|
|
| |||||||||||||||||||||||||||||||
Background Prolonged corticosteroid therapy increases the risk of osteoporosis and fracture. We studied whether corticosteroid-induced osteoporosis could be prevented by treatment with calcium, calcitriol (1,25-dihydroxyvitamin D3), and calcitonin.
Methods One hundred three patients starting long-term corticosteroid therapy were randomly assigned to receive 1000 mg of calcium per day orally and either calcitriol (0.5 to 1.0 µg per day orally) plus salmon calcitonin (400 IU per day intranasally), calcitriol plus a placebo nasal spray, or double placebo for one year. Data on treatment efficacy were available for 92 of these patients. Bone density was measured every four months for two years by photon absorptiometry. There were no significant differences between groups with respect to age, underlying disease, initial bone density, or corticosteroid dose during the first year.
Results Calcitriol (mean dose, 0.6 µg per day), with or without calcitonin, prevented more bone loss from the lumbar spine (mean rates of change, -0.2 and -1.3 percent per year, respectively) than calcium alone (-4.3 percent per year, P = 0.0035). Bone loss at the femoral neck and distal radius was not significantly affected by any treatment. In the second year, lumbar bone loss did not occur in the group previously treated with calcitonin plus calcitriol (+0.7 percent per year), but it did occur in the group given calcium alone (-2.3 percent per year). The calcitriol group also lost lumbar bone (-3.6 percent per year) but received more corticosteroid in the second year than the other two groups.
Conclusions Calcitriol and calcium, used prophylactically with or without calcitonin, prevent corticosteroid-induced bone loss in the lumbar spine.
Corticosteroids increase bone resorption in several ways6. They decrease calcium absorption7,8 and increase urinary calcium excretion,9,10 causing secondary hyperparathyroidism. They also inhibit bone formation both directly11 and indirectly, by decreasing gonadal steroid secretion12. Vitamin D preparations have been used to treat or prevent corticosteroid-induced bone loss, although their efficacy is unproved. Thus, although treatment with calcitriol increases calcium absorption in patients receiving corticosteroid therapy7,13 and calcitriol has direct stimulatory effects on osteoblasts and bone formation,14,15 it has no beneficial effect on radial bone density. Calcitonin, a potent inhibitor of bone resorption, may improve bone density in patients receiving long-term corticosteroid therapy,16,17,18 but whether it prevents bone loss in patients starting corticosteroid therapy is not known.
Since corticosteroid-induced bone loss appears to be most marked during the first 6 to 12 months of treatment,5,19,20 we studied the effects of calcium, calcitriol, and a nasal formulation of calcitonin on bone mineral density during the first year of therapy in patients receiving long-term corticosteroid therapy.
Methods
Patients
We studied 103 patients with rheumatic, immunologic, or respiratory diseases within four weeks after the initiation of corticosteroid therapy that was expected to continue for at least two years. Patients who had previously received corticosteroid, calcitonin, calcitriol, fluoride, thiazide, or anticoagulant drug therapy were excluded, as were those with any diseases that might affect bone metabolism, renal impairment or calculi, gastrointestinal disease, vasomotor or allergic rhinitis, or acute or chronic sinusitis. The indications for corticosteroid therapy were as follows: rheumatoid arthritis (in 26 patients), polymyalgia rheumatica or temporal arteritis (21 patients), systemic lupus erythematosus (20), dermatomyositis or polymyositis (6), interstitial lung disease (6), Sjogren's syndrome (5), sarcoidosis (4), connective tissue disease (4), vasculitis (3), uveitis (2), eosinophilic fasciitis (2), Waldenstrom's macroglobulinemia (1), Wegener's granulomatosis (1), autoimmune deafness (1), and Weber-Christian disease (1). The corticosteroid therapy was managed by the referring physician independently of the trial. The study was approved by the St. Vincent's Hospital Research Ethics Committee, and informed consent was obtained from each patient.
Study Design
The study was a randomized, double-blind, parallel-group study in which the patients were assigned to one of three groups to be treated for one year with stratification according to sex, age, underlying disease, and initial dose of prednisone or prednisolone (considered to be equipotent)21. Dietary calcium intake22 and physical activity23 were assessed at entry. All the groups received calcium supplementation during the first year, but physical activity was not controlled during the study. The patients in group 1 received 0.5 to 1.0 µg of calcitriol (Rocaltrol, Hoffmann-LaRoche, Basel, Switzerland) daily, plus salmon calcitonin nasal spray (Miacalcic, Sandoz Pharma, Basel), 400 IU per day, plus 1000 mg of elemental calcium daily, in the form of 5.23 g of calcium lactate-gluconate and 0.8 g of calcium carbonate (Sandocal, Sandoz Australia, Sydney). Group 2 received calcitriol plus calcium with a placebo nasal spray. Group 3 received calcium plus both placebo calcitriol and placebo nasal spray. The nasal sprays containing calcitonin or identical placebo and the calcium tablets were supplied by Sandoz Pharmaceuticals (Sydney, Australia). The capsules of calcitriol and identical placebo were provided by Roche Pharmaceuticals (Sydney). Calcitriol treatment was begun at a dose of 0.5 µg per day for two weeks and, in the absence of hypercalcemia, was increased every two weeks by 0.25 µg per day to a maximal dose of 1.0 µg per day. Since mild hypercalcemia (total serum calcium, 10.4 to 11.2 mg per deciliter [2.60 to 2.80 mmol per liter]) occurred at this dose in some patients receiving less than 10 mg of prednisone or prednisolone per day, the dose of calcitriol was subsequently limited to 0.5 µg per day in patients taking 
10 mg per day of either corticosteroid. The average daily dose and the cumulative dose of corticosteroid were determined from diaries kept daily by each patient.
Follow-up
Bone-Density Measurements
Efficacy was evaluated by measurement of the bone density of the lumbar spine, femoral neck, and distal radius at base line and every four months for two years. The bone density of the lumbar spine (L2-4) and femoral neck was measured with a Lunar DP-3 dual-photon absorptiometer (Lunar Radiation, Madison, Wis.)24. The coefficient of variation of replicate measurements on different days in 19 normal subjects 21 to 71 years of age was 1.8 percent in the lumbar spine and 1.9 percent in the femoral neck5. The bone density of the distal radius was measured with a Lunar SP-2 single-photon absorptiometer at a site corresponding to 5 mm of separation between the radius and ulna. The coefficient of variation of replicate measurements on the same day in five normal subjects was 1.0 percent. The bone-density scans for each patient were analyzed by one person who was unaware of the patient's dose of corticosteroid and treatment group. The rates of change in bone density for the first and second years of the study were calculated from regression equations on the basis of measurements obtained at base line and at 4, 8, and 12 months for the first year and at 12, 16, 20, and 24 months for the second year.
The patients were asked about adverse effects at each visit, and a nasal examination was also performed. Serum calcium was measured at one, three, and five weeks and every two months thereafter. If a patient was found to have hypercalcemia, defined as a total serum calcium concentration greater than 10.4 mg per deciliter or a serum ionized calcium concentration greater than 5.2 mg per deciliter (1.30 mmol per liter), both calcium and calcitriol were discontinued. The calcitriol was reintroduced after the serum calcium concentration had returned to normal.
Radiographic Assessment
Lateral radiographs of the thoracic and lumbar spine obtained at base line and at one and two years were analyzed independently by two investigators who were unaware of the patient's status. A vertebral fracture was defined as a reduction of at least 20 percent in the anterior, middle, or posterior vertebral height.
Biochemical Analyses
Serum samples and a timed two-hour morning urine specimen were collected after an overnight fast at base line and every four months thereafter. The base-line samples were obtained before the first dose of corticosteroid in 52 patients. Hematologic and serum biochemical analyses were performed by automated methods by the departments of chemical pathology and hematology at St. Vincent's Hospital. Serum parathyroid hormone (reference range, 4 to 28 pg per milliliter) was measured by immunoradiometric assay (Nichols Institute Diagnostics, San Juan Capistrano, Calif.). Serum 25-hydroxyvitamin D (reference range, 10 to 60 ng per milliliter [25 to 150 nmol per liter]) was measured by competitive protein-binding assay. Serum 1,25-dihydroxyvitamin D (calcitriol) (reference range, 16 to 62 pg per milliliter [38 to 150 pmol per liter]) was measured after extraction and purification by radioreceptor assay (Nichols Institute Diagnostics). Serum osteocalcin (reference range, 3 to 18 ng per milliliter) was measured as described elsewhere25. Urinary calcium was measured by titration with a Corning Calcium Analyzer 940 (Halstead, Essex, United Kingdom), and hydroxyproline was measured with an autoanalyzer (AA1, Technicon, Tarrytown, N.Y.)26. The results were expressed as ratios relative to the urinary creatinine concentration as measured with an Astra autoanalyzer (Beckman Instruments, Brea, Calif.). The urinary calcium: creatinine ratio was considered elevated if it exceeded 0.16 mg per milligram (0.46 mmol per millimole), and the hydroxyproline: creatinine ratio if it exceeded 17 mg per gram (0.15 µmol per millimole)27.
Statistical Analysis
Data management and computations were performed with the SAS statistical software package (SAS Institute, Cary, N.C.). All measurements of bone density and other results were used in the analyses of patients while they continued corticosteroid therapy, whether or not they completed the two-year study.
The results were analyzed in a two-stage model constrained to a common value at one year28 and a mixed-effect analysis of variance. The values for the percent change were compared by an analysis of covariance with age, weight, years after menopause, corticosteroid dose, and the presence or absence of rheumatoid arthritis used as covariates. To increase the precision of the estimates, we estimated the indexes in the model by the least-squares method, weighted by the individual residual mean square. Pairwise differences between treatment groups were derived from the analysis of covariance.
Results
The characteristics of the 92 patients in the three groups for whom data on efficacy were available at base line are shown in Table 1. Of the 103 patients enrolled, 11 had no measurements after base line, 69 completed one year of the study, and 60 completed two years. The reasons for discontinuation were cessation of corticosteroid therapy (in 21 patients, 12 of whom stopped taking corticosteroid in the first year); noncompliance in 10 patients; side effects of the study drugs in 5 patients (hypercalcemia in 2, headaches in 2, and nasal symptoms in 1); and the presence of one of the criteria for exclusion from the study in 5 patients (cancer in 2, thiazide diuretic therapy in 2, and oral anticoagulant therapy in 1 patient). Three patients began estrogen therapy during the second year of the study. Two patients had atraumatic rib fractures during the first year (one patient each in groups 2 and 3). No patients had vertebral fractures during the first year, but five patients each had such a fracture in the second year (two in group 1, one in group 2, and two in group 3).
|
|
The initial mean daily dose of prednisone or prednisolone was 25 mg, and the mean daily dose during the first year was 13.5 mg. There were no significant differences between the three groups with respect to the mean cumulative dose of corticosteroid during the first year: 4.41 ±3.11 g in group 1, 4.24 ±2.49 g in group 2, and 4.59 ±3.16 g in group 3. During the second year, the patients in all groups received less corticosteroid (mean daily dose, 7.5 mg; P<0.001), and the cumulative doses differed (group 1, 2.16 ±1.27 g; group 2, 3.37 ±2.67 g; group 3, 2.33 ±1.33 g; P = 0.03 for the comparison between group 1 and group 2).
Bone Densitometry
The patients treated with calcium alone (group 3) lost significantly more bone from the lumbar spine during the first year (-4.3 ±5.5 percent per year) than did groups 1 and 2 (-0.2 ±6.5 and -1.3 ±5.6 percent per year, respectively; P = 0.0035) (Figure 1). There was a dose-response relation in which those receiving more calcitriol lost less bone (P<0.003). In contrast, the rate of bone loss in the femoral neck was similar in all three groups (-2.8 ±13.1, -2.8 ±10.3, and -2.9 ±6.8 percent per year for groups 1, 2, and 3, respectively). The pattern of bone loss in the distal radius was more variable (+1.3 ±24.1, +0.8 ±12.1, and -3.0 ±12.5 percent per year), but the differences were not statistically significant. The changes in bone density in the first year were similar when the patients with rheumatoid arthritis were excluded.
|
Biochemical Measurements
The serum calcium, phosphate, parathyroid hormone, 25-hydroxyvitamin D, and calcitriol concentrations and the urinary hydroxyproline:creatinine ratio did not change materially during the study (Table 3). The mean values for the urinary calcium:creatinine ratio were significantly increased in groups 1 and 2 at 4 months (0.15 ±0.09 mg per milligram and 0.15 ±0.12 mg per milligram, respectively, vs. 0.10 ±0.06 mg per milligram in group 3; P<0.04), but they returned to normal by 12 months. Thirty-seven patients had a urinary calcium:creatinine ratio higher than 0.16 mg per milligram at some time during the first year (11 in group 1, 17 in group 2, and 9 in group 3), but the serum creatinine concentration was stable in all the groups throughout the study. The serum osteocalcin concentration at base line in the samples taken before corticosteroid therapy was initiated (51 patients) differed significantly from those taken afterward (48 patients) (mean serum osteocalcin, 8.5 and 4.5 µg per liter, respectively; P<0.001). The serum osteocalcin concentration moved gradually over a period of 12 months toward the values obtained before corticosteroid treatment, but it differed significantly between groups at 4, 8, and 12 months, being higher in group 3 than in group 1 or 2 (P<0.05 by repeated-measures analysis of variance).
|
change in lumbar bone density = 0.67 - 1.41 log (serum osteocalcin).
There was no significant relation between either corticosteroid dose or menopausal status, rates of change in bone density at any site.
Discussion
Bone loss from the lumbar spine, but not the femoral neck or distal radius, was prevented or reduced by treatment for one year with calcium plus calcitriol, with or without calcitonin, in the patients receiving corticosteroid therapy. In the second year of the study, when the patients received no calcium, calcitriol, or calcitonin, bone loss in the lumbar spine continued in the group that had received calcium alone (group 3), but not in the group that had received calcitonin and calcitriol (group 1). There was bone loss from the lumbar spine in group 2 (the patients who had received calcium plus calcitriol) in the second year, but this group received a higher cumulative dose of corticosteroid during that year than did the other groups. These results suggest that therapy should be extended in patients who continue to receive corticosteroid therapy. There was no difference in fracture rates between the groups, as would be expected on the basis of the sample size and study duration29,30,31. Hypercalciuria, as assessed on the basis of urine samples obtained in the morning,25,27 was common in all the groups, suggesting that it was caused by the corticosteroid therapy. The increases in urinary calcium excretion in the groups treated with calcitriol were not associated with any changes in renal function, a finding consistent with those in studies of women with postmenopausal osteoporosis29,30.
In corticosteroid-treated patients, preparations of vitamin D increase calcium absorption and reduce bone resorption but do not affect radial bone density13,32. Calcitriol also has a direct effect on osteoblasts, opposing the effects of corticosteroid on osteocalcin-gene expression,33 and calcitriol may reverse corticosteroid-induced suppression of serum osteocalcin concentrations14,15,34. In this study serum osteocalcin concentrations increased after four months; surprisingly, the patients in group 3, who lost the most bone, had the most marked increases in serum osteocalcin in the first year. This result is consistent with data indicating that high bone turnover predicts greater bone loss,35,36 which may be related to genetic factors37.
With regard to the prevention of lumbar bone loss, the group treated with calcitonin and calcitriol had no additional benefit during the first year as compared with the group treated with calcitriol, but the power to detect a difference between groups 1 and 2 was only moderate. There appeared to be some persistent benefit of calcitonin in the following year, in a manner consistent with studies in patients receiving long-term corticosteroid and parenteral calcitonin therapy16,17,18. Although the intranasal administration of 200 IU of salmon calcitonin has been reported to be equivalent to 80 IU given intramuscularly,38 the long-term bioavailability of nasal calcitonin is uncertain. Nasal and intramuscular calcitonin alone may both reduce vertebral-bone loss in corticosteroid-treated patients39. It is important to note that calcium alone did not prevent bone loss from the lumbar spine in our study, but the rate of loss in this group was less than that previously reported in patients receiving corticosteroid therapy5,19. Thus, calcium may reduce corticosteroid-related bone loss, in a manner consistent with previous studies suggesting some benefit of calcium alone in patients receiving corticosteroids40,41.
Our finding of decreased bone loss in the lumbar spine but not at the other sites is consistent with a differential effect of corticosteroids at various bone sites3,5. Patients treated with corticosteroids lose more bone from the lumbar spine than from the radius,3 and there are site-specific responses to agents used to treat osteoporosis42,43,44. It is possible that the underlying disease may have had a confounding effect on bone density by independently affecting physical activity or nutrition, but this effect would have been expected in all three groups.
These results have important therapeutic implications for patients starting corticosteroid therapy. Vertebral fracture is a common and important complication of high-dose corticosteroid therapy45. Thus, our finding that bone loss from the lumbar spine can be prevented by treatment with calcium plus calcitriol, with possibly some additional longer-term effect of calcitonin, suggests that the incidence of corticosteroid-related vertebral fractures could be reduced by this treatment.
Supported by grants from Sandoz Pharmaceuticals, Basel, Switzerland, and the National Health and Medical Research Council of Australia.
We are indebted to the departments of chemical pathology, hematology, and nuclear medicine at St. Vincent's Hospital for expert assistance.
Source Information
From the Bone and Mineral Research Division, Garvan Institute of Medical Research (P.S., J.B., P.K., S.K., T.N., J.E.), and the Departments of Endocrinology (P.K., J.E.), Nuclear Medicine (N.P.), Rheumatology (P.S., S.K.), and Gerontology (P.S.), St. Vincent's Hospital and the Schools of Medicine (P.S., P.K., N.P., J.E.) and Community Medicine (P.S.), University of New South Wales, Sydney, Australia.
Address reprint requests to Dr. Sambrook at the Bone and Mineral Research Division, Garvan Institute of Medical Research, St. Vincent's Hospital, Darlinghurst, NSW 2010, Australia.
References
-(OH) D3 administration on bone and bone mineral metabolism in patients on chronic glucocorticoid treatment: a double blind controlled study. Clin Endocrinol (Oxf) 1983;19:265-273. [Medline]
| |||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
