| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Division of Endocrinology, Departments of Medicine (D.M.A., D.E.S.) and Epidemiology (E.V., D.M.B.), University of California, San Francisco, California 94143; and San Francisco Coordinating Center and California Pacific Medical Center Research Institute (T.B.), San Francisco, California 94105
Address all correspondence and requests for reprints to: Diana M. Antoniucci, M.D., 74 New Montgomery Street, Suite 600, San Francisco, California 94105. E-mail: dantoniucci{at}psg.ucsf.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Objective: The objective of this study was to test whether the presence of vitamin D insufficiency at the initiation of raloxifene therapy affected the subsequent response of bone mineral density (BMD).
Design, Setting, and Participants: We studied 7522 postmenopausal participants of the Multiple Outcomes of Raloxifene Evaluation, a placebo-controlled trial of the effects of raloxifene on BMD and fracture.
Intervention: After enrollment, all participants began daily supplements of 500 mg calcium and 400–600 IU cholecalciferol; 1 month later, women were randomly assigned to placebo or raloxifene.
Main Outcome Measure: Serum levels of vitamin D [25-hydroxy vitamin D (25OHD)] were measured at enrollment, randomization, and 6 months later. We categorized participants’ vitamin D status (deficient, insufficient, or sufficient) based on their randomization 25OHD level. We estimated the effects of treatment on BMD within these subgroups using linear regression models.
Results: At enrollment, 3.2% of participants were vitamin D deficient, and 51.8% were insufficient; after 7 months of cholecalciferol supplementation, 0.2% of all participants remained D deficient, and 23.6% remained insufficient. The effects of raloxifene on hip and spine BMD did not vary by vitamin D status at randomization (P = 0.08 and P = 0.7, respectively).
Conclusion: We conclude that vitamin D status at initiation of raloxifene therapy does not affect the subsequent BMD response when coadministered with cholecalciferol and calcium. After 7 months of cholecalciferol therapy, very few women continued to have 25OHD levels in the deficient range; however, 25OHD levels remained suboptimal in nearly one fourth of the cohort. Additional research is needed to determine whether these observations can be generalized to other antiresorptive agents.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Osteomalacia, the most severe manifestation of vitamin D deficiency, is characterized by insufficient mineralization of osteoid, and is quite different from osteoporosis, which is best characterized by an imbalance between bone formation and bone resorption, leading to net bone loss with trabecular thinning. Despite their pathophysiological differences, these disorders are indistinguishable on dual-energy x-ray absorptiometry (DXA) imaging, both appearing as low bone mineral density (BMD). Recommendations for the management of low bone density include supplementation with calcium and vitamin D and, if the bone density is low enough, treatment with antiosteoporosis agents (7). These agents are directed at improving bone density but do not treat vitamin D deficiency, which may be the cause of or a significant contributor to the low BMD (3, 4, 8).
Treatment for vitamin D deficiency consists of pharmacological doses of vitamin D, such as 50,000 IU once or twice a week for 4–6 wk or until repletion is achieved (9, 10). These replacement doses are much higher than the doses used for the prevention or management of osteoporosis, which typically are in the 400–800 IU/d range. High dose replacement rapidly reestablishes vitamin D stores and can lead to significant improvements in BMD ranging from 4–8% per year (11, 12). It is not known whether reducing bone turnover with antiresorptives in a person with some degree of osteomalacia may impair the mineralization of osteoid necessary to resolve the osteomalacia. A study using etidronate suggested that the use of antiresorptives without pharmacological vitamin D replacement in a patient with insufficient vitamin D stores may lead to smaller improvements in BMD than might be observed in a vitamin D sufficient patient undergoing the same treatment (13). Conversely, etidronate is relatively unique among the antiresorptives in that it has significant antimineralization effects when used daily. Other antiresorptives do not impair mineralization and may not affect the bone density response to vitamin D therapy.
Although it has been postulated that having vitamin D deficiency or insufficiency may affect a person’s response to antiresorptive therapy, this theory has not been tested with any antiresorptive agent other than etidronate. We undertook this study to determine whether the effect of raloxifene on BMD after 3 yr of follow-up differed based on whether subjects participating in the Multiple Outcomes of Raloxifene Evaluation (MORE) study were vitamin D insufficient or not when initiating antiresorptive therapy.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
We analyzed data obtained during the MORE study. MORE was a multicenter international placebo controlled trial of the effects of raloxifene on BMD and fracture incidence in postmenopausal women. A total of 7705 subjects were recruited in 25 countries. Details of study recruitment and participant characteristics have been published previously (2, 14). Briefly, participants were at least 2 yr past menopause and met any of the following criteria: osteoporosis defined as a T-score below –2.5 at the lumbar spine or femoral neck; low BMD and one or more moderate (25–40% height loss) vertebral fractures or two or more mild (20–25% height loss) vertebral fractures; or at least two moderate fractures, regardless of BMD. Reasons for exclusion included the following: a history of metabolic bone disease, other than osteoporosis; substantial postmenopausal symptoms; malignancy; recent treatment for osteoporosis (with exception of calcium and vitamin D supplements); glucocorticoid therapy for more than 1 month within the preceding year; treatment with antiepileptic drugs, pharmacological doses of cholecalciferol, or oral estrogen within the previous 2 months; endocrine diseases requiring therapy (except for type 2 diabetes or hypothyroidism); serum creatinine levels greater than 2.5 mg/dl (221 µmol/liter); active nephrolithiasis, abnormal hepatic function tests, or untreated malabsorption; or consumption of more than four alcoholic drinks per day. The protocol was approved by the human studies review board at each study site, and participants provided informed consent.
Vitamin D status was examined in all participants who had randomization 25-hydroxy vitamin D (25OHD) levels available (n = 7522). The analyses of the effects of raloxifene on 3-yr BMD included only women in whom both randomization serum levels of 25OHD and at least one follow-up spine or hip BMD measurement were available (n = 6909). The last BMD value available was carried forward in participants who withdrew before completing the study. The analyses of the effects of raloxifene on incident vertebral fractures included only those women in whom both randomization serum levels of 25OHD and follow-up vertebral x-rays were available (n = 6687). Analyses of the effect of raloxifene on the incidence of any new nonvertebral fracture included those women in whom randomization serum levels of 25OHD and data on the incidence of a nonvertebral fracture was available (n = 7522).
Treatment
After entry into the study, all participants received daily supplements of 500 mg calcium and 400–600 IU cholecalciferol. One month after initiation of calcium and cholecalciferol supplements, women were randomly assigned to treatment groups. Each group was asked to take one of three types of tablets: placebo or 60 or 120 mg raloxifene. Thus, twice as many women received raloxifene as placebo. The drug manufacturer produced randomly numbered kits containing identically appearing raloxifene or placebo tablets. Trial centers dispensed the kits in numerical order to the women enrolled in the study. Participants and clinic and laboratory staff were all blinded to treatment assignment.
Biochemical assessment
Fasting blood samples were obtained at the enrollment visit before any intervention had taken place and at the randomization visit, approximately 1 month after the initiation of calcium and cholecalciferol supplements. Fasting blood samples were obtained again 6 months after randomization, at which time participants had been on calcium and cholecalciferol supplements for approximately 7 months and treatment with either placebo or raloxifene for 6 months. The blood samples were centrifuged and kept frozen until analyzed. Serum 25OHD was measured by RIA (DiaSorin, Stillwater, MN) with an interassay coefficient of variation (CV) between 2.7 and 6.0%. The limit of detection was 3.2 ng/ml (8 nmol/liter). Serum 25OHD levels below the sensitivity of the assay were assigned a value corresponding to half the lower limit of detection (1.6 ng/ml or 4 nmol/liter) for analytical purposes. PTH was assessed using an immunoradiometric assay (DiaSorin); the interassay CV ranged between 8.2 and 11.8%. Serum calcium was measured by colorimetry using a Roche Hitachi 747 or 911 analyzer (Roche, Indianapolis, IN) with an interassay CV between 1.30 and 1.34%. Serum phosphorus was quantified by colorimetry using an Hitachi (Tokyo, Japan) 747-200 analyzer; the interassay CV ranged between 1.34 and 1.68%. Serum alkaline phosphatase was determined by photometric measurement (at 405 nm) of the change in p-nitrophenol over time (Sigma, St. Louis, MO), with an interassay CV between 2.35 and 3.53%.
BMD assessment
BMD measurements of the spine and femoral neck were measured annually by dual-energy x-ray absorptiometry as reported previously (14). A central reading facility provided correction factors to adjust for intersite differences and changes in the performance of the densitometers over time (15, 16).
Fracture assessment
Details of fracture assessment have been described previously (14). Briefly, participants underwent vertebral radiographs at enrollment and 24 and 36 months. In addition, when women experienced symptoms of vertebral fracture, they underwent radiography at interim 6-month visits. All vertebral radiographs were assessed at a central site by radiologists blinded to treatment group assignment, who scored each vertebra using a semiquantitative scale. The grading scores were set as 0 for none, 1 for mild, 2 for moderate, and 3 for severe fractures. When a follow-up radiograph demonstrated a grade change of at least 1, this was defined as an incident fracture.
Nonvertebral fractures were determined by direct questioning every 6 months at each clinic visit. Traumatic fractures, defined as fractures resulting from a traffic collision, a beating, or having been struck by a falling or moving object, were excluded from analyses. Also excluded from analyses were pathological fractures and those involving the fingers, toes, and skull.
Statistical analyses
We examined the effect of randomization vitamin D levels on the annualized percentage change in femoral neck BMD during 3 yr of raloxifene therapy. To account for the seasonal variation in serum 25OHD levels, we added the difference (2.7 ng/ml or 6.8 nmol/liter) in average concentration among subjects randomized during the summer compared with those randomized during the winter to the measured levels of the latter group. In the northern hemisphere, winter was defined as October through March and summer as April through September, whereas in the southern hemisphere, the definition was reversed.
We then categorized adjusted vitamin D levels less than or equal to 10 ng/ml (25 nmol/liter) as "deficient," adjusted levels more than 10 ng/ml (25 nmol/liter) and less than or equal to 30 ng/ml (75 nmol/liter) as "insufficient," and adjusted levels more than 30 ng/ml (75 nmol/liter) as "sufficient." These categorizations were based on the most recent expert opinion (17, 18, 19) and on clinical criteria that take the whole calciotropic system and bone histology into account (20). ANOVA was then used to assess differences in biochemical parameters across the vitamin D subgroups, whereas t tests were used to compare the treatment groups within these three subgroups. ANOVA was also used to assess differences in the mean change in PTH level when participants’ vitamin D status category changed over 6 months. These analyses were only performed when at least three participants changed vitamin D status category over 6 months.
Previous analyses demonstrated that the percentage change in BMD was similar in the groups assigned to 60 and 120 mg raloxifene (14), so these groups were combined within vitamin D subgroups. All analyses were by intention-to-treat, without regard to compliance with assigned regimen.
We used linear regression models to estimate and compare the effects of treatment on change in BMD within the subgroups defined by vitamin D level. We used logistic regression models to estimate and compare the effects of treatment on the risk of fracture within the subgroups defined by vitamin D level.
In the primary analysis, we considered vitamin D status at randomization, about 1 month after supplementation had begun. In secondary analyses, we also considered analogous classifications of vitamin D levels measured before medical therapy or dietary supplement was initiated (
1 month before randomization), as well as levels measured 6 months after randomization.
All analyses were performed using Stata Software, version 8.0 (Stata Corporation, College Station, TX).
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
At randomization, the women in the three vitamin D status subgroups were similar in prevalence of current smoking, previous use of hormone replacement therapy, and femoral neck BMD (Table 1
). However, the women who were vitamin D insufficient were slightly older (66.6 ± 7.1 compared with 66.3 ± 7.0 yr) than those with vitamin D sufficiency. The insufficient group had a higher body mass index (BMI) when compared with the sufficient group (25.9 ± 4.2 vs. 24.6 ± 3.7 kg/m2) but not to the deficient group (26.1 ± 4.7 kg/m2). The women with vitamin D insufficiency had higher BMD at the spine than those with deficiency and sufficiency (0.82 ± 0.14 compared with 0.78 ± 0.14 and 0.81 ± 0.13 g/cm2, respectively).
|
). Serum PTH levels were also inversely correlated with 25OHD levels (r = –0.22; P < 0.001). Compared to the women with vitamin D insufficiency, the sufficient group had a slightly lower serum phosphorus concentration (3.80 ± 0.5 vs. 3.76 ± 0.4 mg/dl, or 1.23 ± 0.2 vs. 1.22 ± 0.1 mmol/liter). Lastly, the deficient participants had higher serum alkaline phosphatase levels compared with the sufficient participants (82.6 ± 24.3 vs. 74.7 ± 20.7 U/liter), as did the insufficient participants (79 ± 22.8 vs. 74.7 ± 20.7 U/liter).
Compliance, defined as taking at least 80% of study medication by pill count, was similar in the deficient (91%) and sufficient (93%; P = 0.6) groups but was significantly lower in the insufficient group (91%) when compared with the women in the sufficient group (P = 0.001).
Vitamin D status
All participants were given 400–600 IU/d cholecalciferol as part of the study. Overall, self-reported compliance with this therapy was 99.7 ± 0.1% at the 1-month visit and 98.3 ± 0.1% at the 6-month visit. As shown in Table 2, 3.2% of participants were vitamin D deficient and 51.8% were vitamin D insufficient before initiating any vitamin D supplementation. Within 1 month, 72 and 7%, respectively, of these women were no longer vitamin D deficient or insufficient. After 6 months of cholecalciferol supplementation, only 0.2% of all participants remained D deficient and 23.6% remained vitamin D insufficient. The subjects who remained D deficient after 6 months of vitamin D supplements had taken cholecalciferol for an average of 224 ± 29 d, whereas those who had become D sufficient had taken the supplements for an average of 208.6 ± 20.7 d (P = 0.004). However, similar proportions of women who remained vitamin D deficient at the 6-month visit and those who repleted reported 100% compliance with this supplementation (P = 0.1). Those participants who were still vitamin D insufficient after 7 months of supplementation had taken the supplement for an average of 211.7 ± 27.9 d and reported 98.8% compliance compared with 207.6 ± 20.6 d (P < 0.001) and 99.5% (P = 0.03) in those who repleted. As depicted in Fig. 1, women whose vitamin D status improved over time, regardless of their original status, experienced a significant decrease in serum PTH, whereas women whose vitamin D status worsened experienced a significant rise in their serum PTH concentration.
|
|
Within the vitamin D status categories, the treatment groups did not differ in age, BMI, femoral neck and spine BMD, serum 25OHD concentrations, frequency of previous hormone replacement therapy, or prevalence of current smoking at randomization (Table 3). Follow-up time averaged 2.73 ± 0.6 yr and did not differ between participants who were on placebo compared with those who were taking raloxifene (P = 0.1). One imbalance was detected in the vitamin D-insufficient subgroup: serum PTH levels were slightly lower among the women assigned to placebo (32 ± 14 pg/ml or 3.4 ± 1.5 pmol/liter) than among those assigned to raloxifene (33 ± 15 pg/ml or 3.5 ± 1.6 pmol/liter).
|
). In pairwise comparisons as well, there was at most only weak evidence for differences in response to treatment (Fig. 2). The effect of raloxifene on spine BMD was similar across the three subgroups (P for heterogeneity 0.2) (Fig. 3). Furthermore, in secondary analyses assessing modification of the treatment effect on femoral neck BMD by vitamin D status before supplementation began, as well as 6 months after randomization, we found similar response to treatment across the three subgroups (P for heterogeneity 0.5 and 0.1, respectively).
|
|
The odds ratio of a new vertebral fracture for the vitamin D-deficient group on raloxifene treatment compared with placebo (n = 39) was 1.3 (95% CI of 0.3–5.6), whereas for the insufficient group (n = 2089) it was 0.6 (95% CI of 0.5–0.8), and for the sufficient group (n = 2316) it was 0.5 (95% CI of 0.4–0.7). However, neither the test for heterogeneity (P = 0.3) nor for trend (P = 0.2) across these three estimates was statistically significant.
There appeared to be no reduction in the odds of nonvertebral fracture in any of the vitamin D groups. The effects of raloxifene on nonvertebral fractures also did not vary by vitamin D status at randomization, because neither the test for heterogeneity (P = 0.6) nor the test for trend (P = 0.4) were statistically significant.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Vitamin D deficiency is associated with decreased calcium absorption and secondary hyperparathyroidism. The elevated serum PTH levels lead to increased bone turnover and bone loss, which tends to be more prominent at cortical bone sites (26). Repletion of vitamin D stores leads to resolution of the secondary hyperparathyroidism and to mineralization of existing osteoid in the case of severe vitamin D deficiency (27, 28). Raloxifene is a selective estrogen receptor modulator, which has been shown to decrease bone turnover (8, 29, 30), although, unlike estrogen, it does not seem to reduce bone formation (31). Raloxifene has also been shown to increase BMD at both the hip and lumbar spine (14, 30, 32) and is generally considered to be an antiresorptive agent for the treatment of osteoporosis in postmenopausal women. However, unlike bisphosphonates, raloxifene also appears to have positive effects on calcium balance via its selective estrogen receptor actions on the intestine (31). We sought to test whether women with vitamin D deficiency at randomization have a smaller BMD response to raloxifene than those who were vitamin D sufficient initially, based on the theory that an antiresorptive agent might reduce the ability of preexisting osteoid to mineralize even in the presence of concurrent sufficient vitamin D repletion. Koster et al. (13) showed that 1 yr of etidronate therapy with supplemental calcium but no vitamin D led to a 0.5% increase in BMD in participants who had vitamin D deficiency at baseline and a significantly greater increase of 5.2% in women with vitamin D sufficiency. The present study demonstrated that raloxifene did not alter the ability of BMD to improve even in the presence of vitamin D deficiency. This effect may have been partly modulated by the positive effect of raloxifene on calcium balance in addition to its selective greater negative effects on bone resorption than bone formation. Additionally, vitamin D supplements were started 4 wk before antiresorptive therapy. The study also found no differences on the effects of raloxifene on fracture risk based on vitamin D status, although there were few events among the vitamin D-deficient participants.
Our data also indicate, as has been shown previously by Lips et al. (2, 27), that almost all of the women who started out being D deficient and more than half of those who started out being D insufficient repleted their vitamin D stores with relatively low doses of vitamin D. The fact that those few women who remained D deficient reported greater compliance and longer courses of D therapy suggests that they might have had difficulty absorbing the cholecalciferol supplement they were provided. Because more than half of participants who began the trial with insufficient vitamin D stores reached vitamin D sufficiency after 6 months of therapy with 400–600 IU cholecalciferol and because their BMD response to raloxifene was the same, one could postulate that, as long as 400–600 IU vitamin D is started along with raloxifene therapy, there is no need to check 25OHD levels at initiation of therapy. Serum tests for 25OHD are relatively expensive, and it would be ideal to minimize how many women with low BMD need to be screened for vitamin D deficiency. However, it would be important to determine whether women are still vitamin D deficient or insufficient after 6–7 months of cholecalciferol supplementation, because those women who remain vitamin D insufficient at this point may need more aggressive vitamin D replacement or may even need a more thorough evaluation for the etiology of their vitamin deficiency. Unfortunately, we did not find any parameters, such as initial serum levels of PTH, vitamin D levels, or BMD response to raloxifene, that could help predict which women would not replete (data not shown).
This study used data from an existing trial that was not originally designed for the current analysis. We maintained the original randomization to drug or placebo in our analysis, but some of our groups were relatively small, thus limiting power in some analyses. We used serum 25OHD levels at randomization, that is 1 month after initiation of cholecalciferol supplementation, to determine our deficient, insufficient, and sufficient groups and their response to raloxifene. It is possible that, although serum 25OHD level had changed after 1 month of cholecalciferol treatment, this did not reflect tissue stores of vitamin D and that some of the participants may have been misclassified by our grouping, although analyses based on enrollment vitamin D levels were not different from our primary analyses.
Results from assays for 25OHD are somewhat variable and dependent on the laboratory performing the analysis (33, 34). However, it has been shown that all assays routinely used are capable of discriminating between low, average, and high 25OHD values but that, to compare results from different laboratories, careful cross-calibration is necessary (33). Serum 25OHD levels in MORE were all measured using the DiaSorin RIA in a single centralized laboratory. This assay has compared well to the gold standard of HPLC (33, 35).
We conclude that vitamin D status at initiation of osteoporosis therapy with raloxifene does not affect BMD response to this medication when it is concomitantly administered with 400–600 IU cholecalciferol and 500 mg calcium. Whereas 55% of subjects had suboptimal vitamin D levels at enrollment, only one fourth of participants remained D insufficient after 7 months of therapy with 400–600 IU cholecalciferol. However, it seems warranted to check 25OHD serum levels after 6 or 7 months of supplementation with cholecalciferol to ensure that those women who have not repleted can be assessed for causes of persistent vitamin D insufficiency and treated with higher doses of vitamin D for repletion. Additional research is needed to determine whether these observations can be generalized to other antiresorptive agents.
|
Footnotes |
|---|
First Published Online May 17, 2005
Abbreviations: BMD, Bone mineral density; BMI, body mass index; CI, confidence interval; CV, coefficient of variation; MORE, Multiple Outcomes of Raloxifene Evaluation; 25OHD, 25-hydroxy vitamin D.
Received February 10, 2005.
Accepted May 10, 2005.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  B. Dawson-Hughes, P. Chen, and J. H. Krege Response to Teriparatide in Patients with Baseline 25-Hydroxyvitamin D Insufficiency or Sufficiency J. Clin. Endocrinol. Metab., December 1, 2007; 92(12): 4630 - 4636. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
10 ng/ml), insufficient (serum vitamin D between 10 and 30 ng/ml), or sufficient (serum vitamin D > 30 ng/ml) at study enrollment, after 1 month (randomization), and after 6 months of 400–600 IU/d cholecalciferol supplements
