- Email: [email protected]
Milk intake increases bone mineral content through inhibiting bone resorption: Meta-analysis of randomized controlled trials
e-SPEN Journal 8 (2013) e1ee7
Contents lists available at SciVerse ScienceDirect
e-SPEN Journal journal homepage: http://www.elsevier.com/locate/clnu
Original article
Milk intake increases bone mineral content through inhibiting bone resorption: Meta-analysis of randomized controlled trials De Fu Ma a, *, c, Wei Zheng b, c, Ming Ding b, Yu Mei Zhang b, Pei Yu Wang a a b
Department of Social Medicine & Health Education, School of Public Health, Peking University, Xueyuan Road 38, Haidian District, Beijing 100191, PR China Department of Nutrition & Food Hygiene, School of Public Health, Peking University, Beijing 100191, PR China
a r t i c l e i n f o
s u m m a r y
Article history: Received 16 December 2010 Accepted 16 October 2012
Background and aims: To clarify the effects of milk intake on bone mineral density (BMD) ，bone mineral content (BMC) and bone metabolism markers. Methods: We identified randomized controlled trials related to urinary N-telopeptide cross-links of type I collagen (NTx), serum osteocalcin, BMD and BMC listed on MEDLINE (January 1966eNovember 2010), Science Citation Index and PUBMED (updated till November 2010), China National Knowledge Infrastructure (1979eNovember 2010) etc. Results: Eleven studies with a total of 2397 subjects were selected for meta-analysis. The osteocalcin in subjects who consumed milk decreased by 5.9 (95% confidence interval (CI) 7.23, 4.57) ng/ml in comparison to that in control treatment. Milk intake vs control treatment significantly decreased urine NTx by 5.41 (95% CI 10.35, 0.47) nmol/mmol. Moreover, the total body BMC in subjects who consumed milk increased significantly by 40.32 (95% CI 17.58, 63.05) g in comparison to that in control treatment. Milk intake vs control treatment increased total body BMD by 0.01 (95% CI -0.02, 0.03) g/cm2 with borderline significance. Conclusions: Milk intervention significantly attenuates bone loss through inhibiting bone metabolism. Crown Copyright Ó 2012 Published by Elsevier Ltd on behalf of European Society for Clinical Nutrition and Metabolism. All rights reserved.
Keywords: Milk Osteoporosis Urinary N-telopeptide cross-links of type I collagen Bone mineral content
Introduction Osteoporosis is an increasing public health problem worldwide. The fractures caused by the bone density reduction and bone microstructure alteration result in a lower life quality. Calcium supplementation, protein supplementation, exercise, and hormone replacement therapy and so on have long been demonstrated to be the main strategies to prevent osteoporosis. Clinical trials have shown that calcium supplementation combined with Vitamin D can increase bone mineral density (BMD) and prevent bone loss in elderly women.1 However, the association between dietary protein and osteoporosis are controversial. A prospective study indicated that higher consumption of animal protein had an increased risk of forearm fracture.2 While a systematic review on 31 cross-sectional
Abbreviations: RCTs, randomized controlled trials; BMD, bone mineral density; BMC, bone mineral content; NTx, N-telopeptide cross links of type I collagen; MBP, milk basic protein. * Corresponding author. Tel.: þ86 10 82801743; fax: þ86 10 82802002. E-mail addresses: [email protected] (D.F. Ma), [email protected] (W. Zheng). c Contributed equally to the work; Wei Zheng is the Co-first author.
surveys showed a small positive effect of protein intake on lumbar spine BMD.3 Historically, milk has been widely consumed because of its excellent nutritional value.4 Milk that contains several factors such as bioactive peptides, calcium, growth factors related to bone metabolism might affect both bone formation and bone resorption.4 In particular, milk is a good source of bioavailable calcium compared with other food sources. During the latest several decades, RCTs on bone health suggested that milk intervention might increase BMD and bone mineral content (BMC). In the large, third National Health and Nutrition Examination Survey (NHANES III), they found that among women aged 20e49 year, BMC was 5.6% lower in those who consumed <1 serving of milk/d (low intake) than that in those who consumed >1 serving/d (high milk) during childhood (p < 0.01).5 Cadogan et al. reported that the intervention group consuming, on average, 300 ml milk a day throughout the intervention trial had greater increases of BMD (p ¼ 0.017) and BMC (p ¼ 0.009) compared with the control group.6 Liu et al. found that 45 g milk powder supplementation could increase serum osteocalcin (a biochemical marker for bone formation, p < 0.05) and decrease urinary hydroxyproline (a biochemical marker for bone
2212-8263/$36.00 Crown Copyright Ó 2012 Published by Elsevier Ltd on behalf of European Society for Clinical Nutrition and Metabolism. All rights reserved. http://dx.doi.org/10.1016/j.clnme.2012.10.005
e2
D.F. Ma et al. / e-SPEN Journal 8 (2013) e1ee7
resorption, p < 0.05) in Chinese women.7 Aoe et al. found that 40 mg of milk basic protein (MBP) supplementation was able to significantly suppressed the urinary excretion of cross-linked Nteleopeptides of type-I collagen (NTx, a biochemical marker for bone resorption) in healthy adult women.8 However, the effects of milk intake on BMD, BMC, and bone metabolism appear inconsistent in RCTs. Moreover, it’s difficult to clarify whether the beneficial effects are from milk or from calcium because that calcium fortified milk was widely used in the studies on bone health and milk. Thus, a statistical method of combining these diverse data is needed to evaluate the usefulness of milk therapy. Meta-analysis combines or integrates the results of several studies to provide an increased statistical power for the quantitative identification of trends.9 To clarify the effects of milk intake on bone health, we identified all RCTs related to the effects of milk on bone mass and bone turnover markers and analyzed the effects of milk or calcium fortified milk on bone metabolism quantitatively. Materials and methods MEDLINE (January 1966eNovember 2011), the Cochrane Controlled Trials Register, EMBASE (1985eNovember 2011), Science Citation Index and PUBMED (updated till November 2011), China National Knowledge Infrastructure (1979eNovember 2011), VIP Database for Chinese Technical Periodicals (1989eNovember 2011), and Wanfang database (1982eNovember 2011) were used to search articles (in English and Chinese) that described RCTs investigating the effect of milk on bone metabolism. In the RCTs, BMD and BMC were generally measured to assess the bone mass, and serum osteocalcin was generally used as a bone formation marker, urine NTx was generally used as bone resorption marker.6,10,11 Hence, titles, abstracts, and subject headings in the databases were searched with the use of the following Boolean phrases: (“bone” or “osteoporosis” or “bone mass” or “BMD” or “BMC” or “osteocalcin” or “NTX”) and (“milk” or “fortified milk”). We carried out a broad search for all studies with the Boolean phrases “diet” and (“osteoporosis” or “bone metabolism”). We also examined all references of related reviews and papers identified by the search. Additionally, we tried to contact the authors for the obtaining of unpublished data. Studies were selected for analysis if they met all of the following criteria: 1) subjects ingested milk products for at least 1 week; 2) the RCTs included a parallel control group; 3) Total body BMD, total body BMC, NTx or osteocalcin was used as an index of bone turnover. Studies were excluded if they are lack of indices of interest, lack of a control group, insufficient original data or baseline values. If the study sample was found to overlap with that in another article or if two articles described aspects of the same study, only the publication with the largest sample was used. If the study reported some comparisons, we included all comparisons in the meta-analysis. Two researchers (De Fu Ma and Wei Zheng) extracted data independently. A data collection form was designed, and data were entered into the form twice to reduce input errors. The items entered in the form included participant characteristics, treatment duration, interventional design, and values of relevant indices (Total body BMD, total body BMC, NTx and osteocalcin) before and after milk or control treatments. Jadad Scores were used to measure the quality of the RCTs.12 A numerical score between 0 and 5 was assigned as a rough measure of study design and reporting quality, 0 being the weakest and 5 the strongest. One point was assigned if the trial was either randomized or double-blind or in the case of an accurate description of the drop-out patients. Moreover, further points were given if randomization and blinding procedures were appropriate, whereas, instead, points were subtracted in the case of inappropriate descriptions of the same procedures. An overall score more
than 3 indicated a good quality study. Two researchers rated study quality independently. There was 90% agreement on Jadad Scores. If the researchers disagreed, a final score was reached by discussion. In this meta-analysis, we obtained the mean differences from the post-randomization baseline to after-treatment values for each trial and calculated the pooled standard deviation of the mean differences according to the standard method of Cochrane handbook.13 Weighted mean difference was calculated by subtracting the mean difference of the control group from that of the treatment group. The inverse variance method was used to pool the weighted mean difference with STATA software (version 9.2; Stata Corp., College Station, TX, USA).14 To assess the heterogeneity (apparent diversity in weighted mean differences across studies), we conducted a test based on c2 distribution (p < 0.05 is considered significant). Random-effects model was used as the method of combination for all the analyses showing significant heterogeneity .The funnel plot was performed to detect publication bias. In addition, we performed subgroup analyses for osteocalcin by 4 variables one at a time: form of intervention, treatment length, race, and intervention subjects to identify impact factors that can influence the effects of milk intake according to the characters of the data. Results The trial flow chart was illustrated in Fig. 1. Our literature search identified 39 RCTs including 2 of them obtained from the reference lists. We also tried to contact the authors for unpublished data, but unfortunately none was obtained. 28 studies were excluded because of lack of indices of interest, lack of a control group, insufficient original data or baseline values. Thus, 11 studies (8 in English, 3 in Chinese) with a total of 2397 subjects were included in this meta-analysis.4,6,10,11,15e22 The characteristics of the trials included were shown in Table 1. One study had quality score of four, 7 studies had quality score of three, and 3 studies had quality score Articles retrieved from the original search (n = 3032) Articles excluded after the preliminary screening by looking through the titles and abstracts (n = 2533) Potentially relevant articles identified and screened for retrieval (n = 499) Articles excluded because not RCTs (n = 460) RCTs retrieved for more detailed evaluation (n = 39) RCTs excluded: did not meet the inclusion criteria (n = 28) RCTs included in meta-analysis (n = 11) RCTs withdrawn because did not report: Serum osteocalcin (n = 6) Urinary N-telopeptide cross-links of type I collagen (n = 8) Total body bone mineral density (n =5) Total body bone content (n =7)
RCTs with usable information on: Serum osteocalcin (n = 5) Urinary N-telopeptide cross-links of type I collagen (n = 3) Total body bone mineral density (n =6) Total body bone content (n =4) Fig. 1. Results of search for eligible studies.
D.F. Ma et al. / e-SPEN Journal 8 (2013) e1ee7
e3
Table 1 Characteristics of the 11 selected randomized controlled trials of milk intake and bone turnover. Study
Length of treatment (months)
No
Subjects (years)
Race
Intervention
Indices
Quality score
Budek et al. (2007) Cadogan et al. (1997)
1/4 18
24 82
Male (8) Female (12.2)
Western Western
M: M 1.5 L/day. Con: meat M: M 0.3 L/day. Con: habitual diet
2a,b 3a
Du et al. (2004)
24
757
Female (10)
Asian
Green et al. (2002)
1
50
Western
NTx
3a
Kristensen et al. (2005) Kruger et al. (2006)
1/3 4
22 82
Female (Post, 68) Male (24) Female (27.5)
Osteocalcin NTx Osteocalcin
3a 3a
Liu et al. (2002)
24
375
Female (10)
Asian
M þ Ca: M 330 ml þ Ca 560 mg/day M þ Ca þ VitD: M 330 ml þ Ca 560 mg þ VitD 5 or 8 ug/day. Con: habitual diet M þ Ca: M 50 mg þ Ca 1200 mg/day. Con: apple drink M: 2.5 l/day. Con: coca cola M þ Ca: M 50 g þ Ca 1000 mg/day M þ Ca þ VitK: M 50 g þ Ca 1000 mg þ VitK 80 ug/day. Con: normal diet M: 330 ml/day. Con: normal diet
Osteocalcin Osteocalcin NTx. Total body BMD Total body BMC. Total body BMD
24
351
Female (11)
Asian
BMD. BMC BMD. BMC
2a,b
Tong et al. (2004)
Total Total Total Total
Volek et al. (2003)
3
28
Male (14.4)
Western
2a,b
Zhang et at. (2006)
60
498
Female (10)
Asian
Total body BMC. Total body BMD Total body BMD
Zhu et al. (2005)
24
128
Female (10)
Asian
Osteocalcin
3a
Western Western
M þ Ca: M 330 ml þ Ca 560 mg/day M þ Ca þ VitD: M 330 ml þ Ca 560 mg þ VitD 8 ug/day Con: habitual diet M: M 708 ml/day. Con: Apple juice/Grape juice M þ Ca: M 330 ml þ Ca 560 mg/day M þ Ca þ VitD: M 330 ml þ Ca 560 mg þ VitD 8 ug/day. Con: normal diet M þ Ca: M 330 ml þ Ca 560 mg/day M þ Ca þ VitD: M 330 ml þ Ca 560 mg þ VitD 5e8 ug/day. Con: habitual diet
body body body body
4b
3a
3a,b
Post, post-menopause; M, milk; Ca, calcium; VitD, vitamin D; VitK, vitamin K1; Con, control group; NTx, N-telopeptide cross-links of type I collagen; BMD, bone mineral density; BMC, bone mineral content. Highest total score is 5. a points were deleted from Quality Score because the method of blinding was either not described or not appropriate; b points were deleted from Quality Score because there was no description of random.
of two. In 6 of these studies, calcium fortified milk was used, and normal milk (whole, semi-skimmed, and skimmed milk) was used in other studies. The duration of treatment varied widely, ranging from one week to 60 months including 6 studies with a duration exceeding 6 months. 6 studies were performed in Caucasian subjects and 5 studies in Asian subjects. 3 of these 11 studies were carried out in male and other studies were carried out in female including 1 studies performed in post-menopausal women. 8 of these 11 studies were carried out in children and other studies were carried out in adults. In all RCTs, subjects were healthy and were not undergoing any other therapy for osteoporosis; additional calcium supplements were restricted during treatment. All studies reported no significant differences regarding baseline characteristics such as age, body mass index, osteocalcin, NTx, BMD and BMC between groups. There were also no significant weight changes or negative side effects reported. In the present analysis, 5 studies with 15 comparisons reported values of osteocalcin before and after milk or control treatments. When we combined the 15 comparisons, milk decreased osteocalcin by 5.90 (95% confidence interval (CI) 7.23, 4.57; p < 0.00001) ng/ml (Fig. 2). Further analysis of the effects of milk on osteocalcin is shown in Table 2.Milk intake significantly decreased osteocalcin when the studies that used calcium fortified milk as treatment were excluded (7.90, 95% CI [14.05, 1.75] ng/ml; p ¼ 0.01). The decrease in osteocalcin was still significant when the studies with milk intake of more than 6 months were not included in the analysis (6.18, 95% CI [7.54, 4.82] ng/ml; p < 0.00001). Interestingly, we found that milk intake had significant effects on osteocalcin in Caucasian (5.98, 95% CI [7.32, 4.63] ng/ml; p < 0.00001) but not Asian people (2.09, 95% CI [11.57, 7.40] ng/ml; p ¼ 0.67) after sub-group analysis. Because the hormone levels between female and male are quite different, we performed sub-group analysis according to the intervention subjects. Milk presented a larger mean difference on decreasing osteocalcin in
male (17.28, 95% CI [25.69, 8.87] ng/ml; p < 0.0001) although the number of intervention subjects is small. In addition, we found that milk has a significant effect on osteocalcin in adults (5.97, 95% CI [7.35, 4.60] ng/ml; p < 0.0001) but not in children (4.67, 95% CI [10.28,0.93] ng/ml; p ¼ 0.1) by performing sub-group analysis by children vs adults. In the present analysis, 3 studies with 8 comparisons reported values of NTx before and after milk or control treatments. When the 8 comparisons were combined, the mean difference was e5.41 (95% CI 10.35, 0.47; p ¼ 0.03) nmol/mmol (Fig. 3). After removing the studies on calcium fortified milk, we found that the NTx-decreasing effect was still significant (18.91, 95% CI [33.73, 4.09] nmol/mmol; p ¼ 0.01). Because of the low number of studies, we did not perform subgroup analysis for NTx. Moreover, we performed meta-analysis for total body BMC and total body BMD. In the present analysis, 4 studies with 8 comparisons reported values of total body BMC before and after milk or control treatments. When we combined the 8 comparisons, milk significantly increased total body BMC by 40.32 (95% CI 17.58, 63.05; P ¼ 0.0005) g (Fig. 4). The effect was still significant (41.99, 95% CI [6.94, 77.03] nmol/mmol; p ¼ 0.02) when the studies on calcium fortified milk were excluded. There were two studies by Du et al. and Zhang et al. with 1255 subjects (about half of the total subjects) related to BMC and/or BMD.15,20 In order to avoid bias from the two studies, we performed analysis except for the two studies and the effect was similar (41.77, 95% CI [14.82, 68.73] nmol/ mmol; p ¼ 0.002) on BMC. When the 6 studies with 13 comparisons reported values of total body BMD were combined, the mean difference was 0.01 (95% CI 0.02, 0.03; P ¼ 0.63) g/cm2 in total body BMD (in random-effects model), showing an increasing trend (Fig. 4). The BMD-increasing effect after excluding the studies on calcium fortified milk or the studies with large number of subjects was 0.01 (95% CI [0.00, 0.02] g/cm2; p ¼ 0.02) and 0.01, (95% CI [0.00, 0.02] g/cm2; p < 0.0001), respectively. Because of the low
e4
D.F. Ma et al. / e-SPEN Journal 8 (2013) e1ee7
Fig. 2. Weighted mean difference of serum osteocalcin concentration after treatment with milk. Lines correspond to 95% CI. *: the value of the mean and 95% CI of each study, # : reference number.
number of studies, we did not perform subgroup analysis for total body BMC and total body BMD. According to the results from c2 test, only one analysis on total body BMD which was based on all 6 related studies showed significant heterogeneity (P < 0.0001). So this analysis was performed using random-effects model and all other analyses were performed using fixed-effects model. The potential for publication bias was examined by construction of a “funnel plot” of the relation between the standard error and the weighted mean difference (Fig. 5). It did not provide strong evidence of publication bias for either index. Discussion To our knowledge, this is the first meta-analysis that clarified the effects of milk intake on total body BMD, BMC, osteocalcin, and NTx. In the present analysis, there are 5 RCTs that reported values of osteocalcin, 3 RCTs that reported values of NTx, 6 RCTs that reported values of total body BMD, and 4 RCTs that reported total body BMC. However, the effects of milk on either index appear inconsistent. This discrepancy may be explained by the different intervention forms of milk, various hormone states of the subjects, and the influence of other foods eaten during the study. Moreover, limited sample sizes and intervention duration often prevent the detection of significant effects in individual studies.23 Optimal management of osteoporosis consists of maximizing peak bone mass in early adulthood and preventing rapid bone loss
at menopause.24 Bone formation and bone resorption are continuous processes that maintain the integrity of bone tissue. In our meta-analysis, we found that milk supplementation significantly decreased bone metabolism. In a randomized controlled trial between MBP group and control group, they reported that there was no significant difference between the groups in serum osteocalcin concentrations, a bone formation marker. However, urinary NTx, a bone resorption marker, were lower after 6 months of MBP supplementation than at baseline, indicating that MBP supplementation led to a reduction in the rate of bone resorption.4 A crossover study comparing the effects of semi-skimmed milk and Coca Cola on bone turnover, showed that during the milk invention period, significant reductions on urinary NTx and another bone resorption marker, serum corss-linked C-telopeptides (CTx), were observed.16 Moreover, we confirmed the healthful effect of milk intake on bone-specific parameters such as total body BMD and BMC. Our results showed that milk intake led to an increase trend in total body BMD and a significant increase in total body BMC. Most of the studies excluded from our meta-analysis support our results. Ma et al.’s study revealed that calcium fortified milk (137 ml milk per day with 233 mg calcium) could significantly increase total body BMD/BMC, distal radius and forearm BMD/BMC in school girls.25 Another study by Hu et al. indicated that 500 ml milk per day could prevent bone loss in lactation women.26 In the present study, the control groups in 4 studies did not use normal diet or habitual diet as control intervention. After excluding the studies that did not
D.F. Ma et al. / e-SPEN Journal 8 (2013) e1ee7
e5
Table 2 Subgroup analysis of the effects of milk on osteocalcin. Subgroup outcome Form of intervention Milk Calcium fortified milk Treatment length (months) <6 6 Race Asian Western Intervention subjects Ⅰ Female Male Intervention subjects Ⅱ Children Adults a b
No of comparisons
No of subjects
Treatment effect on osteocalcin (ng/ml)a
P values
Studies included (reference number)
5 10
292 662
7.90 [14.05, 1.75]b 5.81 [7.17, 4.43]b
0.01 <0.00001
6, 10, 15 17, 21
8 7
370 584
6.18 [7.54, 4.82]b 0.52 [6.02,7.05]
<0.00001 0.88
10, 16, 17 6, 21
4 11
338 616
2.09 [11.57, 7.40] 5.98 [7.32, 4.63]b
0.67 <0.00001
21 6, 10,16, 17
13 2
908 46
5.61 [6.96, 4.25]b 17.28 [25.69, 8.87]b
<0.00001 <0.0001
6, 17, 21 10, 16
8 7
608 346
4.67 [10.28, 0.93] 5.97 [7.35, 4.60]b
0.1 <0.00001
6, 10,21 16, 17
95% CI in square bracket; statistically significant.
use normal diet/habitual diet as control intervention, the pooled weighted mean difference for osteocalcin, NTx, total body BMD and BMC changed little. The mechanism of the NTx-lowering and total body BMCincreasing effects of milk is not well understood. One possible explanation could be related to milk-derived peptides. Milk proteins consist of casein and whey fractions, and the latter has been recognized as a potential bone-beneficial factor. Takada et al. reported that whey can stimulate proliferation and differentiation of murine osteoblastic MC3T3-E1 cells, suppress osteoclastmediated bone resorption and decrease formation of osteoclasts in vitro studies.27 In our meta-analysis, the bone metabolismdecreasing and bone mass-increasing effects are small, which is unlikely to translate into lower fracture risk over a lifetime. A previous meta-analysis reported that milk intake didn’t decrease
hip fracture risk.28 Moreover, most of the studies in our metaanalysis used around a third of liter of milk (330 ml) per day, therefore, the results may not be generalizable to lower intakes less than 330 ml per day. It is nowadays a general viewpoint that calcium supplementation has been a useful strategy to prevent osteoporosis. However, in our meta-analysis, “normal milk” without calcium fortifier presented significant effects on bone metabolism and BMD/BMC. This suggests that normal milk without calcium fortifier may be adequate to present beneficial effects on reducing bone resorption. A study by Budek indicated that a high milk intake (1.5 L/day) decreased bone turnover in prepubertal boys,10 and another study by Volek et al. showed that 236 ml of fluid milk per day significantly increase total body BMC/BMD.19 When the intervention duration was shortened to less than 6 months, the effect of milk was still
Fig. 3. Weighted mean difference of urinary NTx concentration after treatment with milk. Lines correspond to 95% CI. *: the value of the mean and 95% CI of each study, #: reference number.
e6
D.F. Ma et al. / e-SPEN Journal 8 (2013) e1ee7
Fig. 4. Weighted mean difference of total body BMC and total body BMD after treatment with milk. Lines correspond to 95% CI. *: the value of the mean and 95% CI of each study, # : reference number.
significant. Some studies also indicated that treatment lasting less than 6 months may be adequate to produce the effects of milk on indices of bone metabolism.10 Interestingly, we found that milk intake had significant effects on osteocalcin in Caucasian but not Asian subjects after sub-group analysis. In the present analysis, there are only two studies related to osteocalcin conducted on Asian women including post-menopausal women. Therefore, in the present analysis, we could not judge whether the effects become obvious when milk are consumed on Asian women because of the large range of confidence interval. In the current analysis, we found that milk intake significantly decreased osteocalcin concentration in male but not female. A study by Kristensen et al. showed that 10day of high milk intake (2.5 L/day) could significantly decrease bone turnover (osteocalin, phosphate, 1,25-dihydroxycholecalciferol,
NTx and CTx, etc) in young men.16 This suggests that young men may still have the opportunity to increase their peak bone mass by modifying their lifestyle and nutrition. Funnel plot did not provide strong evidence of publication bias. Generally, studies with significant results were easier to publish than those with nonsignificant results.23 After 40 of null results appearing in the analysis were added randomly, the pooled weighted mean difference changed little. This suggested that unpublished studies like published studies with nonsignificant results do not seem to influence this combined effect estimate over a large range. Although the relatively low number of studies limited the power of our meta-analysis, the results suggested that milk is contributed moderately to the decrease of bone metabolism, and has a small influence on increasing of total body BMC and BMD. Intervention
D.F. Ma et al. / e-SPEN Journal 8 (2013) e1ee7
Fig. 5. Heterogeneity of difference in osteocalcin after milk treatment. Dashed showed 95% CI line. Funnel plot showing no significant publication bias.
duration for 6 months may be enough for milk to produce favorable effects on bone turnover. Future RCTs should include a larger sample size and an examination of the long-term effects of milk supplementation on bone mass and fracture risk. Conflicts of interest statement We declare that we have no financial and personal relationships with other people or organizations that can inappropriately influence our work. Acknowledgments This research was supported in part by a grant from the Beijing Natural Science Foundation (7122102) and a new teacher fund of Ministry of Education (20110001120027) to Dr Defu Ma. There is no conflict of interest that could inappropriately influence (bias) our work. De Fu Ma, Pei Yu Wang, and Yu Mei Zhang participated in the design of this manuscript. De Fu Ma, Wei Zheng, and Ming Ding participated in abstracting the data and performing statistical analysis. All authors read and approved the final manuscript. References 1. Karkkainen M, Tuppurainen M, Salovaara K, Sandini L, Rikkonen T, Sirola J, et al. Effect of calcium and vitamin D supplementation on bone mineral density in women aged 65e71 years: a 3-year randomized population-based trial (OSTPRE-FPS). Osteoporos Int 2010;21(12):2047e55. 2. Feskanich D, Willett WC, Stampfer MJ, Colditz GA. Protein consumption and bone fractures in women. Am J Epidemiol 1996;143(5):472e9. 3. Darling AL, Millward DJ, Torgerson DJ, Hewitt CE, Lanham-New SA. Dietary protein and bone health: a systematic review and meta-analysis. Am J Clin Nutr 2009;90(6):1674e92. 4. Aoe S, Koyama T, Toba Y, Itabashi A, Takada Y. A controlled trial of the effect of milk basic protein (MBP) supplementation on bone metabolism in healthy menopausal women. Osteoporos Int 2005;16(12):2123e8. 5. Kalkwarf HJ, Khoury JC, Lanphear BP. Milk intake during childhood and adolescence, adult bone density, and osteoporotic fractures in US women. Am J Clin Nutr 2003;77(1):257e65.
e7
6. Cadogan J, Eastell R, Jones N, Barker ME. Milk intake and bone mineral acquisition in adolescent girls: randomised, controlled intervention trial. BMJ 1997;315(7118):1255e60. 7. Liu Z, Qiu L, Chen YM, Su YX. Effect of milk and calcium supplementation on bone density and bone turnover in pregnant Chinese women: a randomized controlled trail. Arch Gynecol Obstet 2011;283(2):205e11. 8. Aoe S, Toba Y, Yamamura J, Kawakami H, Yahiro M, Kumegawa M, et al. Controlled trial of the effects of milk basic protein (MBP) supplementation on bone metabolism in healthy adult women. Biosci Biotechnol Biochem 2001;65(4):913e8. 9. Brockwell SE, Gordon IR. A comparison of statistical methods for meta-analysis. Stat Med 2001;20(6):825e40. 10. Budek AZ, Hoppe C, Michaelsen KF, Molgaard C. High intake of milk, but not meat, decreases bone turnover in prepubertal boys after 7 days. Eur J Clin Nutr 2007;61(8):957e62. 11. Green JH, Booth C, Bunning R. Impact of supplementary high calcium milk with additional magnesium on parathyroid hormone and biochemical markers of bone turnover in postmenopausal women. Asia Pac J Clin Nutr 2002;11(4): 268e73. 12. Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJ, Gavaghan DJ, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996;17(1):1e12. 13. Higgins J, Green S. Cochrane handbook for systematic reviews of interventions version 5.0.0 [updated February 2008]. The Cochrane Collaboration Q3. Available from:, www.cochrane- handbook.org; 2008. 14. Review Manager (RevMan). Computer program. Version 4.2 for Windows. Oxford, England: The Cochrane Collaboration; 2003. 15. Du X, Zhu K, Trube A, Zhang Q, Ma G, Hu X, et al. School-milk intervention trial enhances growth and bone mineral accretion in Chinese girls aged 10e12 years in Beijing. Br J Nutr 2004;92(1):159e68. 16. Kristensen M, Jensen M, Kudsk J, Henriksen M, Molgaard C. Short-term effects on bone turnover of replacing milk with cola beverages: a 10-day interventional study in young men. Osteoporos Int 2005;16(12):1803e8. 17. Kruger MC, Booth CL, Coad J, Schollum LM, Kuhn-Sherlock B, Shearer MJ. Effect of calcium fortified milk supplementation with or without vitamin K on biochemical markers of bone turnover in premenopausal women. Nutrition 2006;22(11-12):1120e8. 18. Tong R, Zhang X. The effect of fortified milk on growth and bone development in girls aged 10e12 y. In: The 6th national academic conference of the child and adolescent health section of CPMA & the 3rd academic conference of the school section of Chinese health education association. Nanjing, China: 2004. 19. Volek JS, Gomez AL, Scheett TP, Sharman MJ, French DN, Rubin MR, et al. Increasing fluid milk favorably affects bone mineral density responses to resistance training in adolescent boys. J Am Diet Assoc 2003;103(10):1353e6. 20. Zhang L, Tong R, Zhang X. The effect of fortified milk on growth and development in adolescent girls. Chin J School Health 2006;27(7):611e2. 21. Zhu K, Du X, Cowell CT, Greenfield H, Blades B, Dobbins TA, et al. Effects of school milk intervention on cortical bone accretion and indicators relevant to bone metabolism in Chinese girls aged 10e12 y in Beijing. Am J Clin Nutr 2005;81(5):1168e75. 22. Liu JC, Zhang J, Ouyang QH, Liu Q. Effect of short-term milk drinking on bone mineral density in healthy girls during spurt growth period. Chinese J Clin Rehabil 2002;6(11):1618e9. 23. Ma DF, Qin LQ, Wang PY, Katoh R. Soy isoflavone intake inhibits bone resorption and stimulates bone formation in menopausal women: metaanalysis of randomized controlled trials. Eur J Clin Nutr 2008;62(2):155e61. 24. Kulak CA, Bilezikian JP. Osteoporosis: preventive strategies. Int J Fertil Womens Med 1998;43(2):56e64. 25. Ma GS, Zhang CY, Zhang Q, Zhang L, Zhu K, Hu XQ. Bone mass of middle school girls two years after cessation of calcium-vitamin D-fortified milk supplementation: follow-up study. Acta Nutrimenta Sinica 2006;28(2):139e42. 26. Hu Y, Feng YP, Wang RY, Gao XF, Wang RL. The effect of calcium-fortified milk on bone mineral density in pregnant or suckling period woman. Chinese J Coal Industry Med 2002;5(9):898e9. 27. Takada Y, Aoe S, Kumegawa M. Whey protein stimulated the proliferation and differentiation of osteoblastic MC3T3-E1 cells. Biochem Biophys Res Commun 1996;223(2):445e9. 28. Bischoff-Ferrari HA, Dawson-Hughes B, Baron JA, Kanis JA, Orav EJ, Staehelin HB, et al. Milk intake and risk of hip fracture in men and women: a meta-analysis of prospective cohort studies. J Bone Miner Res 2010.
Contents lists available at SciVerse ScienceDirect
e-SPEN Journal journal homepage: http://www.elsevier.com/locate/clnu
Original article
Milk intake increases bone mineral content through inhibiting bone resorption: Meta-analysis of randomized controlled trials De Fu Ma a, *, c, Wei Zheng b, c, Ming Ding b, Yu Mei Zhang b, Pei Yu Wang a a b
Department of Social Medicine & Health Education, School of Public Health, Peking University, Xueyuan Road 38, Haidian District, Beijing 100191, PR China Department of Nutrition & Food Hygiene, School of Public Health, Peking University, Beijing 100191, PR China
a r t i c l e i n f o
s u m m a r y
Article history: Received 16 December 2010 Accepted 16 October 2012
Background and aims: To clarify the effects of milk intake on bone mineral density (BMD) ，bone mineral content (BMC) and bone metabolism markers. Methods: We identified randomized controlled trials related to urinary N-telopeptide cross-links of type I collagen (NTx), serum osteocalcin, BMD and BMC listed on MEDLINE (January 1966eNovember 2010), Science Citation Index and PUBMED (updated till November 2010), China National Knowledge Infrastructure (1979eNovember 2010) etc. Results: Eleven studies with a total of 2397 subjects were selected for meta-analysis. The osteocalcin in subjects who consumed milk decreased by 5.9 (95% confidence interval (CI) 7.23, 4.57) ng/ml in comparison to that in control treatment. Milk intake vs control treatment significantly decreased urine NTx by 5.41 (95% CI 10.35, 0.47) nmol/mmol. Moreover, the total body BMC in subjects who consumed milk increased significantly by 40.32 (95% CI 17.58, 63.05) g in comparison to that in control treatment. Milk intake vs control treatment increased total body BMD by 0.01 (95% CI -0.02, 0.03) g/cm2 with borderline significance. Conclusions: Milk intervention significantly attenuates bone loss through inhibiting bone metabolism. Crown Copyright Ó 2012 Published by Elsevier Ltd on behalf of European Society for Clinical Nutrition and Metabolism. All rights reserved.
Keywords: Milk Osteoporosis Urinary N-telopeptide cross-links of type I collagen Bone mineral content
Introduction Osteoporosis is an increasing public health problem worldwide. The fractures caused by the bone density reduction and bone microstructure alteration result in a lower life quality. Calcium supplementation, protein supplementation, exercise, and hormone replacement therapy and so on have long been demonstrated to be the main strategies to prevent osteoporosis. Clinical trials have shown that calcium supplementation combined with Vitamin D can increase bone mineral density (BMD) and prevent bone loss in elderly women.1 However, the association between dietary protein and osteoporosis are controversial. A prospective study indicated that higher consumption of animal protein had an increased risk of forearm fracture.2 While a systematic review on 31 cross-sectional
Abbreviations: RCTs, randomized controlled trials; BMD, bone mineral density; BMC, bone mineral content; NTx, N-telopeptide cross links of type I collagen; MBP, milk basic protein. * Corresponding author. Tel.: þ86 10 82801743; fax: þ86 10 82802002. E-mail addresses: [email protected] (D.F. Ma), [email protected] (W. Zheng). c Contributed equally to the work; Wei Zheng is the Co-first author.
surveys showed a small positive effect of protein intake on lumbar spine BMD.3 Historically, milk has been widely consumed because of its excellent nutritional value.4 Milk that contains several factors such as bioactive peptides, calcium, growth factors related to bone metabolism might affect both bone formation and bone resorption.4 In particular, milk is a good source of bioavailable calcium compared with other food sources. During the latest several decades, RCTs on bone health suggested that milk intervention might increase BMD and bone mineral content (BMC). In the large, third National Health and Nutrition Examination Survey (NHANES III), they found that among women aged 20e49 year, BMC was 5.6% lower in those who consumed <1 serving of milk/d (low intake) than that in those who consumed >1 serving/d (high milk) during childhood (p < 0.01).5 Cadogan et al. reported that the intervention group consuming, on average, 300 ml milk a day throughout the intervention trial had greater increases of BMD (p ¼ 0.017) and BMC (p ¼ 0.009) compared with the control group.6 Liu et al. found that 45 g milk powder supplementation could increase serum osteocalcin (a biochemical marker for bone formation, p < 0.05) and decrease urinary hydroxyproline (a biochemical marker for bone
2212-8263/$36.00 Crown Copyright Ó 2012 Published by Elsevier Ltd on behalf of European Society for Clinical Nutrition and Metabolism. All rights reserved. http://dx.doi.org/10.1016/j.clnme.2012.10.005
e2
D.F. Ma et al. / e-SPEN Journal 8 (2013) e1ee7
resorption, p < 0.05) in Chinese women.7 Aoe et al. found that 40 mg of milk basic protein (MBP) supplementation was able to significantly suppressed the urinary excretion of cross-linked Nteleopeptides of type-I collagen (NTx, a biochemical marker for bone resorption) in healthy adult women.8 However, the effects of milk intake on BMD, BMC, and bone metabolism appear inconsistent in RCTs. Moreover, it’s difficult to clarify whether the beneficial effects are from milk or from calcium because that calcium fortified milk was widely used in the studies on bone health and milk. Thus, a statistical method of combining these diverse data is needed to evaluate the usefulness of milk therapy. Meta-analysis combines or integrates the results of several studies to provide an increased statistical power for the quantitative identification of trends.9 To clarify the effects of milk intake on bone health, we identified all RCTs related to the effects of milk on bone mass and bone turnover markers and analyzed the effects of milk or calcium fortified milk on bone metabolism quantitatively. Materials and methods MEDLINE (January 1966eNovember 2011), the Cochrane Controlled Trials Register, EMBASE (1985eNovember 2011), Science Citation Index and PUBMED (updated till November 2011), China National Knowledge Infrastructure (1979eNovember 2011), VIP Database for Chinese Technical Periodicals (1989eNovember 2011), and Wanfang database (1982eNovember 2011) were used to search articles (in English and Chinese) that described RCTs investigating the effect of milk on bone metabolism. In the RCTs, BMD and BMC were generally measured to assess the bone mass, and serum osteocalcin was generally used as a bone formation marker, urine NTx was generally used as bone resorption marker.6,10,11 Hence, titles, abstracts, and subject headings in the databases were searched with the use of the following Boolean phrases: (“bone” or “osteoporosis” or “bone mass” or “BMD” or “BMC” or “osteocalcin” or “NTX”) and (“milk” or “fortified milk”). We carried out a broad search for all studies with the Boolean phrases “diet” and (“osteoporosis” or “bone metabolism”). We also examined all references of related reviews and papers identified by the search. Additionally, we tried to contact the authors for the obtaining of unpublished data. Studies were selected for analysis if they met all of the following criteria: 1) subjects ingested milk products for at least 1 week; 2) the RCTs included a parallel control group; 3) Total body BMD, total body BMC, NTx or osteocalcin was used as an index of bone turnover. Studies were excluded if they are lack of indices of interest, lack of a control group, insufficient original data or baseline values. If the study sample was found to overlap with that in another article or if two articles described aspects of the same study, only the publication with the largest sample was used. If the study reported some comparisons, we included all comparisons in the meta-analysis. Two researchers (De Fu Ma and Wei Zheng) extracted data independently. A data collection form was designed, and data were entered into the form twice to reduce input errors. The items entered in the form included participant characteristics, treatment duration, interventional design, and values of relevant indices (Total body BMD, total body BMC, NTx and osteocalcin) before and after milk or control treatments. Jadad Scores were used to measure the quality of the RCTs.12 A numerical score between 0 and 5 was assigned as a rough measure of study design and reporting quality, 0 being the weakest and 5 the strongest. One point was assigned if the trial was either randomized or double-blind or in the case of an accurate description of the drop-out patients. Moreover, further points were given if randomization and blinding procedures were appropriate, whereas, instead, points were subtracted in the case of inappropriate descriptions of the same procedures. An overall score more
than 3 indicated a good quality study. Two researchers rated study quality independently. There was 90% agreement on Jadad Scores. If the researchers disagreed, a final score was reached by discussion. In this meta-analysis, we obtained the mean differences from the post-randomization baseline to after-treatment values for each trial and calculated the pooled standard deviation of the mean differences according to the standard method of Cochrane handbook.13 Weighted mean difference was calculated by subtracting the mean difference of the control group from that of the treatment group. The inverse variance method was used to pool the weighted mean difference with STATA software (version 9.2; Stata Corp., College Station, TX, USA).14 To assess the heterogeneity (apparent diversity in weighted mean differences across studies), we conducted a test based on c2 distribution (p < 0.05 is considered significant). Random-effects model was used as the method of combination for all the analyses showing significant heterogeneity .The funnel plot was performed to detect publication bias. In addition, we performed subgroup analyses for osteocalcin by 4 variables one at a time: form of intervention, treatment length, race, and intervention subjects to identify impact factors that can influence the effects of milk intake according to the characters of the data. Results The trial flow chart was illustrated in Fig. 1. Our literature search identified 39 RCTs including 2 of them obtained from the reference lists. We also tried to contact the authors for unpublished data, but unfortunately none was obtained. 28 studies were excluded because of lack of indices of interest, lack of a control group, insufficient original data or baseline values. Thus, 11 studies (8 in English, 3 in Chinese) with a total of 2397 subjects were included in this meta-analysis.4,6,10,11,15e22 The characteristics of the trials included were shown in Table 1. One study had quality score of four, 7 studies had quality score of three, and 3 studies had quality score Articles retrieved from the original search (n = 3032) Articles excluded after the preliminary screening by looking through the titles and abstracts (n = 2533) Potentially relevant articles identified and screened for retrieval (n = 499) Articles excluded because not RCTs (n = 460) RCTs retrieved for more detailed evaluation (n = 39) RCTs excluded: did not meet the inclusion criteria (n = 28) RCTs included in meta-analysis (n = 11) RCTs withdrawn because did not report: Serum osteocalcin (n = 6) Urinary N-telopeptide cross-links of type I collagen (n = 8) Total body bone mineral density (n =5) Total body bone content (n =7)
RCTs with usable information on: Serum osteocalcin (n = 5) Urinary N-telopeptide cross-links of type I collagen (n = 3) Total body bone mineral density (n =6) Total body bone content (n =4) Fig. 1. Results of search for eligible studies.
D.F. Ma et al. / e-SPEN Journal 8 (2013) e1ee7
e3
Table 1 Characteristics of the 11 selected randomized controlled trials of milk intake and bone turnover. Study
Length of treatment (months)
No
Subjects (years)
Race
Intervention
Indices
Quality score
Budek et al. (2007) Cadogan et al. (1997)
1/4 18
24 82
Male (8) Female (12.2)
Western Western
M: M 1.5 L/day. Con: meat M: M 0.3 L/day. Con: habitual diet
2a,b 3a
Du et al. (2004)
24
757
Female (10)
Asian
Green et al. (2002)
1
50
Western
NTx
3a
Kristensen et al. (2005) Kruger et al. (2006)
1/3 4
22 82
Female (Post, 68) Male (24) Female (27.5)
Osteocalcin NTx Osteocalcin
3a 3a
Liu et al. (2002)
24
375
Female (10)
Asian
M þ Ca: M 330 ml þ Ca 560 mg/day M þ Ca þ VitD: M 330 ml þ Ca 560 mg þ VitD 5 or 8 ug/day. Con: habitual diet M þ Ca: M 50 mg þ Ca 1200 mg/day. Con: apple drink M: 2.5 l/day. Con: coca cola M þ Ca: M 50 g þ Ca 1000 mg/day M þ Ca þ VitK: M 50 g þ Ca 1000 mg þ VitK 80 ug/day. Con: normal diet M: 330 ml/day. Con: normal diet
Osteocalcin Osteocalcin NTx. Total body BMD Total body BMC. Total body BMD
24
351
Female (11)
Asian
BMD. BMC BMD. BMC
2a,b
Tong et al. (2004)
Total Total Total Total
Volek et al. (2003)
3
28
Male (14.4)
Western
2a,b
Zhang et at. (2006)
60
498
Female (10)
Asian
Total body BMC. Total body BMD Total body BMD
Zhu et al. (2005)
24
128
Female (10)
Asian
Osteocalcin
3a
Western Western
M þ Ca: M 330 ml þ Ca 560 mg/day M þ Ca þ VitD: M 330 ml þ Ca 560 mg þ VitD 8 ug/day Con: habitual diet M: M 708 ml/day. Con: Apple juice/Grape juice M þ Ca: M 330 ml þ Ca 560 mg/day M þ Ca þ VitD: M 330 ml þ Ca 560 mg þ VitD 8 ug/day. Con: normal diet M þ Ca: M 330 ml þ Ca 560 mg/day M þ Ca þ VitD: M 330 ml þ Ca 560 mg þ VitD 5e8 ug/day. Con: habitual diet
body body body body
4b
3a
3a,b
Post, post-menopause; M, milk; Ca, calcium; VitD, vitamin D; VitK, vitamin K1; Con, control group; NTx, N-telopeptide cross-links of type I collagen; BMD, bone mineral density; BMC, bone mineral content. Highest total score is 5. a points were deleted from Quality Score because the method of blinding was either not described or not appropriate; b points were deleted from Quality Score because there was no description of random.
of two. In 6 of these studies, calcium fortified milk was used, and normal milk (whole, semi-skimmed, and skimmed milk) was used in other studies. The duration of treatment varied widely, ranging from one week to 60 months including 6 studies with a duration exceeding 6 months. 6 studies were performed in Caucasian subjects and 5 studies in Asian subjects. 3 of these 11 studies were carried out in male and other studies were carried out in female including 1 studies performed in post-menopausal women. 8 of these 11 studies were carried out in children and other studies were carried out in adults. In all RCTs, subjects were healthy and were not undergoing any other therapy for osteoporosis; additional calcium supplements were restricted during treatment. All studies reported no significant differences regarding baseline characteristics such as age, body mass index, osteocalcin, NTx, BMD and BMC between groups. There were also no significant weight changes or negative side effects reported. In the present analysis, 5 studies with 15 comparisons reported values of osteocalcin before and after milk or control treatments. When we combined the 15 comparisons, milk decreased osteocalcin by 5.90 (95% confidence interval (CI) 7.23, 4.57; p < 0.00001) ng/ml (Fig. 2). Further analysis of the effects of milk on osteocalcin is shown in Table 2.Milk intake significantly decreased osteocalcin when the studies that used calcium fortified milk as treatment were excluded (7.90, 95% CI [14.05, 1.75] ng/ml; p ¼ 0.01). The decrease in osteocalcin was still significant when the studies with milk intake of more than 6 months were not included in the analysis (6.18, 95% CI [7.54, 4.82] ng/ml; p < 0.00001). Interestingly, we found that milk intake had significant effects on osteocalcin in Caucasian (5.98, 95% CI [7.32, 4.63] ng/ml; p < 0.00001) but not Asian people (2.09, 95% CI [11.57, 7.40] ng/ml; p ¼ 0.67) after sub-group analysis. Because the hormone levels between female and male are quite different, we performed sub-group analysis according to the intervention subjects. Milk presented a larger mean difference on decreasing osteocalcin in
male (17.28, 95% CI [25.69, 8.87] ng/ml; p < 0.0001) although the number of intervention subjects is small. In addition, we found that milk has a significant effect on osteocalcin in adults (5.97, 95% CI [7.35, 4.60] ng/ml; p < 0.0001) but not in children (4.67, 95% CI [10.28,0.93] ng/ml; p ¼ 0.1) by performing sub-group analysis by children vs adults. In the present analysis, 3 studies with 8 comparisons reported values of NTx before and after milk or control treatments. When the 8 comparisons were combined, the mean difference was e5.41 (95% CI 10.35, 0.47; p ¼ 0.03) nmol/mmol (Fig. 3). After removing the studies on calcium fortified milk, we found that the NTx-decreasing effect was still significant (18.91, 95% CI [33.73, 4.09] nmol/mmol; p ¼ 0.01). Because of the low number of studies, we did not perform subgroup analysis for NTx. Moreover, we performed meta-analysis for total body BMC and total body BMD. In the present analysis, 4 studies with 8 comparisons reported values of total body BMC before and after milk or control treatments. When we combined the 8 comparisons, milk significantly increased total body BMC by 40.32 (95% CI 17.58, 63.05; P ¼ 0.0005) g (Fig. 4). The effect was still significant (41.99, 95% CI [6.94, 77.03] nmol/mmol; p ¼ 0.02) when the studies on calcium fortified milk were excluded. There were two studies by Du et al. and Zhang et al. with 1255 subjects (about half of the total subjects) related to BMC and/or BMD.15,20 In order to avoid bias from the two studies, we performed analysis except for the two studies and the effect was similar (41.77, 95% CI [14.82, 68.73] nmol/ mmol; p ¼ 0.002) on BMC. When the 6 studies with 13 comparisons reported values of total body BMD were combined, the mean difference was 0.01 (95% CI 0.02, 0.03; P ¼ 0.63) g/cm2 in total body BMD (in random-effects model), showing an increasing trend (Fig. 4). The BMD-increasing effect after excluding the studies on calcium fortified milk or the studies with large number of subjects was 0.01 (95% CI [0.00, 0.02] g/cm2; p ¼ 0.02) and 0.01, (95% CI [0.00, 0.02] g/cm2; p < 0.0001), respectively. Because of the low
e4
D.F. Ma et al. / e-SPEN Journal 8 (2013) e1ee7
Fig. 2. Weighted mean difference of serum osteocalcin concentration after treatment with milk. Lines correspond to 95% CI. *: the value of the mean and 95% CI of each study, # : reference number.
number of studies, we did not perform subgroup analysis for total body BMC and total body BMD. According to the results from c2 test, only one analysis on total body BMD which was based on all 6 related studies showed significant heterogeneity (P < 0.0001). So this analysis was performed using random-effects model and all other analyses were performed using fixed-effects model. The potential for publication bias was examined by construction of a “funnel plot” of the relation between the standard error and the weighted mean difference (Fig. 5). It did not provide strong evidence of publication bias for either index. Discussion To our knowledge, this is the first meta-analysis that clarified the effects of milk intake on total body BMD, BMC, osteocalcin, and NTx. In the present analysis, there are 5 RCTs that reported values of osteocalcin, 3 RCTs that reported values of NTx, 6 RCTs that reported values of total body BMD, and 4 RCTs that reported total body BMC. However, the effects of milk on either index appear inconsistent. This discrepancy may be explained by the different intervention forms of milk, various hormone states of the subjects, and the influence of other foods eaten during the study. Moreover, limited sample sizes and intervention duration often prevent the detection of significant effects in individual studies.23 Optimal management of osteoporosis consists of maximizing peak bone mass in early adulthood and preventing rapid bone loss
at menopause.24 Bone formation and bone resorption are continuous processes that maintain the integrity of bone tissue. In our meta-analysis, we found that milk supplementation significantly decreased bone metabolism. In a randomized controlled trial between MBP group and control group, they reported that there was no significant difference between the groups in serum osteocalcin concentrations, a bone formation marker. However, urinary NTx, a bone resorption marker, were lower after 6 months of MBP supplementation than at baseline, indicating that MBP supplementation led to a reduction in the rate of bone resorption.4 A crossover study comparing the effects of semi-skimmed milk and Coca Cola on bone turnover, showed that during the milk invention period, significant reductions on urinary NTx and another bone resorption marker, serum corss-linked C-telopeptides (CTx), were observed.16 Moreover, we confirmed the healthful effect of milk intake on bone-specific parameters such as total body BMD and BMC. Our results showed that milk intake led to an increase trend in total body BMD and a significant increase in total body BMC. Most of the studies excluded from our meta-analysis support our results. Ma et al.’s study revealed that calcium fortified milk (137 ml milk per day with 233 mg calcium) could significantly increase total body BMD/BMC, distal radius and forearm BMD/BMC in school girls.25 Another study by Hu et al. indicated that 500 ml milk per day could prevent bone loss in lactation women.26 In the present study, the control groups in 4 studies did not use normal diet or habitual diet as control intervention. After excluding the studies that did not
D.F. Ma et al. / e-SPEN Journal 8 (2013) e1ee7
e5
Table 2 Subgroup analysis of the effects of milk on osteocalcin. Subgroup outcome Form of intervention Milk Calcium fortified milk Treatment length (months) <6 6 Race Asian Western Intervention subjects Ⅰ Female Male Intervention subjects Ⅱ Children Adults a b
No of comparisons
No of subjects
Treatment effect on osteocalcin (ng/ml)a
P values
Studies included (reference number)
5 10
292 662
7.90 [14.05, 1.75]b 5.81 [7.17, 4.43]b
0.01 <0.00001
6, 10, 15 17, 21
8 7
370 584
6.18 [7.54, 4.82]b 0.52 [6.02,7.05]
<0.00001 0.88
10, 16, 17 6, 21
4 11
338 616
2.09 [11.57, 7.40] 5.98 [7.32, 4.63]b
0.67 <0.00001
21 6, 10,16, 17
13 2
908 46
5.61 [6.96, 4.25]b 17.28 [25.69, 8.87]b
<0.00001 <0.0001
6, 17, 21 10, 16
8 7
608 346
4.67 [10.28, 0.93] 5.97 [7.35, 4.60]b
0.1 <0.00001
6, 10,21 16, 17
95% CI in square bracket; statistically significant.
use normal diet/habitual diet as control intervention, the pooled weighted mean difference for osteocalcin, NTx, total body BMD and BMC changed little. The mechanism of the NTx-lowering and total body BMCincreasing effects of milk is not well understood. One possible explanation could be related to milk-derived peptides. Milk proteins consist of casein and whey fractions, and the latter has been recognized as a potential bone-beneficial factor. Takada et al. reported that whey can stimulate proliferation and differentiation of murine osteoblastic MC3T3-E1 cells, suppress osteoclastmediated bone resorption and decrease formation of osteoclasts in vitro studies.27 In our meta-analysis, the bone metabolismdecreasing and bone mass-increasing effects are small, which is unlikely to translate into lower fracture risk over a lifetime. A previous meta-analysis reported that milk intake didn’t decrease
hip fracture risk.28 Moreover, most of the studies in our metaanalysis used around a third of liter of milk (330 ml) per day, therefore, the results may not be generalizable to lower intakes less than 330 ml per day. It is nowadays a general viewpoint that calcium supplementation has been a useful strategy to prevent osteoporosis. However, in our meta-analysis, “normal milk” without calcium fortifier presented significant effects on bone metabolism and BMD/BMC. This suggests that normal milk without calcium fortifier may be adequate to present beneficial effects on reducing bone resorption. A study by Budek indicated that a high milk intake (1.5 L/day) decreased bone turnover in prepubertal boys,10 and another study by Volek et al. showed that 236 ml of fluid milk per day significantly increase total body BMC/BMD.19 When the intervention duration was shortened to less than 6 months, the effect of milk was still
Fig. 3. Weighted mean difference of urinary NTx concentration after treatment with milk. Lines correspond to 95% CI. *: the value of the mean and 95% CI of each study, #: reference number.
e6
D.F. Ma et al. / e-SPEN Journal 8 (2013) e1ee7
Fig. 4. Weighted mean difference of total body BMC and total body BMD after treatment with milk. Lines correspond to 95% CI. *: the value of the mean and 95% CI of each study, # : reference number.
significant. Some studies also indicated that treatment lasting less than 6 months may be adequate to produce the effects of milk on indices of bone metabolism.10 Interestingly, we found that milk intake had significant effects on osteocalcin in Caucasian but not Asian subjects after sub-group analysis. In the present analysis, there are only two studies related to osteocalcin conducted on Asian women including post-menopausal women. Therefore, in the present analysis, we could not judge whether the effects become obvious when milk are consumed on Asian women because of the large range of confidence interval. In the current analysis, we found that milk intake significantly decreased osteocalcin concentration in male but not female. A study by Kristensen et al. showed that 10day of high milk intake (2.5 L/day) could significantly decrease bone turnover (osteocalin, phosphate, 1,25-dihydroxycholecalciferol,
NTx and CTx, etc) in young men.16 This suggests that young men may still have the opportunity to increase their peak bone mass by modifying their lifestyle and nutrition. Funnel plot did not provide strong evidence of publication bias. Generally, studies with significant results were easier to publish than those with nonsignificant results.23 After 40 of null results appearing in the analysis were added randomly, the pooled weighted mean difference changed little. This suggested that unpublished studies like published studies with nonsignificant results do not seem to influence this combined effect estimate over a large range. Although the relatively low number of studies limited the power of our meta-analysis, the results suggested that milk is contributed moderately to the decrease of bone metabolism, and has a small influence on increasing of total body BMC and BMD. Intervention
D.F. Ma et al. / e-SPEN Journal 8 (2013) e1ee7
Fig. 5. Heterogeneity of difference in osteocalcin after milk treatment. Dashed showed 95% CI line. Funnel plot showing no significant publication bias.
duration for 6 months may be enough for milk to produce favorable effects on bone turnover. Future RCTs should include a larger sample size and an examination of the long-term effects of milk supplementation on bone mass and fracture risk. Conflicts of interest statement We declare that we have no financial and personal relationships with other people or organizations that can inappropriately influence our work. Acknowledgments This research was supported in part by a grant from the Beijing Natural Science Foundation (7122102) and a new teacher fund of Ministry of Education (20110001120027) to Dr Defu Ma. There is no conflict of interest that could inappropriately influence (bias) our work. De Fu Ma, Pei Yu Wang, and Yu Mei Zhang participated in the design of this manuscript. De Fu Ma, Wei Zheng, and Ming Ding participated in abstracting the data and performing statistical analysis. All authors read and approved the final manuscript. References 1. Karkkainen M, Tuppurainen M, Salovaara K, Sandini L, Rikkonen T, Sirola J, et al. Effect of calcium and vitamin D supplementation on bone mineral density in women aged 65e71 years: a 3-year randomized population-based trial (OSTPRE-FPS). Osteoporos Int 2010;21(12):2047e55. 2. Feskanich D, Willett WC, Stampfer MJ, Colditz GA. Protein consumption and bone fractures in women. Am J Epidemiol 1996;143(5):472e9. 3. Darling AL, Millward DJ, Torgerson DJ, Hewitt CE, Lanham-New SA. Dietary protein and bone health: a systematic review and meta-analysis. Am J Clin Nutr 2009;90(6):1674e92. 4. Aoe S, Koyama T, Toba Y, Itabashi A, Takada Y. A controlled trial of the effect of milk basic protein (MBP) supplementation on bone metabolism in healthy menopausal women. Osteoporos Int 2005;16(12):2123e8. 5. Kalkwarf HJ, Khoury JC, Lanphear BP. Milk intake during childhood and adolescence, adult bone density, and osteoporotic fractures in US women. Am J Clin Nutr 2003;77(1):257e65.
e7
6. Cadogan J, Eastell R, Jones N, Barker ME. Milk intake and bone mineral acquisition in adolescent girls: randomised, controlled intervention trial. BMJ 1997;315(7118):1255e60. 7. Liu Z, Qiu L, Chen YM, Su YX. Effect of milk and calcium supplementation on bone density and bone turnover in pregnant Chinese women: a randomized controlled trail. Arch Gynecol Obstet 2011;283(2):205e11. 8. Aoe S, Toba Y, Yamamura J, Kawakami H, Yahiro M, Kumegawa M, et al. Controlled trial of the effects of milk basic protein (MBP) supplementation on bone metabolism in healthy adult women. Biosci Biotechnol Biochem 2001;65(4):913e8. 9. Brockwell SE, Gordon IR. A comparison of statistical methods for meta-analysis. Stat Med 2001;20(6):825e40. 10. Budek AZ, Hoppe C, Michaelsen KF, Molgaard C. High intake of milk, but not meat, decreases bone turnover in prepubertal boys after 7 days. Eur J Clin Nutr 2007;61(8):957e62. 11. Green JH, Booth C, Bunning R. Impact of supplementary high calcium milk with additional magnesium on parathyroid hormone and biochemical markers of bone turnover in postmenopausal women. Asia Pac J Clin Nutr 2002;11(4): 268e73. 12. Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJ, Gavaghan DJ, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996;17(1):1e12. 13. Higgins J, Green S. Cochrane handbook for systematic reviews of interventions version 5.0.0 [updated February 2008]. The Cochrane Collaboration Q3. Available from:, www.cochrane- handbook.org; 2008. 14. Review Manager (RevMan). Computer program. Version 4.2 for Windows. Oxford, England: The Cochrane Collaboration; 2003. 15. Du X, Zhu K, Trube A, Zhang Q, Ma G, Hu X, et al. School-milk intervention trial enhances growth and bone mineral accretion in Chinese girls aged 10e12 years in Beijing. Br J Nutr 2004;92(1):159e68. 16. Kristensen M, Jensen M, Kudsk J, Henriksen M, Molgaard C. Short-term effects on bone turnover of replacing milk with cola beverages: a 10-day interventional study in young men. Osteoporos Int 2005;16(12):1803e8. 17. Kruger MC, Booth CL, Coad J, Schollum LM, Kuhn-Sherlock B, Shearer MJ. Effect of calcium fortified milk supplementation with or without vitamin K on biochemical markers of bone turnover in premenopausal women. Nutrition 2006;22(11-12):1120e8. 18. Tong R, Zhang X. The effect of fortified milk on growth and bone development in girls aged 10e12 y. In: The 6th national academic conference of the child and adolescent health section of CPMA & the 3rd academic conference of the school section of Chinese health education association. Nanjing, China: 2004. 19. Volek JS, Gomez AL, Scheett TP, Sharman MJ, French DN, Rubin MR, et al. Increasing fluid milk favorably affects bone mineral density responses to resistance training in adolescent boys. J Am Diet Assoc 2003;103(10):1353e6. 20. Zhang L, Tong R, Zhang X. The effect of fortified milk on growth and development in adolescent girls. Chin J School Health 2006;27(7):611e2. 21. Zhu K, Du X, Cowell CT, Greenfield H, Blades B, Dobbins TA, et al. Effects of school milk intervention on cortical bone accretion and indicators relevant to bone metabolism in Chinese girls aged 10e12 y in Beijing. Am J Clin Nutr 2005;81(5):1168e75. 22. Liu JC, Zhang J, Ouyang QH, Liu Q. Effect of short-term milk drinking on bone mineral density in healthy girls during spurt growth period. Chinese J Clin Rehabil 2002;6(11):1618e9. 23. Ma DF, Qin LQ, Wang PY, Katoh R. Soy isoflavone intake inhibits bone resorption and stimulates bone formation in menopausal women: metaanalysis of randomized controlled trials. Eur J Clin Nutr 2008;62(2):155e61. 24. Kulak CA, Bilezikian JP. Osteoporosis: preventive strategies. Int J Fertil Womens Med 1998;43(2):56e64. 25. Ma GS, Zhang CY, Zhang Q, Zhang L, Zhu K, Hu XQ. Bone mass of middle school girls two years after cessation of calcium-vitamin D-fortified milk supplementation: follow-up study. Acta Nutrimenta Sinica 2006;28(2):139e42. 26. Hu Y, Feng YP, Wang RY, Gao XF, Wang RL. The effect of calcium-fortified milk on bone mineral density in pregnant or suckling period woman. Chinese J Coal Industry Med 2002;5(9):898e9. 27. Takada Y, Aoe S, Kumegawa M. Whey protein stimulated the proliferation and differentiation of osteoblastic MC3T3-E1 cells. Biochem Biophys Res Commun 1996;223(2):445e9. 28. Bischoff-Ferrari HA, Dawson-Hughes B, Baron JA, Kanis JA, Orav EJ, Staehelin HB, et al. Milk intake and risk of hip fracture in men and women: a meta-analysis of prospective cohort studies. J Bone Miner Res 2010.