Abstract
Purpose:
It is unclear if midlife consumption of dairy and soy food intake, and their components of calcium and isoflavones (in soy), is related to cognitive impairment in elderly.
Methods:
We used baseline data on lifestyle and habitual diet of 16,948 participants collected during their recruitment into the Singapore Chinese Health Study from1993 to 1998, and data on their cognitive function, measured using a 30-item Singapore modified Mini-Mental State Examination, during follow-up interviews from 2014 to 2016. We used multivariable logistic regression models to estimate the odds ratios (ORs) and 95% confidence intervals (CIs) for developing cognitive impairment.
Results:
Higher dairy intake was associated with a lower risk of cognitive impairment in a dose-dependent manner (P for trend= 0.009). Compared to the lowest quartile of dairy intake, ORs (95% CIs) were 0.93 (0.81-1.07) for the second, 0.88 (0.76-1.01) for the third, and 0.82 (0.72-0.94) for the fourth quartiles of intake. Similar results were found for dairy calcium intake (P for trend= 0.008). However, there was no statistically significant association for intake of soy (OR comparing extreme quartiles 0.99, 95% CI 0.87-1.14, P for trend= 0.92), isoflavones (OR 1.01, 95% CI 0.88-1.15, P for trend= 0.90) or non-dairy calcium (OR 1.06, 95% CI 0.86-1.30, P for trend= 0.81) with risk of cognitive impairment.
Conclusions:
Dairy intake at midlife could have a protective association against cognitive impairment that may not be attributed to its calcium content alone, while soy or isoflavone intake was not associated with cognition of elderly in our study.
Keywords: Dairy, milk, soy, isoflavones, calcium, cognitive impairment, Chinese
Introduction
Given the lack of an effective cure for cognitive impairment, its rising prevalence has made prevention an urgent public health challenge [1]. Consistent findings from observational studies have shown associations between nutritional factors and cognition in the elderly [2]. Milk and dairy products have been reported to be associated with a lower risk of hypertension, diabetes, and stroke, which are known predisposing factors for cognitive impairment [3]. The link between milk intake and cognitive disorders has been assessed in some epidemiological studies, but the results are inconsistent [3,4].
Investigators have proposed that calcium may be a potential mediating factor for the inverse association between dairy intake and cardiometabolic risk in observational studies [5-8]. Since cardiovascular diseases and cognitive impairment share common risk factors [9], calcium may also be postulated to have an impact on cognitive function through shared mechanisms with cardiovascular diseases. The “calcium hypothesis” ascribes many of the neural dysfunctions underlying chronic brain disorders of aging to mechanisms that regulate cellular calcium homeostasis [10]. However, it remains unclear if dietary calcium plays a role because circulating calcium is tightly regulated [11]. The observational evidence is scarce, conflicting and admixed with inverse [12] and null [13] associations. Moreover, it is difficult, if not impossible, to conduct a study to differentiate the effect of dietary calcium from that of dairy food in populations with high dairy intake, since the latter constitutes the main source of dietary calcium [14,15]. Such a study is practically feasible only in populations where dietary calcium comes from a variety of non-dairy foods [16].
The historically lower prevalence of dementia in East Asia (4.2%) than in Western countries (6.9% in Europe and 6.5% in the USA) [17] has directed attention to regional dietary differences, and one such difference is the higher intake of soy food in Asians [18]. The most bioactive compounds of soy food are isoflavones. Although animal and cell studies have elucidated potential neuroprotective benefits of isoflavones, studies in human are largely inconsistent, with approximately half in support of such benefits while another half suggesting null or even adverse effect [19,20]. Hence, there is a huge knowledge gap in this area that needs to be filled by findings from well-designed large-scale studies in Asian populations with relatively higher intake of soy food than Western populations [19].
In this study, we aimed to study the association of midlife intake of dairy and soy foods, as well as their components of calcium (in both foods) and isoflavones (in soy food), with the risk of cognitive impairment at late life among Chinese in Singapore, a population that has generally low intake of dairy but high intake of soy, and diverse sources of dietary calcium. While dairy contributed to 17.3% of dietary calcium, soy foods contributed to 11.8% of dietary calcium in this population, with the remaining diet calcium coming from a variety of foods such as vegetables and grain products [16].
Methods
Study population and design
We used data from the Singapore Chinese Health Study, a population-based cohort study that recruited 63,257 Chinese men and women aged 45 to 74 years from April 1993 to December 1998. This recruitment was restricted to permanent residents or citizens of Singapore living in government-built housing estates, where 86% of Singapore residents resided during the time of recruitment. The participants were from the two major dialect groups in Singapore, the Hokkiens and Cantonese, who mostly originated from the Fujian and Guangdong provinces in Southern China, respectively. At recruitment, participants were interviewed in-person and information on demographic characteristics, lifestyle factors (physical activity, tobacco use, and alcohol intake), habitual diet, and medical history was collected through structured questionnaires. After the baseline interview, participants were re-contacted for follow-up 1 (1999 to 2004), follow-up 2 (2006 to 2010), and follow-up 3 (2014 to 2016) interviews. A total of 17,107 surviving participants aged 61-96 years participated in the follow-up 3 interviews, which included testing of cognitive function (Supplemental Figure 1). The Institutional Review Board at the National University of Singapore approved the study and informed consent was obtained from each study participant. Further details of study design have been published previously [21].
Assessment of diet and covariates
A semi-quantitative food-frequency questionnaire (FFQ) including 165 commonly consumed food items in this population was administered during the baseline interview. The respondents were required to choose the portion size (small, medium, large) from provided photographs and select from 8 categories of frequency (ranging from “never or hardly ever” to “two or more times a day”). The FFQ was subsequently validated using two 24-hour dietary recall interviews and a repeat administration of it among a subset of 810 cohort participants. Similar intake distributions were observed for these two methods, with most mean pairs for energy and nutrients within 10% of each other's values. Moreover, the correlation coefficient for energy and nutrients ranged from 0.24 to 0.79 [21], which is comparable with previous validation studies in diverse populations [22]. Specifically, correlation coefficients for calcium intake between FFQ and 24-hour recalls ranged from 0.51 to 0.62 [21]. In another validation study, dietary intake of individual soy food item was monotonously associated with urinary excretions of soy isoflavones, namely genistein, daidzein and glycitein [23]. Moreover, a statistically significant dose-dependent positive correlation was observed between soy intake and urine isoflavones measured in specimens collected on average 6.5 years after the baseline interview [24].
Participants were asked about the consumption of six commonly consumed dairy products in this population, comprising of: 1) milk including powdered, whole, low fat, and chocolate but excluding addition to coffee or tea; 2) Milo, Ovaltine, or Horlicks (different brand names of malted milk powders); 3) Yakult and Vitagen (probiotic cultured milk product made by fermentation); 4) added milk to coffee or tea; 5) butter used as bread spread; and 6) ice cream and frozen yogurt. The small amounts of dairy products used in cooking procedures of local dishes were also taken into account to enhance the accuracy of total dairy intake. Soy intake was assessed from the intake of seven items that are commonly consumed in the Chinese population: plain tofu, tau pok, tau kwa, foo pei, foo jook, tofu far, and soybean drink. As described previously, we calculated equivalent amounts of tofu per day to facilitate comparison with a known dietary item while taking into account the varying water contents in the different soy food items [25]. We further assessed nutrient intakes including dairy and non-dairy calcium, as well as soy protein and isoflavones, from the FFQ data using the Singapore Food Composition Database that was specifically developed for this cohort study [21].
Self-reported information about age, weight, height, educational level, smoking status, and physical activity was also collected at baseline. Body mass index (BMI, in kg/m2) was calculated by body weight in kg divided by square of height in meter. Information on medical history of type 2 diabetes, hypertension, coronary artery disease, and stroke were collected by asking if participants have been diagnosed by a doctor to have these conditions. Cancer history was obtained either from self-report or through data linkage with the nationwide Singapore Cancer Registry database.
Assessment of cognitive function
We assessed cognitive function of participants face-to-face using Singapore Modified Mini-Mental State Examination (SM-MMSE) [26]. The SM-MMSE test includes orientation, attention, immediate and delayed recall, language, pentagon copying, and the ability to follow verbal and written instructions. The test could be administered in the Chinese dialects of Hokkien or Cantonese, or in Chinese Mandarin, according to the preference of the participants. The total SM-MMSE score ranges from 0 to 30, with higher score representing better global cognition [27]. After being systematically trained by a psychiatric and geriatric epidemiologist who had expertise and extensive experience in cognitive assessment, the interviewers had to pass the assessment to perform the SM-MMSE test. All the interviews were recorded and 20% were randomly selected for quality control. The interviewers were re-trained and re-assessed if they did not follow the protocol.
The traditional cut-off to select patients with suspected cognitive impairment is 23/24 [28]. Previous studies [29,30] have shown that education levels significantly affected the MMSE score. Low educated participants might be classified as cognitively impaired due to low literacy skills, whereas higher education could counterbalance subtle impairment. Thus, we defined the cut-off points for cognitive impairment according to education levels using the cutoff points from the Shanghai Dementia Survey [30], which had participants with comparable education levels to our participants: less than 18 for those without schooling, less than 21 for those with 1-6 years of education, and less than 25 for those with more than 6 years education.
Statistical Analysis
We performed the final analysis on 16,948 participants (6914 men and 10,034 women) after excluding those with missing value on cognitive tests (n = 55), and participants who were mute (n = 1), blind (n = 55), or deaf (n = 48) at the follow-up 3. Energy-adjusted values were used for all foods and nutrients using the residual method [31]. We employed logistic regression models to estimate odds ratios (OR) and 95% confidence intervals (CI) for associations of dairy, soy, and calcium intake quartiles and categories of milk intake frequency with cognitive impairment.
In multivariable models, we first adjusted for potential confounders as suggested in the literature [32]. The first model included age (years), sex, interview year at baseline (1993-1995, 1996-1998), dialect group (Hokkien, Cantonese), level of education (none, primary school, secondary school, A level and more), and marriage status (married, widowed, separated, never married); and the second model further included moderate physical activity level (<0.5, 0.5-3.9, ≥4 h/week), BMI (kg/m2), cigarette smoking (never, past, current smoker), alcohol consumption (never/monthly, weekly, daily), self-reported history of hypertension, diabetes, coronary artery disease, and stroke, cancer history, sleep duration (≤5 h, 6-8 h, ≥9 h) and total energy intake (kcal/day). In the third model we further adjusted for intakes of red meat, poultry, fish, vegetables, and fruits, as well as tea (≤monthly, weekly, daily), and coffee (<daily, 1 cups/day, ≥2 cups/day), and soda (glasses/week) to control for confounding role of other dietary items [2,33]. In the fourth model, we adjusted for the scores of “vegetable, fruit, and soy-rich” dietary pattern (quartiles) as a measure of overall diet quality instead of the individual food items. This pattern was previously identified in this population and found to be inversely associated with diabetes [34], mortality [35], and hip fracture [36]. Finally, we added quartile intakes of potassium, magnesium, phosphorus, vitamins D, E, and folate to the second model to test if the associations were mediated (for milk, dairy, and soy) or confounded (for calcium) by these nutrients that were both correlated with our dietary exposures of interest, and observed to be related with cognitive impairment in previous studies [2,13,12].
We tested linear trends across quartiles of intakes by assigning the median intake to each of the four categories and using this as a continuous variable in regression models. Interactions were tested through cross-product terms between dietary factors as quartiles of intake (median of each category) and gender, age at baseline (<median age of 52 years vs ≥52 years), BMI (<23 and ≥ 23 kg/m2), daily vitamin D intake (<median intake of 59.8 IU vs ≥59.8 IU), and a composite co-morbidity variable for the self-report of any medical history of diabetes, hypertension, coronary artery disease, and stroke. All statistical analyses were conducted using Stata Statistical Software, Release 14.2 (Stata Corporation, College Station, TX), with 2-sided P value less than 0.05 as the threshold for statistical significance.
Results
The median SM-MMSE score was 26 (interquartile range = 23-28) after a mean 20.2 (SD: 1.9) years of follow-up since baseline interview (ranged 15.8-23.8 years). There were 5,333 (31.5%) individuals with SM-MMSE score below 24. Using education-specific cut-off points, 2,443 (14.4%) participants who had low SM-MMSE scores were defined as having cognitive impairment.
Dairy and milk
Median value (interquartile range) of total dairy intake was 28.7 (11.0-83.7) g/day in all participants included in the present analysis, with milk as a major component (R2, coefficient of determination between dairy and milk intake=0.79). About 15.9% reported to drinking milk daily, while 64.8% hardly drank any milk (less than once a month). Almost all participants who drank milk daily were categorized into the highest quartile of dairy intake. Women, as well as those with higher education or higher level of physical activity, were more likely to consume higher dairy products than their respective counterparts, whereas smokers or alcohol drinkers consumed less amount of dairy (Table 1). Participants with higher dairy intake had a lower intake of coffee and soda, also ate less red meat, but had a higher intake of fruits and several nutrients, including potassium, magnesium, phosphorous, vitamin D, folate, and vitamin E.
Table 1:
Participant characteristics according to quartiles of dairy intake in the Singapore Chinese Health Study
| Quartiles of dairy intake |
||||
|---|---|---|---|---|
| Q1 | Q2 | Q3 | Q4 | |
| Median intake, g/d | 5.68 | 9.41 | 36.6 | 252 |
| n | 4,423 | 4,059 | 4,046 | 4,420 |
| Age at baseline, y | 52.4 ± 5.9 | 53.3 ± 6.3 | 53.2 ± 6.4 | 53.1 ± 6.3 |
| Female sex, n (%) | 1,725 (39.0) | 2,576 (63.5) | 2,701 (66.8) | 3,032 (68.6) |
| Dialect group, n (%) | ||||
| Hokkien | 2,240 (50.6) | 1,933 (47.6) | 1,930 (47.7) | 2,342 (53.0) |
| Cantonese | 2,183 (49.4) | 2,126 (52.4) | 2,116 (52.3) | 2,078 (47.0) |
| Education, n (%) | ||||
| No | 652 (14.7) | 914 (22.5) | 902 (22.3) | 725 (16.4) |
| Primary | 2,131 (48.2) | 1,789 (44.1) | 1,790 (44.2) | 1,877 (42.5) |
| Secondary | 1,640 (37.1) | 1,356 (33.4) | 1,354 (33.5) | 1,818 (41.1) |
| Married, n (%) | 4,034 (91.2) | 3,591 (88.5) | 3,539 (87.5) | 3,852 (87.2) |
| Ever smoker, n (%) | 1,465 (33.1) | 834 (20.6) | 800 (19.8) | 663 (15.0) |
| Weekly/daily alcohol drinker, n (%) | 807 (18.3) | 376 (9.3) | 335 (8.3) | 428 (9.7) |
| Weekly moderate activity, n (%) | ||||
| <0.5 hours/wk | 3,482 (78.7) | 3,220 (79.3) | 3,212 (79.4) | 3,157 (71.4) |
| 0.5-3 hours/wk | 612 (13.8) | 538 (13.2) | 567 (14.0) | 820 (18.6) |
| ≥4 hours/wk | 329 (7.4) | 301 (7.4) | 267 (6.6) | 443 (10.0) |
| Sleep, n (%) | ||||
| ≤ 5 hours/d | 347 (7.9) | 324 (8.0) | 346 (8.6) | 373 (8.4) |
| 6-8 hours/d | 3,819 (86.3) | 3,513 (86.6) | 3,478 (86.0) | 3,822 (86.5) |
| ≥ 9 hours/d | 257 (5.8) | 222 (5.5) | 222 (5.5) | 225 (5.1) |
| Hypertension, n (%) | 827 (18.7) | 795 (19.6) | 828 (20.5) | 835 (18.9) |
| Diabetes, n (%) | 199 (4.5) | 189 (4.7) | 170 (4.2) | 270 (6.1) |
| Coronary artery disease, n (%) | 83 (1.9) | 90 (2.2) | 86 (2.1) | 94 (2.1) |
| Stroke, n (%) | 17 (0.4) | 17 (0.4) | 21 (0.5) | 27 (0.6) |
| Cancer, n (%) | 57 (1.29) | 67 (1.65) | 88 (2.17) | 117 (2.65) |
| Daily tea drinker, n (%) | 1,160 (26.2) | 849 (20.9) | 936 (23.1) | 1,031 (23.3) |
| Coffee drinker, n (%) | ||||
| Less than daily | 1,061 (24.0) | 1,211 (29.8) | 1,243 (30.7) | 1,650 (37.3) |
| 1 cups/d | 1,517 (34.3) | 1,610 (39.7) | 1,415 (35.0) | 1,617 (36.6) |
| ≥2 cups/d | 1,845 (41.7) | 1,238 (30.5) | 1,388 (34.3) | 1,153 (26.1) |
| Soda > 2 times/w, n (%) | 669 (15.1) | 420 (10.4) | 359 (8.9) | 353 (8.0) |
| Body mass index, kg/m2 | 23.3 ± 3.3 | 23.3 ± 3.2 | 23.1 ± 3.3 | 22.9 ± 3.1 |
| Total energy intake, kcal/d | 1970 ± 525 | 1366 ± 380 | 1332 ± 484 | 1686 ± 549 |
| Dairy, g/d | −5.13 ± 15.8 | 20.3 ± 4.8 | 43.5 ± 11.6 | 227 ± 117 |
| Red meat, g/d | 31.2 ± 23.4 | 31.5 ± 15.5 | 31.2 ± 14.9 | 26.3 ± 18.9 |
| Poultry, g/d | 22.0 ± 20.9 | 22.4 ± 14.3 | 21.7 ± 13.5 | 19.1 ± 16.9 |
| Fish and seafood, g/d | 57.9 ± 31.8 | 56.3 ± 24.0 | 54.5 ± 22.7 | 54.2 ± 27.3 |
| Total vegetables, g/d | 116 ± 66.4 | 117 ± 48.0 | 115 ± 47.5 | 121 ± 59.3 |
| Total fruit, g/d | 201 ± 182 | 215 ± 130 | 216 ± 131 | 236 ± 161 |
| Total calcium intake, mg/d | 303 ± 130 | 358 ± 113 | 406 ± 112 | 630 ± 203 |
| Non-dairy calcium intake, mg/d | 292 ± 108 | 297 ± 72.4 | 293 ± 69.7 | 297 ± 89.1 |
| Dairy calcium intake, mg/d | 4.31 ± 32.5 | 49.3 ± 23.2 | 102 ± 38.6 | 310 ± 146 |
| Potassium, mg/d | 1665 ± 490 | 1755 ± 333 | 1822 ± 329 | 2074 ± 432 |
| Magnesium, mg/d | 235 ± 40.0 | 241 ± 29.7 | 245 ± 29.4 | 264 ± 36.7 |
| Phosphorus, mg/d | 767 ± 131 | 803 ± 92.3 | 829 ± 88.6 | 977 ± 137 |
| Vitamin D, IU/d | 71.0 ± 50.9 | 83.4 ± 36.7 | 97.9 ± 35.9 | 166 ± 61.5 |
| Folate, mcg/d | 152 ± 58.3 | 158 ± 41.7 | 160 ± 42.5 | 172 ± 52.2 |
| Vitamin E, mg/d* | 5.89 (5.67-6.11) | 7.72 (7.48-7.98) | 7.84 (7.59-8.10) | 8.35 (8.01-8.70) |
| MMSE total score | 24.2 ± 4.1 | 24.7 ± 4.0 | 25.2 ± 3.7 | 25.4 ± 3.6 |
The data are expressed as n (%) or mean ± standard deviation.
Geometric mean (95% CI)
All the observed differences were statistically significant at P <0.001 except for sleep (P = 0.67), hypertension (P = 0.16), and coronary artery disease (P = 0.72).
Dairy intake was inversely associated with risk of cognitive impairment in a dose-response manner (P for trend=0.008) (Table 2), and the association was slightly attenuated after adjustment for potential confounders that included other dietary items; OR (95% CI) comparing highest vs. lowest quartiles was 0.82 (0.72-0.94). Further adjustment for dietary pattern using the “vegetable, fruit, and soy-rich” scores in quartiles did not materially change our findings. We did not observe a mediation role after further adjustment for aforementioned minerals and vitamins (OR comparing extreme quartiles 0.80, 95% CI 0.68-0.95, P for trend=0.02). Compared to non- milk-drinkers, those who drank milk daily had a 14% lower risk of cognitive impairment in a multivariable model (OR 0.86, 95% CI 0.76–0.98, P for trend=0.03).
Table 2:
Odds ratio (95% Confidence Interval) of cognitive impairment according to intakes of dairy products and frequency of milk intake
| Quartiles of dairy intake |
P for trend* |
||||
|---|---|---|---|---|---|
| Q1 | Q2 | Q3 | Q4 | ||
| Median intake, g/d | 5.68 | 9.41 | 36.6 | 252 | |
| Cases/non-cases | 644/3,779 | 612/3,447 | 576/3,470 | 611/3,809 | |
| Model 1† | 1.00 | 0.89 (0.78-1.01) | 0.84 (0.74-0.96) | 0.79 (0.70-0.90) | 0.004 |
| Model 2‡ | 1.00 | 0.92 (0.80-1.05) | 0.87 (0.75-1.00) | 0.82 (0.72–0.93) | 0.009 |
| Model 3§ | 1.00 | 0.93 (0.81-1.07) | 0.88 (0.76-1.01) | 0.82 (0.72-0.94) | 0.009 |
| Model 4** | 1.00 | 0.92 (0.80-1.05) | 0.87 (0.75-1.00) | 0.82 (0.72-0.94) | 0.01 |
| Milk intake frequency |
P for trend |
||||
| None | ≤ once a week | 2-6 times/week | Daily | ||
| n(%) | 10,986 (64.8) | 1,602 (9.5) | 1,660 (9.8) | 2,700 (15.9) | |
| Cases/non-cases | 1,630/9,356 | 206/1,396 | 215/1,445 | 392/2,308 | |
| Model 1† | 1.00 | 0.99 (0.84-1.17) | 0.92 (0.78-1.07) | 0.87 (0.77-0.99) | 0.03 |
| Model 2‡ | 1.00 | 1.00 (0.85-1.18) | 0.92 (0.78-1.08) | 0.87 (0.77-0.99) | 0.03 |
| Model 3§ | 1.00 | 1.00 (0.85-1.18) | 0.93 (0.79-1.10) | 0.86 (0.75-0.98) | 0.03 |
| Model 4** | 1.00 | 1.00 (0.85-1.18) | 0.92 (0.78-1.08) | 0.88 (0.77-1.00) | 0.04 |
Linear trend was tested by treating the median values of quartiles or frequency of milk intake as a continuous variable
Multivariable model 1: adjusted for age, sex, dialect, year of interview, educational level, and marriage status;
Multivariable model 2: further adjusted for body mass index, physical activity, smoking status, alcohol use, baseline history of self-reported hypertension, diabetes, heart attack, and stroke, history of cancer, sleep status, and total energy intake;
Multivariable model 3: further adjusted for soy, red meat, poultry, fish, vegetables, fruits, tea, coffee, and soda;
Multivariable model 4: model 2 plus vegetable-fruit-soy (VFS) dietary pattern.
We did not find any significant interaction, whether for total dairy food or milk alone, with sex, age, BMI, and vitamin D intake (all Ps for interaction >0.18). When stratified by the baseline history of comorbidities (coronary artery disease, stroke, hypertension, or diabetes), the observed inverse association was only evident in subjects without a history of these diseases (OR 0.78, 95% CI 0.67-0.91, P for trend=0.002), while the association was not significant in those with a prior medical history (OR 0.92, 95% CI 0.71-1.18, P for trend=0.98) (P for interaction=0.04, Supplemental table 1).
Soy foods
Participants with higher soy intake had lower prevalence of smoking, alcohol and coffee drinking, but had higher level of moderate physical activity. While soy intake was directly correlated with fish, vegetable, fruit, and vitamin D intake, it was inversely correlated with magnesium, potassium, phosphorous, vitamin E, and folate intake (Supplemental table 2). We did not observe any significant association between tofu equivalent, soy protein or isoflavones, and odds of cognitive impairment (Table 3).
Table 3:
Odds ratio (95% Confidence Interval) of cognitive impairment according to intakes of soy products
| Quartiles of dairy intake |
P for trend* |
||||
|---|---|---|---|---|---|
| Q1 | Q2 | Q3 | Q4 | ||
| Tofu equivalent | |||||
| Median intake, g/d | 39.4 (18.5-53.0) | 81.1 (73.0-89.3) | 116 (107-128) | 190 (162-241) | |
| Cases/non-cases | 587/3502 | 591/3439 | 640/3624 | 625/3940 | |
| Model 1† | 1.00 | 0.94 (0.82-1.07) | 0.97 (0.85-1.10) | 0.90 (0.79-1.03) | 0.16 |
| Model 2‡ | 1.00 | 0.97 (0.85-1.11) | 0.99 (0.87-1.13) | 0.91 (0.80-1.04) | 0.19 |
| Model 3§ | 1.00 | 1.01 (0.88-1.16) | 1.06 (0.93-1.22) | 0.99 (0.87-1.14) | 0.92 |
| Model 4** | 1.00 | 0.98 (0.86-1.12) | 1.02 (0.89-1.16) | 0.95 (0.83-1.09) | 0.48 |
| Isoflavones | |||||
| Median intake, mg/d | 6.06 (2.58-8.23) | 13.1 (11.6-14.5) | 19.3 (17.5-21.3) | 32.4 (27.2-41.9) | |
| Cases/non-cases | 591/3482 | 599/3437 | 619/3696 | 634/3890 | |
| Model 1† | 1.00 | 0.92 (0.81-1.05) | 0.90 (0.79-1.02) | 0.91 (0.80-1.03) | 0.18 |
| Model 2‡ | 1.00 | 0.95 (0.83-1.08) | 0.93 (0.81-1.06) | 0.92 (0.81-1.05) | 0.24 |
| Model 3§ | 1.00 | 0.99 (0.87-1.14) | 0.99 (0.86-1.13) | 1.01 (0.88-1.15) | 0.90 |
| Model 4** | 1.00 | 0.96 (0.84-1.10) | 0.95 (0.83-1.08) | 0.95 (0.83-1.09) | 0.56 |
| Soy protein | |||||
| Median intake, g/d | 2.15 (1.04-2.85) | 4.33 (3.89-4.75) | 6.17 (5.65-6.72) | 10.0 (8.50- 12.5) | |
| Cases/non-cases | 571/3514 | 603/3437 | 655/3622 | 614/3932 | |
| Model 1† | 1.00 | 0.98 (0.86-1.12) | 1.02 (0.90-1.16) | 0.91 (0.80-1.04) | 0.16 |
| Model 2‡ | 1.00 | 1.01 (0.88-1.15) | 1.06 (0.92-1.21) | 0.92 (0.81-1.05) | 0.18 |
| Model 3§ | 1.00 | 1.06 (0.92-1.21) | 1.13 (0.99-1.30) | 1.00 (0.87-1.15) | 0.87 |
| Model 4** | 1.00 | 1.02 (0.89-1.17) | 1.08 (0.95-1.24) | 0.96 (0.83-1.10) | 0.50 |
Linear trend was tested by treating the median values of quartiles or frequency of milk intake as a continuous variable
Multivariable model 1: adjusted for age, sex, dialect, year of interview, educational level, and marriage status;
Multivariable model 2: further adjusted for body mass index, physical activity, smoking status, alcohol use, baseline history of self-reported hypertension, diabetes, heart attack, and stroke, history of cancer, sleep status, and total energy intake;
Multivariable model 3: further adjusted for dairy, red meat, poultry, fish, vegetables, fruits, tea, coffee, and soda;
Multivariable model 4: model 2 plus vegetable-fruit-soy (VFS) dietary pattern.
No significant interaction was observed between soy intake and sex, age, BMI or vitamin D intake. Isoflavones subtypes (daidzein, genistein, and glycitein) were not significantly associated with cognitive impairment in men and women after adjustment for other nutrients or dietary pattern (Supplemental table 3).
Calcium
The mean intake (±SD) of dietary calcium in this study population was 413 ± 165 mg/d, and the major sources were non-dairy foods, which accounted for a median 80.2% of total calcium intake (295 ± 86.8 mg/d). In contrast, dairy products only contributed 18.7% (118 ± 143 mg/d). The differences in characteristics of participants across quartiles of dietary calcium were similar as those observed across quartiles of dairy intake (Supplemental table 4). We observed an inverse association between total dietary calcium and cognitive impairment that was independent of dietary pattern (Table 4). When we differentiated between dairy and non-dairy sources of calcium, only dairy calcium was associated with lower risk (OR comparing highest vs. lowest quartiles 0.86, 95% CI 0.76-0.98, P for trend=0.01). In contrast, calcium from non-dairy sources was not significantly associated with cognitive impairment (OR 1.06, 95% CI 0.86-1.30, P for trend=0.81). Dairy calcium was highly correlated with dairy intake (r = 0.73). When dairy intake was added as a covariate to the model, dairy calcium was no longer associated with cognitive impairment (OR 1.04 95% CI 0.78-1.40, P for trend=0.88). Further adjustment for other nutrients did not materially change these results (data not shown).
Table 4:
Odds ratio (95% Confidence Interval) of cognitive impairment according to intakes of calcium
| Quartiles of calcium intake |
P for Trend* |
||||
|---|---|---|---|---|---|
| Q1 | Q2 | Q3 | Q4 | ||
| Dietary calcium | |||||
| Median intake, mg/d | 257 | 339 | 412 | 598 | |
| Cases/non-cases | 609/3,427 | 593/3,352 | 602/3,787 | 639/3,939 | |
| Model 1† | 1.00 | 0.86 (0.76-0.98) | 0.81 (0.71-0.92) | 0.78 (0.69-0.89) | 0.001 |
| Model 2‡ | 1.00 | 0.88 (0.77-1.01) | 0.82 (0.72-0.93) | 0.80 (0.70-0.91) | 0.002 |
| Model 3** | 1.00 | 0.92 (0.80-1.06) | 0.86 (0.75-1.00) | 0.84 (0.72-0.97) | 0.03 |
| Model 4†† | 1.00 | 0.89 (0.78-1.02) | 0.83 (0.72-0.96) | 0.81 (0.71-0.93) | 0.007 |
| Dairy calcium | |||||
| Median intake, mg/d | 7.2 | 45.7 | 101 | 314 | |
| Cases/non-cases | 626/3,717 | 624/3,478 | 588/3,575 | 605/3,735 | |
| Model 1† | 1.00 | 0.97 (0.85-1.10) | 0.90 (0.79-1.02) | 0.84 (0.74-0.95) | 0.005 |
| Model 2‡ | 1.00 | 1.01 (0.88-1.15) | 0.93 (0.82-1.07) | 0.87 (0.76-0.99) | 0.01 |
| Model 3** | 1.00 | 1.02 (0.89-1.16) | 0.94 (0.83-1.08) | 0.86 (0.75-0.98) | 0.008 |
| Model 4†† | 1.00 | 1.00 (0.88-1.15) | 0.93 (0.82-1.07) | 0.87 (0.76-0.99) | 0.01 |
| Nondairy calcium | |||||
| Median intake, mg/d | 202 | 260 | 303 | 378 | |
| Cases/non-cases | 562/3,262 | 625/3,277 | 615/3,709 | 641/4,157 | |
| Model 1† | 1.00 | 1.01 (0.88-1.15) | 0.87 (0.76-1.00) | 0.89 (0.78-1.01) | 0.02 |
| Model 2‡ | 1.00 | 1.03 (0.90-1.17) | 0.89 (0.78-1.02) | 0.89 (0.78-1.01) | 0.02 |
| Model 3** | 1.00 | 1.11 (0.96-1.30) | 1.01 (0.85-1.21) | 1.06 (0.85-1.30) | 0.81 |
| Model 4†† | 1.00 | 1.04 (0.91-1.19) | 0.91 (0.79-1.05) | 0.92 (0.79-1.07) | 0.14 |
Linear trend was tested by treating the median values of quartiles or frequency of milk intake as a continuous variable
Multivariable model 1: adjusted for age, sex, dialect, year of interview, educational level, and marriage status;
Multivariable model 2: further adjusted for body mass index, physical activity, smoking status, alcohol use, baseline history of self-reported hypertension, diabetes, heart attack, and stroke, history of cancer, sleep status, and total energy intake;
Multivariable model 3: further adjusted for red meat, poultry, fish, vegetables, fruits, tea, coffee, and soda;
Multivariable model 4: model 2 plus vegetable-fruit-soy (VFS) dietary pattern.
While we did not find any significant interaction with sex, age, BMI, and vitamin D intake (P > 0.3), dietary calcium showed an inverse association in those without a history of coronary artery disease, stroke, hypertension, and diabetes (OR 0.81, 95% CI 0.69-0.96; P-trend=0.01) but not in participants with any one of these pre-existing comorbidities (OR 0.94, 95% CI 0.74-1.23, P for trend=0.98) (P for interaction=0.04).
In sensitivity analyses, we used the SM-MMSE score as a continuous outcome and observed significantly higher values at the highest quartiles of dairy and dairy calcium intake compared to the lowest quartiles. In contrast, there were no significant differences in SM-MMSE scores across quartiles of soy and non-dairy calcium intake (Supplemental table 5).
Discussion
In this longitudinal study among Chinese in Singapore, we observed a lower risk of cognitive impairment associated with higher mid-life intake of dairy food, and this was more evident in the subgroup without history of cardiometabolic diseases. Correspondingly, while dairy calcium was associated with reduced risk, non-dairy calcium did not influence the risk of cognitive impairment. Conversely, the intake of soy food and isoflavones was not associated with cognitive impairment.
To our knowledge, there are four cohort studies published to date about the association between milk or dairy intake and cognitive disease that have reported controversial findings [37-40]. Among them, one small study in a Japanese population found a significant inverse association between quartiles of dairy intake and risk of Alzheimer’s disease but not vascular dementia [37], whereas another small Japanese study reported that daily milk intake was inversely associated with risk of vascular dementia but not Alzheimer’s disease [38]. Conversely, in a study of Australian men, regular intake of full-cream milk was associated with poorer cognitive function assessed by MMSE [39]; and a study of French women reported an association for cognitive decline with intakes of milk and yoghurt, cheese, and dairy desserts and ice-cream, although only the association with ice-cream reached statistical significance [40]. Only in two of these studies, the associations were appropriately controlled for potential confounders such as BMI, smoking, and physical activity in their statistical models [37,40], and among them, only one [37] adjusted for intake of other foods such as vegetable, fruit, fish, and red meat, all of which might be related to cognitive impairment. In a meta-analysis involving a total 10,941 participants from these four cohort studies and three cross-sectional studies, the authors found that a higher milk intake was inversely associated with a composite outcome comprising cognitive impairment/decline, dementia, and Alzheimer’s disease. However, an inverse but statistically non-significant association was observed in pooled analysis of the five studies that had specifically used cognitive impairment/decline as outcome (OR 0.76, 95% CI 0.50-1.17). The authors attributed the inconsistency of results across these observational studies to the difference in outcome measures, study designs, small sample sizes and lack of full adjustment for potential confounders (I2 = 77%) [3].
Risk factors of cardiovascular disease, whether as composite risk score [41] or as individual factors [42], have been shown to be useful indicators of future cognitive status. We have previously reported that dairy intake could reduce the risk of hypertension [15], diabetes [14] and stroke, but not coronary artery disease [43] in this Singapore Chinese cohort. The inverse association between dairy intake and cognitive impairment in our study was observed in the subgroup without history of these diseases. Taken together, our results suggest that dairy intake could reduce the risk of cognitive impairment through the prevention of these diseases that have adverse impact on cognition [9], and is therefore less effective in those who already suffer from these diseases. More research is warranted to confirm our findings and elucidate biological mechanism for the effect of dairy intake on cognitive function.
Dairy products are one of the richest sources of calcium that provide high absorbability [44]. Prior epidemiologic studies suggested that calcium may be a mediating factor for the inverse association observed between dairy products and cardiometabolic disease risk [5-8]. Given an established relationship between cardiometabolic diseases and cognitive state [41,42], we speculated that calcium may also mediate the association between dairy intake and cognitive impairment. The observational evidence about the relationship between dietary calcium and cognition is scarce and controversial. Consistent with our finding on dietary calcium, a 17-year follow-up study in a Japanese population also reported that a higher calcium intake was associated with a lower risk of all-cause dementia. In contrast, a null association was observed between dietary calcium and mild cognitive impairment (HR per 100 mg increments 1.01, 95% CI 0.89–1.15) in an Australian population [13]. Aside from our comprehensive adjustment, we separately assessed calcium intake from its two major sources in the present analysis. In populations with high dairy intake, up to 80% of dietary calcium may come from dairy products [6], which makes it difficult to evaluate the relation of calcium independent from that of dairy. Conversely, dairy products only contributed to approximately 17% of total calcium intake in this study [16]. We observed an inverse association of cognitive impairment with dairy calcium even in a fully adjusted model, but not with non-dairy calcium. Therefore, we concluded that the observed protective association for calcium was mainly accounted for by the intake of dairy products. This is in line with our previous findings regarding the protective association between calcium intake and risk of hypertension [15] or diabetes [14]. Although we cannot rule out a potential role of calcium that is derived specifically from dairy food, the lack of association between non-dairy calcium and cognitive impairment in the present study points to other components present in dairy food being the mediating factors for the protective role in cognitive function. Potential candidates for these factors are branched-chain amino acids, lactotripeptides, and other minerals such as potassium in dairy [45], which have been shown in studies to possibly modify neurovascular dysfunction by reducing metabolic risks [3].
Soy foods are the major source of isoflavones, which have been shown in experimental studies to have neuroprotective potentials that could be mediated via their antioxidant and antiinflammatory effects, as well as their interactions with estrogen receptors localized throughout the brain, especially in areas vulnerable to age-related cognitive decline [19]. However, findings of observational studies in human have been controversial. A study of Japanese-American men that used the Cognitive Abilities Screening Instrument to assess cognition reported that a consistently higher midlife intake of tofu assessed at two time points nine years apart was associated with a higher risk of cognitive decline in late life [46]. In contrast, a study in Japan using MMSE for cognitive assessment reported an inverse association of cognitive impairment with intake of total soy products and total or subtypes of isoflavones among women but not men [47]. On the contrary, another longitudinal study of American Japanese and American Chinese women that assessed cognition using four different cognitive function tests did not report any association between intake of genistein and the cognitive function scores [48]. Comparing these two studies [47,48], the study with significant findings had relatively higher genistein intake (a mean of 24.6 mg/d versus 7.17 mg/d in the mid-tertile), larger sample size, older participants (68 vs. 46 years), and longer follow-up (8 vs. 4 years). In comparison, the mean age (53 years) and genistein intake (9.9 mg/d) of the women in our study were in the ranges between these two Japanese studies. Yet, in another multiethnic longitudinal study in the US, higher isoflavone intake was associated with a better processing speed but a worse verbal memory in postmenopausal Asian women [49], suggesting that the observed heterogeneity might be due to differential effects of soy isoflavones on different cognitive vulnerability. Hence, we acknowledge that results from the aforementioned observational studies and our study are conflicting, and comparisons of results across different studies may be complicated by the differences in the follow-up period, use of outcome measurement tools, and dose and type of soy food intake, as well as the lack of proper adjustment for potential confounders.
The strengths of our study were its large sample size, long follow-up, and accurate measures of foods and nutrients in a population with high soy and low dairy intake, which gave us a unique opportunity to study independent effects of dairy and calcium intake in the same study population. However, we acknowledge the limitations in our study. First, we only assessed cognitive function at the follow-up 3 visits and did not exclude people with cognitive impairment at recruitment, which may have resulted in reverse causation. Nevertheless, since participants in our study were required to complete a set of complex questionnaires (including a 165-item FFQ) at baseline, it was unlikely for those with severely impaired cognitive function to have been included in this cohort at recruitment. Furthermore, if we had recruited those with mild cognitive impairment at baseline, any misclassification of soy or dairy was likely to be random and non-differential in this group, and thus leading to a possible underestimation of the true risk estimates. Second, the role of MMSE is still controversial as a stand-alone single-administration test in the identification of mild cognitive impairment patients who could later develop dementia [28]. Therefore, future studies are still needed to explore if our findings can be extrapolated to clinical dementia. Third, dietary habits were only explored at baseline and we were unable to take into account the impact of potential changes in dietary intake after baseline interview. The FFQs rely on participants’ view of their own long-term diet, and this could be a source of misclassification, but the misclassification is less likely to be differential given the longitudinal nature of our study with its long follow-up. Nevertheless, the intrinsic limitations of FFQs should be considered. Fourth, our findings were confined to the ranges of intake in our population, and their generalizability for other populations should be tested in future studies. Finally, despite our multivariable adjustment for covariates that included dietary pattern, the possibility of bias introduced by residual confounding cannot be eliminated.
In conclusion, our findings suggest that regular dairy intake, such as one glass of milk per day, may protect against the development of cognitive impairment later in life but the potential benefits could not be attributed to its calcium content alone. Conversely, our data does not support a relationship, whether inverse or direct, between soy or isoflavone intake and cognitive impairment. Our findings from this population-based cohort of Chinese in Singapore provide evidence for the recommendation of dairy foods in the prevention of cognitive impairment at levels of intake that are applicable to other Asian populations. Given the observational nature of this study, future intervention studies are required to substantiate these findings.
Supplementary Material
Acknowledgment
We thank Siew-Hong Low of the National University of Singapore for supervising the fieldwork in the Singapore Chinese Health Study and Renwei Wang for the maintenance of the cohort study database. Finally, we acknowledge the founding Principal Investigator of the Singapore Chinese Health Study, Mimi C. Yu.
Grant Support: This work was supported by the Singapore National Medical Research Council (NMRC/CSA/0055/2013) and the United States National Cancer Institute, National Institutes of Health (UM1 CA182876 and R01 CA144034), and the Saw Swee Hock School of Public Health, National University of Singapore. M Talaei is supported by Cambridge-NUHS Seed Fund (NUHSRO/2017/015/Cambridge/01). A Pan is supported by the National Key Research and Development Program of China (2017YFC0907504). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Footnotes
Names for Pubmed Index: Talaei M, Feng L, Yuan JM, Pan A, Koh WP.
Conflict(s) of Interest/Disclosure: None to be declared. On behalf of all authors, the corresponding author states that there is no conflict of interest.
Publisher's Disclaimer: This Author Accepted Manuscript is a PDF file of an unedited peer-reviewed manuscript that has been accepted for publication but has not been copyedited or corrected. The official version of record that is published in the journal is kept up to date and so may therefore differ from this version.
References
- 1.Ulamek-Koziol M, Czuczwar SJ, Pluta R (2016) Nutrients and mild cognitive impairment: Food for thought. Nutrition 32 (7-8):910–912. doi: 10.1016/j.nut.2016.01.006 [DOI] [PubMed] [Google Scholar]
- 2.Vandewoude M, Barberger-Gateau P, Cederholm T, Mecocci P, Salva A, Sergi G, Topinkova E, Van Asselt D (2016) Healthy brain ageing and cognition: Nutritional factors. European Geriatric Medicine 7 (1):77–85. doi: 10.1016/j.eurger.2015.12.005 [DOI] [Google Scholar]
- 3.Wu L, Sun D (2016) Meta-Analysis of Milk Consumption and the Risk of Cognitive Disorders. Nutrients 8 (12). doi: 10.3390/nu8120824 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Crichton GE, Bryan J, Murphy KJ, Buckley J (2010) Review of dairy consumption and cognitive performance in adults: findings and methodological issues. Dement Geriatr Cogn Disord 30 (4):352–361. doi: 10.1159/000320987 [DOI] [PubMed] [Google Scholar]
- 5.Soedamah-Muthu SS, Verberne LD, Ding EL, Engberink MF, Geleijnse JM (2012) Dairy consumption and incidence of hypertension: a dose-response meta-analysis of prospective cohort studies. Hypertension 60 (5):1131–1137. doi: 10.1161/HYPERTENSIONAHA.112.195206 [DOI] [PubMed] [Google Scholar]
- 6.McGrane MM, Essery E, Obbagy J, Lyon J, Macneil P, Spahn J, Van Horn L (2011) Dairy Consumption, Blood Pressure, and Risk of Hypertension: An Evidence-Based Review of Recent Literature. Curr Cardiovasc Risk Rep 5 (4):287–298. doi: 10.1007/s12170-011-0181-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Ralston RA, Lee JH, Truby H, Palermo CE, Walker KZ (2012) A systematic review and metaanalysis of elevated blood pressure and consumption of dairy foods. J Hum Hypertens 26 (1):3–13. doi: 10.1038/jhh.2011.3 [DOI] [PubMed] [Google Scholar]
- 8.Astrup A (2014) Yogurt and dairy product consumption to prevent cardiometabolic diseases: epidemiologic and experimental studies. Am J Clin Nutr 99 (5 Suppl):1235S–1242S. doi: 10.3945/ajcn.113.073015 [DOI] [PubMed] [Google Scholar]
- 9.Picano E, Bruno RM, Ferrari GF, Bonuccelli U (2014) Cognitive impairment and cardiovascular disease: so near, so far. Int J Cardiol 175 (1):21–29. doi: 10.1016/j.ijcard.2014.05.004 [DOI] [PubMed] [Google Scholar]
- 10.Alzheimer's Association Calcium Hypothesis W (2017) Calcium Hypothesis of Alzheimer's disease and brain aging: A framework for integrating new evidence into a comprehensive theory of pathogenesis. Alzheimers Dement 13 (2):178–182 e117. doi: 10.1016/j.jalz.2016.12.006 [DOI] [PubMed] [Google Scholar]
- 11.Wang L, Manson JE, Sesso HD (2012) Calcium intake and risk of cardiovascular disease: a review of prospective studies and randomized clinical trials. Am J Cardiovasc Drugs 12 (2):105–116. doi: 10.2165/11595400-000000000-00000 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Ozawa M, Ninomiya T, Ohara T, Hirakawa Y, Doi Y, Hata J, Uchida K, Shirota T, Kitazono T, Kiyohara Y (2012) Self-reported dietary intake of potassium, calcium, and magnesium and risk of dementia in the Japanese: the Hisayama Study. J Am Geriatr Soc 60 (8):1515–1520. doi: 10.1111/j.1532-5415.2012.04061.x [DOI] [PubMed] [Google Scholar]
- 13.Cherbuin N, Kumar R, Sachdev PS, Anstey KJ (2014) Dietary Mineral Intake and Risk of Mild Cognitive Impairment: The PATH through Life Project. Front Aging Neurosci 6:4. doi: 10.3389/fnagi.2014.00004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Talaei M, Pan A, Yuan JM, Koh WP (2018) Dairy intake and risk of type 2 diabetes. Clin Nutr 37 (2):712–718. doi: 10.1016/j.clnu.2017.02.022 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Talaei M, Pan A, Yuan JM, Koh WP (2017) Dairy Food Intake Is Inversely Associated with Risk of Hypertension: The Singapore Chinese Health Study. J Nutr 147 (2):235–241. doi: 10.3945/jn.116.238485 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Butler LM, Wong AS, Koh WP, Wang R, Yuan JM, Yu MC (2010) Calcium intake increases risk of prostate cancer among Singapore Chinese. Cancer Res 70 (12):4941–4948. doi: 10.1158/0008-5472.CAN-09-4544 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Prince M, Bryce R, Albanese E, Wimo A, Ribeiro W, Ferri CP (2013) The global prevalence of dementia: a systematic review and metaanalysis. Alzheimers Dement 9 (1):63–75 e62. doi: 10.1016/j.jalz.2012.11.007 [DOI] [PubMed] [Google Scholar]
- 18.Zhao L, Brinton RD (2007) WHI and WHIMS follow-up and human studies of soy isoflavones on cognition. Expert Rev Neurother 7 (11):1549–1564. doi: 10.1586/14737175.7.11.1549 [DOI] [PubMed] [Google Scholar]
- 19.Soni M, Rahardjo TB, Soekardi R, Sulistyowati Y, Lestariningsih, Yesufu-Udechuku A, Irsan A, Hogervorst E (2014) Phytoestrogens and cognitive function: a review. Maturitas 77 (3):209–220. doi: 10.1016/j.maturitas.2013.12.010 [DOI] [PubMed] [Google Scholar]
- 20.Thaung Zaw JJ, Howe PRC, Wong RHX (2017) Does phytoestrogen supplementation improve cognition in humans? A systematic review. Ann N Y Acad Sci 1403 (1):150–163. doi: 10.1111/nyas.13459 [DOI] [PubMed] [Google Scholar]
- 21.Hankin JH, Stram DO, Arakawa K, Park S, Low SH, Lee HP, Yu MC (2001) Singapore Chinese Health Study: development, validation, and calibration of the quantitative food frequency questionnaire. Nutr Cancer 39 (2):187–195. doi: 10.1207/S15327914nc392_5 [DOI] [PubMed] [Google Scholar]
- 22.Stram DO, Hankin JH, Wilkens LR, Pike MC, Monroe KR, Park S, Henderson BE, Nomura AM, Earle ME, Nagamine FS, Kolonel LN (2000) Calibration of the dietary questionnaire for a multiethnic cohort in Hawaii and Los Angeles. Am J Epidemiol 151 (4):358–370 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Seow A, Shi CY, Franke AA, Hankin JH, Lee HP, Yu MC (1998) Isoflavonoid levels in spot urine are associated with frequency of dietary soy intake in a population-based sample of middle-aged and older Chinese in Singapore. Cancer EpidemiolBiomarkers Prev 7 (2):135–140 [PubMed] [Google Scholar]
- 24.Talaei M, Lee BL, Ong CN, van Dam RM, Yuan JM, Koh WP, Pan A (2016) Urine phytooestrogen metabolites are not significantly associated with risk of type 2 diabetes: the Singapore Chinese health study. Br J Nutr 115 (9):1607–1615. doi: 10.1017/S0007114516000581 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Talaei M, Koh WP, van Dam RM, Yuan JM, Pan A (2014) Dietary soy intake is not associated with risk of cardiovascular disease mortality in Singapore Chinese adults. J Nutr 144 (6):921–928. doi: 10.3945/jn.114.190454 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Feng L, Chong MS, Lim WS, Ng TP (2012) The Modified Mini-Mental State Examination test: normative data for Singapore Chinese older adults and its performance in detecting early cognitive impairment. Singapore Med J 53 (7):458–462 [PubMed] [Google Scholar]
- 27.Folstein MF, Folstein SE, McHugh PR (1975) "Mini-mental state". A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12 (3):189–198 [DOI] [PubMed] [Google Scholar]
- 28.Arevalo-Rodriguez I, Smailagic N, Roque IFM, Ciapponi A, Sanchez-Perez E, Giannakou A, Pedraza OL, Bonfill Cosp X, Cullum S (2015) Mini-Mental State Examination (MMSE) for the detection of Alzheimer's disease and other dementias in people with mild cognitive impairment (MCI). Cochrane Database Syst Rev (3):CD010783. doi: 10.1002/14651858.CD010783.pub2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Tse CS, Chang JF, Leung GT, Fung AW, Hau KT, Chiu HF, Lam LC (2013) Effects of education on very mild dementia among Chinese people in Hong Kong: potential mediators in the Cantonese Mini-Mental State Examination tasks. Aging Ment Health 17 (3):310–318. doi: 10.1080/13607863.2012.743962 [DOI] [PubMed] [Google Scholar]
- 30.Katzman R, Zhang MY, Ouang Ya Q, Wang ZY, Liu WT, Yu E, Wong SC, Salmon DP, Grant I (1988) A Chinese version of the Mini-Mental State Examination; impact of illiteracy in a Shanghai dementia survey. J Clin Epidemiol 41 (10):971–978 [DOI] [PubMed] [Google Scholar]
- 31.Willett W, Stampfer MJ (2013) Implications of total energy intake for epidemiologic analysis In: Willett W (ed) Nutritional epidemiology. vol 30, 3rd edn. Oxford University Press, Oxford, USA, p 260 [Google Scholar]
- 32.Imtiaz B, Tolppanen AM, Kivipelto M, Soininen H (2014) Future directions in Alzheimer's disease from risk factors to prevention. Biochem Pharmacol 88 (4):661–670. doi: 10.1016/j.bcp.2014.01.003 [DOI] [PubMed] [Google Scholar]
- 33.Yusufov M, Weyandt LL, Piryatinsky I (2017) Alzheimer's disease and diet: a systematic review. Int J Neurosci 127 (2):161–175. doi: 10.3109/00207454.2016.1155572 [DOI] [PubMed] [Google Scholar]
- 34.Odegaard AO, Koh WP, Butler LM, Duval S, Gross MD, Yu MC, Yuan JM, Pereira MA (2011) Dietary patterns and incident type 2 diabetes in chinese men and women: the singapore chinese health study. Diabetes Care 34 (4):880–885. doi: 10.2337/dc10-2350 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Odegaard AO, Koh WP, Yuan JM, Gross MD, Pereira MA (2014) Dietary patterns and mortality in a Chinese population. Am J Clin Nutr 100 (3):877–883. doi: 10.3945/ajcn.114.086124 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Dai Z, Butler LM, van Dam RM, Ang LW, Yuan JM, Koh WP (2014) Adherence to a vegetable-fruit-soy dietary pattern or the Alternative Healthy Eating Index is associated with lower hip fracture risk among Singapore Chinese. J Nutr 144 (4):511–518. doi: 10.3945/jn.113.187955 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Ozawa M, Ohara T, Ninomiya T, Hata J, Yoshida D, Mukai N, Nagata M, Uchida K, Shirota T, Kitazono T, Kiyohara Y (2014) Milk and dairy consumption and risk of dementia in an elderly Japanese population: the Hisayama Study. J Am Geriatr Soc 62 (7): 1224–1230. doi: 10.1111/jgs.12887 [DOI] [PubMed] [Google Scholar]
- 38.Yamada M, Kasagi F, Sasaki H, Masunari N, Mimori Y, Suzuki G (2003) Association between dementia and midlife risk factors: the Radiation Effects Research Foundation Adult Health Study. J Am Geriatr Soc 51 (3):410–414 [DOI] [PubMed] [Google Scholar]
- 39.Almeida OP, Norman P, Hankey G, Jamrozik K, Flicker L (2006) Successful mental health aging: results from a longitudinal study of older Australian men. Am J Geriatr Psychiatry 14 (1):27–35. doi: 10.1097/01.JGP.0000192486.20308.42 [DOI] [PubMed] [Google Scholar]
- 40.Vercambre MN, Boutron-Ruault MC, Ritchie K, Clavel-Chapelon F, Berr C (2009) Long-term association of food and nutrient intakes with cognitive and functional decline: a 13-year follow-up study of elderly French women. Br J Nutr 102 (3):419–427. doi: 10.1017/S0007114508201959 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.DeRight J, Jorgensen RS, Cabral MJ (2015) Composite cardiovascular risk scores and neuropsychological functioning: a meta-analytic review. Ann Behav Med 49 (3):344–357. doi: 10.1007/s12160-014-9681-0 [DOI] [PubMed] [Google Scholar]
- 42.Gottesman RF, Albert MS, Alonso A, Coker LH, Coresh J, Davis SM, Deal JA, McKhann GM, Mosley TH, Sharrett AR, Schneider ALC, Windham BG, Wruck LM, Knopman DS (2017) Associations Between Midlife Vascular Risk Factors and 25-Year Incident Dementia in the Atherosclerosis Risk in Communities (ARIC) Cohort. JAMA Neurol 74 (10):1246–1254. doi: 10.1001/jamaneurol.2017.1658 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Talaei M, Koh WP, Yuan JM, Pan A (2017) The association between dairy product intake and cardiovascular disease mortality in Chinese adults. Eur J Nutr 56 (7):2343–2352. doi: 10.1007/s00394-016-1274-1 [DOI] [PubMed] [Google Scholar]
- 44.Gueguen L, Pointillart A (2000) The bioavailability of dietary calcium. J Am Coll Nutr 19 (2 Suppl):119S–136S [DOI] [PubMed] [Google Scholar]
- 45.Lee M, Lee H, Kim J (2018) Dairy food consumption is associated with a lower risk of the metabolic syndrome and its components: a systematic review and meta-analysis. Br J Nutr 120 (4):373–384. doi: 10.1017/S0007114518001460 [DOI] [PubMed] [Google Scholar]
- 46.White LR, Petrovitch H, Ross GW, Masaki K, Hardman J, Nelson J, Davis D, Markesbery W (2000) Brain aging and midlife tofu consumption. J Am Coll Nutr 19 (2):242–255 [DOI] [PubMed] [Google Scholar]
- 47.Nakamoto M, Otsuka R, Nishita Y, Tange C, Tomida M, Kato Y, Imai T, Sakai T, Ando F, Shimokata H (2018) Soy food and isoflavone intake reduces the risk of cognitive impairment in elderly Japanese women. Eur J Clin Nutr. doi: 10.1038/s41430-017-0061-2 [DOI] [PubMed] [Google Scholar]
- 48.Huang MH, Luetters C, Buckwalter GJ, Seeman TE, Gold EB, Sternfeld B, Greendale GA (2006) Dietary genistein intake and cognitive performance in a multiethnic cohort of midlife women. Menopause 13 (4):621–630. doi: 10.1097/01.gme.0000227336.35620.8f [DOI] [PubMed] [Google Scholar]
- 49.Greendale GA, Huang MH, Leung K, Crawford SL, Gold EB, Wight R, Waetjen E, Karlamangla AS (2012) Dietary phytoestrogen intakes and cognitive function during the menopausal transition: results from the Study of Women's Health Across the Nation Phytoestrogen Study. Menopause 19 (8):894–903. doi: 10.1097/gme.0b013e318242a654 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.