Bone turnover is increased during weight loss in postmenopausal women and can be suppressed with calcium supplementation. In this study, we assessed the influence of energy restriction with and without calcium supplementation (1 g/day) in premenopausal women. Thirty-eight obese premenopausal women (body mass index [BMI] of 35.0 ± 3.9 kg/m2) completed a 6-month study of either moderate weight loss or weight maintenance. During weight loss, women were randomly assigned to either a calcium supplementation (n = 14) or placebo group (n = 14) and lost 7.5 ± 2.5% of their body weight. The control group of women (n = 10) maintained their body weight. Total body and lumbar bone mineral density (LBMD) and content were measured by dual-energy X-ray absorptiometry (DXA) at baseline and after weight loss. Throughout the study, blood and urine samples were collected to measure bone turnover markers and hormones. During moderate energy restriction, dietary calcium intake decreased (p < 0.05) and the bone resorption marker deoxypyridinoline (DPD) increased slightly (p ≤ 0.05) without evidence of bone loss. Calcium supplementation during weight loss tended to increase lumbar BMD by 1.7% (p = 0.05) compared with the placebo or weight maintenance groups. In contrast to our previous findings in postmenopausal women, premenopausal obese women who consume a low calcium diet do not lose bone over a 6-month period, whether their weight is stable or decreasing moderately.
HIGH BODY weight is protective against osteoporosis(1, 2) and weight loss has been shown to decrease bone mass in postmenopausal women. In addition, epidemiological studies show that low body weight or weight loss increase fracture risk in older women.(3–5) However, studies of weight loss in premenopausal women do not show consistent findings of bone loss.(6–8) A particular concern in studies of obesity and weight loss is the potential measurement artifact of bone densitometry because of the excessive amount of overlying fat tissue and weight change.(8–11) Markers of bone turnover can support changes of bone density but are limited by their variability, especially when a single measurement is taken only before and after the treatment. Therefore, we include serial measures of bone resorption and formation in the present controlled trial of obese premenopausal women to determine whether bone is mobilized during energy restriction and can explain changes in bone density.
One potential method to reduce the increased rate of bone resorption associated with energy restriction is to increase dietary calcium. An increase in dietary calcium suppresses the rate of bone turnover in weight-stable premenopausal women(12–14) and in postmenopausal women during weight loss.(15) To our knowledge, this is the first prospective study to examine the role of dietary calcium during energy restriction in premenopausal women. In this double-blind placebo-controlled trial, calcium supplementation (1 g/day) was provided to determine its regulation of bone turnover and mass in obese premenopausal women during 6 months of moderate energy restriction.
MATERIALS AND METHODS
Sixty premenopausal women (aged 42.1 ± 6.2 years; body mass index [BMI] = 34.0 ± 3.9 kg/m2) were recruited for a 6-month weight loss study that began in either the early spring or the fall. Women who had not been pregnant or lactating within the previous year and had a history of a regular menstrual cycle were included. Women who were ill or taking medication known to interfere with bone metabolism, including oral contraceptives, were excluded. Each participant also had to be weight stable for at least 3 months before the start of the study.
Women were randomly assigned in a double-blind manner to one of two weight loss groups: (a) 1000-mg elemental calcium in the form of calcium citrate or (b) placebo tablets (Mission Pharmaceuticals, San Antonio, TX, USA). Subjects were asked to consume two divided doses taken at breakfast and dinner during this 6-month study. The placebo group was divided on completion of the study into those subjects who did and did not lose >2.5% of their initial body weight. Those who did not lose weight were included in a second control group that consisted of women who were recruited to maintain their body weight. At baseline, subjects' food frequency and 24-h diet recalls were assessed to determine usual calorie and nutrient intake. A reduced calorie diet was then individually created using the American Diabetic Association Exchange List. There were weekly group meetings led by a registered dietitian to encourage weight loss using behavior modification and nutrition education. During the diet period, the women were encouraged to continue their regular exercise patterns. Each subject was required to keep a diary containing daily diet recalls, physical activity, and menstrual cycles. Diet analysis is reported at baseline and at week 3 and week 20 of energy restriction.
Total body and lumbar spine dual-energy X-ray absorptiometry (DXA) were taken at baseline and on completion of the weight loss regimen to determine changes in bone mineral density (BMD), bone mineral content (BMC), and total body fat mass (FM) and lean mass (LM). At weeks 0-4, 6, 9, 12, 16, 20, and 25, fasting blood and urine samples were taken to determine markers of bone resorption (urinary pyridinium cross-links), bone formation (serum osteocalcin), and serum parathyroid hormone (PTH). Serum levels of 25-hydroxyvitamin D [25(OH)D], N-telopeptide of type I collagen (NTx), sex hormone-binding globulin, and estrone were measured at week 0 and week 25.
Bone markers and hormones:
Pyridinium cross-links (pyridinoline [PYD] and deoxypyridinoline [DPD]) were measured by high-performance liquid chromatography after hydrolysis and column fractionation.(16, 17) The CV for PYD and DPD was 4% and 6%, respectively, as determined in four subjects on 3 consecutive days. The values are expressed as nanomoles per millimoles of urinary creatinine (no. 555; Sigma Diagnostics], St. Louis, MO, USA). N-telopeptide of type I collagen was measured in serum samples using an ELISA with a monoclonal antibody to N-telopeptide (NTx; Ostex International, Inc., Seattle, WA, USA) with a CV < 11%. Serum osteocalcin was determined by radioimmunoassay (RIA; Diagnostic Systems Laboratories [DSL], Webster, TX, USA) with an inter- and intra-assay CV < 4%. Serum 25(OH)D was determined by RIA (Diasorin, Stillwater, MN; CVs < 15%). Serum intact PTH was measured using immunoradiometric assay (DSL). The intra- and interassay CVs for serum PTH were 3% and 4%, respectively. Serum estrone levels were measured by RIA (DSL; CVs < 9%) and are reported for women at baseline and final measurements when women were in the same stage of their menstrual cycle (for paired analysis). Sex hormone-binding globulin was measured by two-site immunoradiometric assay (DSL; CVs < 12%).
Bone density and body composition:
Total body, lumbar BMD (LBMD), BMC, FM, and LM were measured using DXA (DPX; Lunar Corp., Madison, WI, USA) at previously reported CVs (<1.0% for BMD, 1.1% for BMC, 1.9% for % fat, and 1.1% for LM in 20 subjects measured once per day for four days).(18)
To assess changes in urinary pyridinium cross-links, serum osteocalcin, and PTH between groups over time, a two-way analysis of variance (ANOVA) with repeated measures (on time) at three levels of treatment was performed (Super ANOVA; SAS Institute, Berkeley, CA, USA). If main effect or interaction was significant at p < 0.05, linear contrast analysis (Scheffè's test) was performed. For comparisons before and after treatment (delta) between groups, we performed a one-way ANOVA. To examine whether the season of initiating weight loss influenced the change of measurements from baseline to final, a two-way ANOVA was performed using season and treatment as the main effects. Paired t-tests were used to determine differences between measurements before and after treatment within a group. Values are reported as 0 ± SD in text and tables and 0 ± SEM in figures.
Thirty-eight of the 60 women recruited completed this 6-month study (37% dropped out). The principal reason for discontinuing the study was lack of commitment to the weight loss program because of personal reasons. In addition, one subject who completed the study was dropped because of a reported dramatic increase in physical activity during the weight loss period. Eleven of the subjects recruited to lose weight were unable to be included because of insufficient weight loss, as assessed at weeks 7-8 of the study. Of these 11 women, 6 qualified (<2.5% weight loss and had been randomly assigned to the placebo group) and agreed to participate in the weight maintenance control group. There was 10 women (9 white women and 1 Hispanic woman) in the weight maintenance control group (4 women were recruited specifically for this group). There were 14 women in the calcium group (all white women) and 14 women in the placebo group (9 white women, 3 black women, 1 Asian woman, and 1 Hispanic woman). None of the women in the weight maintenance group dropped out over the 6-month period. There were no characteristic differences between the women who dropped out (n = 22; aged 42.5 ± 5.3 years; BMI of 34.1 ± 4.7 kg/m2) and those who remained in the study (n = 38; average age was 40.4 ± 5.8 years; BMI of 35.0 ± 4.0 kg/m2). The degree of compliance with calcium supplements, as determined by pill counts, was 91 ± 8%. Menstrual cycle remained regular for women in all groups during the study period.
Nutrient intakes were not significantly different in any group at baseline (Table 1). During weight loss, there was a significant decrease in energy intake in both weight loss groups, as expected (Table 1). Calcium intake from the diet and supplement was significantly increased (p < 0.001) in the calcium group, and there was a small but significant decrease in dietary calcium in the placebo group (p < 0.05). Other nutrients did not change significantly between groups and during the study period. Dietary energy intake did not vary significantly from baseline or during the diet period for women who maintained their body weight (Table 1).
Table Table 1.. Estimated Intake of Nutrients from 3-Day Food Records at Baseline and After Weight Loss or Maintenance
Weight loss, bone turnover, and hormonal changes
After 6 months, the women in both the placebo and the calcium groups lost at least 5% of their baseline weight (range, 6.1-16.6%) while women in the weight maintenance group lost 0.6 ± 1.1% of their baseline weight (Table 2). Women who were energy restricted lost 7.9 ± 4.1% (calcium-supplemented) and 7.1 ± 2.5% (placebo) of their body weight. Compared with weight maintenance, both total body BMD and BMC did not change significantly in either weight loss group (Fig. 1; Table 2). BMD in the calcium group increased (p < 0.05) in the lumbar region from baseline to final and compared with the placebo and weight maintenance groups (p = 0.05).
Table Table 2.. Bone and Body Mass Measurements at Baseline and Percentage Change After Weight Loss and Maintenance
At baseline and after weight loss, markers of bone turnover were not significantly different between groups (Table 3). In the placebo group, but not the calcium or weight maintenance groups, urinary DPD values increased from baseline to final (p = 0.05; Fig. 2). There were no significant differences in urinary PYD (Fig. 2), serum N-telopeptides, or osteocalcin levels (Table 3) over time because of energy restriction or calcium supplementation.
Table Table 3.. Bone Turnover and Serum Hormones at Baseline and Percentage Change After Weight Loss and Maintenance
At baseline, serum PTH and estrone levels were not significantly different between groups (Table 3; Fig. 3). In all three groups, the levels of PTH and estrone did not change significantly during energy restriction. Serum estrone did not differ significantly between groups at baseline (200 ± 144 pmol/liter) and did not decrease significantly with weight loss (−21 ± 38%) compared with the weight maintenance group. Sex hormone-binding globulin increased by 17 ± 26% with weight loss (p < 0.005) from 150.1 ± 85.7 nmol/liter to 165.5 ± 85.7 nmol/liter and was not significantly affected by calcium intake.
Serum 25(OH)D levels also were similar between groups at baseline and did not differ significantly after weight loss (Table 3). However, when weight loss subjects were divided by season of the year in which they participated (rather than with or without calcium supplementation) there was a significant increase (p < 0.01) in serum 25(OH)D levels for those who started in the spring and finished at the end of the summer (March-August; n = 17) compared with those who started in fall and ended in the winter (September-February; n = 11; Fig. 4).
In the current study of moderate energy restriction in premenopausal women, there was little or no change in measures of bone turnover. Bone mass was not significantly altered by a moderate weight loss of 7-8%. In addition, in this double-blind placebo-controlled trial, we found that a 1-g calcium supplement showed a small increase in LBMD by 1.7% during weight loss. To our knowledge, this is the first controlled study to examine the effect of moderate weight loss and calcium intake in obese premenopausal women. Importantly, the findings differ markedly from our studies of weight loss and bone in postmenopausal women.
Most studies examining weight loss and bone have included heterogenous populations of pre- and postmenopausal women and men.(9, 19–21) The specific effects on the premenopausal population are not clear from these studies with mixed populations.(9, 21) Unlike the data for premenopausal women, there are more consistent findings of bone loss after weight reduction in postmenopausal women.(22–25) In studies that have examined only premenopausal women, the findings vary and are likely to be related to different study designs or population characteristics including differences in weight and age.(6–8) In a study by Ramsdale et al.,(6) a 5% weight loss after 6 months resulted in a small decrease in bone density at some sites (total body and lumbar spine of −0.7% and −0.5%, respectively) but not at hip. Without a control group in this study(6) it is suggested that the interpretation of these small changes in bone density is likely to be limited by the sensitivity of the DXA measurement (i.e., CV of BMD is 0.9%). In the study by Van Loan et al.,(8) there was a large weight loss of 16 kg in a short period of time (15 weeks). Although there was a small decrease in BMD, it was attributed to measurement artifact associated with changes in the amount of soft tissue surrounding the bone.(8) In addition, Tothill et al.(26) found that total body measurements of BMD and BMC are uncertain during weight loss, but that the Lunar instrument (used in this study) shows less interference with fat change than the Hologic instrument (Hologic, Bedford, MD, USA) that has been used by others examining weight loss.(7)
The largest weight loss and bone study and the only one that included a control group was conducted by Salamone et al.(7) In this study of overweight premenopausal women,(7) a small amount of weight loss (3.2 kg) over 18 months resulted in a significant loss of BMD at the hip (but not at the spine) as compared with a control group who showed a small weight gain (0.4 kg) over the course of the study. These findings are consistent with the present 6-month study showing no significant loss of spine BMD. Because the women studied by Salamone(7) were leaner (BMI of 25) and older (47 years of age at the start of the 1.5-year study) than those in this study, these characteristics could potentially affect bone mass during weight reduction. Leaner women who lose weight are at greater risk of bone loss and fracture risk than heavier women.(3, 4, 23) Also, older premenopausal women(7) often experience hormonal changes and bone loss even before showing clinical symptoms of menopause(27) and could be more susceptible to the detrimental effects of weight loss on bone.
This study has a number of limitations. The number of women losing weight (n = 28) may not be large enough to detect changes in bone turnover or density. However, many other studies have used a similar or smaller sample size and have had the statistical power to detect significant bone loss in pre- and postmenopausal women.(8, 15 19 20) Another limitation of the study is that we did not measure bone density at the hip, which would have been interesting because of the inconsistent findings at this site in previous studies.(7, 6, 19) Also, it is possible that with a longer period of caloric restriction, bone loss may have occurred (due to evidence of a slight rise in DPD, Table 3).(7) However, our goal was to examine bone when weight loss reached a nadir. In our experience, we find that most of the women reach 80% of their weight loss in the first 4-5 months and then lose either very slowly or not at all during the latter months of caloric restriction. Therefore, although it was expected that the most detrimental effects of weight loss on bone mass would occur at ∼6 months(24, 25) during the transient bone remodeling period,(28) no bone loss was observed; yet, calcium supplementation tended to increase spine bone density. Hence, the absence of bone mobilization and loss during weight loss in premenopausal women differs from our findings in postmenopausal women.(15, 24)
A comparison of bone loss in this study of premenopausal women with data from our previous studies in postmenopausal women(15, 24) shows that total body bone loss was less (p < 0.05, two-way ANOVA) in the premenopausal women. We considered whether the greater loss of body weight in our previous study(15) of postmenopausal women (10%) compared with 7-8% in this study could explain the absence of bone loss in the premenopausal women (both groups had low dietary calcium intakes during caloric restriction). An analysis of covariance was performed and showed that the greater bone loss in post- than premenopausal women could have been caused by a greater weight loss in postmenopausal women. Nevertheless, the more marked reduction in calcium intake during weight loss in premenopausal women (from 810 to 459 mg/day; Table 1) compared with postmenopausal women (602-515 mg/day(15)) should have had a more detrimental effect on bone. Therefore, it remains possible that estrogen-replete premenopausal women are at lower risk of bone mobilization and loss with moderate weight loss than postmenopausal women.
The levels of serum PTH did not change with weight loss or calcium supplementation. However, serum PTH levels were less variable in the calcium compared with placebo and control groups (Fig. 3), suggesting that 1.6 g calcium/day stabilizes levels of serum PTH. We had anticipated that serum levels of PTH would decrease with increased calcium intake, as has been found in weight stable pre- and postmenopausal women(29) and weight-losing postmenopausal women.(15) In the present study, the absence of a serum PTH response to calcium supplementation may be caused by an unusually high baseline intake(15, 29, 30) of approximately 1 g calcium/day (Table 1). Increasing calcium intake above this already high baseline level (that met the calcium requirements)(31) could attenuate the serum PTH response. Nevertheless, the additional calcium in the supplemented group appeared to have some beneficial effect by showing a small increase in LBMD by 1.7% compared with the placebo and weight maintenance groups. In weight-stable premenopausal women supplemented with only 600 mg calcium/day (in the form of dairy products), bone loss is prevented compared with controls(32); yet, no bone gain was observed. In the present study, it is possible that the higher level of total daily calcium intake (or the form of calcium as citrate malate) contributes to the trend to increase in bone density.
Serum 25(OH)D levels showed no significant change during the study period in any group. However, there was significant seasonal variation in serum 25(OH)D, with an increase in levels for those who lost weight during the spring and summer months (March-August) compared with those who began their weight loss regimen in the fall and ended the study in the winter (September-February). This phenomenon has been well documented in other young healthy subjects.(33, 34) Unlike findings in weight-stable subjects with low or average calcium intakes,(33–36) we found no seasonal variation for other measurements, such as bone turnover and mass, and hormones. The dynamic state of energy restriction or greater fat stores in obese subjects may mask the response of bone parameters(37) to seasonal variations in serum 25(OH)D.
Because of the high prevalence (>40%) of dieting in the United States,(38) determining the risk for bone loss has important implications. Older obese women who lose weight are at risk for bone mobilization and loss(24, 25) and benefit by increasing calcium intake.(15) However, this study shows that obese premenopausal women who lose a moderate amount of weight (of 7-8%), which is typical of many standard weight loss programs, do not lose bone within the total body or lumbar region, even when consuming a calcium-deficient diet (Table 1). However, calcium supplementation during weight loss may be beneficial in premenopausal women by increasing spine BMD. It is possible that high circulating estrogen levels associated with both obesity and the premenopausal state protects against the bone loss that is associated with moderate weight loss.
We thank A. Herzig, in the Department of Statistics and Computer Science at University of Wisconsin, for her consultation on the study design and analysis. The clinical assistance of the dietitians and the technical support of H. Chowdhury, M. Cifuentes, and A. Charles were greatly appreciated. S.A.S. was supported by the NIH (grant AG12161), the Feasibility Grant from the NY Obesity Research Center (DK26687), and a Johnson and Johnson Discovery Award.