RSS Feeds
Body Weight and Bone Mineral Density in Postmenopausal Women with Primary Hyperparathyroidism
- Andrew B. Grey, MB, ChB;
- Margaret C. Evans, BSc;
- Joanne P. Stapleton, RGON; and
- Ian R. Reid, MD
- From the University of Auckland, Auckland, New Zealand. Requests for Reprints: Ian R. Reid, MD, Department of Medicine, University of Auckland, Private Bag 92019, Auckland 1, New Zealand. Grant Support: In part by the Auckland Medical Research Foundation, the Health Research Council of New Zealand, and the New Zealand Lottery Board.
Abstract
Objective: To assess bone mineral density and body composition in postmenopausal women with primary hyperparathyroidism.
Design: Cross-sectional study with an age-matched control group.
Setting: University teaching hospital.
Patients: 41 postmenopausal women with mild primary hyperparathyroidism and 43 eucalcemic, age-matched controls.
Measurements: Total body, lumbar spine, and proximal femoral (femoral neck, Ward's triangle, and trochanter) bone mineral density; body composition; and fat distribution were measured using dual-energy x-ray absorptiometry.
Results: Women with primary hyperparathyroidism were heavier (75.5 kg compared with 66.3 kg; difference, 9.2 kg [95% CI, 3.7 to 14.7 kg]; P = 0.002), had a higher fat mass (33.3 kg compared with 26.1 kg; difference, 7.2 kg [CI, 3.0 to 11.4 kg]; P = 0.001), and had a more android pattern of fat distribution (android-to-gynoid fat ratio, 1.05 compared with 0.84; difference, 0.21 [CI, 0.1 to 0.32]; P = 0.0004) than the controls. Unadjusted bone mineral density was similar in patients and controls at all sites: total body, 0.990 compared with 1.023 g/cm2 (difference, 0.033; CI, −0.004 to 0.070); posteroanterior lumbar spine, 1.032 compared with 1.018 g/cm2 (difference, 0.014; CI, −0.031 to 0.059); lateral lumbar spine, 0.569 compared with 0.528 g/cm2 (difference, 0.041; CI, −0.022 to 0.104); femoral neck, 0.799 compared with 0.825 g/cm2 (difference, 0.026; CI, −0.072 to 0.124); Ward's triangle, 0.653 compared with 0.677 g/cm 2 (difference, 0.024; CI, −0.035 to 0.089); trochanter, 0.734 compared with 0.733 g/cm2 (difference, 0.001; CI, −0.024 to 0.026); and arms, 0.720 compared with 0.739 g/cm2 (difference, 0.019; CI, −0.015 to 0.053). After adjustment for body weight, bone mineral density in women with primary hyperparathyroidism was lower than that in controls for total body (P = 0.0004), femoral neck (P = 0.001), Ward's triangle (P = 0.01), trochanter (P = 0.02), and arms (P = 0.0006). Spinal bone mineral density did not differ between groups.
Conclusions: Body weight, total body fat mass, and proportion of android fat are increased in postmenopausal women with primary hyperparathyroidism; these unexplained factors may be relevant to the increased incidence of cardiovascular disease in this condition. Unadjusted bone mineral density values are similar in patients with primary hyperparathyroidism and in controls, suggesting that this condition is not associated with an increased risk for fracture.
Primary hyperparathyroidism is the third most common endocrine disorder and has its highest incidence in postmenopausal women [1]. The advent of multichannel biochemical analysis has led to the recognition that mild, asymptomatic disease occurs frequently [2], and the correct approach to the management of persons with such disease is much debated [3, 4].
In recent years, attention has focused on the long-term skeletal effects of primary hyperparathyroidism. Asymptomatic primary hyperparathyroidism is considered to be one of the four indications for bone mineral density measurement [5], and osteopenia in patients with primary hyperparathyroidism is regarded as an indication for surgical intervention [6]. Several investigators, using single-photon absorptiometry, have reported reductions in bone mineral content in the proximal forearm, a site of predominantly cortical bone [3, 7-17], but these reductions have not been a universal finding [18]. Few data have been reported on bone mineral density in the proximal femur, an important site of osteoporotic fracture, where a combination of cortical and trabecular bone is found. No data on total body bone mineral density in primary hyperparathyroidism have been reported. Both reduced [11, 12, 18] and normal [8, 16] values have been reported for bone mineral density at the trabecular-rich lumbar spine. Many of the investigators who have done bone mineral density studies [3, 8-15, 17, 18] have neither reported body weight in the study groups nor indicated what adjustment was made for this important determinant of bone density [19].
In our study, we did a comprehensive assessment of bone mineral density (including that for the proximal femur, the lumbar spine in the posteroanterior and lateral projections, and the total body) in a cohort of postmenopausal women with primary hyperparathyroidism and compared the results with those observed in healthy eucalcemic women.
Methods
Participants
Postmenopausal women with mild, asymptomatic primary hyperparathyroidism were recruited by postal invitation from the Auckland Hospital Endocrinology Clinic and local general practices. Invitations were sent to 70 women, 7 of whom had moved and were not locatable, and 3 of whom were ineligible because they were taking estrogen. Of the 60 women who were both contactable and eligible, 41 (68%) agreed to participate. In each participant, hypercalcemia was detected incidentally during routine blood testing and primary hyperparathyroidism was confirmed by the presence of a concomitant increase in serum ionized calcium and intact parathyroid hormone. No participant had evidence of malignancy or a family history of hypercalcemia. None was taking any medication or had any disease other than primary hyperparathyroidism known to influence bone metabolism. Twenty-five patients (61%) were taking antihypertensive medications, and 7 (17%) had known ischemic heart disease. Five (12%) were currently employed.
Controls
Forty-three normal postmenopausal women from the same community as the patients with primary hyperparathyroidism provided control data. These participants were part of a larger group of healthy postmenopausal women who were recruited by newspaper advertisement and whose clinical and demographic characteristics have previously been reported [20]. Those who were similar in age to the patients with primary hyperparathyroidism were selected by an independent statistician, who was unaware of their bone mineral density and body weight, to provide an age-matched control group. None had any condition or was taking any medication known to influence bone metabolism. Three (7%) were taking antihypertensive medications, and 2 (5%) had known ischemic heart disease. Seven (16%) were currently employed.
Bone Density and Body Composition
Bone mineral density was assessed using a Lunar DPX-L dual-energy x-ray absorptiometer (Lunar Radiation Corporation, Madison, Wisconsin). Separate scans of the whole body, the lumbar spine in both the posteroanterior (L2-L4) and lateral projections (L3), and the proximal femur (femoral neck, Ward's triangle, and trochanter) were done and analyzed using the manufacturer's version 1.3 software. Because previous studies had often measured mid-radius bone mineral content, analyses of the bone mineral density of the arms subregion of the total body scans were done to provide an assessment of appendicular cortical bone. Total body fat mass and lean mass were also quantified from the whole body scans [21]. The precision (coefficient of variation) of the bone mineral density measurements in our laboratory was 0.4% for total body, 1.0% for posteroanterior lumbar spine, 3.1% for lateral lumbar spine, and 1.4% for femoral neck. The precision of the body composition measurements was 2.7% for total body fat mass and 0.8% for lean body mass. Android (waist) and gynoid (thigh) fat were measured by regional analysis of the total body scans [22]. The waist region was defined by a box whose superior border was the uppermost part of the 12th thoracic vertebra, whose inferior border was at the level of the iliac crests, and whose lateral borders were the outermost soft tissue to either side. The thigh region was defined by a box of equal dimensions, positioned so that its uppermost border was level with the inferior pubic rami. Fat distribution was assessed by calculating the ratio of android to gynoid fat in each participant. In the 43 controls, the android-to-gynoid fat ratio derived in this manner correlated with the waist-to-hip ratio as assessed by tape measure (r = 0.73, P < 0.0001).
Radiologic Studies
Lateral radiographs of the lumbar spine were obtained from each patient with primary hyperparathyroidism and each control. Any vertebrae affected by fracture were excluded from bone mineral density analysis.
Biochemical Studies
Intact parathyroid hormone concentrations were measured using a two-site immunoradiometric assay (Nichols Institute, San Juan Capistrano, California) with a coefficient of variation of 8% (normal range, 1 to 5 pmol/L). Ionized calcium was measured using an ion-specific electrode (Radiometer, Copenhagen, Denmark; normal range, 1.17 to 1.28 mmol/L).
Body Mass
Weight was measured using electronic scales; body mass index was calculated by dividing weight (kg) by the square of height (m).
Socioeconomic Status
Population census data were used to assign each study participant to one of five groups, according to average household income in their residential suburbs within the Auckland urban area [23].
Statistical Analysis
Baseline data in the two groups were compared using the Student t-test and the chi-square test. Further analysis of the bone mineral density data was done using the GLM procedures of the SAS statistical package (SAS Institute, Cary, North Carolina). Least-squares mean bone mineral density values at each site were generated by analysis of covariance, with body weight as a covariate, and then the values of the controls were compared with those of patients with primary hyperparathyroidism. A significance level of α = 0.05 was used for all analyses. Results are presented as mean ±SE unless otherwise specified.
The study was approved by the Auckland Area Health Board Ethics Committee, and each participant gave written, informed consent.
Results
The mean (±SD) ionized calcium level for patients with primary hyperparathyroidism was 1.42 ±0.08 mmol/L (range, 1.30 to 1.63 mmol/L); the mean level of parathyroid hormone was 9.4 ±4.7 pmol/L (range, 3.4 to 25.5 pmol/L).
Clinical and body composition data for patients with primary hyperparathyroidism and controls are shown in Table 1. The two groups were comparable for age, height, and cigarette smoking. Patients with primary hyperparathyroidism weighed, on average, 9 kg more than controls. This difference was almost entirely due to an increased total body fat mass in the patients with primary hyperparathyroidism. Lean mass did not differ between the groups. The ratio of android-to-gynoid fat in patients with primary hyperparathyroidism was greater than that in the controls. No difference existed between patients with primary hyperparathyroidism and controls in socioeconomic status, which was assessed according to residential area (P = 0.3).
Body weight correlated with bone mineral density at all sites in the controls (0.52 < r < 0.69, P < 0.001) and with total body and proximal femoral bone mineral density in the patients with primary hyperparathyroidism (0.45 < r < 0.58, P < 0.005). Figure 1 shows the bone mineral density results, unadjusted for weight, in each of the two groups. There were no significant differences between the primary hyperparathyroidism group and the control group at any site.
Bone mineral density results adjusted for body weight are shown in (Figure 2). After adjustment for body weight, total body, proximal femoral, and arm bone mineral densities were significantly lower in patients with primary hyperparathyroidism than in controls. The mean reduction was 6% in total body bone mineral density, 12% in femoral neck bone mineral density, 10% in Ward's triangle bone mineral density, 7% in bone mineral density in the trochanteric region, and 7% in arm bone mineral density. No difference was found between patients with primary hyperparathyroidism and controls in spinal bone mineral density assessed in either projection.
Discussion
Our study showed that postmenopausal women with primary hyperparathyroidism are significantly heavier, have greater total body fat mass, and have proportionally more android fat than do age-matched, eucalcemic controls. Unadjusted bone mineral density measurements in the women with primary hyperparathyroidism were similar at all sites to control values. Adjustment for body weight, however, showed modest reductions in bone mass in the women with primary hyperparathyroidism at sites containing a substantial proportion of cortical bone (arms, proximal femur, and total body), whereas bone mineral density at the predominantly trabecular site (lumbar spine) remained similar to that observed in controls.
Obesity has not previously been reported as a feature of primary hyperparathyroidism, and our study appears to be the first to address this issue. Body weight has been reported infrequently in the literature on primary hyperparathyroidism. High body weight was present in four studies of primary hyperparathyroidism [7, 24-26], only one of which included age-matched controls [25]. However, increased body weight has been overlooked or regarded as a chance finding. Studies that recorded normal body weight in patients with primary hyperparathyroidism either involved small (n < 10), heterogeneous groups of participants [27-30] or failed to adjust for a difference in age between the patients with primary hyperparathyroidism and the controls [16].
The mechanism underlying the association between primary hyperparathyroidism and obesity is uncertain. It is unlikely that our finding is a chance event, given the highly significant differences between women with primary hyperparathyroidism and controls in both weight (P = 0.002) and fat mass (P = 0.001) and the absence of a plausible mechanism for the introduction of a weight bias during patient selection. Nonetheless, it is possible that an unrecognized selection bias may have influenced the findings. The mean weight of controls in our study is similar to that of local [31] and international [32-34] populations of normal postmenopausal women. Obesity itself, through its association with other diseases, may increase the chance of diagnosing primary hyperparathyroidism simply by increasing the likelihood of blood testing. If this ascertainment bias applies, however, it should do so in other persons with primary hyperparathyroidism who come to medical attention and would not negate the significance of our finding of normal unadjusted bone mineral density.
The insulin resistance that occurs in persons with primary hyperparathyroidism [27-30] might promote increased fat mass by selectively affecting skeletal muscle and not adipose tissue, thereby diverting carbohydrate to adipocytes. Alternatively, because adipocytes and osteoblasts share a common progenitor cell-type [35], parathyroid hormone, which directly activates osteoblasts, may also influence adipocyte differentiation and function. A third possibility is that the effects of obesity on calcium metabolism (decreased serum 25-hydroxyvitamin D and urinary calcium concentrations with secondary hyperparathyroidism) [36] may, over many years, promote the development of autonomous hyperparathyroidism.
Our data indicate that postmenopausal women with primary hyperparathyroidism have proportionally more android fat than do their eucalcemic peers. An android pattern of fat distribution is an independent risk factor for cardiovascular disease [37, 38]. The android-to-gynoid fat ratio assessed by dual-energy x-ray absorptiometry is increased in patients with angiographically proven coronary artery disease when compared with those with no coronary artery disease [39]. Our data may in part explain the increased cardiovascular mortality in primary hyperparathyroidism, which is observed even after adjusting for other vascular risk factors [40].
Recommendations for management of mild primary hyperparathyroidism include assessment of bone mass and surgical intervention if osteopenia is shown [6]. Single-photon absorptiometry studies of forearm bone mineral content in primary hyperparathyroidism [3, 7-17] have often shown a reduction in cortical bone mass that is considerably greater than the 6% to 7% deficit in weight-adjusted values at the cortical sites (total body and arm) assessed in our study. An increased fat mass in primary hyperparathyroidism may explain this discrepancy because increased adipose tissue results in artifactual underestimation of bone mass when single-photon absorptiometry techniques are used [41]. Thus, previous studies of primary hyperparathyroidism may have overestimated the deficit in cortical bone mass by failing to control for adiposity. Thus, it is notable that two recent histomorphometric studies failed to show a reduction in cortical width in primary hyperparathyroidism [42, 43].
Controversy exists about whether trabecular bone mass is diminished in primary hyperparathyroidism [7, 8, 11, 12, 14, 16, 18, 42]. Our finding of normal spinal bone mineral density is consistent with histomorphometric studies of primary hyperparathyroidism, which have shown either normal or increased trabecular bone mass [14, 42], and with some densitometric studies [8, 16]. One of the studies reporting spinal osteopenia in primary hyperparathyroidism used single-energy computed tomography to assess bone mass [11], rendering the results subject to artifactual underestimation caused by an increased marrow fat content. The other two studies, which used population-derived normal values rather than matched controls, do not state whether the data were weight adjusted [12, 18].
Studies that have shown measurement of bone mineral density to be predictive of fracture risk have used absolute bone mineral density rather than weight-adjusted values [44, 45]. Our results, therefore, suggest that postmenopausal women with primary hyperparathyroidism, by virtue of their normal unadjusted bone mineral density, are not at increased risk for fracture compared with their eucalcemic peers. Controversy exists about whether fracture incidence is increased in primary hyperparathyroidism. Studies that included only participants with mild, asymptomatic primary hyperparathyroidism found no increase in fracture prevalence [46-48]. An increased prevalence of vertebral fractures has been reported in patients with primary hyperparathyroidism undergoing parathyroidectomy [49, 50], but a bias toward referring those with fractures for surgery may have influenced these findings. Greater disease severity may explain the excess of both axial and appendicular fractures in another study, in which 28% of the participants had serum calcium levels above 3 mmol/L [51].
In summary, our findings suggest that osteopenia may not be as important a feature of primary hyperparathyroidism in postmenopausal women as has previously been suggested, and this may have important implications for the management of this common condition. The observations that postmenopausal women with primary hyperparathyroidism have increased fat mass and an android fat distribution raise new questions about the effects of chronic mild parathyroid hormone excess and the pathogenesis of this condition.
- Copyright ©2004 by the American College of Physicians
Related Clinical Content
