Prevalence and Progression of Subjective Cognitive Decline Among Rural Chinese Older Adults: A Population-Based Study

Abstract

Background:

Few community-based studies have examined occurrence and progression of subjective cognitive decline (SCD).

Objective:

To investigate prevalence and progression of SCD among rural-dwelling Chinese elderly people.

Methods:

This cohort study included 2,488 cognitively unimpaired adults (age≥65 years) who were examined at baseline (2014-2015) and followed in 2018. Demographic, epidemiological, clinical, and neuropsychological data were collected via in-person interviews and clinical examinations following a structured questionnaire. At baseline, SCD was assessed using the self-rated Ascertain Dementia 8-item Questionnaire. At follow-up, Alzheimer’s disease (AD) and vascular dementia (VaD) were clinically diagnosed following the international criteria. Data were analyzed using logistic regression models.

Results:

The prevalence of SCD was 40.07%. SCD at baseline was associated with the multivariable-adjusted odds ratio (OR) of 1.51 (95% confidence interval 1.10–2.07) for incident cognitive impairment, no dementia (CIND) and 3.11 (1.64–5.93) for incident AD. Among people with SCD at baseline, the multivariable-adjusted OR of incident CIND was 0.55(0.32–0.96) for hyperlipidemia; the multivariable-adjusted OR of incident AD was 1.21 (1.14–1.30) for older age, 0.32 (0.12–0.88) for high education, 2.60 (1.11–6.08) for carrying APOE ɛ4 allele, and 0.34 (0.13–0.86) for high social support, whereas the multivariable-adjusted OR of incident VaD was 6.30 (1.71–23.18) for obesity.

Conclusion:

SCD affects over 40% of rural-dwelling cognitively unimpaired older adults in China. SCD is associated with accelerated progression to CIND and AD. Older age, lack of school education, APOE ɛ4 allele, and low social support are associated with an increased risk of progression from SCD to AD, whereas obesity is related to accelerated progression to VaD.

Keywords

Alzheimer’s disease dementia population-based study prevalence risk factors subjective cognitive decline

INTRODUCTION

In 2014, the Subjective Cognitive Decline Initiative (SCD-I) Working Group defined subjective cognitive decline (SCD) as a diagnostic entity characterized by self-perceived persistent decline in cognitive capacity with no evidence of objective cognitive impairment [1]. SCD has been increasingly recognized as an early marker of subsequent cognitive impairment and may represent an “at-risk” stage for dementia. Thus, identifying risk and protective factors that contribute to fast progression from SCD to cognitive impairment and dementia may facilitate the development of intervention strategies to slow cognitive decline or conversion to dementia.

Several population-based studies have reported that the prevalence of SCD among cognitively unimpaired older adults ranges from 17% to 69% [2 –5]. This wide range of SCD prevalence is due in part to differences in demographic characteristics of the study populations and the methodologies used to define SCD. Notably, despite the relatively high prevalence of cognitive impairment in rural residents [6], most of the previous studies have targeted urban populations with relatively high education and socioeconomic status [3, 4], while prevalence data of SCD from rural populations are scarce. This is important because compared with urban populations, rural residents usually have low socioeconomic status (i.e., low income and low education) and different healthcare systems, and thus may have different epidemiological features and related factors of cognitive disorders in older adults [7].

There is currently no consensus on the standard approach to define SCD. Most studies have used a single or double questions on memory decline and associated concerns or carried out clinical interviews to define SCD, although others have assessed SCD through a specific questionnaire [8, 9]. In addition, most of the previous studies on SCD did not include an objective assessment of cognitive function when defining SCD [10]. This is important because objective neuropsychological testing is mandatory to exclude objective cognitive impairment in individuals with SCD [1].

Previously, community-based longitudinal studies have suggested that SCD is associated with increased risk for mild cognitive impairment (MCI) (odds ratio ranges from 1.42 to 2.03) and dementia (odds ratio ranges from 1.05 to 4.26) [5]. Advanced age [11] and apolipoprotein E (APOE) ɛ4 allele [12] have been associated with accelerated progression of SCD. However, evidence from longitudinal studies that explore other demographic features, lifestyles, cardiovascular risk factors, and social support for SCD progression is lacking.

Thus, in this community-based study of rural Chinese older adults who were living in western Shandong province, we sought to 1) investigate the prevalence of SCD; 2) examine the associations of SCD with subsequent development of cognitive impairment, no dementia (CIND), Alzheimer’s disease (AD), and vascular dementia (VaD); and 3) explore the risk and protective factors associated with progression from SCD to CIND, AD, and VaD.

METHODS

Study participants

This is a population-based prospective cohort study. The baseline participants of this study were derived from the Shandong Yanggu Study of Aging and Dementia (SYS-AD) that engaged rural adults who were aged≥65 years and residing in Yanlou town, Yanggu county, western Shandong province in 2014, as previously described [13, 14]. Briefly, at baseline (August 2014-September 2015), 3,189 participants underwent a series of in-person interviews, clinical examinations, and neuropsychological screening tests. Of these, 701 were excluded due to prevalent dementia (n = 199), prevalent CIND (n = 293), or incomplete information to determine cognitive status (n = 209). Therefore, the analytical sample for determining the prevalence of SCD included 2,488 participants who were free from CIND and dementia at baseline; of these, 997 were defined with SCD. Among the 1,491 participants who did not have SCD in the baseline survey, 393 were not included in the follow-up study due to death (n = 102), loss of contacts (n = 264), and incomplete follow-up data (n = 27), thus, 1,098 persons without SCD at baseline underwent the follow-up examination in 2018. Of the 997 participants who were defined as having SCD at baseline, 252 were not included in the 2018 follow-up examination because 80 persons died before the 2018 follow-up examination, 147 lost contact or refused to participate, and 25 had incomplete data on follow-up, leaving 745 participants for the analysis regarding progression from SCD to CIND and dementia. Of the 745 participants with SCD at baseline, 97 were diagnosed with CIND, 43 with AD, 13 with VaD, and 1 with other type of dementia at the 2018 follow-up examination. In addition, we excluded 1 person who was diagnosed with other type of dementia, leaving 744 participants for the analysis of risk and protective factors associated with the progression from SCD to AD and VaD. Finally, 688 of the 745 participants did not progress to dementia at the time of follow-up, and this sample was used to analyze the risk and protective factors associated with progression from SCD to CIND. Figure 1 presents the flowchart of the study participants.

Fig. 1

Flowchart of the study participants, 2014-2015 to 2018. SYS-AD, Shandong Yanggu Study of Aging and Dementia; SCD, subjective cognitive decline; CIND, cognitive impairment, no dementia; AD, Alzheimer’s disease; VaD, vascular dementia.

The Ethics Committee on Human Experiments of Shandong Provincial Hospital affiliated to Shandong University in Jinan, Shandong, China reviewed and approved the protocol for the SYS-AD study. Written informed consent was obtained from all the participants, or a proxy (usually a guardian or family member) if the participants were unable to provide the consent.

Data collection

At baseline, trained staff collected data through face-to-face interviews, clinical examinations, neuropsychological testing, and laboratory tests. The two-phase examination procedure was implemented during the baseline survey, as previously reported [13, 14]. In brief, at phase I (August-December 2014), we collected data on sociodemographic characteristics (e.g., age, sex, and education), lifestyles (e.g., smoking and alcohol intake), medical history (e.g., chronic health conditions), clinical symptoms, and use of medication. Height and weight were measured in light clothing and with no shoes, and body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters (kg/m²), and obesity was defined as BMI≥28 kg/m². The Chinese Social Support Rating Scale (SSRS) was administered to examine social support [15]. The SSRS was a 10-item scale comprising 3 subscales, i.e., subjective support, objective support, and utility of support, with the total score ranging from 0 to 40. The SSRS score≥20 indicates satisfactory social support. Sleep quality and disturbances in the past 1 month were assessed using the Pittsburgh Sleep Quality Index (PSQI), which is a sum of scores for the 7 domains (i.e., subjective sleep quality, sleep latency, sleep duration, sleep efficiency, sleep disturbance, use of sleeping medication, and daytime dysfunction). The total PSQI score ranged from 0 to 21, with a higher score indicating worse sleep quality. Poor sleep quality is defined as a total PSQI score≥7 [16]. The administration of the Chinese version of the Mini-Mental State Examination (MMSE) and detailed definitions of cardiovascular risk factors were described previously [14]. Subjective cognitive decline was assessed using the self-rated Ascertain Dementia 8-item Questionnaire (AD8) [17]. At phase II (December 2014-September 2015), for participants with the MMSE score≤24 or AD8 score≥2 at phase I, we further assessed comprehensive cognitive function using a neuropsychological test battery and depressive symptoms with the 17-item Hamilton Depression Rating Scale (HAMD-17) (n = 783). The total HAMD-17 ranged from 0 to 54, with a higher score indicating more severe depression. HAMD-17 score≥8 indicates the presence of depression [18].

Follow-up survey

The follow-up assessments were conducted in March-September 2018 as part of the baseline evaluations of the Multimodal Interventions to Delay Dementia and Disability in Rural China (MIND-China) [19]. During the follow-up examination, data on sociodemographic features, lifestyles, cardiometabolic risk factors, clinical conditions, and comprehensive neuropsychological testing were collected. Peripheral blood samples were taken, and genomic DNA was extracted from venous blood leukocytes. APOE genotype was detected by TaqMan single nucleotide polymorphism method [20].

Diagnosis of CIND and dementia

The same three-step procedure was used to diagnose dementia by senior neurologists specialized in dementia care and treatment at both baseline and follow-up examinations [19]. In brief, trained medical staff first conducted the face-to-face interviews and clinical examinations to collect data on health history, cognitive function, and activities of daily living (Chinese version), and recorded all information following the structured questionnaires. Then, neurologists reviewed all the records and made a preliminary diagnosis for participants who were suspected to have dementia. Finally, senior neurologists conducted the second face-to-face interview with those who were suspected to have dementia and their caregivers/informants or who had insufficient data for making a diagnosis of dementia status, and reassessed their medical history, cognitive status, daily living ability, and whenever available, neuroimaging data, and made a final diagnosis of dementia according to the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (DSM-IV) criteria [21]. We further divided dementia into AD and VaD according to the NIA-AA criteria for probable AD [22] and the NINDS-AIREN criteria for probable VaD [23], respectively. CIND was defined as having subjective cognitive complaints (AD8 score≥2) and global cognitive impairment (MMSE score≥1 standard deviation [SD] below the age- and education-specific mean score) as previously reported [14].

Assessment of SCD

We defined SCD based on principles outlined in the SCD-I recommendations [1]: 1) a self-experienced consistent decline in cognitive capacity compared to a previous normal cognitive status; and 2) normal age-adjusted and education-adjusted performance on cognitive tests after standardization. Participants with major psychiatric diseases, neurological conditions, and other systemic diseases that may cause cognitive impairment were excluded. Among participants who were free of CIND or dementia, self-rated AD8 [17] was used to identify SCD. The AD8 test consists of 8 items that rate changes in memory, problem-solving abilities, orientation, and daily activities. In the AD8 test, subjects were asked about changes in self-perception of cognitive problems in the past year compared with their previous normal state. The number of items that scored as “Yes, a change” was summed to obtain the total AD8 score (score range, 0–8). A higher AD8 score indicates poorer subjective cognitive ability. People with the AD8 score≥2 but without objective cognitive impairment were considered to have SCD [10].

Statistical analysis

Baseline characteristics of study participants by SCD status are described. Continuous variables were analyzed with the Student’s t-test and categorical variables with the χ² test. The crude and the age- and sex-specific prevalence of SCD among cognitively unimpaired participants was calculated. Multinomial logistic regression models were used to examine the risk of progression to CIND, AD, and VaD associated with baseline SCD. Binary and multinomial logistic regression models were used to explore baseline various factors associated with the conversion from SCD to CIND, AD, and VaD. We reported the main results from 2 models: Model 1 was controlled for age, sex, and education; and Model 2 was additionally controlled for obesity, APOE genotype, smoking, alcohol drinking, hypertension, diabetes, hyperlipidemia, ischemic heart disease, stroke, SSRS, PSQI, and HAMD scores, which were able to find 1) whether the results were present independent of major demographic confounders and 2) to what extent the results were affected by additional potential confounder other than demographic factors. All analyses were performed using IBM SPSS Statistics for Windows, Version 26.0 (IBM Corp, Armonk, NY, USA). Statistical significance was set at two-tailed p≤0.05.

RESULTS

Baseline characteristics of study participants

When compared the total sample of participants in the SYS-AD examination in 2014 (n = 3,189) with those of the MIND-China survey in 2018 (age≥65 years, n = 5,246), there was no significant difference in the mean age (71.79 versus 71.74, p = 0.497) and distribution of sex (female: 57.54% versus 57.17%, p = 0.736), but the SYS-AD participants had a higher proportion of lack of school education (44.12% versus 40.68; p = 0.007) [19].

The mean age of the 2,488 baseline participants included in this analysis was 71.09 (SD = 5.34) years, and 55.18% were women (Table 1). Compared to those without SCD, participants with SCD were older, more likely to be women, less educated, less likely to smoke and drink alcohol, and more likely to be a farmer and have ischemic heart disease, a lower SSRS score, a higher PSQI score, and a lower MMSE score (p < 0.05). There was no significant difference between the 2 groups in the proportions of obesity, APOE ɛ4 allele, hypertension, diabetes, hyperlipidemia, and stroke, and the mean HAMD score (Table 1).

Table 1

Baseline characteristics of study participants in the total sample and by subjective cognitive decline (n = 2,488)

Characteristics^*	Total sample	Subjective cognitive decline at baseline
	(n = 2,488)	No (n = 1,491)	Yes (n = 997)	p
Age (y), mean (SD)	71.09 (5.34)	70.55 (4.97)	71.91 (5.75)	< 0.001
Women, n (%)	1,373 (55.18)	741 (49.70)	632 (63.39)	< 0.001
Educational level, n (%)				< 0.001
No school education	992 (39.87)	481 (32.26)	511 (51.25)
Elementary school	1,099 (44.17)	710 (47.62)	389 (39.02)
Middle school or above	397 (15.96)	300 (20.12)	97 (9.73)
Obesity, n (%)	451 (20.41)	258 (19.62)	193 (21.56)	0.266
APOEɛ4, n (%)	284 (15.56)	166 (15.26)	118 (16.01)	0.663
Ever smoking, n (%)	937 (37.71)	616 (41.37)	321 (32.23)	< 0.001
Drinking, n (%)	912 (36.70)	599 (40.23)	313 (31.43)	< 0.001
Comorbidity, n (%)
Hypertension	1,771 (71.33)	1,045 (70.28)	726 (72.89)	0.158
Diabetes mellitus	324 (13.04)	188 (12.63)	136 (13.64)	0.465
Hyperlipidemia	650 (26.18)	372 (25.03)	278 (27.88)	0.113
Ischemic heart disease	472 (19.02)	253 (17.05)	219 (21.97)	0.002
Stroke history	230 (9.27)	124 (8.36)	106 (10.63)	0.055
Questionnaire, mean (SD)
SSRS score	24.68 (4.03)	25.05 (4.02)	24.13 (3.98)	< 0.001
PSQI score	5.58 (3.88)	4.83 (3.55)	6.68 (4.08)	< 0.001
MMSE score	22.74 (4.61)	23.64 (4.52)	21.40 (4.41)	< 0.001
HAMD score	4.37 (3.54)	4.37 (3.83)	4.36 (3.41)	0.961

APOE, apolipoprotein E gene; SSRS, Social Support Rating Scale; PSQI, Pittsburgh Sleep Quality Index; MMSE, Mini-Mental State Examination; HAMD, Hamilton Depression Scale. ^*The number of subjects with missing value was 278 in obesity, 663 in APOE genotype, 3 in smoking, 3 in alcohol drinking, 5 in hypertension, 3 in diabetes, 5 in hyperlipidemia, 7 in coronary heart disease, 7 in stroke, 57 in SSRS, 238 in PSQI, and 1705 in HAMD score.

Prevalence of SCD

Of the 2,488 cognitively unimpaired elderly people at baseline, 997 were defined as having SCD. The overall prevalence of SCD was 40.07% (95% CI 38.15% –42.00%), with the prevalence being higher in women than in men (46.03% versus 32.74%, p < 0.001); controlling for age and education, women were 1.37 times (95% CI 1.13–1.65) more likely than men to have SCD. The prevalence of SCD was increased with advanced age from 34.93% in participants aged 65–69 years to 55.11% in those aged 80 years and above, and women had a higher prevalence of SCD than men across all age groups(Fig. 2).

Fig. 2

Age- and sex-specific prevalence (%) of subjective cognitive decline.

Association of baseline SCD with incident CIND and dementia subtypes

Out of the 1,843 participants who were free of cognitive impairment (i.e., free of CIND and dementia) at baseline (2014) and who underwent the follow-up survey in 2018, 1,098 had no SCD and 745 were defined as having SCD at baseline; at the 2018 follow-up examination (mean follow-up time per person = 3.80 years; SD = 0.15), incident CIND, AD, and VaD were diagnosed in 189, 58, and 25 persons, respectively. Multinomial logistic regression analysis suggested that controlling for age, sex, education, obesity, APOE genotype, smoking, alcohol drinking, hypertension, diabetes, hyperlipidemia, ischemic heart disease, stroke, SSRS, and PSQI, having SCD at baseline was significantly associated with a 1.51-fold increased risk of CIND (95% CI 1.10–2.07) and a 3.11-fold increased risk of AD (95% CI 1.64–5.93). However, the association between baseline SCD and an increased OR of incident VaD at follow-up was not statistically significant (adjusted OR = 1.37, 95% CI 0.57–3.28).

Risk and protective factors for progression of SCD

Of the 688 participants who had SCD at baseline and who did not progress to dementia at the follow-up in 2018, 97 progressed to CIND in 2018. Binary logistic regression analysis showed that controlling for age and education, female sex was associated with an increased risk of progression from SCD to CIND (p < 0.05), but this association became non-significant in the multivariable-adjusted model. Hyperlipidemia was significantly associated with a decreased likelihood of progression from SCD to CIND, even in the fully-adjusted model (p < 0.05) (Table 2). Stratifying analyses by sex showed that higher education was significantly associated with a reduced OR of the conversion from SCD to CIND in males (demographic-adjusted OR = 0.22, 95% CI 0.08–0.58; multivariable-adjusted OR = 0.15, 95% CI 0.05–0.48), but not in females (demographic-adjusted OR = 1.30, 95% CI 0.77–2.20; multivariable-adjusted OR = 1.38, 95% CI 0.80–2.39).

Table 2

Associations of conversion from subjective cognitive decline to cognitive impairment, no dementia with demographic, lifestyle, and clinical factors (n = 688)

Characteristics^a	No. of	No. of	Odds ratio (95% confidence interval), CIND
	subjects	cases	Model 1^b	Model 2^b
Age (y)			1.00 (0.95–1.05)	1.00 (0.95–1.05)
Sex
Male	239	23	1.00 (Reference)	1.00 (Reference)
Female	449	74	1.79 (1.03–3.13)^*	1.73 (0.64–4.67)
Educational level
No school education	316	51	1.00 (Reference)	1.00 (Reference)
Elementary school or above	372	46	0.94 (0.57–1.53)	0.97 (0.58–1.62)
Obesity
No	494	64	1.00 (Reference)	1.00 (Reference)
Yes	144	22	1.14 (0.67–1.95)	1.23 (0.71–2.14)
APOEɛ4 allele carrier
No	564	78	1.00 (Reference)	1.00 (Reference)
Yes	101	15	1.07 (0.58–1.95)	1.12 (0.60–2.08)
Smoking
Never	478	77	1.00 (Reference)	1.00 (Reference)
Ever	210	20	0.80 (0.32–1.97)	0.75 (0.28–1.97)
Alcohol drinking
Never	473	74	1.00 (Reference)	1.00 (Reference)
Ever	215	23	1.06 (0.51–2.20)	1.24 (0.58–2.67)
Hypertension
No	170	22	1.00 (Reference)	1.00 (Reference)
Yes	518	75	1.05 (0.62–1.76)	1.33 (0.74–2.36)
Diabetes mellitus
No	588	84	1.00 (Reference)	1.00 (Reference)
Yes	100	13	0.91 (0.48–1.71)	0.99 (0.52–1.91)
Hyperlipidemia
No	480	76	1.00 (Reference)	1.00 (Reference)
Yes	208	21	0.54 (0.32–0.91)^*	0.55 (0.32–0.96)^*
Ischemic heart disease
No	545	78	1.00 (Reference)	1.00 (Reference)
Yes	143	19	0.85 (0.50–1.47)	0.88 (0.50–1.56)
Stroke history
No	618	91	1.00 (Reference)	1.00 (Reference)
Yes	70	6	0.57 (0.24–1.36)	0.61 (0.25–1.48)
SSRS score
< 20	64	14	1.00 (Reference)	1.00 (Reference)
≥20	611	82	0.54 (0.28–1.03)	0.52 (0.27–1.01)^c
PSQI score
< 7	330	45	1.00 (Reference)	1.00 (Reference)
≥7	293	43	1.02 (0.65–1.62)	1.03 (0.64–1.66)
HAMD score
< 8	340	46	1.00 (Reference)	1.00 (Reference)
≥8	63	13	1.50 (0.75–3.00)	1.47 (0.71–3.03)

CIND, Cognitive impairment, no dementia; APOE, apolipoprotein E gene; SSRS, Social Support Rating Scale; PSQI, Pittsburgh sleep quality index; HAMD, Hamilton Depression Scale. ^aThe number of subjects with missing value was 50 in obesity, 23 in APOE genotype, 13 in SSRS, 65 in PSQI, and 285 in HAMD. Dummy variables were created to indicate those with missing values. ^bModel 1 was adjusted for age, sex, and education; and in model 2, additional adjustment was made for obesity, APOE genotype, smoking, alcohol drinking, hypertension, diabetes, hyperlipidemia, ischemic heart disease, stroke, SSRS, PSQI, and HAMD scores. ^*p < 0.05. ^cp = 0.055.

Of the 744 participants who had SCD at baseline, 43 were diagnosed with AD and 13 with VaD in the 2018 follow-up examination. Multinomial logistic regression analysis suggested that advanced age and APOE ɛ4 allele were significantly associated with an increased risk of progression from SCD to AD, whereas higher education and SSRS score were associated with a reduced likelihood of progression to AD (p < 0.05). Obesity was associated with a higher risk of progression from SCD to VaD (p < 0.05). There were no significant associations of any of other examined factors with the progression of SCD (Table 3). We were not able to perform meaningful statistical analysis by sex due to too few AD and VaD cases in the sex-specific subgroups.

Table 3

Associations of conversion from subjective cognitive decline to Alzheimer’s disease and vascular dementia with demographic, lifestyle, and clinical factors (n = 744)

Characteristics^a	No. of subjects	Odds ratio (95% confidence interval), AD (n = 43)			Odds ratio (95% confidence interval), VaD (n = 13)
		No. of cases	Model 1^b	Model 2^b	No. of cases	Model 1^b	Model 2^b
Age (y)			1.19 (1.12–1.26)^*	1.21 (1.14–1.30)^*		0.99 (0.88–1.11)	1.03 (0.90–1.18)
Sex
Male	251	6	1.00 (Reference)	1.00 (Reference)	6	1.00 (Reference)	1.00 (Reference)
Female	493	37	2.01 (0.71–5.67)	3.83 (0.67–21.72)	7	0.31 (0.08–1.18)	0.49 (0.03–7.66)
Educational level
No school education	359	35	1.00 (Reference)	1.00 (Reference)	8	1.00 (Reference)	1.00 (Reference)
Elementary school or above	385	8	0.43 (0.17–1.10)	0.32 (0.12–0.88)^*	5	0.28 (0.07–1.12)	0.24 (0.04–1.24)
Obesity
No	532	33	1.00 (Reference)	1.00 (Reference)	5	1.00 (Reference)	1.00 (Reference)
Yes	157	5	0.71 (0.26–1.95)	0.73 (0.25–2.11)	8	6.26 (1.95–20.04)^*	6.30 (1.71–23.18)^*
APOEɛ4 allele carrier
No	603	30	1.00 (Reference)	1.00 (Reference)	9	1.00 (Reference)	1.00 (Reference)
Yes	113	10	2.22 (1.00–4.97)^c	2.60 (1.11–6.08)^*	2	1.37 (0.29–6.49)	1.25 (0.24–6.56)
Smoking
Never	521	36	1.00 (Reference)	1.00 (Reference)	7	1.00 (Reference)	1.00 (Reference)
Ever	223	7	1.67 (0.37–7.56)	1.17 (0.23–6.01)	6	2.47 (0.28–22.11)	2.64 (0.20–35.06)
Alcohol drinking
Never	517	35	1.00 (Reference)	1.00 (Reference)	9	1.00 (Reference)	1.00 (Reference)
Ever	227	8	2.05 (0.67–6.26)	2.90 (0.80–10.48)	4	0.56 (0.11–2.72)	0.50 (0.08–3.11)
Hypertension
No	182	9	1.00 (Reference)	1.00 (Reference)	3	1.00 (Reference)	1.00 (Reference)
Yes	562	34	0.87 (0.39–1.96)	0.98 (0.39–2.44)	10	1.33 (0.35–5.05)	0.75 (0.16–3.49)
Diabetes mellitus
No	636	38	1.00 (Reference)	1.00 (Reference)	10	1.00 (Reference)	1.00 (Reference)
Yes	108	5	1.18 (0.43–3.30)	1.34 (0.44–4.09)	3	1.79 (0.47–6.80)	1.05 (0.22–4.95)
Hyperlipidemia
No	515	28	1.00 (Reference)	1.00 (Reference)	7	1.00 (Reference)	1.00 (Reference)
Yes	229	15	1.12 (0.56–2.25)	0.98 (0.45–2.17)	6	2.39 (0.76–7.51)	1.48 (0.41–5.37)
Ischemic heart disease
No	587	32	1.00 (Reference)	1.00 (Reference)	10	1.00 (Reference)	1.00 (Reference)
Yes	157	11	1.28 (0.60–2.74)	1.31 (0.56–3.02)	3	1.23 (0.33–4.60)	0.65 (0.14–3.01)
Stroke history
No	668	41	1.00 (Reference)	1.00 (Reference)	9	1.00 (Reference)	1.00 (Reference)
Yes	76	2	0.78 (0.18–3.44)	0.75 (0.16–3.54)	4	4.02 (1.17–13.81)^*	3.33 (0.76–14.51)
SSRS score
<20	73	8	1.00 (Reference)	1.00 (Reference)	1	1.00 (Reference)	1.00 (Reference)
≥20	654	32	0.42 (0.17–1.03)	0.34 (0.13–0.86)^*	11	1.37 (0.17–11.07)	1.43 (0.15–13.77)
PSQI score
<7	357	19	1.00 (Reference)	1.00 (Reference)	8	1.00 (Reference)	1.00 (Reference)
≥7	316	19	1.07 (0.53–2.14)	0.92 (0.43–2.00)	4	0.63 (0.18–2.13)	0.57 (0.14–2.31)
HAMD score
<8	362	14	1.00 (Reference)	1.00 (Reference)	8	1.00 (Reference)	1.00 (Reference)
≥8	72	7	2.36 (0.87–6.41)	2.79 (0.93–8.40)	2	1.31 (0.26–6.49)	2.01 (0.34–12.03)

AD, Alzheimer’s disease; VaD, vascular dementia; APOE, apolipoprotein E gene; SSRS, Social Support Rating Scale. ^aThe number of subjects with missing value was 55 in obesity, 28 in APOE genotype, 17 in SSRS, 71 in PSQI, and 310 in HAMD score. Dummy variables were created to indicate those with missing values. ^bModel 1 was adjusted for age, sex, and education; and in model 2, additional adjustment was made for obesity, APOE genotype, smoking, alcohol drinking, hypertension, diabetes, hyperlipidemia, ischemic heart disease, stroke, SSRS, PSQI, and HAMD scores. ^*p < 0.05. ^cp = 0.051.

DISCUSSION

In this prospective cohort study, we examined the prevalence of SCD and factors associated with progression from SCD to CIND, AD, and VaD among a rural-dwelling Chinese elderly population. The major findings can be summarized as follows: 1) The overall prevalence of SCD was 40.07%, with the prevalence being increased with advanced age and being higher in women than in men; 2) Older adults with SCD had a 1.51-fold increased risk of developing CIND and a 3.11-fold increased risk of developing AD during the follow-up period of about 4 years; 3) Older age, lack of school education, APOE ɛ4 allele, and low social support were associated with accelerated progression from SCD to AD, while obesity were related to fast progression to VaD. To our knowledge, this is the first prospective cohort study exploring the prevalence and progression of SCD among the rural Chinese elderly population.

Previously, population-based studies have shown a substantial variation in the prevalence of SCD, depending on age of study population, geographic regions, and methods used to screen and define SCD. Although the SCD-I Working Group has introduced a consensus definition of pre-MCI SCD, there is currently still a lack of standardized operational methods for assessing SCD, which makes it difficult to compare the prevalence rates across studies. Indeed, subjective cognitive complaint can be assessed by different methods. The Brazilian Longitudinal Study of Aging (age≥50 years) showed that the prevalence of SCD, defined by having persistent cognitive complaints in the past 2 years, was estimated to be 29% [24]. The Shanghai Study of Health Promotion showed that the prevalence of SCD, defined using a simplified questionnaire and 2 additional questions involving memory and other domains, was 39% among older adults aged≥60 years [3]. The Mayo Clinic Study of Aging showed that SCD affected over 50% of participants aged 70–95 years where SCD was defined according to questions derived from the Blessed memory test and Everyday Cognition scale [25]. Different criteria for defining objective cognitive impairment further affect the prevalence of SCD. A population-based study in Beijing showed that the prevalence of SCD in cognitively unimpaired elderly (age 60–80 years) was 18% and 25% when applying neuropsychological standards of Jak/Bondi and ADNI2, respectively [4].

Our study showed that SCD was associated with an increased risk of subsequent conversion to CIND and AD, but not VaD, which is overall consistent with several previous reports from USA and Europe. For instance, the community-based cohort study that involved 4 study cohorts in USA, with an average follow-up period of 6 years, indicated that memory complaints were related to a faster cognitive decline and a higher risk of developing AD [26]. A longitudinal study in Germany showed that subjective memory impairment was predictive of the conversion to AD, but not to VaD [27]. The Amsterdam Study of the Elderly, with a mean follow-up of 3.2 years, found that subjective memory complaint was associated with a nearly 3-fold increased risk of AD among cognitively unimpaired elderly [28], which is in line with the observation in our study.

We found a higher prevalence of SCD in women than in men, and female sex was associated with an increased risk of progression from SCD to CIND after adjustment for age and education, although the association was attenuated when controlling for additional potential confounding factors. However, the sex-specific association between SCD and subsequent risk of dementia has been inconsistent across previous studies [29, 30], an issue that deserves further investigation in future studies.

Advanced age is the most consistent risk factor for AD [31]. Our study showed that each 1-year increase in age was associated with an approximately 21% increased risk of conversion from SCD to AD after 65 years of age. Furthermore, our results suggested that high education was associated with reduced risk of SCD progression. In the UK Cognitive Function and Ageing Study-Wales, cognitive reserve attenuated the risk of developing dementia associated with SCD [32]. These studies support the possibility that higher education may act as a proxy for cognitive reserve, which could help individuals with SCD to compensate for earlier cerebral pathological changes and thus reduce cognitive decline and deterioration.

Previous studies in older adults found that hyperlipidemia was associated with a reduced risk of cognitive decline [33], which is consistent with our study. However, other studies have found no association [34] or even an opposite association [35]. It is worth noting that owing to the relative short period of follow-up, reverse causation cannot be completely ruled out because serum cholesterol may start to decline over a decade prior to onset of clinical dementia and preclinical brain Alzheimer pathologies may lead to hypolipidemia [36].

The APOE ɛ4 allele is a known genetic risk factor not only for MCI and AD, but also for the conversion from MCI to dementia [37]. Our study identified the APOE ɛ4 allele was associated with a 2.60-fold increased risk of progression from SCD to AD. Similarly, a population-based study from South Korea [38] showed that older adults with SCD carrying the APOE ɛ4 allele had an almost 3-fold increased risk of subsequent cognitive impairment. A recent review supported an interaction effect between APOE ɛ4 status and SCD on cognitive decline such that APOE ɛ4 allele is a robust risk factor for accelerated cognitive decline in individuals with SCD [39].

Social support is closely linked to psychosocial processes and cognitive functions. Our study showed that a higher SSRS score was related to a 66% reduced risk of progression from SCD to AD. Lack of supportive social relationships may promote negative psychological emotions (e.g., loneliness and depression) [40, 41] and unhealthy lifestyles (e.g., smoking, drinking alcohol, and lack of physical activity) [42], which may increase the risk of developing cognitive impairment and dementia.

Midlife obesity has been associated with dementia, but studies of late-life obesity have yielded mixed results [43]. A previous meta-analysis found that a high BMI in midlife was a risk factor for late-life cognitive impairment and dementia, while a high BMI in late-life appeared to confer protective effects against dementia [44]. There is a lack of relevant longitudinal studies examining the association of obesity with SCD progression among cognitively unimpaired older adults. Our study found that obesity was associated with a 6.30-fold increased risk of conversion from SCD to VaD, whereas no significant association was observed with the progression to AD. Specific fat components are associated with distinct metabolic profiles [45] and may have different impacts on cognitive impairment and dementia. Central obesity may have a more significant impact on structural abnormalities in the brain than general obesity [46]. The association between reduced central obesity and peripheral adiposity and the risk of cognitive impairment and dementia deserves further investigation.

The main strength of this study is the community-based prospective cohort design that involved Chinese rural residents, a sociodemographic group that has rarely been targeted in research of SCD and dementia. In addition, dementia and CIND were defined following the international criteria, and objective cognitive impairment was excluded when defining SCD, consistent with the Jessen’s 2014 criteria for SCD [1]. Nevertheless, several limitations should be considered when interpreting our results. First, when defining CIND, we assessed objective cognitive function using MMSE rather than a comprehensive neuropsychological test battery. Second, the AD8 scale is generally used as a screening tool for dementia, although it includes questions to assess the self-reported subjective changes in some cognitive functions. Further studies should validate the use of AD8 as a screening tool for SCD in nondemented rural population with very limited education. Third, the study participants were recruited from only one rural area and around 16% of the participants at baseline were lost to the follow-up examination. This should be kept in mind when generalizing our results to other populations. Finally, our single-center study has a relatively short follow-up period. Future large-scale, multi-center, and long-term prospective cohort studies are warranted to verify our study findings.

Conclusion

In conclusion, SCD affects more than 40% of rural-dwelling older adults in China who are free of cognitive impairment, and SCD is associated with a nearly 1.5- and over 3-fold increased risk of developing CIND and AD, respectively. Older age, lack of education, APOE ɛ4 allele, and low social support are associated with an increased risk of progression from SCD to AD, while obesity is implicated in its progression to VaD. These findings, which contribute to the understanding of the distribution and progression of SCD in the rural population of China, if confirmed, may have important implications in terms of prognosis of SCD and prevention of dementia.

Footnotes

ACKNOWLEDGMENTS

We would like to thank all the participants of the SYS-AD and MIND-China Projects as well as the staff who were involved in the data collection and management.

FUNDING

SYS-AD was financially supported by the Science and Technology Program for Public Wellbeing of Shandong Province, China (grant no. 2013kjhm180405). MIND-China was supported in part by grants from the National Key R&D Program of China (grant no.: 2017YFC1310100), the National Natural Science Foundation of China (grants no.: 81861138008 and 8191101618), the Academic Promotion Program of Shandong First Medical University, and the Taishan Scholar Program of Shandong Province, China. This work was also supported by the Shandong Provincial Key Research and Development Program (grant no.: 2021LCZX03), the Integrated Traditional Chinese and Western Medicine Program in Shandong Province (grant no.: YXH2019ZXY008), the Brain Science and Brain-like Intelligence Technology Research Projects of China Ministry of Sciences and Technology (grant no.: 2021ZD0201808). C Qiu received grants from the Swedish Research Council (VR, grants no.: 2017-00740, 2017-05819, and 2020-01574) for the Sino-Sweden Network and Research Projects, the Swedish Foundation for International Cooperation in Research and Higher Education (STINT, grant no.: CH2019-8320) for the Joint China-Sweden Mobility program, and Karolinska Institutet, Stockholm, Sweden. The funding agency had no role in the study design, data collection and analysis, writing of this manuscript, and in the decision to submit the work for publication.

CONFLICT OF INTEREST

The authors have no conflict of interest to report.

DATA AVAILABILITY

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

References

Jessen

, Amariglio

, van Boxtel

, Breteler

, Ceccaldi

, Chetelat

, Dubois

, Dufouil

, Ellis

, van der Flier

, Glodzik

, van Harten

, de Leon

, McHugh

, Mielke

, Molinuevo

, Mosconi

, Osorio

, Perrotin

, Petersen

, Rabin

, Rami

, Reisberg

, Rentz

, Sachdev

, de la Sayette

, Saykin

, Scheltens

, Shulman

, Slavin

, Sperling

, Stewart

, Uspenskaya

, Vellas

, Visser

, Wagner

, Subjective Cognitive Decline Initiative Working G (2014) A conceptual framework for research on subjective cognitive decline in preclinical Alzheimer’s disease. Alzheimers Dement 10, 844–852.

Rohr

, Pabst

, Riedel-Heller

, Jessen

, Turana

, Handajani

, Brayne

, Matthews

, Stephan

BCM

, Lipton

, Katz

, Wang

, Guerchet

, Preux

, Mbelesso

, Ritchie

, Ancelin

, Carriere

, Guaita

, Davin

, Vaccaro

, Kim

, Han

, Suh

, Shahar

, Din

, Vanoh

, van Boxtel

, Kohler

, Ganguli

, Jacobsen

, Snitz

, Anstey

, Cherbuin

, Kumagai

, Chen

, Narazaki

, Ng

, Gao

, Gwee

, Brodaty

, Kochan

, Trollor

, Lobo

, Lopez-Anton

, Santabarbara

, Crawford

, Lipnicki

, Sachdev

, for Cohort Studies of Memory in an International Consortium (COSMIC) (2020) Estimating prevalence of subjective cognitive decline in and across international cohort studies of aging: A COSMIC study. Alzheimers Res Ther 12, 167.

Ruan

, Xiao

, Gong

, Zhang

, Ruan

, Zhang

, Chen

, Yu

(2020) Prevalence of cognitive frailty phenotypes and associated factors in a community-dwelling elderly population. J Nutr Health Aging 24, 172–180.

Hao

, Wang

, Zhang

, Xing

, Guo

, Hu

, Mu

, Chen

, Cao

, Zhi

, Liu

, Li

, Yang

, Li

, Du

, Sun

, Wang

, Liu

, Zhao

, Li

, Yu

, Wang

, Jia

, Han

(2017) Prevalence, risk factors, and complaints screening tool exploration of subjective cognitive decline in a large cohort of the Chinese population. J Alzheimers Dis 60, 371–388.

Pike

, Cavuoto

, Li

, Wright

, Kinsella

(2021) Subjective cognitive decline: Level of risk for future dementia and mild cognitive impairment, a meta-analysis of longitudinal studies. Neuropsychol Rev 32, 703–735.

Jia

, Quan

, Fu

, Zhao

, Li

, Wei

, Tang

, Qin

, Wang

, Qiao

, Shi

, Wang

, Du

, Zhang

, Luo

, Qu

, Zhou

, Gauthier

, Jia

, Group for the Project of Dementia Situation in China (2020) Dementia in China: Epidemiology, clinical management, and research advances. Lancet Neurol 19, 81–92.

Liu

, Li

, An

, Cheng

, Chen

, Zou

, Zhang

, Gan

, Xu

, Ou

, Wu

, Wang

, Zeng

(2021) Urban-rural disparities in mild cognitive impairment and its functional subtypes among community-dwelling older residents in central China. Gen Psychiatr 34, e100564.

Jessen

, Amariglio

, Buckley

, van der Flier

, Han

, Molinuevo

, Rabin

, Rentz

, Rodriguez-Gomez

, Saykin

, Sikkes

SAM

, Smart

, Wolfsgruber

, Wagner

(2020) The characterisation of subjective cognitive decline. Lancet Neurol 19, 271–278.

Slot

RER

, Sikkes

SAM

, Berkhof

, Brodaty

, Buckley

, Cavedo

, Dardiotis

, Guillo-Benarous

, Hampel

, Kochan

, Lista

, Luck

, Maruff

, Molinuevo

, Kornhuber

, Reisberg

, Riedel-Heller

, Risacher

, Roehr

, Sachdev

, Scarmeas

, Scheltens

, Shulman

, Saykin

, Verfaillie

SCJ

, Visser

, Vos

SJB

, Wagner

, Wolfsgruber

, Jessen

, Alzheimer’s Disease NeuroimagingInitiative; DESCRIPA working group; INSIGHT-preAD study group; SCD-Iworking group; Wiesje M van der Flier (2019) Subjective cognitivedecline and rates of incident Alzheimer’s disease andnon-Alzheimer’s disease dementia. Alzheimers Dement 15, 465–476.

10.

Wei

, Huang

, Chen

, Lin

, Shyu

, Chen

, Huang

, Lin

(2019) Subjective cognitive decline in the community is affected at multiple aspects of mental health and life quality: A cross-sectional study of the community medicine of Keelung Chang Gung Memorial Hospital. Dement Geriatr Cogn Dis Extra 9, 152–162.

11.

Fonseca

, Ducksbury

, Rodda

, Whitfield

, Nagaraj

, Suresh

, Stevens

, Walker

(2015) Factors that predict cognitive decline in patients with subjective cognitive impairment. Int Psychogeriatr 27, 1671–1677.

12.

Bessi

, Balestrini

, Bagnoli

, Mazzeo

, Giacomucci

, Padiglioni

, Piaceri

, Carraro

, Ferrari

, Bracco

, Sorbi

, Nacmias

(2020) Influence of ApoE genotype and clock T3111C interaction with cardiovascular risk factors on the progression to Alzheimer’s disease in subjective cognitive decline and mild cognitive impairment patients. J Pers Med 10, 45.

13.

, Cong

, Hou

, Chang

, Zhang

, Tang

, Han

, Wang

, Kalpouzos

, Du

, Qiu

(2021) Characterizing global and regional brain structures in amnestic mild cognitive impairment among rural residents: A population-based study. J Alzheimers Dis 80, 1429–1438.

14.

Ren

, Dong

, Hou

, Han

, Liu

, Wang

, Xu

, Wang

, Monastero

, Cong

, Du

, Qiu

(2022) Prevalence, incidence, and progression of cognitive impairment, no dementia among rural-dwelling Chinese older adults. J Alzheimers Dis 85, 1583–1592.

15.

Xiao

(1994) The theoretical basis and applications of Social Support Rating Scale (SSRS) (in Chinese). J Clin Psychiatry 004, 98–100.

16.

Buysse

, Reynolds

3rd , Monk

, Berman

, Kupfer

(1989) The Pittsburgh Sleep Quality Index: A new instrument for psychiatric practice and research. Psychiatry Res 28, 193–213.

17.

Galvin

, Roe

, Powlishta

, Coats

, Muich

, Grant

, Miller

, Storandt

, Morris

(2005) The AD8: A brief informant interview to detect dementia. Neurology 65, 559–564.

18.

Zimmerman

, Martinez

, Young

, Chelminski

, Dalrymple

(2013) Severity classification on the Hamilton Depression Rating Scale. J Affect Disord 150, 384–388.

19.

Wang

, Han

, Zhang

, Cong

, Tang

, Hou

, Liu

, Han

, Zhang

, Feng

, Yin

, Song

, Dong

, Liu

, Li

, Ngandu

, Kivipelto

, Snyder

, Carrillo

, Persson

, Fratiglioni

, Launer

, Jia

, Du

, Qiu

(2022) Health status and risk profiles for brain aging of rural-dwelling older adults: Data from the interdisciplinary baseline assessments in MIND-China. Alzheimers Dement (N Y) 8, e12254.

20.

Liang

, Liu

, Cong

, Wang

, Liu

, Fa

, Tian

, Cheng

, Wang

, Hou

, Du

, Qiu

(2022) Association and interaction of TOMM40 and PVRL2 with plasma amyloid-beta and Alzheimer’s disease among Chinese older adults: A population-based study. Neurobiol Aging 113, 143–151.

21.

American Psychiatric Association (1994) Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV). American Psychiatric Association, Washington, DC.

22.

McKhann

, Knopman

, Chertkow

, Hyman

, Jack

Jr. , Kawas

, Klunk

, Koroshetz

, Manly

, Mayeux

, Mohs

, Morris

, Rossor

, Scheltens

, Carrillo

, Thies

, Weintraub

, Phelps

(2011) The diagnosis of dementia due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 263–269.

23.

Román

, Tatemichi

, Erkinjuntti

, Cummings

, Masdeu

, Garcia

, Amaducci

, Orgogozo

, Brun

, Hofman

, Moody

, O’Brien

, Yamaguchi

, Grafman

, Drayer

, Bennett

, Fisher

, Ogata

, Kokmen

, Bermejo

, Wolf

, Gorelick

, Bick

, Pajeau

, Bell

, DeCarli

, Culebras

, Korczyn

, Bogousslavsky

, Hartmann

, Scheinberg

(1993) Vascular dementia: Diagnostic criteria for research studies. Report of the NINDS-AIREN International Workshop. Neurology 43, 250–260.

24.

Borelli

, Zimmer

, Bieger

, Coelho

, Pascoal

, Chaves

MLF

, Amariglio

, Castilhos

(2022) Subjective cognitive decline in Brazil: Prevalence and association with dementia modifiable risk factors in a population-based study. Alzheimers Dement (Amst) 14, e12368.

25.

van Harten

, Mielke

, Swenson-Dravis

, Hagen

, Edwards

, Roberts

, Geda

, Knopman

, Petersen

(2018) Subjective cognitive decline and risk of MCI: The Mayo Clinic Study of Aging. Neurology 91, e300–e312.

26.

Arvanitakis

, Leurgans

, Fleischman

, Schneider

, Rajan

, Pruzin

, Shah

, Evans

, Barnes

, Bennett

(2018) Memory complaints, dementia, and neuropathology in older blacks and whites. Ann Neurol 83, 718–729.

27.

Jessen

, Wiese

, Bachmann

, Eifflaender-Gorfer

, Haller

, Kolsch

, Luck

, Mosch

, van den Bussche

, Wagner

, Wollny

, Zimmermann

, Pentzek

, Riedel-Heller

, Romberg

, Weyerer

, Kaduszkiewicz

, Maier

, German Study on Aging, Cognition and Dementia in Primary Care Patients Study Group (2010) Prediction of dementia by subjective memory impairment: Effects of severity and temporal association with cognitive impairment. Arch Gen Psychiatry 67, 414–422.

28.

Geerlings

, Jonker

, Bouter

, Ader

, Schmand

(1999) Association between memory complaints and incident Alzheimer’sdisease in elderly people with normal baseline cognition. Am JPsychiatry 156, 531–537.

29.

Heser

, Kleineidam

, Wiese

, Oey

, Roehr

, Pabst

, Kaduszkiewicz

, van den Bussche

, Brettschneider

, Konig

, Weyerer

, Werle

, Fuchs

, Pentzek

, Mosch

, Bickel

, Maier

, Scherer

, Riedel-Heller

, Wagner

(2019) Subjective cognitive decline may be a stronger predictor of incident dementia in women than in men. J Alzheimers Dis 68, 1469–1478.

30.

Abner

, Kryscio

, Caban-Holt

, Schmitt

(2015) Baseline subjective memory complaints associate with increased risk of incident dementia: The PREADVISE trial. J Prev Alzheimers Dis 2, 11–16.

31.

Chan

, Wang

, Wu

, Liu

, Theodoratou

, Car

, Middleton

, Russ

, Deary

, Campbell

, Wang

, Rudan

, Global Health Epidemiology Reference Group (GHERG) (2013) Epidemiology of Alzheimer’s disease and other forms of dementia in China, 1990-2010: A systematic review and analysis. Lancet 381, 2016–2023.

32.

Jia

, Li

, Cao

(2021) Subjective cognitive decline, cognitive reserve indicators, and the incidence of dementia. J Am Med Dir Assoc 22, 1449–1455e4.

33.

Mielke

, Zandi

, Sjogren

, Gustafson

, Ostling

, Steen

, Skoog

(2005) High total cholesterol levels in late life associated with a reduced risk of dementia. Neurology 64, 1689–1695.

34.

Reitz

, Luchsinger

, Tang

, Manly

, Mayeux

(2005) Impact of plasma lipids and time on memory performance in healthy elderly without dementia. Neurology 64, 1378–1383.

35.

, Yin

, Zhu

, Luo

, Shi

, Gao

(2017) Blood cholesterol in late-life and cognitive decline: A longitudinal study of the Chinese elderly. Mol Neurodegener 12, 24.

36.

Stewart

, White

, Xue

, Launer

(2007) Twenty-six-year change in total cholesterol levels and incident dementia: The Honolulu-Asia Aging Study. Arch Neurol 64, 103–107.

37.

, Caracciolo

, Wang

, Santoni

, Winblad

, Fratiglioni

(2013) Accelerated progression from mild cognitive impairment to dementia among APOE epsilon4epsilon4 carriers. J Alzheimers Dis 33, 507–515.

38.

Hong

, Yoon

, Shim

, Kim

, Choi

, Jeong

, Yoon

, Yang

, Lee

(2015) Predictors of clinical progression of subjective memory impairment in elderly subjects: Data from the Clinical Research Centers for Dementia of South Korea (CREDOS). Dement Geriatr Cogn Disord 40, 158–165.

39.

Ali

, Smart

, Gawryluk

(2018) Subjective cognitive decline and APOE epsilon4: A systematic review. J Alzheimers Dis 65, 303–320.

40.

Chen

, Hicks

, While

(2014) Loneliness and social support of older people in China: A systematic literature review. Health Soc Care Community 22, 113–123.

41.

Sonnenberg

, Deeg

, van Tilburg

, Vink

, Stek

, Beekman

(2013) Gender differences in the relation between depression and social support in later life. Int Psychogeriatr 25, 61–70.

42.

Croezen

, Picavet

, Haveman-Nies

, Verschuren

, de Groot

, van’t Veer

(2012) Do positive or negative experiences of social support relate to current and future health? Results from the Doetinchem Cohort Study. BMC Public Health 12, 65.

43.

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.

44.

, Hu

, Ou

, Shen

, Xu

, Wang

, Dong

, Tan

, Yu

(2020) Association of body mass index with risk of cognitive impairment and dementia: A systematic review and meta-analysis of prospective studies. Neurosci Biobehav Rev 115, 189–198.

45.

Carr

, Utzschneider

, Hull

, Kodama

, Retzlaff

, Brunzell

, Shofer

, Fish

, Knopp

, Kahn

(2004) Intra-abdominal fat is a major determinant of the National Cholesterol Education Program Adult Treatment Panel III criteria for the metabolic syndrome. Diabetes 53, 2087–2094.

46.

Tang

, Zhao

, Lu

, Zhang

, Xin

, Sun

, Tian

, Cardoso

, Yang

, Simo

, Zhou

, Stehouwer

CDA

(2021) Relationship between central obesity and the incidence of cognitive impairment and dementia from cohort studies involving 5,060,687 participants. Neurosci Biobehav Rev 130, 301–313.