Presented at the 22nd Annual Meeting of the European Thyroid Association, Vienna August 27–September 1, 1994, and at the 68th Annual Meeting of the American Thyroid Association, Chicago September 27–October 1, 1994.
We measured bone mineral density (BMD) (lumbar spine, femoral neck, Ward's triangle, and trochanter) in 34 men given suppressive doses of levothyroxine (L-T4) for a mean of 10.2 years. Indications for treatment were nontoxic goiter (n = 5) or thyroidectomy for differentiated thyroid cancer (n = 6) or nontoxic goiter (n = 3). Patients were followed at our institution and treated with the minimal amount of L-T4 able to suppress thyroid-stimulating hormone (TSH). At the time of evaluation, free T3 was normal in all cases, whereas free T4 was increased in 14 men (41.2%). The mean daily dose of L-T4 was 172 ± 6 μg, and the cumulative dose of LT4 was 673 ± 71 mg. We found no significant difference between patients and age- and weight-matched controls in BMD (g/cm2) at any site of measurement (lumbar spine 1.144 ± 0.12 vs. 1.168 ± 0.15; femoral neck 0.979 ± 0.13 vs. 1.001 ± 0.13; Ward's triangle 0.854 ± 0.17 vs. 0.887 ± 0.15; and trocanther 0.852 ± 0.13 vs. 0.861 ± 0.13). BMD was not correlated with the duration of therapy, cumulative or mean daily dose of L-T4, serum levels of free T4, free T3, osteocalcin, and bone alkaline phosphatase. Serum calcium and osteocalcin were slightly but significantly elevated in patients compared with controls, whereas there was no difference in intact parathyroid hormone, bone alkaline phosphatase, and sex hormone-binding globulin (marker of thyroid hormone action). Our data suggest that L-T4 suppressive therapy, if carefully carried out and monitored, using the smallest dose necessary to suppress TSH secretion, has no significant effects on bone metabolism and bone mass in men.
It is well established that thyrotoxicosis is associated with increased bone turnover.1 The resorption rate exceeds the formation rate, resulting in bone loss.2 Indeed, several cross-sectional studies have found a reduced bone density in hyperthyroid patients,3–5 in whom an increased risk of fracture has also been demonstrated.6,7
The question of whether long-term thyroid hormone administration is associated with bone loss is still debated.8–10 In the last 10 years, several authors have addressed this issue in women. Most of the earlier studies have indicated that doses of thyroid hormones able to suppress thyroid-stimulating hormone (TSH) secretion (suppressive doses) may cause bone loss, particularly in postmenopausal women11–14; on the contrary, more recent studies, including our own, have failed to demonstrate any detrimental effect.15–18
Emerging evidence indicates that male osteoporosis also represents an important health problem.19 Although thyroid diseases are more common in females, a significant number of male patients are also submitted to treatment with thyroid hormones. Only limited information is available on the effect of thyroid hormone administration on bone metabolism in men.
In the present study, we evaluated bone mineral density (BMD) and several parameters of calcium metabolism in a group of men chronically treated with suppressive doses of thyroid hormones. Treatment was carefully monitored to avoid even mild thyrotoxicosis. We found no evidence of decreased bone density at any site, indicating that in men, as well as in premenopausal women, carefully carried out levothyroxine (L-T4) treatment, even in suppressive doses, is devoid of any significant effect on bone mass and metabolism.
MATERIALS AND METHODS
Thirty-four men treated for more than 4 years (up to 21 years) with suppressive doses of thyroid hormones (thyroid extracts up to 1980 and subsequently L-T4) were included in the study. Indications for treatment were nontoxic goiter (n = 5) or near total or subtotal thyroidectomy for differentiated thyroid cancer (n = 26) or nontoxic goiter (n = 3), respectively. The rationale for the use of L-T4 suppressive therapy in these conditions has been discussed elsewhere.20 All patients were seen at least annually at our institution to titrate L-T4 doses in order to use in each subject the minimal dose of L-T4 able to suppress TSH. For this purpose, we attempted to reduce the daily dose of L-T4 even in those patients who had suppressed TSH and normal thyroid hormone levels.
Thirty-four men matched for age and body weight were used as controls. No patient or control subject had a history of metabolic disorders or was taking other medications known to affect bone metabolism. All subjects gave informed consent, and the study was approved by the ethical committee of the University of Pisa.
Venous blood samples were collected between 0800–0900 h, and serum was stored at −20°C until use. Serum FT4 and FT3 were measured by the Liso-Phase Kits (Laboratory Bouty S.p.A., Milan, Italy), following chromatographic separation of the hormone by Sephadex LH-20 chromatography.21 Free thyroid hormone levels measured by this method compare extremely well with the results obtained by the equilibrium dialysis method.21,22 Before 1982, total T4 and T3 were measured, and the free FT4 and FT3 indexes (FT4I and FT3I) were calculated multiplying the serum T4 and T3 levels by the percent T3 resin uptake.
Serum TSH was measured up to 1985 by radioimmunoassay (RIA) using different commercial kits and subsequently by immunoradiometric assays. In the last 3 years, we used the AutoDELFIA hTSH Ultra Kit (Wallac Oy, Turku, Finland) having a detection limit of 0.01 μU/ml. Before 1985, the individual dose of L-T4 was adjusted to have no response of TSH to intravenous injection of 200 μg of TRH; later the finding of undetectable basal TSH was sufficient to define suppression of TSH secretion.
Serum 1-84 parathyroid hormone (PTH) (Nichols Institute, San Juan, CA, U.S.A.) and bone-specific alkaline phosphatase (B-ALP) (Hybritech, San Diego, CA, U.S.A.) were measured by immunoradiometric assays; osteocalcin by a human radioimmunoassay (Nichols Institute); carboxy-terminal cross-linked telopeptide of type I collagen (ICTP) by RIA (Farmos Diagnostica, Turku, Finland); and serum calcium by the o-cresolphthalein reaction (Merck, Darmastad, Germany). As an index of the peripheral effect of thyroid hormones,23 we also measured serum sex hormone-binding globulin (SHBG) by an immunoradiometric assay (Orion Diagnostica, Espoo, Finland). Normal values for men in our laboratory were as follows: FT4 8.4–21.3 pmol/l; FT3 3.9–8.5 pmol/l; TSH 0.4–3.8 mU/l; serum calcium 2.2–2.6 pmol/l; PTH 11.1–51.6 ng/l; osteocalcin 1.0–9.3 μg/l; B-ALP 1.4–15.3 μg/l; ICTP 1.8–5.0 μg/l; SHBG 27–92 nmol/l.
Bone density measurements
BMD values were measured by dual-energy X-ray absorptiometry (DXA) with a dual-photon X-ray system (Lunar DPX, Lunar Corporation, WI, U.S.A.) at the lumbar spine (anterior-posterior, L2–L4), femoral neck, Ward's triangle, and trochanter. BMD was expressed as g/cm2. The in vivo precision was 1.1% for measurement of the lumbar spine, 1.2% for femoral neck, 1.6% for Ward's triangle, and 2.1% for trochanter.17 A Z score was calculated for each bone density measurement from the mean and SD of the control group [Z score = (patient's value − control mean)/control SD].
Unless otherwise indicated, all results were expressed as mean ± SE. Statistical analysis was performed using the Student's t-test and the Sperman's rank correlation test. Linear regressions were performed using FT3, FT4, SHBG, osteocalcin, B-ALP, duration of L-T4 therapy, and cumulative dose of L-T4 as predictor of the lumbar spine, femoral neck, Ward's triangle, and trochanter BMDs.
Clinical and biochemical data
Selected clinical data of the 34 patients included in this study are shown in Table 1. The compliance with L-T4 therapy could not be entirely guaranteed, although significant differences from the prescribed treatment schedule could be excluded on the basis of the periodic assessment of thyroid function tests.
Table Table 1. Selected Clinical Data in Men Chronically Treated with Suppressive Dosesof L-T4 and Controls*
The data reported in Table 2 refer to the time of BMD measurement. Serum FT3 levels were normal in all patients, but significantly higher compared with controls (p = 0.04). On the contrary, 14 patients (41.2%) had FT4 values above the upper limit of the normal range. Mean FT4 was significantly higher in patients compared with controls (p < 0.001). When all the available data on thyroid hormone measurements were evaluated (268 determinations each of FT4 or FT4I and FT3 or FT3I), abnormally elevated levels of FT4 or FT4I were found in 120 of 268 (44.8%) determinations; nine (26.5%) patients had increased values more than half of the time they were followed. Conversely, FT3 or FT3I were increased in only 39 of 268 (14.6%) estimations; in these patients, the daily dose of L-T4 was promptly reduced. The latter finding was more frequently observed in patients with longer follow-up, in whom thyroid extracts instead of L-T4 were initially used.
Table Table 2. Serum Biochemical and Hormonal Data in L-T4–Treated Patients and Controls*
TSH levels were undetectable or unresponsive to TRH in almost all determinations. At the time of BMD measurement, TSH values were undetectable in 26 patients, 0.1 mU/l in 6, and 0.2 mU/l in 2.
Osteocalcin was normal in all but one patient; however, the mean value was significantly higher compared with controls (6.2 ± 0.4 μg/l vs. 4.1 ± 0.3 μg/l, p < 0.001). Serum concentration of total serum calcium was slightly elevated, whereas concentration of intact PTH, B-ALP, ICTP, and SHBG did not differ between patients and controls. No correlation was found between osteocalcin, B-ALP, ICTP, SHBG, and FT3 or FT4 (data not shown).
Bone density measurements
As shown in Table 3, there was no significant difference between patients and age-matched controls in mean BMD at any site of measurements. Individual data, expressed as Z score, are shown in Fig. 1. Analysis of data from the group of patients and controls again showed no difference; mean patient Z scores were −0.16 ± 0.14 for lumbar spine, −0.17 ± 0.18 for femoral neck, −0.22 ± 0.20 for Ward's triangle, and −0.07 ± 0.17 for trocanther.
Table Table 3. BMD (g/cm2) of the Lumbar Spine, Femoral Neck, Ward's Triangle and Trocanther in L-T4–Treated Patients and Controls*
The BMD data were further analysed according to the levels of FT4 or FT4I during the follow up. BMD values did not differ in men with increased FT4 or FT4I levels more than half of the time they were followed (n = 9) compared with the remaining patients (n = 25) whose hormonal values were normal in more than half determinations (data not shown). Despite the higher mean FT4 concentration in the former group (24.6 ± 1.3 pmol/l vs. 18.7 ± 1.0 pmol/l, p = 0.003), none of the clinical and biochemical parameters evaluated in this study was significantly different in these two subgroups of patients, with the exception of SHBG that was unexpectedly higher in the group of patients with lower FT4 (27.0 ± 3.7 vs. 39.2 ± 3.3, p = 0.04).
The BMD at any site was not correlated with duration of L-T4 therapy, mean daily dose of L-T4, cumulative L-T4 intake, nor with serum levels of FT4, FT3, osteocalcin, and B-Alp.
Thyroid hormone preparations are widely used for treatment of patients with hypothyroidism and nodular goiter or to prevent recurrence after surgery for goiter or thyroid cancer.20,24–27
In recent years, clinicians have been concerned that treatment with thyroid hormones in doses able to suppress TSH secretion (suppressive doses) may result in potentially serious health problems at the cardiac and bone levels.9,20,24 Recent studies using sophisticated methods of evaluation have shown that there are minimal but definite cardiac effects of L-T4 suppressive therapy,28,29 even though it remains to be established whether these changes cause detrimental effects.30
The potentially deleterious effect of L-T4 therapy on bone has been extensively investigated, especially in women.8 Although there is no obvious consensus, the current general opinion shared by us and substantiated by a meta-analysis of Faber and Galloe31 is that in premenopausal women L-T4 treatment, even in suppressive doses, is not associated with significant effect on bone mass. However, in postmenopausal women, most studies have indicated that L-T4 treatment may contribute to osteopenia. It is worth noting that concomitant estrogen therapy may counteract T4-induced bone loss, even at suppressive doses.32
Despite the controversy on the possible osteopenic effect of L-T4 suppressive therapy, at present there is no evidence of an increased rate of osteoporotic fracture in L-T4–treated patients, provided that patients with a previous history of hyperthyroidism are excluded.33,34 This is the sole important issue; indeed, a decrease of bone mass by itself is meaningless unless it is accompanied by an increased risk of fracture.
Although thyroid disorders are definitely more common in women, a substantial proportion of male patients suffer from thyroid diseases and may be submitted to long-term treatment with thyroid hormones. This finding raises the question of whether male patients treated by L-T4 might be at risk of osteoporosis. This issue has been evaluated only in a small number of patients16,35–37 and no conclusions can be drawn. Two authors have reported no difference in bone mass compared with controls,16,36 one a significant increase at the lumbar spine,37 and one a decrease at the mid- and distal radius.36
In the present study, we found no evidence of bone loss at any site of measurement in a relatively large group of men submitted to long-term treatment with suppressive doses of L-T4. Total serum calcium and osteocalcin, but not B-ALP, were slightly but significantly elevated. The meaning of this finding is uncertain. Indeed a slight increase of circulating levels of thyroid hormone, and particularly FT3, might stimulate calcium mobilization from bone and increase osteoblastic activity10; however, no correlation was observed between circulating levels of FT4 and FT3 and calcium or osteocalcin concentrations.
With the exception of total alkaline phosphatase, whose levels have been found to be comparable between L-T4–treated patients and controls,16,36 no data on other markers of bone turnover are available in men. Osteocalcin and other markers of bone turnover have been measured in most studies evaluating patients under L-T4 suppressive therapy in women. On the one hand, in premenopausal women, several authors have reported that serum osteocalcin and urinary excretion of collagen-derived pyridinium cross-links levels did not vary significantly in patients that had been taking L-T4 sufficient to suppress TSH compared with controls.38–40 On the other hand, postmenopausal women had increased excretion of cross-links, whereas osteocalcin changes were not statistically significant.39
The finding of skeletal integrity in men treated with suppressive doses of L-T4 is in agreement with the results of a previous cross-sectional study of our group in premenopausal women,17 and reassuring data are also emerging from preliminary results of a prospective study going on in premenopausal women.41 When comparing our data with the results of other authors, we would like to underscore the modality of treatment monitoring in our patients.17 First, L-T4 therapy was carefully monitored and all patients were followed by the same group of physicians who employed a well-defined treatment strategy. In addition, the dose of L-T4 was individually adjusted in order to use the lowest dose suppressing TSH secretion. Finally, an attempt to reduce further the daily dose of L-T4 was performed also in those cases in which thyroid hormones were in the normal range.
The discrepancies between our results and those of other investigators might be related to several factors. In most studies, a fixed dose of L-T4 was prescribed, without individual adjustments,12,13 and no data were provided on the hormonal status during the follow-up period.11–16 The doses of L-T4 used in some studies are definitely higher that those currently recommended; it is therefore conceivable that many patients may have been overtreated during therapy.42,43 The importance of a careful monitoring of L-T4 therapy has been recently stressed by Greenspan et al. who reported that when FT4I was maintained in the physiologic range, there was no evidence of decreased bone density in L-T4–treated women.15 In agreement with other studies,12,13,15–17 we found no correlation between duration of L-T4 therapy, cumulative or daily dose of L-T4, serum FT4 or FT3 on one hand, and BMD values on the other.
In conclusion, our data appear to suggest that L-T4 suppressive therapy, if carefully carried out and monitored, using the smallest dose necessary to suppress TSH secretion, does not cause significant changes in bone mass in men nor in premenopausal women. Since there is some evidence that L-T4 treatment may be associated with a decrease in the bone mass in postmenopausal women, it seems reasonable to evaluate the presence of other risk factors for osteoporosis in postmenopausal women submitted to L-T4 therapy, to restrict the use of suppressive doses to patients with differentiated thyroid cancer, and to allow incomplete TSH suppression to patients with benign diseases.
This work was supported by grants from the Ministero dell'Università e della Ricerca Scientifica e Tecnologica, 40% (Rome, Italy) Target Project: “Biologia e Patologia delle Interazioni Cellula-Matrice,” subproject “Osteoporosi e Osteopatie Metaboliche,” from the National Research Council (CNR Rome, Italy) Target Project: “Biotechnology and Bioinstrumentation.” Grant No. 91.01219.PF70; Target Project: “Prevention and Control of Disease Factors” (FATMA) Grant No. 93.00689.