Development and Evaluation of a Reference
Measurement Procedure for the Determination of
Estradiol-17â in Human Serum Using
Isotope-Dilution Liquid Chromatography-Tandem
Mass Spectrometry
Susan S.-C. Tai* and Michael J. Welch
Analytical Chemistry Division, National Institute of Standards and Technology, Gaithersburg, Maryland 20899-8392
Estradiol is the most potent natural estrogen and is
derived from the ovaries. Its concentration in blood is
measured to determine ovarian function. A reference
measurement procedure for estradiol in serum involving
isotope-dilution coupled with liquid chromatography-
tandem mass spectrometry (LC/MS/MS) has been devel-
oped and critically evaluated. A deuterated estradiol
(estradiol-d3) was used as an internal standard. The
estradiol and its internal standard were extracted from
serum matrix using solid-phase extractions and deriva-
tized with dansyl chloride prior to reversed-phase LC/MS/
MS. The accuracy of the measurement was evaluated by
a comparison of results of this reference method on
lyophilized human serum reference materials for estradiol
[Certified Reference Materials (CRMs) 576, 577, and
578] with the certified values determined by gas chro-
matography/mass spectrometry (GC/MS) reference meth-
ods and by a recovery study for the added estradiol. The
results of this method for estradiol agreed with the
certified values within the uncertainty of the measure-
ments for the three CRMs. The recovery of the added
estradiol ranged from 100.7 to 101.8%. This method was
applied to the determination of estradiol in frozen serum
samples from three individual female donors. Excellent
reproducibility was obtained with within-set coefficient of
variations (CVs) ranging from 0.6 to 2.2% and between-
set CVs ranging from 0.2 to 0.4%. Excellent linearity was
also obtained, with correlation coefficients of all linear
regression lines (measured intensity ratios vs mass ratios)
ranging from 0.998 to 1.000. The detection limit at a
signal-to-noise ratio of �3 was 0.6 pg of estradiol (or 1
ng/L, as expressed as a concentration). This well-
characterized LC/MS/MS method for serum estradiol,
which demonstrates good accuracy and precision, low
susceptibility to interferences, and comparability with GC/
MS reference methods, qualifies as a reference measure-
ment procedure that can be used to provide an accuracy
base to which routine methods for estradiol can be
compared and that will serve as a standard of higher order
for measurement traceability.
Estrogens are hormones that are responsible for the develop-
ment and maintenance of the female reproductive organs and
female secondary gender characteristics. Estrogens also partici-
pate in the regulation of the menstrual cycle and in the mainte-
nance of pregnancy. Estradiol is the most potent natural estrogen
and is derived from the ovaries. Estradiol is largely bound to
protein in circulation. Its concentration in blood is measured to
determine ovarian function.1
The major challenge of serum estradiol measurement is the
low concentrations of estradiol normally found in serum (pico-
grams/milliliter). Liquid chromatography/mass spectrometry
(LC/MS) and liquid chromatography/tandem mass spectrometry
(LC/MS/MS) have recently been used for low-concentration
thyroid hormone analysis due to the high sensitivity, reproduc-
ibility, and selectivity of these approaches.2-4 However, the
ionization efficiencies for most steroid hormones which lack highly
ionizable functional groups are relatively low for different ionization
methods; therefore, the sensitivities of steroids at picogram levels
are difficult to achieve. Recently, new techniques which utilize
the introduction of a highly ionizable group into steroid molecules
to significantly enhance the ionization efficiencies have been
developed for various steroid hormones.5-9 The derivatization of
estrogens with a dansyl functional group resulted in reported
signal enhancements of up to 1000-fold using LC/MS/MS,6-9 thus
making it possible to measure serum estradiol at picogram levels
by LC/MS/MS.
Serum estradiol measurements are routinely performed using
methods based upon immunoassays, an approach with high sensi-
tivity, but which is prone to nonspecificity for many analytes. There
* Author for correspondence. E-mail: susan.tai@nist.gov.
(1) Tietz, N. W. In Tietz Textbook of Clinical Chemistry, 2nd ed.; Burtis, C. A.,
Ashwood, E. R., Eds.; Saunders: Philadelphia, 1994; pp 1857-1864.
(2) Tai, S. S.; Bunk, D. M.; White, E. V.; Welch, M. J. Anal. Chem. 2004, 76,
5092-5096.
(3) Thienpont, L. M.; Fierens, C.; De Leenheer, A. P.; Przywara, L. Rapid
Commun. Mass Spectrom. 1999, 13, 1924-1931.
(4) Tai, S. S.; Sniegoski, L. T.; Welch, M. J. Clin. Chem. 2002, 48, 637-642.
(5) Higashi, T.; Shimada, K. Anal. Bioanal. Chem. 2004, 378, 875-882.
(6) Nelson, R. E.; Grebe, S. K.; O’Kane, D. J.; Singh, R. J. Clin. Chem. 2004,
50, 373-384.
(7) Xia, Y.; Chang, S. W.; Patel, S.; Bakhtiar, R.; Karanam, B.; Evans, D. C.
Rapid Commun. Mass Spectrom. 2004, 18, 1621-1628.
(8) Shou, W. Z.; Jiang, X.; Naidong, W. Biomed. Chromatogr. 2004, 18, 414-
421.
(9) Anari, M. R.; Bakhtiar, R.; Zhu, B.; Huskey, S.; Franklin, R. B.; Evans, D. C.
Anal. Chem. 2002, 74, 4136-4144.
Anal. Chem. 2005, 77, 6359-6363
10.1021/ac050837i CCC: $30.25 © 2005 American Chemical Society Analytical Chemistry, Vol. 77, No. 19, October 1, 2005 6359
Published on Web 08/26/2005
is a need for critically evaluated reference measurement proce-
dures (RMPs) for serum estradiol. RMPs can be used to directly
assess the accuracy of routine methods10 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 Organization for
Standardization (ISO) published ISO 15193 (In Vitro Diagnostic
SystemssMeasurement of Quantities in Samples of Biological
OriginsPresentation of Reference Measurement Procedures)11
that describes the requirements of an RMP for clinical diagnostic
markers. The Joint Committee for Traceability in Laboratory
Medicine (JCTLM)12 is reviewing potential RMPs and compiling
a list of those that meet the requirements of ISO 15193. At present,
there are three RMPs for serum estradiol approved by the JCTLM.
Two of these methods13-15 involve liquid/liquid extraction to iso-
late estradiol from the serum and then fractionation by Sephadex
LH-20 chromatography13,14 or reversed-phase LC15 to further purify
estradiol from the serum extract prior to gas chromatography/
mass spectrometry (GC/MS) analysis. The third method involves
liquid/liquid extraction to isolate estradiol prior to LC/MS/MS
analysis.6 This method was developed for use as a high-throughput
routine method with coefficient of variations (CVs) of 2-20%.
Recently, the National Institute of Standards and Technology
(NIST) developed an isotope-dilution (ID) coupled with LC/MS/
MS-based method for serum estradiol that meets the requirements
of an RMP as defined by ISO 15193. This method measures total
unconjugated estradiol in serum, both free and protein-bound
forms. The method involves two solid-phase extractions (SPEs),
one with a retention mechanism of reversed phase and the other
with anion exchange, to isolate estradiol from serum, and
derivatization of estradiol with dansyl chloride6,7 to enhance
electrospray ionization efficiencies prior to reversed-phase LC/
MS/MS analysis. This RMP was found to be free from interfer-
ences by testing the structure analogues and conjugated metabo-
lites of estradiol under the same assay conditions for estradiol
measurement. The accuracy of the measurement was evaluated
by a comparison of results of this RMP on CRMs 576, 577, and
578 [from the Community Bureau of Reference (BCR)] with the
certified values determined by GC/MS reference methods16 and
by a recovery study for the added estradiol. The reproducibility
of this method was evaluated by repeated measurements of frozen
serum samples from three individual female donors and found to
be significantly improved as compared to the previously reported
LC/MS/MS method.6
EXPERIMENTAL SECTION
This method measures total unconjugated estradiol in serum,
both free and protein-bound forms.
Materials. The estradiol reference compound used for this
work was obtained from Sigma (St. Louis, MO). The impurities
in this estradiol material were evaluated at NIST by liquid
chromatography/ultraviolet (LC/UV), and moisture content was
measured by Karl Fischer titration. Appropriate corrections were
made for impurities and moisture. A stable deuterium-labeled
internal standard, estradiol-d3 (16,16,17-d3), with an isotopic purity
of 98% was obtained from Aldrich (Milwaukee, WI). 3H-labeled
estradiol was obtained from Perkin-Elmer NEN (Boston, MA).
C18 Sep-Pak solid-phase extraction cartridges (Vac 3 mL, 500 mg)
were obtained from Waters (Milford, MA). Solid-phase XTRX type
AX/S extraction columns were obtained from Creative Technology
Systems, Inc. (Newark, DE). A Zorbax Eclipse XDB-C18 column
[15 cm � 2.1 mm (i.d.); 5-ím particle diameter] was obtained from
Agilent Technologies (Palo Alto, CA). Dansyl chloride was
obtained from Sigma (St. Louis, MO). Solvents used for LC/MS/
MS measurements were HPLC grade, and all other chemicals
were reagent grade. Frozen human serum samples from individual
female donors were obtained from Interstate Blood Bank, Inc.
(Memphis, TN). Samples of Seracon II, defibrinated, dialyzed,
delipidated, and charcoal-stripped frozen human serum were
obtained from Seralogicals Corporation (Norcross, GA). CRMs
576, 577, and 578, lyophilized human serum reference materials
for estradiol, manufactured by the European Commission Institute
for Reference Materials and Measurements (formerly BCR), were
obtained from Resource Technology Corporation (Laramie, WY).
The following compounds were used for interference testing:
Estrone, 17R-estradiol, 17â-estradiol-3-(â-D-glucuronide) sodium
salt, 17â-estradiol-3-sulfate sodium salt, 17â-estradiol 3-(â-D-glu-
curonide)-17-sulfate dipotassium salt, and 17â-estradiol-3,17-di-
sulfate dipotassium salt were obtained from Sigma (St. Louis, MO).
17â-Estradiol-17-sulfate sodium salt and 17â-estradiol-17-(â-D-
glucuronide) were obtained from Steraloids (Newport, RI).
Preparation of Calibrators. Two independently weighed
standard stock solutions of estradiol were prepared. Approximately
2 mg of the estradiol reference compound for each stock solution
was accurately weighed on an analytical balance and dissolved in
methanol in a 1000-mL volumetric flask. A working solution was
prepared from each stock by diluting 1.0 mL in methanol in a
500-mL volumetric flask. The concentrations of estradiol in the
working standard solutions were �4 pg/íL. A solution of isoto-
pically labeled internal standard, estradiol-d3, at a concentration
of �2 pg/íL was prepared in the same way as the unlabeled
estradiol.
Three aliquots from each working standard solution of estradiol
were spiked with estradiol-d3, yielding six standards with the mass
ratios of unlabeled to labeled compound ranging from 0.5 to 1.5.
The mixtures (510-700 íL) were dried under nitrogen at 45 °C;
derivatized with 300 íL of 1.0 g/L dansyl chloride in acetone-
0.1 mol/L sodium bicarbonate buffer, pH 10.4 (50:50 by volume)
at 60 °C for 5 min; and cooled to refrigerated temperatures for
LC/MS/MS analysis. The estradiol concentration range of the
calibration curve was �1.3 to 3.9 pg/íL (for mass ratio range of
0.5-1.5).
Sample Preparation. Commercially available frozen serum
samples from three individual female donors (concentrations 1,
2, and 3, respectively) were used for this study. Samples were
prepared in three different sets. Each set consisted of triplicate
5.0-mL aliquots from each of concentrations 1 and 2, and triplicate
3.0-mL aliquots from concentration 3. Each aliquot was placed into
(10) Thienpoint, L. M.; De Leenheer, A. P. Clin. Chem. 1998, 44, 671-674.
(11) ISO 15193. 2003; http://www.iso.ch/iso/en/CatalogueListPage; CatalogueL-
ist.
(12) JCTLM. 2005; http://www.bipm.org/en/committees/jc/jctlm/jctlm-db.
(13) Siekmann, L. J. Clin. Chem. Clin. Biochem. 1984, 22, 551-557.
(14) Siekmann, L.; Siekmann, A.; Breuer, H. Fresenius’ J. Anal. Chem. 1978,
290, 122-123.
(15) Thienpont, L. M.; Verhaeghe, P. G.; Kristien, A.; Brussel, V.; De Leenheer,
A. P. Clin. Chem. 1988, 34, 2066-2069.
(16) Thienpont, L. M.; De Leenheer, A. P.; Dirscherl C. D. EUR 17540 EN, 1997.
6360 Analytical Chemistry, Vol. 77, No. 19, October 1, 2005
a 50-mL glass centrifuge tube containing 2 íg of R-naphathol as
a carrier13 and 62 íg (200 nmol) of D-norgestrel for displacing
estradiol from the binding protein,15,17 and an appropriate amount
of estradiol-d3 was added to give an approximate 1:1 ratio of analyte
to internal standard. Each sample was adjusted to pH 5.0 by adding
0.5 mol/L sodium acetate buffer, pH 5.0. After equilibration at
room temperature for 2 h, estradiol was isolated from the serum
matrix using C18 Sep-Pak SPE cartridges. Each sample was loaded
onto a cartridge previously conditioned by wetting with 5 mL of
methanol, followed by 5 mL of water. The cartridge was then
washed with 12 mL of water, followed by 8 mL of water-
acetonitrile (70:30 by volume). The estradiol was eluted from the
cartridge with 2.5 mL of methanol, and the eluate was dried under
nitrogen at 45 °C. The estradiol was further purified using solid-
phase XTRX type AX/S extraction columns.18 The residue was
reconstituted with 1.0 mL of 2-propanol-methanol (50:50 by
volume) and 2.5 mL of 2.0 mol/L sodium hydroxide, and the
solution was applied to an AX/S column. The column was then
washed sequentially with 4.0 mL of methanol, 1.5 mL of water,
3.0 mL of 50 mL/L acetic acid in water, followed by 2.0 mL of
methanol-water (20:80 by volume). The estradiol was eluted from
the column with 4.0 mL of methanol. The eluate was dried and
derivatized following the procedure described above for calibrator
preparation. The derivative is stable at refrigerated temperatures
for at least 4 weeks.
The estradiol concentrations of samples were first tested by a
preliminary experiment in which arbitrary amounts of the internal
standard were chosen, and a wider range of mass ratios (e.g.,
0.3-4.3) for the standards were used. Once the approximate
estradiol concentration was determined, the quantity of internal
standard can be calculated to get a 1:1 ratio.
[3H]-Labeled estradiol was used as a tracer to evaluate the
recovery of estradiol from serum using this extraction method.
Equilibration. Commercially available frozen human serum
samples from a single donor were used for this study. The
concentration of estradiol in this serum material was 170 ng/L.
Eight 3.0-mL aliquots were taken for the equilibration study. An
appropriate amount of estradiol-d3 was added to each aliquot, and
duplicate aliquots for each of the four time intervals (1, 2, 3, and
4 h) were equilibrated at room temperature. The samples were
processed as described above for LC/MS/MS measurement.
Recovery of the Added Estradiol. Commercially available
frozen human serum samples from a single donor were used for
this study. Three 5.0-mL (for samples without unlabeled estradiol
added) and nine 3.0-mL aliquots (for samples with unlabeled
estradiol added) were taken for a study of the accuracy of the
method. A known amount of unlabeled estradiol was added to
the nine 3.0-mL aliquots, three each with 91.7, 183.4, and 275.1
ng/L of estradiol. No estradiol was added to the three 5.0-mL
aliquots. An appropriate amount of estradiol-d3 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 an Applied
Biosystems API 4000 (Foster City, CA) equipped with an Agilent
1100 Series LC system (Palo Alto, CA). A Zorbax Eclipse XDB-
C18 column was used for the analysis. Aliquots of standards (20
íL) or sample extracts (25 íL) were separated by LC with a
gradient mobile phase consisting of 1 mL/L acetic acid in water-
acetonitrile. The gradient was initially set at water-acetonitrile
(35:65 by volume) for 25 min, ramped to 100% acetonitrile at 26
min, and held for 10 min to wash the column. The flow rate was
0.25 mL/min. The column temperature was set at 25 °C. The
autosampler tray temperature was set at 10 °C. Electrospray
ionization in the positive ion mode and multiple reaction monitor-
ing (MRM) mode were used for LC/MS/MS. The transitions at
m/z 506 f 171 and m/z 509 f 171 were monitored for dansyl
estradiol and dansyl estradiol-d3, respectively.6,7 The dwell times
were 0.25 s for MRM. The curtain gas and collision gas were
nitrogen at settings of 207 kPa (30 psi) and 34 kPa (5 psi),
respectively. The ion source gases 1 and 2 were air at settings of
448 kPa (65 psi) and 483 kPa (70 psi), respectively. The
electrospray voltage was set at 5500 V, and the turbo gas
temperature was maintained at 400 °C. The declustering potential,
entrance potential, collision energy, and collision exit potential
were set at 126, 10, 47, and 12 V, respectively.
The following measurement protocol was used for LC/MS/
MS analysis. The standards were analyzed first, followed by the
samples, then by the samples and standards in reversed order.
By combining the data of standards run before and after the
samples, a 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.
Interference Testing. The metabolites or structural analogues
of estradiol having molecular masses close to that of estradiol,
estrone (270 Da) and estradiol-17R (272 Da), were tested for
interference using the LC/MS/MS method described above for
estradiol measurement. Estrone and 17R-estradiol were derivatized
with dansyl chloride, and the transitions at m/z 504 f 171 and
m/z 506 f 171 were monitored for dansyl estrone and dansyl
estradiol, respectively.
The conjugated (glucuronidated and sulfated) metabolites of
estradiol listed in the materials section were tested for interfer-
ence. Stripped serum samples spiked with each conjugated
estradiol were extracted and derivatized following the procedure
described in the sample preparation section. The transition at m/z
506 f 171 was monitored for dansyl estradiol.
Statistical Treatment. Statistical treatment of the data was
in accord with NIST guidelines,19 which conform with the ISO
Guide to the Expression of Uncertainty in Measurement.20
Potential sources of uncertainty were evaluated, and those factors
that could contribute significantly were used to calculate the
standard uncertainty. The type A uncertainty, calculated from the
imprecision of the measurements, was combined quadratically
with the type B uncertainty components to determine the standard
uncertainty uc. For type A, an analysis of variance calculation was
performed on the measurement data to determine whether set-
set differences were statistically significant. This analysis deter-
mined the number of independent measurements, n, used for
calculating the measurement standard deviation of the mean. The
(17) Heikkinen, R.; Fotsis, T.; Adlercreutz, H. Clin. Chem. 1981, 27, 1186-
1189.
(18) Wu, H.; Ramsay, C.; Ozaeta, P.; Liu, L.; Aboleneen, H. Clin. Chem. 2002,
48, 364-366.
(19) Taylor, B. N.; Kuyatt, C. E. NIST Technical Note 1297, 1994 (available at
http://physics.nist.gov/Pubs/guidelines/contents.html).
(20) ISO. Guide to the Expression of Uncertainty in Measurement; International
Organization for Standardization: Geneva, Switzerland, 1993.
Analytical Chemistry, Vol. 77, No. 19, October 1, 2005 6361
type B factor contributions used were based on knowledge about
uncertainty i