点击显示 收起
1 From the Department of Human Nutrition, Centre for Advanced Food Studies, Royal Veterinary and Agricultural University, Frederiksberg, Denmark
2 Supported by the Danish government, through the Food Technology Research and Development Program (FØTEK) and the Danish Dairy Research Foundation. 3 Address reprint requests and correspondence to C Mølgaard, Department of Human Nutrition, Centre for Advanced Food studies, Royal Veterinary and Agricultural University, Rolighedsvej 30, DK-1958 Frederiksberg C, Denmark. E-mail: cm{at}kvl.dk.
ABSTRACT
Background: Single-meal studies have established that calcium has an acute inhibitory effect on the absorption of iron. However, there is growing evidence that high calcium intakes do not compromise iron status.
Objective: We evaluated whether long-term calcium supplementation taken with the main meal affected biomarkers of iron status in adolescent girls with high requirements of both iron and calcium.
Design: The study was a randomized, double-blind, placebo-controlled trial of supplementation with 500 mg Ca/d for 1 y among 113 adolescent girls aged 13.2 ± 0.4 y at enrollment. Participants were advised to take the supplement with their evening meal, which usually contributes the majority of dietary iron. Iron status was assessed at baseline and after 1 y of supplementation by measuring hemoglobin and serum concentrations of ferritin and transferrin receptors (TfRs).
Results: The mean (±SD) hemoglobin at enrollment was 134 ± 9 g/L, geometric mean serum ferritin was 26.3 µg/L (interquartile range: 18.6–39.4 µg/L), and serum TfR was 4.19 mg/L (3.52–5.10 mg/L). Daily calcium supplementation had no effect on the least-squares mean concentrations of iron-status markers adjusted for their baseline values (hemoglobin: 136 and 134 g/L, P = 0.31; ferritin: 25.4 and 26.1 µg/L, P = 0.73; TfR: 4.1 and 4.4 mg/L, P = 0.12; and the ratio of TfR to ferritin: 160 and 161 in the calcium and placebo groups, respectively; P = 0.97).
Conclusion: Although it remains to be shown in iron-deficient persons, long-term iron status does not seem to be compromised by high calcium intakes.
Key Words: Iron status • intervention • calcium supplementation • calcium intake • adolescence • puberty • girls
INTRODUCTION
Calcium intake during the rapid growth in adolescence is considered important for primary prevention of osteoporosis, and calcium supplementation is often recommended to young girls with low dietary calcium intake (1). Also during adolescence, iron requirements increase because of rapid growth (increased skeletal muscle and blood volume); in females, moreover, they increase because of menstrual blood loss after menarche (2, 3). It has been shown in single-meal studies that calcium has an acute inhibitory effect on the absorption of both heme and nonheme iron (4-6).
It remains unclear whether whole-body retention of radiolabeled iron from complete diets is affected by calcium intake. Some studies suggest that calcium intake does not affect whole-body retention from complete diets for 4–5 d (7, 8), whereas consumption of dairy products with iron-rich meals during a 10-d period was shown to reduce iron absorption significantly (9). On the basis of the potentially negative influence of high calcium intake on iron absorption, it has been suggested that the intake of calcium (ie, milk or supplements) together with iron-rich meals should be avoided (1, 4). However, iron absorption data from single-meal or short-term studies may not be directly translated into the influence of long-term calcium supplementation on iron status. Indeed, long-term studies of calcium supplementation were conducted in different population groups and unequivocally showed no effect of long-term calcium supplementation on iron-status indicators (3, 6, 10-12). However, as far as we are informed, only one study has evaluated the effect on adolescent girls (3), a vulnerable group regarding iron status. We measured iron-status markers in 12–14-y-old girls participating in a randomized controlled study providing a daily dose of 500 mg calcium for a period of 1 y. As previously reported, calcium supplementation induced a modest increase in bone mineralization (13).
SUBJECTS AND METHODS
Subjects
All girls (n = 1213) with Danish names (both first names and surnames) aged 12 ± 0.5 y from Frederiksberg and Copenhagen municipalities were sent a food-frequency questionnaire (FFQ) for recording of dietary calcium intake. The FFQ included 88 food items from which usual daily intake (in mg/d) during the past month was estimated. The FFQ was previously validated against weighed food records (14). The girls were asked to complete the FFQ and return it together with information on their health, weight, and height. Six hundred eight girls (50%) returned the FFQ. Two groups were selected according to their dietary calcium intake: a group with intakes between the 40th and 60th centiles (1000–1304 mg/d; medium-intake group; n = 121) and a group with intakes below the 20th centile (<713 mg/d; low-intake group; n = 120). After exclusion because of nonwhite racial origin, abnormal weight for height (<3rd or > 97th centile (15)), diseases or intake of drugs with a potential effect on bones, 105 (medium-intake group) and 83 (low-intake group) were eligible for the present study. From these intake groups, 60 and 53 girls, respectively, agreed to participate.
Written informed consent was obtained from all of the participants and their parents. The study was approved (approval no. J nr (KF) 01–033/95) by the Ethics Committee for Copenhagen and Frederiksberg.
Design
The study was a randomized controlled trial designed to evaluate interactions between calcium supplementation and habitual calcium intake by stratifying the randomization according to habitual calcium intake. Girls from each of the 2 calcium intake groups were randomly assigned to receive either 500 mg Ca/d as CaCO3 or placebo (microcrystalline cellulose) daily for 1 y. The girls were advised to take their supplement with their main (evening) meal to increase the compliance. The supplements were identical in appearance, and the intervention code was unknown to the study participants and investigators. At baseline and after 1 y, height and weight were measured, and dietary calcium intake was assessed by FFQ. In addition, blood was analyzed for concentrations of hemoglobin, serum ferritin, and transferrin receptors (TfRs). Compliance was evaluated by tablet count and expressed in percentage [(number of tablets eaten/number of tablets that should have been eaten) x 100].
Anthropometry, body composition, and pubertal stage
Height was measured to the nearest millimeter by using a wall-mounted stadiometer. While the subjects were wearing underclothes, weight was measured to the nearest 0.1 kg with an electronic digital scale. Pubertal development was evaluated as before or after menarche. Most of the girls underwent menarche before the end of the study period, and menarche at baseline was used to indicate the pubertal stage of the girls. Lean body mass (LBM) was measured by using dual-energy X-ray absorptiometry (DXA) with a Hologic 1000/W scanner (Hologic Inc, Waltham, MA) and by using HOLOGIC software (version 5.61; Hologic Inc). Details about the procedure were reported previously (13).
Laboratory methods
Hemoglobin was measured with a Sysmex analyzer (KX-21; Sysmex Corporation, Kobe, Japan). The cutoff for anemia was 110 g/L, according to Danish guidelines (16). The concentration of soluble TfRs was measured in duplicate by using an enzyme immunoassay kit (catalog no. TFX-94; Ramco Laboratories Inc, Houston, TX). Serum ferritin was measured in duplicate by using a fluoroimmunoassay kit (B069-101, DELFIA Ferritin; Wallac, Turku, Finland) that had a detection system based on a europium-labeled monoclonal antibody to human ferritin. Low iron status and iron deficiency were defined as ferritin <20 and <12 µg/L, respectively (17). The ratio of TfR to ferritin (both concentrations in µg/L) was calculated as a measure reflecting total body iron, with TfR:ferritin being negatively associated with body iron.
Statistical analysis
All statistical analyses were performed with SPSS software (version 12.0; SPSS Inc, Chicago). Normally distributed variables are given in means ± SDs, and variables that did not confine to normality are given in geometric means and interquartile ranges. Values of ferritin and TfRs were log10 transformed to obtain normally distributed variables. Difference in iron-status markers were normally distributed and presented as mean differences and 95% CIs. Iron status after 1 y was compared between the 2 intervention groups by using analysis of covariance, in which the baseline value of the dependent variable was included as a covariate to adjust for baseline imbalances. P values < 0.05 were considered significant. Habitual calcium intake x intervention, low iron status x intervention, and change in LBM x intervention interactions were tested, and those with P values < 0.10 were considered significant. The prevalence of low iron status and iron deficiency after 1 y was compared between intervention groups by using logistic regression after adjustment for baseline prevalence.
RESULTS
Of the 113 girls who were included, 111 (98.2%) completed the trial. Median compliance during the intervention period was 86% (ie, 430 mg calcium/d), and there were no significant differences among the 4 groups (P = 0.44, one-way ANOVA). The mean (±SD) hemoglobin concentration at enrollment was 134 ± 9 g/L, the geometric mean (interquartile range) serum ferritin concentration was 26.3 (18.6–39.4) µg/L, and the geometric mean (interquartile range) TfR was 4.19 (3.52–5.10) mg/L. Depleted iron stores (ferritin <12 µg/L) were evident in 11 girls (9.8%), of whom 2 had TfR > 8.3 mg/L, which indicated tissue iron deficiency. According to the Danish cutoff for anemia in children aged 7 to
View this table:
TABLE 1. Baseline characteristics of 113 adolescent girls according to intervention group and habitual calcium intake1
Iron status did not differ according to habitual calcium intake, and the effect of calcium supplementation did not depend on habitual calcium intake (no significant interactions, Table 2). Least-squares means (LSMs) adjusted for baseline values did not differ between the 2 intervention groups for any of the iron-status markers. In other words, there was no effect of calcium supplementation on either of the evaluated iron-status markers after 1 y of calcium supplementation.
View this table:
TABLE 2. Iron-status markers after 1 y of calcium supplementation among 111 young girls according to intervention group and habitual calcium intake1
Furthermore, calcium supplementation did not significantly affect the proportion of girls having low iron status [odds ratio (OR) for ferritin <20 µg/L: 0.94; 95% CI: 0.33, 2.73] or iron deficiency (OR <12 µg/L: 0.58; 95% CI: 0.12, 2.85) after adjustment for baseline iron deficiency.
To explore whether the results were likely to be biased by the skewed distribution of iron status, subjects were stratified according to whether they had replete or low iron status at baseline (the cutoff was 20 µg ferritin/L). The effect of calcium supplementation did not appear to depend on initial iron status according to the estimates in the different strata (data not shown).
LBM expansion during growth may negatively affect the iron stores of adolescent girls, as shown in a study with a 4-y follow-up (3). In our study, the mean absolute change in LBM during 1 y was 1.24 ± 1.94 kg, which corresponds to a monthly gain of 103 g LBM. There was no association between serum ferritin and the 1-y change in LBM either with or without adjustment for baseline ferritin (unadjusted estimate for log ferritin: B = –0.004; 95% CI: –0.023, 0.016; P = 0.71). Similarly, there were no effects of the change in LBM on TfR (unadjusted estimate for log TfR: B = –0.004; 95% CI: –0.017, 0.010; P = 0.62), and nor was change in LBM an effect modifier of supplementation for any of the iron-status markers (LBM x intervention group interaction for hemoglobin, ferritin, and TfR, P > 0.10).
DISCUSSION
The long-term effect of daily calcium supplementation on iron status was investigated in a group of adolescent girls around the time of menarche, when the requirements of iron and calcium are high. Daily calcium supplementation for 1 y had no effect on iron stores assessed by serum ferritin, no effect on tissue iron status assessed by TfR, and no effect on TfR:ferritin, which is closely associated with total body iron estimated from several iron-status markers (18, 19). Moreover, hemoglobin was unaffected by calcium supplementation.
Even though large sample sizes are generally needed to verify negative findings, we believe that it is unlikely that the lack of effect was due to limited power, because the LSM suggested very small differences, if any. Thus, we do not completely reject the hypothesis that calcium supplementation impairs iron status, but our data suggest that any effects are negligible, as indicated by the very small differences in the LSMs. Furthermore, the estimates for the different iron-status markers point in different directions, which supports the argument that the effects are random errors around unity.
The girls were in the age around the time of menarche, when iron stores begin to be used to compensate for menstrual losses. At baseline, menarcheal status was significantly associated with iron status but not with iron deficiency. Furthermore, circulating TfRs tended to be higher, although not significantly (P = 0.10), which indicated a lower tissue iron availability among girls who had undergone menarche.
The study participants were advised to take the supplement with their main (evening) meal, which usually contributes the highest amount of both heme and nonheme iron. Thus, the girls were not advised to separate the calcium supplements from the iron-rich meals as has been recommended (4), and our study illustrates a high iron absorption–inhibiting potential. However, this potential may have been further amplified by supplementing at 2 or 3 meals daily. The inhibiting effect of calcium on iron absorption seen in single-meal studies appears to be dose related in the range from 40 to 300 mg, and there is no further effect at higher doses (20). Therefore, a higher iron-inhibiting potential could theoretically be obtained by distributing the total amount of 500 mg/d across more of the daily meals.
The main concern of our study is whether the estimates are biased by the baseline imbalance in iron status between the intervention groups. If baseline iron status does not interact with the effect of calcium supplementation on iron status, then control for initial values in the analyses will provide valid estimates. If however, the effect of calcium supplementation is influenced by iron status, then our results may be biased.
It is well known that iron status is the main determinant of the efficiency of iron absorption and that increasing absorption is a consequence of decreasing iron status (6, 21). Thus, there is reason to believe that the effect of modifiers of iron bioavailability is more pronounced with decreasing iron status. If this is the case, then the effect of calcium supplementation could be underestimated in our study because iron status was lower in the placebo group than in the calcium group. Because of this imbalance, the study had little power to assess the interaction between iron status and calcium supplementation. On the other hand, restriction of the analysis to subjects with low iron status at baseline did not show a negative influence of calcium supplementation on iron status. In conclusion, we have no reason to doubt the validity of our estimates.
The long-term effect of calcium on iron status was previously investigated in different population groups, including healthy adults of both sexes (6), lactating and nonlactating women (10, 22), and adolescent females (3). None of those studies found any evidence that calcium supplementation compromised long-term iron status assessed by serum ferritin. Similarly, there was no difference in the incorporation of iron in red blood cells between preschool children who consumed low-calcium diets for 5 wk and those who consumed high-calcium diets (12).
None of the above studies used the TfR as an indicator of iron status. However, particularly in populations with low iron stores, it may be useful, in addition to serum ferritin, because the concentration of TfR in serum increases with decreasing iron status beyond the point of depleted iron stores (serum ferritin <12 µg/L), whereas serum ferritin is not very sensitive to further depletion of body iron stores (17, 23). One of the above studies was conducted among Gambian lactating women with a high prevalence of iron deficiency (40% had <12 µg serum ferritin/L; 22). Unfortunately, hemoglobin and serum ferritin were the only iron-status markers used in that study. In the current study, we performed further evaluations of persons with low iron stores (<20 µg/L), among whom the effect would probably be most pronounced. This approach did not suggest that there was a differential effect, for either ferritin or TfRs, among susceptible individuals with low iron stores.
We evaluated whether there was an association between change in LBM and iron status assessed by serum ferritin and TfR. An inverse relation between an increase in LBM and iron status would be plausible because iron stores are used to cover LBM requirements in periods with increased needs due to, for example, construction of muscle tissue. This has been shown in a group of American girls aged 11–y (3). The mean increase in LBM in the current study was only 1.2 kg/y, whereas it was 2.9 kg/y in the American study. Age at enrollment was 13.2 y in the current study and 10.8 y in the American study, and the reason that we did not observe the association may be that most of the girls in our study were old enough to have passed the time for maximum gain of LBM.
The lack of effect on long-term iron status of calcium supplementation, despite a well-established inhibiting effect of iron absorption observed in single-meal studies, may represent a phenomenon that is not specific to the iron x calcium interaction. For example, the enhancing effect of ascorbic acid on iron absorption does not appear to extrapolate to an effect on iron status of long-term high intakes of ascorbic acid (24). Short-term studies of mineral metabolism probably are primarily of mechanistic interest, and public health recommendations should always be based on long-term studies.
In conclusion, this study conducted among adolescent females during a period with high requirements of both iron and calcium did not suggest any adverse effects on iron status of daily calcium supplementation with 500 mg Ca in addition to a relatively high dietary calcium intake. However, before the long-term effects of calcium supplementation on iron status can be fully rejected, it remains to be shown that there is no effect in iron-deficient persons (11).
ACKNOWLEDGMENTS
We are grateful to Birgitte Hermansen for carrying out most of the practical work in the study.
CM and KFM designed the study. CM was responsible for the collection of data. PK was responsible for data analysis and prepared the first draft manuscript. All authors participated in the discussion of results, commented on the manuscript, and approved the final manuscript. None of the authors had any financial or personal conflicts of interests.
REFERENCES