Introduction

Osteoporosis and cardiovascular disease (CVD) are two important causes of morbidity and mortality in the elderly. They present great public health challenges worldwide. Rather than being two unrelated diseases exclusively related to age, increasing evidence suggests that these two chronic conditions are associated with each other [1–3], sharing risk factors (e.g., smoking, physical activity, and vitamin D deficiency) [1] and several common pathophysiological mechanisms (e.g., oxidative stress, inflammation, and dyslipidemia) [2, 4]. Epidemiological studies have shown that subjects exposed to either of these two diseases are more likely to suffer a higher risk of the other disorder [5–9]. Evidence has also shown that several approaches, such as the Mediterranean diet, could help prevent both chronic diseases [10–12], and strategies to reduce CVD risk may also benefit bone health.

The 2006, the American Heart Association Diet and Lifestyle Recommendations (AHA-DLR) was released by the AHA to reduce CVD risk in the general population [13]. A total of nine main components make up the AHA-DLR, including balancing caloric intake and physical activity for a healthy body weight; consuming a high-fiber diet rich in vegetables, fruit, whole grain, and fish (especially oily fish); limiting the intake of saturated fat, trans fat, cholesterol, salt, and added sugar; consuming alcohol moderately; and following these principles when eating outside the home [13]. Bhupathiraju et al. developed an original AHA Diet and Lifestyle Score (AHA-DLS) based on the principles of the AHA-DLR [14], with higher scores representing better adherence to the AHA-DLR. They found that higher scores were associated with a better dyslipidemia condition [14], a higher bone mineral density (BMD), and a lower osteoporosis risk (all p values <0.05) in older Puerto Ricans [15]. However, these results were based on a relatively small sample (fewer than 1000) of a particular Hispanic population with health disparities and high-risk factors for both CVD and osteoporosis [16, 17]. More studies of other populations are needed to better understand the problem, but no other studies on this field have been found.

The purpose of this cross-sectional study is to investigate the association between the AHA-DLR and BMD at the whole body, lumbar spine, and hip sites in middle-aged and elderly Chinese.

Methods

Study populations

This study is based on the Guangzhou Nutrition and Health Study (GNHS), a community-based prospective cohort study designed to investigate the nutritional determinants of cardiometabolic outcomes and osteoporosis. The study was first conducted between July 2008 and July 2010 (baseline) in urban Guangzhou. Detailed information can be found in previous articles [18]. Briefly, a total of 3169 subjects aged 40–75 years completed the baseline survey, and 2520 of them were retained and completed the same survey and BMD measurements during a 3-year follow-up. At follow-up, 832 subjects were newly recruited and completed the corresponding survey and measurements. After eliminating 209 subjects for the following reasons: (i) a history of serious chronic disease, such as hyperthyroidism or malignancy; (ii) use of related medication for osteoporosis; (iii) core related data missing; and (iv) extreme energy intake (<600 or >3500 kcal/day for women and <800 or >4200 kcal/day for men), a total of 3143 subjects (including 2371 follow-up subjects and 772 newly recruited subjects) provided complete BMD data and surveys and were included in this cross-sectional analysis. Informed consent was obtained from all of the subjects. The Ethics Committee of the School of Public Health at Sun Yat-sen University approved the study.

Measurements and data collection

During both the baseline and the follow-up, we invited subjects to the School of Public Health at Sun Yat-sen University for the relevant measurements and face-to-face interviews. A structured questionnaire survey was conducted to collect information related to demographics (e.g., age, gender, marital status, household income, education status), lifestyle factors (e.g., smoking, drinking, physical activity), history of disease and medications, and habitual dietary intake and use of calcium supplements. Physical activity was measured and translated into MET h/day as described previously [19]. Height and weight were measured with the subjects in light clothes and no shoes, and body mass index (BMI, using kg/m2) was then calculated.

Assessment of dietary intake

Dietary information was collected using a prevalidated 79-item food frequency questionnaire (FFQ) [20]. The subjects were asked to report the approximate frequencies and portion sizes of the foods they consumed during the preceding year based on provided pictures. The average daily intake of total energy and specific nutrients was then calculated according to the China Food Composition Table 2002 [21]. Trans fatty acids were calculated based on the research data available for the Chinese diet in China [22], and added sugars were calculated from major sources of sugars in the food categories of sugars, cakes, and sugar-sweetened beverages. The average values were calculated for the 2342 subjects who provided both baseline and follow-up dietary data, and only one kind of dietary data was provided and used in the analysis of the 801 new subjects.

American Heart Association Diet and Lifestyle Scores

The 2006 AHA-DLR were released by the AHA to reduce the CVD risk in Americans over 2 years of age. The AHA-DLS was calculated as in a previous article [14] except for the following. We assigned sodium intake a score of 0–10 according to the participants’ urinary sodium/creatinine (Na/Cr, mmol/mmol) levels because of a lack of dietary sodium data. Scores ranging from a minimum of 0 to maximums of 4, 6, and 10 were assigned to subjects according to their adherence to each of the subcomponents. Because BMD is strongly affected by weight, the modified AHA-DLS excluded the BMI components from the original scales, and we further adjusted it in the statistical analyses. The highest possible score for the modified AHA-DLS was 100, where a higher score represented better adherence to the AHA-DLR.

Assessment of bone mineral density

During the follow-up (2011–2013), the subjects’ BMD (g/cm2) at the whole body (WB), lumbar spine (LS), total hip (TH), femur neck (FN), and trochanter (TR) was measured using dual-energy X-ray absorptiometry (DXA) (Discovery W, Hologic Inc., Waltham, MA, USA) and analyzed with Hologic Discovery software version 3.2. The in vivo coefficients of variation of the duplicated BMD measurements in 30 subjects after repositioning were 1.18% (WB), 0.87% (LS), 1.02% (TH), 1.92% (FN), and 1.82% (TR). The long-term CV of the measurements was 0.26%, a value found by testing the phantom daily between March 2011 and May 2015.

Statistical analysis

Common characteristics were presented as means and standard deviations (SDs) for the continuous variables and as frequencies and percentages for the categorical variables. All of the analyses were performed with SPSS 17.0 for Windows (SPSS, Inc., Chicago, USA).

Subjects with a higher total energy intake tended to have a higher intake of nutrients, so we adjusted the modified AHA-DLS for total energy intake using the residual method [23]. We assessed the associations between the modified AHA-DLS—both as a continuous (five-unit increases) measure and as a categorical (sex-specific energy-adjusted quartiles) measure—and BMD by using general linear regression and analysis of covariance (ANCOVA). Two covariance models were used in all analyses with model I adjusted for age and sex and model II further adjusted for BMI, marital status, education, household income, smoking status, calcium supplement use, and daily total energy intake and dietary calcium intake (calcium from AHA-DLR components was excluded). Stratified analyses were performed according to gender, and years since menopause and use of estrogen were added as factors for females only. Bonferroni tests were conducted to make multiple comparisons between quartiles.

We also assessed the associations between each of the modified AHA-DLS subcomponents (one-unit increases) in the linear regression. Analyses were adjusted for the covariates in model II.

Results

Our study included 2092 women and 1051 men (Table 1). The mean (SD) age was 59.7 (5.5) years for women (95.4% of whom were postmenopausal) and 62.4 (6.5) years for men. Subjects with higher modified energy-adjusted AHA-DLS scores tended to be younger, to have a lower BMI and higher dietary calcium intake, were more likely to use calcium supplements, and smoked less (all p trends <0.01).

Table 1 Characteristics of study participants by quartiles of energy-adjusted AHA-DLR score
Full size table

Higher modified AHA-DLS scores were positively and dose-dependently associated with 1.84–2.61% higher BMD (quartile 4 vs. 1, all p values <0.01) at all of the bone sites after adjusting for age and gender in the total subjects (Table 2). The associations were strengthened when other variables (e.g., BMI, education, smoking) were further adjusted for in model II (all p trends <0.001). The BMD values were 1.93–3.11% higher in the top (vs. bottom) quartiles of the modified AHA-DLS (all p values <0.001). Similar associations were found in both women and men (p interactions 0.118–0.395) (Table 3).

Table 2 Comparisons of covariate-adjusted mean of bone mineral density (g/cm2) by quartiles of energy-adjusted AHA-DLR scores
Full size table
Table 3 Comparisons of covariate-adjusted mean of bone mineral density by quartiles of energy-adjusted AHA-DLR scores stratified by gender
Full size table

In the linear regression models, after adjusting for potential covariates, every five-unit increase in the AHA-DLS score was associated with 4.20–6.07, 4.44–8.51, and 3.36–4.67 mg/cm2 higher BMD at different bone sites for the total subjects, men, and women, respectively (all p values <0.05) (Table 4). For the subcomponent analyses, higher scores for physical activity, fruit and vegetables, whole grains, fish, and urinary Na/Cr levels and lower scores for dietary saturated fat and cholesterol were associated with a higher BMD at several bone sites after adjusting for the potential covariates (all p values <0.05) (Table 5).

Table 4 Associations between energy-adjusted modified AHA-DLS and bone mineral density (mg/cm2)
Full size table
Table 5 Associations between subcomponents of the modified AHA-DLS and bone mineral density (mg/cm2)
Full size table

Discussion

In this community-based cross-sectional study, we observed a favorable association between higher modified AHA-DLS scores and BMD at multiple sites in middle-aged and elderly Chinese adults. Previous studies suggested that one SD decrease in the total hip BMD was associated with an 85% (95% CI, 70–101%) increase in the risk of total osteoporotic fractures [24]. Therefore, a 0.021 g/cm2 (or 0.18 and 0.19 SDs for men and women) higher BMD at the total hip in quartile 4 (vs. 1) of AHA-DLS score in our study would be associated with a 15.4–15.9% decrease in osteoporotic fractures in future. Our findings suggest that the AHA-DLR, which was first introduced to reduce CVD risk, could also benefit bone health in this population.

An increasing number of epidemiological studies support a close biological relationship between osteoporosis and CVD [5–7], although whether the strategies conducted specifically for CVD prevention could also benefit bone health has received less attention. To our knowledge, there is only one direct study in this field whose results are consistent with ours. A five-unit increase in the modified AHA-DLS score was associated with a 0.005–0.008 g/cm2 higher BMD and lower risk (OR, 0.83–0.96, all p values <0.05) for osteoporosis/osteopenia at both the femur and lumbar spine sites in 933 Puerto Ricans aged 47–79 years [15]. Nevertheless, our results are consistent with those of several other studies that focused on healthy dietary patterns with similar principles. For instance, the Mediterranean diet (which is characterized by a higher intake of fiber, fruits and vegetables, and fish; moderate drinking; and limited saturated fat intake), which first received attention for CVD protection, was also associated with lower hip fracture incidence in 48,814 men and 139,981 women at a follow-up in the EPIC study 9 years later [11] and was also associated with a higher calcareous bone BMD (p trend = 0.001) in a cross-sectional study of 200 Spanish women [25]. In addition, increasing evidence suggests that the Dietary Approaches to Stop Hypertension (DASH) diet (which is characterized by a higher intake of high-fiber grains, vegetables, and fruits and restrained intake of fat and cholesterol) might also benefit bone health by blocking the bone resorption process [26, 27]. These results highlight the possibility of addressing two major public health problems with a single set of guidelines.

In the subcomponent analysis, after adjusting for potential covariates, a higher level of physical activity was associated with a higher BMD at all bone sites (all p values <0.001). This favorable association is consistent with the results of previous studies. More vigorous physical activity was associated with a higher femur neck BMD (β = 0.09, p < 0.001) in a cross-sectional study of 1228 70-year-old Swedish subjects [28] and a 39% lower risk of hip fractures (95% CI, 0.54–0.69) in a meta-analysis of 1,235,768 subjects [29]. The positive association between a higher intake of fruits and vegetables, whole grains, fish, and sodium restriction and BMD found in our study is consistent with previous studies [30–33], which reinforces the role of nutrition in protecting bone health. However, higher scores (a lower intake) for saturated fat and cholesterol were associated with a lower BMD at several hip sites. These results are in conflict with several former studies that found detrimental associations with bone health [34, 35]. More studies are needed to clarify the issue. Nevertheless, a positive association with BMD was found at all bone sites when dietary subcomponents were combined. Our results reinforce the importance of combining positive exercise and diet to better protect bone health.

Previous evidence suggests that several important biological mechanisms of CVD [36], including oxidative stress, inflammation (e.g., C-reaction protein (CRP) and interleukin-6 (IL-6)), and dyslipidemia, also play important roles in the development of osteoporosis by inhibiting the osteoblastic bone formation process and promoting osteoclast bone resorption [37–40]. Although few direct data were found, the favorable association of AHA-DLR with BMD might stem from several components that mitigate these detrimental factors. Increased physical activity was inversely associated with a lower risk of elevated CRP in 14,461 American adults [41] and positively associated with an increase in total antioxidant capacity (TAC) in 3042 Greek adults [42]. The greater Mediterranean diet score, which shares several components with AHA-DLR, was associated with lower levels of blood CRP and IL-6 in a meta-analysis of randomized clinical trials (RCT) [43] and with greater TAC in Greek adults [42]. Better adherence to the DASH diet and sodium reduction improved the serum profile of CRP and glutathione in an RCT [44] and decreased bone turnover markers, such as serum osteocalcin and the C-terminal telopeptide of type I collagen, in another RCT of 186 adults [26]. These results suggest an association between the AHA-DLR and these mechanisms, although further interventional studies are needed to identify the causal relationship.

The strength of this study is that we examined the association between the AHA-DLR and BMD at multiple sites using a large sample size. In addition, the dietary data averages for the majority of our population were used for analysis, providing a better estimation of habitual dietary intake. However, our study has several limitations. First, the cross-sectional design cannot identify a causal relationship, although we used the average values of dietary intake to better estimate habitual consumption over the period before the BMD assessment, attenuating the possibility of causal inversion in the majority of the subjects. Second, dietary calcium was adjusted for in the analysis for a more prudent consideration, which might have resulted in an excessive correction and underestimation of the association, although we excluded calcium from the AHA-DLR components in the analysis. Nevertheless, positive results were still observed at most bone sites. Third, blood vitamin D status of the participants was not available for the adjustments in the analyses of this study, which might increase the possibility of residual confounding. Finally, the subjects, who were recruited as volunteers, might have led healthier lifestyles or engaged in more healthy activities, which might attenuate the positive association. However, health-related factors, such as education and economic status, smoking, and the use of calcium supplements, did not significantly modify the AHA-DLR–BMD association (p interaction range 0.069–0.993).

In conclusion, we observed that better adherence to the AHA-DLR (as indicated by a higher modified AHA-DLS score) was favorably associated with BMD at multiple sites in middle-aged and elderly Chinese. Our results support the hypothesis that strategies to reduce the CVD risk might also benefit bone health and that a single set of guidelines might thus contribute to the prevention of these two major public problems.