We measured bone mineral density in a large group of patients with Addison disease to determine the importance of these factors in the development of osteoporosis.

Methods

Top Methods Results Discussion Author & Article Info References

Patients

The patients were recruited from the members of a Dutch association of patients with Addison disease. All 155 members, who represent approximately 40% of all patients with Addison disease in the Netherlands [1], were invited to visit the outpatient clinic of our hospital for an interview, physical examination, blood sampling, and measurement of bone mineral density. Ninety-three patients accepted and 2 patients appeared to have no primary adrenocortical failure, so the results of 91 patients with Addison disease were available for analysis. Thirty-one men and 60 women composed the group. Ages were 42.4 ±14.2 years (mean ±SD) in the men and 46.9 ±13.9 years in the women. Glucocorticoid replacement therapy had been given for 11.0 ±9.4 years in the men and for 10.3 ±8.3 years in the women. Addison disease was caused by autoimmune adrenalitis in 83.5% of the patients; in the other patients, the cause was tuberculosis or was unknown. Nearly all patients were treated with either hydrocortisone or cortisone acetate; three patients were treated with prednisone and three patients with dexamethasone. For calculations, the doses of the glucocorticoids were converted to milligrams of hydrocortisone (30 mg hydrocortisone = 37.5 mg cortisone acetate = 7.5 mg prednisone = 1 mg dexamethasone). The daily hydrocortisone dose at the time of the study was 29.2 ±7.0 mg in the men and 28.5 ±6.9 mg in the women. Most of the patients (n = 85) were also treated with mineralocorticoid replacement (fludrocortisone in a dose of 101 ±42 µg [range, 9 to 200 µg] daily, except for 1 patient who used 437.5 µg daily). Thirty eight women (63.3%) had a normal menstrual cycle or were using oral contraceptives (estrogen content, 30 to 50 µg per tablet) for birth control only, except for 1 patient who was treated with oral contraceptives to make an irregular menstrual cycle more regular; 14 (23.3%) women were postmenopausal and had never received estrogen replacement therapy, and 8 (13.3%) were also postmenopausal but were receiving or had received estrogen replacement therapy. Four women had premature menopause (at the ages of 15, 27, 34, and 38 years).

Bone Mineral Density

Bone mineral density of the lumbar spine (L2 to L4) and both femoral necks were measured using a dual-energy x-ray absorptiometer (Hologic QDR-1000 X-ray Bone Densitometer; Hologic Inc., Waltham, Massachusetts). The results of the measurements were compared with age-matched reference population data originating from the United States (lumbar spine, 605 women and 294 men; proximal femur, 747 women and 725 men; all were white) that were provided by the manufacturer of the equipment. Bone mineral density was considered to be decreased if it was less than 2 standard deviations of the mean value of the age-matched reference population in at least one of the assessed regions.

Hormone Assays

Blood samples were taken between 1100 h and 1600 h. Plasma adrenocorticotropin was measured by immunoradiometric assay (Nichols Diagnostics, San Juan Capistrano, California); normal values at 0900 h are 4 to 18 pmol/L. Follicle-stimulating hormone, testosterone, androstenedione, and dehydroepiandrosterone sulfate (DHEA-S) were measured by commercially available radioimmunoassay kits. Normal values were as follows: testosterone, 9 to 35 nmol/L in men and 0.5 to 2.0 nmol/L in women; androstenedione, 3 to 7 nmol/L in both men and women; DHEA-S, 3 to 11 µmol/L in men and 1 to 9 µmol/L in women.

Statistical Analysis

All values are given as mean ±SD unless stated otherwise. Differences between groups were analyzed using the Student t-test. The Pearson correlation coefficient was calculated to assess the association between bone mineral density and plasma levels of various hormones. Linear regression analysis and stepwise multiple linear regression analysis were used to determine the effects of age and hydrocortisone dose on bone mineral density. All statistical analyses were done with the SPS-PC (version 4.01) statistical software package. The study was approved by the hospital ethics committee.

Results

Top Methods Results Discussion Author & Article Info References

Decreased bone mineral density in at least one of the measured regions was found in 10 of 31 men (32%; 95% CI, 17% to 51%) and in 4 of 60 women (7%; CI, 2% to 16%). In the men, decreased bone mineral density was found either in three regions (n = 5) or in one region (lumbar spine, n = 4; right femoral neck, n = 1); in the women, decreased bone mineral density was found in three regions (n = 1), in two regions (both femoral necks, n = 1), or in one region (lumbar spine, n = 2). Regression analysis showed a significant linear correlation between age and bone mineral density of the lumbar spine (r = –0.0046;CI, –0.0071 to –0.0020;P = 0.0009), left femoral neck (r = –0.0051;CI, –0.0069 to –0.0033;P < 0.001), and right femoral neck (r = –0.0051;CI, –0.0069 to –0.0034;P < 0.001). No statistically significant differences were found between men and women with regard to age, duration of glucocorticoid substitution, or glucocorticoid dose expressed in absolute quantities or per kilogram of body weight.

Men

Characteristics of the men with and without decreased bone mineral density are shown in Table 1. No difference was found between the two groups in age, duration of glucocorticoid substitution therapy, body mass index, or serum concentrations of adrenocorticotropin, testosterone, androstenedione, or DHEA-S. The hydrocortisone dose, when expressed per kilogram of body weight, was significantly (P = 0.032) higher in the men with decreased bone mineral density than in the men with normal bone mineral density. The independent effects of age and hydrocortisone dose per kilogram of body weight on bone mineral density were further evaluated using stepwise multiple linear regression analysis. A significant linear correlation was found between the hydrocortisone dose per kilogram of body weight and bone mineral density of the lumbar spine (r = –0.86;CI, –1.60 to –0.13;P = 0.029) but not of the left (r = –0.40;CI, –0.99 to 0.18; P = 0.19) and right (r = –0.41;CI, –1.00 to 0.18; P = 0.18) femoral neck. Age appeared to contribute significantly to the variance of bone mineral density of both femoral necks (left: r = –0.0056;CI, –0.0097 to –0.0015,P = 0.012; right: r = –0.0054;CI, –0.0096 to –0.0013,P = 0.016) but not of the lumbar spine (r = –0.0005;CI, –0.0059 to 0.0050; P > 0.2). No statistically significant correlation was found between bone mineral density and the plasma concentrations of testosterone, androstenedione, or DHEA-S (r between –0.26 and 0.10).

The characteristics of the women with and without decreased bone mineral density are given in Table 1. Because of the small number of women with decreased bone mineral density, no statistically significant differences between the two groups could be shown. The women with decreased bone mineral density were more frequently postmenopausal than were the women with normal bone mineral density. Age had a statistically significant influence on bone mineral density of the lumbar spine (r = –0.0060;CI, –0.0087 to –0.0033;P = 0.001) and both femoral necks (left: r = –0.0045;CI, –0.0063 to –0.0028;P < 0.001; right: r = –0.0047;CI, –0.0064 to –0.0031;P < 0.001), whereas the hydrocortisone dose per kilogram of body weight did not contribute significantly to the variance of bone mineral density of the lumbar spine (r = 0.035; CI, –0.30 to 0.37; P = 0.84) or the femoral necks (left: r = 0.057; CI, –0.16 to 0.28; P = 0.62; right: r = 0.023; CI, –0.20 to 0.24; P = 0.84). A positive correlation was found between bone mineral density and the plasma concentrations of testosterone and androstenedione but not of DHEA-S, whereas the plasma concentration of follicle-stimulating hormone was inversely correlated with bone mineral density (Table 2). The postmenopausal women (n = 14) had a significantly lower bone mineral density than did the premenopausal women (n = 46) with regard to the lumbar spine (0.87 ±0.19 compared with 1.04 ±0.14 gBA/cm²; P = 0.006), left femoral neck (0.67 ±0.12 compared with 0.75 ±0.11 gBA/cm²; P = 0.01), and right femoral neck (0.67 ±0.09 compared with 0.76 ±0.11 gBA/cm²; P = 0.006).

Discussion

Top Methods Results Discussion Author & Article Info References

Our study shows that decreased bone mineral density occurred in more than 30% of male patients receiving long-term glucocorticoid replacement therapy for Addison disease. The daily hydrocortisone dose per kilogram of body weight in the patients with decreased bone mineral density was significantly higher than in men with normal bone mineral density, and bone mineral density of the lumbar spine was inversely correlated with glucocorticoid dose in the entire group. We conclude that long-term, mild glucocorticoid over-replacement is probably responsible for the decreased bone mineral density of the lumbar spine in these patients, despite the absence of clinically recognizable signs or symptoms of hypercortisolism. The finding, with regression analysis, that no correlation existed between the glucocorticoid dose and bone mineral density of the femoral necks can be explained by the observation that glucocorticoid-induced bone loss preferentially involves the spine and ribs [2].

Because no extensive reference data originating from a Dutch population for measurement of bone mineral density using a dual-energy x-ray absorptiometer are available, we used the age-matched reference database from a white, U.S. population. This may have influenced our results. However, there are no reasons to assume that both populations differ significantly in bone mineral density. It has also been shown that the reference data of lumbar spine bone mineral density of a British population do not differ significantly from the North American reference database [3].

Because the incidence of osteoporotic fractures is about five times less in men than in women [4], the absence of decreased bone mineral density in our female patients using similar doses of glucocorticoid replacement is remarkable. Because most (76.6%) of our female patients were premenopausal or had been treated with estrogen for several years after the onset of menopause, we think estrogen may exert a protective effect against the development of osteoporosis caused by excess glucocorticoid replacement. This is in accord with a study [5] showing that estrogen and progesterone replacement therapy reduced glucocorticoid-induced bone loss, as measured by bone mineral density, in a group of postmenopausal or amenorrheic women receiving long-term prednisone treatment for asthma. In another study, no histomorphometric evidence of osteoporosis was found in premenopausal women after prolonged treatment with moderate doses of prednisolone for asthma [6].

Our findings in men differ from the conclusions of one earlier study done in 12 men and 23 women with Addison disease [7]. Bone mineral density of the midshaft and distal parts of the radius was measured using a single-photon absorptiometer. Normal bone mineral density was reported in premenopausal women and in men. The different results in the two studies might be explained by the fact that glucocorticoid-induced bone loss (which preferentially involves the spine and ribs) in the radius may be detected less suitably using a single-photon absorptiometer [2]. Furthermore, the number of patients in that study was considerably smaller than in our study, and the average dose of glucocorticoid replacement (equivalent to 20 to 30 mg of hydrocortisone daily) may have been lower than in our study, although no precise data were provided. Finally, the bone mineral density of 3 of 12 men (25%) in that study was less than 1 SD of the mean of the control group. In postmenopausal women, the authors found increased bone loss associated with low plasma levels of testosterone, androstenedione, and DHEA-S [7]; a similar tendency was found in our study. It was shown previously that higher androgen levels in women are associated with increased skeletal mass [8].

Our finding that long-term, mild glucocorticoid over-replacement may result in decreased bone mineral density is comparable to the observation that mild over-replacement with thyroid hormones during an extended period is also associated with accelerated bone loss [9].

Correct determination of the dose required for adequate glucocorticoid replacement is empirical. A hydrocortisone replacement dose of 30 mg daily is widely accepted [10]. Our male patients with decreased bone mineral density received an average hydrocortisone dose of 31 mg daily, but our findings strongly suggest that this dose is too high for many patients. Recent studies of cortisol production by stable isotope dilution and mass spectrometry suggest that cortisol production may be lower than previously believed. Using this technique in healthy volunteers, Esteban and colleagues [11] found a daily cortisol production rate of 9.9 ±2.7 mg or an average 5.7 mg/m² body surface area. For comparison, our male patients with normal bone mineral density received 13.6 mg hydrocortisone/m² per day, and those with decreased bone mineral density received 16.4 mg hydrocortisone/m² per day. These observations may explain why standard replacement doses of hydrocortisone may be associated with increased bone loss.

Based on the normal cortisol production rate, the replacement dose of glucocorticoids in patients with adrenocortical insufficiency may need adjustment to accommodate the bioavailability of an intermittently administered oral dose. For practical purposes, the correct dose is the lowest possible dose that allows the patient to feel well without producing symptoms of hypocortisolism.

Studies in patients with the Cushing syndrome have shown that steroid-induced osteoporosis is often reversible after restoration of normal steroid levels [12]. Therefore, measurement of bone mineral density (particularly of the spine) using a dual-energy x-ray absorptiometer may be helpful in assessing the appropriateness of glucocorticoid replacement therapy during follow-up of male patients.