Development and Evaluation of a Reference Measurement Procedure

Development and Evaluation of a Reference Measurement Procedure for the Determination of Total 3,3',5-Triiodothyronine in Human Serum Using ...
0 downloads 0 Views 58KB Size
Download PDF
Anal. Chem. 2004, 76, 5092-5096

Development and Evaluation of a Reference Measurement Procedure for the Determination of Total 3,3′,5-Triiodothyronine in Human Serum Using Isotope-Dilution Liquid Chromatography-Tandem Mass Spectrometry Susan S.-C. Tai,* David M. Bunk, Edward White, V, and Michael J. Welch

Analytical Chemistry Division, National Institute of Standards and Technology, Gaithersburg, Maryland 20899-8392

3,3′,5-Triiodothyronine (T3) is an important diagnostic marker for thyroid function. A reference measurement procedure (RMP) for total T3 in serum involving isotope dilution coupled with liquid chromatography-tandem mass spectrometry (LC/MS/MS) has been developed and critically evaluated. The method uses solid-phase extraction with mixed-mode retention mechanisms of reversed phase and ion exchange prior to reversed-phase LC/MS/ MS. In addition to a labeled T3 internal standard (T3-13C9), labeled thyroxine (T4-d5) is also added to serum samples in order to monitor the degradation of T4 to T3. The accuracy of the measurement was evaluated by a recovery study for added T3 and was supported by a comparison study with the other RMP. The recovery of the added T3 ranged from 98.9% to 99.4%. The results of this method and the other RMP agreed to within 1%. Samples of frozen serum pools were prepared and measured in three separate sets. Excellent reproducibility was obtained with within-set coefficients of variation (CVs) ranging from 0.8% to 1.6% and between-set CVs ranging from 1.9% to 2.6%. Excellent linearity was also obtained with correlation coefficients of all linear regression lines (measured intensity ratios vs mass ratios) ranging from 0.9995 to 0.9996. The detection limit at a signal-to-noise ratio of ∼3 was 1 pg of T3. The T4 degradation during sample preparation was minimized to a small percentage (no more than 3% of the T3 values) by use of antioxidants (ascorbic acid, dithiothreitol, citric acid) and can be accounted for in the T3 measurement process. This wellcharacterized LC/MS/MS method for total serum T3, which demonstrates good accuracy and precision, low susceptibility to interferences, accountability of the conversion of T4 to T3, and comparability with the other RMP, qualifies as a reference measurement procedure and can be used to provide an accuracy base to which routine methods for T3 can be compared. 3,3′,5-Triiodothyronine (T3) and thyroxine (T4) are thyroid hormones. T4 is secreted by the thyroid gland, and T3, a triiodo * To whom correspondence should be addressed. E-mail: [email protected].

5092 Analytical Chemistry, Vol. 76, No. 17, September 1, 2004

analogue of T4, is mainly produced by peripheral deiodination of T4. T4 and T3 are largely bound to protein in circulation, and their concentrations in blood (typically 50-110 µg/L for total T4, and 0.5-2.0 µg/L for total T3) are measures of thyroid function. T3, present at much lower levels in serum, has more biological activities than T4.1 The major challenges of serum T3 measurement are the low concentration of T3 normally found in serum and the degradation of serum T4 into T3, induced by light, heat, or chemical reaction during sample preparation.2-5 Because the concentration of serum T4 is ∼100-fold higher than that of T3, even slight degradation of T4 into T3 can have a strong influence on measured serum T3 values. Therefore, minimizing T4 degradation and accounting for it in the T3 measurement process is crucial. Total serum T4 and T3 measurements are routinely performed using methods based upon immunoassays, an approach with high sensitivity, but which is prone to nonspecificity for many analytes. There is a need for critically evaluated reference measurement procedures (RMPs) for total serum T4 and T3. RMPs can be used to directly assess the accuracy of routine methods or can be used to assign or verify the concentrations of controls and calibrators used in routine methods. They also provide a means for demonstrating traceability of routine methods and materials to high-order reference materials. Recently the International Standards Organization published ISO 15193 (In vitro diagnostic systemss measurement of quantities in samples of biological origins presentation of reference measurement procedures)6 that describes the requirements of an RMP for clinical diagnostic markers. The Joint Committee on Traceability in Laboratory Medicine (JCTLM) is reviewing potential RMPs and compiling a list of those that meet the requirements of ISO 15193. At present, there is one RMP for total serum T3 approved for this list. This (1) Tietz, N. W. In Tietz Textbook of Clinical Chemistry, 2nd ed.; Burtis, C. A., Ashwood, E. R. Eds.; Saunders: Philadelphia, 1994; pp 1698-1717. (2) Thienpont, L. M.; Fierens, C.; De Leenheer, A. P.; Przywara, L. Rapid Commun. Mass Spectrom. 1999, 13, 1924-1931. (3) Ramsden, D. B. In Mass Spectrometry; Lawson, A. M., Ed.; De Gruyter: Berlin, 1989; Chapter 12. (4) Kazemifard, A. G.; Moore, D. E.; Aghazadeh, A. J. Pharm. Biomed. Anal. 2001, 25, 697-711. (5) Koya, S.; Matsuura, K.; Kubota, E. Yakugaku Zasshi. 1982, 102, 923-934. (6) ISO 15193. 2003; www.iso.ch/iso/en/CatalogueListPage. CatalogueList. 10.1021/ac049516h Not subject to U.S. Copyright. Publ. 2004 Am. Chem. Soc.

Published on Web 08/05/2004

Figure 1. Structures of T4, T3, and their isotopically labeled compounds: (A) T4; (B) T3; (C) T4-d5; (D) T3-13C9; /, 13C labeling.

method, developed at the University of Ghent, involves two-step liquid/liquid extraction to isolate T3 from the serum and then liquid chromatography (LC) fractionation to further purify T3 from the serum extract prior to gas chromatography/mass spectrometry (GC/MS) or LC/MS/MS analysis.2 Recently, the National Institute of Standards and Technology (NIST) developed an RMP for total serum T4 using LC/MS.7 We report here a method for total serum T3 that also meets the requirements of an RMP as defined by ISO 15193. This method involves an isotope-dilution LC/MS/MS method for serum T3 using a solid-phase extraction (SPE) procedure with mixed-mode retention mechanisms of reversed-phase and cation exchange to isolate T3 from serum.7 Due to the specificity and selectivity of the SPE extraction, the extract produced was free from interference at ions monitored for T3 by LC/MS/MS. The T4 degradation in this method was minimized to a small percentage of the T3 values and can be accounted for in the T3 measurement process by monitoring the conversion of T4-d5 to T3-d5. The accuracy of the measurement was evaluated by a recovery study for added T3 and was supported by a comparison study with the other RMP. The reproducibility of this method was evaluated by repeated measurements of three frozen serum pools. EXPERIMENTAL SECTION Samples and standards containing T3 were handled with minimal exposure to light (incandescent light at reduced intensity was used) and were stored at refrigerator temperatures (4 °C). Materials. The T3 reference compound used for this work was obtained from Aldrich (Milwaukee, WI). The impurities in this T3 material were evaluated at NIST by LC/MS, and moisture content was measured by Karl Fischer titration. Appropriate corrections were made for impurities. A carbon-13-labeled internal standard, T3-13C9 (β-[4-(3-iodo-4-hydroxyphenoxy)-3,5-diiodo(1,2,3,4,5,6-13C6) phenyl]-L-13C3-alanine), was a gift from Professor Linda Thienpont (University of Ghent). T4-d5 was synthesized at NIST.7 Structures of T4, T3, T4-d5, and T3-13C9 are shown in Figure 1. Bond-Elut Certify SPE cartridges (LRC, 10 mL, 300 mg) were obtained from Varian (Harbor City, CA). A Zorbax Eclipse XDBC18 column (15 cm × 2.1 mm i.d., 5-µm particle diameter) was obtained from Agilent Technologies (Palo Alto, CA). Solvents used for LC/MS/MS measurements were HPLC grade, and all other chemicals were reagent grade. Samples of College of American Pathologists (CAP) frozen serum pools were provided by the CAP (7) Tai, S. S.; Sniegoski, L. T.; Welch, M. J. Clin. Chem. 2002, 48, 637-642.

from their 1999 survey. Samples of Standard Reference Material (SRM) 1951b, Lipids in Frozen Human Serum, Level 1 were obtained from the Standard Reference Materials Program at NIST. Samples of lyophilized human serum were obtained from Hematronics, Inc. (Benicia, CA). Samples of Seracon II, defibrinated, dialyzed, delipidated, and charcoal-stripped frozen serum were obtained from Seralogicals Corp. (Norcross, GA). These serum materials were used for method development for T3, and none of them have certified T3 values. Preparation of Calibrators. Two independently weighed standard stock solutions of T3 were prepared. Approximately 2.5 mg of the T3 reference compound for each stock solution was accurately weighed on an analytical balance (Mettler Toledo model ME22 with a readability of 1 µg) and dissolved in 20 mL of methanol containing several drops of 1 mol/L hydrochloric acid in a 100-mL volumetric flask. After complete dissolution of the substance, the flask was filled to volume with methanol. Immediately after preparation of the first standard stock solutions, a second stock solution was prepared from each first stock by diluting 5.0 mL in 0.05 mol/L sodium phosphate dibasic (Na2HPO4) buffer (pH 11.6) in a 100-mL volumetric flask containing 5 mg of diiodotyrosine (10 mL of 0.5 g/L in methanol) as a protective carrier substance. A final working solution was prepared from each second stock by diluting 4.0 mL in 0.05 mol/L sodium phosphate dibasic buffer (pH 11.6) in a 100-mL volumetric flask containing 5 mg of diiodotyrosine. The final concentrations of T3 in the working standard solutions were ∼0.05 ng/µL. A solution of isotopically labeled internal standard, T3-13C9, at a concentration of ∼0.05 ng/µL was prepared in the same way as the unlabeled T3. The concentrations of the two working standard solutions of T3 were cross-checked against each other by LC/MS/MS and were within 0.5% of each other. Two aliquots from each working standard solution of T3 were spiked with T3-13C9, yielding four standards with the mass ratios of unlabeled to labeled compound ranging from 0.6 to 1.5. The mixtures were diluted with a solvent consisting of 10 mL/L acetic acid in water/methanol (48:52 by volume) to a concentration of ∼0.01 ng/µL T3 for LC/MS/MS analysis. Aliquotting of the stock solution was carried out using a Rainin EDP-2 motorized pipet. The volumes dispensed were calibrated by weighing of water. Sample Preparation. Throughout the sample preparation, the sample containers were wrapped with aluminum foil to minimize the exposure to light. CAP frozen serum pools at two T3 concentrations and NIST SRM 1951b at one T3 concentration (level 1) were used for this study. Samples were prepared in three different sets (each set on a different day), each set consisting of two vials (10 mL per vial) each of the two concentrations of CAP materials (1.4 and 2.4 µg/L for concentrations 1 and 2, respectively) and seven vials (1 mL per vial) of SRM 1951b (0.9 µg/L). For the CAP materials, duplicate 3.0-mL aliquots were taken from each of two vials for sample workup. For SRM 1951b, the contents from seven vials were combined and duplicate aliquots were taken. Each aliquot was placed into a 50-mL Teflon centrifuge tube containing 5 µg of diiodotyrosine (50 µL of 0.1 g/L in methanol), 3 mg each of the three antioxidants (60 µL each of ascorbic acid and citric acid at concentrations of 50 g/L in water and 120 µL of dithiothreitol at a concentration of 25 g/L in water), and an appropriate amount of T3-13C9 was added to give an approximate Analytical Chemistry, Vol. 76, No. 17, September 1, 2004

5093

1:1 ratio of analyte to internal standard. Additionally, an aliquot containing 225 ng of T4-d5 in 0.05 mol/L sodium phosphate dibasic buffer (pH 11.6) was added to each sample to monitor the possible conversion of T4 to T3 during sample preparation. To each sample, 180 mg of sodium chloride was added, and the mixture was equilibrated at room temperature for 1 h and then deproteinized with 6 mL of acetone in an ice bath for 30 min.2,8After centrifugation of the tube at 2000g for 15 min, the supernatant was transferred to another 50-mL Teflon tube containing 5 µg of diiodotyrosine and 3 mg each of the three antioxidants and evaporated to ∼3 mL using nitrogen at 30 °C. T3 was then isolated from the serum matrix using Bond-Elut Certify SPE.7 To this concentrated supernatant, 7 mL of 0.1 mol/L potassium acetate buffer (pH 4.0) was added, and each sample was loaded at a rate of 3-4 mL/min onto a cartridge previously conditioned by wetting sequentially with 6 mL of methylene chloride/2propanol (75:25 by volume) and 6 mL of methanol, followed by 6 mL of 0.1 mol/L potassium acetate buffer (pH 4.0). The cartridge was then washed sequentially with 10 mL of water, 5 mL of 0.1 mol/L hydrochloric acid, and 20 mL of methanol, followed by 10 mL of methylene chloride/2-propanol (75:25 by volume). The T3 was eluted from the cartridge with 3.5 mL of methylene chloride/ 2-propanol/ammonium hydroxide (70:26.5:3.5 by volume) to a 10mL Teflon tube containing 5 µg of diiodotyrosine and 3 mg each of the three antioxidants. The eluate was dried under nitrogen at 30 °C and reconstituted with 10 mL/L acetic acid in water/ methanol (48:52 by volume) to a concentration of ∼0.01 ng/µL of T3. Stripped serum samples spiked with known amounts of T3 were used to evaluate the recovery of T3 from serum with this extraction method. Two groups of samples were prepared. For the first group, stripped serum was spiked with known amount of T3 before extraction and T3-13C9 after extraction. For the second group, both T3 and T3-13C9 were spiked before extraction. Recovery of T3 from serum was calculated from the comparison of the intensity ratios of the two groups. Equilibration. Commercially available lyophilized human serum samples (Hematronics level 2) were used for this study. Vials of reconstituted serum samples were combined, and six 3.0mL aliquots were taken for the equilibration study. An appropriate amount of T3-13C9 was added to each aliquot, and duplicate aliquots for each of the three time intervals (1, 2, and 3 h) were equilibrated at room temperature. The samples were processed as described above for LC/MS/MS measurement. Recovery of the Added T3. Commercially available lyophilized human serum samples (Hematronics level 1) were used for this study. Vials of reconstituted serum samples were combined and 12 3.0-mL aliquots were taken for a study of the accuracy of the method. A known amount of unlabeled T3 was added to 9 of the 12 aliquots, 3 each with 0.747, 1.494, and 2.241 µg/L T3. No T3 was added to the other three aliquots. An appropriate amount of T3-13C9 was added to each aliquot, and the aliquots were processed as described above for LC/MS/MS measurement. LC/MS/MS Analysis. Analysis was performed on a Quattro Ultima (Waters/Micromass, Milford, MA) equipped with a model 2795 LC (Waters). A Zorbax Eclipse XDB-C18 column was used (8) De Brabandere, V. I.; Hou, P.; Sto ¨ckl, D.; Thienpont, L. M.; De Leenheer, A. P. Rapid Commun. Mass Spectrom. 1998, 12, 1099-1103.

5094

Analytical Chemistry, Vol. 76, No. 17, September 1, 2004

for the analysis. Aliquots of standards or sample extracts (∼0.4 ng of T3) were analyzed by LC with a gradient mobile phase consisting of 1 mL/L acetic acid in water/methanol. The gradient was initially set at water/methanol (48:52 by volume) for 15 min, ramped to 100% methanol at 15.5 min, and held for 10 min to wash the column. The flow rate was 0.25 mL/min. The column temperature was set at 30 °C. Electrospray ionization in the positive ion mode and multiple reaction monitoring (MRM) mode were used for LC/MS/MS. The transitions of (M + H)+ f [(M + H)+ - HCOOH] at m/z 652 f 606 and m/z 661 f 614 were monitored for T3 and T3-13C9, respectively.2 The collision gas was argon at a collision cell pressure of 0.25 Pa (2.5 × 10-3 mbar), and the collision energy was 18 eV. The dwell times were 0.3 s for MRM. Measurement Protocol. The following measurement protocol was used for LC/MS/MS analysis. For each set of samples, a single analysis of each of the four standards was run first. Subsequently, duplicate analyses of each sample were run. Finally, the four standards were run again in reverse order. By combining the data of standards run before and after the samples, a composite linear regression was calculated, which was used to convert the measured intensity ratios of analyte to mass ratios. The mass ratios were then used along with the amounts of the internal standard added to calculate analyte concentrations. Monitoring T4 to T3 Conversion during Sample Preparation. Antioxidants (dithiothreitol, ascorbic acid, citric acid) were added to each sample at various steps to minimize the conversion of T4 into T3 during sample preparation. A given amount of T4-d5 was added to each sample to monitor the possible conversion of T4 to T3. The transitions of (M + H)+ f [(M + H)+ - HCOOH] at m/z 783 f 737 and m/z 657 f 611 were monitored for T4-d5 and T3- d5, respectively. The LC/MS/MS conditions described above were used except that the mobile phase of water/methanol (45:55 by volume) was used instead of water/methanol (48:52 by volume). RESULTS AND DISCUSSION Evaluation of Critical Parameters. Reference measurement procedures must be thoroughly tested for sources of bias. The following sections discuss the evaluation of critical parameters that potentially could bias the results. (1) Extraction. The solid-phase extraction with mixed-mode retention mechanisms of reversed-phase and cation ion exchange produced a clean extract with no interference detected at ions monitored for T3 by LC/MS/MS. T3 was completely separated from T4 and biologically inactive 3,3′,5′-triiodothyronine (reverse T3) with retention times of 10, 15, and 18 min for T3, reverse T3, and T4, respectively, under the LC conditions for T3 measurement. The overall recovery of T3 from the serum with this extraction method was determined to be approximately 65%. With complete equilibration between the labeled and unlabeled forms, as discussed below, relative recoveries of the native T3 and the T3-13C9 internal standard should be equal. Therefore, the absolute recoveries are not critical, since it is the ratio of unlabeled to labeled T3 that is measured. (2) Equilibration. Complete equilibration of T3 in serum with spiked internal standard is necessary for accurate measurement. The time required for T3 to be liberated from its protein binding and then equilibrated with the internal standard, T3-13C9, was

Table 1. Recovery of Added T3a

added 0 0.747 1.494 2.241 a

concentration (µg/L) expected dectected nab 1.803 2.550 3.297

1.056 1.783 2.534 3.278

Table 2. Reproducibility of LC/MS/MS Measurements of T3 in Serum (µg/L) recovery (%)

CV, % (n ) 6)

nab 98.9 99.4 99.4

2.3 1.4 1.7 1.5

Based on known additions to a serum pool. b na, not applicable.

investigated. It was found that the equilibration was complete within 1 h at room temperature and the ratio of T3 to T3-13C9 was unchanged for up to 3 h. Therefore, 1 h was the time chosen for equilibration. (3) Recovery of the Added T3. The recoveries of the added T3 are listed in Table 1. The average concentration of endogenous T3 was 1.056 µg/L. The amount of T3 recovered and added was in very good agreement for all three concentrations with mean recoveries of 98.9%, 99.4%, and 99.4% for 0.747, 1.494, and 2.241 µg/L T3 added, respectively. (4) Conversion of T4 to T3 during Sample Preparation. In preliminary experiments, the measured T3 values were irreproducible when antioxidants were not used. An antioxidant mixture (ascorbic acid, dithiothreitol, citric acid) added to samples at various steps minimized the conversion of T4 to T3 during sample preparation and produced reproducible T3 results. The ascorbic acid and dithiothreitol in the mixture were intended to suppress free radical oxidation and prevent oxidation by lipid hydroperoxides likely present in processed sera. The citric acid, a metal complexing agent, was intended to suppress metal-assisted oxidation. The percent conversion of T4 to T3 was calculated for each sample. For each concentration in a given set, the mean percent conversion along with the concentrations of T4 (previously measured using the method described in ref 7) and T3 was used to calculate the correction factor caused by the conversion of T4 to T3. They were determined to be 0.91%-1.46% for CAP serum (concentration 1), 1.77%-2.01% for CAP serum (concentration 2), and 2.76%-3.06% for SRM 1951b. For each concentration in a given set, the true T3 values were determined by applying a correction factor to the measured T3 values. Measurement of Frozen Serum Materials. The LC/MS/ MS method for the determination of T3 was applied to three frozen serum pools. Samples were prepared and analyzed in three different sets, each set consisting of two concentrations of CAP materials and one concentration of SRM 1951b. The measured T3 values were corrected by the correction factor caused by the conversion of T4 to T3 as mentioned above and results are shown in Table 2. Excellent reproducibility was obtained for all three concentrations with within-set CVs ranging from 0.8% to 1.6% and between-set CVs ranging from 1.9% to 2.6%. A linear regression line (measured intensity ratios vs mass ratios) was generated for each set of samples. Excellent linearity was obtained with correlation coefficients (R) of all linear regression lines ranging from 0.9995 to 0.9996. The detection limit at a signal-to-noise ratio of ∼3 was 1 pg. Selected ion chromatograms for T3 and T3-13C9 are shown in Figure 2. Statistical Analysis of Results. A summary of the statistical analysis for the results is shown in Table 3. An analysis of variance

overall serum CAP conc 1 CAP conc 2 SRM 1951b level 1

set

mean

SDa

CV (%)

1 2 3 1 2 3 1 2 3

1.423 1.379 1.377 2.453 2.373 2.339 0.940 0.900 0.897

0.022 0.015 0.018 0.020 0.027 0.019 0.009 0.014 0.013

1.6 1.1 1.3 0.8 1.2 0.8 0.9 1.5 1.4

mean

SDb

CV (%)

1.393

0.026

1.9

2.389

0.059

2.5

0.912

0.024

2.6

a Standard deviation (SD) of a single measurement within a set. b SD of the mean for that level.

Figure 2. Selected ion chromatograms by LC/MS/MS for T3 and T3-13C9 from a serum sample.

found that set-set differences were significant for all three materials. Thus, the three set means for each material were averaged. The standard deviation of the mean for each material was calculated by dividing the standard deviation of the set means for that level by the square root of 3. To calculate the standard uncertainty uc, the standard deviation of the mean from the measurements was combined quadratically with the uncertainty in the conversion of T4 to T3 in the sample preparation, a measured value, and the type B factors, which were uncertainties related to the use of volumetric measurements, estimated to be 1%, and uncertainty in the purity of the reference compound, estimated to be 0.5%. The uncertainty in the T4 to T3 conversion for each material was taken as the standard deviation in the percentage conversion measured for each set. The WelchSatterthwaite approximation was used to calculate the effective degrees of freedom. Since the measurement uncertainty is the largest single contributor to the overall uncertainty, the calculated degrees of freedom are relatively small, resulting in the coverage factor, k, values being greater than 2.0. The expanded uncertainties, U, corresponding to 95% confidence intervals, were calculated by multiplying uc by k.9 (9) Taylor, B. N.; Kuyatt, C. E. NIST Technical Note 1297, 1994.

Analytical Chemistry, Vol. 76, No. 17, September 1, 2004

5095

Table 3. Estimation of Expanded Uncertainties for LC/MS/MS Measurements of T3 in Serum CAP conc 1

CAP conc 2

SRM 1951b level 1

set 1 mean, µg/L set 2 mean, µg/L set 3 mean, µg/L

1.423 1.379 1.377

2.453 2.373 2.339

0.940 0.900 0.897

mean, µg/L SD, µg/L SD of mean, µg/L

1.393 0.026 0.0150

2.389 0.059 0.0338

0.912 0.024 0.0137

1% uncertainty of vol error, µg/L uncertainty of T4 to T3 convrsn, µg/L 0.5% uncertainty of purity of ref compd, µg/L

0.0139 0.0040 0.0070

0.0239 0.0031 0.0119

0.0091 0.0014 0.0046

combined SD uncertainty (uc), µg/L effective degrees of freedom k factor

0.0220 9.2 2.26

0.0432 5.3 2.57

0.0171 4.9 2.78

expanded uncertainty (U), µg/L relative expanded uncertainty,a %

0.050 3.6

0.111 4.7

0.048 5.2

a

Uncertainty of 95% confidence interval.

Comparison with Other High-Order RMP. A comparison study between RMPs of the University of Ghent and our method on two lyophilized human serum pools found that the two methods agreed to within 1%, demonstrating comparability of the two methods. Further details of the study will be published elsewhere. CONCLUSIONS This well-characterized LC/MS/MS method for total serum T3, which demonstrates good accuracy and precision, low

5096

Analytical Chemistry, Vol. 76, No. 17, September 1, 2004

susceptibility to interferences, accountability of a small amount of conversion of T4 to T3, and comparability with the other RMP, qualifies as a reference measurement procedure as defined by ISO 15193. Use of this reference procedure can provide an accuracy base to which routine methods for T3 can be compared. NIST is planning to use this method to certify the concentration of T3 in a new SRM for hormones in frozen serum pools. This SRM will be used by laboratories to test the accuracy of their methods, calibrators, and controls and to establish traceability of their measurement results to this NIST reference measurement procedure.

ACKNOWLEDGMENT We gratefully acknowledge Professor Linda Thienpont for a gift of the internal standard, T3-13C9, in conjunction with the European project G6RD-CT2004-00587. We thank Sam Margolis for the Karl Fischer analysis. Certain commercial equipment, instruments, and materials are identified in this paper to adequately specify the experimental procedure. Such identification does not imply recommendation or endorsement by NIST nor does it imply that the equipment, instruments, or materials are necessarily the best available for the purpose.

Received for review March 28, 2004. Accepted June 21, 2004. AC049516H