Association of Healthy Lifestyles With Risk of Alzheimer Disease and Related Dementias in Low-Income Black and White Americans

Abstract

Background and Objectives

Although the importance of healthy lifestyles for preventing Alzheimer disease and related dementias (ADRD) has been recognized, epidemiologic evidence remains limited for non-White or low-income individuals who bear disproportionate burdens of ADRD. This population-based cohort study aims to investigate associations of lifestyle factors, individually and together, with the risk of ADRD among socioeconomically disadvantaged Americans.

Methods

In the Southern Community Cohort Study, comprising two-thirds self-reported Black and primarily low-income Americans, we identified incident ADRD using claims data among participants enrolled in Medicare for at least 12 consecutive months after age 65 years. Five lifestyle factors—tobacco smoking, alcohol consumption, leisure-time physical activity (LTPA), sleep hours, and diet quality—were each scored 0 (unhealthy), 1 (intermediate), or 2 (healthy) based on the health guidelines. A composite lifestyle score was created by summing all scores. Cox regression was used to estimate hazard ratios (HRs, 95% CIs) for incident ADRD, treating death as a competing risk.

Results

We identified 1,694 patients with newly diagnosed ADRD among 17,209 participants during a median follow-up of 4.0 years in claims data; the mean age at ADRD diagnosis was 74.0 years. Healthy lifestyles were individually associated with an 11%–25% reduced risk of ADRD: multivariable-adjusted HR (95% CI) was 0.87 (0.76–0.99) for never vs current smoking, 0.81 (0.72–0.92) for low-to-moderate vs no alcohol consumption, 0.89 (0.77–1.03) for ≥150 minutes of moderate or ≥75 minutes of vigorous LTPA each week vs none, 0.75 (0.64–0.87) for 7–9 hours vs >9 hours of sleep, and 0.85 (0.75–0.96) for the highest vs lowest tertiles of the Healthy Eating Index. The composite lifestyle score showed a dose-response association with up to 36% reduced risk of ADRD: multivariable-adjusted HRs (95% CIs) across quartiles were 1 (ref), 0.88 (0.77–0.99), 0.79 (0.70–0.90), and 0.64 (0.55–0.74); p trend <0.001. The beneficial associations were observed regardless of participants' sociodemographics (e.g., race, education, and income) and health conditions (e.g., history of cardiometabolic diseases and depression).

Discussion

Our findings support significant benefits of healthy lifestyles for ADRD prevention among socioeconomically disadvantaged Americans, suggesting that promoting healthy lifestyles and reducing barriers to lifestyle changes are crucial to tackling the growing burden and disparities posed by ADRD.

In 2021, ∼6.2 million older Americans live with Alzheimer disease and related dementias (ADRD), and this number is projected to double by 2050.¹ Notably, ADRD disproportionately affect Black Americans^1,2 and individuals with low socioeconomic status (SES)^3,4 who are subjected to a wide array of systemic and structural disadvantages and social inequalities.⁵ For example, Black people, particularly those with limited formal education, may face challenges such as unemployment or underemployment, racism and discrimination, and lack of access to safe neighborhoods, nutritious foods, and needed medical care, which in turn may increase their risk of developing ADRD. Yet, despite racial and socioeconomic disparities in ADRD, Black and low-SES populations have been underrepresented in epidemiologic research; thus, identifying modifiable factors for ADRD prevention among these disadvantaged populations remains a critical public health issue.

The 2020 Lancet Commission on Dementia Prevention, Intervention, and Care has reported 12 modifiable risk factors for ADRD prevention and highlighted the importance of healthy lifestyles.⁶ Indeed, the protective role of healthy lifestyles, including nonsmoking,^7-10 regular exercise,^7,8,11,12 limited alcohol use,^13,14 adequate sleep,^7,15,16 and high-quality diets,^8,17,18 in ADRD has been widely discussed. Recently, some cohort studies have reported synergistic effects of lifestyle factors,^19-22 showing promising associations of combined scores with >30% reduced risk of ADRD, even among individuals genetically predisposed to the disease. However, previous studies mainly focused on White general populations, thus not fully addressing race-specific and SES-specific features in the lifestyle and ADRD associations. A better understanding of this issue among non-White Americans and low-SES individuals is crucial to tackling the growing burden and disparities posed by ADRD.

In this study, we evaluated the associations of 5 major lifestyle factors, that is, tobacco smoking, alcohol drinking, leisure-time physical activity (LTPA), sleep duration, and diet quality, and composite scores of those factors with incident ADRD among predominantly low-income Black and White Americans. We also examined whether the associations were modified by other ADRD risk factors such as participants' age, sociodemographics, and preexisting health conditions (i.e., cardiometabolic diseases and depression).

Methods

Study Population

This analysis is based on the Southern Community Cohort Study (SCCS), which aims to investigate racial, socioeconomic, and other health disparities in the United States.²³ Between March 2002 and September 2009, >84,000 men and women, aged 40–79 years, were recruited across 12 southeastern states in the United States. Most participants (∼86%) were enrolled at community health centers (CHCs) that served medically uninsured or underinsured individuals. Two-thirds of the SCCS participants were self-reported Black individuals; over half had annual household income <$15,000. After obtaining written informed consent, baseline interviews were conducted to collect data on sociodemographics, lifestyles, medical history, and anthropometrics. Dietary intakes were assessed using a validated 89-item food frequency questionnaire.^24,25

Assessment of Lifestyle Factors

Five lifestyle factors assessed at baseline—tobacco smoking, alcohol drinking, LTPA, sleep duration, and overall diet quality—were scored as 2 (healthy), 1 (intermediate), or 0 (unhealthy). Specifically, smoking status was defined as nonsmokers who never smoked regularly, former smokers who had smoked >100 cigarettes in their lifetime but ceased, and current smokers who continued to smoke. Alcohol consumption was classified according to the Dietary Guidelines for Americans²⁶ into low-to-moderate (>0 to ≤2 and >0 to ≤1 drink per day for men and women, respectively; 1 drink = 14 g of ethanol), heavy (>2 and >1 drink per day, respectively), and nondrinkers. Physical activity was defined by the 2018 Physical Activity Guidelines for Americans,²⁷ recommending ≥150 minutes moderate or ≥75 minutes vigorous LTPA per week, corresponding to ≥8.3 metabolic equivalent hours (MET-h/wk). Sleep duration, as the weighted average of sleep hours during weekdays and weekends, was divided into 7–9, >9, and <7 hours, based on the National Sleep Foundation's recommendations for adults.²⁸ Healthy diet was assessed by 2010 Healthy Eating Index (HEI), representing an overall diet quality and adherence to the Dietary Guidelines for Americans,^29,30 and categorized into tertiles.

A composite lifestyle score was created by summing the score of all 5 factors (ranging from 0 to 10) and categorized by its quartile distribution (score 0–3, 4, 5–6, and 7–10). In addition, a weighted score was estimated using beta coefficients of each lifestyle factor in the Cox regression model after mutual adjustment for all 5 factors and all potential confounders (see more details in Covariates): Each lifestyle score (0, 1, and 2) was weighted by multiplying by its beta coefficient, summed, and categorized into quartiles.

Clinical Diagnosis of ADRD

Clinical diagnosis of ADRD was based on Medicare claims data, which has 85% sensitivity and 89% specificity in identifying ADRD cases compared with gold-standard in-person cognitive assessments.³¹ Full descriptions of ADRD ascertainment in the SCCS have been reported elsewhere.³² In brief, using the International Classification of Diseases 9/10 codes defined by Centers for Medicare and Medicaid Chronic Conditions Warehouse,^33,34 we identified when the first presence of ADRD diagnosis occurred in claims data, including inpatient, outpatient, or carrier claims, for years 1999–2016. To capture incident cases only, we restricted our analysis to participants who enrolled in traditional Medicare for ≥12 consecutive months without an ADRD claim after age 65 years. Time-to-event analysis was started after 12 continuous months of Medicare enrollment and censored on the date of death, loss to follow-up, termination of traditional Medicare coverage, or the end of claims data linkage (December 2016), whichever came first.

Covariates

Potential confounders were selected a priori and added to the model sequentially. The basic model was adjusted for enrollment age (years), calendar years of starting follow-up in claims data (<2003, 2003–2007, 2007–2011, 2011–2016), and the time interval between enrollment and starting follow-up (years); the latter 2 were selected considering possible changes in health behaviors and providers' coding practices over time. The second model further included sex (men and women), race (self-identified Black, White, and Other), educational attainment (<high school, high school graduation, some college, and ≥university degree), annual household income (<$15,000, ≥$15,000 to <$25,000, ≥$25,000 to <$50,000, and ≥$50,000), marital status (currently, previously, and never married), and enrollment source (CHCs and general population). In the final model, given the potential influence of certain prevalent diseases on both lifestyles and ADRD risk, we further included the history of cardiometabolic diseases (i.e., ischemic heart disease, diabetes, hypertension, and dyslipidemia; yes/no) and depression (yes/no), based on self-reported doctor diagnosis or use of prescription medicine to treat those conditions at baseline survey. Proportions of missing covariates were mostly <1%, and thus, missing was assigned by the median (continuous) or mode (categorical) values of nonmissing covariates.

Statistical Analysis

Among the total SCCS participants, we first selected beneficiaries aged 65 years or older between study enrollment and 2016 (n = 31,484). To minimize possible reverse causation and enhance the accuracy of claims-based diagnosis, we excluded individuals who did not continuously enroll in traditional Medicare for ≥12 months (n = 10,716), who had an ADRD diagnosis before age 65 years (n = 275) or within their first 12 months of Medicare enrollment (n = 642). Additional exclusions were made (n = 2,642) if participants had a self-reported history of Parkinson disease, multiple sclerosis, stroke, or brain cancer at baseline or missing on all lifestyle variables (i.e., smoking, alcohol drinking, LTPA, sleep hours, and HEI). Finally, 17,209 participants, including 9,641 Black Americans and 6,755 White Americans, remained for the current analysis (see study flowchart in eFigure 1, links.lww.com/WNL/C89).

Baseline characteristics of incident cases of ADRD and noncases were compared using the χ² test or t test. Cox proportional hazard regression models, treating death as a competing risk, were used to estimate hazard ratios (HRs) and 95% CIs for incident ADRD. The lowest score group (unhealthy) was modeled as the reference. For the individual analysis of each lifestyle factor, participants with respective missing data were excluded (n = 242 for smoking, 476 for alcohol drinking, 631 for LTPA, 228 for sleep hours, and 872 for HEI). Meanwhile, for the composite score analysis, missing 1–4 lifestyle factors were imputed using regression models based on the nonmissing lifestyle data and all covariates. Age at starting follow-up in claims data and age at censoring was used as the time-scale. The global goodness-of-fit test by Schoenfeld residuals confirmed no violation against the proportional hazard assumption. Linear trends were tested using median values of each quartile of the composite lifestyle scores. Stratified analyses were conducted to evaluate whether the associations varied by individuals' characteristics (e.g., age, sex, race, education, and income), health conditions (e.g., history of cardiometabolic diseases or depression), or follow-up time. Interactions between lifestyle factors and stratification variables were tested using a multiplicative interaction term; p interaction was corrected by a false discovery rate for multiple comparisons. A series of sensitivity analyses were conducted by excluding participants who reported a history of cancer or heart disease at baseline, who switched from traditional Medicare to a Medicare Advantage managed care plan during the follow-up, or ADRD cases diagnosed within the first 2 years of starting follow-up in claims data, by requiring at least 2 claims on ADRD diagnosis on 2 different dates, and by restricting to participants without any missing data on lifestyles. Statistical analyses were performed using SAS 9.4 (SAS Institute, Inc., Cary, NC). Two-sided p < 0.05 was considered statistically significant.

Standard Protocol Approvals, Registrations, and Patient Consents

All participants provided written informed consent before entering the SCCS. The study protocol of SCCS was approved by the Institutional Review Boards of Vanderbilt University Medical Center and Meharry Medical College. This study was approved by the SCCS Data and Biospecimen Use Committee and used the deidentified data only.

Data Availability

Data could be shared in a deidentified fashion after acquiring permission from the SCCS Data and Biospecimen Use Committee.

Results

During a median follow-up of 4.0 years in claims data (interquartile range 2.0–7.0 years) among 17,209 participants (mean age at baseline survey = 62.8 years), we identified 1,694 individuals with newly diagnosed ADRD (mean age at diagnosis = 74.0 years). Individuals with ADRD were more likely to have lower educational attainment and household income and live alone after marriage (i.e., separated, divorced, and widowed) than noncases (all p < 0.001). Individuals with ADRD who smoked cigarettes reported greater pack-years than noncase counterparts (p = 0.001). Individuals with ADRD had a lower amount of LTPA, a higher score of depressive symptoms, and a higher prevalence of ischemic heart disease, diabetes, hypertension, and depression than individuals without ADRD (all p < 0.001; Table 1).

Table 1.

Baseline Characteristics of Study Participants

Living a healthy lifestyle was associated with a significantly reduced risk of ADRD (Table 2). Never smoking, low-to-moderate alcohol consumption, ≥150/75 minutes of moderate/vigorous LTPA per week (8.3 MET-h/wk), 7–9 hours of sleep, and the highest tertile of HEI were each associated with an 11%–25% reduced risk of ADRD after adjustment for all covariates (model 3: HR [95% CI] = 0.87 [0.76–0.99], 0.81 [0.72–0.92], 0.89 [0.77–1.03], 0.75 [0.64–0.87], and 0.85 [0.75–0.96], respectively). When considering all 5 factors together, individuals with ≥7 or 5–6 of 10 points of a composite healthy lifestyle score showed a 32% and 22% lower risk of ADRD, respectively, than those with 0–3 points (HR [95% CI] = 0.68 [0.58–0.81] for ≥7 points and 0.78 (0.69–0.88) for 5–6 points in Model 3; p trend <0.001). The inverse association was more evident when using a weighted composite score: HRs (95% CIs) from the highest to second-lowest quartiles were 0.64 (0.55–0.74), 0.79 (0.70–0.90), and 0.88 (0.77–0.99) vs the lowest quartile; p trend <0.001. The risk of developing ADRD decreased gradually with increasing scores in a dose-response manner (eTable 1, links.lww.com/WNL/C89).

Table 2.

Associations of Specific Health Behaviors With Risk of ADRD

The beneficial associations of healthy lifestyles were observed consistently regardless of participants' sociodemographic factors, obesity status, medical history, or follow-up time (Figure 1 and eTable 2, links.lww.com/WNL/C89). HRs (95% CIs) for the highest vs lowest quartiles of the weighted lifestyle score were similar between Black and White participants: 0.69 (0.57–0.85) and 0.62 (0.50–0.78), respectively. When comparing the highest vs lowest quartiles of lifestyle score, individuals with higher SES showed a 40%–42% lower risk of ADRD (HR [95% CI] = 0.58 [0.46–0.72] for household income ≥$15,000 per year and 0.60 [0.50–0.72] for ≥ high school graduation), while those with low SES yielded a 28%–31% lower risk (0.72 [0.59–0.88] and 0.69 [0.54–0.90], respectively). No evidence of a significant interaction was found by race, income, or education (all p interaction >0.05). The association seemed stronger among individuals without a history of cardiometabolic disease or depression (HRs ranged from 0.48 to 0.63); however, those with an existing chronic condition(s) also benefit from healthy lifestyles for ADRD prevention (HRs ranged from 0.66 to 0.68; all p interaction >0.05).

Figure 1. Risk of ADRD by Highest vs Lowest Quartiles of a Weighted Composite Lifestyle Score: Stratified Analysis.

The weighted lifestyle score was based on β coefficients of each lifestyle factor in competing risk models with all 5 lifestyle factors and all covariates. HRs (95% CIs) were estimated from competing risk models adjusted for enrollment age, time interval from enrollment to starting follow-up, calendar years of starting follow-up, sex, race, education, income, marital status, enrollment source, history of cardiometabolic disease (i.e., ischemic heart disease, diabetes, hypertension, and dyslipidemia), and history of depression. p interaction was corrected for multiple comparisons by controlling the false discovery rate. The CES-D score ranged from 0 to 30, and a cutoff score of 10 or higher indicates the presence of significant depressive symptoms. Cardiometabolic disease includes ischemic heart disease, diabetes, hypertension, or dyslipidemia. Time interval indicated years from the date of study enrollment to the date of starting follow-up in claims data and was divided by median distribution (4 years). Total observable time indicated years from the date of study enrollment to the date of the end of follow-up in claims data and was divided by median distribution (10 years). ADRD = Alzheimer disease and related dementias; BMI = body mass index; CES-D = Center for Epidemiologic Studies Depression Scale; HR = hazard ratio.

The lifestyle and ADRD associations were not changed in a series of sensitivity analyses, confirming the robustness of our main findings (Table 3).

Table 3.

Risk of ADRD by a Composite Lifestyle Score^a: Sensitivity Analysis

Discussion

In this prospective investigation of primarily low-income Black and White Americans, healthy lifestyles—never smoking, low-to-moderate alcohol consumption, regular exercise, adequate sleep, and healthy eating—were each associated with a reduced risk of ADRD independent of sociodemographic factors and health conditions. The risk of ADRD decreased with an increased number of healthy lifestyles: Individuals with ≥7 or 5–6 of 10 points of the composite score showed a 32% and 22% reduced risk, respectively, compared with those with 0–3 points. Inverse associations remained robust regardless of other ADRD risk factors, including age, race, SES, and history of cardiometabolic disease and depression. Our findings provide novel evidence that healthy lifestyles, individually and together, can benefit all Americans, including those facing systemic disadvantages and disproportionate burdens of ADRD, highlighting the importance of promoting healthy lifestyles and making lifestyle modifications achievable for all.

Modifiable lifestyle factors have attracted increasing attention to prevent or delay the development of ADRD. Evidence from large cohort studies and meta-analysis has accumulated regarding increased risks of cognitive decline and ADRD because of tobacco smoking,^9,10 excessive alcohol drinking,¹⁴ physical inactivity,^11,12 inadequate/disturbed sleep,^15,16 and unhealthy diets.^17,18 Hence, the Alzheimer Association and World Health Organization have recommended lifestyle interventions to reduce the risk of ADRD, including smoking cessation, physical activity, healthy diet, moderate alcohol consumption, and management of obesity, diabetes, hypertension, dyslipidemia, and depression.^8,35 Recently, the combined effect of multiple lifestyle factors has also been suggested. Among 196,383 UK Biobank participants (1,769 incident dementia cases), a favorable lifestyle (≥3 adherence to nonsmoking, regular exercise, healthy diets, and moderate alcohol intake) was associated with a 32% lower dementia risk, including among those with a high ADRD polygenic risk score (HR [95% CI] = 0.68 [0.51–0.90]).¹⁹ Similarly, the Rotterdam Study (n = 6,352 including 915 incident dementia cases) showed that participants with an unfavorable lifestyle (≤2 of 6 factors: nonsmoking, no depression, no diabetes, regular exercise, social activity, and healthy diet) had a 29% higher dementia risk than those with ≥5 factors (HR [95% CI] = 1.29 [1.05–1.59]).²¹ In a pooled analysis of 2 US studies (the Chicago Health and Aging Project and Rush Memory and Aging Project; n = 2,765 with 608 incident ADRD), participants who self-reported ≥4 healthy behaviors (i.e., nonsmoking, physical activity, light-to-moderate alcohol intake, high-quality diet, and engagement in late-life cognitive activities) had a 60% lower risk of ADRD than those with 0 or 1 healthy behavior (HR [95% CI] = 0.40 [0.28–0.56]).²⁰ Of note, this analysis also indicated that the beneficial associations did not differ by APOE genotype, in line with the UK Biobank study. Furthermore, a systematic review and meta-analysis of 18 studies with >40,000 participants summarized that greater numbers of risk factors (e.g., cardiometabolic biomarkers, history of diabetes, hypertension and depression, smoking, and physical inactivity) gradually increased the risk of cognitive decline and ADRD.³⁶ Taken together, epidemiologic evidence to date has supported a protective role of healthy lifestyles against ADRD but remains limited to predominantly White adults. Our study incorporating large numbers of Black and low-SES Americans adds novel and robust evidence on the promising role of healthy lifestyles in ADRD prevention even among disadvantaged populations.

In addition, several large randomized controlled trials have evaluated multidomain lifestyle interventions on cognitive function or dementia risk in older people.^37-39 In the Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability,³⁷ a 2-year intervention of diet, exercise, cognitive training, and vascular risk monitoring among 1,260 adults in Finland who were age 60–77 years and at risk for dementia, a small but significantly better cognitive performance was found in the intervention than in the control group who received general health advice. However, in the Prevention of Dementia by Intensive Vascular care trial,³⁸ a 6-year cluster-intervention targeting vascular and lifestyle risk factors among 3,526 adults aged 70–78 years in the Netherlands, no significant difference was found for dementia incidence or disability score compared with control (usual care). Similarly, the Multidomain Alzheimer Preventive Trial,³⁹ a 3-year lifestyle intervention involving physical activity, cognitive training, and nutritional guidance with or without n-3 polyunsaturated fatty acid supplementation among 1,680 adults in France and Monaco showed no significant effect on cognitive decline vs the placebo. Nevertheless, all those trials did not raise safety concerns, and some potentially beneficial effects have been suggested in participants with cardiometabolic conditions from post hoc analyses.⁴⁰ Overall, evidence from clinical trials regarding lifestyle modifications for ADRD prevention is emerging but remains inconsistent. Meanwhile, most trials were conducted in developed European countries with high standards of living and medical care; thus, future studies are needed among populations with different geographic, racial/ethnic, and socioeconomic backgrounds.

In our current observational study, we found that healthy lifestyles were consistently linked to reduced risk of ADRD in stratified analyses by participants' race, SES, and chronic disease status; no evidence of a significant interaction was found. Although previous studies reported higher likelihoods of ADRD and faster cognitive decline among Black than White Americans after accounting for the SES effect,^41,42 we found comparable beneficial lifestyle and ADRD associations between Black and White participants independent of educational attainment and household income. Of note, all our risk estimates were derived after controlling the competing risk of death, a potential bias for older individuals, which has been proposed for aging-related research but not often controlled in previous studies on lifestyle and ADRD risk. Biologically, healthy lifestyles can enhance immune function, cardiometabolic conditions, and antioxidant and anti-inflammatory properties and reduce neuroinflammation and psychological stress, which in turn help prevent or delay the onset of ADRD.^{1,7,9,13,14,19,20,35} Given that low SES is more likely associated with adverse lifestyles progressing to neural/physical degeneration and cognitive impairment,^1,43-45 providing resources for low-SES populations to promote healthy lifestyles and address their barriers to lifestyle changes may have a substantial impact to reduce racial and socioeconomic disparities in ADRD. Notably, we should recognize that it is challenging for people facing systemic and structural disadvantages to maintain healthy lifestyles or make lifestyle changes. It is imperative to establish public health strategies and social efforts to make lifestyle modifications more achievable for disadvantaged populations.

This study sought to provide novel evidence on healthy lifestyles and incident ADRD among Black Americans and low-income Americans who are disproportionately affected by ADRD but are not as well represented in many cohort studies. Our study population, prospective design, comprehensive data, and clinically diagnosed ADRD allow us to conduct thorough analyses, including adjustment for a series of potential confounders and stratification by race, SES, and other ADRD risk factors. Furthermore, we applied Cox models accommodating the marginal probability of death as a competing risk. Those findings based on enhanced scientific rigors may help lay the groundwork to reduce public health burdens and disparities posed by ADRD in the United States. Nevertheless, we acknowledge several limitations in our current study. First, we could not account for heritable factors (e.g., APOE genotype) because of a lack of data, but previous studies have indicated that healthy lifestyles would benefit older people regardless of genetic risk.^{19,20,43,46,47} Second, although Medicare claims-based algorism has been shown to have high validity for identifying ADRD cases,³¹ we may miss some cases if providers failed to diagnose or document the diagnosis, especially for underserved populations. Indeed, previous studies indicated that low-SES individuals were less likely to be diagnosed with ADRD in claims data even if they had cognitive impairment.^48-50 Third, lifestyle factors were assessed once at baseline; thus, our analyses could not consider any changes in lifestyles. In addition, despite using validated questionnaires, measurement errors in self-reported lifestyle factors may lead to regression dilution, and residual confounding from imperfectly measured or unmeasured confounders might influence our results. Reverse causation may be another concern because we did not have data on cognitive status at baseline, and some individuals may change their lifestyles because of preclinical conditions or comorbidity of other diseases. To address those concerns, we adjusted for existing common chronic diseases, performed stratified analyses by disease status, excluded ADRD cases diagnosed within 2 years of follow-up, and excluded participants with cancer or heart disease—all these analyses confirm the robustness of our findings. Finally, our study was unable to evaluate the lifestyle and ADRD association in subpopulations adequately (e.g., participants with races/ethnicities other than Black or White and sex identities other than men or women) because of insufficient sample sizes or a lack of data. Thus, whether our findings can be widely generalizable to diverse populations in the United States remains to be evaluated.

In summary, our prospective investigation in a population-based cohort of mostly low-income Black and White Americans indicates that healthy lifestyles, including nonsmoking, low-to-moderate alcohol consumption, adequate sleep, regular physical activity, and high-quality diets, are associated with significantly reduced risk of ADRD, regardless of participants' race, SES, and preexisting health conditions. Our findings support the impact of achievable, healthy lifestyles on preventing ADRD that could potentially benefit everyone to eventually reduce the health burdens and disparities posed by ADRD.

Glossary

ADRD

Alzheimer disease and related dementias

CHC

community health center

HEI

Healthy Eating Index

HR

hazard ratio

LTPA

leisure-time physical activity

MET-h/wk

metabolic equivalent hours of LTPA per week

SCCS

Southern Community Cohort Study

SES

socioeconomic status

Appendix. Authors

Study Funding

The Southern Community Cohort Study (SCCS) is funded by grant U01 CA202979 from the National Cancer Institute (NCI) at the NIH, including a supplement for research related to Alzheimer's disease and related dementias (3U01 CA202979-03S1). SCCS data collection was performed by the Survey and Biospecimen Shared Resource, which is supported in part by the Vanderbilt-Ingram Cancer Center (P30 CA68485). This analysis was partly supported by Vanderbilt Memory and Alzheimer's Center Pilot and Feasibility Funding to D. Yu.

Disclosure

The authors report no disclosures relevant to the manuscript. Go to Neurology.org/N for full disclosures.