| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Clinical Studies |
Departments of Endocrinology (A.S., I.J., D.F.W., J.P.M.) and Clinical Biochemistry (A.S., J.M.B., K.N., C.P.P.), St. Bartholomew’s and the Royal London Hospital, London E1 1BB, England
Address all correspondence and requests for reprints to: Dr. A. Siddiqi, Department of Metabolism and Endocrinology, Alexandra Wing, St. Bartholomew’s and the Royal London School of Medicine and Dentistry, Whitechapel, London E1 1BB, England. E-mail: ASiddiqi{at}mds.qmw.ac.uk
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Coinciding with euthyroidism at 4–8 weeks, there was a peak in b-ALP and OC and a prompt fall into the normal range in Udpd and Sdpd, but not Spyr, levels. Mean b-ALP continued to be raised at week 52 when it was inversely correlated with BMD. Mean BMD rose approximately 6%, P < 0.01, over 1 yr. Coupling indices were calculated as a measure of bone balance and, at diagnosis, was [minus4.26 in favor of bone resorption and rose with treatment in favor of bone formation: weeks 2: -0.23; 4: +4.01; 8: +4.37; 12: +4.44; 24: +2.32; and 52: +1.56.
Bone turnover is balanced within 2 weeks of starting treatment for thyrotoxicosis. Udpd accurately indicates thyrotoxic bone resorption. Serum b-ALP indicates continuing bone formation and, at 1 yr, may provide a marker for low BMD. OC, Sdpd, and Spyr are less sensitive in documenting bone remodeling during treatment of thyrotoxicosis.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Retrospective and cross-sectional studies suggest that a significant proportion of patients with a history of thyrotoxicosis continue to be at risk of reduced bone mineral density (BMD) up to 18 yr later (7). Longitudinal studies have been few, and estimation of BMD has been hindered by imprecise and indirect techniques. Of the four studies reported, one has used calcium bone index (8), and the remaining three, dual photon (153Gd) absorptiometry for either all (9, 10) or at least one (11) of the serial estimations of bone mineral content (BMC). Although it is established that thyrotoxicosis leads to reduced BMC or density, the issues that remain unclear are: 1) the capacity of antithyroid medication to restore bone mass; and 2) the prediction of which individuals are more likely to suffer a long-term deficit in BMD.
Changes in BMD occur only slowly, making it difficult to judge the benefits, if any, of antithyroid medication. Early identification of those individuals at risk of long-term reduction in BMD caused by thyrotoxicosis may enable early therapeutic intervention. A number of biochemical markers are thought to reflect the rapid bone turnover in thyrotoxicosis, although their predictive value remains to be established.
Such markers include serum ALP, OC, and serum and urinary collagen cross-links. Previous studies of these markers in thyrotoxicosis have been either cross-sectional (12, 13, 14) or limited by cumbersome techniques (15, 16) and absence of simultaneous BMD measurements (17, 18). Total ALP and OC are raised in approximately 30% and 65–90% of thyrotoxic patients, respectively. Methodologies such as polyacrylamide and agarose gel electrophoresis, heat inactivation, lectin reactivity, and neuraminidase susceptibility have been used to differentiate skeletal and liver ALP isoenzymes; and they exhibit poor resolution, require sample pretreatment, and offer indirect quantitation. Correlation of OC with b-ALP and FT3 have produced conflicting results and OC’s precise relationship to b-ALP is unclear.
Recently, novel markers related to collagen molecules have been established as specific markers of bone resorption. These include the nonreducible cross-links of mature collagen, serum and urinary pyridinoline and deoxypyridinoline, and serum carboxy-terminal telopeptide of Type 1 collagen (ICTP). Urinary pyridinoline (Upyr) is raised in 99% of thyrotoxic patients but is less specific for bone than urinary deoxypyridinoline (Udpd). The usefulness of Udpd and ICTP as markers of bone turnover has been assessed in Cushing’s Syndrome (19), neoplastic disease (20), GH deficiency (21), postmenopausal status (22), anticonvulsants (23), aging (24), and hyperparathyroidism (25), but data on thyrotoxicosis is limited to cross-sectional studies (23, 26). Serum pyridinoline and deoxypyridinoline to date have not been studied in hyperthyroidism, although their correlation with OC, ALP, Upyr, and Udpd has been established in renal osteodystrophy (27).
The high prevalence of thyrotoxicosis and its potential effect on long-term BMD necessitate longitudinal studies to determine whether the behavior of sensitive biochemical markers during the first year of treatment can predict the ultimate degree of restoration of BMD.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Seventeen patients (14 premenopausal women 28–48 yr old; 3 men
27–42 yr old) with Graves’ disease were studied longitudinally.
Informed consent was obtained, and the study was approved by the East
London and City Health Authority Ethical Committee. Diagnosis of
thyrotoxicosis was made on the basis of clinical examination, elevated
serum free thyroxine (FT4) or total T3, and
suppressed TSH (<0.01 mU/L), and etiology was determined by
extrathyroidal manifestations of Graves’ disease or positive thyroid
antiperoxidase antibodies. Duration of disease was estimated from the
time of onset of symptoms. Exclusion criteria included previous thyroid
disease, amenorrhea, postmenopausal status, vegetarian diet, cigarette
smoking, and any medical conditions known to affect bone metabolism.
All patients were treated with 30 mg carbimazole daily for 6 weeks and
20 mg daily for 4 weeks, followed by titration of the dose of
carbimazole according to the biochemical thyroid status. During
follow-up, 1 woman (patient 4) received radioiodine at week 40 and 1
man (patient 13) was treated with block-and-replace treatment using
carbimazole and thyroxine. Patient 13 was additionally treated with
propranolol (120 mg daily) for the initial 8 weeks. One woman (patient
9) had a relapse at week 12 but was euthyroid again by week 24, and 1
woman (patient 14) remained thyrotoxic from poor compliance with
therapy. The others remained euthyroid on carbimazole. The clinical
characteristics of these patients are shown in Table 1
.
|
Blood was taken just before (week zero) and 2, 4, 8, 12, 24, and 52 weeks after commencement of carbimazole. The following serum measurements were made: FT4, T3, TSH, b-ALP, OC, and serum pyridinoline (Spyr) and deoxypyridinoline (Sdpd). On each occasion, 24-h urine collections were made also for Udpd and creatinine (Cr).
b-ALP was assayed using a solid-phase 2-site immunoradiometric assay using mouse monoclonal Ig G to skeletal ALP (Tandem-R Ostase assay kit, HybriTech Europe, Liege, Belgium). The intra- and interassay coefficients of variation (CVs) were 3.7–6.7% and 7.0–8.1%, respectively. The reference range, obtained from a population of 261 females and 217 males 20–79 yr old, was 8–16.6 µg/L. Cross-reaction with total ALP was 16.5%.
Serum OC is subject to proteolysis and was thus stored in the presence of the protease inhibitor, Trasylol, before measurement by competitive immunoassay using monoclonal antibody directed against intact bovine OC (Metra Biosystems, Mountainview, CA). The intra- and interassay CVs were 4.8–8.0% and 4.8–7.6%, respectively. The reference value, obtained from a population of 79 premenopausal females and 61 males over 25 yr old, was 3.8–10.0 ng/mL.
Spyr and Sdpd were assayed by high-performance liquid chromatography after acid hydrolysis (28). The intraassay variations for Spyr and Sdpd were 7.8% and 9.9%, respectively. Interassay variations were 16.2% and 30%, respectively. The reference value obtained from a population of 153 males and females was 1.89–4.85 nmol/L and 0.25–1.64 nmol/L, respectively.
Udpd was assayed by competitive immunoassay using a monoclonal antibody with selective, high affinity against free deoxypyridinoline and with negligible binding to Dpd peptides and free or bound pyridinoline (Metra Biosystems). The intra- and interassay CVs were 3.6–9.5% and 6.3–10.3%, respectively. The reference range, obtained from 281 premenopausal females and 86 males, was 2.5–6.0 nmol/L Dpd/mM Cr.
Thyroid function tests were assayed using the Immulite autoanalyzer (Euro/DPC Limited, Gwynedd, UK). Both FT4 and T3 were measured using a solid-phase chemiluminescent immunoassay and TSH with a solid-phase two-site chemiluminescent assay. The intra- and interassay CVs were less than 4%. The reference range was 8.8–24 pmol/L for FT4, 0.8–2.5 nmol/L for T3, and 0.4–4.0 mU/L for TSH .
BMD of the spine (lumbar vertebrae 2–4) and femur [femoral neck (FN) and trochanter] was measured using dual-energy x-ray absorptiometry (DEXA, lunar DPX, Madison, WI) within 4 weeks of both starting and finishing the study. The intra- and interassay CVs were 0.4% and 1.1%, respectively. BMD (g per cm2), t-scores (units of SD from peak adult mass) and z-scores (units of SD from an age-, sex-, height- and weight-adjusted mean) were reported for each subject. Data was analyzed with DXA software 3.6, Madison, WI. Because our patients were near their peak adult bone mass, there was no significant difference in correlations performed using BMD reported as either t-score or z-score. All calculations in this study subsequently refer to z-scores.
Statistics
All results are expressed as mean ± SEM. Data were compared using ANOVA followed by a Fisher’s test. P < 0.05 was regarded as statistically significant. To evaluate the performance and estimate the discriminatory power of the five biochemical parameters, we calculated the z-scores in relation to reference values established in our laboratory. To investigate bone balance in treated and untreated thyrotoxic patients, we calculated the coupling index [CI, (29)] based on the formula: z-score of formation markers (b-ALP + OC) minus z-score of resorption markers (Udpd + Spyr). Only two of the three resorption markers were used in this calculation to match the number of formation markers. Sdpd was not used because it had a low z-score throughout the study, implying poorer discrimination power compared with Udpd and Spyr. Simple regression analyses were performed between each marker and thyroid hormone concentration, each marker and BMD, and between markers.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Mean BMD (Table 2) was similar in the lumbar spine
(LS) and FN; z-scores: LS 0.176 ± 0.203 and FN
0.09 ± 0.18, P > 0.05]. LS BMD was lower than
-0.5 SD in three patients and lower than -1
SD in one patient, whereas FN z-score was less
than -0.5 SD in five and less than -1 SD in
two patients when compared with age-, sex-, and height-matched
controls. No patient fulfilled the current World Health Organization
definition of osteoporosis, i.e. a t-score of
less than -2.5.
|
BMD after 1 yr of treatment of thyrotoxicosis
After treatment with carbimazole, BMD increased significantly in
16 patients (z-score: LS 0.18 ± 0.20 to 0.38 ±
0.21, P < 0.003; FN 0.09 ± 0.18 to 0.407 ±
0.16, P < 0.01) (Fig. 1). There was no
difference if patient 13, receiving ß-blockers, was excluded. In
patient 14, who remained thyrotoxic throughout the study,
z-score decreased in the LS from -0.34 to -0.94 and, in
the FN, from 0.69 to -0.40. If this patient were excluded from the
calculations, the mean percentage rise in BMD, compared with
pretreatment values, was similar (P > 0.05) in the LS
(5.71 ± 1.23%) and at the FN (6.27 ± 1.68%). No patient
had a BMD lower than -1 SD in either region, although
z-score in the LS remained below -0.5 SD in two
patients and in the FN in one patient.
|
)Serum levels of specific markers of osteoblast function were raised in over half the patients. Mean values for b-ALP and OC were elevated above the assay mean in 10 of 17 patients by 2.32 ± 0.43- and 2.20 ± 0.31-fold, respectively, and were significantly correlated (r = 0.5; P < 0.05). In comparison, the elevation in total ALP was lower at 1.20 ± 0.16-fold, although there was significant positive correlation with b-ALP (r = 0.95, P < 0.001). Of the markers of bone resorption, mean Udpd was elevated in all patients by 3.49 ± 0.45-fold to 15.7 ± 2.04 nmol/L Dpd/mM Cr, mean Sdpd was elevated in 8 of 13 patients (in whom it was measured) by 1.07 ± 0.14-fold to 1.03 ± 0.14 nmol/L, and Spyr (in all but 1 of 14 patients) by 2.26 ± 0.21-fold to 7.63 ± 0.72 nmol/L.
|
Biochemical parameters of bone metabolism after treatment of thyrotoxicosis
We determined the above mentioned parameters at regular intervals
during the first year of treatment of thyrotoxicosis (Figs. 2, 3, 4, and 5). Patient 14, who remained thyrotoxic, was excluded
from all analyses subsequent to week 2. The remaining patients achieved
biochemical euthyroidism between weeks 4 and 8. Patient 9 had a relapse
of thyrotoxicosis between weeks 14 and 24, and measurements from week
24 consequently were omitted from relevant analyses.
|
|
|
|
Also coinciding with the attainment of euthyroidism was the fall into the normal range of Udpd levels in all 16 patients by week 8 (5.36 ± 0.47 nmol/L Dpd/mM Cr, P < 0.001). Sdpd and Spyr fell in parallel. At euthyroidism, mean Sdpd was within the normal range (0.65 ± 0.13 nmol/L), although 2 patients continued to have raised levels for a further 4 weeks. Mean Spyr did not fall to normal till week 12 (3.2 ± 0.44 nmol/L), and 4 patients continued to have raised levels till week 24.
At 1 yr, lumbar BMD was significantly negatively correlated with b-ALP (r = -0.82, P < 0.001). Patients were stratified according to the serum b-ALP at week 52 into those with levels either below (Group 1, n = 8) or above (Group 2, n = 8) the assay mean. Comparison of LS z-scores in patients in these 2 groups revealed a significant difference (Group 1: 0.974 ± 0.17 vs. Group 2: -0.097 ± 0.29, P < 0.01). FN BMD (z-score) also was inversely correlated with b-ALP (r = 0.63, P < 0.01). Similar stratification of FN z-scores according to b-ALP revealed a significant difference: (Group 1: 0.779 ± 0.16 vs. Group 2: 0.035 ± 0.19, P < 0.05). Elevation in serum b-ALP, therefore, seemed to identify patients likely to have low BMD (z-score). BMD was not correlated with the other osteoblast marker, OC.
Relationship between markers of bone metabolism
The results of markers expressed as z-scores are shown
in Table 4. Before treatment, there were greater
increases in the resorption markers, Udpd and Spyr, than in the
formation markers, b-ALP and OC. Of the resorption markers, Udpd had
the highest z-score and, of the formation markers, b-ALP.
After treatment, whereas the z-scores of the resorption
markers rapidly fell, b-ALP and OC z-scores rose further to
peak at 2 and 4 weeks, respectively.
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Normalizing data with the calculation of z-scores allowed us to determine the relationships between the individual parameters. b-ALP and OC, both markers of mature osteoblast function, had pretreatment z-scores that were raised to a similar degree and can be said to be equivalent in indicating thyrotoxic induced osteoblastic activity. Interestingly, both b-ALP and OC rose after treatment. This has not been demonstrated so convincingly before and may imply increase in osteoid seam width, which occurs as mineralization lag time (the time interval between formation and subsequent mineralization of osteoid) lengthens with resolution of thyrotoxicosis (6). Similar rises after treatment have been observed in OC and phosphate after surgery for Cushing’s syndrome and primary hyperparathyroidism, respectively. Subsequent changes in the serum levels of b-ALP and OC were not similar during the course of treatment. OC fell within the reference range in all but one patient by 24 weeks, whereas 50% of patients continued to have raised b-ALP at the end of 1 yr of treatment. This discrepancy probably reflects distinct functions of the two markers, although the precise nature of these is unknown.
Alternative parameters such as PICP, a cleavage product of the carboxyterminal extension released during bone formation, and procollagen type III N-terminal propeptide, a marker for soft tissue growth, have been studied in postmenopausal women, in patients with vertebral osteoporosis (22), and in patients with Cushing’s syndro (19) but seem not to be superior to b-ALP. Indeed, as a biochemical predictor of growth in children, b-ALP was found to be superior to both PICP and procollagen type III (30). In these same studies, when markers of bone resorption have been compared, ICTP has been found to be less sensitive than both Udpd and Upyr. Urinary galactosyl hydroxylysine, initially thought a promising marker of bone collagen degradation (31), is now believed to be less sensitive than both hydroxyproline and Udpd. Its value in thyroid disease remains to be established.
Of all three markers of bone resorption, Udpd had the best performance and highest z-score. Ongoing increased bone resorption before treatment was reflected in elevated Udpd levels in 100% of patients, and cessation of bone resorption upon attainment of euthyroidism by the normalization of Udpd levels at 4–8 weeks. In contrast, pretreatment Sdpd indicated increased bone resorption in 62% of patients, and the low mean z-score did not reflect the ongoing high resorptive state. The elevation of Spyr in a higher percentage of patients (93%) and the slower normalization of mean levels (at 12 weeks) reflects, as with Upyr, the sensitive but nonspecific nature of this marker. The strong correlation between the resorption markers is not surprising because Spyr is an analogue, and Sdpd the serum equivalent, of Udpd. The lack of correlation of these markers with those of bone formation may reflect both the imbalance in bone turnover and the lag phase between bone resorption and bone formation. The correlation between initial FT4 and Udpd and the prompt fall in levels of Udpd implies the possibility of a direct action of thyroid hormone on bone resorption. A similar correlation might be expected with markers of bone formation as thyroid hormones directly stimulate osteoblasts. We did not find such a relationship, in contrast to previous work (18). Such lack of consistency between studies may reflect circulating thyroid hormone levels such that, after a certain critical degree of stimulation by thyroid hormones, these markers are induced maximally and are no longer correlated with T3 and FT4.
We adapted the formula proposed by Eastell (29) to determine the CI. The positive values calculated by this formula indicate net bone turnover in favor of bone formation and negative net bone turnover in favor of bone resorption. Our data show that bone resorption, which initially is two times higher than bone formation, falls rapidly, such that bone turnover is balanced within 2 weeks of starting carbimazole and is in favor of bone formation thereafter. Stratification, according to serum b-ALP at 1 yr, revealed that patients in Group 1 with predominant bone formation (CI: + 3.33) had significantly lower BMD (z-score) than patients in Group 2 with balanced turnover (CI: - 0.09). The inverse correlation between b-ALP and BMD must be interpreted with caution because b-ALP largely reflects surface skeletal activity of bone and not bone mass. Nevertheless, the widespread use of b-ALP in laboratories and its superior performance as an osteoblastic marker in thyrotoxic bone disease make it a useful biochemical parameter to indicate those individuals likely to have poor restoration of BMD.
Thirty-five percent (6 of 17 patients) of our patients had a BMD (z-score) at least -0.5 SD below that expected for their age and sex, 3 had z-scores less than -1.0, and no patient had a z-score less than -2.0. In keeping with the concept of gradient of risk (32), in which the lower the z-score, the greater the risk of fracture, a World Health Organization report defines a t-score of less than -1 as low bone mass and of less than -2.5 as osteoporosis. Thus, none of our patients had osteoporosis. Rosen and Adler (11) reported similar results with a pretreatment mean z-score of -0.52, whereas Krolner (10) reported pretreatment BMD assessed by single- or dual-photon absorptiometry as 2–17% lower compared with controls. These two authors found a significant median gain in BMC of 3.7% by 1 yr and, in BMD, of 11% at 5 yr, respectively, with final estimated bone mass remaining lower, although no longer significantly different, than controls. Both studies included a heterogeneous group of thyrotoxic patients with a variety of diagnoses and receiving a variety of treatments. Krolner did not clarify the ages of his female patients, whereas Rosen included 40% postmenopausal women. Toh (9), in contrast, studied radial BMC in male patients over 2 yr of treatment and found that BMC was lower at 1 yr after treatment, rose negligibly by 2 yr, and remained significantly different from controls. Using DEXA scanning, we have confirmed that BMD rises by approximately 6% overall, after 1 yr of treatment for thyrotoxicosis, and that BMD of the proximal femur, which represents cortical bone and in which microscopic changes are more prominent, is no different from that of lumbar bone, in which trabecular bone predominates. We evaluated the changes in BMD with treatment from the percentage change in the raw data. For a precision error of 1.1%, as in our DEXA measurements, the least significant change in two measurements of the AP spine is 3.1%, using the formula 2 x 2 x precision error (33), and our overall gain in BMD of approximately 6%, therefore, represents an increase in BMD after treatment of thyrotoxicosis.
Because the values for bone mass in patients with fractures overlap substantially with the values in those with no fractures, it has been argued that measuring bone mass is not helpful. However, a number of prospective studies (34, 35, 36) have shown that low bone mass is an important risk factor for fractures, and in most studies, a decrease of 1 SD in bone mass was associated with an increase of 50–100% in the incidence of fracture.
In conclusion, we have shown that both trabecular and cortical BMD significantly increase after 1 yr of antithyroid treatment. b-ALP at 1 yr seems to be associated with a lower BMD and, as such, may identify those individuals who may benefit from additional intervention against osteoporosis. Larger studies, comparing BMD and b-ALP at 1 yr after treatment of thyrotoxicosis, are needed to establish the precise role of b-ALP in predicting low BMD. Cessation of increased bone resorption is indicated highly accurately by Udpd. OC and serum collagen cross-links are less sensitive in the documentation of bone remodeling during the treatment of thyrotoxicosis.
|
Acknowledgments |
|---|
Received July 30, 1996.
Revised October 28, 1996.
Accepted November 19, 1996.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  S. J. Hoffman, J. Vasko-Moser, W. H. Miller, M. W. Lark, M. Gowen, and G. Stroup Rapid Inhibition of Thyroxine-Induced Bone Resorption in the Rat by an Orally Active Vitronectin Receptor Antagonist J. Pharmacol. Exp. Ther., July 1, 2002; 302(1): 205 - 211. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Siddiqi, M. P. Parsons, J. L. Lewis, J. P. Monson, G. R. Williams, and J. M. Burrin TR Expression and Function in Human Bone Marrow Stromal and Osteoblast-Like Cells J. Clin. Endocrinol. Metab., February 1, 2002; 87(2): 906 - 914. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. W. Vesper, L. M. Demers, R. Eastell, P. Garnero, M. Kleerekoper, S. P. Robins, A. K. Srivastava, G. R. Warnick, N. B. Watts, and G. L. Myers Assessment and Recommendations on Factors Contributing to Preanalytical Variability of Urinary Pyridinoline and Deoxypyridinoline Clin. Chem., February 1, 2002; 48(2): 220 - 235. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Pantazi and P. D. Papapetrou Changes in Parameters of Bone and Mineral Metabolism during Therapy for Hyperthyroidism J. Clin. Endocrinol. Metab., March 1, 2000; 85(3): 1099 - 1106. [Abstract] [Full Text]
|
|
|
|
|
|
A. Siddiqi, J. P. Monson, D. F. Wood, G. M. Besser, and J. M. Burrin Serum Cytokines in Thyrotoxicosis J. Clin. Endocrinol. Metab., February 1, 1999; 84(2): 435 - 439. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
) and osteocalcin (ng/mL, •) with time (weeks)
following treatment of thyrotoxicosis in 16 patients. Normal range
b-ALP 8.0–16.6 µg/L; osteocalcin 3.8–10.0 ng/mL.
