Factors Associated with Lower Cognitive Performance Scores Among Older Japanese Men in Hawaii and Japan

Abstract

Background:

Few studies have compared factors related to cognitive function among people with similar genetic backgrounds but different lifestyles.

Objective:

We aimed to identify factors related to lower cognitive scores among older Japanese men in two genetically similar cohorts exposed to different lifestyle factors.

Methods:

This cross-sectional study of community-dwelling Japanese men aged 71–81 years included 2,628 men enrolled in the Kuakini Honolulu-Asia Aging Study based in Hawaii and 349 men in the Shiga Epidemiological Study of Subclinical Atherosclerosis based in Japan. We compared participant performance through Cognitive Abilities Screening Instrument (CASI) assessment in Hawaii (1991–1993) and Japan (2009–2014). Factors related to low cognitive scores (history of cardiovascular disease, cardiometabolic factors, and lifestyle factors) were identified with questionnaires and measurements. Multivariable logistic regression analysis was used to calculate the adjusted odds ratios (ORs) of a low (< 82) CASI score based on different factors.

Results:

CASI scores were lower in Hawaii than in Japan [21.2%(n = 556) versus 12.3%(n = 43), p < 0.001], though this was not significant when adjusted for age and educational attainment (Hawaii 20.3%versus Japan 17.9%, p = 0.328). History of stroke (OR = 1.65, 95%confidence interval = 1.19–2.29) was positively associated with low cognitive scores in Hawaii. Body mass index ≥25 kg/m² tended to be associated with low cognitive scores in Japan; there was a significant interaction between the cohorts.

Conclusion:

Cognitive scores differences between cohorts were mostly explained by differences in educational attainment. Conversely, cardiovascular diseases and cardiometabolic factors differentially impacted cognitive scores among genetically similar older men exposed to different lifestyle factors.

Keywords

Aged cognitive decline community dwelling Japanese Japanese Americans men

INTRODUCTION

Dementia is one of the major causes of disability and dependency among older adults worldwide [1]. The prevalence of dementia is increasing; however, this prevalence varies by region. The change in age-standardized prevalence from 1990 to 2016 has increased in East Asia and high-income Asia-Pacific countries, including Japan, but not in western Europe and high-income North American countries [2]. Lifestyle choices associated with poor health outcomes, such as reduced exercise and increased fat intake, have become more common among East Asian populations, concurrent with the economic development of their respective countries [3, 4]. This, along with longer life expectancy and the rise of chronic age-related diseases, could lead to an increased prevalence of dementia [5, 6].

Previous observational studies, originating mainly from western countries, have reported that the principle modifiable risk factors associated with dementia were cardiometabolic characteristics, such as diabetes, midlife obesity, and midlife hypertension [7], and lifestyle factors, such as physical inactivity; smoking; lower intake of fish, fruits, and vegetables; and higher intake of red meat [8]. Lifestyle choices are often determined by cultural and social constraints. Cardiometabolic disorders associated with cognitive decline develop as a result of both lifestyle choices and genetic factors [9]. One approach to differentiate between the effects of lifestyle and genetic factors on cognitive function involves a comparison of lifestyle and cardiometabolic risk factors among people with similar genetic backgrounds. Few studies have compared factors related to cognitive function among people who have similar genetic backgrounds, but different lifestyles.

Japanese-Americans who have emigrated from Japan to Hawaii are exposed to westernized life-styles from an early age. Studies from the 1980s have reported that this population has worse cardiometabolic performance (with the exception of hypertension rate), including a high body mass index (BMI), serum lipid level, and diabetes rate than do their Japan-based counterparts [10]. The older adults in Japan are individuals who started to experience westernized lifestyles around the 1970s, which for them, was around mid-life. This study investigated factors associated with low cognitive function in two groups of genetically similar Japanese men exposed to westernized lifestyles from different starting points in life using cross-sectional data from the Kuakini Honolulu-Asia Aging Study (HAAS) and the Shiga Epidemiological Study of Subclinical Atherosclerosis (SESSA). This comparison might help develop strategies to mitigate risks of cognitive decline among East Asian populations, where the prevalence of dementia is increasing because of modifiable lifestyle risk factors.

METHODS

Study populations

The Kuakini Honolulu Heart Program (HHP) is a population-based cohort study of cardiovascular diseases among 8,006 Japanese-American men born during 1900–1919 and living on the island of Oahu, Hawaii, at the time of enrollment in 1965 [11]. The Kuakini HAAS baseline examination began concurrently with the fourth Kuakini HHP examination (1991–1993), launched to study aging-related conditions, such as cognitive function and decline, and included 3,741 men, aged 71–93 years, representing approximately 80%of surviving members of the original cohort [12].

The SESSA is a study of subclinical atherosclerosis and its determinants in a random sample of Japanese residents; the details of the study’s methods, including recruitment, have been reported previously [13, 14]. In brief, the study enrolled 1,094 Japanese men aged 40–79 years who lived in Kusatsu City, Shiga, Japan, in 2006, based on the data from the city’s Basic Residents’ Register. Located in central Japan, Kusatsu City has an industrial structure, representative of an average city in Japan. From 2009 to 2014, all participants were invited for follow-up examination, including cognitive function assessment; of the study cohort, 853 (78.0%) participated in this assessment [15, 16].

For the present study, we limited our analyses to persons aged 71–81 years to ensure comparability, as participants in this age range were found in both cohorts. Moreover, we limited the Hawaii sample to second-generation immigrants whose parents moved from Japan to Hawaii, as these men were exposed to a westernized lifestyle from an early age. The institutional review board of Shiga University of Medical Science reviewed and approved the SESSA study. The Kuakini Medical Center Institutional Review Board reviewed and approved the Kuakini HAAS. All participants or their proxies provided written informed consent.

A total of 1,113 men from the Kuakini HAAS baseline exam (1991–93) and 504 men from the SESSA cohort were excluded from analysis because of meeting any of the following exclusion criteria that had been set a priori: missing the Cognitive Abilities Screening Instrument (CASI) test data (n = 4 in HAAS, n = 34 in SESSA), older than 82 years or younger than 71 years (n = 767 in HAAS, n = 470 in SESSA), place of examination other than a clinic (n = 304 in HAAS), or first-generation immigrants in the Hawaii sample (n = 38) (Fig. 1). Ultimately, 2,977 participants were included in the analysis (n = 2,628 in HAAS and n = 349 in SESSA).

Fig. 1

Flow chart capturing the participant selection process in the present study. CASI, Cognitive Abilities Screening Instrument.

Assessment of risk factors and covariates

Data on risk factors and covariates were obtained at Exam 4 (1991–1993) and using a mail questionnaire (1987–1990) in the Kuakini HAAS cohort and during 2009–2014 in the SESSA cohort follow-up study. The number of years of formal education completed by each participant was obtained from participants’ self-report. Smoking and drinking status were self-reported as “never,” “former,” or “current.” Detailed alcohol consumption data were collected among current drinkers by asking the type (e.g., wine and beer), number (e.g., number of cans of beer), and portion size (e.g., 100 mL of wine) of alcohol consumed. We then calculated the total pure ethanol consumption per day. “Current drinking” was categorized by alcohol consumption volume (≤20 g/day versus > 20 g/day of pure ethanol). Data on exercise and food frequency as a measure of dietary habits were obtained by self-reported questionnaires. High vegetable intake was defined as ≥2 times/day, high fruit intake was defined as ≥1 time/day, and low fish or meat intake was defined as < 3 times/month. We calculated the participants’ BMI as weight (kg) divided by height-squared (m²) and categorized BMI as underweight (< 18.5 kg/m²), normal (18.5–25.0 kg/m²), and obese (≥25 kg/m²). In the HAAS cohort, blood pressure was measured 3 times with measurements 5 min apart on the left arm, after the participant had rested in a seated position for at least 10 min; a mercury-based sphygmomanometer with an appropriate-sized cuff was used. The mean of three measurements was used in the analysis. In the SESSA cohort, blood pressure was measured twice consecutively on the right arm of a seated participant, after the participant had emptied his bladder for urinalysis and had sat quietly for 5 min; an automated oscillometric sphygmomanometer (BP-8800; Omron Colin, Tokyo, Japan) with an appropriately sized cuff was used. The average of two measurements was used in the analysis.

In the HAAS cohort, blood specimens were obtained after an overnight fast of at least 12 h. To estimate the 2-h post-load glucose level, a 75-g glu-cose drink was administered to participants, regardless of their diabetes status. Separated plasma was frozen at −70°C for later measurement of total and high-density lipoprotein cholesterol and plasma glucose at the University of Vermont research laboratory. Lipids were measured using an Olympus Demand System (Olympus Corp., Lake Success, NY). The procedures for taking and preparing blood specimens for laboratory analysis were standardized by the guidelines of the lipid standardization laboratory of the US Centers for Disease Control and Prevention (CDC) [17]. Fasting and 2-h glucose levels were measured by the glucose oxidase method.

In the SESSA cohort, blood samples were obtained early in the clinic visit after a 12-h fast. The sam-ples were sent for routine laboratory tests, including testing for lipids and glucose levels. Serum total cholesterol levels were determined using enzymatic assays, and high-density lipoprotein cholesterol levels were measured using a direct method (Deter-miner-C-TC, and Determiner-L HDL-c, respectively; Kyowa Medix, Tokyo, Japan). Serum lipid levels were determined at a single laboratory (Shiga Laboratory; MEDIC, Shiga, Japan) that had been certified for standardized lipid measurements, according to the protocol of the CDC/Cholesterol Reference Method Laboratory Network (CRMLN). Plasma glucose levels were determined from sodium fluoride–treated plasma, using a hexokinase/glucose-6-phosphate dehydrogenase enzymatic assay.

Hypertension was defined as systolic/diastolic blood pressure ≥140/90 mm Hg or relevant medication use. In the HAAS cohort, diabetes was defined as medication use, fasting glucose level ≥126 mg/dL, or 2-h post-load glucose level ≥200 mg/dL. In the SESSA cohort, diabetes was defined as fasting blood glucose ≥126 mg/dL, HbA1c ≥6.5%(NGSP), or medication use. The non-high-density lipoprotein (HDL) cholesterol level was calculated by subtracting the high-density lipoprotein cholesterol level from the total cholesterol level.

Cognitive function

Cognitive function was evaluated by participants’ performance on the CASI, a validated comprehensive measure of global cognitive function developed for use in cross-cultural and cross-national studies [18]. The CASI comprises 25 questions and was used in its full form to assess the Kuakini HAAS cohort at the fourth Kuakini HHP examination and the SESSA cohort at a follow-up examination for the first time [16]. The CASI is a combination of the Hasegawa Dementia Screening Scale [19], Folstein Mini-Mental State Examination [20], and Modified Mini-Mental State (3-MS) test [21]. The CASI includes 9 domains, with tasks measuring attention, concentration, orientation, short- and long-term memory, language, visual construction, list-generating fluency, and abstraction/ judgment. Total CASI scores range from 0 to 100 points, where higher scores represent better cognitive function. CASI scores < 82 correspond to a score of 25–26 on the Folstein Mini-Mental State Examination, which indicated cognitive impairment [22]. We defined cognitive function as a dichotomous outcome: normal (CASI score ≥82) and low (CASI score < 82), similar to a previous study [22]. By using a CASI score cutoff of < 82, the sensitivity for dementia as defined by the Diagnostic and Statistical Manual of Mental Disorders, Revised Third Edition [23] was 80%and the specificity was 77%in this cohort [24].

Statistical analysis

Crude and adjusted estimates of the prevalence of low cognitive scores for each cohort were calculated using standard analysis of covariance and binary logistic regression models [25]. Crude and adjusted odds ratios (ORs) and 95%confidence intervals (CIs) for the low cognitive score (CASI score < 82) in the cohort were calculated using logistic regression analysis among the combined population. The ORs of each related factor for low cognitive scores, adjusted for sex and age, were calculated with logistic regression analysis in a separate model for each cohort. Since age and educational attainment were un-modifiable factors that largely impacted both cohorts, we also tested for interactions between each factor and low cognitive scores by inserting a product term (cohort ^* related factor, e.g., hypertension) into each model of age and educational attainment in the combined population. We considered p-values < 0.05 statistically significant, and all analyses were two-tailed. SAS version 9.4 (SAS Institute, Cary, NC) was used for all analyses.

RESULTS

The average CASI score was lower for older men based in Hawaii than for older men based in Japan (86.9 versus 88.3, p < 0.001). Moreover, 556 (21.2%) of the older men in Hawaii had low cognitive scores, compared to 43 (12.3%) of the older men in Japan (p < 0.001).

In Hawaii, the men were on average 0.6 years older than those in Japan, while the average educational attainment was 1.5 years less among the men in Hawaii than those in Japan (Table 1). There were fewer current smokers and drinkers among the men in Hawaii than among the men in Japan. Vegetable, meat, and fish intake was greater in Japan than in Hawaii. However, fruit intake was greater in Hawaii than in Japan. In Hawaii, the prevalence of obesity, hypertension, and levels of non-HDL cholesterol was higher than Japan, while there were no statistically significant differences in the prevalence of stroke, coronary heart disease (CHD), or diabetes. Concurrently, the prevalence of taking lipid-lowering drugs was higher in Japan than in Hawaii.

Table 1

Baseline demographic characteristics and lifestyle and disease risk factors among older men in the Hawaii (1991–1993) and Japan (2009–2014) cohorts

Factor	Hawaii (n = 2,628)	Japan (n = 349)	p ^a
CASI score	86.9±8.6	88.3±6.3	0.004
Age, y	75.7±2.7	75.1±2.8	< 0.001
Educational attainment, y	10.8±3.0	12.3±2.5	< 0.001
History of cardiovascular disease
Stroke, %	8.3 (218)	9.2 (32)	0.580
CHD, %	15.9 (417)	14.3 (50)	0.457
Cardiometabolic factors Body mass index, kg/m²
< 18.5, %	3.3 (87)	5.7 (20)	<0.001
18.5–25, %	63.4 (1666)	71.4 (249)
≥25, %	33.1 (869)	22.9 (80)
Height, cm	162.5±5.7	163.6±5.3	0.001
Body weight, kg	62.9±9.1	61.7±8.6	0.022
Hypertension, %	74.1 (1948)	66.2 (231)	0.001
Diabetes, %	29.3 (770)	27.8 (97)	0.467
Non-HDL cholesterol, mg/dL	141.7±32.3	136.1±31.0	0.002
Lipid-lowering medication, %	13.0 (342)	23.8 (83)	< 0.001
Lifestyle factors Smoking status
Never, %	35.8 (941)	20.9 (73)	< 0.001
Past, %	56.9 (1495)	63.6 (222)
Current, %	7.3 (191)	15.5 (54)
Drinking status
Never, %	39.5 (1037)	18.3 (64)	<0.001
Past, %	23.8 (626)	6.3 (22)
Current≤20 g ethanol/day, %	23.7 (624)	43.8 (153)
Current > 20 g ethanol/day, %	12.9 (340)	31.5 (110)
Regular exercise, %	67.1 (1763)	63.0 (220)	0.127
Frequency of dietary intake
High vegetable intake, %	14.5 (382)	56.5 (197)	< 0.001
High fruit intake, %	59.4 (1561)	43.8 (153)	< 0.001
Low meat intake, %	11.4 (299)	21.2 (74)	< 0.001
Low fish intake, %	30.3 (796)	11.5 (40)	< 0.001

Values are means±standard deviation or %(numbers of subjects with the factor). ^aDifferences were evaluated using t-tests or chi-square tests. Hypertension was defined as systolic/diastolic blood pressure ≥140/90 mm Hg or medication use. Diabetes was defined by disease history, medication use, fasting glucose ≥126 mg/dL, or 2 hours post-load glucose ≥200 mg/dL in Hawaii cohort. Diabetes was defined as fasting blood glucose ≥126 mg/dL or HbA1c ≥6.5%(NGSP) or medication use in Japan cohort. High vegetable intake was defined as ≥2 times/day, high fruit intake was defined as ≥1 time/day, and low fish or meat intake was defined as < 3 times/month. CASI, Cognitive Abilities Screening Instrument; CHD, coronary heart disease; HDL, high-density lipoprotein.

The proportion of older men with a low cognitive score in Hawaii was 8.9%higher than in Japan; however, after adjusting for age and educational attainment, these differences were no longer significant (Hawaii 20.3%versus Japan 17.9%, p = 0.328), while the OR of low cognitive scores in Hawaii versus Japan was reduced from 1.91 to 1.19 (Table 2). After further adjusting for history of cardiovascular diseases (stroke and CHD), cardiometabolic factors (BMI, presence of hypertension and diabetes, and lipid-lowering medication use), and lifestyle factors (smoking and drinking status, and regular exercise), the differences became negligible (20.1%versus 18.4%, p = 0.507).

Table 2

Unadjusted and adjusted proportions and odds ratios of CASI scores < 82 among older men based in the Hawaii versus Japan cohorts

Model	Hawaii (n = 2628)	Japan (n = 349)	Odds Ratio^a (95%CI)	p ^b
Crude	21.2%	12.3%	1.91 (1.37–2.66)	< 0.001
Age adjusted	21.0%	13.1%	1.78 (1.27–2.49)	0.001
Age and educational attainment adjusted	20.3%	17.9%	1.19 (0.84–1.69)	0.328
Final model	20.1%	18.4%	1.13 (0.79–1.63)	0.507

Final model was adjusted for age, educational attainment, presence of hypertension, diabetes, lipid-lowering medication use, obesity (body mass index ≥25 kg/m²), history of stroke and coronary heart disease, regular exercise, smoking status, and alcohol status. ^aOdds ratio of CASI score < 82 for men based in Hawaii versus men based in Japan. ^bp-value for odds ratio of CASI score < 82. CASI, Cognitive Abilities Screening Instrument; CI, confidence interval.

Table 3 shows the ORs of low cognitive scores for various factors within each cohort. An increase in age by 5 years was associated with a significant increase in the ORs (95%CI) of low cognitive score, with an OR of 1.97 (1.66–2.35) in Hawaii and 2.58 (1.41–4.71) in Japan. Educational attainment was significantly negatively associated with the age-adjusted ORs of low cognitive score, with an OR of 0.44 (0.39–0.50) in Hawaii and 0.34 (0.21–0.55) in Japan.

Table 3

Age- and educational attainment-adjusted odds ratios of CASI scores < 82 among elderly men in the Hawaii (1991–1993) and Japan (2009–2014) cohorts

Factor	References or factor categories	Odds ratio (95%CI)		p of the interaction
		Hawaii (n = 2628)	Japan (n = 349)
Age (unadjusted)	Every 5 years	1.97 ^a (1.66–2.35)	2.58^b (1.41–4.71)	–
Educational attainment (age-adjusted)	Every 3 years	0.44^a (0.39–0.50)	0.34^a (0.21–0.55)	–
History of cardiovascular disease
Stroke ^d	Absent	1.65 ^a (1.19–2.29)	1.10 (0.35–3.48)	0.472
CHD ^d	Absent	1.06 (0.81–1.39)	1.90 (0.83–4.31)	0.173
Metabolic factors Body mass index ^d
< 18.5 kg/m^2d	18.5 – < 25 kg/m²	1.28 (0.75–2.16)	2.08 (0.58–7.44)	0.485
≥25 kg/m^2d	18.5 – < 25 kg/m²	0.87 (0.70–1.08)	2.04^c (0.97–4.29)	0.032
Hypertension^d	Yes versus No	1.07 (0.85–1.35)	1.23 (0.59–2.56)	0.754
Diabetes^d	Yes versus No	1.20^c (0.97–1.49)	1.35 (0.65–2.80)	0.787
Non-HDL cholesterol^d	Every 40 mg/dL	0.94 (0.83–1.07)	0.74 (0.47–1.16)	0.341
Lipid-lowering medication^d	Absent	0.76 (0.55–1.05)	1.33 (0.62–2.85)	0.197
Lifestyle factors Smoking status
Past^d	Never	0.93 (0.75–1.15)	0.58 (0.24–1.38)	0.376
Current ^d	Never	1.37 (0.93–2.00)	1.10 (0.40–3.03)	0.816
Alcohol status
Past^d	Never	0.93 (0.72–1.20)	0.92 (0.19–4.34)	1.000
Current≤20 g ethanol/day^d	Never	0.68^a (0.52–0.89)	0.99 (0.41–2.38)	0.427
Current > 20 g ethanol/day^d	Never	0.89 (0.65–1.22)	0.72 (0.27–1.96)	0.637
Regular exercise^d	Absent	0.87 (0.71–1.07)	0.51^c (0.26–1.01)	0.156
Frequency of dietary intake
High vegetable intake^d	< 2 /day	1.02 (0.76–1.35)	0.62 (0.31–1.24)	0.205
High fruit intake^d	< 1 /day	1.11 (0.90–1.37)	0.69 (0.34–1.40)	0.191
Low meat intake^d	> 3 /mon	0.82 (0.60–1.13)	0.81 (0.35–1.86)	0.939
Low fish intake^d	> 3 /mon	1.09 (0.87–1.35)	1.45 (0.58–3.67)	0.552

p^a < 0.001, ^b < 0.05, ^c < 0.10. ^dAge and educational attainment adjusted. p-values were calculated for the interaction of cohort^*each factor. The interactions between each factor and low cognitive scores were tested by inserting a product term (cohort ^* related factor, e.g., hypertension) into each model of age and educational attainment in the combined population. Definitions of hypertension, diabetes, and frequency of dietary intake are presented in the Table 1 footnote. CASI, Cognitive Abilities Screening Instrument; CI, confidence interval; CHD, coronary heart disease; HDL, high-density lipoprotein.

All remaining ORs were adjusted for age and educational attainment in each cohort. History of cardiovascular disease and cardiometabolic variables had a different impact on cognitive scores in Hawaii and Japan. For example, the OR for low cognitive score was 1.65 (1.19–2.29) for men based in Hawaii with history of stroke. On the other hand, the age-adjusted OR for low cognitive score was 2.01 (0.92–4.40, p = 0.079) for men based in Japan with history of CHD; however, this was no longer significant after adjusting for educational attainment. The OR of lower cognitive score for obesity (BMI ≥25 kg/m²) was 2.04 (0.97–4.29) in Japan and 0.87 (0.70–1.08) in Hawaii, and there was a significant interaction between Japan and Hawaii (p = 0.032). Diabetes tended to be positively associated with low cognitive scores in Hawaii. Regarding lifestyle factors, the OR of lower cognitive score for current drinking (≤20 g ethanol) was 0.68 (0.52–0.89) in Hawaii. Regular exercise in both Japan and Hawaii tended to be negatively associated with the risk of a low cognitive score. There were no significant associations among smoking status, dietary intake, and low cognitive scores.

DISCUSSION

In this comparison of genetically similar older men based in Hawaii and Japan, a low cognitive score was more common among older men in Hawaii than among older men in Japan (21.2%versus 12.3%, p < 0.001). However, after adjusting for age and educational attainment, these differences were no longer significant. Age and educational attainment were important factors associated with low cognitive scores within each cohort. In addition, stroke was significantly associated with low cognitive scores in Hawaii-based men. Diabetes tended to be associated with low cognitive scores in Hawaii, while BMI ≥25 kg/m² tended to be associated with low cognitive scores in Japan. Light drinking (≤20 g ethanol) was negatively associated with low cognitive scores in Hawaii. Regular exercise tended to be negatively associated with low cognitive scores in Japan, a trend that was also observed in Hawaii.

The proportion of older men with low cognitive scores was 8.9%higher in Hawaii than in Japan. This difference was reduced after adjusting for age and was no longer statistically significant after additionally adjusting for educational attainment. The difference in the proportion of participants with low cognitive scores reduced further after adjusting for cardiometabolic factors, history of cardiovascular disease, and lifestyle factors. In the present study, lower cognitive function among older Japanese men was mostly explained by differences in educational attainment between the cohorts, rather than by differences in lifestyle factors, cardiometabolic factors, and history of cardiovascular disease. Moreover, as more highly educated people tend to make healthier lifestyle choices, such as engaging in more physical activity and consuming higher amounts of vegetables and fruits, educational attainment might be a mediator of healthier lifestyles [26 –28]. Although a previous study has reported the prevalence of dementia among the Japanese-American population in Hawaii as being higher than that among the Japanese population in Japan, this comparison was not adjusted for educational attainment [12]. Therefore, previously reported differences in dementia prevalence between Hawaii and Japan-based populations of older men might be explained by differences in educational attainment. In the present study, on the other hand, part of the effect of educational attainment on lower cognitive scores may also be due to the “Flynn effect,” which states that a year of birth close to recent years is associated with a higher intelligence quotient (IQ) [29, 30], although this effect, thus far, has mostly been examined in the western countries [31, 32].

Previous reports, mainly from western countries, have shown that age, educational attainment, history of cardiovascular diseases (stroke and CHD), presence of diabetes, hypertension, obesity in mid-life, and smoking are risk factors for low cognitive scores. In contrast, regular exercise and greater vegetable, fruit, and fish intake alongside lower meat intake have been negatively associated with the risk of cognitive impairment [8 , 34]. The findings from our study are consistent with previously reported associations, with some subtle differences.

Except for age, educational attainment, and BMI, the strongest risk factor for low cognitive scores in the present study was a history of cardiovascular diseases; however, this impact was different between cohorts. In Hawaii, stroke was linked to lower cognitive scores; in contrast, in Japan, CHD tend to be linked to lower cognitive scores after adjusting for age; there were no significant interactions between Hawaii and Japan. After the rapid economic growth of the 1970s in Japan, lifestyle, and specifically diet, became westernized. The older Japan-based men included in the present study likely experienced lifestyle westernization from mid-life, around the same time that cardiometabolic diseases would have been developing. Therefore, Japan-based men experienced lifestyle westernization and were exposed to cardiometabolic diseases over a shorter time than Hawaii-based men. This could explain why history of cardiovascular disease was not associated with low cognitive score after adjusting for age and educational attainment.

Moreover, an important difference between the Kuakini HAAS and SESSA cohorts is that the HAAS participants had already been enrolled in the study for over 25 years during the time of first cognitive testing, whereas the SESSA participants were recruited approximately only four years before cognitive testing. Additionally, approximately 15%of the participants in the Kuakini HAAS cohort were excluded from the present analysis, as they were examined during home or nursing home visits, rather than at a clinic. Such participants tended to be sicker and frailer. As the excluded participants were likely less healthy and more cognitively impaired than the included participants, sampling may account for the lack of association between CHD history and low cognitive scores among older men in Hawaii. Similarly, it is possible that Japan-based men who had suffered a stroke and CHD declined the offer to participate in the SESSA cohort.

Lifestyle and cardiometabolic factors are risk factors for stroke and CHD, which are major risk factors for cognitive impairment and dementia [33, 34]. Many studies have shown that midlife obesity is a risk factor for dementia in later life [35]. However, the association between dementia and later life obesity is still controversial. Although some studies suggest that obesity later in life is protective against dementia [35], other studies suggest that it is merely a reflection of reverse causation [36]. We showed that there was a positive association between obesity (≥25 kg/m²) and low cognitive scores in Japan, but not in Hawaii-based men, and there was significant interaction between Japan and Hawaii. Japan-based men were exposed to a westernized lifestyle and obesity after midlife. The duration of obesity in Japanese men was shorter than in western men; therefore, the relationship between obesity and low cognitive scores could be different in Hawaii and Japan.

Previous studies have shown that light-to-moderate alcohol consumption either reduced or had no effect on the risk of dementia or cognitive impairment; our findings from Hawaii-based men were consistent with this [37 –39]. However, we did not find significant associations among smoking, dietary habits, and low cognitive scores in both Hawaii- and Japan-based men. It is possible that the association between dietary habits and cognitive function were weakened because we focused solely on the frequency of food intake in the analysis, as the SESSA food frequency questionnaire did not define food intake portion size. Moreover, this discrepancy could also be explained by our small sample size, particularly of the Japan-based cohort.

The present study has some limitations. First, there was an approximately 20-year gap in the timing of examination of the two cohorts and we did not account for the birth cohort effect. Moreover, in Japan, medications for cardiometabolic risk factors might have become more efficacious and relevant diagnostics might have become more available in recent years. Second, all study participants were men. Both cohorts consisted of men because the initial goals of the Kuakini HHP and SESSA were to identify cardiovascular disease risk factors that are common in men. Because the risk factors for dementia and cardiovascular diseases are similar, cognitive performance tests were also added. Therefore, the results of the present study cannot be generalized to women. Third, the definition of diabetes was not the same in both cohorts, as the Hawaii-based cohort used stricter criteria. As a result, the association among low cognitive scores, cardiometabolic factors, and cardiovascular diseases might have been underestimated in Japan compared with those in Hawaii. Fourth, some self-reported data were not solicited using the same questions in the two cohorts. Despite our efforts to use the same categories for lifestyle factors and the same definitions for covariates, misclassification might have occurred. Fifth, as the SESSA cohort did not have available information on the APOE ɛ4 allele, one of the genetic risk factors for developing dementia, we could not use it as a covariate in the present analysis. However, as approximately 42%of persons with Alzheimer’s disease do not have the APOE ɛ4 allele [40], the lack of adjustment for the APOE ɛ4 allele was unlikely to affect our findings. Sixth, since the cognitive performance test has been performed only once to date in the SESSA cohort, we were unable to conduct an examination of longitudinal changes in cognitive functioning and associated factors.

In this comparison of genetically similar older Japanese men based in Hawaii and Japan, we have demonstrated that demographic, lifestyle, cardiometabolic, and cardiovascular disease risk factors are associated with cognitive function in both Japanese populations. Moreover, these factors had different strengths of association with cognitive function, depending on lifestyles (i.e., diet and other potentially modifiable factors) in Hawaii and Japan, over different time periods. Hawaii-based Japanese men were exposed to a more westernized lifestyle starting from birth and lasting a lifetime, while Japan-based Japanese men only had such exposure beginning in mid-life, thus lasting for a shorter duration. In countries that are gradually adopting westernized lifestyles, such as East Asian countries, public health efforts to prevent cognitive impairment should consider different risk factors, duration of exposure, and strengths of association with particular outcomes when designing interventions aimed at cognitive aging.

Footnotes

ACKNOWLEDGMENTS

Research for Kuakini HAAS (Honolulu-Asia Aging Study) was supported by the Kuakini Medical Center; the John A. Hartford Center of Excellence in Geriatrics, Department of Geriatric Medicine, John A. Burns School of Medicine, University of Hawaii; the National Institutes of Health (NIH) [Contract N01-AG-4–2149, Grants U01 AG019349, R01AG027060, and R01AG038707 from the National Institute on Aging, and Contract N01-HC-05102 from the National Heart, Lung, and Blood Institute]; and the Office for Research and Development, Department of Veterans Affairs. The views expressed in this paper do not necessarily represent those of the U.S. federal government. Research for SESSA (Shiga Epidemiological Study of Subclinical Atherosclerosis) was supported by Grants-in-aid for Scientific Research [(A) 13307016, (A) 17209023, (A) 21249043, (A) 23249036, (A) 25253046, (A) 15H02528, (C) 23590790], and the Promotion of Joint International Research (Fostering Joint International Research) [15KK0342] from the Ministry of Education, Culture, Sports, Science, and Technology, Japan, by National Institutes of Health in USA [R01HL068200], and by Glaxo-Smith Kline GB.

Authors’ disclosures available online ().

References

World Health Organization (2008) The global burden of disease: 2004 update, WHO Press, Geneva.

Collaborators GBDD (2019) Global, regional, and national burden of Alzheimer’s disease and other dementias, 1990–2016: A systematic analysis for the Global Burden of Disease Study 2016. Lancet Neurol 18, 88–106.

Kim

, Moon

, Popkin

(2000) The nutrition transition in South Korea. Am J Clin Nutr 71, 44–53.

Feng

, Purser

, Zhen

, Duncan

(2011) Less exercise and more TV: Leisure-time physical activity trends of Shanghai elders, 1998-2008. J Public Health (Oxf) 33, 543–550.

Dodge

, Buracchio

, Fisher

, Kiyohara

, Meguro

, Tanizaki

, Kaye

(2012) Trends in the prevalence of dementia in Japan. Int J Alzheimers Dis 2012, 956354.

, Brayne

, Matthews

(2015) Prevalence of dementia in East Asia: A synthetic review of time trends. Int J Geriatr Psychiatry 30, 793–801.

National Academies of Sciences, Engineering, and Medicine; Health and Medicine Division; Board on Health Sciences Policy; Committee on Preventing Dementia and Cognitive Impairment (2017) Preventing Cognitive Decline And Dementia: A Way Forward. Downey A, Stroud C, Landis S, Leshner AI, eds. National Academies Press (US), Washington, DC.

Morris

, Tangney

, Wang

, Sacks

, Bennett

, Aggarwal

(2015) MIND diet associated with reduced incidence of Alzheimer’s disease. Alzheimers Dement 11, 1007–1014.

Kim

, Kim

, Daily

, Park

(2018) High genetic risk scores for impaired insulin secretory capacity doubles the risk for type 2 diabetes in Asians and is exacerbated by Western-type diets. Diabetes Metab Res Rev 34, e2944.

10.

Nakanishi

, Okubo

, Yoneda

, Jitsuiki

, Yamane

, Kohno

(2004) A comparison between Japanese-Americans living in Hawaii and Los Angeles and native Japanese: The impact of lifestyle westernization on diabetes mellitus. Biomed Pharmacother 58, 571–577.

11.

Syme

, Marmot

, Kagan

, Kato

, Rhoads

(1975) Epidemiologic studies of coronary heart disease and stroke in Japanese men living in Japan, Hawaii and California: Introduction. Am J Epidemiol 102, 477–480.

12.

White

, Petrovitch

, Ross

, Masaki

, Abbott

, Teng

, Rodriguez

, Blanchette

, Havlik

, Wergowske

, Chiu

, Foley

, Murdaugh

, Curb

(1996) Prevalence of dementia in older Japanese-American men in Hawaii: The Honolulu-Asia Aging Study. JAMA 276, 955–960.

13.

Ueshima

, Kadowaki

, Hisamatsu

, Fujiyoshi

, Miura

, Ohkubo

, Sekikawa

, Kadota

, Kadowaki

, Nakamura

, Miyagawa

, Okamura

, Kita

, Takashima

, Kashiwagi

, Maegawa

, Horie

, Yamamoto

, Kimura

, Kita

,Access, Groups SR (2016) Lipoprotein-associated phospholipase A2 is related to risk of subclinical atherosclerosis but is not supported by Mendelian randomization analysis in a general Japanese population. Atherosclerosis 246, 141–147.

14.

Kadota

, Miura

, Okamura

, Fujiyoshi

, Ohkubo

, Kadowaki

, Takashima

, Hisamatsu

, Nakamura

, Kasagi

, Maegawa

, Kashiwagi

, Ueshima

, Group SR, Group NDR (2013) Carotid intima-media thickness and plaque in apparently healthy Japanese individuals with an estimated 10-year absolute risk of CAD death according to the Japan Atherosclerosis Society (JAS) guidelines 2012: The Shiga Epidemiological Study of Subclinical Atherosclerosis (SESSA). J Atheroscler Thromb 20, 755–766.

15.

Yamazoe

, Hisamatsu

, Miura

, Kadowaki

, Zaid

, Kadota

, Torii

, Miyazawa

, Fujiyoshi

, Arima

, Sekikawa

, Maegawa

, Horie

, Ueshima

, SESSA Research Group (2016) Relationship of insulin resistance to prevalence and progression of coronary artery calcification beyond metabolic syndrome components: Shiga Epidemiological Study of Subclinical Atherosclerosis. Arterioscler Thromb Vasc Biol 36, 1703–1708.

16.

Fujiyoshi

, Miura

, Ohkubo

, Miyagawa

, Saito

, Miyazawa

, Shiino

, Kadota

, Kadowaki

, Hisamatsu

, Torii

, Takashima

, Tooyama

, Ueshima

(2020) Proteinuria and reduced estimated glomerular filtration rate are independently associated with lower cognitive abilities in apparently healthy community-dwelling elderly men in Japan: A cross-sectional study. J Epidemiol 30, 244–252.

17.

Schatz

, Masaki

, Yano

, Chen

, Rodriguez

, Curb

(2001) Cholesterol and all-cause mortality in elderly people from the Honolulu Heart Program: A cohort study. Lancet 358, 351–355.

18.

Teng

, Hasegawa

, Homma

, Imai

, Larson

, Graves

, Sugimoto

, Yamaguchi

, Sasaki

, Chiu

, et al. (1994) The Cognitive Abilities Screening Instrument (CASI): A practical test for cross-cultural epidemiological studies of dementia. Int Psychogeriatr 6, 45–58; discussion 62.

19.

Hasegawa

(1983) The clinical assessment of dementia in the aged: A dementia screening scale for psychogeriatric patients. In Aging in the Eighties and Beyond, Bergener M, Lehr U, Lang E, Schmitz-Scherzer R, eds. Springer, New York, pp. 207–218.

20.

Folstein

, Folstein

, McHugh

(1975) “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12, 189–198.

21.

Teng

, Chui

(1987) The Modified Mini-Mental State (3MS) examination. J Clin Psychiatry 48, 314–318.

22.

Galanis

, Petrovitch

, Launer

, Harris

, Foley

, White

(1997) Smoking history in middle age and subsequent cognitive performance in elderly Japanese-American men. The Honolulu-Asia Aging Study. Am J Epidemiol 145, 507–515.

23.

American Psychiatric Association (1987) Diagnostic and Statistical Manual of Mental Disorders, Revised Third Edition. American Psychiatric Association, Washington, DC.

24.

Launer

, Masaki

, Petrovitch

, Foley

, Havlik

(1995) The association between midlife blood pressure levels and late-life cognitive function. The Honolulu-Asia Aging Study. JAMA 274, 1846–1851.

25.

Lane

, Nelder

(1982) Analysis of covariance and standardization as instances of prediction. Biometrics 38, 613–621.

26.

Lubs

, Peplies

, Drell

, Bammann

(2018) Cross-sectional and longitudinal factors influencing physical activity of 65 to 75-year-olds: A pan European cohort study based on the survey of health, ageing and retirement in Europe (SHARE). BMC Geriatr 18, 94.

27.

Andreeva

, Alles

, Feron

, Gonzalez

, Sulmont-Rosse

, Galan

, Hercberg

, Mejean

(2016) Sex-specific sociodemographic correlates of dietary patterns in a large sample of French elderly individuals. Nutrients 8, 484.

28.

Murakami

, Hashimoto

, Lee

, Kawakubo

, Mori

, Akabayashi

(2011) Distinct impact of education and income on habitual exercise: A cross-sectional analysis in a rural city in Japan. Soc Sci Med 73, 1683–1688.

29.

Flynn

(1984) The mean IQ of Americans: Massive gains 1932 to 1978. Psychol Bull 95, 29–51.

30.

Lynn

, Hampson

(1986) The rise of national intelligence: Evidence from Britain, Japan and the U.S.A. Pers Individ Dif 7, 23–32.

31.

Dodge

, Zhu

, Lee

, Chang

, Ganguli

(2014) Cohort effects in age-associated cognitive trajectories. J Gerontol A Biol Sci Med Sci 69, 687–694.

32.

Sacuiu

, Gustafson

, Sjogren

, Guo

, Ostling

, Johansson

, Skoog

(2010) Secular changes in cognitive predictors of dementia and mortality in 70-year-olds. Neurology 75, 779–785.

33.

Deckers

, Schievink

SHJ

, Rodriquez

MMF

, van Oostenbrugge

, van Boxtel

MPJ

, Verhey

FRJ

, Kohler

(2017) Coronary heart disease and risk for cognitive impairment or dementia: Systematic review and meta-analysis. PLoS One 12, e0184244.

34.

Lee

, Saver

, Hong

, Wu

, Liu

, Rao

, Ovbiagele

(2014) Cognitive impairment and risk of future stroke: A systematic review and meta-analysis. CMAJ 186, E536–546.

35.

Pedditzi

, Peters

, Beckett

(2016) The risk of overweight/obesity in mid-life and late life for the development of dementia: A systematic review and meta-analysis of longitudinal studies. Age Ageing 45, 14–21.

36.

Bowman

, Thambisetty

, Kuchel

, Ferrucci

, Melzer

(2019) Obesity and longer term risks of dementia in 65-74 year olds. Age Ageing 48, 367–373.

37.

Galanis

, Joseph

, Masaki

, Petrovitch

, Ross

, White

(2000) A longitudinal study of drinking and cognitive performance in elderly Japanese American men: The Honolulu-Asia Aging Study. Am J Public Health 90, 1254–1259.

38.

Neafsey

, Collins

(2011) Moderate alcohol consumption and cognitive risk. Neuropsychiatr Dis Treat 7, 465–484.

39.

, Wang

, Wan

, Tan

, Li

, Tan

, Yu

(2017) Alcohol consumption and dementia risk: A dose-response meta-analysis of prospective studies. Eur J Epidemiol 32, 31–42.

40.

Bird

(1993) Alzheimer disease overview. In GeneRe- views((R)), Adam MP, Ardinger HH, Pagon RA, Wallace SE, Bean LJH, Stephens K, Amemiya A, eds. University of Washington, Seattle, WA.