Methods

Top Methods Results Discussion Author & Article Info References

Fifteen consecutive white women with recent-onset Graves thyrotoxicosis attending our thyroid unit were enrolled. Six women were premenopausal; nine were postmenopausal. Of these, nine presented with a 3- to 6-month history of weight loss, palpitations, and diaphoresis, and were clinically thyrotoxic. Two more women had apathetic thyrotoxicosis, and the remaining four were asymptomatic but were referred because of biochemical thyrotoxicosis. At the time of diagnosis, all patients had an elevated total serum thyroxine (serum T4 > 168 nmol/L), a suppressed serum thyroid-stimulating hormone (serum TSH < 0.05 mU/L), an elevated technetium-99m thyroid scan (>4%), and a positive test for thyroid antibodies. No patient had a history of rheumatoid arthritis, diabetes mellitus, or alcohol abuse, a family history of osteoporosis, or any other serious medical disorder. None had been treated with estrogen, thiazide diuretics, calcium, or vitamin D for at least 12 months before enrolling in the study. Antithyroid therapy (carbimazole or thiouracil) was initiated after baseline studies had been completed and was continued for 12 months to maintain a euthyroid status (serum TSH > 1 mU/L). Propranolol was only used during the first month for patients with clinically significant sympathoadrenergic symptoms. The mean age of the patients was 52 ±4 years (range, 28 to 80 years). Serum biochemical tests were done, and bone mineral density was measured at baseline and after 12 months. Fifteen controls matched for age (within 2 years) and years since the menopause (within 2 years) were selected from a group of 80 healthy white women recruited for longitudinal densitometric assessment from the suburbs surrounding the hospital.

Venous blood was drawn for biochemical and hormonal measurements before beginning antithyroid therapy at 0800 h. Serum calcium, inorganic phosphate, and alkaline phosphatase levels were determined by routine laboratory methods. Serum bone Gla-protein, a serologic marker of bone formation, was measured by radioimmunoassay as previously described [13]. Total serum thyroxine (GammaCoat total T4; Incstar Corporation; Stillwater, Minnesota), total serum triiodothyronine (Amerlex total T3; Amerlite Diagnostics; Buckinghamshire, England), and serum TSH (hTSH immunoradiometric assay; Dainabot Corporation; Tokyo, Japan) were measured using routine kits. The lower limit of detection for serum TSH in this assay was 0.05 mU/L. Undetectable serum TSH values were assigned this value for the calculation of group mean values. Plasma estradiol was measured by radioimmunoassay with an antibody obtained from Sorin Biomedica (Saluggia, Italy). Menopause was arbitrarily defined as a plasma estradiol concentration less than 50 pmol/L together with a duration of amenorrhea of at least 6 months. Technetium-99m pertechnetate thyroid uptake was calculated using a camera interfaced to a dedicated nuclear medicine computer (GE Starcam; Milwaukee, Wisconsin).

The density of the lumbar spine (anteroposterior; L2 to L4), femoral neck, the Ward triangle, and trochanter was measured by dual-energy x-ray absorptiometry (DXA) (Lunar DPX-L, Lunar Corporation; Wisconsin) [14]. The in vivo precision was 1.7% for measurements of the lumbar spine, 3.6% for the femoral neck, 3.3% for the Ward triangle, and 2.2% for the femoral trochanter. Bone mineral density measurements were expressed in g/cm (²) and as a percentage of age- and sex-matched controls. Longitudinal changes in bone mineral density measurements were expressed as percentage change per year and were calculated by subtracting the final from the initial bone mineral density and dividing this difference by the initial bone mineral density.

All results were expressed as the mean ±SE and, where appropriate, 95% CIs were cited. Data were compared using analysis of variance and by paired or unpaired Student t-tests where appropriate. Logarithmic transformations were used for data with skewed distributions. Forced-entry multiple regression analysis was used to determine independent variables associated with changes in bone mineral density measurements. Variables entered included those known to affect bone mineral density (patient's age, menopausal status, body weight, and serum calcium level) and those hypothesized to affect bone mineral density (total serum T4 and total serum T3). The initial total serum T4 and initial total serum T3 were chosen as markers of the severity of the thyrotoxicosis. Only variables with P values of less than 0.2 were retained as predictors in the final regression model. The partial regression coefficient (b), the 95% CIs, and P values are shown in the Results section.

Results

Top Methods Results Discussion Author & Article Info References

Patients with active Graves thyrotoxicosis weighed less and had higher serum calcium, serum alkaline phosphatase, and serum bone Gla-protein concentrations compared with control volunteers (Table 1). We found no statistically significant correlations between the total serum T4 levels and either the serum alkaline phosphatase or serum bone Gla-protein concentrations. Table 1 also compares the initial bone mineral density measurements in patients with active Graves thyrotoxicosis with those of the matched controls. Lumbar spine and femoral trochanter bone mineral density measurements differed between the two groups (P < 0.05). The mean lumbar spine bone mineral density was 12.7% lower and the mean femoral neck bone mineral density was 9.9% lower in patients with active Graves thyrotoxicosis than in matched controls. Five patients had lumbar spine bone mineral density measurements less than 85% of age- and sex-matched controls (Figure 1).

View larger version (13K): [in this window] [in a new window]   Figure 1. Individual bone mineral values of the lumbar spine (squares) and femoral neck (circles) before (open symbols) and after 12 months of antithyroid therapy (closed symbols) in patients with Graves thyrotoxicosis, expressed as a percentage of age- and sex-matched controls. The horizontal bars represent the mean values of the two groups.

After 3 months of antithyroid therapy, total serum T4, total serum T3, and serum TSH concentrations were in the normal reference range in all patients treated for active Graves thyrotoxicosis (data not shown) and remained within the reference range throughout the remainder of the study. After 12 months of therapy, the weight of patients with Graves thyrotoxicosis increased (P < 0.001 compared with baseline values), and their serum alkaline phosphatase and serum bone Gla-protein concentrations returned to normal (P < 0.001 compared with baseline values) (Table 1). The mean lumbar spine bone mineral increased from an initial value of 1.01 g/cm (²) to 1.07 g/cm², an increment of 6.6.% per year (CI, 3.6% to 9.6%; P < 0.001 compared with controls). Increases in femoral neck bone mineral (1.2% per year; CI, –2.1% to 4.5%; P = 0.2 compared with controls) and femoral trochanter bone mineral (3.2% per year; CI, –2.4% to 8.7%; P = 0.4 compared with controls) were also documented (Table 2). The individual changes in bone mineral measurements in patients and controls are outlined in Figure 2. In the forced-entry multiple regression analysis, the initial total serum T4 (b = 0.002; P = 0.01) was the only independent determinant of the final lumbar spine bone mineral measured at 12 months. On the other hand, the patient's age (b = 0.55; P = 0.005), menopausal status (b = –11.1;P = 0.006), and initial total serum T3 (b = 1.4; P = 0.04) were independently related to the percentage increment in lumbar spine bone mineral density after 12 months of antithyroid therapy.

View larger version (14K): [in this window] [in a new window]   Figure 2. Individual bone mineral values of the lumbar spine (squares), femoral neck (circles), Ward triangle (diamonds), and trochanter in patients with Graves thyrotoxicosis (closed symbols) and in controls (open symbols), expressed as a percentage change of bone mineral over a 12-month period.

Discussion

Top Methods Results Discussion Author & Article Info References

Our study shows that women with active Graves thyrotoxicosis have reduced lumbar spine bone mineral density measurements that increased after 12 months of antithyroid therapy. Changes in the femoral neck and femoral trochanter bone mineral density measurements, however, did not differ statistically. These data suggest that thyrotoxic bone loss may therefore be reversible.

The first case of thyrotoxic osteopathy was reported in 1891 by Von Recklinghausen [15]. Several later cross-sectional studies have shown reduced bone mineral measurements in thyrotoxic patients [1-7]. Fraser and colleagues [2] recorded a 7% reduction in forearm cortical bone using densitometry, and Bouillon and colleagues [7] recorded a 13% reduction in lumbar spine bone mineral using dual-photon absorptiometry. These data are similar to the 12.7% reduction in lumbar spine bone mineral and the 9.9% reduction in femoral neck bone mineral seen in our patients with active Graves thyrotoxicosis using dual-energy x-ray absorptiometry. Although no longitudinal studies evaluating bone mineral changes in patients with untreated thyrotoxicosis have been done, longitudinal data have been reported that describe increased bone mineral loss in patients receiving overzealous T4 replacement [5, 1621]. Ribot and colleagues [18] recorded a 5.4% reduction in lumbar spine, a 7.3% reduction in femoral trochanter, and a 7% reduction in femoral neck bone mineral in patients receiving T4 replacement for 12 months, and Krolner and colleagues [5] recorded a 8.9% reduction in lumbar spine bone mineral over a similar period.

The benefits of antithyroid therapy on the skeleton in thyrotoxicosis have recently been addressed because of the worldwide interest in the prevention of osteoporosis. Longitudinal studies done on patients successfully treated for thyrotoxicosis have produced conflicting data but, on the whole, suggest that thyrotoxic bone loss is potentially reversible [3, 5, 6, 11]. Studies reported by Bayley and colleagues [3] using neutron-activated analysis, by Toh and colleagues [6] using single-photon absorptiometry, and by Krolner and colleagues [5] using dual-photon absorptiometry, have all shown increases in bone mineral varying from 3.7% to 12.9%. In our study, using dual-energy x-ray absorptiometry, lumbar spine bone mineral increased by 6.6% during a 12-month period. Increases in bone mineral were also observed in the femoral neck (1.2%) and femoral trochanter regions (3.2%), but those changes were not statistically significant. Although these studies are only short-term, a recent long-term study by Rosen and colleagues [12] has suggested a longer lasting and beneficial effect of antithyroid therapy on bone mineral. These investigators found that 5 years after antithyroid therapy, the lumbar spine bone mineral in their patients increased 11%. More invasive studies using iliac crest bone histomorphometry have shown a complete return to normal of cortical porosity and bone turnover after antithyroid therapy [10].

Thyroid hormones affect bone cells both in vitro [10, 22, 23] and in vivo [24, 25] by stimulating osteoclastic bone resorption and increasing skeletal remodeling [10, 23, 25]. In the untreated thyrotoxic state, this may eventually lead to both cortical and cancellous bone mineral loss [11, 25] and occasionally to an increase in fracture rates [26, 27]. Noninvasive markers of bone turnover [5-9, 11, 16-21, 26] and histologic features of bone remodeling [11] have been evaluated in patients with both exogenous and endogenous thyrotoxicosis. In our study, serum calcium, serum alkaline phosphatase, and serum bone Gla-protein concentrations were significantly elevated in patients with active Graves thyrotoxicosis and returned to normal after 12 months of antithyroid therapy. No correlation was found between the total serum T4 level and either the serum alkaline phosphatase or serum bone Gla-protein concentrations. These findings are consistent with the results of others [3, 5, 6, 12] and suggest that the high bone turnover state resolves with successful antithyroid therapy. Whether the osteoclastic bone resorption occurs directly because of thyroid hormones or whether it occurs because of cytokines released by activated osteoblasts remains unresolved [22, 23].

Several factors may be incriminated in thyrotoxic bone loss [7-11]. The severity and duration of the hyperthyroid state are related to bone turnover and bone mass. We found that the severity of the thyrotoxicosis (as measured by the initial total serum T3) was independently and positively associated (b = 1.4; CI, 0.08 to 2.8; P = 0.04) with the percentage increment in lumbar spine bone mineral after 12 months of antithyroid therapy. The duration of the thyrotoxic state, however, was extremely difficult to define, especially in some of the elderly patients in whom the cardinal manifestations of their disease include an apathetic state. In these patients, prolonged undetected thyrotoxicosis may result not only in trabecular plate thinning but also in irreversible trabecular plate perforation. Hypogonadism may also contribute to the skeletal changes occurring in postmenopausal women with thyrotoxicosis, suggesting a direct synergistic effect between excess thyroid hormones and estrogen deficiency [3, 17-21].

Thyrotoxic bone loss may be reversible. Successful treatment of the thyrotoxic state with antithyroid therapy may result not only in a return to normal of bone turnover but also in an improvement in bone mineral. We found that lumbar spine bone mineral, which is composed predominantly of cancellous bone, showed the greatest increases.

Abbreviations

T3: triiodothyronine

T4: thyroxine

TSH: thyroid-stimulating hormone (thyrotropin)