Clinical Chemistry 48:4 637– 642 (2002)
Endocrinology and Metabolism
Candidate Reference Method for Total Thyroxine in Human Serum: Use of Isotope-Dilution Liquid Chromatography–Mass Spectrometry with Electrospray Ionization Susan S-C. Tai,* Lorna T. Sniegoski, and Michael J. Welch 110, and 168 g/L). The detection limits (at a signal-tonoise ratio of ⬃3 to 5) were 30 and 20 pg for positive and negative ions, respectively. The results from the LC/MSESI method were within 1 SD of the composite means from many routine clinical methods, although it appears that the clinical method means may be biased high by 4 –5 g/L across the concentrations. Some routine clinical methods may be biased by up to 20% at low concentrations. Conclusions: This well-characterized LC/MS-ESI method for total serum thyroxine with a theoretically sound approach, demonstrated good accuracy and precision, and low susceptibility to interferences qualifies as a candidate reference method. Use of this reference method as an accuracy base may reduce the apparent biases in routine methods along with the high interlaboratory imprecision.
Background: There is a need for a critically evaluated reference method for thyroxine to provide an accuracy base to which routine methods can be traceable. We describe a candidate reference method involving isotope-dilution coupled with liquid chromatography/ mass spectrometry. Methods: An isotopically labeled internal standard, thyroxine-d5, was added to serum, followed by equilibration, protein precipitation, and ethyl acetate and solid-phase extractions to prepare samples for liquid chromatography–mass spectrometry electrospray ionization (LC/MS-ESI) analysis. For separation, a Zorbax Eclipse XDB-C18 column was used with a mobile phase consisting of 1 mL/L acetic acid in acetonitrile–water (32:68 by volume) for positive ions and a Zorbax ExtendC18 column with a mobile phase consisting of 2 mL/L ammonium hydroxide in methanol–water (32:68 by volume) for negative ions. [M ⴙ H]ⴙ ions at m/z 778 and 783 for thyroxine and its labeled internal standard were monitored for positive ions and [M ⴚ H]ⴚ ions at m/z 776 and 781 for negative ions. Samples of frozen serum pools were prepared and measured in three separate sets. Results: Within-set CVs were 0.2–1.0%. The correlation coefficients of all linear regression lines (measured intensity ratios vs mass ratios) were 0.999 –1.000. Positive- and negative-ion measurements agreed with a mean difference of 0.45% at three concentrations (50,
© 2002 American Association for Clinical Chemistry
Thyroxine (T4) is secreted by the thyroid gland and is largely bound to protein in the circulation. Its concentration in blood, typically 50 –110 g/L, is a measure of thyroid function. NIST provides methods and reference materials to support accuracy and traceability in clinical chemistry. NIST has developed isotope-dilution (ID)1 mass spectrometric definitive methods for 12 clinical analyte, including cholesterol, glucose, and calcium, and has used these methods to certify concentrations in serum-based Standard Reference Materials (1 ). This report describes the development of a method for total thyroxine in serum using liquid chromatography–mass spectrometry with electrospray ionization (LC/MS-ESI).
Analytical Chemistry Division, National Institute of Standards and Technology, Gaithersburg, MD 20899-8392. *Author for correspondence. Fax 301-977-0685; e-mail
[email protected]. Certain commercial equipment, instruments, and materials are identified in this article 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 September 4, 2001; accepted January 4, 2002.
1 Nonstandard abbreviations: ID, isotope dilution; LC/MS-ESI, liquid chromatography–mass spectrometry with electrospray ionization; SPE, solidphase extraction; and CAP, College of American Pathologists.
637
638
Tai et al: LC/MS Candidate Reference Method for Thyroxine
Various ID gas chromatography–mass spectrometry methods have been published for the determination of T4 in serum (2–5 ). These methods involve liquid-liquid extraction or purification by multiple LC columns, and a two-step derivatization of thyroxine before gas chromatographic–mass spectrometric analysis. The resulting derivatives have very high relative molecular weights (⬃1000) and are difficult to measure by gas chromatography. Recent developments of highly sensitive LC/MS instrumentation with robust interfaces such as electrospray or atmospheric pressure chemical ionization make it possible to develop highly reproducible, selective, and sensitive LC/MS methods for low-concentration analytes (6, 7 ). We developed a solid-phase extraction (SPE) procedure that uses mixed-mode retention mechanisms of reversed-phase and ion-exchange to isolate thyroxine from serum. Because of the specificity and selectivity of the SPE extraction, the extract produced was clean with no interference at ions monitored by LC/MS. The twostep liquid-liquid extraction used by the earlier LC-tandem MS proposed reference method (6 ) was not as selective as the SPE used here, and thus tandem MS has been needed to provide interference-free ion chromatograms. In addition to positive-ion measurement for thyroxine, we also developed a negative-ion measurement method with a LC separation using a column stable at the high pH necessary for generation of sufficient negative ions for good sensitivity. This LC/MS-ESI method was applied to the determination of thyroxine in samples of frozen serum pools. The method was evaluated by measuring thyroxine in both positive and negative modes. The LC/MS-ESI method was compared with routine clinical methods.
Materials and Methods materials The l-thyroxine reference compound used for this work was obtained from Aldrich. The impurities in this thyroxine material were evaluated at NIST by LC/MS-ESI, LC with ultraviolet detection, and direct-probe MS and moisture content by Karl Fischer titration. Appropriate corrections were made for impurities and moisture content. A stable deuterium-labeled internal standard, l-thyroxine-d5 {-[3,5-diiodo-4-hydroxyphenoxy)-3,5-diiodo-(2,5d2-phenyl)-]-l-alanine-2,3,3-d3}, was synthesized at NIST by the oxidative coupling of 4-hydroxy-3,5-diiodophenylpyruvic acid and diiodo-tyrosine-d5 as described by Ramsden and Farmer (8 ) and Nishinaga et al. (9 ). 125Ilabeled l-thyroxine was obtained from DuPont NEN. Bond-Elut CertifyTM SPE cartridges (LRC; 10 mL; 300 mg) were obtained from Varian. A Zorbax Eclipse XDB-C18 column [15 cm ⫻ 2.1 mm (i.d.); 5-m particle diameter] and a Zorbax Extend-C18 column [15 cm ⫻ 2.1 mm (i.d.); 5-m particle diameter] were obtained from Agilent Technologies. Solvents used for LC/MS measurements were HPLC grade, and all other chemicals were reagent grade.
preparation of calibrators Two independently weighed stock solutions of thyroxine were prepared. Approximately 5 mg of the thyroxine reference compound for each stock solution was accurately weighed on an analytical balance (Mettler ME22; a readability of 1 g) and dissolved in 20 mL of methanol containing several drops of 1 mol/L hydrochloric acid in a 100-mL volumetric flask. The balance was calibrated and accurate to 1 g. After complete dissolution of the thyroxine, the flask was filled to the mark with methanol. Immediately after preparation of the stock solutions, a working solution was prepared from each stock by diluting 5.0 mL in 0.05 mol/L Na2HPO4 buffer (pH 11.6) in a 100-mL volumetric flask containing 5 mg of diiodotyrosine as a protective carrier substance. The final concentrations of thyroxine in the working solutions were ⬃2.5 mg/L. A solution of isotopically labeled internal standard, thyroxine-d5, at a concentration of ⬃2.5 mg/L was prepared in the same way as the unlabeled thyroxine. The two working solutions of thyroxine were cross-checked against each other by LC/MS and were within 0.5% of each other. Thyroxine-d5 was added to two aliquots of each working solution of thyroxine, yielding four calibrators with mass ratios of unlabeled to labeled compound ranging from 0.68 to 1.32. The mixtures were diluted with a solvent consisting of 10 mL/L acetic acid in acetonitrile– water (32:68 by volume) to a thyroxine concentration ⬃0.25 mg/L for LC/MS analysis. The stock solution was aliquoted with a Rainin EDP-2 motorized pipette. The volumes were calibrated by weighing.
sample preparation Samples of frozen serum pools were prepared in three different sets (each set on a different day), each set consisting of two vials each of three concentrations: 50, 110, and 168 g/L for concentrations 1, 2, and 3, respectively. Duplicate 3.0-mL aliquots were taken from each vial for sample work-ups. Each aliquot was placed in a 50-mL Teflon centrifuge tube containing 5 g of diiodotyrosine as a protective carrier substance, and an appropriate amount of thyroxine-d5 was added to give an ⬃1:1 ratio of analyte to internal standard. Each sample was acidified to pH 2 with 1 mol/L hydrochloric acid and equilibrated at 37 °C for 2 h. After equilibration, the mixture was deproteinized with 5 mL of 150 g/L trichloroacetic acid in an ice bath for 30 min. The thyroxine was then extracted from the deproteinized sample with 5 mL of ethyl acetate (10 ). Each tube was shaken vigorously for 10 min on a mechanical shaker to allow complete mixing. After centrifugation of the tube for 10 min at 2000g, the upper ethyl acetate layer was transferred to another 50-mL Teflon tube containing 5 g of diiodotyrosine. Ethyl acetate extraction was repeated twice more with 4 mL and 3 mL of ethyl acetate, respectively. The combined ethyl acetate extract was reduced to ⬃0.5 mL under nitrogen at 40 °C. The thyroxine was then
639
Clinical Chemistry 48, No. 4, 2002
further isolated from the serum matrix by Bond-Elut Certify SPE. To this concentrated extract was added 1 mL of methanol and 4 mL of 0.01 mol/L hydrochloric acid, and the pH was adjusted to 1.5 ⫾ 0.5 with 3 mol/L potassium hydroxide. Each sample was loaded at a rate of 3– 4 mL/min on a cartridge previously conditioned by wetting with 6 mL of methylene chloride–2-propanol (75:25 by volume), 6 mL of methanol, and 6 mL of 0.01 mol/L hydrochloric acid. The cartridge was then washed sequentially with 10 mL of water, 10 mL of 0.1 mol/L hydrochloric acid, 20 mL of methanol, and 10 mL of methylene chloride–2-propanol (75:25 by volume). The thyroxine was eluted from the cartridge with 5 mL of methylene chloride–2-propanol–ammonium hydroxide (70:26.5:3.5 by volume). The eluate was dried under nitrogen at 40 °C and reconstituted with a solvent consisting of 10 mL/L acetic acid in acetonitrile–water (32:68 by volume) to a final concentration of 0.25 mg/L thyroxine for LC/MS analysis. 125 I-labeled l-thyroxine was used as a tracer to evaluate the recovery of thyroxine from serum with this extraction method.
equilibration Vials of frozen serum samples were combined, and six 3.0-mL aliquots were taken for the equilibration study. Each aliquot was equilibrated at 37 °C for various times (0.5, 1, 1.5, 2, 3, and 4 h). The samples were processed as described above for positive-ion measurement.
recovery of added thyroxine Vials of concentration 1 frozen serum samples were combined, and twelve 3.0-mL aliquots were taken for a test for the accuracy of the method. Unlabeled thyroxine was added to 9 of the 12 aliquots, 3 each with 56.7, 85.1, and 113.4 g/L thyroxine. No thyroxine was added to the other three aliquots. A given amount of thyroxine-d5 was added to each aliquot, and the aliquots were then processed as described above for positive-ion measurement.
lc/ms-esi analysis Analysis was performed on a Hewlett Packard 1100 Series LC/MSD from Agilent Technologies with an ESI interface. An autosampler was used. The flow rate was 0.3 mL/min. The nitrogen drying gas temperature was set at 350 °C, and the flow was 12 L/min. The nebulizer pressure was set at 172 kPa (25 psi), Vcap at 3500 V, and fragmentor at 100 V. [M ⫹ H]⫹ ions at m/z 778 and 783 for thyroxine and thyroxine-d5 were monitored for positive ions. A Zorbax Eclipse XDB-C18 column with double endcapping was used to provide good peak shape for thyroxine. Aliquots (20 L) of calibrators or sample extracts (⬃5 ng of thyroxine) were analyzed by LC with an isocratic mobile phase consisting of 1 mL/L acetic acid in acetonitrile– water (32:68 by volume). [M ⫺ H]⫺ ions at m/z 776 and 781 for thyroxine and
thyroxine-d5 were monitored for negative ions. A LC column with stability at high pH is required for negativeion measurements. Such a column, a Zorbax Extend-C18 column with high pH stability (up to pH 11.5), was used. The calibrators and sample extracts used for positive-ion measurements were adjusted to pH 10 with 5 mol/L ammonium hydroxide (1 mL of NH4OH/5 mL of calibrator or sample) for negative-ion measurements. Only one aliquot from each vial was selected. The pH was adjusted immediately before negative-ion LC/MS analysis. Aliquots (20 L) of calibrators or sample extracts (⬃4 ng of thyroxine) were analyzed by LC with an isocratic mobile phase consisting of 2 mL/L ammonium hydroxide in methanol–water (32:68 by volume).
measurement protocol The following measurement protocol was used for LC/MS analysis. For positive-ion measurements, four samples from one concentration of a given set were analyzed as a group. For negative-ion measurements, six samples from all three concentrations of a given set were analyzed as a group. For each group of samples, a single analysis of each of the four calibrators was run first. Subsequently, duplicate analyses of each sample were run. Finally, the four calibrators were run again in reverse order. By combining the data of calibrators run before and after the samples, we calculated a composite linear regression, 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.
Results and Discussion extraction The combination of deproteinization and liquid-liquid and solid-phase extractions produced a clean extract with no interference detected at the ions monitored. 125I-labeled l-thyroxine was used as a tracer to evaluate the extraction procedures. The overall recovery of thyroxine from the serum with this extraction method was ⬃74%. As long as equilibration is complete, as discussed below, relative recoveries of the native thyroxine and thyroxine-d5 internal standard should be equal. Therefore, the absolute recoveries are not critical because it is the ratio of unlabeled to labeled thyroxine that is measured. Experiments were performed to support a lack of ion suppression. The extracts from two kinds of biologic matrices, human serum albumin (Sigma No. A3782) and bovine serum albumin (Standard Reference Material 927c), were prepared, and known amounts of thyroxine were added. The response of thyroxine in the extracts was compared with the matching concentration of thyroxine in the absence of extracts. No differences in responses were found (within 0.3%), demonstrating that there is no evidence of ion suppression.
640
Tai et al: LC/MS Candidate Reference Method for Thyroxine
Table 1. Recovery of added thyroxine.
Table 3. Negative-ion LC/MS-ESI measurements of thyroxine in serum.
Thyroxine, g/L Added
Detected
Added amount recovered
Recovery, %
CV, % (n ⴝ 6)
0 56.7 85.1 113.4
49.4 105.8 135.0 162.2
NAa 56.4 85.6 112.8
NA 99.5 100.6 99.5
0.6 0.5 0.2 0.6
a
NA, not applicable.
equilibration Complete equilibration of thyroxine in serum with the added internal standard is necessary for accurate measurement. The time required for thyroxine to be liberated from its protein binding and then equilibrated with the internal standard, thyroxine-d5, was investigated. It was found that the equilibration was complete within 30 min at 37 °C, and the ratio of thyroxine to thyroxine-d5 was unchanged for up to 4 h. Two hours was the time chosen for equilibration.
recovery of added thyroxine The recovery data for the added thyroxine are listed in Table 1. The mean concentration of endogenous thyroxine was 49.4 g/L. The amounts of thyroxine recovered and added were in very good agreement for all three concentrations, with mean recoveries of 99.5%, 100.6%, and 99.5% for 56.7, 85.1, and 113.4 g/L of added thyroxine, respectively.
serum results The LC/MS-ESI method for the determination of thyroxine was applied to samples of frozen serum pools. Samples were prepared and analyzed in three different sets, each set consisting of three concentrations of thyroxine. The results are shown in Table 2 for positive ions and Table 3 for negative ions. Excellent reproducibility was obtained for all three concentrations, with within-set CVs Table 2. Positive-ion LC/MS-ESI measurements of thyroxine in serum. Thyroxine
Overall
Conca
Set
Mean, g/L
SD,b g/L
CV, %
1 1 1 2 2 2 3 3 3
1 2 3 1 2 3 1 2 3
49.4 49.4 49.6 108.0 109.2 110.1 167.0 166.8 169.2
0.24 0.31 0.47 0.55 0.34 0.52 0.96 1.07 1.20
0.5 0.6 1.0 0.5 0.3 0.5 0.6 0.6 0.7
a
Conc, concentration. b SD of the single measurements within a set.
Mean, g/L
49.5
SD, g/L
CV, %
0.1
0.2
109.1
0.6
0.5
167.7
0.8
0.5
Thyroxine
Overall
Conca
Set
Mean, g/L
SD,b g/L
CV, %
1 1 1 2 2 2 3 3 3
1 2 3 1 2 3 1 2 3
50.0 49.2 50.3 109.1 109.8 110.5 168.3 167.1 167.2
0.34 0.49 0.18 0.32 0.24 1.05 0.86 0.35 1.33
0.7 1.0 0.4 0.3 0.2 1.0 0.5 0.2 0.8
Mean, g/L
SD, g/L
CV, %
49.9
0.3
0.6
109.8
0.4
0.4
167.6
0.4
0.2
a
Conc, concentration. b SD of the single measurements within a set.
of 0.2–1.0% and between-set CVs of 0.2– 0.6%. Excellent linearity was also obtained, with the correlation coefficients (r) for all linear regression lines (measured intensity ratios vs mass ratios) ranging from 0.999 to 1.000. A linear regression line was generated for each group of samples. A typical regression line was: y ⫽ ⫺0.0654 ⫹ 1.2206x (r ⫽ 0.9998; SE ⫽ 0.0070; n ⫽ 8). The statistical analysis, as given in Table 4, showed that positive- and negative-ion measurements were in good agreement with a mean difference of 0.45%, demonstrating that there is no significant undetected bias. The detection limits (at a signal-tonoise ratio of ⬃3 to 5) were 30 and 20 pg for positive and negative ions, respectively. Single-ion chromatograms for thyroxine and thyroxine-d5 are shown in Figs. 1 and 2 for positive and negative ions, respectively.
statistical analysis The overall uncertainty in the measurements includes the measurement repeatability (type A) and other factors Table 4. Calculation of expanded uncertainties for LC/MS-ESI measurements of thyroxine in serum. Positive ion Mean, g/L Relative SD of mean, % Negative ion Mean, g/L Relative SD of mean, % Bias between methods,a % Uncertainty of purity of reference compound, % Combined relative SD uncertainty, % Degrees of freedom K factor Relative expanded uncertainty, % Mean, g/L Uncertainty,b g/L a b
Conc 1
Conc 2
Conc 3
49.5 0.15
109.1 0.55
167.7 0.46
49.9 0.63 0.45 0.2
109.8 0.39 0.45 0.2
167.6 0.22 0.45 0.2
0.81 5.58 2.57 2.1 49.7 1.0
0.84 8.38 2.31 1.9 109.5 2.1
0.71 10.52 2.23 1.6 167.7 2.7
Mean bias across three concentrations (see text). Uncertainty of 95% confidence interval.
641
Clinical Chemistry 48, No. 4, 2002
Table 5. Comparison of LC/MS-ESI method with routine clinical methods, mean results. Conc
LC/MS-ESI, g/L
Routine methods,a g/L
Difference, g/L
Relative difference, %
1 2 3
49.7 ⫾ 1.0b 109.5 ⫾ 2.1b 167.7 ⫾ 2.6b
54.4 ⫾ 5.9c 113.9 ⫾ 11.1c 172.6 ⫾ 19.2c
4.7 4.4 4.9
9.5 4.0 2.9
a
CAP, private communication. Uncertainty of 95% confidence interval. c Standard deviation of the composite results from all the methods. b
Fig. 1. Single-ion chromatograms obtained by positive-ion LC/MS-ESI for thyroxine and thyroxine-d5 from a serum sample, concentration 2.
(type B) that include the bias between the two measurement approaches and the purity of the reference compound. The calculation of the uncertainty is shown in Table 4. The standard deviations from the positive- and negative-ion measurements were quite similar, thus contributing about equally to the uncertainty overall. As with most high-precision ID/MS measurements, within-set variability was small compared with set-to-set variability for most cases, as shown by analyses of variance. Consequently, for each concentration under each measurement condition, n was taken as 3 (the number of sets). The data from the positive- and negative-ion measurements sug-
Fig. 2. Single-ion chromatograms obtained by negative-ion LC/MS-ESI for thyroxine and thyroxine-d5 from a serum sample, concentration 1.
gest that there may be a small bias between the results from the two types of measurements. The mean difference was 0.45% for the negative-ion results vs the positive-ion results. It is possible that there was a trend toward smaller differences at higher concentrations, but the differences for individual samples were scattered sufficiently to preclude any definitive conclusions about biases between the two types of measurements. Therefore, for the purpose of calculating the uncertainty, the mean difference was used. The reference compound, the unlabeled thyroxine calibrator, was evaluated for purity. The estimated purity with its 95% confidence interval, based on results from multiple techniques plus an allowance for undetected impurities, was 98.7% ⫾ 0.4%. This uncertainty in the purity was included in determining the overall uncertainty in the measurements.
comparison with routine clinical methods The results obtained by the LC/MS-ESI method and the mean value of the routine clinical methods [⬃2000 laboratories from College of American Pathologists (CAP) surveys] are compared in Table 5, which shows that the relative standard deviations are ⱖ10% for the composite results from many routine clinical methods in the CAP survey for the three concentrations measured in this study, demonstrating that there is considerable variability among the clinical methods in use today. The results from the candidate reference method were within 1 SD of the composite means for these same materials in the CAP survey, although it appears that the clinical method means may be biased high by 4 –5 g/L across the concentrations. If this is true, then some of the methods in routine use may be biased high by up to 20% at the low concentrations. In conclusion, this well-characterized LC/MS-ESI method for total serum thyroxine, with its theoretically sound approach, demonstrated good accuracy and precision and low susceptibility to interferences, which qualifies it as a candidate reference method. Use of this reference method as an accuracy base may reduce the apparent biases in routine methods along with the high interlaboratory imprecision. NIST is planning to develop a serum-based Standard Reference Material for thyroxine. Because there is considerable interest in the measurement of free
642
Tai et al: LC/MS Candidate Reference Method for Thyroxine
(non-protein-bound) thyroxine, research will be conducted into the isolation of free thyroxine from bound thyroxine. If a scientifically sound and reproducible approach to this isolation can be found, this method will be applied to the measurement of free thyroxine. Methods for thyroxine determination by ID/MS are under consideration for comparison studies involving national metrologic institutes from around the world.
We gratefully acknowledge the CAP for partial support for the development of this method. We thank Alex Cohen for the synthesis of thyroxine-d5 and Sam Margolis for the Karl Fischer analysis.
References 1. Welch M, Colbert J. Certification strategies for health-related SRMs at NIST. Fresenius J Anal Chem 1998;360:312– 6. 2. Siekmann L. Measurement of thyroxine in human serum by isotope dilution mass spectrometry. Definitive methods in clinical chemistry, V. Biomed Environ Mass Spectrom 1987;14:683– 8. 3. Thienpont LM, De Brabandere VI, Sto¨ckl D, De Leenheer AP. Development of a new method for the determination of thyroxine in serum based on isotope dilution gas chromatography mass spectrometry. Biomed Mass Spectrom 1994;23:475– 82. 4. Mo¨ller B, Falk O, Bjo¨rkhem I. Isotope dilution-mass spectrometry
of thyroxine proposed as a reference method. Clin Chem 1983; 29:2106 –10. 5. De Brabandere VI, Sto¨ckl D, Thienpont LM, De Leenheer AP. On the use of trimethylchlorosilane in methanol for methylation of thyroxine prior to perfluoroacylation and isotope dilution-gas chromatography/mass spectrometry. J Mass Spectrom Lett 1998;33: 1032. 6. De Brabandere VI, Hou P, Sto¨ckl D, Thienpont LM, De Leenheer AP. Isotope dilution-liquid chromatography/electrospray ionization-tandem mass spectrometry for the determination of serum thyroxine as a potential reference method. Rapid Commun Mass Spectrom 1998;12:1099 –103. 7. Tai SS, Welch MJ. Determination of 11-nor-⌬9-tetrahydrocannabinol-9-carboxylic acid in a urine-based standard reference material by isotope-dilution liquid chromatography-mass spectrometry with electrospray ionization. J Anal Toxicol 2000;24:385–9. 8. Ramsden DB, Farmer MJ. Development of a gas chromatographic selected ion monitoring assay for thyroxine (T4) in human serum. Biomed Mass Spectrom 1984;11:421–7. 9. Nishinaga A, Cahnmann HJ, Kon H, Matsuura T. Model reactions for the biosynthesis of thyroxine. XII. The nature of a thyroxine precursor formed in the synthesis of thyroxine from diiodotyrosine and its keto acid analog. Biochemistry 1968;7:388 –97. 10. Hay ID, Annesley TM, Jiang NS, Gorman CA. Simultaneous determination of D- and L-thyroxine in human serum by liquid chromatography with electrochemical detection. J Chromatogr 1981;226: 383–90.
