Hostname: page-component-7c8c6479df-xxrs7 Total loading time: 0 Render date: 2024-03-19T06:32:05.869Z Has data issue: false hasContentIssue false

Associations of dietary intakes of calcium, magnesium and soy isoflavones with osteoporotic fracture risk in postmenopausal women: a prospective study

Published online by Cambridge University Press:  01 August 2022

Yong Cui
Affiliation:
Division of Epidemiology, Department of Medicine, Vanderbilt-Ingram Cancer Center, Vanderbilt Epidemiology Center, Vanderbilt University Medical Center, Nashville, TN, USA
Hui Cai
Affiliation:
Division of Epidemiology, Department of Medicine, Vanderbilt-Ingram Cancer Center, Vanderbilt Epidemiology Center, Vanderbilt University Medical Center, Nashville, TN, USA
Yutang Gao
Affiliation:
Department of Epidemiology, Shanghai Cancer Institute, Renji Hospital, Shanghai Jiaotong University School of Medicine, Shanghai, China
Qi Dai
Affiliation:
Division of Epidemiology, Department of Medicine, Vanderbilt-Ingram Cancer Center, Vanderbilt Epidemiology Center, Vanderbilt University Medical Center, Nashville, TN, USA
Gong Yang
Affiliation:
Division of Epidemiology, Department of Medicine, Vanderbilt-Ingram Cancer Center, Vanderbilt Epidemiology Center, Vanderbilt University Medical Center, Nashville, TN, USA
Wei Zheng
Affiliation:
Division of Epidemiology, Department of Medicine, Vanderbilt-Ingram Cancer Center, Vanderbilt Epidemiology Center, Vanderbilt University Medical Center, Nashville, TN, USA
Xiao-Ou Shu*
Affiliation:
Division of Epidemiology, Department of Medicine, Vanderbilt-Ingram Cancer Center, Vanderbilt Epidemiology Center, Vanderbilt University Medical Center, Nashville, TN, USA
*
*Corresponding author: Xiao-Ou Shu, fax 615-936-8291, email xiao-ou.shu@vanderbilt.edu

Abstract

The role of dietary factors in osteoporotic fractures (OFs) in women is not fully elucidated. We investigated the associations between incidence of OF and dietary calcium, magnesium and soy isoflavone intake in a longitudinal study of 48 584 postmenopausal women. Multivariable Cox regression was applied to derive hazard ratios (HRs) and 95 % confidence intervals (CIs) to evaluate associations between dietary intake, based on the averages of two assessments that took place with a median interval of 2⋅4 years, and fracture risk. The average age of study participants is 61⋅4 years (range 43⋅3–76⋅7 years) at study entry. During a median follow-up of 10⋅1 years, 4⋅3 % participants experienced OF. Compared with daily calcium intake ≤400 mg/d, higher calcium intake (>400 mg/d) was significantly associated with about a 40–50 % reduction of OF risk among women with a calcium/magnesium (Ca/Mg) intake ratio ≥1⋅7. Among women with prior fracture history, high soy isoflavone intake was associated with reduced OF risk; the HR was 0⋅72 (95 % CI 0⋅55, 0⋅93) for the highest (>42⋅0 mg/d) v. lowest (<18⋅7 mg/d) quartile intake. This inverse association was more evident among recently menopausal women (<10 years). No significant association between magnesium intake and OF risk was observed. Our findings provide novel information suggesting that the association of OF risk with dietary calcium intake was modified by Ca/Mg ratio, and soy isoflavone intake was modified by history of fractures and time since menopause. Our findings, if confirmed, can help to guide further dietary intervention strategies for OF prevention.

Type
Research Article
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.
Copyright
Copyright © The Author(s), 2022. Published by Cambridge University Press on behalf of The Nutrition Society

Introduction

Osteoporosis is characterised by reduced bone mass and fragmentation of bone architecture, with a consequent increase in bone fragility and susceptibility to fractures(Reference Sambrook and Cooper1,Reference Watts and Manson2) . Osteoporotic fractures (OFs) are a major cause of disability and reduced quality of life, particularly in postmenopausal women(Reference Watts and Manson2). It has been estimated that approximately 1 in 2 women aged 50 years or older will experience an OF in their remaining lifetime, imposing a considerable economic burden on health services(Reference Melton3Reference Burge, Dawson and Solomon5). Previous studies have investigated relationships between OF and several nutritional and/or dietary factors; however, the results are mixed(Reference Zheng, Lee and Chun6,Reference Zheng, Lee and Chun7) . More information is needed on the roles of these modifiable factors in OF for the development of non-pharmacologic preventive strategies, accordingly.

Both calcium and magnesium are major components of bone and essential micronutrients required to maintain bone homeostasis and play important roles in bone health(8,Reference Jahnen-Dechent and Ketteler9) . Dietary calcium intake has been recommended for osteoporosis prevention; however, the recommendations vary by country(Reference Yao, Bennett and Mafham10,Reference Francis, Anderson and Patel11) . Furthermore, studies have shown that magnesium may interact with calcium and/or vitamin D, interfere with calciotropic hormones, and has been known as a natural calcium antagonist(Reference Rosanoff, Dai and Shapses12Reference Iseri and French14). To date, information on the relationship between dietary magnesium intake and OF risk is sparse, and whether dietary magnesium intake is related to the risk of OF or may influence the estimation of the association between dietary calcium intake and OF is largely unknown.

Estrogen plays an important role in skeletal homeostasis and regulates bone remodelling(Reference Nakamura, Imai and Matsumoto15,Reference Manolagas, O'Brien and Almeida16) , while the decline in estrogen levels associated with menopause causes bone loss in women(Reference Raisz17). Estrogen exerts its effect on target cells through binding estrogen receptors (ERs);(Reference Deroo and Korach18) thus, modulators of ERs play an important role in bone health. Isoflavones, rich in soybeans and soy-based products, are major types of phytoestrogen, with a noticeable property as a natural selective ER modulator(Reference Oseni, Patel and Pyle19). In vitro experiments and in vivo animal studies have shown that isoflavones have potential bone-specific effects via estrogenic/antiestrogenic effects and other biologic mechanisms(Reference Gómez-Zorita, González-Arceo and Fernández-Quintela20,Reference Pabich and Materska21) . Epidemiological and clinical evidence support that dietary isoflavones may attenuate menopause-induced osteoporotic bone loss(Reference Alekel, Germain and Peterson22Reference Abdi, Alimoradi and Haqi24). However, information is still limited regarding whether the association between dietary isoflavone intake and OF risk is modified by timing regarding menopause and other host factors as previously suggested(Reference Zheng, Hsieh and Cai25).

In the present study, we investigated the incidence of OF and evaluated their associations with dietary calcium, magnesium and soy isoflavone intake in a prospective observational cohort of over 48 000 postmenopausal women.

Methods

Study population

Participants of the study were drawn from the Shanghai Women's Health Study (SWHS), a large population-based prospective cohort study conducted in urban Shanghai, China. Detailed descriptions of the study design and methods have been published elsewhere(Reference Zheng, Chow and Yang26). Briefly, 74 940 women aged 40–74 years were recruited from seven typical urban communities in Shanghai between 1996 and 2000, with a 92 % response rate. At the study enrolment, each participant signed a consent form and completed an in-person survey conducted by trained interviewers. The information collected at enrolment included socio-demographic characteristics, dietary habits including soy intake, physical activity, menstrual and reproductive history, smoking, drinking and other lifestyle factors and exposures, as well as medical history. Anthropometric measurements were also taken. The cohort has been followed up between 2000 and 2016, through in-person surveys every 2–4 years, a total of five times to update exposure information and collect information on changes of health status, including bone fractures and vital status. Annual record linkages with the vital statistics registry were carried out to ensure a complete ascertainment of mortality information. Response rates for the five in-person follow-up surveys were over 91 % (99⋅7, 98⋅8, 95⋅0, 92⋅5 and 91⋅0 %, respectively). In the first and second follow-up surveys, participants were only asked about time (month and year) and anatomic site(s) of a bone fracture, but no information was collected on the cause of the fracture; thus, the type of bone fracture could not be assessed. Therefore, for the current analysis, we only included OF and non-osteoporotic fractures (non-OFs) that occurred after the second follow-up survey as outcomes of the study. In other words, the cohort follow-up for the current analysis started at the completion of the second follow-up. Fractures developed between the SWHS enrolment and the second follow-up survey were treated as history of bone fractures and the menopausal status was based on the data of the second follow-up survey. We excluded 3248 women who did not participate in the first and/or second follow-up surveys, 22 778 women with premenopausal status at the second follow-up survey, and 270 postmenopausal women who developed both OF and non-OF after the second follow-up survey, leaving a total of 48 584 postmenopausal women for the current analysis (Fig. 1). The SWHS was approved by the institutional review boards of all participating institutions. This study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving human subjects/patients were approved by the institutional review boards of all participating institutions. Written informed consent was obtained from all subjects/patients.

Fig. 1. Flowchart for the participant selection process.

Study variables and covariates assessment

Dietary information was collected using a validated food frequency questionnaire (FFQ)(Reference Villegas, Yang and Liu27) at the study enrolment and at the first FFQ follow-up for the present study. A total of eighty-one food items were included in the SWHS FFQ. For each food item or food group, subjects were asked how frequently (daily, weekly, monthly, annually or never) they consumed the food or food group, which was followed by a question on the amount consumed in lians per unit of time. Lian is a unit of weight in China (1 lian = 50 g). Soy intake assessed in the study included consumption of tofu, soy milk, fresh soybeans and other soy products. Daily intakes of calories, macro- and micronutrients, vitamin D, calcium, magnesium, protein, soy isoflavones and major isoflavone components (genistein, daidzein and glycitein) were derived from FFQ data by summing the products of individual food intake amounts and nutrient contents of food items based on the Chinese Food Composition Tables(Reference Villegas, Yang and Liu27,Reference Yang, Wang and Pan28) . To improve the dietary assessment, we averaged the dietary intake data collected at study enrolment and the first follow-up and applied them in the present study. For those women (n = 2217, 4⋅5 %) who missed information on the dietary assessment at the first FFQ follow-up and/or were diagnosed with cancer, diabetes, stroke or myocardial infarction between the study enrolment and the first follow-up survey, only dietary intake data at enrolment were used due to the concern that these diseases and associated treatments would change dietary habits. Dietary intakes were categorised by quartile distributions for the analyses. Calcium and magnesium intakes were categorised from ≤400 to >800 mg/d and from ≤200 to >400 mg/d, respectively, while soy isoflavone intake was categorised by quartile distributions.

Outcomes

The primary outcome of interest was the occurrence of OF. During the in-person follow-up surveys, participants were asked if they had a bone fracture since the last survey. If a participant answered ‘yes’, she was asked to provide further information on time (month and year), anatomic site(s) and cause of fracture. Fracture sites were coded with ICD-9 or ICD-10. Fractures in anatomic sites commonly associated with osteoporosis include ICD-9 codes 805, 806, 807, 808, 810, 812, 813, 818, 819, 820, 821, 822, 823, 824 and ICD-10 codes S22, S32, S42, S52, S72, S82 and M80. For causes of fractures, participants could select (1) car accident or physical trauma, (2) fall when riding a bicycle, (3) fall by sliding/fall from standing height, (4) fall down from a high place (providing height in m), and (5) others (specify the cause). OFs were defined as low-trauma bone fractures (e.g. due to falls by sliding/from standing height) and occurring in anatomic sites commonly associated with osteoporosis, with exception of those caused by malignancy or other pathologic fractures not related osteoporosis(Reference Zheng, Hsieh and Cai25,Reference Warriner, Patkar and Curtis29) . Non-OFs included any fracture other than OF. Because trauma, which is unrelated to diet, is the major cause of non-OF, we also included non-OF as a comparison group to evaluate the validity of the outcome assessment and study findings(Reference Prasad and Jena30,Reference Groenwold31) .

Statistical analysis

Included in the current analysis were 48 584 postmenopausal women. We used the Kaplan–Meier product-limit method to estimate the 10-year occurrence rate, and a log-rank test to compare 10-year occurrence rates among participant groups defined by covariates. Multivariable Cox regression model was applied to evaluate the associations of dietary calcium intake, dietary magnesium intake and soy isoflavone intake with incidence of OF and non-OF, measured by hazard ratios (HRs) and 95 % confidence intervals (CIs). Entry time was the date of the second follow-up survey completion, and exit time was the date of fracture occurrence, date of cancer, stroke or myocardial infarction diagnosis (due to a concern that these events and their associated treatments may change a participant's dietary habits), date at death, or last follow-up, whichever came first. Covariates adjusted for in the models included known/suspected risk factors for bone fractures, based on literature and factors that were significantly associated with bone fracture risk in the univariate analysis of our own data. These included age (continuous variable), educational level (less than high school/high school graduate/higher than high school), cigarette smoking status (never/ever), alcohol consumption (yes/no), regular exercise (yes/no), body mass index (BMI, continuous variable), calcium supplement use (non-user/ever-user), comorbidity (Charlson's score, 0/1/≥2), fracture history (yes/no) and dietary intakes of calories and vitamin D (quartiles). All covariates were based on information collected at study entry except age, which was collected at the second follow-up survey completion, and dietary intakes of calories, calcium, magnesium and vitamin D, which were the averages of baseline and first FFQ follow-up survey data. Dietary calcium and magnesium intakes were mutually adjusted in the model. The same models were applied in multiplicative interactions between isoflavone/each composition and bone fracture history or year(s) since menopause. Proportional hazard assumption was assessed by adding an interaction term between an exposure and time (exposure*log [time]) in the model for testing. No major violation was discovered for the proportionality assumption. All statistical tests were based on two-tailed probability and a significance level set at alpha (α) < 0⋅05.

Results

During a median follow-up of 10⋅1 years (interquartile range: 9⋅3–11⋅0) since the completion of the second follow-up survey, among 48 584 postmenopausal women, 2075 developed OF, 3221 developed non-OF and 43 288 remained free of any fractures. Table 1 shows the 10-year occurrence of two types of bone fractures by participant characteristics among postmenopausal women. For OF, a higher 10-year occurrence rate was observed among women with a prior history of fractures, older age (especially older than 69 years), low income, ever cigarette smoking, BMI less than 18⋅5, regular exercise, comorbidity (especially Chalson's comorbidity score ≥2), longer breast-feeding (especially >36 months) and ever use of calcium supplements. For non-OF, however, the rate did not increase with age. A higher 10-year occurrence rate was seen among women with a prior history of fractures, high educational level, normal BMI (18⋅5–25⋅0), regular exercise, breast-feeding for 1–36 months and ever use of calcium supplements.

Table 1. 10-year occurrence rate by participant characteristics and fracture type among postmenopausal women, the SWHS (N = 48 584)

Table 2 shows the associations of dietary calcium and magnesium intakes with OF overall and by calcium/magnesium (Ca/Mg) intake ratio and calcium supplement use. Overall, no significant association was found between dietary calcium intake and OF. Analysis stratified by the dietary Ca/Mg ratio median (ratio <1⋅7 or ≥1⋅7, median cut), however, showed that, compared with lower calcium intake (≤400 mg/d), higher calcium intake (>400 mg/d) was associated with about a 40–50 % reduction of OF risk among women with a dietary Ca/Mg ratio ≥1⋅7 (HR 0⋅63, 95 % CI 0⋅48, 0⋅82 to HR 0⋅52, 95 % CI 0⋅34, 0⋅80). This association was observed among women with (range from HR 0⋅67, 95 % CI 0⋅44, 1⋅01 to HR 0⋅51, 95 % CI 0⋅28, 0⋅94) and without (range from HR 0⋅65, 95 % CI 0⋅41, 1⋅02 to HR 0⋅53, 95 % CI 0⋅29, 0⋅99) use of calcium supplements. However, no significant association was observed between dietary calcium intake and OF among women with a dietary Ca/Mg ratio <1⋅7. No significant associations were observed between dietary magnesium intake and OF or non-OF overall or by Ca/Mg ratio and calcium supplement use. Our observations showed that the associations of calcium or magnesium intakes did not vary by history of bone fractures and no significant association was found between calcium intake and non-OF, regardless of Ca/Mg ratio and calcium supplement use (data not shown).

Table 2. Associations of calcium and magnesium intakes with osteoporotic fracture by Ca/Mg ratio and Ca supplement use among postmenopausal women

a In the group with Ca/Mg ratio < median (1⋅7), mean Ca/Mg ratio = 1⋅44; in the group with Ca/Mg ratio ≥ median (1⋅7), mean Ca/Mg ratio = 2⋅07.

b Adjusting for: Age, history of bone fracture, income, educational level, cigarette smoking status, alcohol consumption, regular exercise, BMI, Charlson's Score, breasting time, calcium supplement use, daily dietary intake of calories and vitamin D. Dietary calcium and magnesium intakes were mutually adjusted in the model.

Table 3 shows the associations of soy isoflavone intake with OF overall and by prior fracture history among postmenopausal women. Overall, no associations were observed between soy isoflavone intake and OF; however, a significant interaction between soy isoflavone intake and prior fracture history was detected (P for interaction = 0⋅002). Among women with a prior bone fracture history, compared with the lowest quartile (<1⋅7 mg/d), the highest quartile (>42⋅0 mg/d) of soy isoflavone intake was associated with a 28 % reduction of OF risk (HR 0⋅72, 95 % CI 0⋅55, 0⋅93). By contrast, high isoflavone intake was associated with an elevated OF risk, seen among women without a prior fracture history (HR 1⋅22, 95 % CI 1⋅01, 1⋅48). Similar association patterns were also observed between major components of soy isoflavones (daidzein, genistein and glycitein) and OF. No significant association was observed between soy isoflavone intake and non-OF, regardless of prior history of bone fractures (data not shown).

Table 3. Associations of soy isoflavones intake with osteoporotic fracture by any bone fracture history among postmenopausal women

a Adjusting for: Age, BF history, income, educational level, cigarette smoking status, alcohol consumption, regular exercise, BMI, Charlson's Score, breasting time, calcium supplement use, daily dietary intake of calories, vitamin D, calcium and magnesium.

b P-value for interaction between soy isoflavones and BF history.

We further examined associations of soy isoflavone intake with bone fractures by year(s) since menopause among postmenopausal women (Table 4). For women with a prior fracture history, soy isoflavone intake was found to be inversely associated with OF risk among women who became menopausal within 10 years (HR 0⋅80, 95 % CI 0⋅56, 1⋅16 for isoflavone intake between 18⋅7 and 29⋅0 mg/d, HR 0⋅65, 95 % CI 0⋅43, 1⋅00 for isoflavone intake between 29⋅1 and 42⋅0 mg/d, HR 0⋅59, 95 % CI 0⋅37, 0⋅96 for isoflavone intake >42⋅0 mg/d), but not among those who had been menopausal for over 10 years. No association was observed between soy isoflavone intake and non-OF risk among postmenopausal women, regardless of prior history of bone fracture and year(s) since menopause (data not shown).

Table 4. Associations of soy isoflavones intake with osteoporotic fractures by year(s) since menopause among postmenopausal women

a Adjusting for: Age, income, educational level, cigarette smoking status, alcohol consumption, regular exercise, BMI, Charlson's Score, breasting time, calcium supplement use, daily dietary intake of calories, vitamin D, calcium and magnesium.

b P-value for interaction between soy isoflavones and year(s) since menopause.

Discussion

In this large-scale longitudinal study, we found that dietary calcium intake was inversely associated with OF among postmenopausal women who had a dietary Ca/Mg ratio above 1⋅7, regardless of calcium supplement use. We also found that high levels of soy isoflavone intake and intakes of major isoflavone components (genistein, daidzein or glycitein) were associated with a reduced risk of OF in postmenopausal women who had a prior history of bone fracture, particularly in those who became menopausal within 10 years of dietary assessments.

Although dietary calcium intake has been recommended for osteoporosis prevention, previous studies examining the associations of dietary calcium intake with OF risk have yielded mixed results(Reference Yao, Bennett and Mafham10,Reference Cumming and Nevitt32Reference Nieves, Barrett-Connor and Siris36) . In an early systematic review and meta-analysis(Reference Cumming and Nevitt32), Cumming et al. analysed fourteen studies of calcium supplements (including four randomised trials), eighteen studies of dietary calcium and hip fracture (no randomised trials), five studies of dietary calcium and other fracture sites (no randomised trials), as well as sixteen additional observational studies of dietary calcium and hip fracture (eleven case-control studies and five cohort studies). Meta-analysis of sixteen observational studies on dietary calcium and hip fracture revealed a significant heterogeneity across studies. The pooled estimate on dietary calcium intake and hip structure was 0⋅96 (95 % CI 0⋅93, 0⋅99) per 300 mg/d increase in calcium intake. The authors concluded that the overall result supports the current clinical and public health policy of recommending increased calcium intake among older women for fracture prevention. A recent systematic review and meta-analysis(Reference Yao, Bennett and Mafham10) showed that vitamin D supplementation alone was associated with a reduced risk of fractures, and daily use of both calcium and vitamin D supplementation reduced the risk of hip fracture. However, several other studies found no strong evidence that dietary calcium intake was associated with reduced risk of OFs(Reference Warensjö, Byberg and Melhus33Reference Nieves, Barrett-Connor and Siris36). For example, using data from the National Osteoporosis Risk Assessment (NORA) study(Reference Nieves, Barrett-Connor and Siris36), Nieves et al. reported that calcium and vitamin D intakes influenced bone mass but not short-term fracture risk in Caucasian postmenopausal women. In the present study, we found that among postmenopausal women with a dietary Ca/Mg ratio >1⋅7, among whom the mean Ca/Mg intake ratio was 2⋅07, dietary calcium intake was inversely associated with OF. However, no significant association was found among women whose Ca/Mg intake ratio was ≤1⋅7, among whom the mean Ca/Mg intake ratio was 1⋅44. Ca/Mg ratio has been previously suggested to modify the effect of Ca on mortality or cancer incidence. Dai and his colleagues reported that among men with a Ca/Mg ratio >1⋅7, increased intake of Ca was associated with reduced risk of total mortality and mortality due to cancer(Reference Dai, Shu and Deng13). In a recent study, it was also reported that the inverse association between calcium intake and distal colorectal cancer was modified by the Ca/Mg ratio(Reference Zhao, Giri and Zhu37). Our results are in line with findings from these previous observational studies supporting that the Ca effect on health may be modified by Ca/Mg ratio. Additionally, a recent precision-based randomised trial showed that optimal magnesium status was important for optimising vitamin D status, which is essential to calcium absorption(Reference Dai, Zhu and Manson38,Reference Heaney39) . Our study finding suggests that a proper Ca/Mg ratio may be required for bone health effects of calcium, and failure to take this property into consideration when evaluating the health effects of calcium may lead to erroneous conclusions.

There were several reports on dietary magnesium intake and risk of osteoporosis fractures; however, findings were conflicting. In the Women's Health Initiative Observational Study, Orchard et al. reported that a lower magnesium intake is associated with lower bone mineral density (BMD) of the hip, but a magnesium intake slightly greater than the Recommended Dietary Allowance is associated with increased lower-arm and wrist fractures(Reference Orchard, Larson and Alghothani40). A systematic review revealed a positive correlation between magnesium intake and BMD, but no association was found between high magnesium intake and fracture(Reference Farsinejad-Marj, Saneei and Esmaillzadeh41). Consistently, we also did not find a significant association between magnesium intake and OF or non-OF. Interestingly, a recent study investigated both magnesium intake and serum level of magnesium, and found an inverse association between magnesium concentration in serum and fracture risk in middle-aged Caucasian men, but a null association for magnesium intake(Reference Kunutsor, Whitehouse and Blom42). Additional research is needed to fully understand the role of magnesium intake and absorption in fracture.

Studies have well documented that a prior history of fractures is an important risk factor for osteoporosis and OF(Reference Toth, Banefelt and Åkesson43,Reference Johansson, Siggeirsdottir and Harvey44) . For example, Toth et al. (Reference Toth, Banefelt and Åkesson43) evaluated a retrospective, observational study of a total of 35 146 women aged 55–90 years. The authors found that among 7180 hip, 2786 clinical vertebral, and 25 180 non-hip/non-vertebral (NHNV) index fractures, 38 % of women with hip, 38 % with clinical vertebral and 25 % with NHNV index fractures had one or more previous fractures. Following any index fracture, cumulative incidence of a new fracture was over 11 % over the subsequent 24 months. These data demonstrate that almost one-third of women aged 55–90 years suffering a new fracture have had a previous fragility fracture. Among postmenopausal women from the SWHS who had a prior history of bone fracture, 7⋅1 % developed OF during the 10-year follow-up, while among those without a history of bone fracture, the 10-year occurrence rate was 3⋅9 %. Furthermore, we found that among women with a prior history of fractures, a higher level of soy isoflavone intake (>42 mg/d) was associated with a 28 % reduction of OF risk. Similar association patterns were also observed between major components of soy isoflavones (genistein, daidzein and glycitein) and OF. Thus, our finding supports the notion that prior history of fractures is an important risk factor for OF, and further provides the first evidence that prior history of fractures modifies the association of soy isoflavone intake with OF in postmenopausal women. On the other hand, we found that high dietary isoflavone intake was associated with an increased risk of OF among women without a bone fracture history.

Menopause is associated with increased bone resorption and decreased BMD because estrogen levels drop. Isoflavone has both estrogenic/antiestrogenic effects(Reference Gómez-Zorita, González-Arceo and Fernández-Quintela20,Reference Pabich and Materska21) . Several studies have reported that soy isoflavones attenuate menopause-induced osteoporotic bone loss and fractures among women(Reference Alekel, Germain and Peterson22Reference Zheng, Hsieh and Cai25). A review of randomised controlled trials suggested that soy isoflavone consumption during menopausal transition may prevent a reduction in BMD and promote bone health(Reference Abdi, Alimoradi and Haqi24). Indeed, our previous study found that soy isoflavone intake was associated with a reduced risk of incident fractures in pre-/perimenopausal breast cancer survivors(Reference Zheng, Hsieh and Cai25). Findings of the present study support the notion that soy isoflavone intake may protect against menopause-related osteoporosis and OF among high-risk women, i.e. those with a prior history of bone fractures. We have no ready explanation for the opposite association, i.e. an increase of OF with high isoflavone intake found for women without a prior history of bone fractures. We speculate that women without a history of bone fracture may have high levels of endogenous estrogen, and among them, isoflavones may exert their antiestrogenic effects giving dual estrogen agonist and antagonistic effects of isoflavones. This hypothesis will need to be evaluated in further studies.

The strengths of the present study include the prospective study design, large sample size, high response rates, repeated dietary assessments using a validated FFQ and parallel analyses of non-OF. However, our study also has several limitations. First, our information on fracture incidences and exposures was self-reported. No information was available on BMD and history of OF. Thus, misclassification on outcome assessment is possible, although a major bone fracture such as OF is more likely be accurate. The latter is supported by a validity study which compared self-reported fractures with data from medical records and found that the validity of self-reports for hip and forearm/wrist fractures was high(Reference Chen, Kooperberg and Pettinger45). In our study, we defined OF as low-trauma bone fractures (e.g. due to falls by sliding/fall from standing height) and occurring in anatomic sites commonly associated with osteoporosis, which takes into consideration both cause and anatomic site(s) to minimise the outcome misclassification bias. We found that OF and non-OF had different associations with age and several other socio-economic and lifestyle factors, in addition to the different association patterns with calcium and isoflavones. These data provide indirect support to the validity of our assessment of OF, as well as our study findings. Second, since our study was not originally designed for investigating OF, information on a number of potential risk factors for osteoporosis and OF, such as a family history of osteoporosis and hyperthyroidism or hyperparathyroidism, was not collected in the study. Thus, residual confounding may remain. Third, while we used data from two FFQs administered at the baseline survey and the first follow-up survey to improve the dietary assessment, changes in dietary intake during subsequent follow-ups were not captured, which may affect the associations between nutrients under study and fracture risk.

In summary, our study shows that among postmenopausal women, the calcium and OF association was modified by the Ca/Mg ratio. Dietary calcium intake more than 400 mg daily was associated with about a 40–50 % reduction of OF risk only when a daily dietary Ca/Mg ratio is equal to or above 1⋅7. Calcium intake was not associated with OF when Ca/Mg ratio was less than 1⋅7. We also provide evidence that a high level of soy isoflavone intake is associated with a reduced risk of OF, especially in postmenopausal women who had a prior bone fracture history and were within 10 years of menopause. These novel findings, if confirmed, would not only have a direct impact on the development strategies to prevent osteoporosis and OF among Chinese women, but also contribute to our understanding of the interplays among nutrients and between host characteristics and nutrient intake, as well as time of dietary intake on bone health, for postmenopausal women in general.

Acknowledgements

The authors wish to thank participants and research team members, particularly the Shanghai Cancer Institute research team members, for their contributions to the study. We thank Mary Shannon Byers for her assistance with manuscript preparation.

The SWHS was supported by the U.S. National Institutes of Health (UM1 CA182910 to W. Z.). The funding source had no role in the design and conduct of the study; collection, management, analysis and interpretation of the data; preparation, review or approval of the manuscript; and decision to submit the manuscript for publication.

Concept and design: X. O. S. and W. Z. Drafting of the manuscript: Y. C. and X. O. S. Statistical analysis: Y. C., H. C. and X. O. S. Interpretation of data and critical revision of the manuscript for important intellectual content: all authors. Administrative, technical or material support: X. O. S. and W. Z. All authors had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. All authors have read and approved the final manuscript.

The authors declare no conflicts of interest.

Data will be provided upon request.

The SWHS was approved by the institutional review boards of all participating institutions.

All participants provided written informed consent.

All authors consent to the publication of this manuscript.

SWHS data used in this research project will be made available following the study data sharing policy posted online at: https://swhs-smhs.app.vumc.org/smhs_index.php.

Footnotes

These two authors contributed to the study equally and share the first authorship.

References

Sambrook, P & Cooper, C (2006) Osteoporosis. Lancet 367, 20102018. doi:10.1016/S0140-6736(06)68891-0.CrossRefGoogle ScholarPubMed
Watts, NB & Manson, JE (2017) Osteoporosis and fracture risk evaluation and management: shared decision making in clinical practice. JAMA 317, 253254. doi:10.1001/jama.2016.19087.CrossRefGoogle ScholarPubMed
Melton, LJ III (2003) Adverse outcomes of osteoporotic fractures in the general population. J Bone Miner Res 18, 11391141. doi:10.1359/jbmr.2003.18.6.1139.CrossRefGoogle ScholarPubMed
Ensrud, KE (2013) Epidemiology of fracture risk with advancing age. J Gerontol A Biol Sci Med Sci 68, 12361242. doi:10.1093/gerona/glt092.CrossRefGoogle ScholarPubMed
Burge, R, Dawson, HB, Solomon, DH, et al. (2007) Incidence and economic burden of osteoporosis-related fractures in the United States, 2005–2025. J Bone Miner Res 22, 465475. doi:10.1359/jbmr.061113.CrossRefGoogle ScholarPubMed
Zheng, X, Lee, SK & Chun, OK (2016) Soy isoflavones and osteoporotic bone loss: a review with an emphasis on modulation of bone remodeling. J Med Food 19, 114. doi:10.1089/jmf.2015.0045.CrossRefGoogle ScholarPubMed
Zheng, X, Lee, SK & Chun, OK (2016) Soy isoflavones and osteoporotic bone loss: a review with an emphasis on modulation of bone remodeling. J Med Food 19, 114. doi:10.1089/jmf.2015.0045.CrossRefGoogle ScholarPubMed
Calcium and Vitamin D: Important at Every Age. Internet: https://www.bones.nih.gov/health-info/bone/bone-health/nutrition/calcium-and-vitamin-d-important-every-age (accessed 12 February 2020).Google Scholar
Jahnen-Dechent, W & Ketteler, M (2012) Magnesium basics. Clin Kidney J 5, i3i14. doi:10.1093/ndtplus/sfr163.CrossRefGoogle ScholarPubMed
Yao, P, Bennett, D, Mafham, M, et al. (2019) Vitamin D and calcium for the prevention of fracture: a systematic review and meta-analysis. JAMA Netw Open 2, e1917789. doi:10.1001/jamanetworkopen.2019.17789.CrossRefGoogle ScholarPubMed
Francis, RM, Anderson, FH, Patel, S, et al. (2006) Calcium and vitamin D in the prevention of osteoporotic fractures. QJM 99, 355363. doi:10.1093/qjmed/hcl031.CrossRefGoogle ScholarPubMed
Rosanoff, A, Dai, Q & Shapses, SA (2016) Essential nutrient interactions: does low or suboptimal magnesium status interact with vitamin D and/or calcium status? Adv Nutr 7, 2543. doi:10.3945/an.115.008631.CrossRefGoogle ScholarPubMed
Dai, Q, Shu, XO, Deng, X, et al. (2013) Modifying effect of calcium/magnesium intake ratio and mortality: a population-based cohort study. BMJ Open 3, e002111. doi:10.1136/bmjopen-2012-002111.CrossRefGoogle ScholarPubMed
Iseri, LT & French, JH (1984) Magnesium: Nature's physiologic calcium blocker. Am Heart J 108, 188193. doi: 10.1016/0002-8703(84)90572-6.CrossRefGoogle ScholarPubMed
Nakamura, T, Imai, Y, Matsumoto, T, et al. (2007) Estrogen prevents bone loss via estrogen receptor alpha and induction of Fas ligand in osteoclasts. Cell 130, 811823. doi:10.1016/j.cell.2007.07.025.CrossRefGoogle ScholarPubMed
Manolagas, SC, O'Brien, CA & Almeida, M (2013) The role of estrogen and androgen receptors in bone health and disease. Nat Rev Endocrinol 9, 699712. doi:10.1038/nrendo.2013.179.CrossRefGoogle ScholarPubMed
Raisz, L (2005) Pathogenesis of osteoporosis: concepts, conflicts, and prospects. J Clin Invest 115, 33183325. doi:10.1172/JCI27071.CrossRefGoogle ScholarPubMed
Deroo, BJ & Korach, KS (2006) Estrogen receptors and human disease. J Clin Invest 116, 561570. doi:10.1172/JCI27987.CrossRefGoogle ScholarPubMed
Oseni, T, Patel, R, Pyle, J, et al. (2008) Selective estrogen receptor modulators and phytoestrogens. Planta Med 74, 16561665. doi:10.1055/s-0028-1088304.CrossRefGoogle ScholarPubMed
Gómez-Zorita, S, González-Arceo, M, Fernández-Quintela, A, et al. (2020) Scientific evidence supporting the beneficial effects of isoflavones on human health. Nutrients 12, 3853. doi:10.3390/nu12123853.CrossRefGoogle ScholarPubMed
Pabich, M & Materska, M (2019) Biological effect of soy isoflavones in the prevention of civilization diseases. Nutrients 11, 1660. doi:10.3390/nu11071660.CrossRefGoogle ScholarPubMed
Alekel, DL, Germain, AS, Peterson, CT, et al. (2000) Isoflavone-rich soy protein isolate attenuates bone loss in the lumbar spine of perimenopausal women. Am J Clin Nutr 72, 844852. doi:10.1093/ajcn/72.3.844.CrossRefGoogle ScholarPubMed
Chen, YM, Ho, SC, Lam, SS, et al. (2003) Soy isoflavones have a favorable effect on bone loss in Chinese postmenopausal women with lower bone mass: a double-blind, randomized, controlled trial. J Clin Endocrinol Metab 88, 47404747. doi:10.1210/jc.2003-030290.CrossRefGoogle Scholar
Abdi, F, Alimoradi, Z, Haqi, P, et al. (2016) Effects of phytoestrogens on bone mineral density during the menopause transition: a systematic review of randomized, controlled trials. Climacteric 19, 535545. doi:10.1080/13697137.2016.1238451.CrossRefGoogle Scholar
Zheng, N, Hsieh, E, Cai, H, et al. (2019) Soy food consumption, exercise, and body mass index and osteoporotic fracture risk among breast cancer survivors: The Shanghai Breast Cancer Survival Study. JNCI Cancer Spectr 3, pkz017. doi:10.1093/jncics/pkz017.CrossRefGoogle ScholarPubMed
Zheng, W, Chow, WH, Yang, G, et al. (2005) The Shanghai Women's Health Study: rationale, study design, and baseline characteristics. Am J Epidemiol 162, 11231131. doi:10.1093/aje/kwi322.CrossRefGoogle ScholarPubMed
Villegas, R, Yang, G, Liu, D, et al. (2007) Validity and reproducibility of the food-frequency questionnaire used in the Shanghai men's health study. Br J Nutr 97, 9931000. doi:10.1017/S0007114507669189.CrossRefGoogle ScholarPubMed
Yang, XY, Wang, GY & Pan, XC (2002) Chinese Food Composition Tables 2002, Chinese edition. Beijing: Beijing Medical University Pub.Google Scholar
Warriner, AH, Patkar, NM, Curtis, JR, et al. (2011) Which fractures are most attributable to osteoporosis? J Clin Epidemiol 64, 4653. doi:10.1016/j.jclinepi.2010.07.007.CrossRefGoogle ScholarPubMed
Prasad, V & Jena, AB (2013) Prespecified falsification end points. JAMA 309, 241242. doi:10.1001/jama.2012.96867.CrossRefGoogle ScholarPubMed
Groenwold, RHH (2013) Falsification end points for observational studies. JAMA 309, 17691771. doi:10.1001/jama.2013.3089.CrossRefGoogle ScholarPubMed
Cumming, RG & Nevitt, MC (1997) Calcium for prevention of osteoporotic fractures in postmenopausal women. J Bone Miner Res 12, 13211329. doi:10.1359/jbmr.1997.12.9.1321.CrossRefGoogle ScholarPubMed
Warensjö, E, Byberg, L, Melhus, H, et al. (2011) Dietary calcium intake and risk of fracture and osteoporosis: prospective longitudinal cohort study. Br Med J 24, 342d1473. doi:10.1136/bmj.d1473.Google Scholar
Michaëlsson, K, Melhus, H, Bellocco, R, et al. (2003) Dietary calcium and vitamin D intake in relation to osteoporotic fracture risk. Bone 32, 694703. doi:10.1016/s8756-3282(03)00048-6.CrossRefGoogle ScholarPubMed
Feskanich, D, Willett, WC & Colditz, GA (2003 Feb) Calcium, vitamin D, milk consumption, and hip fractures: a prospective study among postmenopausal women. Am J Clin Nutr 77, 504511. doi:10.1093/ajcn/77.2.504.CrossRefGoogle Scholar
Nieves, JW, Barrett-Connor, E, Siris, ES, et al. (2008 May) Calcium and vitamin D intake influence bone mass, but not short-term fracture risk, in Caucasian postmenopausal women from the National Osteoporosis Risk Assessment (NORA) study. Osteoporos Int 19, 673679. doi:10.1007/s00198-007-0501-2.CrossRefGoogle Scholar
Zhao, J, Giri, A, Zhu, X, et al. (2019) Calcium:magnesium intake ratio and colorectal carcinogenesis, results from the prostate, lung, colorectal, and ovarian cancer screening trial. BJC 121, 796804. doi: 10.1038/s41416-019-0579-2.CrossRefGoogle ScholarPubMed
Dai, Q, Zhu, X, Manson, JE, et al. (2018) Magnesium status and supplementation influence vitamin D status and metabolism: results from a randomized trial. Am J Clin Nutr 108, 12491258. doi:10.1093/ajcn/nqy274.CrossRefGoogle ScholarPubMed
Heaney, RP (2008) Vitamin D and calcium interactions: functional outcomes. Am J Clin Nutr 88, 541S544S. doi:10.1093/ajcn/88.2.541S.CrossRefGoogle ScholarPubMed
Orchard, TS, Larson, JC, Alghothani, N, et al. (2014) Magnesium intake, bone mineral density, and fractures: results from the Women's Health Initiative Observational Study. Am J Clin Nutr 99, 926933. doi:10.3945/ajcn.113.067488.CrossRefGoogle ScholarPubMed
Farsinejad-Marj, M, Saneei, P & Esmaillzadeh, A (2016) Dietary magnesium intake, bone mineral density and risk of fracture: a systematic review and meta-analysis. Osteoporos Int 27, 13891399. doi:10.1007/s00198-015-3400-y.CrossRefGoogle ScholarPubMed
Kunutsor, SK, Whitehouse, MR, Blom, AW, et al. (2017) Low serum magnesium levels are associated with increased risk of fractures: a long-term prospective cohort study. Eur J Epidemiol 32, 593603. doi:org/10.1007/s10654-017-0242-2.CrossRefGoogle ScholarPubMed
Toth, E, Banefelt, J, Åkesson, K, et al. (2020) History of previous fracture and imminent fracture risk in Swedish women aged 55 to 90 years presenting with a fragility fracture. J Bone Miner Res 35, 861868. doi:10.1002/jbmr.3953.CrossRefGoogle ScholarPubMed
Johansson, H, Siggeirsdottir, K, Harvey, NC, et al. (2017) Imminent risk of fracture after fracture. Osteoporos Int 28, 775780. doi:10.1007/s00198-016-3868-0.CrossRefGoogle ScholarPubMed
Chen, Z, Kooperberg, C, Pettinger, MB, et al. (2004) Validity of self-report for fractures among a multiethnic cohort of postmenopausal women: results from the Women's Health Initiative observational study and clinical trials. Menopause 11, 264274. doi:10.1097/01.gme.0000094210.15096.fd.CrossRefGoogle ScholarPubMed
Figure 0

Fig. 1. Flowchart for the participant selection process.

Figure 1

Table 1. 10-year occurrence rate by participant characteristics and fracture type among postmenopausal women, the SWHS (N = 48 584)

Figure 2

Table 2. Associations of calcium and magnesium intakes with osteoporotic fracture by Ca/Mg ratio and Ca supplement use among postmenopausal women

Figure 3

Table 3. Associations of soy isoflavones intake with osteoporotic fracture by any bone fracture history among postmenopausal women

Figure 4

Table 4. Associations of soy isoflavones intake with osteoporotic fractures by year(s) since menopause among postmenopausal women