- Email: [email protected]
Determination of iohexol in human serum by a semi-automated liquid chromatography tandem mass spectrometry method
Clinical Biochemistry 48 (2015) 679–685
Contents lists available at ScienceDirect
Clinical Biochemistry journal homepage: www.elsevier.com/locate/clinbiochem
Determination of iohexol in human serum by a semi-automated liquid chromatography tandem mass spectrometry method Faye B. Vicente a, Gina K. Vespa a, Fabiola Carrara b, Flavio Gaspari b, Shannon Haymond a,⁎ a b
Department of Pathology, Northwestern University Feinberg School of Medicine, Ann & Robert H. Lurie Children's Hospital of Chicago, USA Laboratory of Pharmacokinetics Clinical Chemistry, IRCCS—Istituto di Ricerche Farmacologiche Mario Negri, Clinical Research Center for Rare Diseases “Aldo e Cele Daccò”, Ranica, BG, Italy
a r t i c l e
i n f o
Article history: Received 26 January 2015 Received in revised form 20 March 2015 Accepted 24 March 2015 Available online 31 March 2015 Keywords: LC-MS/MS Iohexol Kidney function GFR Pediatrics
a b s t r a c t Objectives: Measured glomerular filtration rate (mGFR) is the best indicator of renal function in children and adolescents. GFR determination using iohexol clearance has been increasingly accepted and applied in clinical practice because it is accurate, readily available, non-radioactive, safe and is used intravenously even in the presence of renal disease. This study describes the development and evaluation of a semi-automated method for determination of iohexol in human serum using liquid chromatography coupled with electrospray ionization (ESI) tandem mass spectrometry (LC-MS/MS). Design and methods: Iohexol was extracted from serum using a MICROLAB® NIMBUS4 automation robot and supernatant was dried under nitrogen gas and reconstituted in mobile phase. Ioversol was used as the internal standard. Chromatography was performed using a C-8 analytical column (Phenomenex, 3 μm, 50 × 3.0 mm I.D.) at room temperature and a gradient LC method on a Waters 2795 Alliance HT HPLC system. The flow rate was 0.5 mL/min and the retention times were 2.36 min and 2.14 min for iohexol and ioversol, respectively. Detection by MS/MS was achieved using a (Micromass Quattro Micro) tandem mass spectrometer operated in the ESI-positive mode. The multiple-reaction monitoring (MRM) method used ion transitions m/z 821.9 to 803.7 for iohexol and m/z 807.9 to 588.7 for ioversol. Method validation studies were conducted to determine the linearity, accuracy, precision, matrix effects and stability. A method comparison of blinded, residual patient samples was conducted with a well-established method. Results: The method was linear from 7.7 μg/mL to 2000.0 μg/mL. The low limit of quantification and the detection limit were established at 7.7 and 3.0 μg/mL, respectively. Within-run and between-run precisions were found to be b6% CV and measured values deviated no more than 5% from target concentrations. Carryover and matrix effects were not significant. Comparison to a well-established method showed very good agreement with correlation coefficient of 0.996 for iohexol and 0.993 for GFR/1.73 m2. Conclusions: This method accurately and precisely quantifies iohexol in 50 μL of serum, enabling determination of mGFR by iohexol clearance. The method is highly correlated to a reference method. Use of an automated liquid handler reduces labor-intensive, manual sample preparation steps. The stability of this analyte and the robustness of this assay fit well within our clinical workflow and we have successfully applied this method to determine mGFR in pediatric patients. © 2015 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.
Introduction Measured glomerular filtration rate (mGFR) is the best indicator of renal function in children and adolescents. Accurate assessment of GFR is critical for diagnosing acute and chronic kidney disease, providing early intervention to prevent end-stage renal failure, safely prescribing nephrotoxic and renally cleared drugs and monitoring for
⁎ Corresponding author at: Ann & Robert H. Lurie Children's Hospital of Chicago, Northwestern University Feinberg School of Medicine, 225 East Chicago Avenue, Box 17, Chicago, IL 60611-2605, USA. Fax: +1 312 227 9616. E-mail address: [email protected] (S. Haymond).
adverse side effects from medications. Estimates of GFR are commonly calculated using equations based on creatinine and other parameters (e.g., BUN, cystatin C, race, gender, weight and height). Although relatively inexpensive and convenient, creatinine-based clearance is limited due to dependence on muscle mass, relative insensitivity to detect small changes in renal function and the assumption that extra renal clearance of creatinine is small. In children and adolescents, these equations are particularly problematic [1]. The largest study in children that has directly compared estimating equations with iohexol clearance mGFR showed that the best eGFR formula yielded 87.7% of eGFR within 30% of the iohexol-based mGFR and 45.6% within 10% [2]. Efforts continue to refine and improve estimating equations for use in pediatrics but there are frequently cases where an accurate and clinically useful
http://dx.doi.org/10.1016/j.clinbiochem.2015.03.017 0009-9120/© 2015 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.
680
F.B. Vicente et al. / Clinical Biochemistry 48 (2015) 679–685
method for determination of measured GFR is needed to assess pediatric kidney function. Iohexol (Omnipaque™, GE Healthcare) is an iodinated, watersoluble, nonionic monomeric contrast medium, and it is a suitable marker for GFR, as it is not secreted, metabolized, or reabsorbed by the kidney [3]. The ‘gold standard’ for GFR measurement has been by inulin clearance. However, inulin clearance requires timed urine collections and technically difficult assays and inulin is no longer readily available. Radioactive agents are also used as surrogate markers for GFR, with 51 Cr-EDTA clearance being the most common radioactive agent used. Although 51Cr-EDTA is an excellent marker, it is not available in the United States. Other radioactive agents have been used, but are not as adequate as inulin and are not suitable for repeated measures. Iohexol clearance to determine GFR is a comparable alternative to both inulin and radioactive agents as studies show a close agreement between iohexol and inulin 51Cr-EDTA clearance [4], [5]. GFR determination using iohexol clearance has been increasingly accepted and applied in clinical practice because it is accurate, readily available, nonradioactive, safe and is used intravenously even in the presence of renal disease [2,6]. Methods describing measurement of iohexol in human specimens have been reported using HPLC with UV detection [2,4,7–10] and liquid chromatography mass spectrometry (LC/MS) [11]. Measurement using a semi-automated sample preparation with an LC-MS/MS method provides advantages over other methods that would otherwise require considerable time and labor and allows for minimal sample volume with the highest degree of sensitivity and specificity inherent to LCMS/MS methods. This is the first report of an iohexol method using an automated liquid handling system for sample preparation. Measurement of iohexol-based GFR is not standardized; therefore, results for this ‘gold standard’ assessment may vary across laboratories. Since GFR is based on the analytical measurement of iohexol and the subsequent calculation of its clearance, both components may introduce bias. Little has been reported on the comparison of iohexol methods across laboratories with one recent description of external quality assurance [12]. In this report a candidate method is compared to that used at a reference laboratory, further demonstrating that agreement between analytical methods is possible for iohexol-based GFR [4].
constant at 0.500 mL/min. An instrument-controlled gradient was applied as follows: 0 min, 2% B; 3 min, 100% B; 5 min, 100% B; 5.5 min, 2% B. The retention times were 2.36 min and 2.14 min for iohexol and ioversol, respectively. Solvent flow was diverted from the source to waste at 0 to 1 min and at 5 to 7.5 min. Total sample data acquisition time was 4 min. After analytical runs are completed, the column is flushed for 45 min at a flow rate of 0.250 mL/min and stored with 70% methanol in water. The mass spectrometer was operated by the MassLynx v4.1 software and data analysis was conducted using QuanLynx v4.1. Unit resolution for Q1 and Q3 was used for detection. MS/MS conditions were optimized to the parameters shown in Table 1. Ion transitions monitored using MRM were m/z 821.9 → 803.7 for iohexol and m/z 807.9 → 588.7 for ioversol. Sample collection and storage A BD Vacutainer® Serum Tube containing a clot activator was used to collect 3.0 mL of whole blood (minimum 0.3 mL). After mixing by gentle tube inversion (5–10 times), the blood was allowed to clot for at least 30 min while standing at room temperature. Serum was separated from the cells using a high-speed, horizontal bench-top centrifuge (StatSpin® Express 4, Iris Sample Processing, Westwood, MA, USA) operated for 3 min to produce a force of 4000 ×g at 5100 rpm. Serum was transferred to a labeled aliquot tube and stored at −20 °C prior to analysis. Sample and reagent preparation Stock solutions, calibrators and quality controls Stock solutions for calibrators (20 mg/mL) and controls (7.5 mg/mL) were prepared by dissolving iohexol in water using 2 separately weighed standard materials. These solutions were used to prepare serum-based calibrators at 10, 100, 500 and 1000 μg/mL and controls at 15, 150 and 750 μg/mL, respectively. Charcoal dextran stripped serum was selected as the base matrix for calibrators and control materials because it is commercially available and represents a close match to the clinical sample matrix. Ioversol was dissolved in methanol to create a 5 mg/mL stock solution. This solution was used to prepare the protein precipitation solution in methanol at 40 μg/mL. Stock solutions were stored at −70 °C and working solutions were stored at −20 °C.
Experimental Chemicals Iohexol (C19H26I3N3O9) and ioversol (C18H24I3N3O9) standards were from U.S. Pharmacopeial Convention (Rockville, MD, USA). Water (Omnisolv, LC/MS grade) and methanol (Optima, LC/MS grade) were obtained from EMD Millipore Corporation (Billerca, MA, USA) and Fisher Scientific (Fair Lawn, NJ, USA), respectively. Ammonium acetate (HPLC grade) and formic acid (ACS-certified, 99+% purity) were obtained from Sigma-Aldrich (St. Louis, MO, USA). Charcoal dextran stripped human serum was from Innovative Research (Novi, MI, USA). Instrument and method details LC-MS/MS conditions LC-MS/MS was performed using a Waters 2795 Alliance HT HPLC system coupled to a Micromass Quattro Micro tandem quadrupole mass spectrometer (Waters Corporation, Milford, MA) operated in the electrospray positive ion mode with multiple reaction monitoring (MRM). Chromatographic separation was performed at room temperature with a Luna C8 analytical column (3 μm, 50 × 3.0 mm I.D.; Phenomenex, USA) preceded by a C8 (4 × 2.0 mm I.D.; Phenomenex, USA) guard column. Mobile phase A was 2 mM ammonium acetate and 0.1% formic acid in water and mobile phase B was 2 mM ammonium acetate and 0.1% formic acid in methanol. The flow rate was held
Semi-automated sample preparation Sample preparation involved protein precipitation of serum with methanol containing ioversol internal standard followed by vortex mixing and centrifugation. The supernatant was dried down,
Table 1 MS/MS instrument parameters. Capillary voltage Cone voltage Source T Desolvation T Cone gas Desolvation gas Collision gas P LM1 resolution HM1 resolution Ion energy 1 MS/MS entrance Collision (iohexol) Collision (ioversol) MS/MS exit LM2 resolution HM2 resolution Ion energy 2 Dwell Inter-channel delay Inter-scan delay
1.4 kV 38 V 130 °C 400 °C 35 L/h 700 L/h 3.70 E−3 mbar 14.5 14.5 0.5 −2 21 eV 24 eV 1 13.2 13.2 2.3 0.2 s 0.03 s 0.03 s
F.B. Vicente et al. / Clinical Biochemistry 48 (2015) 679–685
reconstituted with mobile phase A and centrifuged prior to analysis. A multi-channel pipetting workstation was used to automate some steps of the sample preparation as follows. The MICROLAB® NIMBUS4 (Hamilton Company, Reno, NV, US) automated multi-channel pipetting workstation, controlled by a computer running Venus software, was used for aliquotting samples, protein precipitation solution and reconstitution solvent. The following supplies were loaded on the deck of the pipetting workstation: 1.5 mL conical polypropylene microcentrifuge tubes, borosilicate glass test tubes (13 × 100 mm) and HPLC vials. Serum samples, calibrators, quality control materials and troughs filled with the respective reagents were placed in the carriers and pedestals on the deck. 50 μL of samples, calibrators and QCs were transferred from the appropriate source tubes by the pipetting workstation into each microcentrifuge tube followed by addition of the 400 μL of precipitation solution. The tubes were capped and taken offline to be vortexed manually for 30 s followed by centrifugation at 9015 rcf for 1 min at room temperature. The microcentrifuge tubes were then returned to the NIMBUS4 deck and 100 μL from each tube was transferred to a labeled glass tube. The glass tubes were removed from the pipetting workstation and the supernatant was evaporated until dryness under nitrogen gas at room temperature for 15 min using a Reacti-Therm III Heating/ Stirring module (Thermo Scientific). Using the pipetting workstation, the residue was reconstituted with 2 mL of mobile phase A. The tubes were capped and vortexed mixed followed by centrifugation at 1430– 1500 rcf for 5 min. 1 mL of the reconstituted solution was transferred by the pipetting workstation to a labeled Maximum Recovery HPLC autosampler vial (Waters Corporation, Milford, MA). 3 μL was injected into the LC-MS/MS system. Method validation Linearity and lower limit of quantitation Linearity was confirmed over the calibration range using least square linear regression analysis and calculation of coefficient of determination (r2). The material used to verify linearity above the calibration range was prepared by adding approximately 2000 μg/mL (1996.9 μg/mL) iohexol to charcoal-stripped serum and performing serial dilution to at least 6 levels. The lower limit of quantification (LLOQ) was defined as the lowest iohexol concentration with signal-to-noise ≥ 10, when repeatedly sampled (n = 3) demonstrated a CV ≤ 20% and recovery within 95–105% of the accepted value. The lower limit of detection (LOD) was defined as the lowest concentration resulting in a signal-to-noise response ≥ 3 and a mean value 3SD greater than that measured in blank samples. Precision and accuracy Precision was evaluated following a modified CLSI EP10-A3AMD protocol, using 3 concentrations of QCs prepared in serum (13.2, 130.5 and 786.3 μg/mL) [13]. To evaluate intra-day and inter-day precision, 4 replicates of the mid level QC and 3 replicates of each low and high level QCs were extracted and analyzed in 2 batches on each of 4 days. The acceptance criterion for precision was ≤10% CV at all concentrations. There is no higher order reference method or certified reference material available for assessing trueness of iohexol measurements. Therefore, accuracy of the method was assessed by measurement of QCs at multiple, clinically relevant concentrations (as described above for precision). Intra-day and inter-day accuracy was determined as the percent relative error by comparing measured iohexol concentrations against nominal target concentrations and acceptability was defined as a mean deviation of no more than 10% for all concentrations. Method comparison A method comparison was conducted by sending blinded residual specimens (n = 80) to the Mario Negri Institute for parallel measurement of iohexol using a well-established method [4]. A
681
comparison of the GFR calculated using iohexol values from both methods was also performed. Data were analyzed by Deming regression and Bland–Altman plots. The concordance correlation coefficient for the methods was calculated. A Wilcoxon test and a t-test were performed to assess for statistical differences in the data sets. Statistical analysis was performed using MedCalc Statistical Software version 14.12.0 (MedCalc Software bvba, Ostend, Belgium; 2014). Selectivity, carryover and matrix effects Confirmation of analyte identity was determined through retention time and the monitoring of unique mass transitions. On each of 4 days, three different levels of quality control material were run in a specific order following a modified EP10-A3AMD protocol (mid, high, low, mid, mid, low, low, high, high, mid) to determine the presence of carryover [13]. Matrix effects were evaluated by a method based on that of Matuszewski et al. [14]. Iohexol and ioversol were spiked at 2 concentrations into serum from 5 different sources after extraction and resultant peak areas were compared to those obtained from the same concentration spiked into mobile phase A. Iohexol was evaluated at 653.1 and 13.1 μg/mL and ioversol at 495.8 and 9.9 μg/mL. Iohexolfree serum pools were created using the following: (a) charcoalstripped serum from a commercial source, (b) leftover hemolyzed patient serum, (c) leftover lipemic patient serum, (d) leftover serum from pediatric patients with kidney disease, and (e) leftover serum from general pediatric patients. Matrix effect was calculated as the ratio of the area of the analyte peak post-extraction to the area of the analyte peak in mobile phase A, multiplied by 100. To assess the possibility of ion suppression for iohexol and ioversol, commercial serum and serum pools from kidney and stem cell transplant pediatric patients were extracted without internal standard. A 100.8 μg/mL solution of iohexol in water was introduced into the ion source by a syringe pump at 10 μL/min through a T-connector and then the extracted sample was introduced by injection from the HPLC autosampler. Each chromatogram was inspected for the presence of signal reduction at the retention time of iohexol and ioversol. The same procedure was performed using a 101.0 μg/mL of ioversol in water. Stability The stability of iohexol in human serum was assessed using various conditions that are relevant to a typical clinical laboratory workflow. Effect of storage temperature on pre- and post-extracted samples was evaluated (ambient for 8 h, 4 °C for 24 and 36 h and − 20 °C for 1 mo). Stability after freeze–thaw cycles when stored at − 20 °C was determined for pre-extracted samples. Clinical application An iohexol clearance procedure derived from that of Schwartz and colleagues in the NIDDK CKiD study [2] was used to determine GFR in mL/min/1.73 m2. This involved administration of 5 mL of Omnipaque 300™ with serial blood samples collected over 300 min. Sample collection, processing, storage and analysis followed the procedures described above. Iohexol dose (mg I) was calculated by weighing the syringe preand post-infusion. Iohexol was measured by LC-MS/MS at a total of four sampling intervals (10, 30, 120, and 300 min post-iohexol infusion). The disappearance of iohexol was resolved into two curves using the logarithm of the concentration of iohexol as a function of time. The slow (renal) curve was determined from the concentrations obtained at 120 and 300 min. The fast curve was determined from 10 and 30 min concentrations, corrected for the contribution of the slow curve. Formulae for calculations of GFR were programmed into an Excel™ spreadsheet and validated with a two-compartment pharmacokinetic mathematical model over a wide range of compartment volumes and elimination constants. GFR was calculated as:
682
F.B. Vicente et al. / Clinical Biochemistry 48 (2015) 679–685
(a)
(b)
(c)
(d)
Fig. 1. Representative mass spectra of iohexol (a–b) and ioversol (c–d).
GFR = I / (exp A/α + exp B/β) × 1.73 / BSA, in mL/min per 1.73 m2; I: the dose of iohexol (mg I), calculated using concentration of 647 mg/mL with density 1.345 g/L; exp A: the intercept of the slow curve, and α its corresponding slope; exp B: the intercept of the fast curve, and β its corresponding slope; GFR was corrected to 1.73 m2 body surface area (BSA) by the ratio 1.73 / BSA. BSA was calculated using the Haycock formula with weight in kg and height in cm: BSA(m2) = 0.024265 × Wt0.5378 × Ht0.3964 [15].
Results and discussion LC and MS/MS method selection Mass spectra The selected precursor (m/z 821.9) and product (m/z 803.7) ions used for detection of iohexol are shown in the representative mass spectra (Figs. 1a and b). The internal standard (ioversol) was detected using precursor and product ions of m/z 807.9 and 588.7, respectively (Figs. 1c
Fig. 2. Representative chromatograms of (a) charcoal-stripped human serum, (b) iohexol (21.5 μg/mL) and (c) ioversol (39.7 μg/mL) in human serum.
F.B. Vicente et al. / Clinical Biochemistry 48 (2015) 679–685
and d). Use of ioversol as an internal standard is limited in that it is not a stable isotope-labeled form of iohexol. Therefore, its physicochemical properties are not identical to that of iohexol, creating potential for variability in signal response to ion suppression or enhancement that may be encountered during sample processing and analysis. This may explain the differences observed between iohexol and ioversol during the matrix effect study at the lowest measured concentration. The product ions selected for the multiple reaction monitoring (MRM) experiment were the most stable and abundant peaks observed during MS optimization and correspond to a water loss (m/z 18) for iohexol and a loss of m/z 219 for ioversol. Although the chosen transitions are different, are relatively non-specific and an analog is used as the internal standard, results of the method validation suggest that the impact of these design limitations on the quantitation of iohexol is minimal and is not clinically significant. Chromatograms Fig. 2 shows the representative chromatograms of (a) serum sample not containing iohexol (i.e., pre-dose), (b) serum sample collected at 300 min post-iohexol infusion iohexol (21.5 μg/mL) and (c) the internal standard, ioversol (39.7 μg/mL), added to a serum sample collected at 300 min post-iohexol infusion. Retention times were assigned as iohexol: 2.36 min (acceptable range: 2.16 to 2.56 min) and ioversol: 2.14 min (acceptable range: 1.94 to 2.34 min). Linearity and lower limit of quantitation AMR and regression analysis The iohexol assay was linear from 7.7 to 2000 μg/mL, which is sufficient for GFR determination over a wide range of kidney function (normal to late-stage CKD). The linear regression analysis of measured versus expected concentrations showed excellent correlation with r2 N 0.999. Calibrators were prepared at 10 – 1000 μg/mL, meeting the clinical requirements for this assay. The AMR is verified every 6 months. QC concentrations were selected at 15, 150 and 750 μg/mL to cover the range typically measured in our population and represent medically relevant intervals observed during the iohexol clearance protocol.
683
Method comparison A correlation of this iohexol assay and that of an international reference method, previously described [4] was conducted for 80 pediatric specimens ranging from 15 to 694 μg/mL iohexol. Fig. 3 shows the analysis of the correlation data. The concordance correlation coefficient (CCC) of the methods, a parameter that reflects both accuracy and precision, was 0.996. The method comparison of iohexol concentration showed a Deming slope of 1.019 and intercept of −1.385. A Bland–Altman plot shows good agreement between the 2 methods. A Wilcoxon test (data are not normally distributed) showed no statistical difference between the reference and candidate methods (p = 0.1321). A t-test performed on the log transformed data yielded the same results. The GFR (per 1.73 m2) values obtained using iohexol concentrations from the 2 laboratories were also compared and showed high correlation (CCC = 0.993). Fig. 4 contains the analysis of GFR values from 20 patients, ranging from 25 to 135 mL/min/1.73 m2. Regression analysis demonstrated a Deming slope of 1.030 with intercept of −2.408. The Deming regression analyses for both the iohexol concentration and the GFR showed slopes of 1 and intercepts near zero with 95% confidence intervals including 1 and 0, respectively. Therefore, neither a constant difference nor a proportional difference exists between these methods. The concordance correlation coefficient (CCC) of the methods demonstrated a high degree of agreement between the methods for iohexol concentration and GFR. This agreement was also observed in the Bland–Altman plots.
(a) R = 0.996 Deming fit: Y= 1.019X – 1.385
LOD and LLOQ with imprecision The LOD was defined by diluting the lowest level calibrator and measuring in replicate. The LOD for the iohexol assay was 3.0 μg/mL, as this was the lowest concentration at which the analyte peak was detected with a S/N N 3 and the mean value was 3SD above the value of the blank samples. The LLOQ was determined from linearity experiments as 7.7 μg/mL, since this concentration met the defined criteria with average S/N 10.2, CV 5.1% and recovery of 97%. Precision and accuracy Results from intra- and inter-day precision experiments are summarized in Table 2. Intra-day precision ranged from 1.0 to 3.8% and interday precision was 1.5–3.9% for serum-based QC samples. Table 2 also contains mean accuracy data for the multiple concentrations, showing that all measured concentrations were within 2.2–5.0% and 2.1–4.4%% of the target concentration within a run and between runs, respectively. Table 2 Within run and run-to-run precision and accuracy results for quantitation of iohexol in human serum. Sample
Conc. (μg/mL)
Intra-day (within run)
Inter-day (run-to-run)
SD
CV
% Relative error
SD
CV
% Relative error
QC1 QC2 QC3
13.2 130.5 786.3
0.52 3.27 8.22
3.8% 2.5% 1.0%
5.0% 2.3% 2.2%
0.54 2.13 12.30
3.9% 1.6% 1.5%
4.4% 2.3% 2.1%
(b) +1.96 SD
Mean
-1.96 SD
Fig. 3. Deming linear regression (a) and Bland–Altman (b) plots of serum iohexol results obtained by 2 methods. Mean percent difference = 0.1% with upper limit of agreement (+1.96 SD) = 12.1% and lower limit of agreement (−1.96 SD) = −11.9%.
684
F.B. Vicente et al. / Clinical Biochemistry 48 (2015) 679–685
(a)
Table 3 Matrix effect for iohexol and ioversol in 5 pools of human serum. Matrix effect
Iohexol (653.1 μg/mL)
Iohexol (13.1 μg/mL)
Ioversol (495.8μg/mL)
Ioversol (9.9 μg/mL)
(b)
A B C D E A B C D E A B C D E A B C D E
%ME
Mean
Std. dev.
%CV
101 100 105 102 103 90 97 98 102 101 100 97 100 105 102 91 84 82 85 83
102
1.94
1.9
97
4.14
4.3
101
2.91
2.9
85
3.50
4.1
The 5 matrix pools evaluated were: (A) charcoal-stripped serum from a commercial source, (B) hemolyzed patient pool, (C) lipemic patient pool, (D) kidney disease patient pool, and (E) pediatric patient pool. +1.96 SD
Mean -1.96 SD
Fig. 4. Deming linear regression (a) and Bland–Altman (b) plots of calculated GFR results obtained by 2 methods. Mean percent difference = −0.7% with upper limit of agreement (+1.96 SD) = 12.1% and lower limit of agreement (−1.96 SD) = −13.4%.
A difference in GFR mL/min/1.73 m2 greater than 10% was observed for one patient. In this case, the candidate method (slightly) underestimated the first 3 concentrations and overestimated that at the final timepoint. This resulted in a slower elimination phase, per the 2-compartment model and, as a consequence, a lower GFR (22.4 vs 28.3 mL/min/1.73 m2). Although the methods are highly correlated, small over- or underestimations in iohexol concentration may lead to important differences in GFR when determined using a 2-compartment kinetic model based on only 4 timepoints.
This may be attributed to the physicochemical differences between these molecules and dissimilarities in selected MRM transitions. Stability Iohexol stability in serum samples was assessed under a variety of conditions relevant to the workflow in our clinical laboratory. Serum samples stored 24 h at 4 °C, 1 mo at − 20 °C and 3–4 mo at − 20 °C after 3 freeze–thaw cycles showed b 3% difference in mean values obtained when prepared and analyzed promptly after processing. Postextraction samples were stable for 8 h ambient, after 36 h at 5 °C in the autosampler and after 1 mo when frozen at −20 °C, deviating less than 5% from the mean value obtained when analyzed promptly after extraction. These stability results compare well to that previously published for LC-MS/MS and LC-UV methods for iohexol analysis [16,17]. Clinical applicability The serum concentration–time profiles of 3 representative pediatric patients are depicted in Fig. 5, demonstrating the clinical applicability of this method. Each patient required measured GFR due to commonly acknowledged limitations in estimated GFR calculations in children.
Selectivity, carryover and matrix effects Extracted serum, free of iohexol and ioversol (evaluated across 15 lots), and LLOQ samples (n = 6) were free from coeluting peaks. Carryover was assessed by monitoring the effect of multiple high concentration injections followed by low concentration injections in a defined pattern. The level of carryover in this assay was determined to be insignificant by an independent sample t-test comparing the means of the results for the high and low sequential samples. Matrix effects were evaluated by comparing the instrument response of postextraction spiked samples to that of mobile-phase spiked samples. Table 3 contains the detailed results for the matrix effect experiments. Response in biological matrix compared to neat solutions showed acceptable recovery and CV across the 5 matrix pools ranged from 2 to 5%, indicating the relative matrix effect in the method is minimal and clinically acceptable. Recovery of ioversol at 9.9 μg/mL is slightly reduced compared to that of iohexol at 13.1 μg/mL, suggesting that the former is more susceptible to matrix effect at lower concentrations.
Fig. 5. Iohexol serum concentration–time profile for 3 pediatric patients after administration of 5 mL Omnipaque 300™. GFRs were determined to be 92.4 (◆), 49.4 (∎) and 22.4 (▲) mL/min/1.73 m2, respectively.
F.B. Vicente et al. / Clinical Biochemistry 48 (2015) 679–685
Conclusions This report describes the development and validation of a LC-MS/MS method with semi-automated sample preparation for determination of iohexol in human serum. This method accurately and precisely quantifies iohexol in 50 μL of serum. The method is highly correlated to a reference method. We have determined that the stability of this analyte and the robustness of this assay fit well within our clinical workflow. We have successfully applied this method to clinical practice as the iohexol values measured have been fit to an elimination module that enables determination of the measured GFR in pediatric patients. Acknowledgments The authors wish to thank Dr. Frederick Smith for his support and contributions to this project. References [1] Zappitelli M, Parvex P, Joseph L, et al. Derivation and validation of cystatin C-based prediction equations for GFR in children. Am J Kidney Dis: Off J Natl Kidney Found 2006;48:221–30. [2] Schwartz GJ, Furth S, Cole SR, Warady B, Munoz A. Glomerular filtration rate via plasma iohexol disappearance: pilot study for chronic kidney disease in children. Kidney Int 2006;69:2070–7. [3] Olsson B, Aulie A, Sveen K, Andrew E. Human pharmacokinetics of iohexol. A new nonionic contrast medium. Invest Radiol 1983;18:177–82. [4] Gaspari F, Perico N, Ruggenenti P, et al. Plasma clearance of nonradioactive iohexol as a measure of glomerular filtration rate. J Am Soc Nephrol: JASN 1995;6: 257–63. [5] Brown SC, O'Reilly PH. Iohexol clearance for the determination of glomerular filtration rate in clinical practice: evidence for a new gold standard. J Urol 1991;146: 675–9.
685
[6] Schwartz GJ, Work DF. Measurement and estimation of GFR in children and adolescents. Clin J Am Soc Nephrol: CJASN 2009;4:1832–43. [7] Brandstrom E, Grzegorczyk A, Jacobsson L, Friberg P, Lindahl A, Aurell M. GFR measurement with iohexol and 51Cr-EDTA. A comparison of the two favoured GFR markers in Europe. Nephrol Dial Transplant: Off Publ Eur Dial Transplant Assoc— Eur Ren Assoc 1998;13:1176–82. [8] Castagnet S, Blasco H, Vourc'h P, et al. Routine determination of GFR in renal transplant recipients by HPLC quantification of plasma iohexol concentrations and comparison with estimated GFR. J Clin Lab Anal 2012;26:376–83. [9] Cavalier E, Rozet E, Dubois N, et al. Performance of iohexol determination in serum and urine by HPLC: validation, risk and uncertainty assessment. Clin Chim Acta; Int J Clin Chem 2008;396:80–5. [10] Slack A, Tredger M, Brown N, Corcoran B, Moore K. Application of an isocratic methanol-based HPLC method for the determination of iohexol concentrations and glomerular filtration rate in patients with cirrhosis. Ann Clin Biochem 2014; 51(Pt 1):80–8. [11] Denis MC, Venne K, Lesiege D, et al. Development and evaluation of a liquid chromatography-mass spectrometry assay and its application for the assessment of renal function. J Chromatogr A 2008;1189:410–6. [12] Luis-Lima S, Gaspari F, Porrini E, et al. Measurement of glomerular filtration rate: internal and external validations of the iohexol plasma clearance technique by HPLC. Clin Chim Acta; Int J Clin Chem 2014;430:84–5. [13] Krouwer JSC GS, Tholen DW. Preliminary Evaluation of Quantitative Clinical Laboratory. Measurement Procedures; Approved Guideline. 3rd ed.; 2014. [14] Matuszewski BK, Constanzer ML, Chavez-Eng CM. Strategies for the assessment of matrix effect in quantitative bioanalytical methods based on HPLC-MS/MS. Anal Chem 2003;75:3019–30. [15] Haycock GB, Schwartz GJ, Wisotsky DH. Geometric method for measuring body surface area: a height-weight formula validated in infants, children, and adults. J Pediatr 1978;93:62–6. [16] Lee SY, Chun MR, Kim DJ, Kim JW. Determination of iohexol clearance by highperformance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS). J Chromatogr B Analyt Technol Biomed Life Sci 2006;839:124–9. [17] Soman RS, Zahir H, Akhlaghi F. Development and validation of an HPLC-UV method for determination of iohexol in human plasma. J Chromatogr B Analyt Technol Biomed Life Sci 2005;816:339–43.
Contents lists available at ScienceDirect
Clinical Biochemistry journal homepage: www.elsevier.com/locate/clinbiochem
Determination of iohexol in human serum by a semi-automated liquid chromatography tandem mass spectrometry method Faye B. Vicente a, Gina K. Vespa a, Fabiola Carrara b, Flavio Gaspari b, Shannon Haymond a,⁎ a b
Department of Pathology, Northwestern University Feinberg School of Medicine, Ann & Robert H. Lurie Children's Hospital of Chicago, USA Laboratory of Pharmacokinetics Clinical Chemistry, IRCCS—Istituto di Ricerche Farmacologiche Mario Negri, Clinical Research Center for Rare Diseases “Aldo e Cele Daccò”, Ranica, BG, Italy
a r t i c l e
i n f o
Article history: Received 26 January 2015 Received in revised form 20 March 2015 Accepted 24 March 2015 Available online 31 March 2015 Keywords: LC-MS/MS Iohexol Kidney function GFR Pediatrics
a b s t r a c t Objectives: Measured glomerular filtration rate (mGFR) is the best indicator of renal function in children and adolescents. GFR determination using iohexol clearance has been increasingly accepted and applied in clinical practice because it is accurate, readily available, non-radioactive, safe and is used intravenously even in the presence of renal disease. This study describes the development and evaluation of a semi-automated method for determination of iohexol in human serum using liquid chromatography coupled with electrospray ionization (ESI) tandem mass spectrometry (LC-MS/MS). Design and methods: Iohexol was extracted from serum using a MICROLAB® NIMBUS4 automation robot and supernatant was dried under nitrogen gas and reconstituted in mobile phase. Ioversol was used as the internal standard. Chromatography was performed using a C-8 analytical column (Phenomenex, 3 μm, 50 × 3.0 mm I.D.) at room temperature and a gradient LC method on a Waters 2795 Alliance HT HPLC system. The flow rate was 0.5 mL/min and the retention times were 2.36 min and 2.14 min for iohexol and ioversol, respectively. Detection by MS/MS was achieved using a (Micromass Quattro Micro) tandem mass spectrometer operated in the ESI-positive mode. The multiple-reaction monitoring (MRM) method used ion transitions m/z 821.9 to 803.7 for iohexol and m/z 807.9 to 588.7 for ioversol. Method validation studies were conducted to determine the linearity, accuracy, precision, matrix effects and stability. A method comparison of blinded, residual patient samples was conducted with a well-established method. Results: The method was linear from 7.7 μg/mL to 2000.0 μg/mL. The low limit of quantification and the detection limit were established at 7.7 and 3.0 μg/mL, respectively. Within-run and between-run precisions were found to be b6% CV and measured values deviated no more than 5% from target concentrations. Carryover and matrix effects were not significant. Comparison to a well-established method showed very good agreement with correlation coefficient of 0.996 for iohexol and 0.993 for GFR/1.73 m2. Conclusions: This method accurately and precisely quantifies iohexol in 50 μL of serum, enabling determination of mGFR by iohexol clearance. The method is highly correlated to a reference method. Use of an automated liquid handler reduces labor-intensive, manual sample preparation steps. The stability of this analyte and the robustness of this assay fit well within our clinical workflow and we have successfully applied this method to determine mGFR in pediatric patients. © 2015 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.
Introduction Measured glomerular filtration rate (mGFR) is the best indicator of renal function in children and adolescents. Accurate assessment of GFR is critical for diagnosing acute and chronic kidney disease, providing early intervention to prevent end-stage renal failure, safely prescribing nephrotoxic and renally cleared drugs and monitoring for
⁎ Corresponding author at: Ann & Robert H. Lurie Children's Hospital of Chicago, Northwestern University Feinberg School of Medicine, 225 East Chicago Avenue, Box 17, Chicago, IL 60611-2605, USA. Fax: +1 312 227 9616. E-mail address: [email protected] (S. Haymond).
adverse side effects from medications. Estimates of GFR are commonly calculated using equations based on creatinine and other parameters (e.g., BUN, cystatin C, race, gender, weight and height). Although relatively inexpensive and convenient, creatinine-based clearance is limited due to dependence on muscle mass, relative insensitivity to detect small changes in renal function and the assumption that extra renal clearance of creatinine is small. In children and adolescents, these equations are particularly problematic [1]. The largest study in children that has directly compared estimating equations with iohexol clearance mGFR showed that the best eGFR formula yielded 87.7% of eGFR within 30% of the iohexol-based mGFR and 45.6% within 10% [2]. Efforts continue to refine and improve estimating equations for use in pediatrics but there are frequently cases where an accurate and clinically useful
http://dx.doi.org/10.1016/j.clinbiochem.2015.03.017 0009-9120/© 2015 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.
680
F.B. Vicente et al. / Clinical Biochemistry 48 (2015) 679–685
method for determination of measured GFR is needed to assess pediatric kidney function. Iohexol (Omnipaque™, GE Healthcare) is an iodinated, watersoluble, nonionic monomeric contrast medium, and it is a suitable marker for GFR, as it is not secreted, metabolized, or reabsorbed by the kidney [3]. The ‘gold standard’ for GFR measurement has been by inulin clearance. However, inulin clearance requires timed urine collections and technically difficult assays and inulin is no longer readily available. Radioactive agents are also used as surrogate markers for GFR, with 51 Cr-EDTA clearance being the most common radioactive agent used. Although 51Cr-EDTA is an excellent marker, it is not available in the United States. Other radioactive agents have been used, but are not as adequate as inulin and are not suitable for repeated measures. Iohexol clearance to determine GFR is a comparable alternative to both inulin and radioactive agents as studies show a close agreement between iohexol and inulin 51Cr-EDTA clearance [4], [5]. GFR determination using iohexol clearance has been increasingly accepted and applied in clinical practice because it is accurate, readily available, nonradioactive, safe and is used intravenously even in the presence of renal disease [2,6]. Methods describing measurement of iohexol in human specimens have been reported using HPLC with UV detection [2,4,7–10] and liquid chromatography mass spectrometry (LC/MS) [11]. Measurement using a semi-automated sample preparation with an LC-MS/MS method provides advantages over other methods that would otherwise require considerable time and labor and allows for minimal sample volume with the highest degree of sensitivity and specificity inherent to LCMS/MS methods. This is the first report of an iohexol method using an automated liquid handling system for sample preparation. Measurement of iohexol-based GFR is not standardized; therefore, results for this ‘gold standard’ assessment may vary across laboratories. Since GFR is based on the analytical measurement of iohexol and the subsequent calculation of its clearance, both components may introduce bias. Little has been reported on the comparison of iohexol methods across laboratories with one recent description of external quality assurance [12]. In this report a candidate method is compared to that used at a reference laboratory, further demonstrating that agreement between analytical methods is possible for iohexol-based GFR [4].
constant at 0.500 mL/min. An instrument-controlled gradient was applied as follows: 0 min, 2% B; 3 min, 100% B; 5 min, 100% B; 5.5 min, 2% B. The retention times were 2.36 min and 2.14 min for iohexol and ioversol, respectively. Solvent flow was diverted from the source to waste at 0 to 1 min and at 5 to 7.5 min. Total sample data acquisition time was 4 min. After analytical runs are completed, the column is flushed for 45 min at a flow rate of 0.250 mL/min and stored with 70% methanol in water. The mass spectrometer was operated by the MassLynx v4.1 software and data analysis was conducted using QuanLynx v4.1. Unit resolution for Q1 and Q3 was used for detection. MS/MS conditions were optimized to the parameters shown in Table 1. Ion transitions monitored using MRM were m/z 821.9 → 803.7 for iohexol and m/z 807.9 → 588.7 for ioversol. Sample collection and storage A BD Vacutainer® Serum Tube containing a clot activator was used to collect 3.0 mL of whole blood (minimum 0.3 mL). After mixing by gentle tube inversion (5–10 times), the blood was allowed to clot for at least 30 min while standing at room temperature. Serum was separated from the cells using a high-speed, horizontal bench-top centrifuge (StatSpin® Express 4, Iris Sample Processing, Westwood, MA, USA) operated for 3 min to produce a force of 4000 ×g at 5100 rpm. Serum was transferred to a labeled aliquot tube and stored at −20 °C prior to analysis. Sample and reagent preparation Stock solutions, calibrators and quality controls Stock solutions for calibrators (20 mg/mL) and controls (7.5 mg/mL) were prepared by dissolving iohexol in water using 2 separately weighed standard materials. These solutions were used to prepare serum-based calibrators at 10, 100, 500 and 1000 μg/mL and controls at 15, 150 and 750 μg/mL, respectively. Charcoal dextran stripped serum was selected as the base matrix for calibrators and control materials because it is commercially available and represents a close match to the clinical sample matrix. Ioversol was dissolved in methanol to create a 5 mg/mL stock solution. This solution was used to prepare the protein precipitation solution in methanol at 40 μg/mL. Stock solutions were stored at −70 °C and working solutions were stored at −20 °C.
Experimental Chemicals Iohexol (C19H26I3N3O9) and ioversol (C18H24I3N3O9) standards were from U.S. Pharmacopeial Convention (Rockville, MD, USA). Water (Omnisolv, LC/MS grade) and methanol (Optima, LC/MS grade) were obtained from EMD Millipore Corporation (Billerca, MA, USA) and Fisher Scientific (Fair Lawn, NJ, USA), respectively. Ammonium acetate (HPLC grade) and formic acid (ACS-certified, 99+% purity) were obtained from Sigma-Aldrich (St. Louis, MO, USA). Charcoal dextran stripped human serum was from Innovative Research (Novi, MI, USA). Instrument and method details LC-MS/MS conditions LC-MS/MS was performed using a Waters 2795 Alliance HT HPLC system coupled to a Micromass Quattro Micro tandem quadrupole mass spectrometer (Waters Corporation, Milford, MA) operated in the electrospray positive ion mode with multiple reaction monitoring (MRM). Chromatographic separation was performed at room temperature with a Luna C8 analytical column (3 μm, 50 × 3.0 mm I.D.; Phenomenex, USA) preceded by a C8 (4 × 2.0 mm I.D.; Phenomenex, USA) guard column. Mobile phase A was 2 mM ammonium acetate and 0.1% formic acid in water and mobile phase B was 2 mM ammonium acetate and 0.1% formic acid in methanol. The flow rate was held
Semi-automated sample preparation Sample preparation involved protein precipitation of serum with methanol containing ioversol internal standard followed by vortex mixing and centrifugation. The supernatant was dried down,
Table 1 MS/MS instrument parameters. Capillary voltage Cone voltage Source T Desolvation T Cone gas Desolvation gas Collision gas P LM1 resolution HM1 resolution Ion energy 1 MS/MS entrance Collision (iohexol) Collision (ioversol) MS/MS exit LM2 resolution HM2 resolution Ion energy 2 Dwell Inter-channel delay Inter-scan delay
1.4 kV 38 V 130 °C 400 °C 35 L/h 700 L/h 3.70 E−3 mbar 14.5 14.5 0.5 −2 21 eV 24 eV 1 13.2 13.2 2.3 0.2 s 0.03 s 0.03 s
F.B. Vicente et al. / Clinical Biochemistry 48 (2015) 679–685
reconstituted with mobile phase A and centrifuged prior to analysis. A multi-channel pipetting workstation was used to automate some steps of the sample preparation as follows. The MICROLAB® NIMBUS4 (Hamilton Company, Reno, NV, US) automated multi-channel pipetting workstation, controlled by a computer running Venus software, was used for aliquotting samples, protein precipitation solution and reconstitution solvent. The following supplies were loaded on the deck of the pipetting workstation: 1.5 mL conical polypropylene microcentrifuge tubes, borosilicate glass test tubes (13 × 100 mm) and HPLC vials. Serum samples, calibrators, quality control materials and troughs filled with the respective reagents were placed in the carriers and pedestals on the deck. 50 μL of samples, calibrators and QCs were transferred from the appropriate source tubes by the pipetting workstation into each microcentrifuge tube followed by addition of the 400 μL of precipitation solution. The tubes were capped and taken offline to be vortexed manually for 30 s followed by centrifugation at 9015 rcf for 1 min at room temperature. The microcentrifuge tubes were then returned to the NIMBUS4 deck and 100 μL from each tube was transferred to a labeled glass tube. The glass tubes were removed from the pipetting workstation and the supernatant was evaporated until dryness under nitrogen gas at room temperature for 15 min using a Reacti-Therm III Heating/ Stirring module (Thermo Scientific). Using the pipetting workstation, the residue was reconstituted with 2 mL of mobile phase A. The tubes were capped and vortexed mixed followed by centrifugation at 1430– 1500 rcf for 5 min. 1 mL of the reconstituted solution was transferred by the pipetting workstation to a labeled Maximum Recovery HPLC autosampler vial (Waters Corporation, Milford, MA). 3 μL was injected into the LC-MS/MS system. Method validation Linearity and lower limit of quantitation Linearity was confirmed over the calibration range using least square linear regression analysis and calculation of coefficient of determination (r2). The material used to verify linearity above the calibration range was prepared by adding approximately 2000 μg/mL (1996.9 μg/mL) iohexol to charcoal-stripped serum and performing serial dilution to at least 6 levels. The lower limit of quantification (LLOQ) was defined as the lowest iohexol concentration with signal-to-noise ≥ 10, when repeatedly sampled (n = 3) demonstrated a CV ≤ 20% and recovery within 95–105% of the accepted value. The lower limit of detection (LOD) was defined as the lowest concentration resulting in a signal-to-noise response ≥ 3 and a mean value 3SD greater than that measured in blank samples. Precision and accuracy Precision was evaluated following a modified CLSI EP10-A3AMD protocol, using 3 concentrations of QCs prepared in serum (13.2, 130.5 and 786.3 μg/mL) [13]. To evaluate intra-day and inter-day precision, 4 replicates of the mid level QC and 3 replicates of each low and high level QCs were extracted and analyzed in 2 batches on each of 4 days. The acceptance criterion for precision was ≤10% CV at all concentrations. There is no higher order reference method or certified reference material available for assessing trueness of iohexol measurements. Therefore, accuracy of the method was assessed by measurement of QCs at multiple, clinically relevant concentrations (as described above for precision). Intra-day and inter-day accuracy was determined as the percent relative error by comparing measured iohexol concentrations against nominal target concentrations and acceptability was defined as a mean deviation of no more than 10% for all concentrations. Method comparison A method comparison was conducted by sending blinded residual specimens (n = 80) to the Mario Negri Institute for parallel measurement of iohexol using a well-established method [4]. A
681
comparison of the GFR calculated using iohexol values from both methods was also performed. Data were analyzed by Deming regression and Bland–Altman plots. The concordance correlation coefficient for the methods was calculated. A Wilcoxon test and a t-test were performed to assess for statistical differences in the data sets. Statistical analysis was performed using MedCalc Statistical Software version 14.12.0 (MedCalc Software bvba, Ostend, Belgium; 2014). Selectivity, carryover and matrix effects Confirmation of analyte identity was determined through retention time and the monitoring of unique mass transitions. On each of 4 days, three different levels of quality control material were run in a specific order following a modified EP10-A3AMD protocol (mid, high, low, mid, mid, low, low, high, high, mid) to determine the presence of carryover [13]. Matrix effects were evaluated by a method based on that of Matuszewski et al. [14]. Iohexol and ioversol were spiked at 2 concentrations into serum from 5 different sources after extraction and resultant peak areas were compared to those obtained from the same concentration spiked into mobile phase A. Iohexol was evaluated at 653.1 and 13.1 μg/mL and ioversol at 495.8 and 9.9 μg/mL. Iohexolfree serum pools were created using the following: (a) charcoalstripped serum from a commercial source, (b) leftover hemolyzed patient serum, (c) leftover lipemic patient serum, (d) leftover serum from pediatric patients with kidney disease, and (e) leftover serum from general pediatric patients. Matrix effect was calculated as the ratio of the area of the analyte peak post-extraction to the area of the analyte peak in mobile phase A, multiplied by 100. To assess the possibility of ion suppression for iohexol and ioversol, commercial serum and serum pools from kidney and stem cell transplant pediatric patients were extracted without internal standard. A 100.8 μg/mL solution of iohexol in water was introduced into the ion source by a syringe pump at 10 μL/min through a T-connector and then the extracted sample was introduced by injection from the HPLC autosampler. Each chromatogram was inspected for the presence of signal reduction at the retention time of iohexol and ioversol. The same procedure was performed using a 101.0 μg/mL of ioversol in water. Stability The stability of iohexol in human serum was assessed using various conditions that are relevant to a typical clinical laboratory workflow. Effect of storage temperature on pre- and post-extracted samples was evaluated (ambient for 8 h, 4 °C for 24 and 36 h and − 20 °C for 1 mo). Stability after freeze–thaw cycles when stored at − 20 °C was determined for pre-extracted samples. Clinical application An iohexol clearance procedure derived from that of Schwartz and colleagues in the NIDDK CKiD study [2] was used to determine GFR in mL/min/1.73 m2. This involved administration of 5 mL of Omnipaque 300™ with serial blood samples collected over 300 min. Sample collection, processing, storage and analysis followed the procedures described above. Iohexol dose (mg I) was calculated by weighing the syringe preand post-infusion. Iohexol was measured by LC-MS/MS at a total of four sampling intervals (10, 30, 120, and 300 min post-iohexol infusion). The disappearance of iohexol was resolved into two curves using the logarithm of the concentration of iohexol as a function of time. The slow (renal) curve was determined from the concentrations obtained at 120 and 300 min. The fast curve was determined from 10 and 30 min concentrations, corrected for the contribution of the slow curve. Formulae for calculations of GFR were programmed into an Excel™ spreadsheet and validated with a two-compartment pharmacokinetic mathematical model over a wide range of compartment volumes and elimination constants. GFR was calculated as:
682
F.B. Vicente et al. / Clinical Biochemistry 48 (2015) 679–685
(a)
(b)
(c)
(d)
Fig. 1. Representative mass spectra of iohexol (a–b) and ioversol (c–d).
GFR = I / (exp A/α + exp B/β) × 1.73 / BSA, in mL/min per 1.73 m2; I: the dose of iohexol (mg I), calculated using concentration of 647 mg/mL with density 1.345 g/L; exp A: the intercept of the slow curve, and α its corresponding slope; exp B: the intercept of the fast curve, and β its corresponding slope; GFR was corrected to 1.73 m2 body surface area (BSA) by the ratio 1.73 / BSA. BSA was calculated using the Haycock formula with weight in kg and height in cm: BSA(m2) = 0.024265 × Wt0.5378 × Ht0.3964 [15].
Results and discussion LC and MS/MS method selection Mass spectra The selected precursor (m/z 821.9) and product (m/z 803.7) ions used for detection of iohexol are shown in the representative mass spectra (Figs. 1a and b). The internal standard (ioversol) was detected using precursor and product ions of m/z 807.9 and 588.7, respectively (Figs. 1c
Fig. 2. Representative chromatograms of (a) charcoal-stripped human serum, (b) iohexol (21.5 μg/mL) and (c) ioversol (39.7 μg/mL) in human serum.
F.B. Vicente et al. / Clinical Biochemistry 48 (2015) 679–685
and d). Use of ioversol as an internal standard is limited in that it is not a stable isotope-labeled form of iohexol. Therefore, its physicochemical properties are not identical to that of iohexol, creating potential for variability in signal response to ion suppression or enhancement that may be encountered during sample processing and analysis. This may explain the differences observed between iohexol and ioversol during the matrix effect study at the lowest measured concentration. The product ions selected for the multiple reaction monitoring (MRM) experiment were the most stable and abundant peaks observed during MS optimization and correspond to a water loss (m/z 18) for iohexol and a loss of m/z 219 for ioversol. Although the chosen transitions are different, are relatively non-specific and an analog is used as the internal standard, results of the method validation suggest that the impact of these design limitations on the quantitation of iohexol is minimal and is not clinically significant. Chromatograms Fig. 2 shows the representative chromatograms of (a) serum sample not containing iohexol (i.e., pre-dose), (b) serum sample collected at 300 min post-iohexol infusion iohexol (21.5 μg/mL) and (c) the internal standard, ioversol (39.7 μg/mL), added to a serum sample collected at 300 min post-iohexol infusion. Retention times were assigned as iohexol: 2.36 min (acceptable range: 2.16 to 2.56 min) and ioversol: 2.14 min (acceptable range: 1.94 to 2.34 min). Linearity and lower limit of quantitation AMR and regression analysis The iohexol assay was linear from 7.7 to 2000 μg/mL, which is sufficient for GFR determination over a wide range of kidney function (normal to late-stage CKD). The linear regression analysis of measured versus expected concentrations showed excellent correlation with r2 N 0.999. Calibrators were prepared at 10 – 1000 μg/mL, meeting the clinical requirements for this assay. The AMR is verified every 6 months. QC concentrations were selected at 15, 150 and 750 μg/mL to cover the range typically measured in our population and represent medically relevant intervals observed during the iohexol clearance protocol.
683
Method comparison A correlation of this iohexol assay and that of an international reference method, previously described [4] was conducted for 80 pediatric specimens ranging from 15 to 694 μg/mL iohexol. Fig. 3 shows the analysis of the correlation data. The concordance correlation coefficient (CCC) of the methods, a parameter that reflects both accuracy and precision, was 0.996. The method comparison of iohexol concentration showed a Deming slope of 1.019 and intercept of −1.385. A Bland–Altman plot shows good agreement between the 2 methods. A Wilcoxon test (data are not normally distributed) showed no statistical difference between the reference and candidate methods (p = 0.1321). A t-test performed on the log transformed data yielded the same results. The GFR (per 1.73 m2) values obtained using iohexol concentrations from the 2 laboratories were also compared and showed high correlation (CCC = 0.993). Fig. 4 contains the analysis of GFR values from 20 patients, ranging from 25 to 135 mL/min/1.73 m2. Regression analysis demonstrated a Deming slope of 1.030 with intercept of −2.408. The Deming regression analyses for both the iohexol concentration and the GFR showed slopes of 1 and intercepts near zero with 95% confidence intervals including 1 and 0, respectively. Therefore, neither a constant difference nor a proportional difference exists between these methods. The concordance correlation coefficient (CCC) of the methods demonstrated a high degree of agreement between the methods for iohexol concentration and GFR. This agreement was also observed in the Bland–Altman plots.
(a) R = 0.996 Deming fit: Y= 1.019X – 1.385
LOD and LLOQ with imprecision The LOD was defined by diluting the lowest level calibrator and measuring in replicate. The LOD for the iohexol assay was 3.0 μg/mL, as this was the lowest concentration at which the analyte peak was detected with a S/N N 3 and the mean value was 3SD above the value of the blank samples. The LLOQ was determined from linearity experiments as 7.7 μg/mL, since this concentration met the defined criteria with average S/N 10.2, CV 5.1% and recovery of 97%. Precision and accuracy Results from intra- and inter-day precision experiments are summarized in Table 2. Intra-day precision ranged from 1.0 to 3.8% and interday precision was 1.5–3.9% for serum-based QC samples. Table 2 also contains mean accuracy data for the multiple concentrations, showing that all measured concentrations were within 2.2–5.0% and 2.1–4.4%% of the target concentration within a run and between runs, respectively. Table 2 Within run and run-to-run precision and accuracy results for quantitation of iohexol in human serum. Sample
Conc. (μg/mL)
Intra-day (within run)
Inter-day (run-to-run)
SD
CV
% Relative error
SD
CV
% Relative error
QC1 QC2 QC3
13.2 130.5 786.3
0.52 3.27 8.22
3.8% 2.5% 1.0%
5.0% 2.3% 2.2%
0.54 2.13 12.30
3.9% 1.6% 1.5%
4.4% 2.3% 2.1%
(b) +1.96 SD
Mean
-1.96 SD
Fig. 3. Deming linear regression (a) and Bland–Altman (b) plots of serum iohexol results obtained by 2 methods. Mean percent difference = 0.1% with upper limit of agreement (+1.96 SD) = 12.1% and lower limit of agreement (−1.96 SD) = −11.9%.
684
F.B. Vicente et al. / Clinical Biochemistry 48 (2015) 679–685
(a)
Table 3 Matrix effect for iohexol and ioversol in 5 pools of human serum. Matrix effect
Iohexol (653.1 μg/mL)
Iohexol (13.1 μg/mL)
Ioversol (495.8μg/mL)
Ioversol (9.9 μg/mL)
(b)
A B C D E A B C D E A B C D E A B C D E
%ME
Mean
Std. dev.
%CV
101 100 105 102 103 90 97 98 102 101 100 97 100 105 102 91 84 82 85 83
102
1.94
1.9
97
4.14
4.3
101
2.91
2.9
85
3.50
4.1
The 5 matrix pools evaluated were: (A) charcoal-stripped serum from a commercial source, (B) hemolyzed patient pool, (C) lipemic patient pool, (D) kidney disease patient pool, and (E) pediatric patient pool. +1.96 SD
Mean -1.96 SD
Fig. 4. Deming linear regression (a) and Bland–Altman (b) plots of calculated GFR results obtained by 2 methods. Mean percent difference = −0.7% with upper limit of agreement (+1.96 SD) = 12.1% and lower limit of agreement (−1.96 SD) = −13.4%.
A difference in GFR mL/min/1.73 m2 greater than 10% was observed for one patient. In this case, the candidate method (slightly) underestimated the first 3 concentrations and overestimated that at the final timepoint. This resulted in a slower elimination phase, per the 2-compartment model and, as a consequence, a lower GFR (22.4 vs 28.3 mL/min/1.73 m2). Although the methods are highly correlated, small over- or underestimations in iohexol concentration may lead to important differences in GFR when determined using a 2-compartment kinetic model based on only 4 timepoints.
This may be attributed to the physicochemical differences between these molecules and dissimilarities in selected MRM transitions. Stability Iohexol stability in serum samples was assessed under a variety of conditions relevant to the workflow in our clinical laboratory. Serum samples stored 24 h at 4 °C, 1 mo at − 20 °C and 3–4 mo at − 20 °C after 3 freeze–thaw cycles showed b 3% difference in mean values obtained when prepared and analyzed promptly after processing. Postextraction samples were stable for 8 h ambient, after 36 h at 5 °C in the autosampler and after 1 mo when frozen at −20 °C, deviating less than 5% from the mean value obtained when analyzed promptly after extraction. These stability results compare well to that previously published for LC-MS/MS and LC-UV methods for iohexol analysis [16,17]. Clinical applicability The serum concentration–time profiles of 3 representative pediatric patients are depicted in Fig. 5, demonstrating the clinical applicability of this method. Each patient required measured GFR due to commonly acknowledged limitations in estimated GFR calculations in children.
Selectivity, carryover and matrix effects Extracted serum, free of iohexol and ioversol (evaluated across 15 lots), and LLOQ samples (n = 6) were free from coeluting peaks. Carryover was assessed by monitoring the effect of multiple high concentration injections followed by low concentration injections in a defined pattern. The level of carryover in this assay was determined to be insignificant by an independent sample t-test comparing the means of the results for the high and low sequential samples. Matrix effects were evaluated by comparing the instrument response of postextraction spiked samples to that of mobile-phase spiked samples. Table 3 contains the detailed results for the matrix effect experiments. Response in biological matrix compared to neat solutions showed acceptable recovery and CV across the 5 matrix pools ranged from 2 to 5%, indicating the relative matrix effect in the method is minimal and clinically acceptable. Recovery of ioversol at 9.9 μg/mL is slightly reduced compared to that of iohexol at 13.1 μg/mL, suggesting that the former is more susceptible to matrix effect at lower concentrations.
Fig. 5. Iohexol serum concentration–time profile for 3 pediatric patients after administration of 5 mL Omnipaque 300™. GFRs were determined to be 92.4 (◆), 49.4 (∎) and 22.4 (▲) mL/min/1.73 m2, respectively.
F.B. Vicente et al. / Clinical Biochemistry 48 (2015) 679–685
Conclusions This report describes the development and validation of a LC-MS/MS method with semi-automated sample preparation for determination of iohexol in human serum. This method accurately and precisely quantifies iohexol in 50 μL of serum. The method is highly correlated to a reference method. We have determined that the stability of this analyte and the robustness of this assay fit well within our clinical workflow. We have successfully applied this method to clinical practice as the iohexol values measured have been fit to an elimination module that enables determination of the measured GFR in pediatric patients. Acknowledgments The authors wish to thank Dr. Frederick Smith for his support and contributions to this project. References [1] Zappitelli M, Parvex P, Joseph L, et al. Derivation and validation of cystatin C-based prediction equations for GFR in children. Am J Kidney Dis: Off J Natl Kidney Found 2006;48:221–30. [2] Schwartz GJ, Furth S, Cole SR, Warady B, Munoz A. Glomerular filtration rate via plasma iohexol disappearance: pilot study for chronic kidney disease in children. Kidney Int 2006;69:2070–7. [3] Olsson B, Aulie A, Sveen K, Andrew E. Human pharmacokinetics of iohexol. A new nonionic contrast medium. Invest Radiol 1983;18:177–82. [4] Gaspari F, Perico N, Ruggenenti P, et al. Plasma clearance of nonradioactive iohexol as a measure of glomerular filtration rate. J Am Soc Nephrol: JASN 1995;6: 257–63. [5] Brown SC, O'Reilly PH. Iohexol clearance for the determination of glomerular filtration rate in clinical practice: evidence for a new gold standard. J Urol 1991;146: 675–9.
685
[6] Schwartz GJ, Work DF. Measurement and estimation of GFR in children and adolescents. Clin J Am Soc Nephrol: CJASN 2009;4:1832–43. [7] Brandstrom E, Grzegorczyk A, Jacobsson L, Friberg P, Lindahl A, Aurell M. GFR measurement with iohexol and 51Cr-EDTA. A comparison of the two favoured GFR markers in Europe. Nephrol Dial Transplant: Off Publ Eur Dial Transplant Assoc— Eur Ren Assoc 1998;13:1176–82. [8] Castagnet S, Blasco H, Vourc'h P, et al. Routine determination of GFR in renal transplant recipients by HPLC quantification of plasma iohexol concentrations and comparison with estimated GFR. J Clin Lab Anal 2012;26:376–83. [9] Cavalier E, Rozet E, Dubois N, et al. Performance of iohexol determination in serum and urine by HPLC: validation, risk and uncertainty assessment. Clin Chim Acta; Int J Clin Chem 2008;396:80–5. [10] Slack A, Tredger M, Brown N, Corcoran B, Moore K. Application of an isocratic methanol-based HPLC method for the determination of iohexol concentrations and glomerular filtration rate in patients with cirrhosis. Ann Clin Biochem 2014; 51(Pt 1):80–8. [11] Denis MC, Venne K, Lesiege D, et al. Development and evaluation of a liquid chromatography-mass spectrometry assay and its application for the assessment of renal function. J Chromatogr A 2008;1189:410–6. [12] Luis-Lima S, Gaspari F, Porrini E, et al. Measurement of glomerular filtration rate: internal and external validations of the iohexol plasma clearance technique by HPLC. Clin Chim Acta; Int J Clin Chem 2014;430:84–5. [13] Krouwer JSC GS, Tholen DW. Preliminary Evaluation of Quantitative Clinical Laboratory. Measurement Procedures; Approved Guideline. 3rd ed.; 2014. [14] Matuszewski BK, Constanzer ML, Chavez-Eng CM. Strategies for the assessment of matrix effect in quantitative bioanalytical methods based on HPLC-MS/MS. Anal Chem 2003;75:3019–30. [15] Haycock GB, Schwartz GJ, Wisotsky DH. Geometric method for measuring body surface area: a height-weight formula validated in infants, children, and adults. J Pediatr 1978;93:62–6. [16] Lee SY, Chun MR, Kim DJ, Kim JW. Determination of iohexol clearance by highperformance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS). J Chromatogr B Analyt Technol Biomed Life Sci 2006;839:124–9. [17] Soman RS, Zahir H, Akhlaghi F. Development and validation of an HPLC-UV method for determination of iohexol in human plasma. J Chromatogr B Analyt Technol Biomed Life Sci 2005;816:339–43.