Abstract
INTRODUCTION
This study aimed to assess whether mid‐life social relationships are associated with a lower risk of late‐life dementia.
METHODS
Participants in the Atherosclerosis Risk in Communities (ARIC) study were assessed for social support and isolation (Visit 2;1990–1992). A composite measure, “social relationships,” was generated. Incident dementia cases were identified following Visit 2 through 2019, using ongoing surveillance. Associations between mid‐life social relationships and incident dementia were evaluated with Cox proportional‐hazard regression models. Formal interaction tests examined whether sex or race modified this association.
RESULTS
Among 13,070 participants without dementia, those with strong social relationships in mid‐life had a lower risk for developing dementia, compared to participants with poor mid‐life social relationships. Neither sex nor race significantly modified this association.
DISCUSSION
Stronger mid‐life social relationships may have a potentially protective effect on dementia risk. Future studies evaluating psychosocial health at multiple time points in ethnically diverse populations are needed.
Highlights
Psychosocial health is a modifiable risk factor for dementia.
Stronger mid‐life social relationships are associated with a lower risk of dementia.
Social relationships may have a potentially protective effect on late‐life outcomes.
This study leveraged data collected from nearly 20 years of follow‐up in 13,070 participants.
Longitudinal data used in this study were captured in a population that is 24% Black.
Keywords: Atherosclerosis Risk in Communities (ARIC) study, dementia, longitudinal follow‐up, mid‐life, psychosocial health, social relationships
1. BACKGROUND
Psychosocial health has gained widespread recognition as a key factor in preserving cognition and sustained functional independence in aging adults. 1 , 2 As such, numerous observational cohort studies have examined the association between different indices of psychosocial health and related social behaviors with long‐term cognitive function and the risk of incident dementia. 3 , 4
Many of these studies examining the association between psychosocial health with dementia and related late‐life outcomes remain limited to small (n <210) 5 , 6 , 7 or racially homogeneous samples, 8 , 9 , 10 , 11 , 12 making findings less generalizable to the broader population. Others are cross‐sectional 13 , 14 or have follow‐up periods of less than 10 years, rendering them susceptible to reverse causation. 6 , 7 , 9 , 15 , 16 , 17 , 18 Still, other studies have evaluated the impact of psychosocial health over longer follow‐up periods, but the outcome of interest is cognitive decline in specified domains (e.g., processing speed, verbal memory, executive function). 7 , 14 , 19 , 20 Altogether, findings from these studies have indicated that psychosocial health may have a potentially protective effect on late‐life outcomes, 4 , 5 , 7 , 10 , 11 , 13 , 21 although the reported magnitude of this association varies likely due to differences in study design, the constructs used to define psychosocial health, and the outcomes examined.
Through data from 13,070 U.S. adults enrolled in the community‐based Atherosclerosis Risk in Communities (ARIC) cohort, we hypothesized that participants with stronger social relationships in mid‐life would have a lower risk of dementia in late‐life. The component of psychosocial health assessed was “social relationships,” operationalized through joint classification of measures used to categorize perceived social support and social isolation. Two factors were key to our investigation: (1) the use of a follow‐up period that was nearly 20 years long, which allowed us to more accurately report observed associations between social relationships and dementia risk; and (2) the examination of incident dementia as the outcome, due to its growing prevalence underscored by increasing necessity to understand how components of psychosocial health relate to clinical change substantial enough to warrant a medical diagnosis.
Secondary analyses examined effect modification of the association between social relationships in mid‐life and dementia risk in late‐life by sex and race. Finally, due to growing interest in understanding the interplay between co‐existing genetic factors and additional lifestyle factors adjacent to psychosocial health, 22 we further examined whether these factors modify the association between social relationships in mid‐life and dementia risk.
2. METHODS
2.1. Study population
The ARIC study recruited 15,792 adults (aged 45 to 64 years) from four U.S. communities at baseline (Visit 1:1987–1989). 23 Self‐reported questionnaires regarding psychosocial measures were collected in 14,348 participants at Visit 2 (1990–1992). Cognition was further assessed at Visit 2 using a three‐instrument cognitive assessment. Only participants who did not have dementia and had psychosocial health assessed at Visit 2 were included in the present study. Follow‐up for development of incident dementia was though December 31, 2019. For roughly 30% of participants seen at the Jackson site, follow‐up data for this study was censored through December 31, 2017, due to administrative delays in data reporting. Further details of the study timeline (Figure S1) and exclusion criteria are described (Figure 1). In total, the study used longitudinal data from 13,070 participants. The study was approved by institutional review boards at each study center, and all participants provided written informed consent to participate in each study visit and follow‐up.
FIGURE 1.
Exclusion criteria for analytic sample (N = 13,070). Lost to follow‐up including but not limited to participants who developed dementia and/or died.
2.2. Psychosocial measures
Perceived social support was evaluated at Visit 2 using the Interpersonal Support Evaluation List‐Short Form (ISEL‐SF; Figure S2). This 16‐item scale was constructed and validated by ARIC investigators from the original 40‐item full scale and assesses perceived social support with four subscales including (1) appraisal support, (2) tangible assets support, (3) belonging support, and (4) self‐esteem support. 24 , 25 The total score is an equally weighted sum, with a range of 0–48. Higher scores indicate greater perceived social support. Total ISEL‐SF score was categorized into distribution‐based tertiles of the full ARIC sample assessed at Visit 2 (high ≥42, intermediate 36–41, and low ≤35). 26
Social isolation was evaluated using the Lubben Social Network Scale (LSNS; Figure S3), also administered at Visit 2. This 10‐item scale assesses the size of the participant's active social network and the perceived social support received by family, friends, and neighbors. The total score is an equally weighted sum, with a range of 0–50. Higher scores indicate a lower risk of social isolation. Scores were divided into four categories based upon the Lubben criteria, which has been used in many analyses conducted in the ARIC cohort: ≤20 = isolated; 21–25 = high risk for isolation; 26–30 = moderate risk for isolation; ≥31 = low risk for isolation. 19 , 26 , 27 , 28
Following the categorization of both psychosocial measures, participants fell into 12 different groups. Performance on the ISEL‐SF and LSNS was paired to create a measure reflecting social relationships (Figure 2). Participants were classified as having strong, average, or poor social relationships in mid‐life. This conversion into a categorical composite measure that encompasses both psychosocial measures was conducted for several reasons. First, the correlation between continuous ISEL‐SF and LSNS scores in this sample is moderate (r = 0.46), which may indicate that the two constructs are not entirely linear with one another. Second, previous studies in the ARIC‐PET (positron emission tomography) cohort have shown stronger associations with dementia risk when using the “social relationships” variable as opposed to independent scales. 29 Finally, the use of an integrative, composite measure allows us to assess multiple facets of psychosocial health more comprehensively, which aligns with recent recommendations to study healthy aging with a holistic approach. 21 , 22 , 30 , 31
FIGURE 2.
Categorization of social relationships (N = 13,070).
2.3. Dementia
The 3‐instrument cognitive battery (Delayed Word Recall, Word Fluency, and Digit Symbol Substitution) initially used at Visit 2 was repeated at Visit 4 (1996–1998), and an expanded battery was administered at Visit 5 as part of the ARIC Neurocognitive Study (ARIC‐NCS; 2011–2013). 32 , 33 , 34 Incident dementia cases were identified in several ways. First, as part of the ARIC‐NCS, participants seen in‐person for Visits 5–7 underwent detailed cognitive testing, and a subset had informant interviews. Participants seen in‐person for Visits 5–7 underwent a physician‐ and neuropsychologist‐led adjudication process and standard diagnostic criteria, which led to a research diagnosis of normal cognition, mild cognitive impairment (MCI), or dementia. 32 , 33 , 34 Second, in addition to in‐person visits, dementia surveillance was ongoing by use of telephone‐based cognitive assessments, informant interviews, and hospitalization codes and death certificates. Additional post‐Visit 5 incident dementia cases were identified through several sources, including administration of the Six Item Screener (SIS) annually by phone and the Ascertain Dementia 8‐Item Informant Questionnaire (AD8), when appropriate. 32 , 33 , 34 Date of dementia onset was defined as the earliest date of SIS/AD8 interviews, date of hospitalization records with dementia diagnosis, or the date of the in‐person visit when a participant was classified as having dementia. Of participants with diagnoses based on SIS/AD8, hospitalization records, or death certificates, dementia onset was defined as 180 days prior to the interview, hospitalization, or death. 34 , 35 , 36
RESEARCH IN CONTEXT
Systematic review: The authors reviewed the literature using PubMed for studies examining psychosocial health and long‐term health outcomes. This is a growing area of interest, yet there remains a strong need for longitudinal studies that specifically assess the association between behaviors underlying social relationships with dementia as the primary outcome. Relevant existing publications are cited.
Interpretation: In this large community‐based cohort of 13,070 individuals with roughly 20 years of follow‐up, our results provide important evidence that stronger social relationships in mid‐life are associated with a lower risk of incident dementia in late‐life.
Future directions: Our findings provide further evidence of the potentially protective role of strong psychosocial health in mid‐life. Future studies can build upon these findings by continuing to evaluate psychosocial health throughout the life course in racial and ethnically diverse populations with additional characterization of related lifestyle factors and social determinants of health.
2.4. Covariates
Covariates included age (at Visit 2), sex, educational attainment (less than high school, high school or equivalent, or more than high school), race, and apolipoprotein E (APOE) ɛ4 genotype (0 or ≥1 allele). Race and study center were combined to create a race‐center variable that accounted for the distribution of racial groups across ARIC centers. Occupational and marital status were self‐reported at Visit 2 and categorized as “employed and working outside the home” or “not working outside the home” (which included being retired, unemployed, or a homemaker) and “married” or “not married” (divorced, separated, widowed, or never married), respectively. The Maastricht Vital Exhaustion Questionnaire, administered at Visit 2 to measure symptoms of depression and fatigue, was dichotomized at ≥14 to indicate depressive symptomology. 37 , 38 , 39 Vascular risk factors, including hypertension (systolic blood pressure >140 mm Hg, diastolic blood pressure >90 mm Hg, or use of antihypertensive medications), diabetes (fasting glucose ≥126 mg/dL, non‐fasting glucose ≥200 mg/dL, HbA1c ≥6.5, self‐report of physician‐diagnosed diabetes, or use of oral diabetes medications or insulin), smoking/drinking (self‐report, dichotomized into ever vs never), obesity (BMI ≥30 kg/m2), and elevated total cholesterol (≥200 mg/dL) were collected at Visit 2.
2.5. Statistical analysis
R version 4.3.1 (R Foundation for Statistical Computing) was used for all analyses. Participant characteristics were evaluated, and Cox proportional‐hazards regression models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for the association between social relationships and incident dementia from ARIC Visit 2 through December 31, 2019 (with 5 years after Visit 2 considered to be time 0). Three covariate‐adjusted models were fit to the data: Model 1 adjusted for demographics and APOE ɛ4; Model 2 additionally adjusted for occupational status, marital status, and depressive symptoms at Visit 2; and Model 3 further adjusted for vascular risk factors measured at Visit 2. Because of the possibility that vascular risk factors are on the causal pathway between psychosocial measures and dementia (and thus are not potential confounders), our primary analyses and inferences were based on Model 2 because Model 3 may be over‐adjusted. Secondary analyses explored the association between social relationships and incident dementia with Wald tests for interaction by sex and race, prior to stratification by both measures. In addition, we evaluated interactions between social relationships with age (dichotomized at the Visit 2 sample median), APOE ɛ4 genotype, depressive symptoms, and educational attainment. P < 0.05 was considered statistically significant; testing was two‐sided. Multiple imputation chained equations (MICE) were used to impute missing covariates in 820 participants.
2.6. Sensitivity analyses
Sensitivity analyses assessed the independent contributions of social support and social isolation as categorical measures, since these factors were paired to create the “social relationships” variable. The purpose of these analyses was to see whether the observed effects of psychosocial measures on dementia risk were congruent across these different behaviors comprising the social relationships variable. Additional sensitivity analyses were conducted, excluding participants whose global cognitive function fell in the lowest fifth percentile at Visit 2 or anyone who had an incident stroke prior to the end of follow‐up. These analyses were conducted to confirm that the associations shown were not driven by those at the highest risk of dementia due to extrinsic factors such as low baseline cognitive function or stroke. We performed a Fine and Gray sub‐distribution competing risk analysis with death as a competing outcome. 40
3. RESULTS
3.1. Participant characteristics
The analytic sample included 13,070 participants (7268 [56%] women, 3127 [24%] Black, Visit 2 median [interquartile range; IQR] age, 57 [10] years). Fifty‐seven percent of participants had “high” or “intermediate” social support and risk of social isolation was “low” or “moderate,” and they were classified as having strong mid‐life social relationships, whereas 28% had average mid‐life social relationships, and 15% had poor mid‐life social relationships (Figure 2). The median (IQR) follow‐up (after accounting for a 5‐year lag exposure period following Visit 2) was 19.3 (9.1) years, during which 2643 (20%) of participants developed dementia (Table 1). The incidence rate (per 1000 person‐years) was 14.2 for participants with poor mid‐life social relationships, 12.7 for those with average mid‐life social relationships, and 11.1 for those with strong mid‐life social relationships.
TABLE 1.
ARIC participant characteristics of the analytic sample at Visit 2 (1990–1992); stratified by level of social relationships (N = 13,070).
| Characteristics |
Full sample (N = 13,070) |
Strong social relationships (n = 7485) |
Average social relationships (n = 3674) |
Poor social relationships (n = 1911) |
p‐value |
|---|---|---|---|---|---|
| Age, y, median (Q1, Q3) | 57 (52, 62) | 56 (52, 61) | 57 (52, 62) | 57 (52, 62) | 0.001 |
| Sex, n (%) | <0.001 | ||||
| Women | 7268 (55.6) | 4417 (59.0) | 1954 (53.2) | 897 (46.9) | |
| Men | 5802 (44.4) | 3068 (41.0) | 1720 (46.8) | 1014 (53.1) | |
| Race‐ARIC Center, n (%) | 0.003 | ||||
| Black – Forsyth | 353 (2.7) | 191 (2.6) | 93 (2.5) | 69 (3.6) | |
| Black – Jackson | 2774 (21.2) | 1514 (20.2) | 820 (22.3) | 440 (23.0) | |
| White – Forsyth | 3119 (23.9) | 1863 (24.9) | 853 (23.2) | 405 (21.2) | |
| White – Minnesota | 3534 (27.0) | 2056 (27.5) | 992 (27.0) | 486 (25.4) | |
| White – Washington | 3290 (25.2) | 1861 (24.9) | 916 (24.9) | 511 (26.7) | |
| APOE ɛ4 status, n (%) | 0.04 | ||||
| Non‐carrier | 8763 (67.0) | 5100 (68.1) | 2413 (65.7) | 1250 (65.4) | |
| Carrier | 3894 (29.8) | 2161 (28.9) | 1145 (31.2) | 588 (30.8) | |
| Education level, n (%) | <0.001 | ||||
| Less than high school | 2682 (20.5) | 1297 (17.3) | 894 (24.3) | 491 (25.7) | |
| High school or equivalent | 4101 (31.4) | 2352 (31.4) | 1173 (31.9) | 576 (30.1) | |
| More than high school | 6269 (48.0) | 3828 (51.1) | 1600 (43.5) | 841 (44.0) | |
| Occupational status, n (%) | 0.22 | ||||
| Employed outside the home | 8694 (66.5) | 5032 (67.2) | 2420 (65.9) | 1242 (65.0) | |
| Not working outside the home a | 4376 (33.5) | 2453 (32.8) | 1254 (34.1) | 669 (35.0) | |
| Marital status, n (%) | <0.001 | ||||
| Married | 10,399 (79.6) | 6275 (83.8) | 2917 (79.4) | 1207 (63.2) | |
| Not married b | 2658 (20.3) | 1202 (16.1) | 753 (20.5) | 703 (36.8) | |
| Depressive symptoms, c n (%) | 3847 (29.4) | 1528 (20.4) | 1400 (38.1) | 919 (48.1) | <0.001 |
| Hypertension, n (%) | 4557 (34.9) | 2511 (33.5) | 1334 (36.3) | 712 (37.3) | 0.004 |
| Diabetes, n (%) | 1840 (14.1) | 983 (13.1) | 561 (15.3) | 296 (15.5) | 0.004 |
| Body mass index (≥30 kg/m2), n (%) | 3758 (28.8) | 2113 (28.2) | 1106 (30.1) | 539 (28.2) | 0.21 |
| Total cholesterol (≥200 mg/dL), n (%) | 7386 (56.5) | 4233 (56.6) | 2094 (57.0) | 1059 (55.4) | 0.70 |
| Ever smoker, n (%) | 7808 (59.7) | 4238 (56.6) | 2302 (62.7) | 1268 (66.4) | <0.001 |
| Ever drinker, n (%) | 10,153 (77.7) | 5719 (76.4) | 2862 (77.9) | 1572 (82.3) | <0.001 |
| Developed dementia by Visit 7, n (%) | 2643 (20.2) | 1449 (19.4) | 770 (21.0) | 424 (22.2) | 0.03 |
Notes: Data for the following covariates were missing and are not included in the proportions shown above: APOE ɛ4 (n = 413), education (n = 18), marital status (n = 13), depressive symptoms (n = 193), hypertension (n = 39), diabetes (n = 56), total cholesterol (n = 61), body mass index (n = 17), smoking (n = 7), and drinking (n = 8). Multiple imputation chained equations were applied to all missing covariates prior to modeling. SI conversion: To convert cholesterol to millimoles per liter, multiply by 0.0259.
Abbreviation: APOE, apolipoprotein E; ARIC, Atherosclerosis Risk in Communities
Not working outside the home including participants who identified as retired and no longer working for pay, homemaking or unemployed at Visit 2.
Not married including participants who identified as being divorced, separated, widowed, or never married at Visit 2.
Depressive symptoms scored using Maastricht Vital Exhaustion Questionnaire.
3.2. Mid‐life social relationships as potentially protective for dementia risk
Across models, having strong (HR: 0.77, 95% CI: 0.69–0.86) or average social relationships (HR: 0.85, 95% CI: 0.75–0.96) in mid‐life was associated with a lower risk for developing dementia, compared to participants with poor social relationships in mid‐life (Table 2). Kaplan–Meier estimates of late‐life dementia risk stratified by mid‐life social relationships are shown (Figure 3).
TABLE 2.
Adjusted hazard ratios for association of mid‐life social relationships with dementia risk (N = 13,070).
| Social relationships |
Model 1 a HR (95% CI) |
Model 2 b HR (95% CI) |
Model 3 c HR (95% CI) |
|---|---|---|---|
| Strong (n = 7458) | 0.71(0.64–0.79) ** | 0.77(0.69–0.86) ** | 0.76(0.67–0.85) ** |
| Average (n = 3674) | 0.81 (0.72–0.91) ** | 0.85 (0.75–0.96) ** | 0.84 (0.74–0.95) ** |
| Poor (n = 1911) | Reference | Reference | Reference |
| Age (per 1 year) | 1.16 (1.15–1.17) ** | 1.15 (1.14–1.16) ** | 1.15 (1.14–1.16) ** |
| Race‐ARIC Center | |||
| White – Minnesota | 1.07 (0.95–1.20) | 1.11 (0.99–1.25) | 1.09 (0.97–1.23) |
| White – Maryland | 1.10 (0.99–1.24) | 1.12 (1.00–1.25) | 1.08 (0.97–1.22) |
| Black – Forsyth | 1.29 (0.98–1.70) | 1.24 (0.94–1.63) | 1.10 (0.83–1.45) |
| Black – Jackson | 1.65 (1.46–1.86) ** | 1.61 (1.43–1.81) ** | 1.41 (1.24–1.60) ** |
| White – Forsyth | Reference | Reference | Reference |
| Sex (Men) | 1.06 (0.98–1.14) | 1.16 (1.06–1.26) ** | 1.14 (1.04–1.25) ** |
| APOE ɛ4 carrier | 1.88 (1.74–2.04) ** | 1.89 (1.75–2.05) ** | 1.90 (1.76–2.06) ** |
| Education level | |||
| Less than high school | Reference | Reference | Reference |
| High school or equivalent | 0.76 (0.68–0.85) ** | 0.78 (0.70–0.87) ** | 0.80 (0.72–0.89) ** |
| More than high school | 0.64 (0.58–0.71) ** | 0.67 (0.61–0.74) ** | 0.70 (0.63–0.77) ** |
| Not working outside the home | – | 1.15 (1.06–1.26) ** | 1.13 (1.04–1.23) ** |
| Not married | – | 1.17 (1.06–1.29) ** | 1.15 (1.04–1.27) ** |
| Depressive symptoms | – | 1.21 (1.11– 1.32) ** | 1.18 (1.08–1.28) ** |
| Hypertension | – | – | 1.16 (1.07–1.26) ** |
| Diabetes | – | – | 1.53 (1.37–1.71) ** |
| Body mass index (≥30 kg/m2) | – | – | 1.12 (1.03–1.23) * |
| Total cholesterol (≥200 mg/dL) | – | – | 1.03 (0.95–1.12) |
| Ever smoker | – | – | 1.12 (1.03–1.22) * |
| Ever drinker | – | – | 0.92 (0.83–1.02) |
Note: SI conversion: To convert cholesterol to millimoles per liter, multiply by 0.0259.
Abbreviations: APOE, apolipoprotein E; HR, hazard ratio.
Model adjusted for age, race‐center, sex, APOE ɛ4, education.
Model adjusted for Model 1 covariates, occupational/marital status, and (having) depressive symptoms as measured at Visit 2.
Model adjusted for Model 2 covariates and vascular risk factors as measured at Visit 2.
P ≤ 0.05
P ≤ 0.01.
FIGURE 3.
Kaplan–Meier curves of dementia‐free survival by mid‐life social relationships. Follow‐up time began 5 years following Atherosclerosis Risk in Communities (ARIC) Visit 2 when psychosocial factors were measured. Rates of dementia incidence differed significantly (log‐rank p < 0.001) in participants with varying levels of social relationships.
Our results did not show statistically significant evidence of effect modification by sex (p‐interaction = 0.14) or race (p‐interaction = 0.69). Stratified sex and race models are shown (Tables S1 and S2). Exploratory analyses that examined whether age (dichotomized at the sample median at Visit 2), APOE ɛ4 status, presence versus absence of depressive symptoms, or educational level modified the association between mid‐life social relationships and dementia were not statistically significant. Due to our prespecified analytic plan to evaluate effect modication by such factors, we further examined model findings from stratified groups to observe potential between‐group differences. Most notably, stronger associations between mid‐life social relationships and dementia risk were shown in participants ≥57 years old (the sample median at Visit 2; Tables S3–S6).
3.3. Sensitivity analyses
Models assessing independent associations between mid‐life measures of social support and social isolation (as assessed via the ISEL‐SF and LSNS, respectively) with incident dementia risk are shown (Tables S7 and S8). In models assessing social support, participants with high and intermediate mid‐life social support had a lower late‐life risk of dementia, relative to those with low mid‐life social support. In models assessing social isolation across four levels, participants with low or moderate mid‐life risk of social isolation had a lower late‐life risk of dementia relative to those who were isolated in mid‐life, yet the association was not significant in those at high risk for isolation. Models excluding participants whose global cognitive function fell in the lowest fifth percentile at Visit 2 or those who had a stroke before the end of follow‐up showed results very similar to the primary models (Tables S9 and S10). Analyses, which accounted for death as a competing risk, showed results similar to the primary models, although the potentially protective effect conferred from mid‐life social relationships was attenuated (Table S11).
4. DISCUSSION
Results from this community‐based cohort study provide evidence that stronger social relationships in mid‐life are associated with a lower risk of incident dementia in late‐life. Findings remained robust to adjustments for demographics, APOE ɛ4, depressive symptoms, occupational status, marital status, and vascular risk factors measured in mid‐life. Secondary and exploratory analyses illustrated no significant effect modification by sex, race, or other examined factors (age, APOE ɛ4 status,, depressive symptoms, or educational level).
Our findings build upon previous studies that have shown the potentially protective effect of social relationships on outcomes, such as sustained cognitive function and/or reduced dementia risk 4 , 5 , 7 , 10 , 11 , 13 , 21 This study significantly adds to existing literature in several ways. First, we examined social relationships in middle‐aged adults with carefully adjudicated dementia diagnoses over a long follow‐up period (median follow‐up: 19.3 years). Second, the design of our longitudinal study, which purposefully excluded any adults with dementia at baseline, and imposed an additional lag time of 5 years following the measurement of psychosocial health, helps account for reverse causation. Next, our examination of the association between mid‐life social relationships with late‐life dementia as the outcome of interest, in contrast to studies examining the rate of cognitive decline, is an important distinction. Incident dementia, as the outcome, highlights the development of a health condition rather than an interindividual rate of cognitive decline, which is highly dependent upon baseline cognition. Finally, we want to highlight that the cumulative incidence of dementia in this sample (20%) is aligned with global estimates of dementia for this age group. 41 This is worth noting as it provides compelling evidence that our sample is not healthier or at lower risk for dementia than the general population, a common shortcoming in cohort studies.
Investigating social relationships in mid‐life relative to dementia in later life is particularly important to intervention and prevention efforts. 21 , 30 Mid‐life represents a significant window when lifestyle choices, including social behaviors, can be altered and prevention is still possible. In accordance with theories of cognitive reserve, the “buffering hypothesis” and related conceptual frameworks, engaging in social, physical, and intellectually stimulating activities in mid‐life may help instill habits of a healthy lifestyle that can be translated into later life, and furthermore, help sustain higher cognitive capabilities in elderly adults. 1 , 42 , 43 Likewise, being socially active in mid‐life can confer simultaneous health benefits (e.g., adults with greater social contact are more likely to have rides to doctor appointments, prescription pick‐ups, company to go for a walk, etc). 44 , 45 , 46
Effect modification by neither sex nor race was statistically significant in this sample. However, results indicate that mid‐life social relationships may be more beneficial in women and White participants relative to men and Black participants, respectively. Previous studies looking at varying aspects of psychosocial health have suggested that women may derive greater benefits of strong psychosocial health, 15 , 25 but this continues to be studied. Taken together, it is possible that these findings reflect that relationships and emotional connections are internalized differently in specified sex and racial groups, although this interpretation is speculative. Continuing to explore whether strong social relationships confer a greater protective effect in women and Black participants could be particularly significant, as both these groups are considered at an increased risk of dementia. 47 , 48 Likewise, related studies investigating sex and race differences are needed to support both intervention efforts and diversity, equity, and inclusion (DEI) research. These efforts are critical as researchers strive to better understand individualized risk profiles for dementia. 49
Relatedly, the ARIC cohort encompasses a unique group of individuals in whom to study mid‐life social relationships in association with dementia risk. The median age of ARIC study participants when psychosocial health was assessed was 57 years old. This evaluation occurred in the early 1990s—a period that marked the introduction of home internet services, which may have influenced social relationships and perceptions of human connectedness. Many ARIC participants were born in the years around the time of the Great Depression and entered adulthood during pivotal events, such as World War II as well as the Korean and Vietnam wars. Racial desegregation further characterized this era and may have largely impacted Black ARIC participants, who were predominantly recruited from two field sites in the South (Jackson, MS and Winston‐Salem, NC). Altogether, this culmination of historical events in addition to societal changes that accompanied the subsequent technological revolution may have shaped psychosocial factors and the worldview of the cohort examined. Thus, our findings of a potentially protective effect of mid‐life social relationships in individuals who collectively lived through these events in both urban and rural communities further speaks to the powerful influence this risk factor may have on incident dementia.
Exploratory analyses in this cohort did not show significant interaction effects when examining factors such as age, APOE ɛ4, depressive symptoms, and education. Although we subsequently conducted stratified analyses, we want to caution against overinterpreting such findings. Our decision to analyze factors related to multiple activity domains (e.g., mental, cognitive, and psychosocial health) despite no evidence of effect modification stems from the scarcity of literature examining how theco‐existing nature and interactions between such characteristics may shape long‐term health outcomes. 22
Sensitivity analyses illustrated that the potentially protective effect observed when measuring social relationships in mid‐life was consistent with that observed when examining social support and the risk of isolation. This provides additional evidence as to the validity of our social relationships measure, previously used in other studies of the ARIC cohort. 29 , 50 Notably, associations between social support or risk of social isolation with global cognitive decline in late‐life were not seen in a previous ARIC study, which measured cognition through use of an abbreviated cognitive battery. 19 Use of a cognitive battery may not reflect ongoing disease development as well as evaluating incident dementia, which further incorporates functional status and informant interviews.
A complex behavior like psychosocial health may be difficult to quantify through means of questionnaires. The rationale that led us to examine social relationships as a composite measure was based upon the idea that social support and social isolation are independent constructs with underlying similarities. 31 , 51 The ISEL‐SF is thought to measure perceived social support, which has been shown to vary little over time, and focuses on the support that participants feel they are receiving with little emphasis on reciprocated social support. 31 The LSNS is largely objective and captures many aspects of one's social network including social support, relationship closeness with friends and family, and aspects of loneliness, all of which may vary situationally. 31 Both scales have a collective emphasis on relationships closeness with friends and family, with a lesser focus on how individuals see themselves and their role within their greater community. These measures capture many facets of psychosocial health but are not absolute. Beyond our efforts to include adjacent lifestyle factors, such as APOE e4 status, depressive symptoms, and educational level, we are hopeful that future work can additionally evaluate robust measures of psychosocial health taken at multiple time points throughout the life course in addition to other social determinants of health.
There are limitations to the present study. Questionnaires in this study were administered solely at Visit 2, which precluded us from measuring changes in psychosocial health as individuals aged from their 50s through 70s. Notably, previous studies have shown social relationships, and scores on these self‐reported scales do not largely fluctuate during this period of life. 31 , 52 We further acknowledge the possibility that some individuals may have experienced an earlier preclinical cognitive decline that influenced the psychosocial factors measured at Visit 2. To avoid this form of reverse causation, we added the 5‐year lag period following Visit 2 to eliminate any individuals who were already very close to having clinical symptoms of dementia at the time psychosocial factors were assessed. We further included a sensitivity analysis, which excluded the bottom 5% of individuals displaying low global cognition at Visit 2 in hopes to avoid the inclusion of individuals who were already in early stages of dementia. Despite these adjustments, survival bias still may have influenced our findings, as those with the poorest social relationships and cognitive function are less likely to be able to come in for study visits. However, having dementia events further identified by phone screening, hospitalization, and death ascertained via International Classification of Diseases (ICD) codes helps address this concern. Finally, we want to reiterate that although these findings were shown in a substantial sample of Black participants, they are not generalizable beyond Black and White participants recruited from within ARIC communities.
In conclusion, the present study illustrates that social relationships in mid‐life may have a potentially protective effect on dementia risk in older adults. In addition to research efforts that aim to reduce dementia risk proactively, emphasis on protective risk factors in mid‐life may serve as an encouraging, non‐pharmacological alternative for prevention efforts. Further evaluation of such factors in parallel to disease processes and disease‐modifying treatments is necessary to study aging and multifactorial diseases, such as dementia.
CONFLICT OF INTEREST STATEMENT
The authors declare no conflicts of interest. Author disclosures are available in the Supporting Information.
ETHICS STATEMENT AND CONSENT
The study was performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. All the institutional review boards of the respective institutions approved the Atherosclerosis Risk in Communities (ARIC) study, and participants signed informed consent for their involvement. Diversity, equity, and inclusion (DEI) was addressed in the study design, execution, and interpretation. Participants came from multiple study centers across the United States characterized by varying economic and racial profiles. When analyzing and interpreting the results, the study accounted for differences in age, sex, and race.
Supporting information
Supporting File 1
Supporting File 2
ACKNOWLEDGMENTS
The authors thank the staff and participants of the Atherosclerosis Risk in Communities (ARIC) study for their important contributions. The ARIC study is carried out as a collaborative study supported by National Heart, Lung, and Blood Institute (NHLBI) contracts (75N92022D00001, 75N92022D00002, 75N92022D00003, 75N92022D00004, and 75N92022D00005). The ARIC Neurocognitive Study is supported by U01HL096812, U01HL096814, U01HL096899, U01HL096902, and U01HL096917 from the National Institutes of Health (NIH; NHLBI, National Institute of Neurological Disorders and Stroke [NINDS], National Institute on Aging [NIA], and National Institute on Deafness and Other Communication Disorders [NIDCD]). R.C.G. and R.F.G. were supported by NINDS Intramural Research Program. P.L.L. is supported by K24 HL159246. P.L.L. reports grants from the NIH to her institution. P.P. is supported by NIH/NHLBI 75N92022D00005. K.W. is supported by the NIA Intramural Research Program (AG000348‐01 and AG000349‐01).
Groechel RC, Liu AC, Lutsey PL, et al. Associations between mid‐life social relationships and the risk of incident dementia: The ARIC study. Alzheimer's Dement. 2025;21:e70365. 10.1002/alz.70365
This work was prepared while Renée C. Groechel was at National Institute of Neurological Disorders and Stroke (NINDS). Renée C. Groechel is now employed at the Injury Outcomes Network (ION) Research. Dr. Groechel's present address and email address are stated above.
REFERENCES
- 1. Fratiglioni L, Marseglia A, Dekhtyar S. Ageing without dementia: can stimulating psychosocial and lifestyle experiences make a difference? Lancet Neurol. 2020;19:533‐543. [DOI] [PubMed] [Google Scholar]
- 2. Livingston G, Huntley J, Sommerlad A, et al. Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. Lancet. 2020;396:413‐446. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Duffner LA, DeJong N, Jansen JF, et al. Associations between social health factors, cognitive activity and neurostructural markers for brain health – a systematic literature review and meta‐analysis. Ageing Res Rev. 2023;89:101986. [DOI] [PubMed] [Google Scholar]
- 4. Penninkilampi R, Casey AN, Singh MF, Brodaty H. The association between social engagement, loneliness, and risk of dementia: a systematic review and meta‐analysis. J Alzheimers Dis. 2018;66(4):1619‐1633. [DOI] [PubMed] [Google Scholar]
- 5. Chan D, Shafto M, Kievit R, et al. Lifestyle activities in mid‐life contribute cognitive reserve in late‐life, independent of education, occupation, and late‐life activities. Neurobiol Aging. 2018;70:180‐183. doi: 10.1016/j.neurobiolaging.2018.06.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Garo‐Pascual M, Gaser C, Zhang L, Tohka J, Medina M, Strange BA. Brain structure and phenotypic profile of superagers compared with age‐matched older adults: a longitudinal analysis from the Vallecas Project. Lancet Healthy Longev. 2023;4(8):e374‐e385. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Heneghan A, Deng F, Wells K, et al. Modifiable lifestyle activities affect cognition in cognitively healthy middle‐aged individuals at risk for late‐life Alzheimer's Disease. J Alzheimers Dis. 2023;91(2):833‐846. [DOI] [PubMed] [Google Scholar]
- 8. Akhter‐Khan SC, Tao Q, Ang TFA, et al. Associations of loneliness with risk of Alzheimer's disease dementia in the Framingham Heart Study. Alzheimers Dement. 2021;17(10):1619‐1627. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Rawtaer I, Gao Q, Nyunt MSZ, et al. Psychosocial risk and protective factors and incident mild cognitive impairment and dementia in community dwelling elderly: findings from the Singapore Longitudinal Ageing Study. J Alzheimers Dis. 2017;57(2):603‐611. [DOI] [PubMed] [Google Scholar]
- 10. Saczynski JS, Pfeifer LA, Masaki K, et al. The effect of social engagement on incident dementia: the Honolulu‐Asia Aging Study. Am J Epidemiol. 2006;163(5):433‐440. [DOI] [PubMed] [Google Scholar]
- 11. Shen C, Rolls ET, Cheng W, et al. Associations of social isolation and loneliness with later dementia. Neurology. 2022;99(2):e164‐e175. [DOI] [PubMed] [Google Scholar]
- 12. Sommerlad A, Sabia S, Singh‐Manoux A, Lewis G, Livingston G. Association of social contact with dementia and cognition: 28‐year follow‐up of the Whitehall II cohort study. PLoSMed. 2019;16(8):e1002862. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Deng JH, Huang K‐Y, Hu X‐X, et al. Midlife long‐hour working and later‐life social engagement are associated with reduced risks of mild cognitive impairment among community‐living Singapore elderly. J Alzheimers Dis. 2019;67(3):067‐1077. [DOI] [PubMed] [Google Scholar]
- 14. Rai P, Sundarakumar JS. Association between social networks and cognition among middle‐aged and older adults in rural India. Neuropsychology. 2024;38(8):740. [DOI] [PubMed] [Google Scholar]
- 15. Huang Y, Zhu X, Liu X, Li J. The effects of loneliness, social isolation, and associated gender differences on the risk of developing cognitive impairment for Chinese oldest old. Aging & Mental Health. 2023;27(7):1360‐1367. [DOI] [PubMed] [Google Scholar]
- 16. Hughes TF, Flatt JD, Fu B, Chang CCH, Ganguli M. Engagement in social activities and progression from mild to severe cognitive impairment: the MYHAT study. Int Psychogeriatr. 2013;25(4):587‐595. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Wilson RS, Krueger KR, Arnold SE, et al. Loneliness and risk of Alzheimer disease. Arch Gen Psychiatry. 2007;64:234‐240. [DOI] [PubMed] [Google Scholar]
- 18. Yang JS, Jeon YJ, Lee GB, Kim HC, Jung SJ. The association between psychological resilience and cognitive function in longitudinal data: results from the community follow‐up survey. J Affect Disord. 2021;290:109‐116. [DOI] [PubMed] [Google Scholar]
- 19. Kats D, Patel MD, Palta P, et al. Social support and cognition in a community‐based cohort: the Atherosclerosis Risk in Communities (ARIC) study. Age Ageing. 2016;45(4):475‐480. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Oh SS, Cho E, Kang B. Social engagement and cognitive function among middle‐aged and older adults: gender‐specific findings from the Korean longitudinal study of aging (2008‐2018). Sci Rep. 2021;11(1):15876. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Sommerlad A, Kivimäki M, Larson EB, et al. Social participation and risk of developing dementia. Nature Aging. 2023;3(5):532‐545. [DOI] [PubMed] [Google Scholar]
- 22. Bielak AA, Gow AJ. A decade later on how to “use it” so we don't “lose it”: an update on the unanswered questions about the influence of activity participation on cognitive performance in older age. Gerontol. 2023;69(3):336‐355. [DOI] [PubMed] [Google Scholar]
- 23. Wright JD, Folsom AR, Coresh J, et al. The ARIC (Atherosclerosis Risk in Communities) Study: JACC Focus Seminar 3/8. J Am Coll Cardiol. 2021;77(23):2939‐2959. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Cohen S, Hoberman HM. Positive events and social supports as buffers of life change stress. J Appl Soc Psychol. 1983;13(2):99‐125. [Google Scholar]
- 25. Payne TJ, Andrew M, Butler KR, Wyatt SB, Dubbert PM, Mosley TH Jr. Psychometric evaluation of the interpersonal support evaluation list – short form in the ARIC Study Cohort. SAGE Open. 2012;2(3):2158244012461923. [Google Scholar]
- 26. Groechel RC, Liu AC, Koton S, et al. Association between mid‐life social factors and estimated late‐life amyloid burden: the Atherosclerosis Risk in Communities (ARIC)‐PET Study. J Alzheimers Dis. 2024;97(4):1901‐1911. doi: 10.3233/JAD-231218 [DOI] [PubMed] [Google Scholar]
- 27. Lubben JE. Assessing social networks among elderly populations. Fam Community Health. 1988;11(3):42‐52. [Google Scholar]
- 28. Lubben J, Blozik E, Gillmann G, et al. Performance of an abbreviated version of the Lubben Social Network Scale among three European community‐dwelling older adult populations. Gerontologist. 2006;46(4):503‐513. [DOI] [PubMed] [Google Scholar]
- 29. Groechel RC, Liu AC, Liu C, et al. Social engagement, amyloid burden, and dementia: the ARIC‐PET study. Alzheimers Dement (Amst). 2024;16(2):e12560. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Holt‐Lunstad J. Why social relationships are important for physical health: a systems approach to understanding and modifying risk and protection. Annu Rev Psychol. 2018;69:437‐458. [DOI] [PubMed] [Google Scholar]
- 31. Sakr‐Ashour F, Concept. Clarification and Assessment of Social Isolation and Social Support in Older Adults. Administration for Community Living Office of Performance and Evaluation. New Editions Consulting, Inc; 2021. [Google Scholar]
- 32. Gottesman RF, Albert MS, Alonso A, et al. Associations between midlife vascular risk factors and 25‐year incident dementia in the Atherosclerosis Risk in Communities (ARIC) cohort. JAMA Neurol. 2017;74(10):1246‐1254. doi: 10.1001/jamaneurol.2017.1658 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Gottesman RF, Wu A, Coresh J, et al. Associations of vascular risk and amyloid burden with subsequent dementia. Ann Neurol. 2022;92(4):607‐619. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Knopman DS, Gottesman RF, Sharrett AR, et al. Mild cognitive impairment and dementia prevalence: the Atherosclerosis Risk in Communities neurocognitive study. Alzheimer Dement (Amst). 2016;2:1‐11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Carpenter CR, DesPain B, Keeling TN, Shah M, Rothenberger M. The Six‐Item Screener and AD8 for the detection of cognitive impairment in geriatric emergency department patients. Ann Emerg Med. 2011;57(6):653‐661. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Galvin JE, Roe CM, Powlishta KK, et al. The AD8: a brief informant interview to detect dementia. Neurology. 2005;65(4):559‐564. [DOI] [PubMed] [Google Scholar]
- 37. Appels A, Hoppener P, Mulder PA. A questionnaire to assess premonitory symptoms of myocardial infarction. Int J Cardiol. 1987;17(1):15‐24. [DOI] [PubMed] [Google Scholar]
- 38. Honda Y, Mok Y, Mathew L, et al. Psychosocial factors and subsequent risk of hospitalizations with peripheral artery disease: the Atherosclerosis Risk in Communities (ARIC) Study. Atherosclerosis. 2021;329:36‐43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Williams JE, Mosley TH Jr, Kop WJ, Couper DJ, Welch VL, Rosamond WD. Vital exhaustion as a risk factor for adverse cardiac events (from the Atherosclerosis Risk in Communities [ARIC] study). Am J Cardiol. 2010;105(12):1661‐1665. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. JASA. 1999;94:496‐509. [Google Scholar]
- 41. Nichols E, Steinmetz JD, Vollset SE, et al. Estimation of the global prevalence of dementia in 2019 and forecasted prevalence in 2050: an analysis for the Global Burden of Disease Study 2019. Lancet Public Health. 2022;7(2):e105‐e125. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Sachdev SP. Social health, social reserve and dementia. Curr Opin Psychiatry. 2022;35(2):111‐117. [DOI] [PubMed] [Google Scholar]
- 43. Stern Y. Cognitive reserve in ageing and Alzheimer's disease. Lancet Neurol. 2012;6(11):1006‐1012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Kang HW, Park M, Wallace JP. The impact of perceived social support, loneliness, and physical activity on quality of life in South Korean older adults. J Sport Health Sci. 2018;7(2):237‐244. doi: 10.1016/j.jshs.2016.05.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Sharifian N, Sol K, Zahodne LB, Antonucci TC. Social relationships and adaptation in later life. Comprehensive Clinical Psychology. 2022:52‐72. doi: 10.1016/B978-0-12-818697-8.00016-9 [DOI] [Google Scholar]
- 46. Zahodne LB. Psychosocial protective factors in cognitive aging: a targeted review. Arch Clin Neuropsychol. 2021;36(7):1266‐1273. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Beam CR, Kaneshiro C, Jang JY, Reynolds CA, Pedersen NL, Gatz M. Differences between women and men in incidence rates of dementia and Alzheimer's disease. J Alzheimers Dis. 2018;64(4):1077‐1083. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48. Mayeda ER, Glymour MM, Quesenberry CP, Whitmer RA. Inequalities in dementia incidence between six racial and ethnic groups over 14 years. Alzheimers Dement. 2016;12(3):216‐224. doi: 10.1016/j.jalz.2015.12.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49. Manly JJ, Gilmore‐Bykovskyi A, Deters KD. Inclusion of Underrepresented Groups in Preclinical Alzheimer Disease Trials – Opportunities Abound. JAMA Netw Open. 2021;4(7):e2114606. doi: 10.1001/jamanetworkopen.2021.14606 [DOI] [PubMed] [Google Scholar]
- 50. Eswaran S, Knopman DS, Koton S, et al. Psychosocial health and the association between cerebral small vessel disease markers with dementia: the ARIC study. Stroke. 2024;55(10):2449‐2458. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51. Hu J, Fitzgerald SM, Owen AJ, et al. Social isolation, social support, loneliness and cardiovascular disease risk factors: a cross‐sectional study among older adults. Int J Geriatr Psychiatry. 2021;36(11):1795‐1809. doi: 10.1002/gps.5601 [DOI] [PubMed] [Google Scholar]
- 52. Ertel KA, Glymour MM, Berkman LF. Social networks and health: a life course perspective integrating observational and experimental evidence. J Soc Pers Relat. 2009;26:73‐92. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supporting File 1
Supporting File 2