Calcium Intake and Bone Mineral Density: Systematic Review
Calcium Intake and Bone Mineral Density: Systematic Review
Increasing calcium intake from dietary sources slightly increased bone mineral density (BMD) (by 0.6–1.8%) over one to two years at all sites, except the forearm where there was no effect. Calcium supplements increased BMD to a similar degree at all sites and all time points (by 0.7–1.8%). In the randomised controlled trials of calcium supplements, the increases in BMD were present by one year, but there were no further subsequent increases. Thus the increases from baseline at both two and over two and half years at each site were similar to the increases at one year. The increases in BMD with dietary sources of calcium were similar to the increases with calcium supplements, except at the forearm, in both direct comparisons of the two interventions in multi-arm studies and in indirect comparisons of the two interventions through subgroup analyses. The increases in BMD were similar in trials of calcium monotherapy and CaD, consistent with a recent meta-analysis reporting that vitamin D monotherapy had no effect on BMD. There were no differences in changes in BMD in our subgroup analyses between trials with calcium doses of ≥1000 mg/day and <1000 mg/day or doses of ≤500 mg/day and >500 mg/day, and in populations with baseline dietary calcium intake of <800 mg/day and ≥800 mg/day. Overall, the results suggest that increasing calcium intake, whether from dietary sources or by taking calcium supplements, provides a small non-progressive increase in BMD, without any ongoing reduction in rates of BMD loss beyond one year. The similar effect of increased dietary intake and supplements suggests that the non-calcium components of the dietary sources of calcium do not directly affect BMD.
The strength of this meta-analysis is its comprehensive nature. We included 59 randomised controlled trials and assessed the effects of both dietary calcium sources and calcium supplements on BMD at five skeletal sites and at three time points. The size of the review permitted a comparison of the effects on BMD of different sources of calcium—dietary sources or supplements—and also the effects in important subgroups such as those defined by dose of calcium, use of co-administered vitamin D, and baseline clinical characteristics. The results are consistent with those from an earlier meta-analysis of 15 randomised controlled trials of calcium supplements, which reported an increase in BMD of 1.6–2.0% over two to four years.
An important limitation is that BMD is only a surrogate for the clinical outcome of fracture. We undertook the review, however, because many of the subgroup analyses in the dataset of trials with fracture as an endpoint have limited power, and a comparison between randomised controlled trials of dietary sources of calcium and calcium supplements with fracture as the endpoint is not possible because only two small randomised controlled trials of dietary sources of calcium reported fracture data. Another limitation is that in 60% of the meta-analyses, statistical heterogeneity between the studies was high (I>50%). This indicates substantial variability in the results of included trials, although this was often because of the presence of a small number of outlying results. Subgroup analyses generally did not substantially reduce or explain the heterogeneity. We used random effects meta-analyses that take heterogeneity into account, and their results should be interpreted as reflecting the average result across the group of trials.
The absence of any interaction with baseline dietary calcium intake or a dose-response relation suggests that increasing intake through dietary sources or through supplements does not correct a dietary deficiency (in which case greater effects would be seen in those with the lowest intakes or the highest doses). An alternative possibility is that increasing calcium intake has a weak anti-resorptive effect. Calcium supplements reduce markers of bone formation and resorption by about 20%, and increasing milk intake also reduces bone turnover by a similar amount. Suppression of bone turnover by this amount might lead to the small observed increases in BMD.
Increases in BMD of about 1–2% over one to five years are unlikely to translate into clinically meaningful reductions in fractures. The average rate of BMD loss in older post-menopausal women is about 1% a year. So the effect of increasing calcium intake is to prevent about one to two years of normal BMD loss, and if calcium intake is increased for more than one year it will slow down but not stop BMD loss. Epidemiological studies suggest that a decrease in BMD of one standard deviation is associated with an increase in the relative risk of fracture of about 1.5–2.0. A one standard deviation change in BMD is about equivalent to a 10% change in BMD. Based on these calculations, a 10% increase in BMD would be associated with a 33–50% reduction in risk of fracture. Therefore, the 1–2% increase in BMD observed with increased calcium intake would be predicted to produce a 5–10% reduction in risk of fracture. These estimates are consistent with findings from randomised controlled trials of other agents. The modest increases in BMD with increased calcium intake are smaller than observed with weak anti-resorptive agents such as etidronate and raloxifene. Etidronate, however, does not reduce vertebral or non-vertebral fractures, and raloxifene reduces vertebral but not non-vertebral fractures. In contrast, potent anti-resorptive agents such as alendronate, zoledronate, and denosumab increase BMD by 6–9% at the spine and 5–6% at the hip over three years. These changes are associated with reductions of 44–70% in vertebral fracture, 35–41% in hip fracture, and 15–25% in non-vertebral fractures. The magnitude of fracture reduction predicted by the small increases in BMD we observed with increased calcium intake are also consistent with the findings of our systematic review of calcium supplements and fracture. We observed small (<15%) inconsistent reductions in total and vertebral fracture overall but no reductions in fractures in the large randomised controlled trials at lowest risk of bias and no reductions in forearm or hip fractures.
The large number of randomised controlled trials that studied increased calcium intake and BMD and the consistency of the results across different populations in studies using higher or lower doses of calcium and in studies of dietary calcium sources or calcium supplements does not reveal any obvious gaps in the evidence. Any future trials conducted should have a strong rationale as to why the results are likely to differ from the large body of existing trial evidence. It is usually recommended that anti-resorptive agents are co-prescribed with calcium and vitamin D, although randomised controlled trials of such agents have shown reductions in risk of fracture and the expected increases in BMD without the co-administration of calcium and vitamin D. Randomised controlled trials clarifying the role of calcium and vitamin D in individuals using anti-resorptive agents might be valuable. In subgroup analyses, we stratified trials by thresholds of baseline dietary calcium intake (800 mg/day) and 25-hydroxyvitamin D (50 nmol/L). The clinical consequences of low calcium intake or vitamin D status such as osteomalacia, however, probably occur only at much lower thresholds, and there might also be interactions between calcium intake and vitamin D status. Analyses of individual patient data would be valuable in exploring these issues further.
In summary, increasing calcium intake from dietary sources increases BMD by a similar amount to increases in BMD from calcium supplements. In each case, the increases are small (1–2%) and non-progressive, with little further effect on BMD after a year. Subgroup analyses do not suggest greater benefits of increasing calcium intake on BMD in any subpopulation based on clinically relevant baseline characteristics. The small effects on BMD are unlikely to translate into clinically meaningful reductions in fractures. Therefore, for most individuals concerned about their bone density, increasing calcium intake is unlikely to be beneficial.
Discussion
Principal Findings
Increasing calcium intake from dietary sources slightly increased bone mineral density (BMD) (by 0.6–1.8%) over one to two years at all sites, except the forearm where there was no effect. Calcium supplements increased BMD to a similar degree at all sites and all time points (by 0.7–1.8%). In the randomised controlled trials of calcium supplements, the increases in BMD were present by one year, but there were no further subsequent increases. Thus the increases from baseline at both two and over two and half years at each site were similar to the increases at one year. The increases in BMD with dietary sources of calcium were similar to the increases with calcium supplements, except at the forearm, in both direct comparisons of the two interventions in multi-arm studies and in indirect comparisons of the two interventions through subgroup analyses. The increases in BMD were similar in trials of calcium monotherapy and CaD, consistent with a recent meta-analysis reporting that vitamin D monotherapy had no effect on BMD. There were no differences in changes in BMD in our subgroup analyses between trials with calcium doses of ≥1000 mg/day and <1000 mg/day or doses of ≤500 mg/day and >500 mg/day, and in populations with baseline dietary calcium intake of <800 mg/day and ≥800 mg/day. Overall, the results suggest that increasing calcium intake, whether from dietary sources or by taking calcium supplements, provides a small non-progressive increase in BMD, without any ongoing reduction in rates of BMD loss beyond one year. The similar effect of increased dietary intake and supplements suggests that the non-calcium components of the dietary sources of calcium do not directly affect BMD.
Strengths and Limitations of the Study
The strength of this meta-analysis is its comprehensive nature. We included 59 randomised controlled trials and assessed the effects of both dietary calcium sources and calcium supplements on BMD at five skeletal sites and at three time points. The size of the review permitted a comparison of the effects on BMD of different sources of calcium—dietary sources or supplements—and also the effects in important subgroups such as those defined by dose of calcium, use of co-administered vitamin D, and baseline clinical characteristics. The results are consistent with those from an earlier meta-analysis of 15 randomised controlled trials of calcium supplements, which reported an increase in BMD of 1.6–2.0% over two to four years.
An important limitation is that BMD is only a surrogate for the clinical outcome of fracture. We undertook the review, however, because many of the subgroup analyses in the dataset of trials with fracture as an endpoint have limited power, and a comparison between randomised controlled trials of dietary sources of calcium and calcium supplements with fracture as the endpoint is not possible because only two small randomised controlled trials of dietary sources of calcium reported fracture data. Another limitation is that in 60% of the meta-analyses, statistical heterogeneity between the studies was high (I>50%). This indicates substantial variability in the results of included trials, although this was often because of the presence of a small number of outlying results. Subgroup analyses generally did not substantially reduce or explain the heterogeneity. We used random effects meta-analyses that take heterogeneity into account, and their results should be interpreted as reflecting the average result across the group of trials.
Implications of Findings
The absence of any interaction with baseline dietary calcium intake or a dose-response relation suggests that increasing intake through dietary sources or through supplements does not correct a dietary deficiency (in which case greater effects would be seen in those with the lowest intakes or the highest doses). An alternative possibility is that increasing calcium intake has a weak anti-resorptive effect. Calcium supplements reduce markers of bone formation and resorption by about 20%, and increasing milk intake also reduces bone turnover by a similar amount. Suppression of bone turnover by this amount might lead to the small observed increases in BMD.
Increases in BMD of about 1–2% over one to five years are unlikely to translate into clinically meaningful reductions in fractures. The average rate of BMD loss in older post-menopausal women is about 1% a year. So the effect of increasing calcium intake is to prevent about one to two years of normal BMD loss, and if calcium intake is increased for more than one year it will slow down but not stop BMD loss. Epidemiological studies suggest that a decrease in BMD of one standard deviation is associated with an increase in the relative risk of fracture of about 1.5–2.0. A one standard deviation change in BMD is about equivalent to a 10% change in BMD. Based on these calculations, a 10% increase in BMD would be associated with a 33–50% reduction in risk of fracture. Therefore, the 1–2% increase in BMD observed with increased calcium intake would be predicted to produce a 5–10% reduction in risk of fracture. These estimates are consistent with findings from randomised controlled trials of other agents. The modest increases in BMD with increased calcium intake are smaller than observed with weak anti-resorptive agents such as etidronate and raloxifene. Etidronate, however, does not reduce vertebral or non-vertebral fractures, and raloxifene reduces vertebral but not non-vertebral fractures. In contrast, potent anti-resorptive agents such as alendronate, zoledronate, and denosumab increase BMD by 6–9% at the spine and 5–6% at the hip over three years. These changes are associated with reductions of 44–70% in vertebral fracture, 35–41% in hip fracture, and 15–25% in non-vertebral fractures. The magnitude of fracture reduction predicted by the small increases in BMD we observed with increased calcium intake are also consistent with the findings of our systematic review of calcium supplements and fracture. We observed small (<15%) inconsistent reductions in total and vertebral fracture overall but no reductions in fractures in the large randomised controlled trials at lowest risk of bias and no reductions in forearm or hip fractures.
The large number of randomised controlled trials that studied increased calcium intake and BMD and the consistency of the results across different populations in studies using higher or lower doses of calcium and in studies of dietary calcium sources or calcium supplements does not reveal any obvious gaps in the evidence. Any future trials conducted should have a strong rationale as to why the results are likely to differ from the large body of existing trial evidence. It is usually recommended that anti-resorptive agents are co-prescribed with calcium and vitamin D, although randomised controlled trials of such agents have shown reductions in risk of fracture and the expected increases in BMD without the co-administration of calcium and vitamin D. Randomised controlled trials clarifying the role of calcium and vitamin D in individuals using anti-resorptive agents might be valuable. In subgroup analyses, we stratified trials by thresholds of baseline dietary calcium intake (800 mg/day) and 25-hydroxyvitamin D (50 nmol/L). The clinical consequences of low calcium intake or vitamin D status such as osteomalacia, however, probably occur only at much lower thresholds, and there might also be interactions between calcium intake and vitamin D status. Analyses of individual patient data would be valuable in exploring these issues further.
Conclusions
In summary, increasing calcium intake from dietary sources increases BMD by a similar amount to increases in BMD from calcium supplements. In each case, the increases are small (1–2%) and non-progressive, with little further effect on BMD after a year. Subgroup analyses do not suggest greater benefits of increasing calcium intake on BMD in any subpopulation based on clinically relevant baseline characteristics. The small effects on BMD are unlikely to translate into clinically meaningful reductions in fractures. Therefore, for most individuals concerned about their bone density, increasing calcium intake is unlikely to be beneficial.
Source...