Whether sodium intake has an impact on bone health has been studied since Goulding observed that sodium intake affected bone mass in animals [1
]. Since then, most studies have observed significant correlations between 24-h measures of urinary calcium and sodium [3
]. This is because renal calcium reabsorption is influenced by the concurrent degree of sodium reabsorption [6
]. Most evidence supports an increase in urinary calcium of about 1 mmol per 100 mmol of sodium excretion, which equates to an increase of 40 mg of calcium excreted in the urine for every 2300 mg of dietary sodium, although variability does exist [3
Over time, sodium-induced calciuria may have the potential to negatively impact bone mineral density (BMD) [3
]. It has been suggested by Heaney [3
] that the impact of sodium on bone may depend on adequacy of calcium intake: at recommended calcium intakes, typical North American sodium intakes would have little effect on calcium balance; however, at low calcium intakes the adaptive increase in intestinal calcium absorption may be insufficient to offset the additional urinary calcium losses. Data are not consistent with regard to whether the association differs at low versus
high calcium intakes. Urinary sodium and calcium were associated at high but not low calcium intakes among healthy elderly persons [8
]. Conversely, in studies of postmenopausal women, urinary sodium and calcium were associated at low but not high calcium intakes [9
]. Few data are available regarding whether calcium intake modulates the relationship between urinary calcium and sodium in young women.
We recently conducted a 2-year study of dietary attitudes, ovulatory function, and bone density in healthy young women [11
]. At the 2-year follow-up, 24-h urine collections were analyzed for sodium and calcium. Herein, we report the cross-sectional relationship between 24-h urinary sodium and urinary calcium excretion, and examine whether the association differs between those with higher versus
lower calcium intakes. Associations with BMD and 2-year BMD change were also assessed.
Numerous studies indicate that higher dietary sodium intake (assessed by 24-h urinary sodium) may increase the excretion of calcium [3
]. Over time, higher urinary calcium losses could negatively impact bone density [7
] although few studies have directly examined associations between sodium and bone, and even fewer of these have been conducted in premenopausal women. As it has been suggested that the attainment of peak bone mass may be more relevant to future osteoporosis risk than bone loss in later life [17
], the relationship between sodium and calcium may be particularly relevant in this group. Thus, we assessed potential relationships between 24-h urinary sodium, urinary calcium, aBMD and 2-year aBMD change among 102 non-obese, healthy young women.
We found significant correlations between urinary sodium and urinary calcium. The magnitude of the association was similar to that previously reported [3
]: for every 100 mmol (2300 mg) increase in sodium excretion, calcium excretion increased by 1.1 mmol (44 mg). In our participants, the relationship was stronger among those with lower calcium intakes and became nonsignificant in women with higher calcium intakes. Our findings are thus consistent with those of Nordin and Polley [9
] and Carbone et al.
], who also observed that sodium and calcium excretion were associated at lower but not higher calcium intakes. Carbone et al.
suggest that sodium-coupled calcium transport in the kidney may predominate at low calcium intakes, but that at higher intakes, sodium-independent mechanisms may become more important, thus decreasing the link between urinary sodium and calcium excretion [10
]. However, the mechanism that could underlie this explanation remains to be clarified. In contrast, Dawson-Hughes et al.
reported that calcium and sodium excretion were correlated at intermediate and high intakes, but not at low intakes [8
]. That observation was most consistent among men; in women, the association was significant across all four quartiles of calcium intake, and disappeared only at calcium intakes below 300 mg/day [8
]. Virtually all of the participants in the present study had intakes above that level.
The potential implications of sodium-induced calciuria for bone are likely to be more serious in those with low calcium intakes, who may be unable to increase calcium absorption to fully compensate for increased urinary losses. For example, Heaney [3
] noted that to offset the average urinary calcium loss of 1 mmol (40 mg) associated with an increased sodium intake of 100 mmol (2300 mg), gross calcium absorption efficiency would need to increase to 34% (from 25%) in those with intakes of 600 mg/day, and to about 50% (from 37%) in those with intakes of 300 mg/day-and that this may not be possible. However, at intakes of 1200 mg/day, absorption efficiency would only need to increase from to 23% (from 20%) [3
]. Empirical support for the idea that high calcium intakes may protect against high sodium intakes is provided by the study of Ilich et al.
]. In a 3-year prospective study, postmenopausal women were randomly assigned to maintain usual sodium intake of about 3000 mg/day or to reduce intake to 1500 mg/day. All women also received calcium supplements, and total calcium intake averaged over 1300 mg/day. Because compliance with the sodium intervention was not high, results were reported by tertile of observed urinary sodium excretion rather than by initial group assignment. No negative associations between urinary sodium and bone density were observed [20
]. This suggests that, at least in postmenopausal women with high calcium intakes, sodium intake does not adversely affect bone.
Most women, however, have calcium intakes well below 1300 mg/day, and the estimated prevalence of inadequacy (intakes below requirements) is high. In both the United States and Canada, about 50% of premenopausal women had calcium intakes from food alone below the Estimated Average Requirement of 800 mg/day [21
], as estimated in NHANES 2001-2002 and the 2004 Canadian Community Health Survey Cycle 2.2 [22
]. To provide context for the potential impact of our findings for women with low calcium intakes, we used our data to predict the effect of reducing sodium intake from our participants’ average intake of 2942 mg/day (based on 24-h urinary excretion) to the Institute of Medicine’s Tolerable Upper Intake Level of 2300 mg/day [24
]. Using the regression equation derived from our data (Urine Ca (mg) = 80.5 + 0.019 × urine sodium (mg)), urinary calcium excretion would be predicted to decrease by 12 mg/day. Assuming no compensatory changes in absorption, that would translate to preventing a loss of about 4 g of calcium on an annual basis (12 mg/day × 365 day/year), or 10 g of bone mineral content, given that bone mineral is about 40% calcium by weight. Whole body bone mineral content of our participants averaged 2300 g; thus, preventing a loss of 10 g equates to a 0.4% bone sparing effect over one year. Clearly, this estimate is fraught with assumptions, but if sustained over decades, suggests that the impact on fracture risk could be meaningful.
Strong experimental or prospective evidence is not available regarding whether sodium intake impacts on bone health in younger women, and the available data are mixed. We found an association between urinary sodium excretion and bone density at the hip in women with calcium intakes below the median of 506 mg/1000 kcal. Supportive data were also obtained by Jones et al.
]: In a combined group of pre-and post-menopausal women, they reported that urinary sodium correlated positively with urinary deoxypyridinoline, a bone resorption marker, and was negatively correlated with BMD in unadjusted but not adjusted analyses. Conversely, in a crossover trial of high versus
low sodium intake for one week, no changes in bone resorption or calcium absorption were observed in 11 premenopausal women, although urinary calcium excretion increased [26
]. Similarly, in a randomized seven-week trial of reduced sodium intake that included 15 young women (as well as 14 men), urinary sodium decreased but no group differences were seen in serum bone turnover markers [27
]. The small sample sizes and short duration of these studies, however, may have limited their power.
Although our sample size was reasonably large, our study also has limitations. Most notably, we had only a single measure of 24-h sodium excretion, and multiple collections are necessary to determine an individual’s “usual” excretion (and thus intake) with desirable accuracy. Although food frequency questionnaires estimate usual nutrient intakes, and sodium excretion was correlated with sodium intake as estimated from the food frequency questionnaire, the association was weak. The Diet History Questionnaire (the FFQ we used) has been validated [28
], but dietary assessment of sodium intake is notoriously challenging. This is largely due to the varying levels of sodium in different varieties of similar products: for example, while different brands or varieties of a particular food (such as tomato sauce for spaghetti) have generally similar energy, macronutrient, and vitamin contents, the sodium content can vary dramatically, by well over 100% [30
]. This is true for almost all foods that have been “processed” to any degree (e.g., foods other than raw fruits and vegetables, unprocessed grains). Furthermore, the FFQ does not assess addition of salt at the table, or in the preparation or cooking of foods. For these reasons, 24-h urine collections are still recommended for assessment of sodium intake, but multiple collections are desirable to capture day-to-day variability in intake.
While our results add to the suggestive evidence that the combination of high sodium intakes and low calcium intakes has adverse effects on bone, controlled trials are needed to provide more definitive conclusions.