Bone mineral density (BMD) is modulated by genetic and environmental factors or certain diseases. In several conditions such as low calcium intake, an influence of vitamin D receptor (VDR) polymorphisms on BMD has been suggested. In the present study, we investigated the relationship of Bsm I and Fok I polymorphisms of the VDR gene and BMD in patients with hyperthyroidism, a disease that often results in low BMD. Bsm I and Fok I genotypes were determined in 76 postmenopausal hyperthyroid patients and 62 healthy postmenopausal women as controls. Patients and controls were matched for age, time since menopause, and lifestyle factors and were free of estrogen medication. BMD evaluation included axial dual X-ray absorptiometry (DXA) and peripheral quantitative computed tomography (PQCT). Low BMD was defined as −2.5 STD below the young adult mean value. Biochemical parameters investigated were thyroid hormones, osteocalcin, and 25-(OH)-vitamin D3 as well as routine laboratory data. Low BMD was found in 61% of hyperthyroid patients and in only 23% of euthyroid controls. In the group of hyperthyroid patients with low bone density, the BB genotype (VDR Bsm I polymorphisms) was significantly more frequent (39%) than in controls (13%; p = 0.003) and hyperthyroid patients with normal BMD (6%; p = 0.013). The odds ratio (OR) for low BMD in patients with BB genotype was 5.7 (95% CI, 1.7-19.1; p < 0.005) as compared with the Bb and bb genotypes and 5.5 (95% CI, 2.3-13.2; p < 0.0001) for hyperthyroidism alone. The cumulative risk for low BMD in patients with hyperthyroidism and BB genotype was 31.4 (95% CI, 3.9-256; p < 0.0003). VDR Fok I genotypes showed no significant relationship with BMD or other general or bone-specific parameters. Thus, hyperthyroidism and the genetic background of a BB genotype may promote synergistically the development of low BMD in hyperthyroid patients. Screening for the BB genotype in these patients therefore could help to identify those with particularly high risk for the development of low BMD and allow early treatment.
BONE FRACTURES as a result of low bone mineral density (BMD) are a major health problem(1) that contributes to social and economic problems of the elderly population. BMD itself is influenced by environmental and genetic factors.(2,3) Efforts have been made in recent years to define the genetic basis of BMD. Recently, Bsm I polymorphisms (Bsm I genotypes) at the 3′ end(4) as well as Fok I polymorphisms (Fok I genotypes) at the 5′ end of the vitamin D receptor (VDR) gene(5) have been described as being associated with low BMD.(6–16) However, this has not been confirmed by other investigators.(17–19)
These VDR Bsm I genotypes (BB, Bb, and bb) were reported to be associated with individual BMD in prepuberty(20) before the onset of additional influences such as gonadal steroids or lifestyle habits. Later in life, specific conditions such as low calcium intake,(7,9,21) menopausal state,(22,23) or low body mass index (BMI)(6) and an individual BB genotype may promote a decrease in BMD. Therefore, VDR Bsm I polymorphisms may influence bone status after the achievement of peak bone mass.
The translation products of the VDR Fok I genotypes differ in three amino acids and also were reported to be associated with BMD in some but not all populations investigated.(11,12)
Hormonal changes like hyperthyroidism—independent of etiology(24–26)—or treatment with excessively high doses of thyroid hormone(27) can lead to high bone turnover and cause impressive and rapid bone loss in affected patients.(28–30) This effect surpasses the effects of menopause(31) and may be further modulated by hereditary conditions of bone metabolism.
In this study, we investigated the incidence of specific Bsm I or Fok I genotypes of the VDR gene in hyperthyroid postmenopausal women who have a particularly high risk for the rapid development of osteoporosis.(31–33)
MATERIALS AND METHODS
Patients and controls
Seventy-six consecutive female white patients with hyperthyroidism awaiting radioiodine therapy were selected from our thyroid outpatient clinic. All patients were postmenopausal women without evidence of other diseases or medication known to influence mineral metabolism. Hyperthyroidism was defined by suppressed thyrotropin (thyroid-stimulating hormone [TSH]; ≤0.01 mEq/liter) and moderately elevated free triiodothyronine (fT3, levels were between 5% and 30% above the normal range). Patients had normal or borderline free thyroxine levels (fT4). The mean duration of hyperthyroidism was estimated to be 5 ± 9 months according to the onset of subjective hyperthyroid symptoms. Hyperthyroidism originated in 80% from thyroid toxic adenomas, in 15% from multinodular toxic goiter, and in 5% from Graves' disease. Twenty-eight percent of patients (n = 21) were on antithyroid medication for a few weeks (1-4 weeks) before radioiodine treatment. These patients had fT3 levels in the upper normal range. Medication was stopped 1 week before radioiodine therapy. Written informed consent was obtained from all patients, and BMD was measured before radioiodine therapy.
As controls, postmenopausal healthy white women (n = 62) from the same geographical region were selected randomly from a routine osteoporosis screening. All patients and controls were nonsmokers and had comparable calcium intake as defined by a routine questionnaire. Patients or controls taking estrogens or with an estrogen medication in the past were excluded from the study.
Blood samples were drawn in the morning after an overnight fast using a Vacutainer system (Beckton-Dickinson, Plymouth, U.K.). Leukocytes and sera were separated from 12 ml of whole blood by centrifugation at 1500× relative centrifugal force (Beckmann centrifuge) and sera were aliquoted. Thyroid hormones (TSH, normal range, [0.1-4 mEq/liter]; fT3, normal range [3-6.3 pmol/liter]; and fT4, normal range [9-23 pmol/liter]) and routine laboratory data were analyzed immediately. Thyroid hormones were measured using commercially available assay kits (TSH and fT3 from Behring, Marburg, Germany, and fT4 from Beckton & Dickinson, Immunodiagnostics, Orangeburg, NY, U.S.A.). The remaining aliquots were stored at −20°C for later analysis of 25-(OH)-vitamin D3 (VitD; Inkstar, Stillwater, MN, U.S.A.) and osteocalcin (OC; CIS-bio-international, Gif-sur Yvette, France).
Genomic DNA was prepared from isolated leukocytes by standard phenol-chloroform extraction. Polymerase chain reaction (PCR) and VDR Bsm I and Fok I genotyping was performed as previously described.(4,5) Alleles were genotyped for the presence (b, f) or the absence (B, F) of the Bsm I or the Fok I restriction site, respectively.
Bone densitometry and radiology
BMD was measured at the lumbar spine by dual X-ray absorptiometry (DXA; Hologic 4000, Waltham, MA, U.S.A.). Lumbar DXA was not performed in cases with severe osteoarthritis of the spine or aortic calcification, as determined by spine radiographs, which also were used to exclude vertebral fractures. At the distal radius, BMD was assessed by peripheral quantitative computed tomography (PQCT XCT900; Stratec, Birkenfeld, Germany). Using these data, trabecular and cortical bone density values were calculated separately in 64 of our patients. The CV of repeated measurements was 2.2% for the lumbar spine and 2% for the distal radius.
Reference population for age-matched comparisons (Z scores) was based on the Hologic and the Stratec database of healthy white women. Low BMD was defined according World Health Organization (WHO) guidelines(2) as a T score less than −2.5 STD below the young healthy adult mean BMD value at the lumbar and/or the distal radius measurement site.
All data are presented as mean ± SD, except when noted. Data normality was shown using the Kolmogorow-Smirnov normality test. For comparison of two groups, numeric values (e.g., age) were analyzed by Student's t-test. Data that showed unequal variance or abnormal distribution was analyzed by the Mann-Whitney rank sum test. Proportions of two groups were compared by χ2-analysis (Fisher's two-tailed exact test). Odds ratios (ORs) and 95% CI were calculated by multiple logistic regression. A value of p < 0.05 is regarded as significant for all analyses. Data were evaluated using SPSS for the PC (Release 8.0, 1998; SPSS, Inc., Chicago, IL, U.S.A.).
When anthropomorphic variables of patients and controls were compared, there was no significant difference in age, time since menopause, or BMI (Table 1). Besides a different thyroid function, OC levels were significantly higher and VitD levels were significantly lower in hyperthyroid patients than in controls. Mean BMD values at the lumbar spine (DXA) and at both the trabecular and the cortical bone of the distal radius (PQCT) were significantly lower in hyperthyroid patients than in controls (Table 1). Forty-six of 76 hyperthyroid patients (61%) and 14 (23%) of controls had a BMD T score value below −2.5 STD. Seventeen of 76 hyperthyroid patients (22%) and 26 (42%) controls were classified as having normal BMD at all sites measured. Osteopenia with T scores between −2.5 and −1 STD was found in 13 (17%) of the hyperthyroid patients and in 22 (35%) of the euthyroid controls.
Table Table 1.. General Characteristics of Hyperthyroid Patients and Controls
Allelic distribution of the Bsm I and Fok I genotypes in both the patient and the control groups agreed with the Hardy-Weinberg law. However, in patients with BMD T scores below −2.5 STD, the frequency of the BB genotype as compared with the two other genotypes was significantly higher in hyperthyroid patients with low bone density (Fisher's two-tailed exact test, p = 0.003, 39%) than in all controls (13%) regardless of BMD and in hyperthyroid patients with normal BMD (BMD T scores above −1 STD; Fisher's two-tailed exact test, p = 0.013, 6%; Table 2 and Fig. 1). In hyperthyroid patients with normal BMD, the frequency of BB genotypes did not differ statistically from the control group.
Table Table 2.. Distribution of Bsm I Genotypes in Subjects with Osteoporosis, Osteopenia, and Normal BMD
Multiple logistic regression analysis of determinants of low bone density is presented in Table 3. ORs for patients homozygous for the BB genotype were 5.7 (95% CI, 1.7-19.1) and for concomitant hyperthyroidism 5.5 (95% CI, 2.3-13.2). This was statistically significant (p < 0.005 and p < 0.0001, respectively). When subjects were compared for the cumulative risk of hyperthyroidism and BB genotype, the OR was 31.4 (95% CI, 3.9-256). This was also statistically significant (p < 0.0003). Other significant factors for low BMD were recent peripheral fractures (OR, 4.4; 95% CI, 1.5-13.1; p = 0.007) and age (OR, 1.1; 95% CI, 1–1.3; p = 0.03). A positive family history for osteoporosis or the duration of the menopause was not a statistically significant variable in this logistic regression model. Fok I genotypes showed no significant association with the above-mentioned parameters.
Table Table 3.. Logistic Regression Analysis of Determinants of Low BMD
Enhanced bone turnover in hyperthyroidism is caused by accelerated bone formation and resorption.(34,35) This promotes bone loss and the development of osteoporosis in 54–57% of hyperthyroid patients(25,36) with an increased risk of bone fractures.(1,30) Hyperthyroid postmenopausal women represent a group of patients who are at particularly high risk for bone fractures as a consequence of low bone density associated with this condition. However, not all hyperthyroid postmenopausal women develop osteoporosis. This raises the question of whether there are differences in metabolic and/or genetic parameters of bone metabolism between hyperthyroid patients with and without low BMD. Therefore, bone density, general and bone-specific parameters, and the Bsm I and Fok I polymorphisms of the VDR gene, which have been reported to be associated with low BMD, were investigated in 76 postmenopausal hyperthyroid women and in 62 postmenopausal euthyroid controls.
Hyperthyroid patients in our study had elevated fT3 and suppressed TSH levels as compared with controls. General parameters, age, time since menopause, and BMI were not different in the two groups. In the group of hyperthyroid patients, OC levels were elevated, indicating high bone turnover.(24) In turn, VitD levels were significantly decreased as compared with controls (Table 1). This may be caused by either insufficient ingestion of VitD or accelerated metabolic clearance of VitD in hyperthyroid patients.(37)
Osteodensitometry revealed BMD T scores below −2.5 STD in 46 of 76 hyperthyroid patients, which is within the reported range.(29,31,36) However, 17 of our hyperthyroid patients had BMD T scores in the normal range and 13 patients showed only osteopenia (BMD T score between −2.5 and −1 STD). With respect to fT3 and fT4 levels, duration of hyperthyroidism, and the BMD values, statistical analysis revealed no differences between osteoporotic, osteopenic, and bone-healthy hyperthyroid patients. Surprisingly, there also was no association between low bone density and elevated OC levels. This may be caused by a decrease in OC levels caused by antithyroid medication in some of the hyperthyroid patients. Finally, differences in environmental factors such as lifestyle and calcium supplementation may explain why low bone density occurs only in approximately one-half of hyperthyroid patients. However, these parameters were beyond the scope of this study.
Comparison of the VDR Bsm I genotypes of hyperthyroid patients and euthyroid controls revealed that the BB genotype was more frequent in hyperthyroid patients with low BMD than in hyperthyroid patients with normal BMD and in euthyroid controls. The frequencies of the Bb and bb genotypes were not significantly different between the groups of hyperthyroid patients with normal or low BMD and controls. Overall, the allelic frequencies of the Bsm I genotypes in patients and controls were according to the Hardy-Weinberg law.
Evaluation of the influence of several factors on BMD, that is, age, time since menopause, recent peripheral fractures, hyperthyroidism, and the Bsm I and Fok I genotypes by logistic regression showed that hyperthyroid patients with a BB genotype are at particularly high risk for low BMD (increased risk, 31.4-fold; p < 0.0003) as compared with hyperthyroid patients and euthyroid controls with Bb or bb genotypes. We therefore conclude that the genetic background including a BB genotype is an important risk factor for the development of low bone density in postmenopausal hyperthyroid women without estrogen medication. An association of hyperthyroidism itself with VDR Bsm I genotypes seems unlikely, because genotype frequencies in hyperthyroid patients with normal BMD did not differ from genotype frequencies in the general population. Fok I genotypes showed no relationship whatsoever with BMD or other general or bone-specific parameters.
However, there are conflicting results on the association of the VDR Bsm I genotypes and low BMD in conditions other than hyperthyroidism. An association of the BB genotype with primary osteoporosis and with low BMD in premenopausal healthy women is reported by some but not all authors.(17,19,38,39) This can be explained in part by the large variety of parameters known to influence bone metabolism, such as differences in lifestyle, race, and environmental factors, as well as certain diseases. Associations of VDR polymorphisms and secondary bone disease were reported for primary(40) and secondary hyperparathyroidism(41,42) or insulin-dependent diabetes mellitus.(43)
It is well known that bone loss in hyperthyroidism is partly caused by a relative lack of calcium, which is caused by an imbalance of calcium resorption and consumption. The intestinal calcium absorption rate is reduced to about 25% in hyperthyroid patients and is even lower in patients with a low calcium diet.(44) Several studies report an association of VDR Bsm I genotypes and impaired intestinal calcium resorption.(9,21,45,46) Calcium resorption studies in postmenopausal women showed that in the BB genotype calcium resorption was significantly lower than in the bb genotypes.(46) Therefore, postmenopausal hyperthyroid women with a BB genotype could be especially sensitive toward a dietary lack of calcium. This may be aggravated by low VitD plasma levels. Bone loss in these patients might be prevented by sufficient calcium(45,47,48) as well as vitamin D supplementation.(49) This gain was most pronounced in study patients with the BB genotype.
In our hyperthyroid patients, there was a highly significant association between the BB genotype and low BMD. Our data support the hypothesis that a genetic background of the BB genotype and stimulated bone turnover in hyperthyroidism may act synergistically and aggravate bone loss. This study analyzed only small groups of patients. Further investigation using larger samples could help to evaluate the usefulness of VDR Bsm I genotyping in hyperthyroid patients for clinical practice.