Marital History and Cognition in a Chinese Longevity Cohort

Abstract

Background:

Marital factor has been associated with dementia and Alzheimer’s disease, but there is limited evidence on the impact of holistic marital history over time.

Objective:

This study aimed to examine association of marital history with cognition.

Methods:

The study included 24,596 dementia-free participants from the Chinese Longitudinal Healthy Longevity Study (CLHLS). Holistic marital history was collected at baseline, categorizing participants into five groups: widow-single, widow-remarried, divorce-single, divorce-remarried and married based on the first two marriages. Dementia was collected at follow-up through self-report or from a delegate if the participant was deceased. For 15,355 participants, the Chinese Mini-Mental Status Examination (CMMSE) was administered at both baseline and follow-ups. Cognitive impairment was defined as a follow-up CMMSE score below 18, and rate of cognitive change was calculated as the change in CMMSE score between consecutive visits divided by the duration.

Results:

Compared with married older adults, widow-single group had significantly higher risk of dementia (HR 1.28, 95% CI 1.05, 1.54), cognitive impairment (HR 1.31, 95% CI 1.17, 1.47) and significantly faster decline of MMSE score (β –0.09, 95% CI –0.17, –0.01). Meanwhile, widow-remarried group had significantly lower risk of dementia, cognitive impairment and slower MMSE score decline than widow-single group, although the differences were only significant among female but not male.

Conclusions:

In this prospective cohort, married older adults and those widowed but with a second marriage had significantly better cognition than widowed individuals who did not remarry.

Keywords

Alzheimer’s disease cognitive dysfunction longitudinal studies marriage social isolation

INTRODUCTION

As the world population ages, the number of people living with dementia and cognitive impairment is increasing. The exploration of dementia risk factors is of vital importance to support an effective prevention strategy and ensuring the proper distribution of health resources to address and manage dementia. Social isolation has long been identified as a contributor to dementia, and it is estimated that 4% of dementia cases worldwide could be prevented by eliminating social isolation [1]. Marriage, as one of the most intimate forms of inter-personal relationships, might play an important role in the dementia risk profile.

With increasing life expectancy, more and more older adults are living alone after their spouses pass away, especially for females, who have a longer life expectancy than males. A meta-analysis of observational studies reported that people who are lifelong single are at a higher risk of dementia than their married counterparts. Moreover, individuals experiencing challenging events like widowhood but not divorce also had an increased risk [2]. A prospective cohort of 2 million Swedes in mid- to old-age found that single, divorced and widowed people are more likely to develop dementia than married individuals during a follow-up of 20 years [3]. Other studies have also reported similar results regarding the association of marital status with both cognitive function at one time point [4 –7] and its change over time [8 –12]. In addition to examining marital status cross-sectionally, there have also been studies considering marital transitions. In a Finland study of 2000 middle-aged adults, people without a partner at mid-life and afterwards have significantly two times higher risk of cognitive impairment and Alzheimer’s disease [13]. However, the number of studies holistically examining marital history over time is still scarce, especially remarriage after widowhood or divorce. Also, there are insufficient studies exploring sex differences in the relationship between marital factors and cognition.

In this study, our aim is to explore the association of longitudinal marital history, especially remarriage after loss of a spouse, with the risk of dementia and cognitive dysfunction using a Chinese longevity cohort. Additionally, sex differences in the relationship between marital status and risk of dementia will be examined.

METHODS

Population

The Chinese Longitudinal Healthy Longevity Study (CLHLS) is a prospective cohort aimed at promoting health for the rapidly ageing population in China. It has been recruiting and following senior residents from 22 out of 31 provinces in 8 waves, which occurred in the years 1998, 2000, 2002, 2005, 2008, 2011, 2014, and 2018. To avoid the problem of small subsample sizes at the more advanced ages, the CLHLS oversampled respondents at more advanced ages, especially males. Consequently, appropriate weights were generated based on the age-sex-rural/urban specific population distribution in the census [14]. In each follow-up wave, new participants were also recruited to compensate for any loss to follow-up of previous participants [15].

Since dementia history was collected in the second wave in the year 2000 onwards, we excluded the first wave (1998) from the analysis. As shown in Fig. 1, 3,797 out of 40,359 participants were excluded due to missing age (n = 383), missing marital history (n = 687), having more than two marriages (n = 408) or having dementia at recruitment (n = 2,319). Among the 36,562 dementia-free seniors with no more than 2 marriages, 4,439 were excluded due to missing baseline Chinese Mini-Mental State Examination (CMMSE) data. Out of the 32,123 participants with baseline information, 24,596 were followed up at least once. For participants with more than two follow-ups, data from the last wave until dementia diagnosis or death were used, and the duration of follow-up was calculated as the age difference from baseline. CMMSE was assessed for 15,355 participants during follow-up and these participants were included in the longitudinal cognition analysis.

Fig. 1

Flow chart of the study participants.

Marital history

At baseline, the whole-life marital history was inquired through a questionnaire, which included information about the participant’s current marital status and total number of marriages throughout their life span. For each marriage, participants provided detailed information, such as the age at which the marriage began, the current status of the marriage (still married, widowed, or divorced) and the age at which the marriage ended. Based on the information from the first two marriages, participants were classified into five groups: married (currently in their first marriage), widow-single (widowed and not remarried), widow-remarried (widowed and subsequently remarried), divorce-single (divorced and not remarried), and divorce-remarried (divorced and subsequently remarried). None of the participants included in the study were lifelong single. The same classification was also applied using marital history before the age of 65 to assess the impact of earlier marital experiences on the outcomes being studied.

Dementia and cognitive function assessment

At each wave of the study, the history of dementia was also collected. Participant or their proxies were asked if the older adult was suffering from dementia, and if the answer was “yes,” they were further asked whether the dementia was diagnosed from a hospital. For deceased participants, their proxies were asked if the older adult had suffered from dementia before death. Participants were considered to have dementia if they received the diagnosis from a hospital or if they were reported to suffer from dementia before death.

Cognitive function was assessed at each wave using CMMSE, which has a total score ranging from 0 to 30, with higher score indicating better cognitive function. For tasks that participants were unable to answer, a score of 0 was assigned [16]. Cognitive impairment was defined as a CMMSE score below the locally validated threshold of 18 [17]. The annual rate of CMMSE change was calculated as the CMMSE score difference between the nearest two follow ups divided by the duration in year. This rate signifies the extent to which cognitive function changes per year, and a rate below zero indicates deteriorating cognitive function of the participant over time.

Covariates

Demographic information such as sex, age, education, residence (city, town, or village), living condition (whether they live alone, with household members, or in a nursing home), and the number of children (0 or at least 1), were collected through a questionnaire. Participants were also asked to rate their own socioeconomic status (very rich, rich, so–so, poor, or very poor). Lifestyle factors were also inquired, including whether the participants engaged in regular exercise and whether they smoke or consume alcohol. Furthermore, participants were asked about their medical history including presence of any chronic heart disease, stroke, hypertension, and diabetes. Missing covariates were imputed to avoid dropping too many records, and details are in the Supplementary Material.

Statistical analysis

Survival analysis was employed to investigate associations between marital history and dementia as well as cognitive impairment. Given that older adults may pass away before the onset of dementia or cognitive impairment and considering the varying mortality risks associated with different marital statuses, we utilized cause-specific Cox regression models from the riskRegression package to accommodate competing events, such as death [18, 19]. The outcomes were categorized into dementia/cognitive impairment, death and censoring. The time of death was obtained from proxies and for individuals diagnosed with dementia or cognitive impairment, the time of event was defined as the year of follow-up when the condition was first identified. Directed acyclic graph was constructed (DAG, Fig. 2) to identify minimally sufficient set of confounders for adjustment, which included demographics (age, sex, education level, and socioeconomic status), lifestyle factors (smoking and alcohol consumption) and health conditions (chronic heart disease and hypertension). To examine the association between martial history and rate of CMMSE score change over time, a mixed-effect linear model was applied to accommodate multiple follow-ups for same participant. Similar adjustments were included in the model. Effect modification by sex was included to understand if these association might be different for male and female. To understand the importance of marriage by midlife, the associations of marital history at age of 65 and below with these outcomes were also examined. All statistical analyses were conducted in R software (version 4.3.3) with married group as reference [20].

Fig. 2

Directed acyclic graph. Green: Exposure; White: Adjusted confounders; Red: Unadjusted confounders; Blue circle with black outline: Outcome of interest; Blue: Predictors of outcome. Covariates taken into consideration include demographics (residence, age, sex, education level, and socioeconomic status), lifestyle (regular exercise, smoking and alcohol consumption) and health conditions (stroke, diabetes, chronic heart disease and hypertension). Their associations with exposure and outcome of interest were based on data used in this study. Diabetes and stroke were predictors of outcome but not associated with exposure and are thus not adjusted. Residence and regular exercise are confounders but not included in minimally sufficient set of confounders after adjustment of other factors.

RESULTS

Table 1 shows the participant characteristics at baseline. For the first marriage, there were 17,369 widowed participants (2,097 remarried), 524 divorced (376 remarried), and 6,703 who were still married. The population had a mean (SD) age of 86.8 (11.3), with the majority (56.8%) being female. Respectively 21.2%, 23.2% and 55.5% of participants were recruited from city, town and rural areas. Most participants (83.6%) lived within a household, approximately half (47.8%) estimated their socioeconomic status as average, and the majority (96.9%) had offspring. Only 31.1% engaged in regular exercise, while 19.4% were smokers and 6.0% consumed alcohol. The prevalence of chronic heart disease, history of stroke, diabetes and hypertension were 8.0%, 4.1%, 1.9%, and 54.3% respectively. The median (P25, P75) CMMSE score was 26.0 (21.0–29.0) and 935 (3.8%) participants were reported to develop dementia during the follow-up.

Table 1

Baseline characteristics of participants

Characteristics	Married	Widowed		Divorced		Total
	(N = 6703)	Single	Remarried	Single	Remarried	(N = 24,596)
		(N = 15,272)	(N = 2,097)	(N = 148)	(N = 376)
Age, mean (SD), y	77.4 (9.7)	90.6 (9.8)	89.6 (9.3)	85.0 (11.1)	84.8 (10.6)	86.8 (11.3)
Female, n (%)	2132 (31.8)	10482 (68.6)	1112 (53.0)	79 (53.4)	157 (41.8)	13962 (56.8)
Received education, n (%)	3870 (57.7)	4232 (27.7)	803 (38.3)	63 (42.6)	210 (55.9)	9178 (37.3)
Residence, n (%)
City	1449 (21.6)	3203 (21.0)	450 (21.5)	40 (27.0)	80 (21.3)	5222 (21.2)
Town	1382 (20.6)	3630 (23.8)	559 (26.7)	35 (23.6)	106 (28.2)	5712 (23.2)
Rural	3872 (57.8)	8439 (55.3)	1088 (51.9)	73 (49.3)	190 (50.5)	13662 (55.5)
Living condition, n (%)
Alone	124 (1.8)	2798 (18.3)	370 (17.6)	39 (26.4)	58 (15.4)	3389 (13.8)
With household	6541 (97.6)	11982 (78.5)	1649 (78.6)	83 (56.1)	301 (80.1)	20556 (83.6)
In nursing home	38 (0.6)	492 (3.2)	78 (3.7)	26 (17.6)	17 (4.5)	651 (2.6)
Socioeconomic status, n (%)
Very rich	898 (13.4)	4861 (31.8)	722 (34.4)	41 (27.7)	128 (34.0)	6650 (27.0)
Rich	1093 (16.3)	2184 (14.3)	353 (16.8)	19 (12.8)	54 (14.4)	3703 (15.1)
So so	3979 (59.4)	6781 (44.4)	779 (37.1)	68 (45.9)	151 (40.2)	11758 (47.8)
Poor	633 (9.4)	1221 (8.0)	189 (9.0)	18 (12.2)	36 (9.6)	2097 (8.5)
Very poor	100 (1.5)	225 (1.5)	54 (2.6)	2 (1.4)	7 (1.9)	388 (1.6)
Have offspring, n (%)	6566 (98.0)	14825 (97.1)	1994 (95.1)	120 (81.1)	336 (89.4)	23841 (96.9)
Regular exercise, n (%)	2464 (36.8)	4282 (28.0)	708 (33.8)	39 (26.4)	147 (39.1)	7640 (31.1)
Current smoker, n (%)	1982 (29.6)	2304 (15.1)	379 (18.1)	38 (25.7)	79 (21.0)	4782 (19.4)
Alcohol consumer, n (%)	753 (11.3)	571 (3.7)	111 (5.3)	8 (5.4)	33 (8.8)	1476 (6.0)
Chronic heart disease, n (%)	682 (10.2)	1075 (7.0)	153 (7.3)	8 (5.4)	38 (10.1)	1956 (8.0)
Stroke history, n (%)	367 (5.5)	547 (3.6)	80 (3.8)	4 (2.7)	15 (4.0)	1013 (4.1)
Diabetes, n (%)	211 (3.1)	237 (1.6)	27 (1.3)	0 (0.0)	3 (0.8)	478 (1.9)
Hypertension, n (%)	3758 (56.1)	8165 (53.5)	1170 (55.8)	66 (44.6)	206 (54.8)	13365 (54.3)
Hypotension, n (%)	85 (1.3)	221 (1.4)	32 (1.5)	4 (2.7)	5 (1.3)	347 (1.4)
CMMSE, median (P25-P75)	29.0 (26.0–30.0)	25.0 (18.0–28.0)	26.0 (20.0–29.0)	27.0 (20.0–29.0)	27.0 (23.0–29.0)	26.0 (21.0–29.0)
Dementia, n (%)	185 (2.8)	645 (4.2)	82 (3.9)	6 (4.1)	17 (4.5)	935 (3.8)

Table 2 shows the association between marital history and dementia risk. Compared to individuals still in their first marriage, those in the widow-single group exhibited a significantly higher hazard of dementia (HR 1.28, 95% CI 1.05, 1.54), with a similar association observed in females but not in males. No significant difference to married group was observed for widow-remarried (HR 1.00, 95% CI 0.66, 1.51), divorce-single (HR 2.16, 95% CI 0.86, 5.41), or divorce-remarried (HR 1.03, 95% CI 0.41, 2.57) groups among female. There was no significant effect modification by sex. Additionally, in post-estimation, the hazard of dementia was significantly lower for widow-remarried females compared to widow-single females, but this association was not significant among males.

Table 2

Association between marital history and dementia risk

Group	Female		Male		Full sample
	HR (95% CI)	p	HR (95% CI)	p	HR (95% CI)	p
Married	Ref.		Ref.		Ref.
Widow-single	1.40 (1.05, 1.88)	0.02	1.17 (0.92, 1.49)	0.20	1.28 (1.05, 1.54)	0.01
Widow-remarried	1.00 (0.66, 1.51)^#	>0.99	1.26 (0.88, 1.82)	0.21	1.08 (0.82, 1.42)	0.59
Divorce-single	2.16 (0.86, 5.41)	0.10	0.51 (0.07, 3.66)	0.50	1.37 (0.61, 3.10)	0.44
Divorce-remarried	1.03 (0.41, 2.57)	0.95	1.79 (0.99, 3.24)	0.06	1.45 (0.88, 2.39)	0.15

Cause-specific Cox regression model. Adjustment of baseline covariates: age, sex, education level, socioeconomic status, smoking, alcohol consumption, chronic heart disease and hypertension. ^#Significantly different from widow-single group.

Table 3 shows the association of marital history with the risk of cognitive impairment. Individuals in the widow-single group had a significantly higher hazard of cognitive impairment compared to the married group (HR 1.31, 95% CI 1.17, 1.47), with this difference being significant in both males and females. However, the widow-remarried (HR 1.12, 95CI 0.96, 1.30), divorce-single (HR 1.22, 95% CI 0.75, 1.98), and divorce-remarried (HR 1.20, 95% CI 0.86, 1.65) groups did not exhibit significantly different hazards of cognitive impairment compared to the married group. There was no significant effect modification by sex observed for these associations. Furthermore, in comparison to the widow-single group, the hazard of cognitive impairment was significantly lower in the widow-remarried group, with this association being significant only in females but not in males.

Table 3

Association between marital history and risk of cognitive impairment

Group	Female		Male		Full sample
	HR (95% CI)	p	HR (95% CI)	p	HR (95% CI)	p
Married	Ref.		Ref.^#		Ref.^#
Widow-single	1.24 (1.06, 1.45)	0.007	1.36 (1.17, 1.58)	<0.001	1.31 (1.17, 1.47)	<0.001
Widow-remarried	1.00 (0.82, 1.23)^#	>0.99	1.29 (1.03, 1.62)	0.03	1.12 (0.96, 1.30)^#	0.16
Divorce-single	1.29 (0.72, 2.32)	0.39	1.00 (0.41, 2.41)	>0.99	1.22 (0.75, 1.98)	0.43
Divorce-remarried	1.04 (0.66, 1.65)	0.87	1.36 (0.86, 2.14)	0.19	1.20 (0.86, 1.65)	0.28

Table 4 presents the association of marital history with the rate of CMMSE score change over time. The widow-single group exhibited a significantly faster decline in MMSE score (β –0.09, 95% CI –0.17, –0.01) compared to married subjects, and this association was only significant in female. However, the widow-remarried (β 0.07, 95% CI –0.06, 0.20), divorce-single (β 0.00, 95% CI –0.25, 0.25) and divorce-remarried (β –0.21, 95% CI –0.57, 0.14) groups did not demonstrate significantly different rates of MMSE score change. No significant effect modification by sex was identified. Through post-estimation, widow-remarried group was found to have a significantly slower decline in MMSE score than the widow-single group, and this association was only significant for females but not males.

Table 4

Association between marital history and rate of change in CMMSE score

Group	Female	Male	Full sample
	β (95% CI)	β (95% CI)	β (95% CI)
Married	Ref.	Ref.	Ref.
Widow-single	–0.16 (–0.27, –0.04)	–0.02 (–0.13, 0.09)	–0.09 (–0.17, –0.01)
Widow-remarried	0.07 (–0.11, 0.26)^#	0.05 (–0.12, 0.23)	0.07 (–0.06, 0.20)^#
Divorce-single	–0.38 (–0.94, 0.18)	–0.06 (–0.65, 0.53)	0.00 (–0.25, 0.25)
Divorce-remarried	0.15 (–0.25, 0.54)	–0.11 (–0.44, 0.22)	–0.21 (–0.57, 0.14)

Mixed-effect linear model. Adjustment of baseline covariates: age, sex, education level, socioeconomic status, smoking, alcohol consumption, chronic heart disease and hypertension. ^#Significantly different from widow-single group.

DISCUSSION

In this study, we found that individuals not experiencing challenging marital events, such as losing a spouse, have lower risk of dementia and cognitive impairment compared to widowed individuals who did not remarry. Additionally, this group experienced a slower decline in cognition. Among the widowed older adults, those who remarried had lower risk of dementia and cognitive impairment, as well as slower decline in cognition compared to those who did not remarry. However, when examined separately by sex, the significant associations in dementia and cognitive decline were observed among only female but not male participants.

Results of this study are consistent with existing literature, that older individuals without a partner (widowed or divorced) are at a higher risk of dementia compared to their married counterparts [2]. Moreover, the study findings suggest that the influence of spousal loss on cognition might accumulate over time instead of being immediately determined, so a second marriage might make some remedy. In this study, widowed individuals who have a second partner (remarried) exhibited better cognition compared to those who remain in widowhood. Several other studies have investigated the association between marital transition and cognitive health, like transition from being married to unmarried [13] and experiencing divorce during study period [21]. To the best of our knowledge, this study is the first to examine the association between remarriage after the loss of a spouse and cognition. Due to the relatively small sample size of divorced individuals (n = 524, with 376 remarried), the study could not fully explore the association between remarriage after divorce and dementia or cognitive impairment.

For dementia risk profile, gender is an important factor that should be taken into consideration, especially when investigating marital factors as predictors of dementia, for interplay of gender with socioeconomic factors [22]. In a hypertensive cohort of Chinese older adults, males showed a significant association between marital status and MMSE score but this association was not observed for female [7]. In this national-representative cohort of older adults, we specifically examined effect modification by sex. It was found that remarriage was associated with significantly lower dementia risk only in widowed females but not males, although a lower dementia risk was observed for married individuals in both genders. For future studies examining the relationship between marital factors and health outcomes, it is essential to exercise more caution and consider males and females separately. This approach will help to better understand the gender-specific effects and implications of marital factors on cognitive function and dementia risk.

In the Finland cohort, marital status was collected at both baseline and follow-up to construct marital transitions [13]. However, information before and during the study was missing with this method. Additionally, dementia diagnosis was made at the time of follow-up, making it challenging to establish temporality between marital transitions and dementia incidence. In another study in Norway, yearly marital status was retrieved from the population registry from age 44 to 68 [21]. This approach helped to avoid bias from self-reported history, but it may not always be applicable due to concerns over privacy breaches related with domestic registration systems. In this study, holistic marital history for the entire lifespan was collected at baseline using a questionnaire. Although information bias from this method of data collection might occur for minor events, significant events like marriage are less likely to be affected. This method allowed gathering of detailed marital history data for each participant, providing valuable insights into the association between marital history and dementia risk.

From 2000 to 2021, the proportion of seniors (aged 65 and above) in China increased from 7% to 14% and this trend is projected to continue [23]. This significant increase of older population has led to a greater focus on health demands for them, prompting increased investment in senior care. According to the concept of Marriage Protection Effect (MPE), unmarried older adults are at a disadvantage in terms of health compared to married ones, which extends to various aspects of health, including general well-being, mental health, and happiness [24, 25]. Consistent with these findings, our study also revealed that widowed older adults are at a higher risk of dementia and cognitive impairment. These results highlight the importance of paying more attention to the health and well-being of older adults who do not have a spouse.

Several limitations in this study that should be acknowledged. Firstly, the dementia outcome was self-reported, and despite efforts to confirm diagnoses made by hospitals through a second question, there might still be undiagnosed dementia subjects. According to China National Sample Survey on Disability (CNSSD), unmarried Chinese older adults are known to have lower utilization of mental health services [26], which may result in a higher proportion of undiagnosed dementia cases in the unmarried group of this study. This differential misclassification of the outcome could potentially bias the associations in the study towards null. However, this is unlikely to be the sole cause of the significant associations observed in this study. Second, the outcome of dementia was collected relatively shortly after the baseline assessment (median 4 years, P25-P75 2–7 years) and it is possible that some participants may have been in the prodromal stage of dementia at the time of assessment. To address this limitation, the study also included outcomes of CMMSE assessment (cognitive impairment and score change) to corroborate the results from the outcome of dementia. Also, we examined marital history before age of 65 to avoid influence of prodromal dementia (Supplementary Tables 1 and 2), and the trend was similar that married or remarried female had lower odds of dementia and cognitive impairment compared with widowed ones who do not have a second partner. Third, only 29.2% of this population received any form of education due to limited accessibility to education during school years of this study population. However, it is worth noting that education coverage and accessibility have significantly improved in China over the years. Therefore, the dementia risk profile may be different for later generations with higher educational backgrounds. Future studies are needed to investigate the association between marriage and dementia in populations with higher education levels to understand how educational attainment may influence dementia risk.

In this national representative cohort of Chinese older adults, our findings indicate that widowed individuals as opposed to divorced individuals, experience faster cognitive decline and a higher risk of dementia and cognitive impairment compared to married individuals. These results highlight the potential importance of marital status as a factor influencing cognitive health in later life. Moreover, our study has uncovered a novel finding that widowed females who enter a new relationship with a new partner experience slower cognitive decline and a lower risk of dementia and cognitive impairment. This observation suggests that remarriage may have a protective effect on cognitive health for this particular subgroup of widowed females. As the demographic landscape of Chinese seniors continues to change, with increasing education coverage, informal cohabitation, and a rise in lifelong singlehood, it becomes crucial to investigate the implications of these changes on cognitive health in younger generations. Future studies focusing on younger age groups will provide valuable evidence and insights into the complex interplay between demographic factors and cognitive health outcomes over time.

AUTHOR CONTRIBUTIONS

Xiangyuan Huang (Conceptualization; Formal analysis; Writing – original draft); Saima Hilal (Conceptualization; Funding acquisition; Methodology; Supervision; Writing – review & editing).

Footnotes

ACKNOWLEDGMENTS

The authors express their special appreciation to Professor Chuen Seng Tan, who has generously provided invaluable expertise in statistical analysis and result interpretation for this study.

FUNDING

S. Hilal was supported by National Medical Research Council Singapore, Transition Award [A-0006310-00-00], Ministry of Education, Academic Research Fund Tier 1 [A-0006106-00-00] and Absence Leave Grant [A-8000336-00-00].

CONFLICT OF INTEREST

Saima Hilal is an Editorial Board Member of this journal but was not involved in the peer-review process of this article nor had access to any information regarding its peer-review.

Xiangyuan Huang has no conflict of interest to report.

DATA AVAILABILITY

The data supporting findings of this study is openly accessible at CLHLS ().

The supplementary material is available in the electronic version of this article: .

References

Livingston

, Huntley

, Sommerlad

, Ames

, Ballard

, Banerjee

, Brayne

, Burns

, Cohen-Mansfield

, Cooper

, Costafreda

, Dias

, Fox

, Gitlin

, Howard

, Kales

, Kivimaki

, Larson

, Ogunniyi

, Orgeta

, Ritchie

, Rockwood

, Sampson

, Samus

, Schneider

, Selbaek

, Teri

, Mukadam

(2020) Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. Lancet 396, 413–446.

Sommerlad

, Ruegger

, Singh-Manoux

, Lewis

, Livingston

(2018) Marriage and risk of dementia: Systematic review and meta-analysis of observational studies. J Neurol Neurosurg Psychiatry 89, 231–238.

Sundstrom

, Westerlund

, Kotyrlo

(2016) Marital status and risk of dementia: A nationwide population-based prospective study from Sweden. BMJ Open 6, e008565.

Jain

, Liu

, Langa

, Farron

, Kabeto

, Lee

(2022) Widowhood and cognition among older women in India: New insights on widowhood duration and mediators. SSM Popul Health 19, 101242.

Jennings

, Farrell

, Liu

, Montana

(2022) Associations between cognitive function and marital status in the United States, South Africa, Mexico, and China. SSM Popul Health 20, 101288.

Perkins

, Lee

, James

, Oh

, Krishna

, Heo

, Lee

, Subramanian

(2016) Marital status, widowhood duration, gender and health outcomes: A cross-sectional study among older adults in India. BMC Public Health 16, 1032.

Shen

, Cheng

, Li

, Xie

, Wang

, Zhou

, Zhu

, Wang

, Tu

, Bao

, Cheng

(2022) Association of marital status with cognitive function in Chinese hypertensive patients: A cross-sectional study. BMC Psychiatry 22, 504.

Biddle

, Jacobs

HIL

, d’Oleire Uquillas

, Zide

, Kirn

, Properzi

, Rentz

, Johnson

, Sperling

, Donovan

(2020) Associations of widowhood and beta-amyloid with cognitive decline in cognitively unimpaired older adults. JAMA Netw Open 3, e200121.

Kim

, Lee

(2019) Social network types and cognitive decline among older Korean adults: A longitudinal population-based study. Int J Geriatr Psychiatry 34, 1845–1854.

10.

Min

, Song

(2023) Spousal loss and cognitive function: The importance of gender and multiple dimensions of marital quality. Aging Ment Health 27, 755–764.

11.

van Gelder

, Tijhuis

, Kalmijn

, Giampaoli

, Nissinen

, Kromhout

(2006) Marital status and living situation during a 5-year period are associated with a subsequent 10-year cognitive decline in older men: The FINE Study. J Gerontol B Psychol Sci Soc Sci 61, P213–219.

12.

Zaheed

, Sharifian

, Morris

, Kraal

, Zahodne

(2021) Associations between life course marital biography and late-life memory decline. Psychol Aging 36, 557–571.

13.

Hakansson

, Rovio

, Helkala

, Vilska

, Winblad

, Soininen

, Nissinen

, Mohammed

, Kivipelto

(2009) Association between mid-life marital status and cognitive function in later life: Population based cohort study. BMJ 339, b2462.

14.

Zeng

, Gu

, Poston

, Vlosky

(2008) Healthy longevity in China: Demographic, socioeconomic, and psychological dimensions, Springer.

15.

Gao

, Kuang

, Qiu

, Wang

, Lv

, Yang

(2017) The time trends of cognitive impairment incidence among older Chinese people in the community: Based on the CLHLS cohorts from 1998 to 2014. Age Ageing 46, 787–793.

16.

Zhang

(2006) Gender differentials in cognitive impairment and decline of the oldest old in China. J Gerontol B Psychol Sci Soc Sci 61, S107–S115.

17.

Zhang

, Zahner

, Roman

, Liu

, Wu

, Hong

, Tang

, Zhou

, Qu

, Zhang

, Li

(2006) Socio-demographic variation of dementia subtypes in China: Methodology and results of a prevalence study in Beijing, Chengdu, Shanghai, and Xian. Neuroepidemiology 27, 177–187.

18.

Ozenne

, Sørensen

, Scheike

, Torp-Pedersen

, Gerds

(2017) riskRegression: Predicting the risk of an event using Cox regression models. R J 9, 440–460.

19.

Wang

, Yi

(2023) Marital status and all-cause mortality rate in older adults: A population-based prospective cohort study. BMC Geriatr 23, 214.

20.

R Core Team (2023) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.

21.

Skirbekk

, Bowen

, Haberg

, Jugessur

, Engdahl

, Bratsberg

, Zotcheva

, Selbaek

, Kohler

, Weiss

, Harris

, Tom

, Krokstad

, Stern

, Strand

(2023) Marital histories and associations with later-life dementia and mild cognitive impairment risk in the HUNT4 70+Study in Norway. J Aging Health 35, 543–555.

22.

Zhang

, Liu

, Choi

(2021) Marital loss and risk of dementia: Do race and gender matter? Soc Sci Med 275, 113808.

23.

Chen

, Giles

, Yao

, Yip

, Meng

, Berkman

, Chen

, Feng

, Glinskaya

, Gong

, Hu

, Kan

, Lei

, Liu

, Steptoe

, Wang

, Yang

, Zhang

, Wu

, Strauss

, Smith

, Zhao

(2022) The path to healthy ageing in China: A Peking University-Lancet Commission. Lancet 400, 1967–2006.

24.

, Noguchi

(2016) Does marriage make us healthier? Inter-country comparative evidence from China, Japan, and Korea. PLoS One 11, e0148990.

25.

Giannelis

, Palmos

, Hagenaars

, Breen

, Lewis

, Mutz

(2021) Examining the association between family status and depression in the UK Biobank. J Affect Disord 279, 585–598.

26.

Guo

, Li

, Chen

, Zheng

(2019) Mental health service utilization and its associated social factors among elderly people with a mental disability in China: A national population-based survey. Scand J Public Health 47, 215–220.