Social determinants of health,genetic biomarkers,and cognitive function in Chinese older adults with subjective cognitive decline and mild cognitive impairment: A cross-sectional analysis

Abstract

Background

Social isolation (SI) is an established risk factor for Alzheimer's disease (AD) and cognitive decline. However, its stage-specific effects across the AD continuum, particularly at subjective cognitive decline (SCD) and mild cognitive impairment (MCI) stages, remain unquantified in Chinese populations. The sex-specific effects of SI on cognitive decline remain incompletely characterized. The apolipoprotein E (APOE) genotype is a driver of cognitive decline and dementia.

Objective

To investigate social connection characteristics and gene distribution in individuals with SCD or MCI, examine their cross-sectional associations with cognitive function, and explore gender differences in SCD or MCI risk/prevention.

Methods

A community-based sample of 164 SCD and 84 MCI patients (July 2021–Dec 2024) was assessed. Demographic, social connectivity, APOE genotype and cognitive data were collected. Social connectivity, APOE genotype and cognitive function were compared between groups. In the overall or male and female genders, Pearson correlation analyzed associations between social connectivity and cognitive scores. Sex-stratified multivariable models tested Lubben Social Network Scale (LSNS-6)-by-sex interactions on MCI risk.

Results

LSNS-6 scores showed weak-to-moderate correlations with Montreal Cognitive Assessment (r = 0.140, p = 0.027), Animal Verbal Fluency Test-Huashan (AVLT-H) delayed recall (r = 0.129, p = 0.043), AVLT-H recognition (r = 0.190, p = 0.003), Trail Making Test-B (r = -0.132, p = 0.038), Boston Naming Test (r = 0.147, p = 0.021), and Animal Verbal Fluency Test (r = 0.148, p = 0.020). Multivariable binary logistic regression analysis revealed the association between social network and MCI risk differed by sex (OR_male = 0.581; OR_female = 0.746; p_interaction = 0.021).

Conclusions

SI may exacerbate cognitive dysfunction in adults with SCD or MCI. Women leverage social connectivity into significantly greater neuroprotective gains compared to men.

Trial registration

ChiCTR2300073429. https://www.chictr.org.cn/bin/project/edit?pid=200381

Keywords

Alzheimer’s disease apolipoprotein E cognitive dysfunction mild cognitive impairment sex differences social isolation subjective cognitive decline

Introduction

Alzheimer's disease (AD) is a slowly progressive neurodegenerative disorder characterized by the widespread deposition of amyloid-β (Aβ) plaques in early stages, followed by tau aggregation in the neocortex, neurodegeneration.^1,2 It manifests through a chronic progressive decline in cognitive function, often accompanied by memory loss, language impairment, compromised problem-solving skills, and other deficits.^3,4 Psychopathological symptoms, such as alterations in personality and social withdrawal, further exacerbate the condition, significantly impeding daily functioning and imposing substantial burdens on both individuals with dementia and their caregivers, as well as on national healthcare systems.^4,5 Presently, there exist no effective therapies or drugs capable of halting or reversing the progression of AD or forestalling neurodegeneration.^4,6

The disease progression is conceived as a continuum, spanning from subjective cognitive decline (SCD) or preclinical Alzheimer's disease to an intermediary stage of mild cognitive impairment (MCI), culminating in dementia.^7,8 MCI, a transitional phase between cognitive decline due to normal aging and the diagnosis of AD dementia, is typified by cognitive deficits alongside generally preserved daily functioning.^9,10 SCD refers to individuals’ self-perceived decline in memory or other cognitive functions compared to their previous performance, without objective neuropsychological deficits.^11,12 Given its proximity to the normal/early pathological boundary in cognitive aging, the impact of study settings on SCD outcomes may be even more significant than in studies focusing on MCI outcomes.^12,13

The strongest genetic risk factor, apolipoprotein E (APOE) ε4, is associated with increased Aβ deposition, hyperphosphorylation and aggregation of tau, and accelerated cognitive decline.^10,14 Its frequency is 13% to 15% worldwide, but among individuals with AD, the prevalence increases to upwards of 40%.¹⁵ However, the risk for cognitive impairment associated with an unhealthy lifestyle far surpasses the risk conferred by APOE ε4 among older adults.^16,17 Fourteen potentially modifiable risk factors, including lower education, hearing impairment, untreated vision loss, smoking, alcohol intake, diabetes, hypertension, obesity, high LDL cholesterol, depression, physical inactivity, poor sleep, air pollution, and low social participation, have been identified.^18,19 Prevention measures may mitigate at least 40% of the risk of AD and related dementia.^5,20 Emerging evidence indicates that modifiable psychosocial and behavioral factors, including lifestyle patterns, perceived social support, and loneliness, serve as significant effect modifiers in the association between the APOE ε4 allele and cognitive trajectories, potentially attenuating or exacerbating genetic risk.²¹

Older adults with SCD often report a withdrawal from social and leisure activities occurring 5–10 years prior.²² Social isolation (SI) and loneliness are two pivotal components of the social determinants of health, encapsulating distinct facets of social interaction.^23,24 SI denotes the objective absence of social connections, characterized by individuals being alone, having few relationships, or engaging infrequently in social interactions.^24–26 A burgeoning epidemic among older populations, SI has prevalence rates of 24% in the USA and 34.9% in China.^27–29 Chinese older adults exhibit unique features of SI, increasingly residing alone in “empty nest” households and often engaging in extensive caregiving for grandchildren.^30,31 Consequently, they report elevated rates of loneliness, depression, and diminished quality of life.^30,31 The COVID-19 pandemic exacerbated these challenges by imposing mobility restrictions, curbing physical and social activities among older adults, thereby accelerating cognitive decline and worsening dementia symptoms.^4,28,32

The prevalence of modifiable dementia risk factors demonstrates significant geographic heterogeneity, driving divergent population-level dementia burdens globally.^33,34 These disparities implicate sex/gender-specific mediation of social, cultural, demographic, and historical contexts in shaping AD risk.^33,35,36 Sex differences in AD are widely recognized.^37,38 Generally, the prevalence of SI varies by gender.²⁶ SI has been linked to future cognitive decline and an increased risk of AD.^19,39 Evidence regarding the gender-specific effects of SI on cognitive decline remains limited and conflicting.^40–42 Nonetheless, the specific impacts of SI across different clinical stages of AD, particularly in SCD, remain inadequately elucidated.^6,43,44 Less is known about the associations between SI and cognitive measures differ by sex in SCD and MCI. We hypothesize that multiple risk factors, including social networks, biomarkers, and sociodemographic variables, differentially impact cognition in older adults with SCD or MCI, modulated by sex. Hence, the purpose of this article is to investigate social connection characteristics, cognitive function, and APOE gene distribution in individuals with SCD or MCI, examine the cross-sectional association of social networks with cognitive function, and explore gender differences in SCD or MCI risk/prevention.

Methods

Participants

A total of 248 individuals were enrolled in the study from July 2021 to December 2024, comprising 164 individuals with SCD and 84 participants with MCI, recruited from communities in Nanjing, China. Subjects were classified as having SCD if they self-reported cognitive decline compared to a previous state, while objective investigations, including a neuropsychological battery (adjusted for age, gender, and education), remained within normal limits, and criteria for MCI, dementia, or other neurologic or psychiatric disorders potentially causing cognitive complaints were not met.⁴⁵ MCI participants fulfilled the core clinical criteria for MCI as outlined by the National Institute on Aging-Alzheimer's Association (NIA-AA) workgroup guidelines.⁴⁶

Exclusion criteria were applied uniformly to all subjects and included: (a) Age under 50 years or over 85 years; (b) Diagnosis of vascular dementia or other central nervous system diseases; (c) Hachinski Ischemic Scale score > 4 points; (d) Inability to complete neuropsychological tests (e.g., due to blindness, deafness, or severe language impairment); (e) History of drug abuse or alcohol dependency within the past 6 months; (f) Concurrent participation in other studies related to cognition; (g) Presence of severe diabetes mellitus, cardiovascular disease, cerebrovascular disease, liver diseases, kidney diseases, or psychiatric disorders.

The study received approval from the ethics committee of Jiangsu Province Hospital, and all participants provided informed consent before enrollment.

Neuropsychological and behavioral assessment

Neuropsychological assessments were conducted by trained psychologists and research assistants. The full battery is displayed below, and we categorized each test into six cognitive domains:

Global cognitive function: Montreal Cognitive Assessment (MoCA).⁴⁷

Memory: the delayed memory (N5) and recognition (N7) dimensions of Auditory Verbal Learning Test-Huashan (AVLT-H),⁴⁸ Wechsler memory scale-revised logical memory (WMS-RLM), Wechsler digit span (WDS).⁴⁹

Executive function: Trail Making Test B (TMT-B),⁵⁰ Symbol Digit Modalities Test (SDMT).⁵¹

Language: Boston Naming Test (BNT),⁵² Animal Verbal Fluency Test (VFT).⁵³

LSNS-6

Social network was assessed using the abbreviated version of the Lubben Social Network Scale (LSNS-6), comprising six questions that evaluate the size of three different aspects of an individual's friends and family social network. The LSNS-6 has been identified as a valuable tool for gauging social contact among older mainland Chinese.^54,55 Scores on the LSNS-6 range from 0 to 30, with higher scores indicating a larger social network. Lubben proposed a score of less than 12 as a clinical cutoff point for the LSNS-6, indicative of social isolation, implying that, on average, respondents have fewer than two people for fulfilling social integration functions.^54,56 A cutoff score of ≤ 12 on the LSNS-6 was used to define SI. Participants scoring above this threshold (13–30) were categorized as not socially isolated (non-SI).

Covariates

All participants underwent a comprehensive survey and assessment protocol to collect data on various demographic variables including age, sex, level of education, living situation. We selected potential confounders of the relationships between social isolation, gender, and cognitive function as covariates. Additionally, lifestyle factors such as physical activity (assessed using the International Physical Activity Questionnaire⁵⁷) and sleep quality (measured with the Pittsburgh Sleepiness Quality Index⁵⁸), quality of life, and health conditions (including clinical history, medication use, neurological examination results, presence of conditions such as coronary heart disease, hypertension, hyperlipidemia, diabetes, lacunar infarction.) were evaluated. The assessment also included the administration of loneliness (using the University of California, Los Angeles Loneliness Scale⁵⁹) and depression (measured with the Geriatric Depression Scale¹²) scales. Furthermore, participants who consented to a blood draw had serum protein and APOE genotype determination. Levels of Serum Aβ₁₋₄₂ and p-Tau-181 were determined via enzyme-linked immunosorbent assay in department of clinical laboratory, The First Affiliated Hospital of Nanjing Medical University. Consumables and reagents were obtained from Shenzhen Anqun Bioengineering Co., Ltd Genotypes for rs7412 and rs429358 were genotyped using the Tiangen Biochemical DP318 of Tiangen Biochemical Technology (Beijing) Co., Ltd Participants were classified as APOE ε4 carriers if they possessed the ε2/ε4, ε3/ε4, or ε4/ε4 genotype.

Statistical analysis

The data were presented as mean ± SD for continuous variables and as frequencies and proportions for categorical variables. Independent t-tests were used for continuous variables, while the Chi-square test was employed for categorical variables. Initially, demographic characteristics, cognitive assessment results, and other observed indicators were compared between the SCD and MCI groups using t-tests and analysis of variance. Subsequently, Pearson correlation analysis was conducted to assess the associations between LSNS-6 scores, cognitive scores, APOE genotype, and other assessment indicators in the overall participant group, as well as within the SCD group, MCI group, and stratified by sex. Following adjustment for covariates including gender, age, education, body mass index (BMI), diabetes, hypertension, lacunar infarction, loneliness, and depression through partial correlation analysis, we assessed the association between LSNS-6 scores and cognitive performance. Finally, to assess potential effect modification by sex in the association between social networks and MCI risk, we tested a sex-by-social network interaction term within multivariable binary logistic regression models, adjusting for established confounders including age, education, APOE ε4 status, cardiovascular comorbidities, and so on. Statistical analyses were performed using SPSS Statistics version 26.0, with significance levels set at p < 0.05.

Results

Demographic characteristics, cognitive assessments, serum biomarkers, and APOE ε4 carrier status

Demographic characteristics, cognitive performance, serum biomarker levels, and APOE ε4 carrier status are summarized in Table 1. A total of 248 participants were included in the analysis, comprising 164 individuals with SCD and 84 with MCI. No significant differences were observed between the two groups in terms of gender, BMI, hypertension, lacunar infarction, serum Aβ/tau protein levels. Significant differences were found in age, education, the scores of cognitive functions and APOE ε4 carrier status between the SCD and MCI groups. Out of 248 participants, 84 cases (33.871%) experienced SI. APOE ε4 carrier status was observed in 7.317% (12/164) of SCD cases versus 19.048% (16/84) of MCI cases. When considering SCD and MCI as outcomes, the risk of MCI in socially isolated participants was 2.119 times higher than in non-socially isolated individuals (Crude Odds Ratio 2.119, 95% Confidence Interval 1.225–3.665).

Table 1.

Basic characteristics of the study population.

		SCD (n = 164)	MCI (n = 84)	TOTAL (n = 248)	p
Gender (male/female)		65/99	26/58	91/157	0.179^a
Age (y)		62.915 (6.835)	65.119 (7.101)	63.661 (6.990)	0.018
BMI (kg/m²)		24.255 (2.770)	24.517 (3.351)	24.344 (2.975)	0.512
Education (y)		11.537 (2.680)	10.262 (2.504)	11.105 (2.685)	<0.001
Hypertension (Yes/No)		74/90	41/43	115/133	0.582^a
Diabetes (Yes/No)		23/141	22/62	45/203	0.019^a
Lacunar Infarction (Yes/No)		45/119	22/62	67/181	0.834^a
LSNS-6		15.713 (5.174)	13.476 (5.324)	14.956 (5.321)	0.002
SI/non-SI		46/118	38/46	84/164	0.007^a
MoCA		23.835 (2.910)	21.119 (2.717)	22.915 (3.119)	<0.001
AVLT-H	delayed recall	4.921 (1.880)	2.821 (1.818)	4.210 (2.106)	<0.001
AVLT-H	recognition	21.049 (1.740)	18.179 (2.658)	20.077 (2.495)	<0.001
TMT-B (s)		146.573 (40.161)	210.440 (63.867)	168.206 (57.900)	<0.001
BNT		25.390 (2.984)	22.988 (3.794)	24.577 (3.466)	<0.001
VFT		17.079 (4.807)	15.107 (4.313)	16.411 (4.730)	0.002
WLM-RLM		12.801 (5.688)	9.608 (4.912)	11.676 (5.627)	<0.001^b
WDS		16.338 (3.893)	14.635 (3.046)	15.738 (3.701)	0.001^b
SDMT		34.630 (9.782)	27.685 (11.881)	32.192 (11.049)	<0.001^c
UCLA		33.667 (10.523)	35.227 (11.820)	34.187 (10.967)	0.347^d
GDS		8.161 (5.807)	9.453 (6.284)	8.618 (5.998)	0.134^e
IPAQ		3256.612 (1914.743)	3918.106 (2264.539)	3477.110 (2056.140)	0.032^d
PSQI		7.520 (4.293)	8.106 (4.990)	7.701 (4.515)	0.382^f
Aβ₁₋₄₂ (pg/L)		75.202 (61.394)	83.400 (97.646)	78.467 (77.589)	0.576^g
Tau-181 (pg/L)		19.233 (12.460)	18.351 (16.520)	18.881 (14.155)	0.742^g
APOE ε4 carrier status (Yes/No)		12/97	16/52	28/149	0.026^{a h}

Chi-square test, ^b210 cases, ^c208 cases, ^d198 cases, ^e212 cases, ^f214 cases, ^g118 cases, ^h177 cases (1 ε4/ε4, 24 ε3/ε4, 3 ε2/ε4, 123 ε3/ε3, 25 ε2/ε3, 1 ε2/ε2).

SCD: subjective cognitive decline; MCI: mild cognitive impairment; MoCA: Montreal Cognitive Assessment; AVLT-H: Auditory Verbal Learning Test-Huashan; TMT-B: Trail Making Test B; BNT: Boston Naming Test; VFT: Animal Verbal Fluency Test; WLM-RLM: Wechsler memory scale-revised logical memory; WDS: Wechsler digit span; SDMT: Symbol Digit Modalities Test; GDS: Geriatric Depression Scale; LSNS-6: Lubben Social Network Scale-6; UCLA: University of California, Los Angeles Loneliness Scale; IPAQ: International Physical Activity Questionnaire; PSQI: The Pittsburgh Sleepiness Quality Index. The parameters boldfaced are with significant difference between SCD and MCI groups.

Correlations of LSNS-6 and cognitive assessments

The heatmap depicts Pearson correlation analysis correlation coefficients among LSNS-6, covariates and neuropsychological test scores. Correlation coefficients (r) are displayed within each cell, with color intensity indicating the strength and direction of the correlation (see color scale). The complete dataset, including detailed sample sizes of per cell, is presented in the Supplemental Material. Statistical significance is denoted as follows: *p < 0.05, **p < 0.01.

LSNS-6: Lubben Social Network Scale-6; GDS: Geriatric Depression Scale; UCLA: University of California, Los Angeles Loneliness Scale; IPAQ: International Physical Activity Questionnaire; PSQI: Pittsburgh Sleepiness Quality Index; MoCA: Montreal Cognitive Assessment; TMT-B: Trail Making Test B; BNT: Boston Naming Test; VFT: Animal Verbal Fluency Test; WLM-RLM: Wechsler memory scale-revised logical memory; WDS: Wechsler digit span; SDMT: Symbol Digit Modalities Test.

The results of Pearson correlation analysis revealed significant correlations between the LSNS-6 score and various cognitive measures (Figure 1), including MoCA (r = 0.140, p = 0.027), AVLT-H (delayed recall) (r = 0.129, p = 0.043), AVLT-H (recognition) (r = 0.190, p = 0.003), TMT-B (r = -0.132, p = 0.038), BNT (r = 0.147, p = 0.021), VFT (r = 0.148, p = 0.020), Geriatric Depression Scale (GDS) (r = -0.398, p < 0.001), and University of California, Los Angeles Loneliness Scale (UCLA) (r = -0.515, p < 0.001). Additionally, borderline associations approaching statistical significance were observed with TMT-B (r = -0.140, p = 0.057) and VFT (r = 0.142, p = 0.053). Furthermore, serum Aβ₁₋₄₂ level exhibited a positive correlation with serum p-Tau-181 level (r = 0.726, p < 0.001) and age (r = 0.214, p = 0.020), while displaying a negative correlation with BNT score (r = -0.238, p = 0.009). No significant associations were observed between the LSNS-6 and cognitive or serological measures in the analysis stratified by SCD and MCI status. Analysis within the SCD group revealed that Aβ₁₋₄₂ was positively correlated with Tau-181 (r = 0.536, p < 0.001) and negatively correlated with scores on the AVLT-H delayed recall (r = -0.286, p = 0.016) and Wechsler memory scale-revised logical memory (WLM-RLM) (r = -0.373, p = 0.003). Furthermore, Tau-181 was also negatively associated with AVLT-H delayed recall (r = -0.248, p = 0.037) and WLM-RLM (r = -0.440, p < 0.001) performances. In the MCI group, Aβ₁₋₄₂ and Tau-181 demonstrated a strong positive correlation (r = 0.873, p < 0.001). Additionally, both Aβ₁₋₄₂ (r = -0.338, p = 0.020) and Tau-181 (r = -0.303, p = 0.039) were significantly correlated with the BNT score.

Figure 1.

Correlation matrix of LSNS-6, covariates, and cognitive assessments.

After adjusting for gender, age, education, BMI, diabetes, hypertension, lacunar infarction, loneliness (UCLA), and depression (GDS) scores by partial correlation analysis, the LSNS-6 score remained a statistically significant correlation with AVLT-H (recognition) (r = 0.161, p = 0.028).

Among 248 participants (91 males, 36.7%; 157 females, 63.3%), sex-stratified analyses revealed differential correlates of social isolation (LSNS-6). In males, LSNS-6 demonstrated broad associations with cognitive and affective measures, including MoCA (r = 0.380, p＜0.001), AVLT-H (delayed recall) (r = 0.271, p = 0.009), AVLT-H (recognition) (r = 0.254, p = 0.015), TMT-B (r = -0.328, p = 0.002), VFT (r = 0.238, p = 0.023), WLM-RLM (r = 0.229, p = 0.046), WDS (r = 0.306, p = 0.007), SDMT (r = 0.311, p = 0.007), and GDS (r = -0.362, p = 0.001), UCLA (r = -0.424, p＜0.001). Conversely, females showed selective links only to BNT (r = 0.172, p = 0.032), GDS (r = -0.463, p < 0.001), and UCLA Loneliness Scale (r = -0.584, p < 0.001).

Correlations of LSNS-6 and the risk of MCI

From Table 2, it is evident that the p values for diabetes (OR = 3.834, 95% CI [1.361, 10.797], p = 0.011), APOE ε4 carrier (OR = 3.239, 95% CI [1.040, 10.092], p = 0.043), and the interaction between gender and LSNS-6 (OR = 1.226, 95% CI [1.031, 1.458], p = 0.021) are all below 0.05, indicating their significant impact on MCI risk. Additionally, separate odds ratios for LSNS-6 were computed for males and females:

{OR}_{male} = e^{- 0.793 + 1 \times 0.25} = 0.581, {OR}_{female} = e^{- 0.793 + 2 \times 0.25} = 0.746, p_{interaction} = 0.021

Table 2.

Logistic regression analysis of risk of MCI: main effects of social network and sex, and their interaction.

	N	n (%)	Event N	OR	95% CI	p
Gender	143
Male		54 (37.762%)	15	0.180	0.017∼1.961	0.159
Female		89 (62.238%)	38	-	-	-
Age		-	-	1.032	0.969∼1.100	0.326
Education		-	-	0.919	0.778∼1.086	0.321
BMI		-	-	1.096	0.957∼1.255	0.187
Hypertension	143
Yes		77 (53.846%)	29	0.801	0.330∼1.945	0.624
No		66 (46.154%)	24	-	-	-
Diabetes	143
Yes		116 (81.119%)	38	3.834	1.361∼10.797	0.011
No		27 (18.881%)	15	-	-	-
Lacunar Infarction	143
Yes		96 (67.133%)	36	0.881	0.361∼2.152	0.781
No		47 (32.867%)	17	-	-	-
APOE genotype	143
APOE ε3/ε3		102 (71.329%)	33	-	-	-
APOE ε2/ε3ⁱ		20 (13.986%)	8	1.269	0.410∼3.926	0.680
APOE ε4 carrier		21 (14.685%)	12	3.239	1.040∼10.092	0.043
IPAQ		-	-	1.000	1.000	0.282
PSQI		-	-	0.960	0.873∼1.056	0.402
GDS		-	-	1.033	0.934∼1.141	0.529
UCLA		-	-	0.996	0.943∼1.052	0.893
LSNS-6		-	-	0.793	0.678∼0.927	0.004
Gender by LSNS-6		-	-	1.226	1.031∼1.458	0.021

OR: odds ratio; 95% CI: 95% confidence interval; LSNS-6: Lubben Social Network Scale-6; GDS: Geriatric Depression Scale; UCLA: University of California, Los Angeles Loneliness Scale; IPAQ: International Physical Activity Questionnaire; PSQI: Pittsburgh Sleepiness Quality Index. The logistic regression demonstrated adequate fit (Hosmer-Lemeshow test: p = 0.797). Explanatory power was quantified by both Cox-Snell R-squared (0.211) and Nagelkerke R-squared (0.288), suggesting the model accounted for approximately 21.1–28.8% of the variance. ⁱ20 cases (19 ε2/ε3, 1 ε2/ε2).

There were statistically significant differences in the effect of LSNS-6 score on cognition between males and females.

Discussion

We observed a statistically significant but small-magnitude correlation between SI and impaired cognition in Chinese males and females with SCD or MCI, whereas loneliness showed no significant association with cognitive function. The independent effects of SI on memory function (AVLT-H) persist even after controlling for various sociodemographic, behavioral, and health factors. The modest association between social isolation and cognition suggests it may not be a direct driver of decline but rather operate through mediators like depression; however, limited statistical power calls for verification in larger, longitudinal studies. SI may influence cognitive health through multiple interconnected pathways. First, SI can lead to depression, chronic stress, and reduced emotional regulation, which are established risk factors for cognitive decline.^31,60,61 Second, it often promotes unhealthy behaviors like physical inactivity.^61–63 Finally, these psychosocial and behavioral factors can trigger biological dysregulation, including chronic inflammation and neuroendocrine stress responses via HPA-axis and sympathetic nervous system activation, which directly damage cardiovascular and brain health.^{19,21,61,64,65}

Htun et al. identified that weak social connection patterns increase dementia risk in men,^40,41 a vulnerability amplified during the pandemic where living alone and deficient interpersonal connections were especially prevalent and deleterious for men, as reported by Wang et al. and Santiago et al..^26,66 Conversely, Guo et al. found that in Chinese populations, SI (distinct from loneliness) was significantly associated with functional disability in women, but not in men.⁴² Our analysis revealed sex-specific patterns: in men, social network scores showed moderate correlations across multiple cognitive domains, while in women, only a weak link to language function was found. These data suggest tentatively that SI affects males more severely, but unmeasured confounding cannot be excluded. Additionally, logistic models demonstrated that the protective effect of stronger social networks against MCI progression was significantly more potent in women, pointing to distinct, sex-specific neuroprotective mechanisms. This may be attributed to several factors: women generally maintain more diverse and emotionally supportive social ties, whereas men often rely on instrumental support and have fewer intimate connections, which may limit the cognitive benefits of social engagement.^{26,40,41,61,65,67} Moreover, social support is more strongly linked to stress reduction in women, and its absence exacerbates unhealthy behaviors, depression, and sleep disturbances—key risk pathways for dementia in females.^31,36,42,60 Sex hormones and immune-inflammatory responses also contribute, as women's heightened immune reactivity may increase their vulnerability to social isolation–induced neuroinflammation.^34,64,65 Thus, while social connectivity benefits both sexes, its role as an active protective factor appears more potent in women. These findings underscore the need for sex-specific interventions, wherein enhancing social networks may offer particular cognitive benefits for older women.

The APOE gene is the strongest genetic risk modifier for AD, with the APOE ε4 allele increasing risk and APOE ε2 decreasing it compared with the common APOE ε3 allele.^68–70 The AD risk conferred by APOE ε4 is dose dependent. Those carrying one ε4 allele have a 2-4-fold increased AD risk, while those carrying two ε4 alleles have a 10-15-fold increased AD risk.^7,16,71 Both APOE ε4 and subjective cognitive decline at baseline were equally and independently associated with dementia or MCI, with the highest risk observed when subjective cognitive decline and APOE ε4 co-occurred.¹⁶ The ApoE protein has a central role in lipid metabolism, and the APOE genotype is associated with cardiovascular disease and cardiovascular disease risks, including stroke, atherosclerosis, white matter hyperintensities, and myocardial infarction, as well as levels of oxidative stress and inflammation.^70,71 In our study, APOE ε4 carrier prevalence was significantly higher in MCI participants than in SCD participants, and APOE ε4 carriers in SCD individuals exhibit a 3.239 times increased risk of MCI. However, the wide confidence interval, indicating risk estimates ranging from a marginal 4% to a substantial tenfold increase, highlights the imprecision of this estimate. This uncertainty reflects limitations in sample size within subgroup analyses, which constrained our statistical power to accurately determine the effect size. Thus, although the point estimate implies a strong association, these results should be interpreted cautiously.

Cognitive brain function is intricately interconnected with cardiovascular health, with the latter acting as a critical modifiable determinant in the risk trajectory for cognitive impairment and all-cause dementia.^72,73 Our results are consistent with prior evidence that diabetes is an independent risk factor for cognitive impairment, with effects mediated through cerebrovascular, metabolic, and inflammatory pathways, thereby supporting the cognitive benefits of its rigorous management.^18,34,73,74

It is well established that the neuropathologic markers of AD (i.e., Aβ plaques and tau neurofibrillary tangles) begin to accumulate for 20 or more years before the onset of cognitive impairment.^75–77 Cognitive change relative to amyloid onset is essential to identify early opportunities for intervention and prevention.⁷⁸ A significant negative correlation was observed between serum Aβ₁₋₄₂ levels and BNT scores across the cohort, consistent with previous reports. Furthermore, when stratified by diagnosis, Aβ₁₋₄₂ and p-Tau-181 were associated with memory scores (AVLT-H, WLM-RLM) in the SCD group and with language function (BNT) in the MCI group, respectively. Plasma Aβ₁₋₄₂ and t-tau levels have predictive utility for cognitive decline from normal cognition to MCI.^75,79 Higher baseline plasma Aβ₁₋₄₂ and t-tau levels may indicate a higher risk of cognitive decline in cognitively normal adults.^75,79,80 Aβ deposition can predict the progression of subsequent tau pathology and the decline in memory and executive function in the early preclinical stage of AD.⁷⁶ Blood biomarkers have the potential to greatly accelerate the development of effective disease modifying treatments in AD by facilitating identification of individuals at the earliest disease stages when the treatments are most likely to be successful.⁸¹

Conclusion

In summary, this study provides evidence suggesting that SI may exacerbate cognitive impairment in Chinese adults with SCD or MCI, which is not entirely mitigated by sociodemographic and behavioral psychological factors. Furthermore, compared to men, women appear to derive greater neuroprotective benefits from robust social networks. This advantage underscores the necessity of incorporating sex-specific assessments of modifiable risk factors in dementia prevention strategies. However, the interaction between APOE ε4 carrier status and SI on cognitive outcomes has not been thoroughly investigated. Furthermore, no significant correlations were identified between social networks and p-Tau or Aβ levels in this study. These results should be interpreted with caution, however, as the limited sample size and cross-sectional design constrain the ability to establish causality or generalize the findings. Reliance solely on LSNS-6 may overlook multidimensional aspects of social connectedness (e.g., online interactions, community engagement). Future investigations should integrate multimodal longitudinal designs with sex stratification to delineate: sex-specific pathways through which SI impacts cognitive trajectories, and interactive effects between SI and APOE genotype on neural vulnerability.

Supplemental Material

sj-xlsx-1-alr-10.1177_25424823251409405 - Supplemental material for Social determinants of health, genetic biomarkers, and cognitive function in Chinese older adults with subjective cognitive decline and mild cognitive impairment: A cross-sectional analysis

Supplemental material, sj-xlsx-1-alr-10.1177_25424823251409405 for Social determinants of health, genetic biomarkers, and cognitive function in Chinese older adults with subjective cognitive decline and mild cognitive impairment: A cross-sectional analysis by Chun-Xu Shi, Jia-Huan Li, Jin Liu, Qiu-Min Zhou, Xi Yang, Yu-Qi Cheng, Nan Xia, Han-Bo Li, Fan-Fan Meng, Tong Wang and Yi Zhu in Journal of Alzheimer's Disease Reports

Footnotes

Acknowledgements

The authors thank the participants who have helped make this research possible.

ORCID iDs

Chun-Xu Shi

Jia-Huan Li

Jin Liu

Xi Yang

Yu-Qi Cheng

Han-Bo Li

Fan-Fan Meng

Yi Zhu

Ethical considerations

The study received approval from the ethics committee of Jiangsu Province Hospital.

Consent to participate

All participants provided informed consent before enrollment.

Consent for publication

Not applicable

Author contribution(s)

Chun-Xu Shi: Conceptualization; Data curation; Formal analysis; Formal analysis; Investigation; Software; Visualization; Writing – original draft.

Jia-Huan Li: Data curation; Investigation; Software; Visualization; Writing – original draft.

Jin Liu: Formal analysis; Investigation; Methodology; Writing – review & editing.

Qiu-Min Zhou: Investigation; Writing – review & editing.

Xi Yang: Data curation; Investigation; Writing – original draft.

Yu-Qi Cheng: Data curation; Investigation; Writing – original draft.

Nan Xia: Data curation; Investigation; Writing – original draft.

Han-Bo Li: Data curation; Investigation; Writing – original draft.

Fan-Fan Meng: Data curation; Investigation; Writing – original draft.

Tong Wang: Conceptualization; Formal analysis; Methodology; Writing – review & editing.

Yi Zhu: Conceptualization; Formal analysis; Formal analysis; Investigation; Methodology; Software; Writing – review & editing.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the the Science & Technology Department of Jiangsu Province, the National Key R& Program of China, (grant number Grant No.BE2023778, Nos.2023YFC3603605).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability statement

The data supporting the findings of this study are available on request from the corresponding author.

Supplemental material

Supplemental material for this article is available online.

References

de Havenon

Stulberg

Littig

, et al. Socioeconomic and medical determinants of state-level subjective cognitive decline in the United States. Alzheimers Dement 2024; 20: 7567–7579.

De Vito

Kunicki

Joyce

, et al. Parallel changes in cognition, neuropsychiatric symptoms, and amyloid in cognitively unimpaired older adults and those with mild cognitive impairment. Alzheimers Dement 2025; 21: e14568.

Ren

Savadlou

Park

, et al. The impact of loneliness and social isolation on the development of cognitive decline and Alzheimer's disease. Front Neuroendocrinol 2023; 69: 101061.

Prommas

Lwin

Chen

, et al. The impact of social isolation from COVID-19-related public health measures on cognitive function and mental health among older adults: a systematic review and meta-analysis. Ageing Res Rev 2023; 85: 101839.

Brodaty

Chau

Heffernan

, et al. An online multidomain lifestyle intervention to prevent cognitive decline in at-risk older adults: a randomized controlled trial. Nat Med 2025; 31: 565–573.

Cheng

Liu

Huang

, et al. Prevalence and risk factors for subjective cognitive decline and the correlation with objective cognition among community-dwelling older adults in China: results from the Hubei memory and aging cohort study. Alzheimers Dement 2023; 19: 5074–5085.

Madsen

Kjeldsen

Ismail

, et al. Capillary dysfunction in healthy elderly APOE ε4 carriers with raised brain aβ deposition. Alzheimers Dement 2024; 20: 459–471.

Cappa

Ribaldi

Chicherio

, et al. Subjective cognitive decline: memory complaints, cognitive awareness, and metacognition. Alzheimers Dement 2024; 20: 6622–6631.

Wang

Zhou

Zhang

, et al. Risk assessment and stratification of mild cognitive impairment among the Chinese elderly: attention to modifiable risk factors. J Epidemiol Community Health 2023; 77: 521–526.

10.

van Olst

Kamermans

Halters

, et al. Adaptive immune changes associate with clinical progression of Alzheimer's disease. Mol Neurodegener 2024; 19: 38.

11.

Jessen

Kleineidam

Wolfsgruber

, et al. Prediction of dementia of Alzheimer type by different types of subjective cognitive decline. Alzheimers Dement 2020; 16: 1745–1749.

12.

Robinson

Klinger

Buckley

, et al. Subjective cognitive decline across ethnoracial groups in the A4 study. Alzheimers Dement 2023; 19: 4084–4093.

13.

Hong

Choi

Kim

, et al. Cognitive and neurodegenerative trajectories of subjective cognitive decline according to baseline biomarkers: results of the CoSCo study. Alzheimers Dement 2025; 21: e14473.

14.

Blanchard

Akay

Davila-Velderrain

, et al. APOE4 Impairs myelination via cholesterol dysregulation in oligodendrocytes. Nature 2022; 611: 769–779.

15.

Polsinelli

Logan

Lane

, et al. APOE Ε4 carrier status and sex differentiate rates of cognitive decline in early- and late-onset Alzheimer's disease. Alzheimers Dement 2023; 19: 1983–1993.

16.

Palmer

Huentelman

Ryan

. More than just risk for Alzheimer's disease: APOE ε4's impact on the aging brain. Trends Neurosci 2023; 46: 750–763.

17.

Wang

Chen

Zhou

, et al. Integrated healthy lifestyle even in late-life mitigates cognitive decline risk across varied genetic susceptibility. Nat Commun 2025; 16: 39.

18.

Yang

Yin

, et al. Temporal trend and attributable risk factors of Alzheimer's disease and other dementias burden in China: findings from the global burden of disease study 2021. Alzheimers Dement 2024; 20: 7871–7884.

19.

TKS

Beck

Dennis

, et al. Social isolation, loneliness, and their joint effects on cognitive decline and incident Alzheimer's disease: findings from the Chicago health and aging project. J Prev Alzheimers Dis 2025; 12: 100046.

20.

Schliep

Barbeau

Lynch

, et al. Overall and sex-specific risk factors for subjective cognitive decline: findings from the 2015-2018 behavioral risk factor surveillance system survey. Biol Sex Differ 2022; 13: 16–30.

21.

Poey

Burr

Roberts

. Social connectedness, perceived isolation, and dementia: does the social environment moderate the relationship between genetic risk and cognitive well-being? Gerontologist 2017; 57: 1031–1040.

22.

Rotenberg

Leung

Quach

, et al. Occupational performance issues in older adults with subjective cognitive decline. Disabil Rehabil 2022; 44: 4681–4688.

23.

Wang

, et al. Joint association of loneliness and traditional risk factor control and incident cardiovascular disease in diabetes patients. Eur Heart J 2023; 44: 2583–2591.

24.

. Social disconnection and mortality: new evidence for old truths. Trends Cogn Sci 2023; 27: 890–891.

25.

Koszalinski

Olmos

. Communication challenges in social isolation, subjective cognitive decline, and mental health status in older adults: a scoping review (2019-2021). Perspect Psychiatr Care 2022; 58: 2741–2755.

26.

Wang

Gao

Han

, et al. A systematic review and meta-analysis of 90 cohort studies of social isolation, loneliness and mortality. Nat Hum Behav 2023; 7: 1307–1319.

27.

Wang

Zhang

Jiang

, et al. Associations of face-to-face and non-face-to-face social isolation with all-cause and cause-specific mortality: 13-year follow-up of the Guangzhou biobank cohort study. BMC Med 2022; 20: 178–192.

28.

Lavenda-Grosberg

Lalzar

Leser

, et al. Acute social isolation and regrouping cause short- and long-term molecular changes in the rat medial amygdala. Mol Psychiatry 2022; 27: 886–895.

29.

Zhang

Liu

. Determinants of social isolation of the Chinese elderly. Popul Res 2016; 40: 75–91.

30.

Yang

Wang

Edelman

, et al. Loneliness as a mediator of the impact of social isolation on cognitive functioning of Chinese older adults. Age Ageing 2020; 49: 599–604.

31.

Steptoe

Chen

, et al. Social isolation, rather than loneliness, is associated with cognitive decline in older adults: the China health and retirement longitudinal study. Psychol Med 2021; 51: 2414–2421.

32.

Morgan-Gouveia

Drago

Harris

, et al. Geriatric update 2022: preventing Alzheimer disease and more. Cleve Clin J Med 2022; 89: 617–624.

33.

Nianogo

Rosenwohl-Mack

Yaffe

, et al. Risk factors associated with Alzheimer disease and related dementias by sex and race and ethnicity in the US. JAMA Neurol 2022; 79: 584–591.

34.

Welberry

Jorm

Kiely

, et al. Sex and socioeconomic differences in 15-year prevalence trends for modifiable dementia risk factors in Australia: a cross-sectional, time series analysis. Lancet Healthy Longev 2025; 6: 100711.

35.

Vila-Castelar

Udeh-Momoh

Aggarwal

, et al. Sex and gender considerations in dementia: a call for global research. Nat Aging 2023; 3: 463–465.

36.

Wittmann

Pabst

Zülke

, et al. Who benefited the most? Effectiveness of a lifestyle intervention against cognitive decline in older women and men - secondary analysis of the AgeWell.de-trial. J Prev Alzheimers Dis 2024; 11: 348–355.

37.

Mielke

Aggarwal

Vila-Castelar

, et al. Consideration of sex and gender in Alzheimer's disease and related disorders from a global perspective. Alzheimers Dement 2022; 18: 2707–2724.

38.

Wang

Sundermann

Buckley

, et al. Sex differences in the association between tau PET and cognitive performance in a non-hispanic white cohort with preclinical AD. Alzheimers Dement 2024; 20: 25–33.

39.

Chen

Grodstein

Capuano

, et al. Late-life social activity and subsequent risk of dementia and mild cognitive impairment. Alzheimers Dement 2025; 21: e14316.

40.

Htun

Teshale

Ryan

, et al. Gender-specific analysis of social connection patterns and risk of dementia in community-dwelling older people. Alzheimers Dement 2024; 20: 4879–4890.

41.

Htun

Teshale

Ryan

, et al. Social determinants of health and risk of dementia among older men and women: a 12-year cohort study in Australia. Alzheimers Dement 2025; 21: e70065.

42.

Guo

Luo

, et al. Social isolation, loneliness and functional disability in Chinese older women and men: a longitudinal study. Age Ageing 2021; 50: 1222–1228.

43.

Spreng

Dimas

Mwilambwe-Tshilobo

, et al. The default network of the human brain is associated with perceived social isolation. Nat Commun 2020; 11: 6393–6403.

44.

Zhang

Tatewaki

Liu

, et al. Perceived social isolation is correlated with brain structure and cognitive trajectory in Alzheimer's disease. Geroscience 2022; 44: 1563–1574.

45.

Han

. Recommendations for diagnosis and treatment of subjective cognitive decline due to preclinical Alzheimer disease in China. J Chin Clin Med Imaging 2018; 29: 534–538.

46.

Jack

Jr. Bennett

Blennow

, et al. NIA-AA Research framework: toward a biological definition of Alzheimer's disease. Alzheimers Dement 2018; 14: 535–562.

47.

Islam

Hashem

Gad

, et al. Accuracy of the Montreal cognitive assessment tool for detecting mild cognitive impairment: a systematic review and meta-analysis. Alzheimers Dement 2023; 19: 3235–3243.

48.

Zhou

Gao

Wang

, et al. Application of auditory verbal learning test-huashan version in patients with subjective cognitive decline and mild cognitive impairment. Chin J Rehabil Med 2024; 39: 191–195.

49.

Wang

Zou

Cui

, et al. Reliability and construct validity of the Chinese version of the wechsler adult intelligence scale-fourth edition. Chin Mental Health J 2013; 27: 692–697.

50.

Wei

Shi

, et al. Diagnostic accuracy of the Chinese version of the trail-making test for screening cognitive impairment. J Am Geriatr Soc 2018; 66: 92–99.

51.

Cherbuin

Sachdev

Anstey

. Neuropsychological predictors of transition from healthy cognitive aging to mild cognitive impairment: the PATH through life study. Am J Geriatr Psychiatry 2010; 18: 723–733.

52.

Harry

Crowe

. Is the Boston naming test still fit for purpose? Clin Neuropsychol 2014; 28: 486–504.

53.

Zhao

Guo

Hong

. Clustering and switching during a semantic verbal fluency test contribute to differential diagnosis of cognitive impairment. Neurosci Bull 2013; 29: 75–82.

54.

Chang

Sha

Chan

, et al. Validation of an abbreviated version of the lubben social network scale (“LSNS-6”) and its associations with suicidality among older adults in China. PLoS One 2018; 13: e0201612.

55.

Hamid

Din

Bagat

, et al.

Do living arrangements and social network influence the mental health status of older adults in Malaysia?

Front Public Health 2021; 9: 624394.

56.

Foong

Ibrahim

Hamid

, et al. Social networks moderate the association between physical fitness and cognitive function among community-dwelling older adults: a population-based study. BMC Geriatr 2021; 21: 679–687.

57.

Rivière

Widad

Speyer

, et al. Reliability and validity of the French version of the global physical activity questionnaire. J Sport Health Sci 2018; 7: 339–345.

58.

Stankeviciute

Falcon

Operto

, et al. Differential effects of sleep on brain structure and metabolism at the preclinical stages of AD. Alzheimers Dement 2023; 19: 5371–5386.

59.

Davidson

Taylor

Ayers

, et al. The relationship between loneliness and positive affect in older adults. Am J Geriatr Psychiatry 2022; 30: 678–685.

60.

Imamura

Kawai

Ejiri

, et al. Association between the use of information and communication technology and cognitive decline stratified by social isolation: the otassha study. J Prev Alzheimers Dis 2025; 12: 100138.

61.

TKS

. Sex differences in the mediating role of chronic inflammation on the association between social isolation and cognitive functioning among older adults in the United States. Psychoneuroendocrinology 2023; 149: 106023.

62.

Sommerlad

Kivimäki

Larson

, et al. Social participation and risk of developing dementia. Nat Aging 2023; 3: 532–545.

63.

Huang

Zhang

Shen

, et al. Analysis of changes in social isolation, loneliness, or both, and subsequent cognitive function among older adults: findings from a nationwide cohort study. Alzheimers Dement 2024; 20: 5674–5683.

64.

Golaszewski

LaCroix

Godino

, et al. Evaluation of social isolation, loneliness, and cardiovascular disease among older women in the US. JAMA Netw Open 2022; 5: e2146461.

65.

Hsiao

Peng

Lee

, et al. Sex-specific impacts of social isolation on loneliness, depressive symptoms, cognitive impairment, and biomarkers: results from the social environment and biomarker of aging study. Arch Gerontol Geriatr 2023; 106: 104872.

66.

Santiago

Quinn

Potashkin

. Co-expression network analysis identifies molecular determinants of loneliness associated with neuropsychiatric and neurodegenerative diseases. Int J Mol Sci 2023; 24: 5909–5929.

67.

Mansour

Greenwood

Francis

, et al. Social network investment of men: cross-sectional and longitudinal associations with mental health problems. Appl Psychol Health Well Being 2024; 16: 138–157.

68.

Martens

Ren

, et al. APOE Genotype determines cell-type-specific pathological landscape of Alzheimer's disease. Neuron 2025; 113: 1380–1397.e7.

69.

Shibata

Chen

Tai

, et al. Impact of APOE, klotho, and sex on cognitive decline with aging. Proc Natl Acad Sci U S A 2025; 122: e2416042122.

70.

Nichols

Brickman

Casaletto

, et al. AD And non-AD mediators of the pathway between the APOE genotype and cognition. Alzheimers Dement 2023; 19: 2508–2519.

71.

Troutwine

Hamid

Lysaker

, et al. Apolipoprotein E and Alzheimer's disease. Acta Pharm Sin B 2022; 12: 496–510.

72.

Ding

Yin

Zhang

, et al. Diabetes accelerates Alzheimer's disease progression in the first year post mild cognitive impairment diagnosis. Alzheimers Dement 2024; 20: 4583–4593.

73.

Mielke

Evans

Neiberg

, et al. Alzheimer disease blood biomarkers and cognition among individuals with diabetes and overweight or obesity. JAMA Netw Open 2025; 8: e2458149.

74.

Bian

Tian

Meng

, et al. Assessing cognitive impairment in home-dwelling Chinese elders aged 80+: a detailed survey of 13,000 participants focusing on demographic factors, social engagement, and disease prevalence. Front Psychiatry 2024; 15: 1355708.

75.

Soldan

Pettigrew

Wang

, et al. Blood-based biomarkers and risk of onset of mild cognitive impairment over the short and long term. Neurology 2025; 104: e210225.

76.

Mao

Wang

, et al. Association of objective subtle cognitive difficulties with amyloid-β and tau deposition compared to subjective cognitive decline. Eur J Nucl Med Mol Imaging 2025; 52: 1481–1495.

77.

Ossenkoppele

Coomans

Apostolova

, et al. Tau PET positivity in individuals with and without cognitive impairment varies with age, amyloid-β status, APOE genotype and sex. Nat Neurosci 2025; 28: 1610–1621.

78.

Cody

Langhough

Zammit

, et al. Characterizing brain tau and cognitive decline along the amyloid timeline in Alzheimer's disease. Brain 2024; 147: 2144–2157.

79.

Huszár

Engh

Pavlekovics

, et al. Risk of conversion to mild cognitive impairment or dementia among subjects with amyloid and tau pathology: a systematic review and meta-analysis. Alzheimers Res Ther 2024; 16: 81.

80.

Lee

Lin

Peng

, et al. Predictive utility of plasma amyloid and tau for cognitive decline in cognitively normal adults. J Prev Alzheimers Dis 2023; 10: 178–185.

81.

Janelidze

Palmqvist

Leuzy

, et al. Detecting amyloid positivity in early Alzheimer's disease using combinations of plasma Aβ42/Aβ40 and p-tau. Alzheimers Dement 2022; 18: 283–293.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.02 MB