Background
The trend towards a rapidly aging society is a manifest global issue that brings subsequent health challenges worldwide. Regardless of geographic location, developed countries and developing countries alike will both face expected increases in health care demand and related socioeconomic burdens [
1]. In 2015, there were 617.1 million people (9 percent of the world population) aged 65 and older. By 2030, this population will increase to approximately 1 billion, equivalent to 12 percent of the predicted total world population. By 2050, this older population is estimated to be 1.6 billion or 17 percent of the entire global population [
2]. According to a report from the US Economics and Statistics Administration, Department of Commerce this rapid growth of an aged society is also being observed in Asia, with an estimation of Asia’s older population almost tripling in size from 341.4 million in 2015 to 975.3 million in 2050 [
2].
An aged society faces numerous health issues. Advances in human civilization have increased human life expectancy, but the subsequent health care problems that accompany ageing will cause a heavy care burden [
3]. For example, cognitive decline is a well-recognized problem in older adults [
3,
4]. When mild cognitive impairment in older adults progresses to dementia, it typically causes disability and is related to a higher mortality risk [
5‐
7]. Cognitive dysfunction is associated with a poor quality of life in older adults [
8]. Thus, finding ways to decelerate or even stop cognitive decline has become an important issue nowadays. In line with the recommendation from the United Nations “Decade of Healthy Ageing (2021–2030)” report, taking action to prevent cognitive function decline is important in achieving successful aging [
9].
There is numerous evidence that age plays a crucial role in the cognitive function decline process [
10,
11]. However, there are still some modifiable factors that are associated with impaired cognitive function [
12]. Studies have linked lower cognitive performance and risk of dementia with diabetic individuals [
12‐
15]. Obesity, on the other hand, has been discussed with controversy. Obesity has been reported to be associated with cognitive decline in some studies [
16], while others propose the opposite point of view [
17,
18]. Additionally, a history of myocardial infarction [
17], hypertension [
19], stroke, and depression [
19] have all been found to be independently associated with a higher cognitive impairment prevalence among different population groups. Lifestyle and nutrition factors discussed in some studies – such as the consumption of different diets, variety of fruit or vegetable intake, or beverage consumption have shown divergent results [
17,
20]. Physical activity, as compared to no exercise, was associated with a lower risk of cognitive impairment, Alzheimer’s disease, and dementia of any type [
21‐
25]. Other independent risk factors for cognitive impairment, like tobacco and alcohol use, have also been studied [
26].
The WHO defines [
27] a hyper-aged society as a society where the aged population accounts for more than 20% of the total population. In the report of the “Population Projections for R.O.C. (Taiwan): 2016–2061” published by the National Development Council, Taiwan is expected to become a hyper-aged society by 2026 [
27]. This aging rate is faster than the rates of other developed countries. Thus, prevention of cognitive impairment is an important health issue, especially in the rapidly ageing societies in Asia. However, studies examining the modifiable factors associated with cognitive decline in community-dwelling and relatively healthy Taiwanese geriatrics are still lacking [
22,
28,
29]. Therefore, this study aimed to investigate the factors associated with cognitive function in community-dwelling Taiwanese older adults patients aged 65 or older.
Results
Our study included 4,578 participants aged 65 years or older (Fig.
1). The average age was 73.5 ± 5.8 years old. Male participants accounted for 27.1% of the total number of participants. The average BMI was 24.4 ± 3.4 (kg/m
2), and the WC was 83.2 ± 10.0 (cm). The mean blood pressure was systolic blood pressure (SBP) = 136.4 ± 19.9(mmHg) and diastolic blood pressure (DBP) = 71.1 ± 11.1(mmHg). Other baseline demographic data, including chronic disease history, social habits, and biochemistry lab data, are summarized in Table
1. The average SPMSQ score was 0.2 ± 0.9 (Table
1). The distribution of participants’ SPMSQ scores is shown in Table S
1.
Table 1
Demographic characteristics of study subjects
Age (years) | 73.5 ± 5.8 |
Gender (male) | 1240 (27.1%) |
Body height (cm) | 155.2 ± 7.6 |
Body weight (kg) | 58.9 ± 9.8 |
Body mass index (kg/m2) | 24.4 ± 3.4 |
Systolic blood pressure (mmHg) | 136.4 ± 19.9 |
Diastolic blood pressure (mmHg) | 71.1 ± 11.1 |
Pulse rate (bpm) | 73.9 ± 11.6 |
Waist circumference (cm) | 83.2 ± 10.0 |
Hypertension (yes) | 2300 (50.2%) |
Diabetes Mellitus (yes) | 698 (15.2%) |
Hyperlipidemia (yes) | 1007 (22.0%) |
Depression (yes) | 30(0.7%) |
Cardiovascular disease (yes) | 787(17.2%) |
Osteoporosis (yes) | 58 (1.3%) |
Hyperthyroidism (yes) | 38 (0.8%) |
Smoking in 6 months(yes) | 157 (3.4%) |
Alcohol in 6 months(yes) | 524 (11.4%) |
Betel nut in 6 months(yes) | 9 (0.2%) |
Exercise in 6 months(yes) | 2820 (61.6%) |
AC sugar (mg/dL) | 106.6 ± 23.2 |
Total protein (g/dL) | 7.3 ± 0.5 |
Albumin (g/dL) | 4.2 ± 0.4 |
GOT (U/L) | 25.5 ± 14.5 |
GPT (U/L) | 22.4 ± 16.0 |
Creatinine (mg/dL) | 1.0 ± 0.4 |
Total cholesterol (mg/dL) | 198.7 ± 34.5 |
Triglyceride (mg/dL) | 118.3 ± 67.1 |
High density lipoprotein (mg/dL) | 58.2 ± 16.1 |
Uric acid (mg/dL) | 5.6 ± 1.4 |
Hemoglobin (g/dL) | 13.2 ± 1.3 |
SPMSQ score (scores) | 0.2 ± 0.9 |
Overall, participants without cognitive impairment (SPMSQ < 3) made up 97.8% of the whole study group, with the majority of the total participants (88.6%) scoring a perfect score of 0 (Table S
1). In total, there were 103 participants with cognitive impairment, comprising 2.3% of the study group. Table
2 presents the differences noted between the normal cognitive function group (SPMSQ < 3) and the cognitive impairment group (SPMSQ≧3). There was a statistically significant difference in age between the group with cognitive impairment, at 79.7 ± 7.5 years old on average, compared with the normal cognitive function group, at 73.3 ± 5.7 years old. Participants in the cognitive impairment group were typically older females with a larger waist circumference. They were more likely to have DM, and less likely to have hyperlipidemia, exercise regularly, or drink alcohol. Lower levels of albumin, high-density lipoprotein (HDL) and hemoglobin were also noted in the cognitive impairment group. Although there were discrepancies in body height and body weight between the two groups, the calculated BMI showed no significant difference (Table
2).
Table 2
Comparisons of cognitive impairmentgroup (SPMSQ≧3) and normal cognitive function group (SPMSQ < 3)
Age (years) | 79.7 ± 7.5 | 73.3 ± 5.7 | < 0.001 |
Gender (male) | 15 (14.6%) | 1225 (27.4%) | 0.004 |
Body height (cm) | 150 ± 7.6 | 155.3 ± 7.6 | < 0.001 |
Body weight (kg) | 54.4 ± 9.5 | 59.0 ± 9.8 | < 0.001 |
Body mass index (kg/m2) | 24.2 ± 3.9 | 24.4 ± 3.4 | 0.546 |
Systolic blood pressure (mmHg) | 138.8 ± 23.4 | 136.4 ± 19.8 | 0.298 |
Diastolic blood pressure (mmHg) | 69.3 ± 12.3 | 71.2 ± 11.1 | 0.093 |
Pulse rate (bpm) | 76.1 ± 13.3 | 73.9 ± 11.5 | 0.098 |
Waist circumference (cm) | 85.6 ± 10.5 | 83.1 ± 10.0 | 0.014 |
Hypertension (yes) | 59 (57.3%) | 2241 (50.1%) | 0.148 |
Diabetes Mellitus (yes) | 26 (25.5%) | 672 (15%) | 0.004 |
Hyperlipidemia (yes) | 12 (11.7%) | 995 (22.2%) | 0.010 |
Depression (yes) | 0 (0%) | 30 (0.7%) | 0.404 |
Cardiovascular disease (yes) | 24 (23.3%) | 763 (17.1%) | 0.096 |
Osteoporosis (yes) | 2 (1.9%) | 56 (1.3%) | 0.536 |
Hyperthyroidism (yes) | 0 (0%) | 38 (0.8%) | 0.348 |
Smoking in 6 months(yes) | 5 (4.9%) | 152 (3.4%) | 0.422 |
Alcohol in 6 months(yes) | 5 (4.9%) | 519 (11.6%) | 0.034 |
Betel nut in 6 months(yes) | 1 (1.0%) | 8 (0.2%) | 0.073 |
Exercise in 6 months(yes) | 31 (30.1%) | 2789 (62.3%) | < 0.001 |
AC sugar (mg/dL) | 110.2 ± 24.9 | 106.5 ± 23.2 | 0.138 |
Total protein (g/dL) | 7.3 ± 0.6 | 7.3 ± 0.5 | 0.529 |
Albumin (g/dL) | 4.0 ± 0.2 | 4.2 ± 0.4 | < 0.001 |
GOT (U/L) | 29.6 ± 35.3 | 25.5 ± 13.6 | 0.240 |
GPT (U/L) | 27.5 ± 48.8 | 22.3 ± 14.3 | 0.285 |
Creatinine (mg/dL) | 1.0 ± 0.2 | 1.0 ± 0.4 | 0.448 |
Total cholesterol (mg/dL) | 194.9 ± 32.9 | 198.8 ± 34.5 | 0.247 |
Triglyceride (mg/dL) | 127.5 ± 54.7 | 118.1 ± 67.3 | 0.161 |
High density lipoprotein (mg/dL) | 53.8 ± 15.4 | 58.3 ± 16.1 | 0.005 |
Uric acid (mg/dL) | 5.5 ± 1.8 | 5.6 ± 1.4 | 0.881 |
Hemoglobin (g/dL) | 12.8 ± 1.4 | 13.2 ± 1.3 | 0.006 |
A multivariate logistic regression analysis was done on the variables mentioned above that presented significant differences between the two groups (Table
3). Age (odds ratio (OR) = 1.16, 95% confidence interval (CI) = 1.13, 1.20) and DM (OR = 1.70, 95% CI = 1.03, 2.82) were found to be positively associated with cognitive impairment. In contrast, male gender (OR = 0.39, 95% CI = 0.21, 0.72), hyperlipidemia (OR = 0.47, 95% CI = 0.25, 0.89), exercise (OR = 0.44, 95% CI = 0.34, 0.56), albumin (OR = 0.37, 95% CI = 0.15, 0.88), and HDL level (OR = 0.98, 95% CI = 0.97, 1.00) were negatively associated with the cognitive impairment group. The former associations of waistline, recent alcohol intake, and hemoglobin level as stated previously in Table
2 did not remain significantly different after multivariate adjustment (Table
3).
Table 3
Multivariate logistic regression of factors associated with cognitive impairment(SPMSQ≧3)
Age (years) | 1.16 (1.13, 1.20)* | < 0.001 |
Gender (male) | 0.39 (0.21, 0.72)* | 0.003 |
Waist circumference (cm) | 1.01 (0.99, 1.03) | 0.440 |
Diabetes Mellitus (yes) | 1.70 (1.03, 2.82)* | 0.038 |
Hyperlipidemia (yes) | 0.47 (0.25, 0.89)* | 0.020 |
Alcohol in 6 months(yes) | 0.80 (0.31, 2.08) | 0.646 |
Exercise in 6 months(yes) | 0.44 (0.34, 0.56)* | < 0.001 |
Albumin (g/dL) | 0.37 (0.15, 0.88)* | 0.024 |
High density lipoprotein (mg/dL) | 0.98 (0.97, 1.00)* | 0.024 |
Hemoglobin (g/dL) | 1.07 (0.91, 1.25) | 0.448 |
In order to obtain a broader perspective, we did a further analysis by including in our data the samples that were originally excluded due to previous missing values. The characteristics of the total 10,944 participants are summarized in Table S
2. The comparison of the two groups (the normal cognitive function and the cognitive impairment group) is shown in Table S
3. Multivariate logistic regression analysis of the factors associated with cognitive impairment is shown in Table S
4. Exercise in six months, higher albumin and HDL levels were significantly associated with a lower risk of cognitive impairment. Increasing age was found to be significantly associated with cognitive impairment. Although not significantly found, male gender, DM and hyperlipidemia also presented a higher risk of cognitive impairment. Overall, the results of the multivariate logistic regression in Table S
4 were comparable to our main results in Table
3. Both seem to have render similar conclusions.
Discussion
Our study is an observational, cross-sectional study examining the modifiable factors associated with cognitive impairment in Taiwan’s community-dwelling older adults. The factors we observed to be associated with increased risk of cognitive impairment were age and diabetes mellitus. Male gender, hyperlipidemia, exercise, albumin level, and HDL level were related to a lower risk of cognitive impairment.
A prevalence of 2.3% older adults with cognitive impairment was observed in our study, which was less than in the other studies. The prevalence of cognitive impairment was 22.2% in a previous Taiwanese study [
37]. In Asia, the prevalence of cognitive impairment ranges from 13.29% to 21.5% [
17,
38]. This disparate finding could be due to our study participants being recruited from the Annual Geriatric Health Examinations Program, which on average comprises only 20 to 30% of all Taipei and New Taipei seniors, with variation for each year, thus there still remained a residual selection bias. Taipei and New Taipei both represent a highly urbanized area. Previous studies analyzing the urban–rural differences in the prevalence of mild cognitive impairment (MCI) of older adults in Taiwan have suggested a lower prevalence of MCI in the urban community than in the rural one [
29,
39]. In addition, seniors that voluntarily partake in the yearly geriatric health exam tend to have better physical and mental health, including autonomy in activities of daily living, higher social participation, and little or no disability. However, even with the small proportion of participants that were found to be have cognitive impairment, we were still able to find significant differences in characteristics between the SPMSQ≧ 3 and SPMSQ < 3 groups.
We observed a higher prevalence of cognitive impairment in females than in males, which is consistent with a previous cross-sectional study done exploring age and sex-specific prevalence among older adults with mild cognitive impairment [
23]. Our study also demonstrated an association between cognitive impairment and ageing. This is consistent with previous studies that have shown ageing to be a key risk factor in cognitive decline [
10,
11]. A history of diabetes mellitus was also shown to have a correlation with our cognitive impairment group. Diabetes mellitus as a risk factor for cognitive impairment has been studied previously [
14,
15]. Suggested mechanisms underlying this relationship include neurotoxic effects on brain cells, increased production of Reactive Oxygen Species (ROS), and accelerated brain microangiopathy development when hyperglycemia is present [
40,
41].
In this study, presence of a history of hyperlipidemia, regular exercise, higher albumin level, and higher HDL level were related to a lower risk of cognitive impairment. Hyperlipidemia in the role of cognitive decline has been up to debate, as previous studies have shown discrepant results. Some studies have shown that elevations in total cholesterol and low-density lipoprotein cholesterol (LDL) were related to decreased cognitive performance [
42,
43] and mild cognitive impairment [
44]. However, other studies had null results and did not find that total cholesterol [
45] or a history of hyperlipidemia [
46] was linked to cognitive decline. Our data revealed that a history of hyperlipidemia and higher levels of HDL were associated with the normal cognitive function group. This is in line with previous research that has shown low HDL to be detrimental to cognition [
47,
48]. Cholesterol dysregulation has been implicated in the development of neurodegenerative diseases, such as Alzheimer’s disease, Parkinson’s disease and Huntington’s disease. An animal study done on Huntington’s disease mice found decreased cholesterol synthesis in the striatum of the brain. Injection of cholesterol directly into the striatum ameliorated some motor symptoms and prevented cognitive decline [
49]. Cholesterol is crucial in ensuring normal brain function, as it is an important component of the cell membrane [
50]. Increase in cholesterol over time was associated with better cognition in a longitudinal study [
51]. A previous study suggested that a lower total cholesterol could be used as a marker to predict cognitive decline in older adults [
52]. Could a higher total cholesterol be an indicator of non-frailty and thus be protective for cognitive function? Could a history of hyperlipidemia in the normal cognitive function group suggest that lipid lowering agents play a role in improving cognition? Prospective cohort studies in the past examining the role of lipid lowering agents in cognitive decline did not find that the medication was preventive against cognitive decline or dementia [
53,
54]. Thus, the precise role of hyperlipidemia on cognitive function needs to be explored further.
Our participants in the cognitive impairment group were less likely to be physically active. Overall, 62.3% of the participants without cognitive impairment compared to only 30.1% of the cognitive impairment group had an exercise habit within the past six months. This is consistent with previous studies that have reported a positive association between exercise and cognitive function [
23,
24]. A meta-analysis of 15 studies concluded that physical activity significantly protected against cognitive decline [
25]. Current and consistent exercise habits among Taiwanese seniors led to better cognitive performance on the SPMSQ over the course of an eight-year follow-up study [
22].
Our results indicated that participants with a higher albumin level had a lower risk of cognitive impairment. Lower albumin levels have been reported to be correlated with poor cognitive performance in older adults [
55‐
57]. In a previous study done in Japan [
58], a positive association between the serum albumin/globulin ratio (A/G ratio) and cognitive function was found in 70-year-old and 80-year-old participants. Similar to our study, the Asian study participants also had normal albumin levels. Since albumin is seen as a marker for nutritional status and inflammation, a low albumin level may indicate malnutrition, chronic hepatitis, nephrotic syndrome or an inflammation status. Thus, higher albumin levels may be associated with healthier individuals and possibly having better cognitive function as well.
Evidence exploring the relationship between waist circumference and cognitive impairment has had conflicting results. Some studies have reported that an increase in waist circumference was associated with cognitive decline risk [
59,
60]. However other studies have shown that greater waist circumference was associated with slower cognitive decline [
61,
62] or produced null results similar to our own [
38]. Further studies are warranted before a consensus is reached regarding the effect of waist circumference on cognitive performance. We also did not observe any significant association between alcohol drinking or hemoglobin levels with cognitive impairment. In addition, SBP, DBP and a clinical history of hypertension were not found to significantly different between the cognitive impairment group and normal cognitive function group. Whereas in a previous study, high blood pressure, hypertension, uncontrolled blood pressure was associated with poorer cognitive function when compared with those whom had normal blood pressure in participants aged 70 and older [
63]. However, this association was not shown in participants aged 60 to 69. In a study done in Japan, high SBP was found to be significantly correlated with reduced cognitive functioning in 70-year-old participants, but not in participants aged 80 years old [
64]. Both of these studies suggest that high blood pressure may be a risk factor for cognitive decline in subjects around 70-year-old, but the results were not consistent amongst other age groups. Blood pressure readings may be affected by the clinic environment, emotional stress and the well-being of the person. Well-controlled hypertension relies on good compliance of antihypertensive drugs, smoking abstinence, regular exercise, and a well-balanced diet. There may have been some confounding factors in play that we did not take in account of and thus, our study did not find any association between hypertension and cognitive function. More studies are needed to verify the relationship between blood pressure and cognition.
Our study focusing on Taiwanese older adults aimed to find out the modifiable risk factors of cognitive impairment to mitigate the subsequent care burden of a rapidly-aging society. Strengths of our study include having a relatively large number of participants and the use of a rigorous and standardized protocol for data collection. The data collected for our study from the Annual Geriatric Health Examinations Program was methodically done with trained nurses and doctors gathering the participants’ information and carrying out the anthropometric measurements. Nevertheless, some limitations should be considered. First, although the SPMSQ is a tool used in cognitive function decline screening, it cannot be used to diagnose dementia. The participants with SPMSQ scores ≧ 3 in our study were identified as having cognitive impairment, but as for its use in screening for dementia, this remains uncertain. Second, other potential risk factors like educational level, income level, diet habits, and history of other comorbidities were not analyzed in this study. Furthermore, participants’ use of anti-hypertensive, anti-hyperglycemic, and lipid-lowering agents was also not examined. Although we included exercise, smoking, and alcohol consumption in our study, more detailed exploration of the type of exercise and its intensity as well as the type and amount of alcohol and cigarette consumption could be useful for future studies and analyses. Finally, only a small percentage (2.3%) of our participants were found to have a score of SPMSQ≧3. This large disparity in sample size between the SPMSQ≧ 3 and SPMSQ < 3 groups could be due to that most community-dwelling older adults who voluntarily participate in the walk-in Annual Geriatric Health Examinations Program usually have little or no disability and are relatively healthy. This may have introduced bias in our results and limits the generalizability of our findings. Perhaps, a different cognitive screening tool with a wider scale distribution such as the Montreal Cognitive Assessment (MoCA) test could be considered for future studies. Further studies are still required to identify the risk factors pertaining to cognitive impairment.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.