241;104), and temazepam (UTEMZ; 301→177;255) and was analyzed by LC-MS/MS with patient and validation samples. Briefly, internal standard solution and urine sample were mixed and enzymatically hydrolyzed at 60 °C with red abalone β-glucuronidase. Then the mixture was diluted with methanol and water prior to direct injection. Calibrator levels were 50, 100, 200, 500, 1000, 5000 ng/ml and QC values were 200 and 1000 ng/ ml. Data analyzed contained TIRs for the calibrators and QCs on a Thermo TSQ™ Quantum Ultra from July 2016 to February 2017. Azole antifungal analysis [10], the second data set, measured fluconazole
deviation (SAD) of 30%. There could be circumstances where these SADs are too large for some COIs and too small for others; hence, the SAD may lead to improper interpretation of the COI or interference. These levels are in place under the assumption that on a specific instrument and for a certain COI these levels do not change. Previously, a study showed that TIRs vary based on ion intensity which was used as a proxy for concentration [9]. Anecdotally, we observed failures of these rules for some of our LC-MS/MS assays and began an investigation into the factors that have impacts on TIR variations.
2. Methods TIRs were collected from assay validation and selected periods after assay implementation. Data sets were non-contiguous, single day preparations. A benzodiazepine data set that contained multiple lots of 206
Clinica Chimica Acta 486 (2018) 205–208
D.R. Bunch et al.
confirm between-day shifts with 2 significant value cut-offs considered, p-value of < 0.05 and < 0.01 (Figs. 1 and 2). The number of statistically significant TIR shift days excluding the service events identified below was normalized to the total number of days for comparison. Percent significant days, calculated based on p < 0.05, of TIR shifts ranged from 5.7 to 27.0% of the time for benzodiazepines and from 5.6 to 27.8% of the time for azoles. If one uses 20% as an acceptable cutoff for selecting stable TIRs, 6 of 7 benzodiazepines, while only 3 of the 6 azoles would be acceptable. However, if CLSI C50-A guidelines were used 4 of 7 benzodiazepines and 5 of 6 azoles would be acceptable. As shown in Table 1, if service events were not excluded there were more unacceptable TIR variations for benzodiazepines. For this analysis, the TIR was not reset to the daily mean calibrator TIR, which has the potential to reduce the number of significant TIR shift days. Based on this data, the CLSI C50-A TIR guidelines did not completely reflect the analyte TIR variations experienced during the routine clinical analysis. The second data analysis identified instrument service events as a cause of TIR shifts in min-max normalized datasets. Development and validation of these two assays spanned a significant time period during which preventative maintenance and other instrument service events occurred. Some of the compounds had shifts that were statistically relevant when single factor ANOVA was applied. Major benzodiazepine service events occurred between days 7 and 8 (Fig. 1). All benzodiazepines tested, except UOXAZP (p-value = 0.96), were shifted with pvalues ranging from 1.79 × 10−6 to 1.53 × 10−39. For the azole compounds, the instrument service event was between days 11 and 12 (Fig. 2) and 4 of the 6 compounds (FLU, ISA, VOR, and ITRA) had a statistically significant shift with p-values ranging from 7.89 × 10−3 to 1.98 × 10−12. This information may prompt the laboratory to check the TIR after an instrument service event and adjust accordingly. The final analysis was conducted to determine if TIR variation changes were concentration dependent. ULORZP, UOHALP, and OHITRA had increasing TIR variation changes (> 20%) as concentrations decreased (Figs. 3–4). The reason for the TIR variation change is not fully understood; but possible reasons include recalibration of the ion optics and pressure gauge in the collision cell of the MS. Another possibility is the absolute number of ions that arrive in the collision cells affects the TIR efficiency which could also explain the concentration-dependent affects. Typically, mass spectrometers receiving preventive maintenance or service are recalibrated. The recalibration may affect the ion optics allowing more ions to be transmitted. In addition, source parameters can affect absolute ion number through ion formation efficiency. These data indicate that TIRs may significantly change during normal operations, instrument service events, and/or variable analyte
Table 1 Initial TIR levels and CLSI C50A recommended maximum tolerances for all analytes. Analyte
UOXZAP ULORZP UNDIAZ UTEMZ UOHALP UOHTRI UAMCLZ POSA VOR FLU ISA ITRA OHITRA
Initial day TIR
5.7 12.8 12.9 13.2 14.9 18.7 26.4 31.4 76.3 78.7 83.8 87.2 114.3
Measured TIR
CLSI recommended
Min
Max
Min
Max
1.8 8.2 0.9 10.2 9.0 12.1 22.1 30.0 72.0 67.1 72.2 76.5 50.7
20.7 44.3 67.7 18.5 96.4 116.4 28.8 34.7 107.4 87.5 92.0 102.6 152.8
2.9 9.0 9.0 9.2 10.4 13.1 19.8 23.6 61.0 63.0 67.0 69.8 91.4
8.6 16.6 16.8 17.2 19.4 24.3 33.0 39.3 91.6 94.4 100.6 104.6 137.2
(FLU; 307→238;220), itraconazole (ITRA; 705→392;432), hydroxyitraconazole (OHITRA; 721→408;430), isavuconazole (ISA; 438→ 224;127), posaconazole (POSA; 701→683;614), and voriconazole (VOR; 350→127;281). The azole data set used a single lot of spiked human serum with calibrators at 0.3, 0.6, 1.3, 2.5, 5.0, and 10.0 μg/ml for all except FLU which had calibrators at 0.5, 0.9, 1.9, 3.8, 7.5, 15.0, and 30.0 μg/ml. The azole QC concentrations were prepared in pooled patient serum at 0.7 and 5.8 μg/ml for all azoles except FLU, which were at 1.9 and 15.0 μg/ml. TIRs for the calibrators and QCs analyzed on a Thermo TSQ™ Vantage from May 2016 to Oct 2016 were obtained. All TIRs were extracted and compiled using a Perl script with statistics and graphs produced using Microsoft Office Excel and GraphPad Prism7. Three different data analyses were performed. First, the between-day variation for the ion ratios was investigated. All daily calibrators and QCs were compared to the next sequential day's calibrators and QCs using a two-tailed t-test. Secondly, ion ratio pre- and postinstrument service were compared using single factor ANOVA with a pvalue < 0.05 being considered significant. Finally, concentration dependent TIR variation change was investigated. TIR variation change that was > 20% at specific analyte concentrations was considered significant. 3. Results and discussion When performing an LC-MS/MS assay, the TIR day-to-day variation can have a significant impact on operation. The initial analysis was to
Fig. 3. Concentration Dependent Average Benzodiazepine TIRs. Concentration averaged TIR were calculated. Error bars are at the min and max TIR mean change for the specific concentration. 207
Clinica Chimica Acta 486 (2018) 205–208
D.R. Bunch et al.
dependent. During routine operation, means of TIR at each concentration level may be recalculated from the calibrators in the same batch. Operators should be aware of the limitations of TIR and utilize this tool properly. References [1] V.M. Carvalho, The coming of age of liquid chromatography coupled to tandem mass spectrometry in the endocrinology laboratory, J. Chromatogr. B Anal. Technol. Biomed. Life Sci. 883–884 (2012) 50–58. [2] H. Ketha, S. Kaur, S.K. Grebe, R.J. Singh, Clinical applications of LC-MS sex steroid assays: evolution of methodologies in the 21st century, Curr. Opin. Endocrinol. Diabetes Obes. 21 (3) (2014) 217–226. [3] G. la Marca, Mass spectrometry in clinical chemistry: the case of newborn screening, J. Pharm. Biomed. Anal. 101 (2014) 174–182. [4] K.S. Leung, B.M. Fong, LC-MS/MS in the routine clinical laboratory: has its time come? Anal. Bioanal. Chem. 406 (9–10) (2014) 2289–2301. [5] J.M. van den Ouweland, I.P. Kema, The role of liquid chromatography-tandem mass spectrometry in the clinical laboratory, J. Chromatogr. B Anal. Technol. Biomed. Life Sci. 883–884 (2012) 18–32. [6] A.H. Wu, D. French, Implementation of liquid chromatography/mass spectrometry into the clinical laboratory, Clin. Chim. Acta 420 (2013) 4–10. [7] C. Heideloff, A.J. McShane, D.R. Bunch, K. Lembright, S. Lawson, S. Wang, Monitoring two transitions by LC-MS/MS may not be sufficient to positively identify benzoylecgonine in patient urine samples, Clin. Chim. Acta 456 (2016) 67–68. [8] C.L.S.I. (CLSI), Mass Spectrometry in the Clinical Laboratory: General Principles and Guidance; Approved Guideline C50-A, Wayne, PA, USA, (2007). [9] H.G. Mol, P. Zomer, M. Garcia Lopez, R.J. Fussell, J. Scholten, A. de Kok, A. Wolheim, M. Anastassiades, A. Lozano, A. Fernandez Alba, Identification in residue analysis based on liquid chromatography with tandem mass spectrometry: experimental evidence to update performance criteria, Anal. Chim. Acta 873 (2015) 1–13. [10] A.J. McShane, S. Wang, Development and validation of a liquid chromatographytandem mass spectrometry assay for the simultaneous quantitation of 5 azole antifungals and 1 active metabolite, Clin. Chim. Acta 474 (2017) 8–13.
Fig. 4. Concentration Dependent Average Azole TIRs. Concentration averaged TIR were calculated for each analyte. Error bars are at the min and max TIR mean change for the specific concentration.
concentrations. During LC-MS/MS method development and validation, the TIRs should be monitored and acceptance criteria should be established for the assay. However, this approach cannot address other factors (e.g., instrument service) that may change the TIRs. Operationally, TIR mean shifts can be partially addressed by adjustment of the analyte TIR means based on the calibrators performed in the same batch. More research is warranted to study the mechanism of changing TIRs in LC-MS/MS assays affected by different factors. At this stage, we recommend that TIRs should be established for each analyte during assay development and validation, which can be concentration
208