Calcium Supplementation with and without Hormone Replacement Therapy To Prevent Postmenopausal Bone Loss

Methods

Healthy, white women between 6 months and 6 years after a natural menopause were recruited to participate in the study. The protocol was approved by the Human Investigation Review Committees of Winthrop-University Hospital and Brookhaven National Laboratory; written informed consent was obtained from each participant. Participants were recruited by announcements in the local press and in hospital and university publications and through a direct mail campaign. All participants had a history and physical examination. Exclusion characteristics included any disorder known to affect bone metabolism such as glucocorticoid use, gastrointestinal disease, or any chronic illness. Previous or current malignancy was an exclusion characteristic as were absolute contraindications to estrogen replacement or calcium supplements. Absolute contraindications to estrogen replacement therapy included estrogen-dependent neoplasm (breast or uterus), undiagnosed vaginal bleeding, thrombophlebitis or thromboembolism, and acute liver disease. Women with the following problems considered by some investigators to be relative contraindications to estrogen therapy were also excluded: gallbladder disease, history of liver disease, first-degree relatives with breast cancer, and hypertension. Calcium urolithiasis was also an exclusion factor. Women with known osteoporosis or with a vertebral compression fracture were not eligible for the study.

One hundred eighteen women entered the study. The women were randomly assigned to three groups: 1) hormonal replacement [estrogen-progesterone-calcium carbonate], 2) calcium carbonate, or 3) placebo. Assignment to the groups was based on computer-generated random numbers provided by the statistician, with stratification for years postmenopause. The women in the hormonal replacement group took conjugated equine estrogens (Premarin, Wyeth-Ayerst Laboratories, Inc.; Philadelphia, Pennsylvania), 0.625 mg daily for 25 days of the month along with medroxyprogesterone (Provera, Upjohn; Kalamazoo, Michigan), 10 mg from days 16 to 25. All women received 400 IU of vitamin D daily in the form of a multivitamin, and calcium supplementation (as Caltrate, Lederle; Clifton, New Jersey) was provided to the two treatment groups. The duration of the study was 2.9 ±1.1 years (mean ±SD).

A 7-day dietary history was reviewed with a nutritionist every 2 months; calcium was provided as calcium carbonate, 600 mg (Caltrate), and used to supplement the diet to approximate a total daily intake of 1700 mg of elemental calcium (the mean + 2 SD found by Heaney and colleagues [7] to result in zero calcium balance in estrogen-deprived women). The calcium supplements were taken with meals in divided doses. The placebo appeared identical to the calcium carbonate tablets. No patients took antacids or histamine-2 blockers. All women had a baseline mammogram.

Measurements

Routine laboratory studies included a complete blood count, urinalysis, and serum fasting calcium, phosphorus, urea nitrogen, creatinine, alkaline phosphatase, cholesterol, and aminotransferase measurements [19, 20]. In addition, follicle-stimulating hormone, estradiol, parathyroid hormone, osteocalcin, free thyroxine, and bone alkaline phosphatase were measured, and a urine specimen was collected after an overnight fast for hydroxyproline, calcium, and creatinine determinations, following a 3-day low-hydroxyproline diet [21-23].

Total body calcium was measured annually in the participants, using the delayed γ neutron activation method at Brookhaven National Laboratory [24, 25]. This method uses a whole-body counter to measure the characteristic γ rays emitted from the neutron capture of Calcium-48 (natural abundance of 0.187%) in the body. The Brookhaven National Laboratory whole-body counter was upgraded in 1987 to use 32 NaI (T1) detectors of 10 cm × 10 cm × 46 cm positioned symmetrically above and below the patient [25]. The activated isotope, Calcium-49, decays with a half-life of 8.72 minutes, emitting a 3.08 MeV characteristic γ line. More than 99.5% of the body calcium is contained in the bone [26]. The method provides total body calcium with a coefficient of variation of about 1.5% when no substantial change in the body weight occurs during the period of repeated studies. The measurements were made annually.

The bone mineral density of the distal radius site was measured using a Lunar Radiation (Madison, Wisconsin) single-photon absorptiometer (SP2). Bone mineral density of the spine (L2-L4) and femur (neck, trochanter, and Ward triangle) was measured using a Lunar Radiation DP4 dual-photon absorptiometer. The software version used for the analysis of scans was DP4 Lunar Corporation Version 1.1. All scans were analyzed using the same software version, which corrects for source decay. Instruments were calibrated daily, and the radioactive source was changed annually. Each measurement was done every 6 months. The coefficient of variation of these measurements was 2%, except for the Ward triangle (2.5%).

Activity was measured using activity monitors (large-scale integrated monitors), which were worn about the waist [27]. The average of 2 weekdays and 1 weekend day was used as an activity score. Activity was measured at baseline and at one other point during the study to ensure that differences among the groups were not due to varied levels of exercise.

Statistical Analysis

Total body calcium was selected as the primary criterion for efficacy for the following reasons: It measures mass rather than density per unit area; it measures calcium balance precisely and accurately in the free living state and may be better related to previous studies using the balance technique; it is more precise than the other measurements; and it avoids sampling error by measuring the entire skeleton rather than a specific region of the appendicular or axial skeleton.

The rate of change in bone mineral was calculated for each woman at each of the sites used in the study. Standard linear regression procedures were used to estimate the rate of bone mineral change for each woman, and the regression intercept was used as the best estimate of the baseline value. Because some women terminated their participation in the study before others, the rate-of-change data were weighted by the inverse variance to reflect the fit of the regression line for each woman [28]. Analyses of covariance were done using body mass index, activity scores, cigarette smoking, calcium intake, age, and years postmenopause as covariates. The data reported in this article are based on all women who provided at least three observations for a particular skeletal site. We considered other criteria, such as using data only from women who had participated in the study for at least 2 years, and all data analyses were done for this subgroup as well. The results of these analyses were invariably similar to those reported here and therefore are not presented separately. The mean rates of change in bone mineral for each condition at each site were characterized in terms of both raw units and percentages; separate analyses were carried out for each. The two indices were similar.

Evidence from recent research is substantial that estrogen replacement therapy is effective, whereas the efficacy of calcium supplements is questionable. Our expectation was that our data would confirm the efficacy of estrogen-progesterone-calcium therapy, and the critical question was whether or not a beneficial effect of calcium supplements given alone could be shown. A separate one-way analysis of covariance was done for each of the bone mineral measurements to compare the mean rates of change in bone mineral for each of the three conditions. We used two a priori contrasts: the first contrasting women taking estrogen with those receiving calcium and the second comparing women receiving calcium supplements with those on placebo. All P values reported are two-tailed.

Results

Baseline data for historical data and bone mineral measurements and chemical studies are given in Table 1. Analysis of variance showed no significant differences in the baseline variables. The initial and final activity scores did not differ significantly.

The range of initial daily calcium intake in the overall study group was 150 to 1263 mg; in the calcium augmentation group, it was 222 to 806 mg. Initial intake in the calcium augmentation group was less than 400 mg in 26%, 400 to 650 mg in 51%, and greater than 650 mg in 23%; 90% had an intake less than 800 mg. Compliance was determined by the number of tablets returned when it was time for a refill of medication (every 2 months). Percentage compliance was calculated as the total number of tablets of calcium, less the number missed, divided by the total, and multiplied by 100. Overall compliance was greater than 98%. The calcium supplements were also tested for dissolution: They were placed in a beaker with 0.1-M HCl solution. At 30 minutes, 482 ±11 mg of calcium was measured; and at 60 minutes, 99.8% (599 ±9.7 mg) was detected. Three patients reported intolerance (gastrointestinal symptoms) to the calcium carbonate tablets, requiring withdrawal from the study by one patient.

Of the women on estrogen-progesterone-calcium, regular menses occurred in all but three patients. Three patients had mid-cycle bleeding, necessitating a dilation and curettage in one patient. Other side effects reported in the estrogen-progesterone-calcium group included breast tenderness (2 patients), decreased libido (1 patient), menstrual cramps (1 patient), eructation (1 patient), constipation (1 patient), headache (2 patients), weight gain (2 patients), and mood disorder (2 patients). Fifteen women smoked cigarettes during the study: 7 in the estrogen group, 4 in the placebo group, and 4 in the calcium augmentation group.

Seventeen women left the study. The reasons for dropping out included stopping hormonal therapy (2 patients), starting hormonal therapy (3 patients), desire to withdraw (3 patients), side effects of calcium augmentation (1 patient), peptic ulcer disease (1 patient), breast cancer (1 patient), hyperthyroidism (1 patient), moved out of the region (2 patients), claustrophobia (1 patient), obesity preventing completion of the neutron activation studies (1 patient), and development of rheumatoid arthritis (1 patient). The woman who developed breast cancer was not in the estrogen-progesterone-calcium group.

Analysis of covariance revealed that years since menopause was related to total body calcium, bone density of the femoral neck, and bone density of the Ward triangle. The means were adjusted for the effect of years since menopause on these sites (Table 2).

The response to calcium augmentation was intermediate between responses to placebo and estrogen-progesterone-calcium at each skeletal site Table 2, Figure 1, Figure 2, Figure 3, Figure 4, Figure 5, and Figure 6. The difference in percentage loss per year in total body calcium for the calcium augmentation group as compared with the placebo group was 1.4%/y (CI, 1.0 to 1.8) and was 1.2%/y (CI, 0.3 to 2.2) for the femoral neck Table 2, Figure 1 and Figure 2. The neutron activation data Figure 1 depicted with annual measurements shows that the calcium augmentation group remained intermediate between the other two groups throughout the study. The femoral neck rates of change Figure 2 also showed a benefit of calcium over placebo throughout the 3 years. The first 2 years of the spine measurements Figure 3 appear to show a transient effect of calcium, but this appearance is probably due to measurement error. The differences between calcium and placebo were not statistically significant for the radius, the Ward triangle, and trochanter Figure 4, Figure 5, and Figure 6.

View larger version:

• In this window
• In a new window

Figure 1. Calcium augmentation remained intermediate in effect between placebo and estrogen-progesterone-calcium throughout the 3 years of the study. HRT = hormone replacement therapy. The change in total body calcium with time.

View larger version:

• In this window
• In a new window

Figure 2. Calcium augmentation had a sustained beneficial effect compared with placebo. BMD = bone mineral density; HRT = hormone replacement therapy. The biannual rate of change in bone mineral density of the femoral neck.

View larger version:

• In this window
• In a new window

Figure 3. The estrogen-progesterone-calcium group did not lose bone density, whereas no statistically significant benefit of calcium over placebo was found. BMD = bone mineral density; HRT = hormone replacement therapy. The biannual rates of change in bone mineral density of the lumbar spine.

View larger version:

• In this window
• In a new window

Figure 4. The rates of change were not statistically different from zero. BMD = bone mineral density; HRT = hormone replacement therapy. Biannual measurements of bone mineral density of the radius.

View larger version:

• In this window
• In a new window

Figure 5. Although estrogen-progesterone-calcium is superior to calcium augmentation, the slope for calcium is intermediate to the other two groups. BMD = bone mineral density; HRT = hormone replacement therapy. Bone mineral density of the trochanter.

View larger version:

• In this window
• In a new window

Figure 6. The response to calcium augmentation is similar to the intermediate effect observed for the trochanter. BMD = bone mineral density; HRT = hormone replacement therapy. Bone mineral density of the Ward triangle.

Rates of bone mineral changes for the estrogen-progesterone-calcium group were stable and differed from calcium for the trochanter by 2.6% (CI, 1.4 to 3.8) and for the total-body calcium measurement by 1.4% (CI, 1.0 to 1.8). The annual plots of these data Figure 1, Figure 2, Figure 3, Figure 4, and Figure 5 show that the estrogen-progesterone-calcium effect was sustained throughout the 3 years of the study.

The placebo group, as expected, lost skeletal mineral at each site, with the highest rates of loss from the Ward triangle ( −2.7%/y).Expression of the total-body calcium data as daily calcium balance yields a value of −46 mg/d for the placebo group as compared with −11 mg/d for the calcium augmentation group.

Discussion

A beneficial effect in retardation of bone loss by calcium augmentation alone in the early postmenopausal years was noted with the total-body calcium measurement but not with all measurements. There are several reasons why this may have occurred: Only the neutron activation technique measures calcium; the total-body calcium measurement is a mass measurement, whereas all other measurements are of areal density; the precision of the neutron activation method is superior to the others; and calcium could have been deposited in extraskeletal tissue. The latter seems unlikely because the magnitude of extraskeletal deposition would have to be great, and no evidence exists that calcium supplementation in this amount could produce extraskeletal calcification. Moreover, total body calcium is an integral measurement of the entire skeleton and is more likely to detect changes if they are not homogeneous (that is, not uniform in the regional skeletal sites measured). We and others have found that the measurement of total body calcium better discriminates women with osteoporotic vertebral fractures from normal women than any regional measurement [29, 30]. Finally, a beneficial effect of calcium augmentation was also found in retardation of bone loss from the femoral neck. This would not be expected if there were some peculiarity specific to the total-body calcium measurement because the femoral neck was measured using an unrelated technique—dual-photon absorptiometry. Because femoral neck fractures have a great effect on morbidity and mortality in the elderly, prevention of loss of bone from this site by increasing calcium intake has great public health importance.

Many studies have suggested that calcium supplements might be effective in retarding postmenopausal bone loss, whereas some studies have had negative results [6-18]. Calcium supplementation in a clinical trial in Danish women also appeared to have had an intermediary effect when one looks at the data graphically, but the response to calcium supplements was not statistically significant when compared with either placebo or hormonal replacement therapy [10]. Because the customary dietary intake in Danish women is high by world standards and the dietary intake of these participants was not reported, the study could be interpreted as indicating that there was no benefit to supplementation of a calcium-sufficient diet, whereas those on a calcium-deficient diet might still benefit.

Dawson-Hughes and colleagues [16] recently reported on a double-blind, placebo-controlled, randomized trial of placebo compared with calcium carbonate or calcium citrate malate in American women whose calcium intake ranged up to 650 mg/d. These investigators proposed that calcium augmentation might be beneficial only in those on a deficient diet. They found that women in the early postmenopausal years did not benefit from a 500-mg supplement. On the other hand, in older, postmenopausal women (>6 years after menopause), bone loss diminished from the radius and femur (but not the spine) with CaCO₃. In those with a low calcium intake, calcium citrate malate retarded spinal bone loss. Reid and colleagues [17] recently reported on a clinical trial with 1000-mg calcium supplements and found retardation of bone loss in women who were 9 to 10 years postmenopausal. Because they used a combination of calcium lactate-gluconate and calcium carbonate, they questioned the amounts and type of calcium supplements that are effective, in particular whether calcium carbonate is effective. We addressed this by using a 1700-mg calcium intake (using carbonate) daily.

The average daily dietary calcium intake in our participants was approximately 500 mg. The average intake found in the National Health and Nutrition Survey (NHANES II) found a median intake of 475 mg for women over 44 years of age. Our calcium-supplement group was representative of American women in calcium intake, with 90% of patients having intakes below the recommended dietary allowance of 800 mg. However, in contrast to the Dawson-Hughes study [16], only one quarter of our participants had a calcium intake less than 400 mg, so our findings cannot be applied only to women on a markedly low calcium intake, although they are the group that would be expected to benefit most from augmentation. Indeed, from our studies we would suggest that a high calcium intake may be beneficial to women who currently ingest amounts equivalent to the recommended dietary allowance.

Prince and colleagues [31] studied 120 postmenopausal women with low forearm bone mineral density, randomized to exercise (n = 41), exercise plus dietary calcium supplementation, or exercise plus hormonal replacement therapy. The control group (not randomized) had normal initial bone mineral density. The control group and exercise group lost bone mineral density at the distal forearm. This study was designed to examine the combined effects of exercise and calcium supplementation or hormonal replacement therapy. The investigators studied the radius alone, using mostly exercises that were non-weight-bearing for the radius, such as walking. The women averaged more than 5 years postmenopause. Although calcium supplementation provided protection, hormonal therapy increased bone density to a greater degree, as we also observed.

Elders and colleagues [8] did a controlled trial of calcium supplementation in perimenopausal women (between 46 and 55 years old) and found evidence for retardation of bone loss from the spine in the first year but not the second year of the study. This transient effect in bone remodeling was supported by the changes they observed in biochemical parameters of bone remodeling (for example, alkaline phosphatase, osteocalcin, urinary hydroxyproline), which were suppressed at 1 year but tended not to continue the suppressive trend at 2 years. We also failed to detect a benefit on spinal bone density beyond 1 year, which raises the question of whether there is a differential effect of calcium augmentation on cortical and cancellous bone [18]. Nonetheless, our study was long enough that it is unlikely that the beneficial effects of calcium that we observed represent a remodeling transient (that is, a temporary change in the calcium balance resulting from a change in bone remodeling). Figure 1 and Figure 2 show that the response to calcium was sustained.

Earlier controlled studies suggested that supplementation of 1000 to 2000 mg/d slowed bone loss from the radius and the metacarpal cortical area. Other controlled studies failed to find an effect on vertebral bone loss. Proponents of calcium augmentation have suggested that calcium supplementation may exert a differential effect on cortical compared with cancellous bone [18]. The absence of any effect of calcium augmentation on spinal bone loss in our study suggests that no benefit of calcium augmentation should be anticipated in the early postmenopausal years in the prevention of vertebral crush fractures. However, the power to detect differences in treatments was the poorest for the spine of all the sites, due to the variability of the measurements.

Our results confirm that estrogen-progesterone replacement therapy with calcium augmentation is effective in retarding early postmenopausal bone loss and support the lack of an equivalent effect of calcium augmentation alone to a total daily intake of 1700 mg. The beneficial effect of estrogen-progesterone-calcium therapy on prevention of loss of bone density from each region of the hip was shown. No loss from the spine was statistically significant.

We cannot conclude from the results of this study whether hormonal therapy in the absence of calcium supplementation is effective in preventing early postmenopausal bone loss or whether estrogen-progestins given separately are effective. Other studies have established the efficacy of hormonal therapy without calcium augmentation. Because we used doses of hormonal therapy that were conventional at the time we designed our study, we also cannot address the efficacy of lower doses, with or without calcium supplements [32].

In summary, in this healthy population of early postmenopausal white women, calcium augmentation of the diet retarded bone loss from the entire skeleton remarkably or from a negative daily balance of 46 mg to less than 12 mg. No effect (or an undetectable effect) of calcium augmentation on the rate of bone loss from the spine was observed, but a beneficial effect on the neck of the femur was shown. Estrogen-progesterone-calcium therapy was superior to calcium augmentation or placebo in preventing bone loss. Women who are at high risk for osteoporosis because of marked osteopenia at menopause would be expected to benefit from estrogen-progesterone-calcium therapy. Women with marked spinal osteopenia should receive this therapy or alternative antiresorptive therapy. However, all women may benefit from a reduction of compact bone loss by ensuring an ample intake of dietary calcium. Although the total-body calcium measurement reflects primarily compact bone, the maintenance of skeletal integrity is to a large extent dependent on compact bone strength. The possible beneficial effects of augmenting dietary calcium around the time of menopause should be considered because in this study calcium balance was remarkably improved and the rate of loss from the neck of the femur was retarded.