点击显示 收起
1 From the Faculty of Veterinary Science, University of Sydney, Sydney, Australia (KZ, QZ, LHF, AT, XD, DRF, and HG); the Institute for Nutrition and Food Safety, Chinese Center for Disease Control and Prevention, Beijing, China (QZ, GM, and XH); and the Institute of Endocrinology and Diabetes, Children's Hospital at Westmead, Westmead, Australia (CTC)
2 Supported by the Nestlé Foundation and Danone China. 3 Address correspondence to H Greenfield, Faculty of Veterinary Science, University of Sydney, Sydney, NSW 2006, Australia. E-mail: heatherg{at}vetsci.usyd.edu.au.
ABSTRACT
Background: A 2-y school milk intervention trial showed that 330 mL of a dietary milk supplement (fortified with calcium alone or with both calcium and vitamin D) enhanced the growth and bone mineral accretion of Chinese girls aged 10 y at baseline. Girls who received milk fortified with both calcium and vitamin D also had better vitamin D status than did girls who received nothing or girls who received milk fortified only with calcium.
Objective: The aim was to evaluate whether these effects were sustained 3 y after supplement withdrawal.
Design: Anthropometric measures and dietary intake were reassessed in 501 of the 698 girls whose data had been studied at the end of the intervention. As in the intervention phase, total-body bone mineral content and bone mineral density and serum 25-hydroxyvitamin D concentrations were measured in half of these subjects.
Results: At follow-up, 99% of girls had reached menarche, at a mean (±SD) menarcheal age of 12.1 ± 1.1 y. No significant differences in the timing of menarche were observed between the 3 groups (P = 0.6). No significant differences in the changes of total-body bone mineral content and bone mineral density since baseline were observed between the groups. The group receiving calcium-fortified milk had significantly greater gains in sitting height (0.9 ± 0.3%; P = 0.02) than did the control group. The group that received calcium- and vitamin D–fortified milk had 17.1 ± 6.7% lower serum 25-hydroxyvitamin D concentrations than did the control group (P = 0.04), but the difference was attenuated by additional adjustment for physical activity level (14.2 ± 6.7%; P = 0.08).
Conclusion: Milk supplementation during early puberty does not have long-lasting effects on bone mineral accretion.
Key Words: Milk supplementation • growth • bone mineral accretion • vitamin D status • Chinese girls • follow-up study
INTRODUCTION
Bone mass at any time during adult life is determined by peak bone mass achieved at skeletal maturity and the rate of bone loss during aging. An optimal peak bone mass is considered the best protection against age-related bone loss and subsequent fracture risk. Dietary calcium intake is an important nutritional factor that influences the achievement of peak bone mass. Dietary supplementation of calcium has been shown to increase bone mineral accretion in children and adolescents from different ethnic backgrounds (1–9). Enhancement of bone mineral accretion has also been reported in white adolescent girls and Chinese children who received milk or dairy product supplementation (10–13). However, whether the effects of short-term supplementation during childhood or adolescence are maintained after milk or calcium supplement withdrawal is uncertain. Only a few studies have made follow-up measurements, and the results were inconsistent. Three follow-up studies of calcium supplementation (in the form of calcium carbonate or calcium phosphate extract from milk) have found that some of the effects on bone mineral accretion were sustained 1.5–3.5 y after the withdrawal of supplements (14–16). In contrast, other studies have shown that the effects disappeared 1–3 y after the withdrawal of calcium (in the form of calcium carbonate or calcium citrate malate) (17–19) or dairy (12) supplements.
In a school milk intervention study conducted on Beijing girls aged 10 y at baseline, we showed that the subjects who received a 330-mL dietary milk supplement (milk fortified with calcium alone or with both calcium and vitamin D) on school days had greater increases in height (by 0.7–0.8%), sitting height (by 0.7–1.2%), total-body-size–adjusted bone mineral content (BMC; by 1.1–2.5%), and total-body bone mineral density (BMD; by 3.1–5.4%) after 2 y than did the unsupplemented control subjects (20, 21). The group that received milk fortified with both calcium and vitamin D had significantly higher plasma 25-hydroxyvitamin D [25(OH)D] concentrations (
SUBJECTS AND METHODS
Subjects and study design
The characteristics of the study subjects and the methods used in the intervention study were detailed previously (20). The important aspects are described below. Seven hundred fifty-seven girls [
Anthropometric measurements and pubertal staging
Subjects were weighed to the nearest 0.1 kg while wearing light clothing and no shoes with an electronic digital scale (Thinner, Fairfield, WI). Height and sitting height were measured by the same observer (QZ) with the use of the same stadiometer in the hospital where the bone densitometry scans were made; the subjects were measured in bare feet to the nearest 0.1 cm with a body height–measuring device (TG-III Type, No. 6; Machinery Plant, Beijing, China). Breast development and pubic hair development were ascertained according to Tanner's definitions of the 5 stages of puberty (22) during an interview by the same observer (KZ). Menstrual status was also determined by interview.
Dietary intake and physical activity
Dietary intakes were assessed from two 3-d food records (including 2 weekdays and 1 weekend day). Chinese measures of bowls, plates, and spoons of standard size were used to quantify food items with the assistance of a set of food-measure models. Chinese food-composition tables and data entry programs were used to calculate nutrient intakes, as previously reported (20). Physical activity levels over the previous 12 mo were estimated by a questionnaire, which included 12 categories of school-organized activities, 25 categories of spare-time activities (ie, activities undertaken during class breaks, lunchtime recesses, and outside school hours), and an open question at the end. The subjects were instructed to report the following: 1) the type, frequency, and duration of school-organized physical activity and spare-time physical activity; and 2) the extent of any training at sports clubs or in sporting teams over the previous 12 mo. The subjects' school physical activity score, the pattern of spare-time allocation, and their means of traveling to and from school were also recorded.
Bone mineral measurement
Total-body BMC and BMD were assessed in a subset of 277 subjects whose total-body bone measurement data were available both at baseline and after 24 mo of the intervention. Total-body BMC, bone area, and BMD were measured by dual-energy X-ray absorptiometry with the same Norland XR-36 densitometer (Norland, Fort Atkinson, WI) used in the intervention study. Software version 3.94 was used. The densitometer had a variation in precision of <1.0% for the measured bone site at standard speed. A daily quality assurance test was performed over the study period with the use of a manufacturer-supplied hydroxyapatite phantom, and the accuracy error was <1.0%. Two technicians performed the measurements during the follow-up study under supervision of the technician who performed the measurements during the intervention study.
Biochemical analysis
Samples of blood (drawn after the subjects had fasted overnight, between 0630 and 0900) were obtained during March and April (ie, the end of winter), similar to sampling performed during the intervention baseline and at the end of the intervention. Serum 25(OH)D and parathyroid hormone (PTH) concentrations were assessed in a subset of subjects, similar to the intervention study. Samples from different schools or groups were analyzed in the same batch. Complete baseline and follow-up 25(OH)D and PTH data were obtained from 213 and 98 girls, respectively. Serum 25(OH)D concentrations were measured with 125I radioimmunoassay (DiaSorin, Stillwater, MN), because the competitive protein-binding assay (CPBA) that was used in the intervention study was not available in China. The intraassay CV was 4.2% and the interassay CV was 6.6%. A cross-calibration study with 18 human samples carried out in both CPBA (in the laboratory of Faculty of Veterinary Science, University of Sydney, where the intervention study samples were assayed) and radioimmunoassay (in the laboratory of 304 Hospital in Beijing, where the follow-up samples were assayed) showed that the serum 25(OH)D concentration measured by the radioimmunoassay method was, on average, 23% higher than that measured by CPBA (LHF, unpublished data). Serum PTH concentrations were measured by immunometric assay (Immulite Intact PTH; Diagnostic Products Corporation, Los Angeles, CA). The intraassay CV was 2.9% and the interassay CV was 5.0%.
Other measurements
The health history of the subjects and family socioeconomic status were obtained by a general information questionnaire. Posteroanterior X-ray radiographs of the nondominant hand and wrist were taken, and bone age was determined to the nearest 0.1 y according to the Chinese standard (23).
Statistical analysis
Descriptive statistics are reported as means ± SDs and differences as means ± SEMs for all variables, unless otherwise indicated. One-factor analysis of variance and 2-factor repeated-measures analysis of variance were used to compare continuous demographical variables at baseline and follow-up. Post hoc analyses were carried out with Tukey's honestly significant difference test. Differences in the frequency data between groups at the same time period were analyzed by a chi-square test.
To allow for clustering by school at both the primary and secondary school levels, a linear mixed model was used to examine the effects of milk supplementation on growth, bone mineral accretion, and serum 25(OH)D and PTH concentrations during follow-up, with schools defined as random effects (24, 25). To investigate any proportional effects of discrete variables (ie, the supplementation group), the continuous variables (ie, bone and anthropometric measures) were natural log-transformed. Because the variable "group" had 3 outcomes (Ca milk, CaD milk, and control groups), 2 dummy variables (Group1 and Group2) were created to represent this variable, where Group1 = 1 for Ca milk and 0 for CaD milk and control, and Group2 = 1 for CaD milk and 0 for Ca milk and control. These 2 dummy variables were treated as a group and were both in each model together. Outcomes were analyzed with adjustment for the baseline value, baseline pubertal stage, and intervention group. Size-adjusted BMC was introduced in a fuller model to study the effects on bone mineral status independently of bone and body size. In this model, BMC was adjusted for bone area, body weight, and height, and we used the mean values and the differences of these variables at baseline and follow-up. Additional modeling was done to adjust for calcium intake, vitamin D intake, and physical activity level. To test whether there were interactions between intervention groups and the baseline value, baseline pubertal stage, calcium intake, vitamin D intake, and physical activity level, the significance of the relevant interaction terms was tested in each model. The regression parameter bgroup for each intervention group, once multiplied by 100, represents the differences between each of the intervention groups and the control group in percentage change from baseline (26). The differences between the 2 supplemented groups were calculated in a similar way. To show whether the absence of effects was due to reduced statistical power from clustering by school in the linear mixed model, the effects of milk supplementation on growth and bone mineral accretion at follow-up were also evaluated at the person level with a multiple regression model with backward elimination (7, 20). The significance level for test statistics was set at P < 0.05. All data were analyzed by SPSS (version 12.0; SPSS Inc, Chicago, IL) and SAS (SAS for WINDOWS version 9.1; SAS Institute Inc, Chicago, IL).
RESULTS
Of the 698 subjects who completed the milk intervention trial, 504 subjects were reexamined 3 y after the end of the intervention (Figure 1). No significant difference in the dropout rate was observed between the 3 groups (27.3%, 28.5%, and 28.7% for the Ca milk, CaD milk, and control groups, respectively). In each study group, there were no significant differences in anthropometric and bone measures between the subjects who were reexamined and those who dropped out at baseline and at the end of the intervention. Data for the 501 girls whose complete data were obtained are shown in the present study.
FIGURE 1.. Study profile and number of subjects at each stage of the investigation. *Numbers in parentheses are the number of subjects in whom measurements of total-body bone mineral and plasma 25-hydroxyvitamin D [25(OH)D] concentrations, respectively, were conducted at that stage. Ca milk group, calcium-fortified milk group; CaD milk group, calcium- and vitamin D–fortified milk group.
The characteristics of the study and the control groups at presupplementation baseline and 36 mo after the supplement withdrawal are shown in Table 1. At presupplementation baseline, there were no significant differences between the 3 groups in age, bone age, pubertal status, and dietary intakes of calcium, phosphorus, vitamin D, and protein, but the control group had significantly higher milk intakes than the CaD milk group and significantly higher energy intake than the Ca milk group. After the supplement withdrawal (at 36 mo), there were no significant differences in bone age, pubertal status, and dietary intake of milk, calcium, phosphorus, energy, and protein between the 3 groups. Compared with the presupplementation baseline, all 3 groups had significantly higher milk and calcium intakes and lower dietary vitamin D intakes 36 mo after the withdrawal of the milk supplement. During the study, the average time spent on spare-time physical activity was higher in the control group than in the supplemented groups. Ninety-nine percent of the girls had reached menarche at follow-up. No significant differences in mean menarcheal age were observed between the 3 groups (
View this table:
TABLE 1. Characteristics and nutrient intakes of the milk-supplemented and control groups at presupplementation baseline and 3 y after supplement withdrawal1
No significant differences in anthropometric and bone measurements were observed between the 3 groups at presupplementation baseline (Table 2). At the end of intervention, both supplemented groups had significantly greater gains in height, sitting-height, and total-body BMD than the control group (Table 3). The CaD milk group also had significantly greater gains in body weight, BMI, total-body BMC, and size-adjusted BMC than the control group, and significantly greater gains in total-body BMD and size-adjusted BMC than the Ca milk group (Table 3).
View this table:
TABLE 2. Anthropometric, bone, and biochemical variables of the milk-supplemented and control groups at presupplementation baseline1
View this table:
TABLE 3. Effects of milk supplementation on anthropometric and total-body (TB) bone measures 3 y after supplement withdrawal1
Three years after the withdrawal of the milk supplements, the Ca milk group had a significantly greater gain in sitting height than the control group, and this persisted after adjustment for the baseline value, baseline pubertal stage, and clustering by school (Table 3
DISCUSSION
Our study showed that the effects of milk supplementation on total-body bone mineral accretion had disappeared 3 y after the supplement withdrawal, whereas the effects on sitting height were still evident in the group that received milk fortified with calcium. A lower vitamin D status compared with controls was also observed in the group that received milk fortified with both calcium and vitamin D.
Our findings on bone mineral accretion are consistent with the results of 4 previous studies, including 2 studies conducted on Chinese prepubertal children that showed that the greater gains in bone mass in children who received calcium or dairy supplementation for 18–36 mo were not sustained when the subjects were followed-up 12–36 mo after the withdrawal of calcium (17–19) or dairy (12) supplementation. However, 3 other studies have found that the some of the benefits of 1-y calcium supplementation on bone mineral accretion were still evident 1.5–3.5 y after the withdrawal of supplements (14–16).
Bonjour et al (14) found that the greater gain in bone mass in prepubertal girls who received 1-y supplementation of calcium phosphate extract from milk was maintained 3.5 y after the cessation of intervention, and the greater gains in bone mass were mainly due to a greater increase in bone size, which possibly resulted from the anabolic effects of milk calcium salt on the bones. Although anabolic effects of milk on bone were also shown in our study (as persistent effects on axial growth after supplement withdrawal), we did not observe a sustained effect on bone mass. One possible reason for the different findings of the study by Bonjour et al and our study could be that the average background calcium intake of the subjects in the Bonjour et al study was twice that of our subjects, which was 450 mg calcium/d. Because the calcium intake of our subjects could not meet the amount required for growth, the bone-remodeling rate increased and the additional bone mineral deposited during the supplementation phase was mobilized when supplementation was discontinued, and the benefits on bone mineral status therefore disappeared (27, 28).
The interaction between puberty and dietary intervention may play a role in the long-term effects of short-term supplementation. Additional follow-up (7.5 y after supplement withdrawal) of the subjects in the study by Bonjour et al suggested that calcium supplementation before puberty was associated with an earlier onset of menarche, and sustained effects on BMD were observed only in the girls with earlier menarche (29). In the present study, although there was a trend for more girls who received supplemented milk to have reached menarche at the end of the intervention than those in the control group (48.3% compared with 40.1%; P = 0.09), there were no significant differences in age of menarche between the supplemented and control groups at follow-up. Whether the intake of milk calcium salt during prepuberty has a greater effect on menarcheal age or the supplementation amount of the present study of 144 mL milk/d (containing 245 mg calcium) is not high enough to affect the age of menarche deserves additional study. In the present study, any residual effects on bone mineral status could also be obscured by the effect of puberty on bone development. Puberty plays a critical role on bone mineral accretion (30), and our subjects progressed from pre-early puberty to mid-late puberty during the 2-y intervention and 3-y follow-up.
A possible long-term effect of milk supplementation during early puberty (from 10- to 12-y-old) on axial growth was observed in the Ca milk group but not in the CaD milk group. A nonsignificant trend toward a lower gain in sitting height was observed for the CaD milk group compared with the Ca milk group (
A recently published 7-y trial conducted in American girls aged 10.9 y at baseline found that calcium supplementation had a significant positive influence on bone accretion during the pubertal growth spurt, but the effects were diminished in late adolescence (31). This finding suggested that BMD could catch up during the bone consolidation phase to compensate for the compromised bone mineral accretion during growth spurts that is caused by inadequate calcium intake. The study conducted on American girls also showed that the subjects with low calcium intakes might not have a complete catch-up (31). The average calcium intake of American girls in that study was 830 mg/d, which approaches the calcium intake of 957 mg/d required for maximum calcium retention in late adolescence (32). We speculate that the catch-up in bone mineral acquisition may not be complete in our Chinese girls, who have habitual calcium intakes of about 450 mg/d, and thus these girls would have a higher possibility of not achieving optimal peak bone mass than would American girls. The average milk intake of 140 g/d at follow-up was higher than both the presupplementation baseline intake of 120 g/d and the reported intake of urban Chinese girls from Beijing aged 12–14 y [
In conclusion, the present study showed that the effects of 2-y milk supplementation during early puberty on bone mineral accretion and the effects of vitamin D–fortified milk on vitamin D status largely disappeared 3-y after supplement withdrawal. Additional studies that include bone structure (ie, measured by peripheral quantitative computed tomography) as one of the outcome measurements over a longer term are needed to examine whether increased milk intake during pubertal years influences bone size and peak bone mass attainment in Chinese girls with low habitual calcium intakes. Meanwhile, we conclude that short-term milk supplementation during early puberty does not have long-lasting effects on bone mineral accretion in Chinese girls.
ACKNOWLEDGMENTS
We thank the staff of Beijing Xicheng Student Health Institute and the Department of School Nutrition, Institute for Nutrition and Food Safety for their help with organizing the fieldwork, and especially X Zhang, L Zhang, Y Tong, CY Zhang, and H Pan. We also thank J Liu, Q Ouyang, Z Duan, and H Xiong from Beijing No. 304 Army Hospital for their reports on bone densitometry and hand X-rays; TA Dobbins of the School of Public Health, University of Sydney for advice on the mixed-model analysis; and M Phillips of Murray Goulburn Co-operative Co Ltd for his great help in developing and ensuring the quality and timely transportation of the milk supplement during the intervention study. We are grateful for the support of the all the school principals and teachers involved and the cooperation of all the participating students and their parents.
KZ was involved in the study design, data collection, and data analysis and drafted the manuscript. QZ, LHF, and XH were involved in the data collection. LHF performed the biochemical analyses. AT participated in laboratory analysis consultations. GM and XD were involved in the conception and design of the study. HG, DRF, and CTC were involved in the conception, design, and interpretation of the study. All authors contributed to the writing of the manuscript. None of the authors had any conflicts of interest.
REFERENCES