Clinical Chemistry 46:2
252–257 (2000)
Endocrinology and
Metabolism
Evaluation of a Bead-based Enzyme Immunoassay
for the Rapid Detection of Osteocalcin in
Human Serum
Alexandra M. Crăciun,1 Cees Vermeer,1 Hans-Georg Eisenwiener,2 Norbert Drees,2
and Marjo H.J. Knapen1*
Background: Circulating osteocalcin is a well-known
marker for bone formation, but none of the commercial
kits currently available can be used in automated systems. Here we present the first semiautomated assay for
human serum osteocalcin.
Methods: Polystyrene beads were coated with antibodies against the COOH terminus of osteocalcin and used
in the COBAS® EIA System. Osteocalcin was detected
with peroxidase-conjugated antibodies against the osteocalcin NH2 terminus.
Results: The time required to analyze an unknown sample was 60 min, with a lower detection limit of 4.5 mg/L
and a linear dose–response curve between 4.5 and 100
mg/L. The intraassay imprecision (CV) was 5– 8% (n 5 21);
the interassay variation was 6 –9% (n 5 14). In samples
from human volunteers and patients, data generated with
the newly developed assay were comparable to those
obtained with standard microtiter plate-based assays.
Conclusions: The coated beads assay may be implemented on fully automated analyzers, which not only may
further reduce imprecision but may also substantially
increase the applicability of osteocalcin as a marker for
bone metabolism in the routine clinical setting.
posttranslational formation of its three g-carboxyglutamate (Gla) residues. Although its function on a molecular
level has remained unclear to date, increased bone formation, including higher bone mass and improved bone
strength, was observed in OC-deficient (knock-out) mice
(5 ). These experiments have demonstrated that OC has an
important role in the regulation of bone growth and in the
correct deposition of the mineral matrix in bone. Because
20 –30% of the de novo synthesized OC is not accumulated in the bone tissue, but is secreted into the blood
stream, circulating OC is widely used as a biomarker for
bone formation (6 – 8 ).
During episodes of vitamin K deficiency or during
treatment with vitamin K antagonists, Gla-containing
proteins are synthesized in an undercarboxylated (Gladeficient) form. Fully carboxylated and undercarboxylated OC may be quantified separately on the basis of
their different affinities for hydroxyapatite (9 ). It has been
reported by various groups that in the general population,
a substantial fraction of circulating OC occurs in its
undercarboxylated form (9 –11 ), and that the concentration of undercarboxylated OC may have an independent
diagnostic value for the assessment of bone mass and
bone fracture risk (12–14 ). During recent years, an increasing number of test kits for serum OC have become
commercially available, all of which are either radioimmunoassays or enzyme-based immunoassays for microtiter plates. Some of the drawbacks of such tests are that
they are laborious and prone to errors by the persons
performing the tests, and that they generally require long
incubation steps (either all day or overnight) before the
data become available. In addition, the various OC kits
© 2000 American Association for Clinical Chemistry
Osteocalcin (OC),3 also known as bone Gla protein, is the
most abundant noncollagenous protein in mature bone
(1, 2 ). Osteoblasts are the exclusive site of OC biosynthesis, which is regulated at the transcription stage by
vitamin D (3, 4 ), whereas vitamin K is required for the
1
Department of Biochemistry and Cardiovascular Research Institute,
Maastricht University, P.O. Box 616, 6200 MD Maastricht, The Netherlands.
2
Hoffmann-La Roche Diagnostics, Basel 4070, Switzerland.
*Address corresponding to this author at: Department of Biochemistry,
University of Maastricht, P.O. Box 616, 6200 MD Maastricht, The Netherlands.
Fax 31-43-367-0992; e-mail m.knapen@bioch.unimaas.nl.
Received September 29, 1999; accepted November 30, 1999.
3
Nonstandard abbreviations: OC, osteocalcin; Gla, g-carboxyglutamate;
BAP, bone-specific alkaline phosphatase; mAb, monoclonal antibody; PTH,
parathyroid hormone; OHPro, hydroxyproline; DPD, deoxypyridinoline; and
CTX, type I collagen C-terminal telopeptide.
252
253
Clinical Chemistry 46, No. 2, 2000
may give widely different values when the same serum
sample is tested (6, 15 ). This may be related to different
specificities of the antibodies used for fully carboxylated
and undercarboxylated OC and whether OC degradation
products are recognized.
To improve the accuracy and reproducibility of OC
quantification, we used commercial antibodies that were
coated onto polystyrene beads and used in the Roche
COBAS® EIA System. We compared the performance of
this semiautomated enzyme immunoassay with commercial radiometric and microtiter plate-based assays for
various markers of bone metabolism.
Materials and Methods
subjects
The reference values for the concentrations of OC and
bone-specific alkaline phosphatase (BAP) were established in serum samples obtained from 151 apparently
healthy subjects (60 men and 91 women) between 19 and
86 years of age who were recruited via a local newspaper.
The within-day and day-to-day variations of OC and
several other markers were examined in 12 apparently
healthy volunteers (6 men and 6 women) between 19 and
27 years of age recruited from the students of the Maastricht University. During the first 24 h of the experiment,
urine and blood samples were collected at 0900 (start) and
at 1100, 1400, 1700, 2100, 2300, 0300, and 0700. Subsequent
samples were collected at 0900 on days 2, 8, 15, 22, 29, and
57. All samples taken at 0900 were obtained after an
overnight fast (only water allowed after 1900 of the
preceding night). Changes in bone markers during the
follow-up of treatment were recorded in serum and urine
samples from 30 osteoporotic women (.65 years of age),
before and after treatment with either bisphosphonate
(Alendronate, 5–10 mg/day for 15 months; Merck), estrogen (Livial, 2.5 mg/day for 6 months; Organon), or
calcitonin (Miacalcic nasal application, 100 IU/day for 12
months; Sandoz). All studies were approved by the University Hospital Medical Ethics Committee.
sample collection and storage
Blood (10 mL) was taken by venipuncture to prepare
serum, 0.5-mL aliquots of which were frozen at 280 °C
within 2 h after sample collection until use. All urine
samples collected at the 0900 time points were obtained
after an overnight fast (14 h) and represent the 2-h second
morning void. Other urine samples collected during the
first day were nonfasting samples. All urine samples were
stored in 0.5-mL aliquots at 230 °C until use.
procedure for the semiautomated oc
immunoassay
The assay (hereafter called the Coated Beads Assay) is
based on commercial monoclonal antibodies (mAbs)
against synthetic peptides homologous to OC residues
1–19 and 20 – 43 (Osteometer), and will be designated here
as mAb1–19 and mAb20 – 43. Polystyrene beads (2 mm
Fig. 1. Calibration curve for OC by Coated Beads Assay.
Calibrators used were from the commercial N-mid Osteocalcin assay (Osteometer). Bars, SD.
diameter) were coated with mAb20 – 43 by Osteometer,
other components for the test were those in the microtiterbased N-midTM Osteocalcin ELISA (Osteometer), and are
described by Rosenquist et al. (16 ). The test procedure, for
which we used the COBAS EIA System from Hoffmann-La Roche (Basel, Switzerland), was as follows: 25
mL of sample (calibrator, control, or serum) was pipetted
into polycarbonate tubes, together with 225 mL of buffer A
(0.14 mol/L NaCl, 0.01 mol/L sodium phosphate, pH 7.4)
and 25 mL of peroxidase-conjugated mAb1–19. Subsequently, one coated bead was added to each tube and
incubated for 30 min at 37 °C while shaking in the COBAS
EIA Incubator. The beads were then washed with distilled
water in the COBAS EIA Washer, and 250 mL of substrate
solution (tetramethylbenzidine) was added, after which
the tubes were incubated for another 15 min at 37 °C with
constant shaking. The reaction was stopped by the addition of 1 mL of 0.1 mol/L H2SO4; within 1 h after
Table 1. Recovery of OC in pathological serum samples
after serial dilution.a
Recovery of OC after dilution, %
Dilution
factor
Sample A
Sample Bc
Sample Cd
1.00
1.25
1.67
2.00
2.50
5.00
100.0 6 1.2
102.7 6 1.8
105.0 6 4.7
99.7 6 4.7
96.6 6 2.4
80.0 6 5.1
100.0 6 2.3
91.9 6 3.4
88.4 6 3.9
90.0 6 8.9
84.0 6 6.2
77.5 6 3.5
100.0 6 2.9
93.6 6 4.2
88.7 6 7.4
85.1 6 5.4
79.9 6 4.5
69.5 6 6.0
b
a
The samples were diluted with calibrator A, and the observed OC concentrations were multiplied by the dilution factor. All samples were tested in triplicate,
and the experiment was repeated on 3 different days. Mean values are given 6
SD.
b
Undiluted serum: OC concentration, 117.4 mg/L.
c
Prediluted twofold: OC concentration, 114.3 mg/L.
d
Prediluted fivefold: OC concentration, 75.3 mg/L.
254
Crăciun et al.: Coated Beads Assay for Human Serum Osteocalcin
Table 2. Bone formation markers in different age groups of men and women.a
OC, mg/L
Age range,
years
Men
19–40
59–86
Women
19–40
41–58
59–86
Mean age,
years
n
Beads
IRMA
BAP, mg/L
27.5 6 6.3
74.4 6 6.5
30
30
21.7 6 10.1
16.8 6 11.9
25.2 6 10.4
16.6 6 9.6b
16.7 6 5.3
13.9 6 8.1
30.5 6 6.1
51.1 6 4.9
75.1 6 6.3
31
28
32
18.8 6 6.0
18.4 6 6.6
22.0 6 10.6
16.6 6 6.1
16.5 6 6.8
17.5 6 9.4
11.5 6 4.7
9.5 6 5.0
15.1 6 6.4c
a
All samples were tested in triplicate, and mean values are given 6 SD. Statistics were performed with the unpaired Student t-test.
Significantly different from men in the 19 – 40 age group.
c
Significantly different from women in the 41–58 age group.
b
termination of the reaction, the absorbance at 450 nm was
recorded with a 25-channel COBAS EIA Photometer.
Serum OC concentrations were calculated using a fourparameter logistic curve fit based on the calibrators of the
assay (0, 6.25, 12.5, 25, 50, and 100 mg/L).
other tests used
The data obtained with the experimental OC assay were
compared with commercial kits for OC (ELSA-Osteo; CIS
Bio-international) and BAP (Tandem-R Ostase; Hybritech). Both tests are two-site IRMAs. Intact parathyroid
hormone (PTH) was quantified in serum with the N-tact
PTH radioimmunoassay from Incstar. Markers tested in
urine were hydroxyproline (OHPro; hypronosticon; Organon Teknika), deoxypyridinoline (DPD; Pyrilinks-D;
Metra Biosystems), and type I collagen C-terminal telopeptide (CTX; CrossLaps; Osteometer BioTech). Creatinine was assessed in urine by standard enzymatic techniques (Boehringer Mannheim) on a Beckman Synchron
CX7-2 automated analyzer. Urinary calcium was determined by atomic absorption spectrophotometry (PerkinElmer). Urinary markers are expressed as the ratio between these markers and creatinine throughout this
report.
data analysis
Statistical analysis was performed with the software package SPSSWin, Ver. 7.5 (SPSS). All results are given as the
mean value 6 SD. Differences between the groups were
investigated with the unpaired Student t-test. The Wilcoxon test was used for the evaluation of differences within
groups. Differences were considered significant at P ,0.05.
Correlations between the data obtained with different test
procedures were evaluated using linear regression.
Results
calibration curve and test characteristics
Calibration curves were constructed on 12 different days
using six calibrator solutions with OC concentrations in
the range of 0 –100 mg/L. Each calibrator was measured in
triplicate, and the mean absorbance at 450 nm (6 SD) was
expressed as a function of the OC concentration (Fig. 1).
The lower limit to detection, defined as the mean absorbance 1 3 SD for calibrator A (0 mg/L), was 4.45 mg/L
[absorbance 5 0.05 1 3 3 0.02 5 0.11]. The intra- and
interassay variation of the test was determined using
three serum pools of known OC concentrations. The
intraassay variation was calculated by expressing the SD
as a percentage of the mean concentration as calculated
Table 3. Day-to-day variation of bone markers during 2 months.
All (n 5 12)
Test
In serum
OC, mg/L (coated beads)
OC, mg/L (ELSA-Osteo)
BAP, mg/L
PTH, ng/L
In urine
DPD/creat,a mmol/mol
OHPro/creat, g/mol
CTX/creat, mg/mol
Calcium/creat, mol/mol
a
creat, creatinine.
Mean 6 SD
23.7 6 6.0
26.0 6 6.1
15.8 6 7.7
31.1 6 5.6
6.9 6 2.8
2.2 6 0.9
160.8 6 104.2
0.25 6 0.09
Men (n 5 6)
Women (n 5 6)
Day-to-day
variation, %
Mean 6 SD
Day-to-day
variation, %
10.8
9.1
11.7
29
23.1 6 6.5
25.0 6 6.4
11.9 6 3.2
28.9 6 6.1
8.3
9.6
13.3
27.5
24.1
42.3
52.4
36.3
5.2 6 2.2
1.7 6 0.5
109.3 6 56.6
0.2 6 0.06
20.7
44.5
58.7
37
Mean 6 SD
24.3 6 2.7
27.2 6 6.1
19.6 6 9.2
33.2 6 4.5
8.6 6 2.5
2.7 6 0.9
212.3 6 119.7
0.3 6 0.1
Day-to-day
variation, %
15.5
8.4
9.9
30.5
27.6
40
46.1
35.5
Clinical Chemistry 46, No. 2, 2000
from 21 replicates of each serum pool, which was repeated on 3 different days. Mean values obtained were
7.6%, 6.0%, and 4.8% for serum pools containing 10, 22,
and 83 mg/L of OC, respectively. The same serum pools
were measured in duplicate on 14 consecutive days, and
interassay variations of the means of duplicates were
calculated by expressing the SDs as percentages of the
means: 6.0%, 9.1%, and 5.9%, respectively.
To determine the linearity on dilution, we used three
human serum samples containing pathologically high OC
concentrations (from patients with renal failure). Samples
with OC concentrations well above the highest calibrator
(calibrator F; 100 mg/L) were prediluted with calibrator A
(0 mg/L) before use to give the following OC concentrations: sample A (undiluted), 97.4 6 2.2 mg/L; sample B
(prediluted twofold), 94.3 6 2.6 mg/L; and sample C
(prediluted fivefold), 75.3 6 0.7 mg/L. On 3 different
days, each sample was diluted serially and tested in
duplicate. Recoveries were calculated and expressed as
percentages of the starting values (Table 1), and it was
apparent that sample A could be diluted twofold without
significant loss of recovery. As was the case with all (n 5
6) of the microtiter plate-based kits and radioimmunoassays we have checked to date (data not shown), the
recovery declined at higher dilutions. Both prediluted
samples (B and C) showed a strong decrease of recovery
after further dilutions, which indicates that more than
twofold dilution of serum samples may lead to substantial
underestimation of the OC concentration in these samples
if the dilutions are prepared with the calibrator A (0
mg/L), which does not have a matrix sufficiently similar
to serum.
To investigate whether the antibodies used in our
assay discriminate between OCs containing different
numbers of Gla residues, full-length synthetic OCs (17 )
containing either 0 or 3 Gla residues were dissolved in
calibrator A and tested in various dilutions. Both synthetic peptides were recognized well, but the antibodies
did not differentiate between fully carboxylated and noncarboxylated OC (data not shown).
255
day-to-day and within-day variations
Twelve subjects (6 men and 6 women) were enrolled in an
experiment in which blood and urine samples were
collected at various time points during the first 24 h and
at weekly intervals during the first month, with a final
sample collection after 2 months. We measured serum
concentrations of OC (by two assays), BAP, and PTH, and
urine concentrations of DPD, CTX, OHPro, total Ca21,
and creatinine. For each subject and each variable sepa-
variations related to age and gender
In this experiment, we assembled serum samples from
apparently healthy men (n 5 60) and women (n 5 91) of
different ages and compared the new assay with commercially available test kits for OC and for BAP. The results
obtained with the two OC assays were very similar, with
all three bone formation markers slightly increased in
men 19 – 40 years of age, possibly because in this group
the peak bone mass was not yet reached (Table 2). Only in
the case of the commercial OC kit was this difference
statistically significant, however. In addition, in women
59 – 86 years of age, the bone formation markers had a
tendency to increase, which may be related to the increased bone turnover frequently seen during postmenopausal bone loss. Only in the case of BAP was this
increase statistically significant, however.
Fig. 2. Within-day variation of bone formation markers.
(Top), individual plots of OC determined with the Coated Beads Assay during
consecutive time points within 24 h; (middle), individual plots of OC determined
with the Elsa-Osteo assay; (bottom), individual plots of BAP. All curves are
presented as the concentration of each marker (mg/L). The mean values per time
point are given by the thick line; bars, SD for each time point.
256
Crăciun et al.: Coated Beads Assay for Human Serum Osteocalcin
Table 4. Bone markers in postmenopausal women before and after therapy.a,b
OC, mg/L OC
Treatment
Biphosphonates
Before
After
Hormone replacement
Before
After
Calcitonin
Before
After
CTX/creat,
mg/mol
0.61 6 0.40
0.63 6 1.14
29.8 6 54.0
17.6 6 26.2
532 6 1698
173 6 332b
6.6 6 2.3
5.6 6 2.9b
0.49 6 0.33
0.31 6 0.29b
19.8 6 26.4
10.1 6 4.8
578 6 307
210 6 144b
14.1 6 2.7
11.9 6 1.7b
0.4 6 0.3
0.4 6 0.3
13.0 6 6.6
11.7 6 5.2
801 6 459
609 6 338b
ELSA-Osteo
BAP, mg/L
28.2 6 13.1
10.3 6 3.07a
29.8 6 12.5
8.9 6 3.91a
12.9 6 4.0
7.0 6 2.06a
23.6 6 7.7
17.6 6 4.2b
28.0 6 8.7
18.5 6 5.0a
21.1 6 8.0
16.0 6 5.1b
23.8 6 8.7
19.1 6 9.7b
a,b
c
Ca/creat,c
mol/mol
DPD/creat,
mol/mol
Beads
Changes within each treatment group were analyzed with the Wilcoxon test:
creat, creatinine.
rately, the individual mean values of the seven fasting
morning samples (taken at 0900) were calculated and
used to calculate the group mean values (Table 3, columns
2, 4, and 6). Individual day-to-day CVs were expressed as
percentages of the corresponding individual mean values
and were used to calculate the mean intraindividual
variation in the group (Table 3, columns 3, 5, and 7). It
turned out that the serum markers (except PTH) were
relatively stable with time and that both the inter- and
intraindividual variations of the Coated Bead Assay were
comparable with commercial kits for bone markers. The
urinary markers, notably OHPro and CTX, had large
intra- and interindividual CVs.
The within-day variation of the various markers was
assessed using the samples obtained during the first 24 h
of the experiment (Fig. 2). The individual variation of the
bone formation markers OC (determined with both assays) and BAP are given in the different plots. No distinct
diurnal variation was found for any of the three assays. A
more pronounced diurnal pattern was observed for the bone
resorption markers, but because of the large interindividual
variation, the difference between zenith and nadir did not
reach statistical significance (data not shown).
follow-up during osteoporosis treatment
To test the ability of the Coated Beads Assay to detect
changes in serum OC concentration during therapy, three
groups of 10 postmenopausal women were followed
during their treatment with bisphosphonates, estrogen, or
calcitonin (Table 4). Changes during therapy were more
pronounced for OC than for BAP, with similar relative
a
P ,0.005;
b
P ,0.05.
changes for the two OC assays. Therapy-induced changes
were also large in the bone resorption markers, but because
of the large standard deviation of urinary markers, the
changes were not statistically significant in all cases.
correlation between bone formation markers
To investigate the correlation between the various tests
for bone formation, the data from the former experiments
were pooled. In total, the data of 181 subjects (60 men, 91
women, and the baseline measurements of 30 postmenopausal women) were used to perform linear regression.
The correlation between both OC assays was r 5 0.879 (P
,0.0001), and the standard deviation of the residuals (Syux)
was 4.66 mg/L. Regression analysis between BAP and
both OC assays produced rather poor regression coefficients and higher Syux values (Table 5).
Discussion
OC is the most abundant noncollagenous protein in bone.
It is synthesized by the osteoblasts, and after its cellular
secretion, ;80% is bound to the hydroxyapatite matrix.
The remainder is released into the blood stream, where it
is available for detection (2 ). Circulating OC is used as a
marker for bone formation, and high concentrations have
been observed in children (2 ) and in patients with high
bone turnover, as is found in Paget disease and postmenopausal osteoporosis (18, 19 ). Commercial test kits for OC
detection are based on one of two principles: radioimmunoassays or enzyme-linked immunoassays. The use of OC
Table 5. Linear regression analysisa of bone formation markers.
Dependent variable
OC (Elsa-Osteo)
BAP
BAP
Independent variable
Slope (SD)
Intercept (SD)
R
P
Syzx, mg/L
OC (coated beads)
OC (coated beads)
OC (Elsa-Osteo)
0.89 (0.036)
0.32 (0.059)
0.22 (0.061)
1.86 (0.81)
7.21 (1.32)
9.41 (1.35)
0.879
0.378
0.259
,0.0001
,0.0001
,0.0001
4.66
7.58
7.91
a
Linear regression analysis was performed according to the equation: y 5 a0(SD) 1 a1(SD) 3 x, in which y is the dependent variable and x is the independent
variable; a0 is the intercept, and a1 is the slope of the equation, both with corresponding SD. For each equation, the regression coefficient (R) and significance (P) are
given. Syux is the standard deviation of the residuals.
Clinical Chemistry 46, No. 2, 2000
as a marker for bone metabolism in routine clinical
chemistry is mainly restricted because of (a) the large
differences between the various kits and (b) the fact that
no automated test is available to date. The former problem may be related to the fact that serum OC occurs in at
least two conformations: fully carboxylated OC, which
contains three residues of the unusual amino acid Gla,
and undercarboxylated OC, which contains 0 –2 Gla residues (12, 20 ). The antibodies (either polyclonal or monoclonal) used in different kits differ from each other in their
relative affinity for fully carboxylated and undercarboxylated OC (15 ), and thus in the amounts that are detected
in the same serum sample. An additional problem is that
some kits detect only intact OC, whereas others recognize both intact OC and OC fragments. Hence, standardization of reference samples and antibodies used is
an absolute requirement before data obtained with
various kits may be compared.
From a technical point of view, a serious drawback of
existing kits is that they are test tube- or microtiter
plate-based assays that must be pipetted by hand. This is
laborious and forms a potential source of imprecision and
mistakes. Therefore, we have used and evaluated a semiautomated method for OC detection, using the antibodies
and calibrators from a commercial enzyme-linked immunosorbent assay. The test is based on the sandwich
principle, with antibodies directed against epitopes at the
NH2 and COOH termini of OC outside the Gla domain,
thus ensuring that only intact OC (both fully carboxylated
and undercarboxylated) is detected. It turned out that the
newly developed assay (provisionally designated as the
Coated Beads Assay) has a high precision and compares
well with the kit from CIS Biointernational in both withinday and day-to-day variation, as well as in patient follow-up
studies. Like other test for bone formation, the standard
deviation of the Coated Beads Assay was much lower than
that of bone resorption markers, and a good correlation was
observed with OC values determined with the CIS kit.
In conclusion, we have demonstrated that with the semiautomated assay, OC may be determined within 1 h, and
that the accuracy of the assay is at least comparable to
existing kits. The new test still requires a few pipetting
steps, but the coated beads may be used equally well in
fully automated analyzers.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
This study was supported by Hoffmann-La Roche Diagnostics (Basel, Switzerland). We thank Dr. P. Proost
(University of Leuven, Belgium) for kindly supplying us
with synthetic OC.
19.
References
1. Price PA. Role of vitamin K-dependent proteins in bone metabolism [Review]. Annu Rev Nutr 1988;8:565– 83.
2. Hauschka PV, Lian JB, Cole DEC, Gundberg CM. Osteocalcin and
20.
257
matrix Gla protein: vitamin K dependent proteins in bone [Review].
Physiol Rev 1989;69:990 –1047.
Kesterson RA, Stanley L, DeMayo F, Finegold M, Pike JW. The
human osteocalcin promotor directs bone-specific vitamin Dregulatable gene expression in transgenic mice. Mol Endocrinol
1993;7:462–7.
Ozono K, Liao J, Kerner SA, Scott RA, Pike JW. The vitamin
D-responsive element in the human osteocalcin gene. Association
with a nuclear proto-oncogene enhancer. J Biol Chem 1990;265:
21881– 8.
Luo G, Ducy P, McKee MD, Pinero GJ, Loyer E, Behringer RR,
Karsenty G. Spontaneous calcification of arteries and cartilage in
mice lacking matrix GLA protein. Nature 1997;386:78 – 81.
Power MJ, Fottrell PF. Osteocalcin: diagnostic methods and clinical applications [Review]. Crit Rev Clin Lab Sci 1991;28:287–335.
Garnero P, Grimaux M, Demiaux B, Preaudat C, Seguin P, Delmas
PD. Measurement of serum osteocalcin with a human-specific twosite immunoradiometric assay. J Bone Miner Res 1992;7:1389 –98.
Eastell R, Robins SP, Colwell T, Assiri AMA, Riggs BL, Russell
RGG. Evaluation of bone turnover in type I osteoporosis using
biochemical markers specific for both bone formation and bone
resorption. Osteoporos Int 1993;3:255– 60.
Knapen MHJ, Hamulyák K, Vermeer C. The effect of vitamin K
supplementation on circulating osteocalcin (bone Gla-protein) and
urinary calcium excretion. Ann Intern Med 1989;111:1001–5.
Knapen MHJ, Jie K-SG, Hamulyák K, Vermeer C. Vitamin K-induced
changes in markers for osteoblast activity and urinary calcium
loss. Calcif Tissue Int 1993;53:81–5.
Plantalech L, Guillaumont M, Vergnaud P, Leclercq M, Delmas PD.
Impairment of gamma carboxylation of circulating osteocalcin (bone
Gla protein) in elderly women. J Bone Miner Res 1991;6:1211– 6.
Szulc P, Chapuy M-C, Meunier PJ, Delmas PD. Serum undercarboxylated osteocalcin is a marker of the risk of hip fracture in
elderly women. J Clin Investig 1993;91:1769 –74.
Szulc P, Arlot M, Chapuy M-C, Duboeuf F, Meunier PJ, Delmas PD.
Serum undercarboxylated osteocalcin correlates with hip bone
mineral density in elderly women. J Bone Miner Res 1994;9:
1591–5.
Knapen MHJ, Nieuwenhuijzen Kruseman AC, Wouters RSME,
Vermeer C. Correlation of serum osteocalcin with bone mineral
density in women during the first 10 years after menopause.
J Bone Miner Res 1998;63:375–9.
Knapen MHJ, Eisenwiener H-G, Vermeer C. Osteocalcin detection
in aging serum and whole blood: stability of different osteocalcin
fractions. Clin Chim Acta 1996;256:151– 64.
Rosenquist C, Qvist P, Bjarnason N, Christiansen C. Measurement of a more stable region of osteocalcin in serum by ELISA
with two monoclonal antibodies. Clin Chem 1995;41:1439 – 45.
Houben RJTJ, Jin D, Stafford DW, Proost P, Ebberink RHM,
Vermeer C, Soute BAM. Osteocalcin binds tightly to the g-glutamyl
carboxylase at a site distinct from that of the other known vitamin
K-dependent proteins. Biochem J 1999;341:265–9.
Gundberg CM, Lian JB, Gallop PM, Steinberg JJ. Urinary gammacarboxyglutamic acid and serum osteocalcin as bone markers:
studies in osteoporosis and Paget’s disease. J Clin Endocrinol
Metab 1983;57:1221–5.
Kaddam IMS, Iqbal SJ, Holland S, Wong M, Manning D. Comparison of serum osteocalcin with total and bone specific alkaline
phosphatase and urinary hydroxyproline:creatinine ratio in patients with Paget’s disease of bone. Ann Clin Biochem 1994;31:
327–30.
Cairns JR, Price PA. Direct demonstration that the vitamin Kdependent bone Gla protein is incompletely g-carboxylated in
humans. J Bone Miner Res 1994;9:1989 –97.