- Email: [email protected]
Simple dilute-and-shoot method for urinary vanillylmandelic acid and homovanillic acid by liquid chromatography tandem mass spectrometry
Accepted Manuscript Simple dilute-and-shoot method for urinary vanillylmandelic acid and homovanillic acid by liquid chromatography tandem mass spectrometry
Zlatuse D. Clark, Jeaneah M. Cutler, Igor Y. Pavlov, Frederick G. Strathmann, Elizabeth L. Frank PII: DOI: Reference:
S0009-8981(17)30078-5 doi: 10.1016/j.cca.2017.03.004 CCA 14674
To appear in:
Clinica Chimica Acta
Received date: Revised date: Accepted date:
13 August 2016 1 March 2017 1 March 2017
Please cite this article as: Zlatuse D. Clark, Jeaneah M. Cutler, Igor Y. Pavlov, Frederick G. Strathmann, Elizabeth L. Frank , Simple dilute-and-shoot method for urinary vanillylmandelic acid and homovanillic acid by liquid chromatography tandem mass spectrometry. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Cca(2016), doi: 10.1016/j.cca.2017.03.004
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Simple dilute-and-shoot method for urinary vanillylmandelic acid and homovanillic acid by liquid chromatography tandem mass spectrometry
Zlatuse D. Clarka, Jeaneah M. Cutlerb, Igor Y. Pavlova, Frederick G. Strathmannc,
IP
b
c
ARUP Laboratories, Salt Lake City, UT 84108
CR
ARUP Institute for Clinical and Experimental Pathology, Salt Lake City, UT 84108
Department of Pathology, University of Utah Health Sciences Center, Salt Lake City,
US
a
T
Elizabeth L. Frankc*
AN
UT 84112
Corresponding author: Elizabeth L. Frank, c/o ARUP Laboratories, Inc., 500 Chipeta
M
Way, Mail Code 115, Salt Lake City, UT 84108, phone +1 (801) 583-2787, ext. 1-2087,
ED
email: [email protected].
PT
Previous presentation of the manuscript: Parts of these data were presented at the
CE
Mass Spectrometry: Applications to the Clinical Lab (MSACL) conference in San Diego, CA, March 28-April 1, 2015.
AC
List of abbreviations: CLSI, Clinical and Laboratory Standards Institute; DLLME, dispersive liquid-liquid microextraction; ECD, electrochemical detection; ESI, electrospray ionization; HVA, homovanillic acid; LLOQ, lower limit of quantification; MRM, multiple reaction monitoring; QIR, qualitative ion ratio; SIL, stable isotope label; VMA, vanillylmandelic acid
1
ACCEPTED MANUSCRIPT Abstract Background: Neuroblastomas are pediatric tumors characterized by overproduction of catecholamines. The catecholamine metabolites, vanillylmandelic acid (VMA) and homovanillic acid (HVA), are used in clinical evaluation of neuroblastoma. Tandem mass spectrometry
T
(LC-MS/MS) is an effective analytical method for measurement of VMA and HVA in urine.
IP
Methods: Dilute-and-shoot sample preparation was performed in a 96-well format using a liquid
CR
handler. Chromatographic separation was achieved using a reverse phase column; detection was accomplished by triple quadrupole mass spectrometry with electrospray ionization in positive
US
mode. Data were acquired by multiple reaction monitoring. Two transitions, quantifier and
AN
qualifier, were monitored for each analyte and its stable isotope-labeled internal standard. Analytical specificity studies were performed.
M
Results: Injection-to-injection time was 4 min. The method was validated for linearity, limit of
ED
quantification, imprecision, accuracy, and interference. Linearity was 0.5 – 100 mg/l for both analytes. Within-run, between-day, and total imprecision were 1.0 – 4.1% for VMA and 0.8 –
PT
3.8% for HVA. The method correlated well with our established HPLC method. Interferences
CE
precluding quantitation of VMA in 3% of specimens were reduced significantly (to 0.1% of specimens) using a modified LC gradient to reanalyze affected samples.
AC
Conclusions: A simple, robust, economical, fast LC-MS/MS method was developed and validated for measurement of urinary VMA and HVA. Keywords: Liquid chromatography; mass spectrometry; vanillylmandelic acid; homovanillic acid; urine
2
ACCEPTED MANUSCRIPT 1. Introduction Laboratory measurement of vanillylmandelic acid (VMA), the end product of epinephrine and norepinephrine metabolism, and homovanillic acid (HVA), a terminal metabolite of dopamine, is used to support clinical diagnosis and management of
IP
T
catecholamine-secreting neurochromaffin tumors such as neuroblastomas [1,2].
CR
Neuroblastoma typically occurs in children and accounts for 7-10% of childhood cancer; approximately 90% of these tumors produce catecholamines [2]. The catecholamine
US
metabolites, VMA and HVA, are used in combination with other prognostic factors to
AN
improve tumor detection and assess treatment response [3,4]. Biochemical tests for VMA and HVA typically are performed using HPLC or GC-
M
MS platforms. LC-MS/MS technology offers several analytical performance advantages
ED
for these small molecules [5]. Use of LC-MS/MS can minimize sample preparation, decrease analysis time, and improve selectivity. Tandem mass spectrometry methods
PT
have been developed to measure urinary VMA [6] and HVA [7] as single analytes and in
CE
panels with other biomarkers related to neuroendocrine tumors [8-10] and other disorders
AC
[11-13]. Assays using plasma and/or serum [10,14,15] circumvent challenges with pediatric 24-h urine collections and variation inherent in random urine specimens. Combined measurement of the analytes allows calculation of the VMA/HVA ratio, which can provide additional prognostic value [16]. In our laboratory, an HPLC method with electrochemical detection (ECD) had been used to analyze urinary VMA and HVA for nearly 30 y. The analytes were measured concurrently using a coulometric electrode array system. Serial coulometric electrodes, 3
ACCEPTED MANUSCRIPT set at increasing potentials, provided electrochemical resolution to augment chromatographic separation of VMA, HVA, and other analytes [17]. Sample analysis was accomplished in 21 minutes. In recent years, increased test volume and frequency of interfering substances in specimens received for analysis, as well as aging of the
IP
T
instruments, limited usefulness of the assay. Modifications to accommodate more
CR
numerous requests for VMA measurement, which could be analyzed separately in approximately five minutes, were implemented, but did not resolve the issues completely.
AN
US
We identified mass spectrometry as a potential replacement method.
2.1. Chemicals and reagents
M
2. Experimental procedures
ED
HPLC-grade methanol was from Burdick & Jackson; formic acid (98%) was from
PT
EM Science; vanillylmandelic acid (VMA, DL-4-hydroxy-3-methoxymandelic acid) and homovanillic acid (HVA, 4-hydroxy-3-methoxyphenyl acetic acid) were from Sigma-
CE
Aldr; (±)-4-hydroxy-3-methoxy-d3-mandelic acid (VMA-d3) was from C/D/N Isotopes;
AC
and homovanillic acid [ring-13C6, 4-hydroxy-18O] (HVA-13C6 18O) was from Cambridge Isotope Laboratories, Inc. Lyphocheck® Quantitative Controls Level I and Level II were purchased from Bio-Rad Clinical Division (Hercules, CA). Clinical Laboratory Reagent (CLR) water was used throughout the study. Synthetic urine was prepared from urea, sodium chloride, creatinine hydrochloride, dipotassium hydrogen phosphate, and sodium dihydrogen phosphate (Sigma-Aldrich). Each lot of synthetic urine was validated by testing; no VMA or HVA was detected in the reagent. 4
ACCEPTED MANUSCRIPT 2.2. Calibration standards and quality control solutions Combined VMA and HVA stock solution was prepared in water at 1000 mg/l. Calibration standards were prepared at 0.5, 2, 10, 50 and 100 mg/l concentrations by dilution of the 1000 mg/l combined stock solution with synthetic urine. VMA-d3 and
IP
T
HVA-13C6 18O individual internal standard stock solutions (1000 mg/l) and combined
CR
working internal standard solution (5 mg/l) were prepared in water. For MS optimization and signal suppression experiments, the stock solutions were diluted with 0.05% (v/v)
US
formic acid in water to 50 and 10 mg/l, respectively.
AN
For the primary VMA-HVA method validation, control materials were prepared at 4 concentrations (Table 2): Levels I and II were Bio-Rad Lyphocheck® Quantitative
M
Controls, reconstituted in 10 ml of CLR water. Levels III and IV were prepared from
ED
patient specimen pools. For the alternate VMA method validation, control materials were
PT
prepared at three concentrations: low (0.8 mg/l) and high (80 mg/l) by dilution of the combined VMA and HVA stock solutions in water; a mid-range control was made by
CE
reconstitution of Lyphocheck® Quantitative Urine Control Level II with water. Aliquots
AC
of stock, calibration, and control solutions were stored frozen at -70 °C. Working internal standard solution was prepared fresh daily by dilution of the stock solutions.
2.3. Sample collection and preparation Urine specimens were selected from samples submitted for clinical testing. Samples were de-identified and managed according to guidelines approved by the Institutional Review Board of the University of Utah.Using a Tecan liquid handler, 50 l of working 5
ACCEPTED MANUSCRIPT internal standard solution and 400 µl of 0.05% (v/v) formic acid in water were added to 50 µl of calibration standards, quality controls, and urine specimens in a 96-well plate and mixed. The plate was centrifuged at 3500×g under refrigeration and 250 l aliquots from each well were transferred to a new 96-well plate from which injections onto the
2.4. LC-MS/MS instrumentation and conditions
CR
IP
T
analytical column were performed.
US
Analysis was performed on a Waters Xevo™ TQ MS Acquity UPLC® instrument,
AN
which consisted of a mass spectrometer with an electrospray ionization (ESI) probe in positive mode and an ACQUITY UPLC® system including a binary solvent manager,
M
sample manager, sample organizer, and column manager. Data were acquired using
ED
multiple reaction monitoring (MRM). Instrument control and data analysis and
PT
quantitation were performed using Waters MassLynx™ software (v4.1). Chromatographic separation was achieved using a Kinetex XB-C18 column (2.1 ×
CE
50 mm, 100 Å, 1.7 µm particles; Phenomenex) with an in-line filter assembly (Waters,
AC
Milford, MA) and 0.05% (v/v) formic acid in water as mobile phase A and 0.05% (v/v) formic acid in methanol as mobile phase B. Gradient elution with 5-30% B in 0 - 2.3 min, followed by a cleanout step to 95% B (2.3 - 2.6 min), was employed in the primary VMA-HVA method. Isocratic elution at 0% B for 1.7 min, followed by a step gradient to 95% B (1.7 - 2.2 min) was used in the alternate VMA method. For both methods, the mobile phase flow rate was 400 µL/min, the column temperature was 30 C, the sample
6
ACCEPTED MANUSCRIPT manager temperature was 10 C, the injection volume was 5 µl, and the injection-toinjection time was 4 min. Electrospray ionization was performed in positive mode. Purified nitrogen was used for the desolvation (1000 l/h) and cone (25 l/h) gases; purified argon was the collision gas
IP
T
(0.20 ml/min). The source temperature was 150 C and the desolvation temperature was
CR
600 C. The capillary voltage was set to 3000 V and the extractor to 3 V. MRM qualifier and quantifier transitions, cone voltages, and collision energies for both analytes and their
US
internal standards in the primary VMA-HVA method are given in Table 1. Dwell time for
monitored with 70 ms dwell time each.
AN
all transitions was 32 ms. For the alternate VMA method, only VMA transitions were
M
A qualitative ion ratio (QIR) was used to confirm identity of the analytes. The QIRs
ED
were calculated as ratios of the peak area for the quantifier transition to the peak area for
PT
the qualifier transition for both the analytes and the internal standards. Patient specimen QIR acceptability range was established as 30% of the mean value of the QIRs obtained
AC
CE
from the calibration standards.
2.5. Method validation 2.5.1. Linearity Linearity was evaluated by analyzing samples prepared by dilution of a high concentration patient specimen pool with a low concentration patient specimen pool at 8 target concentrations ranging from 0.25 to 91.01 mg/l for VMA and from 0.37 to 90.65
7
ACCEPTED MANUSCRIPT mg/l for HVA. Samples were aliquoted in tubes in an amount sufficient for one day of testing and stored frozen at –70 ºC until analysis. A set of 4 replicates from each aliquot was analyzed on 4 separate days, totaling 16 replicates for each concentration. The
IP
T
acceptance criterion for linearity was 10% of expected concentration.
CR
2.5.2. Limit of quantification
Lower limit of quantification (LLOQ) was evaluated by analyzing samples prepared
US
by dilution of a de-identified patient specimen with synthetic urine at concentrations
AN
ranging from 0.18 to 2.91 mg/l for VMA and from 0.18 to 2.82 mg/l for HVA. Samples were aliquoted in tubes in an amount sufficient for one day of testing and stored frozen at
M
–70 ºC until analysis. A set of 4 replicates from each aliquot was analyzed on 4 separate
ED
days, totaling 16 replicates for each concentration. The criteria for establishing the LLOQ
CE
2.5.3. Imprecision
PT
concentration were CV ≤10% and accuracy within 15% of the expected value.
AC
Inter-assay imprecision was evaluated by analyzing control materials prepared at 4 concentrations over 4 days in 4 replicates per run, 1 run per day.
2.5.4. Method comparison The newly developed LC-MS/MS method was compared to the HPLC-ECD method in use. Specimens were stored at ≤–20 °C prior to analysis.
8
ACCEPTED MANUSCRIPT 2.5.5. Carryover Carryover was evaluated as described in EP Evaluator software. A set of high (H) and low (L) samples was assayed in the following consecutive order: 3 L, 2 H, 1 L, 2 H, 4 L, 2 H, 1 L, 2 H, 1 L, 2 H, 1 L. A patient specimen pool with VMA at 3.1 mg/l and
IP
T
HVA at 2.8 mg/l was used for a low concentration sample; a patient specimen pool
CR
spiked to 400 mg/l was used as the high concentration sample.
US
2.5.6. Evaluation of ion suppression
AN
Ion suppression by matrix components was evaluated as the presence of negative peaks at the retention times of VMA and HVA. Urine samples (n = 83) prepared without
M
internal standard were injected onto the LC column and individual 10 mg/l internal
ED
standard solutions were co-infused at 5 µl/min post-column as described by Annesley
PT
[18].
CE
2.5.7. Analytical specificity
AC
A number of endogenous and exogenous compounds were tested for interference in the assay. Individual aliquots of pooled urine with VMA and HVA concentrations of 2.1 mg/l were spiked with solutions of potential interferents at 10% of total volume. Recommendations from CLSI EP7-A2, Interference Testing in Clinical Chemistry, were used to guide prepared interferent concentrations [19]. Appropriate baseline samples were generated by spiking the urine pool with solvents (water, 0.1 mol/l HCl, 0.5 mol/l HCl, 0.1 mol/l NaOH, or methanol) to match the spiking solutions solvents and spiking 9
ACCEPTED MANUSCRIPT volumes. The interference samples were analyzed in 2 separate experiments. In the first, quantitative experiment, VMA and HVA concentrations were measured in a sample spiked with a potential interferent. The result was compared to the respective analyte concentration in a baseline sample. Deviation was calculated as the difference between
IP
T
the measured concentration in each interference sample and the respective baseline
CR
sample mean concentration divided by the baseline sample mean concentration. A substance was considered to interfere if the calculated deviation exceeded ±10%. In the
US
second, qualitative experiment, the interference samples were analyzed in MS scan mode
AN
to obtain retention times and m/z values for the potential interferents.
M
2.5.8. Reference intervals
ED
Reference intervals were evaluated according to the method of Hoffmann [20,21].
PT
Archived data for VMA and HVA measured in urine specimens using the new LCMS/MS method were assessed. Outlier values exceeding the third quartile plus three
CE
times the interquartile range were removed. Simplified computer calculations were
AC
performed with the assumption that maximum error (deviation of the cumulative frequencies curve from the linear regression) did not exceed 5% of the median value. Reference limits were calculated at 2.5% and 97.5% of the linear portion of the data for each defined group. Harris and Boyd [22] criteria were applied for partitioning by age, gender, and creatinine concentration (where applicable).
10
ACCEPTED MANUSCRIPT 2.5.9. Data analysis Statistical evaluation of the data was performed using Microsoft Excel and EP Evaluator software, release 10 (David G. Rhoads Associates, Inc.).
IP
T
3. Results
CR
3.1. Optimization of method parameters 3.1.1. MS parameter optimization
US
MS parameters were determined using post-column co-infusion of compound
AN
solutions into mobile phase solution at 10 µl/min. Both negative and positive ionization modes were explored. Negative mode did not exhibit reduced matrix effects. Positive
M
mode gave slightly higher signal for all compounds and was therefore used in the new
ED
method. While HVA and HVA-13C6 18O ionized to the expected protonated molecular ions, VMA and VMA-d3 readily lost water in the ESI source, yielding the more stable
PT
[M-H2O+H]+ ions, which were selected as the precursor ions. Due to the low molecular
CE
weight of VMA and HVA, the selection of useful MRM transitions was not large. During
AC
method development, all viable MRM transitions were carefully monitored and those with the least occurrence of interferences were included in the MS method (Table 1).
3.1.2. LC parameter optimization Several LC columns were tested for VMA and HVA separation: Waters HSS T3 (2.1×50 mm, 1.8 µm), Waters XBridge C18 (2.1×75 mm, 2.5 µm), Phenomenex Gemini C18 (3×100 mm, 3 µm) and Phenomenex Kinetex XB-C18 (3×50 mm, 2.6 µm and 11
ACCEPTED MANUSCRIPT 2.1×50, 1.7 µm). The Kinetex XB-C18 (3×50 mm, 2.6 µm and 2.1×50 mm, 1.7 µm) columns, with 0.05% formic acid in water as mobile phase A and 0.05% formic acid in methanol as mobile phase B, provided the best combination of retention, separation, and short run times for the compounds of interest. The LC gradient was optimized using post-
IP
T
column infusion of internal standards while analyzing urine specimens to minimize the
CR
impact of sample matrix on analyte quantitation. The injection-to-injection time was shortened from 21 min in the HPLC-ECD method to 4 min. Example chromatograms of
US
VMA and HVA standard solutions and of healthy and abnormal patient specimens are
ED
3.2.1. Linearity
M
3.2. Method validation results
AN
shown in Fig. 1.
PT
Observed concentrations for VMA and HVA were plotted against expected concentrations to give Eqs. (1) and (2), respectively. (1)
y = 0.996x + 0.004; observed error 0.4%
(2)
AC
HVA
y = 0.991x + 0.011; observed error 1.7%
CE
VMA
The assay was linear from 0.5 to 100 mg/l.
3.2.2. Limit of quantification Accuracy (within 15% of the expected value) and imprecision (CV ≤10%) criteria were exceeded at 0.18 mg/l for both VMA and HVA. The LLOQ was set at 0.5 mg/l, the concentration of the lowest calibration standard, for both analytes. 12
ACCEPTED MANUSCRIPT
3.2.3. Imprecision Within-run, between-day, and total imprecision for 4 quality control concentrations ranged from 1.0 – 4.1% for VMA and from 0.8 – 3.8% for HVA. Specific data are listed
CR
IP
T
in Table 2.
3.2.4. Method comparison
US
Specimens with analyte concentrations from 0.5 mg/l to 80 mg/l (the upper limit of
AN
linearity for the HPLC-ECD method) were assayed in the method comparison experiment. Acceptable agreement was observed between the HPLC-ECD method and
M
the newly developed LC-MS/MS method for both VMA and HVA. Deming regression
ED
equations for VMA, Eq. (3), and HVA, Eq. (4), standard errors, correlation coefficients, %bias, and number of specimens were:
y = 1.014x – 0.069; Sy/x = 0.924; R = 0.9904; %bias = –0.356%; n = 374
(3)
HVA
y = 1.153x – 0.983; Sy/x = 2.294; R = 0.9909; %bias = 5.661%; n = 97
(4)
CE
PT
VMA
AC
3.2.5. Carryover
At 400 mg/l, the calculated carryover values for VMA and HVA (0.07 and 0.01 mg/l, respectively) were less than the error limits (0.12 and 0.07 mg/l, respectively), defined as three times the SD of 2 consecutive low results. Carryover was not observed for either analyte.
13
ACCEPTED MANUSCRIPT 3.2.6. Evaluation of ion suppression As the developed method utilizes a dilute-and-shoot sample preparation, a thorough evaluation of matrix effects using a large number (83) of specimens was essential. While the number and intensity of suppression zones varied significantly from sample to
IP
T
sample, VMA and HVA peaks were not adversely affected in the majority of the
CR
specimens (Fig. 2). Distinct suppression zones observed at the retention times of the analytes and affecting analyte signal intensity were found in only 2 specimens for VMA
US
(Fig. 2C and D) and one specimen for HVA (Fig. 2D). The internal standards for both
AN
VMA and HVA elute very closely with their non-labeled analogs and should, therefore, compensate for signal suppression unless it is severe. Of the 83 samples evaluated, severe
PT
ED
less affected by matrix effects.
M
signal suppression was observed only for VMA in one sample (Fig. 2C). HVA was much
3.2.7. Analytical specificity
CE
Compounds, actual and recommended test concentrations, monoisotopic masses,
AC
ions observed in MS scans ([M+1] ions or most intense alternate ions), and retention times, as well as percent deviation of analyte concentrations in test samples versus respective baseline samples are shown in Table 3 for endogenous substances and vitamins and in Table 4 for exogenous substances. Three of the compounds tested eluted very close to VMA (tR = 0.85 min) and could potentially interfere with VMA quantitation at high concentrations. Carbidopa produced a peak (tR = 0.72 min) in the quantitative VMA transition and deviation of 5.7% at the 14
ACCEPTED MANUSCRIPT actual test concentration of 140 mg/l (about 10-fold higher than expected in urine after a 130 mg/l dose [23]). Phenylalanine (tR = 0.89 min) and isoetharine (tR = 0.82 min) induced deviations of -10.9% (at 130 mg/l, recommended test concentration 17 mg/l) and 14.2% (at 100 mg/l, estimated concentration in urine 0.4 mg/l [24]), respectively, but did
IP
T
not appear as peaks in any VMA transitions.
CR
Differential suppression of VMA and its slightly earlier eluting internal standard by the interferents is a most likely explanation for the >±10% VMA concentration deviation
US
observed with phenylalanine and isoetharine. Phenylalanine eluting 0.05 min after VMA
AN
would suppress the analyte more than the internal standard, thus producing a negative deviation, while isoetharine eluting 0.02 min before VMA would suppress the internal
M
standard more than the analyte, thus producing a positive deviation.
ED
None of the 62 tested compounds interfered with HVA quantitation. Caffeine (tR =
PT
2.19 min) and cocaine (tR = 2.12 min) eluted close to HVA (tR = 2.29 min) and could potentially interfere with HVA quantitation at very high concentrations. However, neither
CE
of these compounds appeared as peaks in any of the HVA MRM transitions or produced
AC
HVA deviation from baseline beyond ±10% at their respective test concentrations. In actual use, approximately 3% of VMA results could not be reported due to interference; only one specimen of more than 3000 tested showed an interference that prevented quantification of HVA. The majority of the VMA interferences produced internal standard peak areas below the lower acceptance limit (50% of the batch internal standard peak area median) due to signal suppression by matrix components at the retention time of VMA. 15
ACCEPTED MANUSCRIPT In an effort to improve assay performance, an alternate LC gradient was designed to resolve VMA interferences. Post-column infusion of a VMA-d3 solution was used as a tool to determine the position of suppression zones in six specimens with interference and to manipulate the LC gradient to separate VMA from the suppression zones. Isocratic
IP
T
elution at 0% B for 1.7 min, followed by a step gradient to 95% B (1.7 - 2.2 min) during
CR
which the column effluent (including HVA) is directed to waste was determined to be optimal. No additional method parameters were altered.
US
Validation of the alternate VMA method was performed to assess linearity,
AN
sensitivity, and imprecision (data not shown). All results met assay performance criteria. The alternate VMA method was compared with the primary VMA-HVA method using
M
196 specimens with concentrations spanning the analytical measurement range. Good
ED
agreement between the 2 methods was observed. The Deming regression equation, standard error, and correlation coefficient were: y = 0.978x + 0.092; Sy/x = 0.344; R =
PT
0.9996; Bias = –0.742%.
CE
Reanalysis of 6 specimens that showed interference produced VMA peak areas
AC
approximately 6- to 20-fold higher using the alternate method, indicating that the alternate LC gradient mitigated signal suppression experienced by VMA in the primary method. Use of the alternate VMA method reduced the incidence of non-reportable results from 3.0% to 0.1%. Results for endogenous and exogenous substances evaluated for assay interference using the alternate method parameters are shown in Tables 3 and 4.
16
ACCEPTED MANUSCRIPT 3.2.8. Reference intervals VMA and HVA results extracted from the database were separated according to collection type (random or 24 h) and grouped by known age and gender for evaluation of reference limits. Specimens obtained from adults were primarily 24 h urine collections
IP
T
(80%) and reflected excretion of VMA and HVA per day. Most specimens (90%) from
CR
pediatric patients were random urine collections. Analyte concentrations in random collections are normalized using the creatinine concentration of the sample and reported
US
as mass (milligrams) of VMA or HVA relative to mass (grams) of creatinine. Reference
AN
intervals determined for daily excretion and random urine collections from subsets of the data are shown in Table 5. All partitions by age group and creatinine concentration were
M
justified by Harris-Boyd partitioning criteria except separation of HVA values for the 6 to
ED
11 y old groups based on creatinine concentration.
PT
To determine the extent that a low creatinine concentration might elevate a patient result, we calculated reference limits for the urine acids in pediatric specimens separated
CE
into groups of low [<25 mg/dl (2.2 mmol/l)] and adequate [≥25 mg/dl (2.2 mmol/l)]
AC
measured creatinine (Table 5). Specimens collected from individuals aged ≤2 y were more likely to contain reduced amounts of creatinine. Urine acid ratios determined for these specimens were only slightly higher than those calculated for more concentrated specimens from the same age group (Table 5). Fewer dilute specimens were received for testing from older children. Reference limits assigned by urine creatinine concentration for children aged 3-5 and 6-11 y were slightly higher than those determined for specimens without consideration of specimen creatinine. 17
ACCEPTED MANUSCRIPT 4. Discussion As far as we are aware, all other authors [6-15] have reported ionizing VMA and HVA in negative ESI mode as would be expected for small polar acids. During method development, we tested both positive and negative ion modes and found that positive
IP
T
mode produced slightly higher signal for both compounds and that matrix effects were
CR
similar to those experienced in negative mode.
Magera et al. [6,7] used solid phase extraction for sample cleanup prior to analysis
US
in separate VMA and HVA methods and Konieczna et al. [13] utilized dispersive liquid-
AN
liquid microextraction (DLLME) for a neurotransmitter panel including VMA and HVA, but many others [9-12] have opted for minimal specimen preparation (dilute-and-shoot
M
approach). While sample cleanup certainly contributes to method robustness, it also
ED
increases cost, method complexity, and turnaround time, and in some cases (such as most
PT
forms of liquid-liquid extraction including DLLME) it may complicate or preclude automation. When using a dilute-and-shoot approach, great care must be taken to ensure
CE
that matrix effects are minimized by designing elution gradients to exclude analytes from
AC
suppression zones and/or by the use of well-matched SIL internal standards. Some of the reported methods used no internal standard [8,10,12], while others utilized only 1 or 2 non-isotopically labeled IS for an entire panel of analytes [9,13]. Neither of these options provides sufficient robustness for a quantitative high-throughput clinical assay. In our method, we used the post-column infusion experiment to design an LC gradient separating the analytes from major suppression zones and, as internal standards, we chose SIL analogs that exactly co-elute with their respective analytes, thus ensuring accuracy 18
ACCEPTED MANUSCRIPT and robustness without the added cost of sample preparation consumables. To further improve the assay performance, we designed an alternate LC gradient to resolve matrix interferences observed in 3% of the specimens. We opted to re-inject the affected specimens using a short VMA-only method utilizing isocratic elution followed by a
IP
T
cleanout out step to separate VMA from early eluting matric interferences, rather than
CR
designing a longer chromatographic method for both of the analytes, which would reduce throughput.
US
By moving from an HPLC-ECD to an LC-MS/MS platform we were able to reduce
AN
the injection-to-injection time from 21 to 4 min. Some authors have achieved shorter run times, but only in methods for a single analyte [6,7,25]. Since VMA and HVA have
M
similar characteristics, and there is value in measuring both analytes in the same
ED
specimen [16], it is logical to combine the 2 neuroblastoma biomarkers in one method.
PT
Reports of methods for analysis of VMA and HVA in panels with other analytes not related to neuroblastoma reported longer run times (6-20 min).
CE
While Lionetto et al. [9] and Fang et al. [10] listed multiple fragment ions for the
AC
analytes, only Sadilkova et al. [14] in their method for VMA and HVA in serum used a second MRM transition. It is a standard practice in our laboratory to use 2 MRM transitions, qualitative and quantitative, for both the analyte and internal standard to increase assay specificity and thus ensure accurate identification and quantification of the analytes.
19
ACCEPTED MANUSCRIPT Lionetto et al. [9] investigated ion suppression in the presence of 6 drugs, while Shen et al. [25] tested 4 structurally related endogenous compounds. We investigated 62 substances, including endogenous compounds, vitamins, and frequently used over-thecounter drugs and pharmaceuticals, as potential interferences to ensure a robust
IP
T
performance of our dilute-and-shoot method. Only phenylalanine and isoetharine
CR
produced VMA interference >±10% at test concentrations much higher than would be expected in urine specimens and should not therefore interfere. None of the tested
US
compounds interfered with HVA.
AN
Reference intervals determined from archived patient results analyzed using the new LC-MS/MS method corresponded well with reference intervals established for the
M
HPLC-ECD assay and were similar to published values from HPLC-ECD and GC-MS
ED
methods [26-28]. Reference limits determined using analyte concentrations measured in
PT
dilute random urine specimens and expressed as values normalized to creatinine concentration were slightly higher (<10%) than reference intervals calculated for all
AC
CE
results for children <6 y.
5. Conclusions
We have developed and validated a simple, inexpensive, high-throughput LCMS/MS method for the determination of urinary VMA and HVA. This method utilizes a dilute-and-shoot sample preparation in a 96-well format performed using a liquid handler. Method robustness is ensured by the use of co-eluting SIL analogs as internal standards and by chromatography designed to elute analytes outside of major suppression zones. 20
ACCEPTED MANUSCRIPT For increased specificity and accuracy, 2 transitions, qualifier and quantifier, were monitored for each analyte and its SIL internal standard. VMA interferences precluding quantitation in 3.0% of specimens were reduced significantly (to 0.1% of specimens)
CR
IP
study may be of value to health practitioners ordering the tests.
T
using a modified LC gradient. The information obtained from the analytical specificity
Acknowledgements
US
We thank Seyed Sadjadi, Sung Baek, David Davis, Steven Achelis, and the staff of
AN
the Analytic Biochemistry Laboratory for their assistance during this project. This study
AC
CE
PT
ED
M
was supported by the ARUP Institute for Clinical and Experimental Pathology®.
21
ACCEPTED MANUSCRIPT Figure captions Fig. 1. LC-MS/MS chromatograms of (A) 2 mg/l calibration standard, (B) healthy patient urine, and (C) abnormal patient urine: vanillylmandelic acid (VMA), vanillylmandelic acid-d3 (VMA-d3), homovanillic acid (HVA), homovanillic acid-13C6 18O (HVA-13C6
T
O).
CR
IP
18
Fig. 2. Signal suppression scenarios: (A) very little suppression observed in dilute urines,
US
(B) suppression zones that do not interfere with analyte quantitation, (C) and (D) rare
AN
suppression zones at analyte retention times. VMA-d3 and HVA-13C6 18O were infused
M
post-column.
ED
Fig. 3. Mitigation of VMA signal suppression by the alternate VMA method: (A)
PT
specimen 1 analyzed by the primary VMA-HVA method, (B) specimen 1 analyzed using the alternate VMA method, (C) specimen 2 analyzed by the primary VMA-HVA method,
AC
column.
CE
(D) specimen 2 analyzed using the alternate VMA method. VMA-d3 was infused post-
22
ACCEPTED MANUSCRIPT References [1] G. Eisenhofer, R.J. Whitley, T.G. Rosano, Catecholamines and Serotonin, in: C.A. Burtis, E.R. Ashwood, D.E. Bruns (Eds.) Tietz Textbook of Clinical Chemistry and Molecular Diagnostics, Elsevier Saunders, St. Louis, 2012, pp. 851-894.
IP
T
[2] I. Ora, A. Eggert, Progress in treatment and risk stratification of neuroblastoma:
CR
Impact on future clinical and basic research, Semin. Cancer Biol. 21 (2011) 217-228. [3] M.S. Irwin, J.R. Park, Neuroblastoma: Paradigm for precision medicine, Pediatr. Clin.
US
N. Am. 62 (2015) 225-256.
AN
[4] J.M. Maris, M.D. Hogarty, R. Bagatell, S.I. Cohn, Neuroblastoma, Lancet 369 (2007) 2106-2120.
M
[5] W.H.A. de Jong, E.G.E. de Vries, I.P. Kema, Current status and future developments
ED
of LC-MS/MS in clinical chemistry for quantification of biogenic amines, Clin. Biochem.
PT
44 (2011) 95–103.
[6] M.J. Magera, A.L. Thompson, D. Matern, P. Rinaldo, Liquid chromatography–
CE
tandem mass spectrometry method for the determination of vanillylmandelic acid in
AC
urine, Clin. Chem. 49 (2003) 825-826. [7] M.J. Magera, A.L. Stoor, J.K. Helgeson, D. Matern, P. Rinaldo, Determination of homovanillic acid in urine by stable isotope dilution and electrospray tandem mass spectrometry, Clin. Chim. Acta 306 (2001) 35-41. [8] P. Manini, R. Andreoli, S. Cavazzini, E. Bergamaschi, A. Mutti, W.M.A. Niessen, Liquid chromatography–electrospray tandem mass spectrometry of acidic monoamine metabolites, J. Chromatogr. B 744 (2000) 423-431. 23
ACCEPTED MANUSCRIPT [9] L. Lionetto, A.M. Lostia, A. Stigliano, P. Cardelli, M. Simmaco, HPLC–mass spectrometry method for quantitative detection of neuroendocrine tumor markers: Vanillylmandelic acid, homovanillic acid and 5-hydroxyindoleacetic acid, Clin. Chim. Acta 398 (2008) 53–56.
IP
T
[10] L. Fang, Y. Lv, X. Sheng, S. Yao, Sensitive, rapid and easy analysis of three
CR
catecholamine metabolites in human urine and serum by liquid chromatography tandem mass spectrometry, J. Chromatogr. Sci. 50 (2012) 450–456.
US
[11] B. Crow, M. Bishop, E. Paliakov, D. Norton, J. George, J.A. Bralley, Analysis of
AN
urinary aromatic acids by liquid chromatography tandem mass spectrometry, Biomed. Chromatogr. 22 (2008) 1346–1353.
M
[12] W.-Y. Hsu, C.-M. Chen, F.-J. Tsai, C.-C. Lai, Simultaneous detection of diagnostic
ED
biomarkers of alkaptonuria, ornithine carbamoyltransferase deficiency, and
PT
neuroblastoma disease by high-performance liquid chromatography/tandem mass spectrometry, Clin. Chim. Acta 420 (2013) 140–145.
CE
[13] L. Konieczna, A. Roszkowska, M. Niedzwiecki, T. Baczek, Hydrophilic interaction
AC
chromatography combined with dispersive liquid–liquid microextraction as a preconcentration tool for the simultaneous determination of the panel of underivatized neurotransmitters in human urine samples, J. Chromatogr. A 1431 (2016) 111–121. [14] K. Sadilkova, K. Dugaw, D. Benjamin, R.M. Jack, Analysis of vanillylmandelic acid and homovanillic acid by UPLC–MS/MS in serum for diagnostic testing for neuroblastoma, Clin. Chim. Acta 424 (2013) 253–257.
24
ACCEPTED MANUSCRIPT [15] N. Tohmola, O. Itkonen, U. Turpeinen, S. Joenväärä, R. Renkonen, E. Hämäläinen, Preanalytical validation and reference values for a mass spectrometric assay of serum vanillylmandelic acid for screening of catecholamine secreting neuroendocrine tumors, Clin. Chim. Acta 446 (2015) 206–212.
IP
T
[16] G. Cangemi, G. Reggiardo, S. Barco, et al., Prognostic value of ferritin, neuron-
CR
specific enolase, lactate dehydrogenase, and urinary and plasmatic catecholamine metabolites in children with neuroblastoma, Onco. Targets and Therapy 5 (2012) 417-
US
423.
AN
[17] C.N. Svendsen, Multi-electrode assay detectors in high-performance liquid chromatography: A new dimension in electrochemical analysis, Analyst 118 (1993) 123-
M
129.
ED
[18] T.M. Annesley, Ion suppression in mass spectrometry, Clin. Chem. 49 (2003) 1041-
PT
1044.
[19] CLSI. Interference testing in clinical chemistry; Approved guideline-Second edition.
CE
CLSI document EP7-A2. Wayne, PA: Clinical and Laboratory Standards Institute; 2005.
AC
[20] R.G. Hoffmann, Statistics in the practice of medicine, J. Amer. Med. Assoc. 185 (1963) 864-873.
[21] A. Katayev, C. Balciza, D.W. Seccombe, Establishing reference intervals for clinical laboratory test results: Is there a better way?, Am. J. Clin. Pathol. 133 (2010) 180-186. [22] E.K. Harris, J.C. Boyd, On dividing reference data into subgroups to produce separate reference ranges, Clin. Chem. 36 (1990) 265-270.
25
ACCEPTED MANUSCRIPT [23] R.C. Baselt, Carbidopa, Disposition of Toxic Drugs and Chemicals in Man, 9th ed., Biomedical Publications, Seal Beach, California, 2011, pp. 252-253. [24] B. Spilker, B.S. Watson, J.W. Woods, Drug interference with measurement of metanephrines in urine, Ann. Clin. Lab. Sci. 13 (1983) 16-19.
IP
T
[25] Y. Shen, J. Lu, Q. Tang, Q. Guan, Z. Sun, H. Li, L. Cheng, Rapid, easy analysis of
CR
urinary vanillylmandelic acid for diagnostic testing of pheochromocytoma by liquid chromatography tandem mass spectrometry, J. Chromatogr. B 1002 (2015) 92–97.
US
[26] E. Pussard, M. Neveux, N. Guigueno, Reference intervals for urinary
AN
catecholamines and metabolites from birth to adulthood, Clin. Biochem. 42 (2009) 536539.
M
[27] S. Barco, I. Gennai, G. Reggiardo, B. Galleni, L. Barbagallo, A. Maffia, E. Viscardi,
ED
F. De Leonardis, V. Cecinati, S. Sorrentino, A. Garaventa, M. Conte, G. Cangemi,
PT
Urinary homovanillic and vanillylmandelic acid in the diagnosis of neuroblastoma:
848-852.
CE
Report from the Italian Cooperative Group for Neuroblastoma, Clin. Biochem. 47 (2014)
AC
[28] M.T.C. Tran, J. Baglin, T.T. T. Tran, K.T.Hoang, L.T. Phung, A. Read, R.F. Greaves, Development of a new biochemical test to diagnose and monitor neuroblastoma in Vietman: Homovanllic and vanillylmandelic acid by gas chromatography-mass spectrometry, Clin. Biochem. 47 (2014) 206-215.
26
AC
CE
PT
ED
M
AN
US
CR
IP
T
ACCEPTED MANUSCRIPT
Fig. 1 27
AC
CE
PT
ED
M
AN
US
CR
IP
T
ACCEPTED MANUSCRIPT
Fig. 2 28
M
AN
US
CR
IP
T
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
Fig. 3
29
ACCEPTED MANUSCRIPT Highlights
T IP CR US AN M ED PT CE
Validation of inexpensive, dilute-and-shoot LC-MS/MS method for urine VMA and HVA Comprehensive investigation of signal suppression Comprehensive investigation of interference Unique chromatography design Age and gender specific urine reference intervals for random and timed collections
AC
30
Zlatuse D. Clark, Jeaneah M. Cutler, Igor Y. Pavlov, Frederick G. Strathmann, Elizabeth L. Frank PII: DOI: Reference:
S0009-8981(17)30078-5 doi: 10.1016/j.cca.2017.03.004 CCA 14674
To appear in:
Clinica Chimica Acta
Received date: Revised date: Accepted date:
13 August 2016 1 March 2017 1 March 2017
Please cite this article as: Zlatuse D. Clark, Jeaneah M. Cutler, Igor Y. Pavlov, Frederick G. Strathmann, Elizabeth L. Frank , Simple dilute-and-shoot method for urinary vanillylmandelic acid and homovanillic acid by liquid chromatography tandem mass spectrometry. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Cca(2016), doi: 10.1016/j.cca.2017.03.004
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Simple dilute-and-shoot method for urinary vanillylmandelic acid and homovanillic acid by liquid chromatography tandem mass spectrometry
Zlatuse D. Clarka, Jeaneah M. Cutlerb, Igor Y. Pavlova, Frederick G. Strathmannc,
IP
b
c
ARUP Laboratories, Salt Lake City, UT 84108
CR
ARUP Institute for Clinical and Experimental Pathology, Salt Lake City, UT 84108
Department of Pathology, University of Utah Health Sciences Center, Salt Lake City,
US
a
T
Elizabeth L. Frankc*
AN
UT 84112
Corresponding author: Elizabeth L. Frank, c/o ARUP Laboratories, Inc., 500 Chipeta
M
Way, Mail Code 115, Salt Lake City, UT 84108, phone +1 (801) 583-2787, ext. 1-2087,
ED
email: [email protected].
PT
Previous presentation of the manuscript: Parts of these data were presented at the
CE
Mass Spectrometry: Applications to the Clinical Lab (MSACL) conference in San Diego, CA, March 28-April 1, 2015.
AC
List of abbreviations: CLSI, Clinical and Laboratory Standards Institute; DLLME, dispersive liquid-liquid microextraction; ECD, electrochemical detection; ESI, electrospray ionization; HVA, homovanillic acid; LLOQ, lower limit of quantification; MRM, multiple reaction monitoring; QIR, qualitative ion ratio; SIL, stable isotope label; VMA, vanillylmandelic acid
1
ACCEPTED MANUSCRIPT Abstract Background: Neuroblastomas are pediatric tumors characterized by overproduction of catecholamines. The catecholamine metabolites, vanillylmandelic acid (VMA) and homovanillic acid (HVA), are used in clinical evaluation of neuroblastoma. Tandem mass spectrometry
T
(LC-MS/MS) is an effective analytical method for measurement of VMA and HVA in urine.
IP
Methods: Dilute-and-shoot sample preparation was performed in a 96-well format using a liquid
CR
handler. Chromatographic separation was achieved using a reverse phase column; detection was accomplished by triple quadrupole mass spectrometry with electrospray ionization in positive
US
mode. Data were acquired by multiple reaction monitoring. Two transitions, quantifier and
AN
qualifier, were monitored for each analyte and its stable isotope-labeled internal standard. Analytical specificity studies were performed.
M
Results: Injection-to-injection time was 4 min. The method was validated for linearity, limit of
ED
quantification, imprecision, accuracy, and interference. Linearity was 0.5 – 100 mg/l for both analytes. Within-run, between-day, and total imprecision were 1.0 – 4.1% for VMA and 0.8 –
PT
3.8% for HVA. The method correlated well with our established HPLC method. Interferences
CE
precluding quantitation of VMA in 3% of specimens were reduced significantly (to 0.1% of specimens) using a modified LC gradient to reanalyze affected samples.
AC
Conclusions: A simple, robust, economical, fast LC-MS/MS method was developed and validated for measurement of urinary VMA and HVA. Keywords: Liquid chromatography; mass spectrometry; vanillylmandelic acid; homovanillic acid; urine
2
ACCEPTED MANUSCRIPT 1. Introduction Laboratory measurement of vanillylmandelic acid (VMA), the end product of epinephrine and norepinephrine metabolism, and homovanillic acid (HVA), a terminal metabolite of dopamine, is used to support clinical diagnosis and management of
IP
T
catecholamine-secreting neurochromaffin tumors such as neuroblastomas [1,2].
CR
Neuroblastoma typically occurs in children and accounts for 7-10% of childhood cancer; approximately 90% of these tumors produce catecholamines [2]. The catecholamine
US
metabolites, VMA and HVA, are used in combination with other prognostic factors to
AN
improve tumor detection and assess treatment response [3,4]. Biochemical tests for VMA and HVA typically are performed using HPLC or GC-
M
MS platforms. LC-MS/MS technology offers several analytical performance advantages
ED
for these small molecules [5]. Use of LC-MS/MS can minimize sample preparation, decrease analysis time, and improve selectivity. Tandem mass spectrometry methods
PT
have been developed to measure urinary VMA [6] and HVA [7] as single analytes and in
CE
panels with other biomarkers related to neuroendocrine tumors [8-10] and other disorders
AC
[11-13]. Assays using plasma and/or serum [10,14,15] circumvent challenges with pediatric 24-h urine collections and variation inherent in random urine specimens. Combined measurement of the analytes allows calculation of the VMA/HVA ratio, which can provide additional prognostic value [16]. In our laboratory, an HPLC method with electrochemical detection (ECD) had been used to analyze urinary VMA and HVA for nearly 30 y. The analytes were measured concurrently using a coulometric electrode array system. Serial coulometric electrodes, 3
ACCEPTED MANUSCRIPT set at increasing potentials, provided electrochemical resolution to augment chromatographic separation of VMA, HVA, and other analytes [17]. Sample analysis was accomplished in 21 minutes. In recent years, increased test volume and frequency of interfering substances in specimens received for analysis, as well as aging of the
IP
T
instruments, limited usefulness of the assay. Modifications to accommodate more
CR
numerous requests for VMA measurement, which could be analyzed separately in approximately five minutes, were implemented, but did not resolve the issues completely.
AN
US
We identified mass spectrometry as a potential replacement method.
2.1. Chemicals and reagents
M
2. Experimental procedures
ED
HPLC-grade methanol was from Burdick & Jackson; formic acid (98%) was from
PT
EM Science; vanillylmandelic acid (VMA, DL-4-hydroxy-3-methoxymandelic acid) and homovanillic acid (HVA, 4-hydroxy-3-methoxyphenyl acetic acid) were from Sigma-
CE
Aldr; (±)-4-hydroxy-3-methoxy-d3-mandelic acid (VMA-d3) was from C/D/N Isotopes;
AC
and homovanillic acid [ring-13C6, 4-hydroxy-18O] (HVA-13C6 18O) was from Cambridge Isotope Laboratories, Inc. Lyphocheck® Quantitative Controls Level I and Level II were purchased from Bio-Rad Clinical Division (Hercules, CA). Clinical Laboratory Reagent (CLR) water was used throughout the study. Synthetic urine was prepared from urea, sodium chloride, creatinine hydrochloride, dipotassium hydrogen phosphate, and sodium dihydrogen phosphate (Sigma-Aldrich). Each lot of synthetic urine was validated by testing; no VMA or HVA was detected in the reagent. 4
ACCEPTED MANUSCRIPT 2.2. Calibration standards and quality control solutions Combined VMA and HVA stock solution was prepared in water at 1000 mg/l. Calibration standards were prepared at 0.5, 2, 10, 50 and 100 mg/l concentrations by dilution of the 1000 mg/l combined stock solution with synthetic urine. VMA-d3 and
IP
T
HVA-13C6 18O individual internal standard stock solutions (1000 mg/l) and combined
CR
working internal standard solution (5 mg/l) were prepared in water. For MS optimization and signal suppression experiments, the stock solutions were diluted with 0.05% (v/v)
US
formic acid in water to 50 and 10 mg/l, respectively.
AN
For the primary VMA-HVA method validation, control materials were prepared at 4 concentrations (Table 2): Levels I and II were Bio-Rad Lyphocheck® Quantitative
M
Controls, reconstituted in 10 ml of CLR water. Levels III and IV were prepared from
ED
patient specimen pools. For the alternate VMA method validation, control materials were
PT
prepared at three concentrations: low (0.8 mg/l) and high (80 mg/l) by dilution of the combined VMA and HVA stock solutions in water; a mid-range control was made by
CE
reconstitution of Lyphocheck® Quantitative Urine Control Level II with water. Aliquots
AC
of stock, calibration, and control solutions were stored frozen at -70 °C. Working internal standard solution was prepared fresh daily by dilution of the stock solutions.
2.3. Sample collection and preparation Urine specimens were selected from samples submitted for clinical testing. Samples were de-identified and managed according to guidelines approved by the Institutional Review Board of the University of Utah.Using a Tecan liquid handler, 50 l of working 5
ACCEPTED MANUSCRIPT internal standard solution and 400 µl of 0.05% (v/v) formic acid in water were added to 50 µl of calibration standards, quality controls, and urine specimens in a 96-well plate and mixed. The plate was centrifuged at 3500×g under refrigeration and 250 l aliquots from each well were transferred to a new 96-well plate from which injections onto the
2.4. LC-MS/MS instrumentation and conditions
CR
IP
T
analytical column were performed.
US
Analysis was performed on a Waters Xevo™ TQ MS Acquity UPLC® instrument,
AN
which consisted of a mass spectrometer with an electrospray ionization (ESI) probe in positive mode and an ACQUITY UPLC® system including a binary solvent manager,
M
sample manager, sample organizer, and column manager. Data were acquired using
ED
multiple reaction monitoring (MRM). Instrument control and data analysis and
PT
quantitation were performed using Waters MassLynx™ software (v4.1). Chromatographic separation was achieved using a Kinetex XB-C18 column (2.1 ×
CE
50 mm, 100 Å, 1.7 µm particles; Phenomenex) with an in-line filter assembly (Waters,
AC
Milford, MA) and 0.05% (v/v) formic acid in water as mobile phase A and 0.05% (v/v) formic acid in methanol as mobile phase B. Gradient elution with 5-30% B in 0 - 2.3 min, followed by a cleanout step to 95% B (2.3 - 2.6 min), was employed in the primary VMA-HVA method. Isocratic elution at 0% B for 1.7 min, followed by a step gradient to 95% B (1.7 - 2.2 min) was used in the alternate VMA method. For both methods, the mobile phase flow rate was 400 µL/min, the column temperature was 30 C, the sample
6
ACCEPTED MANUSCRIPT manager temperature was 10 C, the injection volume was 5 µl, and the injection-toinjection time was 4 min. Electrospray ionization was performed in positive mode. Purified nitrogen was used for the desolvation (1000 l/h) and cone (25 l/h) gases; purified argon was the collision gas
IP
T
(0.20 ml/min). The source temperature was 150 C and the desolvation temperature was
CR
600 C. The capillary voltage was set to 3000 V and the extractor to 3 V. MRM qualifier and quantifier transitions, cone voltages, and collision energies for both analytes and their
US
internal standards in the primary VMA-HVA method are given in Table 1. Dwell time for
monitored with 70 ms dwell time each.
AN
all transitions was 32 ms. For the alternate VMA method, only VMA transitions were
M
A qualitative ion ratio (QIR) was used to confirm identity of the analytes. The QIRs
ED
were calculated as ratios of the peak area for the quantifier transition to the peak area for
PT
the qualifier transition for both the analytes and the internal standards. Patient specimen QIR acceptability range was established as 30% of the mean value of the QIRs obtained
AC
CE
from the calibration standards.
2.5. Method validation 2.5.1. Linearity Linearity was evaluated by analyzing samples prepared by dilution of a high concentration patient specimen pool with a low concentration patient specimen pool at 8 target concentrations ranging from 0.25 to 91.01 mg/l for VMA and from 0.37 to 90.65
7
ACCEPTED MANUSCRIPT mg/l for HVA. Samples were aliquoted in tubes in an amount sufficient for one day of testing and stored frozen at –70 ºC until analysis. A set of 4 replicates from each aliquot was analyzed on 4 separate days, totaling 16 replicates for each concentration. The
IP
T
acceptance criterion for linearity was 10% of expected concentration.
CR
2.5.2. Limit of quantification
Lower limit of quantification (LLOQ) was evaluated by analyzing samples prepared
US
by dilution of a de-identified patient specimen with synthetic urine at concentrations
AN
ranging from 0.18 to 2.91 mg/l for VMA and from 0.18 to 2.82 mg/l for HVA. Samples were aliquoted in tubes in an amount sufficient for one day of testing and stored frozen at
M
–70 ºC until analysis. A set of 4 replicates from each aliquot was analyzed on 4 separate
ED
days, totaling 16 replicates for each concentration. The criteria for establishing the LLOQ
CE
2.5.3. Imprecision
PT
concentration were CV ≤10% and accuracy within 15% of the expected value.
AC
Inter-assay imprecision was evaluated by analyzing control materials prepared at 4 concentrations over 4 days in 4 replicates per run, 1 run per day.
2.5.4. Method comparison The newly developed LC-MS/MS method was compared to the HPLC-ECD method in use. Specimens were stored at ≤–20 °C prior to analysis.
8
ACCEPTED MANUSCRIPT 2.5.5. Carryover Carryover was evaluated as described in EP Evaluator software. A set of high (H) and low (L) samples was assayed in the following consecutive order: 3 L, 2 H, 1 L, 2 H, 4 L, 2 H, 1 L, 2 H, 1 L, 2 H, 1 L. A patient specimen pool with VMA at 3.1 mg/l and
IP
T
HVA at 2.8 mg/l was used for a low concentration sample; a patient specimen pool
CR
spiked to 400 mg/l was used as the high concentration sample.
US
2.5.6. Evaluation of ion suppression
AN
Ion suppression by matrix components was evaluated as the presence of negative peaks at the retention times of VMA and HVA. Urine samples (n = 83) prepared without
M
internal standard were injected onto the LC column and individual 10 mg/l internal
ED
standard solutions were co-infused at 5 µl/min post-column as described by Annesley
PT
[18].
CE
2.5.7. Analytical specificity
AC
A number of endogenous and exogenous compounds were tested for interference in the assay. Individual aliquots of pooled urine with VMA and HVA concentrations of 2.1 mg/l were spiked with solutions of potential interferents at 10% of total volume. Recommendations from CLSI EP7-A2, Interference Testing in Clinical Chemistry, were used to guide prepared interferent concentrations [19]. Appropriate baseline samples were generated by spiking the urine pool with solvents (water, 0.1 mol/l HCl, 0.5 mol/l HCl, 0.1 mol/l NaOH, or methanol) to match the spiking solutions solvents and spiking 9
ACCEPTED MANUSCRIPT volumes. The interference samples were analyzed in 2 separate experiments. In the first, quantitative experiment, VMA and HVA concentrations were measured in a sample spiked with a potential interferent. The result was compared to the respective analyte concentration in a baseline sample. Deviation was calculated as the difference between
IP
T
the measured concentration in each interference sample and the respective baseline
CR
sample mean concentration divided by the baseline sample mean concentration. A substance was considered to interfere if the calculated deviation exceeded ±10%. In the
US
second, qualitative experiment, the interference samples were analyzed in MS scan mode
AN
to obtain retention times and m/z values for the potential interferents.
M
2.5.8. Reference intervals
ED
Reference intervals were evaluated according to the method of Hoffmann [20,21].
PT
Archived data for VMA and HVA measured in urine specimens using the new LCMS/MS method were assessed. Outlier values exceeding the third quartile plus three
CE
times the interquartile range were removed. Simplified computer calculations were
AC
performed with the assumption that maximum error (deviation of the cumulative frequencies curve from the linear regression) did not exceed 5% of the median value. Reference limits were calculated at 2.5% and 97.5% of the linear portion of the data for each defined group. Harris and Boyd [22] criteria were applied for partitioning by age, gender, and creatinine concentration (where applicable).
10
ACCEPTED MANUSCRIPT 2.5.9. Data analysis Statistical evaluation of the data was performed using Microsoft Excel and EP Evaluator software, release 10 (David G. Rhoads Associates, Inc.).
IP
T
3. Results
CR
3.1. Optimization of method parameters 3.1.1. MS parameter optimization
US
MS parameters were determined using post-column co-infusion of compound
AN
solutions into mobile phase solution at 10 µl/min. Both negative and positive ionization modes were explored. Negative mode did not exhibit reduced matrix effects. Positive
M
mode gave slightly higher signal for all compounds and was therefore used in the new
ED
method. While HVA and HVA-13C6 18O ionized to the expected protonated molecular ions, VMA and VMA-d3 readily lost water in the ESI source, yielding the more stable
PT
[M-H2O+H]+ ions, which were selected as the precursor ions. Due to the low molecular
CE
weight of VMA and HVA, the selection of useful MRM transitions was not large. During
AC
method development, all viable MRM transitions were carefully monitored and those with the least occurrence of interferences were included in the MS method (Table 1).
3.1.2. LC parameter optimization Several LC columns were tested for VMA and HVA separation: Waters HSS T3 (2.1×50 mm, 1.8 µm), Waters XBridge C18 (2.1×75 mm, 2.5 µm), Phenomenex Gemini C18 (3×100 mm, 3 µm) and Phenomenex Kinetex XB-C18 (3×50 mm, 2.6 µm and 11
ACCEPTED MANUSCRIPT 2.1×50, 1.7 µm). The Kinetex XB-C18 (3×50 mm, 2.6 µm and 2.1×50 mm, 1.7 µm) columns, with 0.05% formic acid in water as mobile phase A and 0.05% formic acid in methanol as mobile phase B, provided the best combination of retention, separation, and short run times for the compounds of interest. The LC gradient was optimized using post-
IP
T
column infusion of internal standards while analyzing urine specimens to minimize the
CR
impact of sample matrix on analyte quantitation. The injection-to-injection time was shortened from 21 min in the HPLC-ECD method to 4 min. Example chromatograms of
US
VMA and HVA standard solutions and of healthy and abnormal patient specimens are
ED
3.2.1. Linearity
M
3.2. Method validation results
AN
shown in Fig. 1.
PT
Observed concentrations for VMA and HVA were plotted against expected concentrations to give Eqs. (1) and (2), respectively. (1)
y = 0.996x + 0.004; observed error 0.4%
(2)
AC
HVA
y = 0.991x + 0.011; observed error 1.7%
CE
VMA
The assay was linear from 0.5 to 100 mg/l.
3.2.2. Limit of quantification Accuracy (within 15% of the expected value) and imprecision (CV ≤10%) criteria were exceeded at 0.18 mg/l for both VMA and HVA. The LLOQ was set at 0.5 mg/l, the concentration of the lowest calibration standard, for both analytes. 12
ACCEPTED MANUSCRIPT
3.2.3. Imprecision Within-run, between-day, and total imprecision for 4 quality control concentrations ranged from 1.0 – 4.1% for VMA and from 0.8 – 3.8% for HVA. Specific data are listed
CR
IP
T
in Table 2.
3.2.4. Method comparison
US
Specimens with analyte concentrations from 0.5 mg/l to 80 mg/l (the upper limit of
AN
linearity for the HPLC-ECD method) were assayed in the method comparison experiment. Acceptable agreement was observed between the HPLC-ECD method and
M
the newly developed LC-MS/MS method for both VMA and HVA. Deming regression
ED
equations for VMA, Eq. (3), and HVA, Eq. (4), standard errors, correlation coefficients, %bias, and number of specimens were:
y = 1.014x – 0.069; Sy/x = 0.924; R = 0.9904; %bias = –0.356%; n = 374
(3)
HVA
y = 1.153x – 0.983; Sy/x = 2.294; R = 0.9909; %bias = 5.661%; n = 97
(4)
CE
PT
VMA
AC
3.2.5. Carryover
At 400 mg/l, the calculated carryover values for VMA and HVA (0.07 and 0.01 mg/l, respectively) were less than the error limits (0.12 and 0.07 mg/l, respectively), defined as three times the SD of 2 consecutive low results. Carryover was not observed for either analyte.
13
ACCEPTED MANUSCRIPT 3.2.6. Evaluation of ion suppression As the developed method utilizes a dilute-and-shoot sample preparation, a thorough evaluation of matrix effects using a large number (83) of specimens was essential. While the number and intensity of suppression zones varied significantly from sample to
IP
T
sample, VMA and HVA peaks were not adversely affected in the majority of the
CR
specimens (Fig. 2). Distinct suppression zones observed at the retention times of the analytes and affecting analyte signal intensity were found in only 2 specimens for VMA
US
(Fig. 2C and D) and one specimen for HVA (Fig. 2D). The internal standards for both
AN
VMA and HVA elute very closely with their non-labeled analogs and should, therefore, compensate for signal suppression unless it is severe. Of the 83 samples evaluated, severe
PT
ED
less affected by matrix effects.
M
signal suppression was observed only for VMA in one sample (Fig. 2C). HVA was much
3.2.7. Analytical specificity
CE
Compounds, actual and recommended test concentrations, monoisotopic masses,
AC
ions observed in MS scans ([M+1] ions or most intense alternate ions), and retention times, as well as percent deviation of analyte concentrations in test samples versus respective baseline samples are shown in Table 3 for endogenous substances and vitamins and in Table 4 for exogenous substances. Three of the compounds tested eluted very close to VMA (tR = 0.85 min) and could potentially interfere with VMA quantitation at high concentrations. Carbidopa produced a peak (tR = 0.72 min) in the quantitative VMA transition and deviation of 5.7% at the 14
ACCEPTED MANUSCRIPT actual test concentration of 140 mg/l (about 10-fold higher than expected in urine after a 130 mg/l dose [23]). Phenylalanine (tR = 0.89 min) and isoetharine (tR = 0.82 min) induced deviations of -10.9% (at 130 mg/l, recommended test concentration 17 mg/l) and 14.2% (at 100 mg/l, estimated concentration in urine 0.4 mg/l [24]), respectively, but did
IP
T
not appear as peaks in any VMA transitions.
CR
Differential suppression of VMA and its slightly earlier eluting internal standard by the interferents is a most likely explanation for the >±10% VMA concentration deviation
US
observed with phenylalanine and isoetharine. Phenylalanine eluting 0.05 min after VMA
AN
would suppress the analyte more than the internal standard, thus producing a negative deviation, while isoetharine eluting 0.02 min before VMA would suppress the internal
M
standard more than the analyte, thus producing a positive deviation.
ED
None of the 62 tested compounds interfered with HVA quantitation. Caffeine (tR =
PT
2.19 min) and cocaine (tR = 2.12 min) eluted close to HVA (tR = 2.29 min) and could potentially interfere with HVA quantitation at very high concentrations. However, neither
CE
of these compounds appeared as peaks in any of the HVA MRM transitions or produced
AC
HVA deviation from baseline beyond ±10% at their respective test concentrations. In actual use, approximately 3% of VMA results could not be reported due to interference; only one specimen of more than 3000 tested showed an interference that prevented quantification of HVA. The majority of the VMA interferences produced internal standard peak areas below the lower acceptance limit (50% of the batch internal standard peak area median) due to signal suppression by matrix components at the retention time of VMA. 15
ACCEPTED MANUSCRIPT In an effort to improve assay performance, an alternate LC gradient was designed to resolve VMA interferences. Post-column infusion of a VMA-d3 solution was used as a tool to determine the position of suppression zones in six specimens with interference and to manipulate the LC gradient to separate VMA from the suppression zones. Isocratic
IP
T
elution at 0% B for 1.7 min, followed by a step gradient to 95% B (1.7 - 2.2 min) during
CR
which the column effluent (including HVA) is directed to waste was determined to be optimal. No additional method parameters were altered.
US
Validation of the alternate VMA method was performed to assess linearity,
AN
sensitivity, and imprecision (data not shown). All results met assay performance criteria. The alternate VMA method was compared with the primary VMA-HVA method using
M
196 specimens with concentrations spanning the analytical measurement range. Good
ED
agreement between the 2 methods was observed. The Deming regression equation, standard error, and correlation coefficient were: y = 0.978x + 0.092; Sy/x = 0.344; R =
PT
0.9996; Bias = –0.742%.
CE
Reanalysis of 6 specimens that showed interference produced VMA peak areas
AC
approximately 6- to 20-fold higher using the alternate method, indicating that the alternate LC gradient mitigated signal suppression experienced by VMA in the primary method. Use of the alternate VMA method reduced the incidence of non-reportable results from 3.0% to 0.1%. Results for endogenous and exogenous substances evaluated for assay interference using the alternate method parameters are shown in Tables 3 and 4.
16
ACCEPTED MANUSCRIPT 3.2.8. Reference intervals VMA and HVA results extracted from the database were separated according to collection type (random or 24 h) and grouped by known age and gender for evaluation of reference limits. Specimens obtained from adults were primarily 24 h urine collections
IP
T
(80%) and reflected excretion of VMA and HVA per day. Most specimens (90%) from
CR
pediatric patients were random urine collections. Analyte concentrations in random collections are normalized using the creatinine concentration of the sample and reported
US
as mass (milligrams) of VMA or HVA relative to mass (grams) of creatinine. Reference
AN
intervals determined for daily excretion and random urine collections from subsets of the data are shown in Table 5. All partitions by age group and creatinine concentration were
M
justified by Harris-Boyd partitioning criteria except separation of HVA values for the 6 to
ED
11 y old groups based on creatinine concentration.
PT
To determine the extent that a low creatinine concentration might elevate a patient result, we calculated reference limits for the urine acids in pediatric specimens separated
CE
into groups of low [<25 mg/dl (2.2 mmol/l)] and adequate [≥25 mg/dl (2.2 mmol/l)]
AC
measured creatinine (Table 5). Specimens collected from individuals aged ≤2 y were more likely to contain reduced amounts of creatinine. Urine acid ratios determined for these specimens were only slightly higher than those calculated for more concentrated specimens from the same age group (Table 5). Fewer dilute specimens were received for testing from older children. Reference limits assigned by urine creatinine concentration for children aged 3-5 and 6-11 y were slightly higher than those determined for specimens without consideration of specimen creatinine. 17
ACCEPTED MANUSCRIPT 4. Discussion As far as we are aware, all other authors [6-15] have reported ionizing VMA and HVA in negative ESI mode as would be expected for small polar acids. During method development, we tested both positive and negative ion modes and found that positive
IP
T
mode produced slightly higher signal for both compounds and that matrix effects were
CR
similar to those experienced in negative mode.
Magera et al. [6,7] used solid phase extraction for sample cleanup prior to analysis
US
in separate VMA and HVA methods and Konieczna et al. [13] utilized dispersive liquid-
AN
liquid microextraction (DLLME) for a neurotransmitter panel including VMA and HVA, but many others [9-12] have opted for minimal specimen preparation (dilute-and-shoot
M
approach). While sample cleanup certainly contributes to method robustness, it also
ED
increases cost, method complexity, and turnaround time, and in some cases (such as most
PT
forms of liquid-liquid extraction including DLLME) it may complicate or preclude automation. When using a dilute-and-shoot approach, great care must be taken to ensure
CE
that matrix effects are minimized by designing elution gradients to exclude analytes from
AC
suppression zones and/or by the use of well-matched SIL internal standards. Some of the reported methods used no internal standard [8,10,12], while others utilized only 1 or 2 non-isotopically labeled IS for an entire panel of analytes [9,13]. Neither of these options provides sufficient robustness for a quantitative high-throughput clinical assay. In our method, we used the post-column infusion experiment to design an LC gradient separating the analytes from major suppression zones and, as internal standards, we chose SIL analogs that exactly co-elute with their respective analytes, thus ensuring accuracy 18
ACCEPTED MANUSCRIPT and robustness without the added cost of sample preparation consumables. To further improve the assay performance, we designed an alternate LC gradient to resolve matrix interferences observed in 3% of the specimens. We opted to re-inject the affected specimens using a short VMA-only method utilizing isocratic elution followed by a
IP
T
cleanout out step to separate VMA from early eluting matric interferences, rather than
CR
designing a longer chromatographic method for both of the analytes, which would reduce throughput.
US
By moving from an HPLC-ECD to an LC-MS/MS platform we were able to reduce
AN
the injection-to-injection time from 21 to 4 min. Some authors have achieved shorter run times, but only in methods for a single analyte [6,7,25]. Since VMA and HVA have
M
similar characteristics, and there is value in measuring both analytes in the same
ED
specimen [16], it is logical to combine the 2 neuroblastoma biomarkers in one method.
PT
Reports of methods for analysis of VMA and HVA in panels with other analytes not related to neuroblastoma reported longer run times (6-20 min).
CE
While Lionetto et al. [9] and Fang et al. [10] listed multiple fragment ions for the
AC
analytes, only Sadilkova et al. [14] in their method for VMA and HVA in serum used a second MRM transition. It is a standard practice in our laboratory to use 2 MRM transitions, qualitative and quantitative, for both the analyte and internal standard to increase assay specificity and thus ensure accurate identification and quantification of the analytes.
19
ACCEPTED MANUSCRIPT Lionetto et al. [9] investigated ion suppression in the presence of 6 drugs, while Shen et al. [25] tested 4 structurally related endogenous compounds. We investigated 62 substances, including endogenous compounds, vitamins, and frequently used over-thecounter drugs and pharmaceuticals, as potential interferences to ensure a robust
IP
T
performance of our dilute-and-shoot method. Only phenylalanine and isoetharine
CR
produced VMA interference >±10% at test concentrations much higher than would be expected in urine specimens and should not therefore interfere. None of the tested
US
compounds interfered with HVA.
AN
Reference intervals determined from archived patient results analyzed using the new LC-MS/MS method corresponded well with reference intervals established for the
M
HPLC-ECD assay and were similar to published values from HPLC-ECD and GC-MS
ED
methods [26-28]. Reference limits determined using analyte concentrations measured in
PT
dilute random urine specimens and expressed as values normalized to creatinine concentration were slightly higher (<10%) than reference intervals calculated for all
AC
CE
results for children <6 y.
5. Conclusions
We have developed and validated a simple, inexpensive, high-throughput LCMS/MS method for the determination of urinary VMA and HVA. This method utilizes a dilute-and-shoot sample preparation in a 96-well format performed using a liquid handler. Method robustness is ensured by the use of co-eluting SIL analogs as internal standards and by chromatography designed to elute analytes outside of major suppression zones. 20
ACCEPTED MANUSCRIPT For increased specificity and accuracy, 2 transitions, qualifier and quantifier, were monitored for each analyte and its SIL internal standard. VMA interferences precluding quantitation in 3.0% of specimens were reduced significantly (to 0.1% of specimens)
CR
IP
study may be of value to health practitioners ordering the tests.
T
using a modified LC gradient. The information obtained from the analytical specificity
Acknowledgements
US
We thank Seyed Sadjadi, Sung Baek, David Davis, Steven Achelis, and the staff of
AN
the Analytic Biochemistry Laboratory for their assistance during this project. This study
AC
CE
PT
ED
M
was supported by the ARUP Institute for Clinical and Experimental Pathology®.
21
ACCEPTED MANUSCRIPT Figure captions Fig. 1. LC-MS/MS chromatograms of (A) 2 mg/l calibration standard, (B) healthy patient urine, and (C) abnormal patient urine: vanillylmandelic acid (VMA), vanillylmandelic acid-d3 (VMA-d3), homovanillic acid (HVA), homovanillic acid-13C6 18O (HVA-13C6
T
O).
CR
IP
18
Fig. 2. Signal suppression scenarios: (A) very little suppression observed in dilute urines,
US
(B) suppression zones that do not interfere with analyte quantitation, (C) and (D) rare
AN
suppression zones at analyte retention times. VMA-d3 and HVA-13C6 18O were infused
M
post-column.
ED
Fig. 3. Mitigation of VMA signal suppression by the alternate VMA method: (A)
PT
specimen 1 analyzed by the primary VMA-HVA method, (B) specimen 1 analyzed using the alternate VMA method, (C) specimen 2 analyzed by the primary VMA-HVA method,
AC
column.
CE
(D) specimen 2 analyzed using the alternate VMA method. VMA-d3 was infused post-
22
ACCEPTED MANUSCRIPT References [1] G. Eisenhofer, R.J. Whitley, T.G. Rosano, Catecholamines and Serotonin, in: C.A. Burtis, E.R. Ashwood, D.E. Bruns (Eds.) Tietz Textbook of Clinical Chemistry and Molecular Diagnostics, Elsevier Saunders, St. Louis, 2012, pp. 851-894.
IP
T
[2] I. Ora, A. Eggert, Progress in treatment and risk stratification of neuroblastoma:
CR
Impact on future clinical and basic research, Semin. Cancer Biol. 21 (2011) 217-228. [3] M.S. Irwin, J.R. Park, Neuroblastoma: Paradigm for precision medicine, Pediatr. Clin.
US
N. Am. 62 (2015) 225-256.
AN
[4] J.M. Maris, M.D. Hogarty, R. Bagatell, S.I. Cohn, Neuroblastoma, Lancet 369 (2007) 2106-2120.
M
[5] W.H.A. de Jong, E.G.E. de Vries, I.P. Kema, Current status and future developments
ED
of LC-MS/MS in clinical chemistry for quantification of biogenic amines, Clin. Biochem.
PT
44 (2011) 95–103.
[6] M.J. Magera, A.L. Thompson, D. Matern, P. Rinaldo, Liquid chromatography–
CE
tandem mass spectrometry method for the determination of vanillylmandelic acid in
AC
urine, Clin. Chem. 49 (2003) 825-826. [7] M.J. Magera, A.L. Stoor, J.K. Helgeson, D. Matern, P. Rinaldo, Determination of homovanillic acid in urine by stable isotope dilution and electrospray tandem mass spectrometry, Clin. Chim. Acta 306 (2001) 35-41. [8] P. Manini, R. Andreoli, S. Cavazzini, E. Bergamaschi, A. Mutti, W.M.A. Niessen, Liquid chromatography–electrospray tandem mass spectrometry of acidic monoamine metabolites, J. Chromatogr. B 744 (2000) 423-431. 23
ACCEPTED MANUSCRIPT [9] L. Lionetto, A.M. Lostia, A. Stigliano, P. Cardelli, M. Simmaco, HPLC–mass spectrometry method for quantitative detection of neuroendocrine tumor markers: Vanillylmandelic acid, homovanillic acid and 5-hydroxyindoleacetic acid, Clin. Chim. Acta 398 (2008) 53–56.
IP
T
[10] L. Fang, Y. Lv, X. Sheng, S. Yao, Sensitive, rapid and easy analysis of three
CR
catecholamine metabolites in human urine and serum by liquid chromatography tandem mass spectrometry, J. Chromatogr. Sci. 50 (2012) 450–456.
US
[11] B. Crow, M. Bishop, E. Paliakov, D. Norton, J. George, J.A. Bralley, Analysis of
AN
urinary aromatic acids by liquid chromatography tandem mass spectrometry, Biomed. Chromatogr. 22 (2008) 1346–1353.
M
[12] W.-Y. Hsu, C.-M. Chen, F.-J. Tsai, C.-C. Lai, Simultaneous detection of diagnostic
ED
biomarkers of alkaptonuria, ornithine carbamoyltransferase deficiency, and
PT
neuroblastoma disease by high-performance liquid chromatography/tandem mass spectrometry, Clin. Chim. Acta 420 (2013) 140–145.
CE
[13] L. Konieczna, A. Roszkowska, M. Niedzwiecki, T. Baczek, Hydrophilic interaction
AC
chromatography combined with dispersive liquid–liquid microextraction as a preconcentration tool for the simultaneous determination of the panel of underivatized neurotransmitters in human urine samples, J. Chromatogr. A 1431 (2016) 111–121. [14] K. Sadilkova, K. Dugaw, D. Benjamin, R.M. Jack, Analysis of vanillylmandelic acid and homovanillic acid by UPLC–MS/MS in serum for diagnostic testing for neuroblastoma, Clin. Chim. Acta 424 (2013) 253–257.
24
ACCEPTED MANUSCRIPT [15] N. Tohmola, O. Itkonen, U. Turpeinen, S. Joenväärä, R. Renkonen, E. Hämäläinen, Preanalytical validation and reference values for a mass spectrometric assay of serum vanillylmandelic acid for screening of catecholamine secreting neuroendocrine tumors, Clin. Chim. Acta 446 (2015) 206–212.
IP
T
[16] G. Cangemi, G. Reggiardo, S. Barco, et al., Prognostic value of ferritin, neuron-
CR
specific enolase, lactate dehydrogenase, and urinary and plasmatic catecholamine metabolites in children with neuroblastoma, Onco. Targets and Therapy 5 (2012) 417-
US
423.
AN
[17] C.N. Svendsen, Multi-electrode assay detectors in high-performance liquid chromatography: A new dimension in electrochemical analysis, Analyst 118 (1993) 123-
M
129.
ED
[18] T.M. Annesley, Ion suppression in mass spectrometry, Clin. Chem. 49 (2003) 1041-
PT
1044.
[19] CLSI. Interference testing in clinical chemistry; Approved guideline-Second edition.
CE
CLSI document EP7-A2. Wayne, PA: Clinical and Laboratory Standards Institute; 2005.
AC
[20] R.G. Hoffmann, Statistics in the practice of medicine, J. Amer. Med. Assoc. 185 (1963) 864-873.
[21] A. Katayev, C. Balciza, D.W. Seccombe, Establishing reference intervals for clinical laboratory test results: Is there a better way?, Am. J. Clin. Pathol. 133 (2010) 180-186. [22] E.K. Harris, J.C. Boyd, On dividing reference data into subgroups to produce separate reference ranges, Clin. Chem. 36 (1990) 265-270.
25
ACCEPTED MANUSCRIPT [23] R.C. Baselt, Carbidopa, Disposition of Toxic Drugs and Chemicals in Man, 9th ed., Biomedical Publications, Seal Beach, California, 2011, pp. 252-253. [24] B. Spilker, B.S. Watson, J.W. Woods, Drug interference with measurement of metanephrines in urine, Ann. Clin. Lab. Sci. 13 (1983) 16-19.
IP
T
[25] Y. Shen, J. Lu, Q. Tang, Q. Guan, Z. Sun, H. Li, L. Cheng, Rapid, easy analysis of
CR
urinary vanillylmandelic acid for diagnostic testing of pheochromocytoma by liquid chromatography tandem mass spectrometry, J. Chromatogr. B 1002 (2015) 92–97.
US
[26] E. Pussard, M. Neveux, N. Guigueno, Reference intervals for urinary
AN
catecholamines and metabolites from birth to adulthood, Clin. Biochem. 42 (2009) 536539.
M
[27] S. Barco, I. Gennai, G. Reggiardo, B. Galleni, L. Barbagallo, A. Maffia, E. Viscardi,
ED
F. De Leonardis, V. Cecinati, S. Sorrentino, A. Garaventa, M. Conte, G. Cangemi,
PT
Urinary homovanillic and vanillylmandelic acid in the diagnosis of neuroblastoma:
848-852.
CE
Report from the Italian Cooperative Group for Neuroblastoma, Clin. Biochem. 47 (2014)
AC
[28] M.T.C. Tran, J. Baglin, T.T. T. Tran, K.T.Hoang, L.T. Phung, A. Read, R.F. Greaves, Development of a new biochemical test to diagnose and monitor neuroblastoma in Vietman: Homovanllic and vanillylmandelic acid by gas chromatography-mass spectrometry, Clin. Biochem. 47 (2014) 206-215.
26
AC
CE
PT
ED
M
AN
US
CR
IP
T
ACCEPTED MANUSCRIPT
Fig. 1 27
AC
CE
PT
ED
M
AN
US
CR
IP
T
ACCEPTED MANUSCRIPT
Fig. 2 28
M
AN
US
CR
IP
T
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
Fig. 3
29
ACCEPTED MANUSCRIPT Highlights
T IP CR US AN M ED PT CE
Validation of inexpensive, dilute-and-shoot LC-MS/MS method for urine VMA and HVA Comprehensive investigation of signal suppression Comprehensive investigation of interference Unique chromatography design Age and gender specific urine reference intervals for random and timed collections
AC
30