Adherence to Life’s Simple 7 and Cognitive Function Among Older Adults: The National Health and Nutrition Examination Survey 2011 to 2014

Abstract

Background

The American Heart Association proposed the Life’s Simple 7 (LS7; including diet, physical activity, smoking, body mass index, blood pressure, plasma fasting glucose, total cholesterol) to promote cardiovascular health. Adherence to LS7 has been found to be associated with better cognitive health as well, but the generalizability of previous studies is limited. We aimed to examine the associations of adherence to LS7 and cognitive function among older adults in a nationally representative sample of population.

Methods and Results

A total of 2585 older adults (≥60 years, 54% female, 80% non‐Hispanic White) in the National Health and Nutrition Examination Survey 2011 to 2014 were included for analysis. Components of LS7 were measured, and adherence to LS7 was calculated on the basis of established cutoff points of individual components. Cognitive function was examined using the Consortium to Establish a Registry for Alzheimer’s Disease Word List Memory Task (immediate and delayed memory), Digit Symbol Substitution Test (DSST), and Animal Fluency Test. Test‐specific and global cognition Z scores were created. Multivariable linear regression models were conducted on the associations of adherence to LS7 with domain‐specific and global cognition Z scores. Each incremental point in adherence to LS7 was associated with higher Z scores for global cognition (β=0.05; 95% CI, 0.02–0.07), Digit Symbol Substitution Test (β=0.05; 95% CI, 0.03–0.07), Consortium to Establish a Registry for Alzheimer’s Disease Word Learning subtest immediate memory (β=0.03; 95% CI, 0.004–0.05), and animal fluency test (β=0.05; 95% CI, 0.02–0.07).

Conclusions

Greater adherence to LS7 metrics is associated with better cognitive function among older US adults in a nationally representative sample of population.

Nonstandard Abbreviations and Acronyms

AFT

Animal Fluency Test

CERAD‐WL

Consortium to Establish a Registry for Alzheimer’s Disease Word Learning subtest

DSST

Digit Symbol Substitution Test

FPG

fasting plasma glucose

LS7

Life’s Simple 7

NHANES

National Health and Nutrition Examination Survey

TC

total cholesterol

Clinical Perspective

What Is New?

• Greater adherence to Life’s Simple 7 metrics is associated with better cognitive function among older US adults in a nationally representative sample of population.

• The adherence to certain components of Life’s Simple 7, including diet and physical activity, has not been satisfactory.

What Are the Clinical Implications?

• Since no threshold was detected for the association between Life’s Simple 7 scores and cognitive function, population strategies that aim for modest improvement of Life’s Simple 7 in the whole population is critical for reducing the burden of dementia.

Dementia is a growing public health epidemic in the United States and globally. ¹ It is estimated that the number of people around the world living with dementia in 2050 will reach 100 million, growing from 43 to 47 million people in 2015/2016. ² Dementia is also ranked among the top 5 causes of death. ² Since no cure for dementia is currently available, prevention of dementia becomes particularly important. Dementia refers to a clinical syndrome characterized by loss of cognitive function, and it is therefore important to identify risk factors associated with poor cognitive function.

The existing literature indicates that dementia is associated with a variety of risk factors, including those related to cardiovascular disease. ³ , ⁴ Particularly, some of the modifiable cardiovascular risk factors such as physical inactivity, cigarette smoking, and high blood pressure are also important to an individual’s brain health. ⁵ The American Heart Association proposed Life’s Simple 7 (LS7), a matrix to promote cardiovascular health, which includes 7 modifiable cardiovascular risk factors, namely, diet, physical activity, cigarette smoking, body mass index (BMI), blood pressure, fasting plasma glucose (FPG), and total cholesterol (TC). ⁶ The associations of 7 individual components with cognitive outcomes have been widely studied, ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ and a few recent studies have also examined the associations between the adherence of LS7 and cognitive outcomes, and higher score of LS7 are associated with better cognitive outcomes in several domains. ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ However, most of these studies were based on individual cohorts with different characteristics, and it is desirable that the associations are examined in a sample representative of the US population.

Therefore, we aimed to examine the association between adherence to LS7 and cognitive function among older adults. The National Health and Nutrition Examination Survey (NHANES) data have been used to monitor the health of the US population and allow examination of prevalent health conditions and their risk factors. With the large and nationally representative sample of the US population, the NHANES study is ideal to examine the association between adherence to LS7 and cognitive function among older adults.

Methods

Anonymized data and materials have been made publicly available at the NHANES website and can be accessed at https://wwwn.cdc.gov/nchs/nhanes/Default.aspx.

The NHANES is a serial ongoing cross‐sectional survey of the civilian, noninstitutionalized US population, conducted by the National Center for Health Statistics at the Centers for Disease Control and Prevention. The 2011 to 2014 NHANES conducted measurements of LS7 components and cognitive performance for participants. We combined data from the 2 survey cycles for analysis. A total of 19 931 people of all ages were enrolled for NHANES interviews from 2011 to 2014. Among these participants, we excluded participants who were aged <60 years (n=16 299), those who had missing information of cognitive tests (n=444), those who missed any information in components of LS7 (diet, physical activity, smoking, BMI, blood pressure, FPG, TC; n=408), those who had missing information in education (n=2), and those who had prevalent stroke (n=193). Our final sample size for analyses included a total of 2585 participants aged ≥60 years.

The NHANES was conducted in accordance with the recommendations of National Center for Health Statistics Research Ethics Review Board with written informed consent from all subjects, according to the Declaration of Helsinki. The protocol was approved by the National Center for Health Statistics Research Ethics Review Board.

The LS7 metrics is defined, per American Heart Association guidelines, ⁶ as shown in Table S1. Dietary intake in NHANES was measured using two 24‐hour dietary recall interviews among all participants. The first interview was conducted in person in the Mobile Examination Center, and the second interview was conducted via telephone 3 to 10 days after the first interview. ²² We also used the intake of food and nutrients using the Food Patterns Equivalents Database from the US Department of Agriculture, 2011 to 2014. ²³ Physical activity was measured using the validated Global Physical Activity Questionnaire, which asked participants to report time spent on moderate and vigorous activity. ²⁴ BMI was calculated with body height and weight measured by trained health technicians. Cholesterol and glucose were analyzed with blood specimens at University of Minnesota (Minneapolis, MN). Three consecutive blood pressure readings were obtained by certified blood pressure in the mobile examination center, and a fourth attempt was made if the measurement was incomplete. An average of blood pressure was calculated for evaluation. Information of smoking (not including secondhand smoking) was self‐reported by participants. To calculate a total score for LS7, a component that achieved ideal, intermediate, and poor were given 2 points, 1 point, and 0 point, respectively. Therefore, the total score ranged from 0 to 14 points, with higher scores indicating better cardiovascular health.

The cognitive tests in NHANES 2011 to 2014 include 3 components: the Digit Symbol Substitution Test (DSST), the Consortium to Establish a Registry for Alzheimer’s Disease Word Learning subtest (CERAD‐WL), and the Animal Fluency test (AFT). The DSST focused on executive function and processing speed, which was conducted using a paper form with a key at the top that contained 9 numbers paired with symbols. Participants were asked to copy the corresponding symbols in the 133 boxes that adjoin the numbers in 2 minutes, and a point was awarded for each correct match. ²⁵ The CERAD‐WL tested immediate (including 3 consecutive learning trials) and delayed verbal memory (1 delayed recall trial). During the learning trial, participants were asked to read 10 unrelated words aloud 1 at a time as presented, with the order of the 10 words being changed for each trial. In the delayed recall test, which occurred after the other 2 cognitive tests, participants were asked to then recall as many words as possible, for those in the same 10‐word list. The score ranges from 0 to 133 as the sum of correct matches. ²⁶ The AFT assessed language ability and executive function, in which participants were asked to name as many animals as possible in 1 minute, and 1 point was awarded for each named animal. ²⁷

In addition to these measurements, we considered age, sex, race and ethnicity, education level, and coronary heart disease as covariates, which were obtained from self‐reported questionnaires.

Descriptive statistics were computed on participants’ characteristics. The prevalence of ideal, intermediate, and poor levels for each component of LS7 were described, and the proportion of different numbers in ideal and poor levels of LS7 components was calculated. Test‐specific Z scores (including DSST, CERAD‐WL delayed memory, CERAD‐WL immediate memory, AFT) were created using sample means and SDs of test scores. A standardized global cognition Z score was then generated by averaging the test‐specific Z scores divided by SD. Multivariable linear regression models were used to examine the associations of different levels of each LS7 component with test‐specific and global cognition Z scores. Furthermore, multivariable linear regression models were used to examine the associations of total LS7 score (continuous and by quintiles) with test‐specific and global cognition Z scores, which were also stratified by sex and race and ethnicity. In addition, we also examined the associations of each incremental ideal and poor item with test‐specific and global cognition Z scores. Finally, we examined the associations of different clusters of behavioral factors (diet, physical activity, smoking, BMI) and biological factors (blood pressure, FPG, TC) with cognitive function. The estimated effect sizes shown as βs and 95% CIs were estimated and presented in the models. The analyses were weighted for the stratified, multistage probability sampling design of NHANES and survey nonresponse. Full sample 2‐year mobile examination center exam weight was used for subsamples in 2011 to 2012 and 2013 to 2014, and the weights were recalculated (divided by 2) after the combination of NHANES 2011 to 2012 with 2013 to 2014. ²⁸ All analyses were conducted using SAS 9.4 (SAS Institute, Cary, NC).

Results

A total of 2585 participants were included in the analysis, including 1251 participants in 2011 to 2012 and 1334 participants in 2013 to 2014, which represented a population size of 47 684 724. The characteristics of participants are shown in Table 1. More than half of the participants were female and 13.1% were aged ≥80 years. About 80% of participants were non‐Hispanic White participants, and 83% had an education level of high school or above. About 9% of all participants had self‐reported coronary heart disease. Compared with those who were aged ≥60 years but not included in the analysis, the included participants had a smaller proportion of older adults aged ≥80 years, a larger proportion of non‐Hispanic White participants, higher level of education, a lower proportion of people smoking, a larger BMI, TC, and diastolic blood pressure. In addition, those who were included had better performance in all cognitive tests (Table S2).

Table 1.

Weighted Characteristics of Participants Included in the Analysis.

	Mean (95% CI) or n (%)
Weighted frequency	47 684 724
Age, ≥80 y	387 (13.1)
Sex, female, n (%)	1319 (54.1)
Race and ethnicity, n (%)
Mexican American	241 (3.5)
Other Hispanic	277 (3.9)
Non‐Hispanic White	1216 (79.7)
Non‐Hispanic Black	621 (8.4)
Other races*	230 (4.6)
Education, n (%)
<9th grade	314 (6.0)
9–11th grade	365 (10.1)
High school graduate	586 (21.6)
Some college	719 (31.5)
College graduate or above	601 (30.7)
Ever smoking, n (%)	1261 (48.7)
Moderate to vigorous physical activity, min/wk	121.8 (107.6–136.1)
Diet
Sodium intake, g/day	3.1 (3.0–3.2)
Sugar sweetened beverage, g/day	246.8 (227.3–266.4)
Fish consumption, oz/day	0.7 (0.6–0.8)
Fruits and vegetables, cups/day	2.0 (1.9–2.2)
Whole grain, oz/day	1.1 (1.1–1.2)
Body mass index, kg/m2	29.3 (28.9–29.7)
Total cholesterol, mg/dL	193.6 (190.9–196.4)
Fasting plasma glucose, mg/dL	112.7 (109.7–115.8)
Systolic blood pressure, mm Hg	132.1 (131.0–133.2)
Diastolic blood pressure, mm Hg	68.6 (67.5–69.6)
Coronary heart disease, n (%)	222 (9.0)
Cognitive function
Digit symbol score	53.2 (52.0–54.3)
Delayed recall score	6.3 (6.1–6.5)
Immediate recall score	19.8 (19.3–20.2)
Animal Fluency Test score	18.3 (17.9–18.7)

^*Other not defined by The NHANES but it does include multi‐racial and non‐Hispanic Asian.

Among all components of the LS7, smoking (88.4%) and FPG (59.1%) had a high prevalence of ideal level, while that of diet was <1%. The prevalence of poor level for both diet and physical activity was >50% (Figure). For the number of ideal or poor level in individual components achieved, no participant had all 7 components ideal or poor. About 80% of the participants achieved ideal levels in 2 to 4 LS7 components, and a similar proportion was observed for participants who achieved poor levels (Figure S1).

Figure 1. Prevalence of poor, intermediate, and ideal levels achieved in each component of Life’s Simple 7 among US adults aged ≥60 years, NHANES 2011 to 2014.

 

After adjustment for age, sex, race and ethnicity, education, other lifestyle components, and coronary heart disease, compared with poor level, intermediate level of physical activity was associated with higher Z scores of global cognition (β=0.17; 95% CI, 0.08–0.27), DSST (β=0.13; 95% CI, 0.03–0.24), CERAD‐WL immediate memory (β=0.16; 95% CI, 0.04–0.27), AFT (β=0.17; 95% CI, 0.06–0.27), and ideal level of physical activity was associated with higher Z scores of global cognition (β=0.22; 95% CI, 0.12–0.33), DSST (β=0.20; 95% CI, 0.08–0.31), and AFT (β=0.33; 95% CI, 0.22–0.43). Ideal level of smoking was associated with a higher Z score of DSST (β=0.20; 95% CI, 0.08–0.32); intermediate level of blood pressure was associated with higher Z scores of global cognition (β=0.11; 95% CI, 0.03–0.19) and CERAD‐WL immediate memory (β=0.14; 95% CI, 0.02–0.25), and ideal level of blood pressure was associated with higher Z scores of global cognition (β=0.19; 95% CI, 0.02–0.35), CERAD‐WL immediate memory (β=0.21; 95% CI, 0.04–0.39), and AFT (β=0.16; 95% CI, 0.02–0.30); ideal level of FPG was associated with higher Z scores of DSST (β=0.22; 95% CI, 0.04–0.41) and AFT (β=0.15; 95% CI, 0.001–0.31) (Table 2).

Table 2.

Multivariable Linear Regression for the Associations of Individual Life’s Simple 7 Components With Test‐Specific and Global Cognition^*

	Diet		PA		Smoking					BMI				BP			FPG		TC
	β (95% CI)	P value	β (95% CI)	P value	β (95% CI)				P value	β (95% CI)			P value	β (95% CI)	P value		β (95% CI)	P value	β (95% CI)	P value
Poor	Reference		Reference		Reference					Reference				Reference			Reference		Reference
Global cognition
Intermediate	0.0005 (−0.10 to 0.11)	0.9925	0.17 (0.08 to 0.27) †	0.0007 †	0.16 (−0.13 to 0.56)				0.3360	‐0.06 (−0.16 to 0.04)			0.2379	0.11 (0.03 to 0.19) †	0.0108 †		0.11 (−0.08 to 0.29)	0.2404	‐0.09 (−0.24 to 0.05)	0.2042
Ideal	0.001 (−0.36 to 0.36)	0.9977	0.22 (0.12 to 0.33) †	0.0001 †	0.08 (−0.07 to 0.22)				0.2871	‐0.05 (−0.17 to 0.08)			0.4534	0.19 (0.02 to 0.35) †	0.0275 †		0.14 (−0.05 to 0.33)	0.1510	‐0.06 (−0.21 to 0.10)	0.4681
DSST
Intermediate	0.05 (−0.06 to 0.15)	0.3667	0.13 (0.03 to 0.24) †	0.0125 †	0.29 (−0.10 to 0.68)			0.1356		0.02 (−0.08 to 0.12)			0.6753	0.06 (−0.03 to 0.16)	0.1636		0.15 (−0.05 to 0.35)	0.1264	‐0.01 (−0.12 to 0.10)	0.8637
Ideal	0.29 (−0.20 to 0.77)	0.2403	0.20 (0.08 to 0.31) †	0.0011 †	0.20 (0.08 to 0.32) †			0.0019 †		‐0.07 (−0.18 to 0.05)			0.2534	0.11 (−0.02 to 0.24)	0.1021		0.22 (0.04 to 0.41) †	0.0177 †	‐0.05 (−0.17 to 0.07)	0.3650
CERAD‐WL delayed memory
Intermediate	0.001 (−0.11 to 0.11)	0.9788	0.08 (−0.04 to 0.20)	0.1957	‐0.19 (−0.60 to 0.23)		0.3647			‐0.10 (−0.21 to 0.01)	0.0666			0.09 (−0.01 to 0.20)		0.0828	0.05 (−0.22 to 0.31)	0.7293	‐0.11 (−0.27 to 0.06)	0.2012
Ideal	0.10 (−0.36 to 0.55)	0.6632	0.09 (−0.02 to 0.20)	0.1219	‐0.003 (−0.16 to 0.15)		0.9719			‐0.06 (−0.18 to 0.06)	0.3011			0.14 (−0.07 to 0.35)		0.1838	0.04 (−0.22 to 0.29)	0.7756	0.01 (−0.15 to 0.16)	0.9074
CERAD‐WL immediate memory
Intermediate	0.03 (−0.08 to 0.14)	0.5667	0.16 (0.04 to 0.27) †	0.0074 †	0.17 (−0.11 to 0.46)	0.2271				‐0.04 (−0.12 to 0.05)		0.4014		0.14 (0.02 to 0.25) †		0.0187 †	0.01 (−0.19 to 0.20)	0.9537	‐0.14 (−0.30 to 0.02)	0.0750
Ideal	‐0.15 (−0.70 to 0.40)	0.5893	0.07 (−0.03 to 0.17)	0.1717	0.03 (−0.13 to 0.20)	0.6850				‐0.01 (−0.12 to 0.10)		0.8490		0.21 (0.04 to 0.39) †		0.0198	‐0.02 (−0.20 to 0.15)	0.8060	‐0.03 (−0.22 to 0.16)	0.7416
AFT
Intermediate	‐0.09 (−0.19 to 0.01)	0.0773	0.17 (0.06 to 0.27) †	0.0026 †	0.26 (−0.15 to 0.68)	0.2063				‐0.06 (−0.18 to 0.05)		0.2782		0.08 (−0.04 to 0.19)	0.1804		0.14 (−0.01 to 0.28)	0.0712	‐0.09 (−0.24 to 0.05)	0.2068
Ideal	‐0.12 (−1.33 to 1.09)	0.8379	0.33 (0.22 to 0.43) †	<0.0001 †	0.01 (−0.16 to 0.18)	0.9211				‐0.04 (−0.20 to 0.12)		0.6224		0.16 (0.02 to 0.30) †	0.0226 †		0.15 (0.001 to 0.31) †	0.0480 †	‐0.10 (−0.28 to 0.07)	0.2479

AFT indicates Animal Fluency Test; BMI, body mass index; BP, blood pressure; CERAD‐WL, Consortium to Establish a Registry for Alzheimer’s Disease Word Learning subtest; DSST, Digit Symbol Substitution Test; FPG, fasting plasma glucose; PA, physical activity; and TC, total cholesterol.

^*Models were adjusted for age, sex, race and ethnicity, education, other Life’s Simple 7 components, coronary heart disease.

^† P<0.05.

After adjustment for age, sex, race and ethnicity, education, and coronary heart disease, each unit increase in LS7 score was associated with higher Z scores for global cognition (β=0.05; 95% CI, 0.02–0.07), DSST (β=0.05; 95% CI, 0.03–0.07), CERAD‐WL immediate memory (β=0.03; 95% CI, 0.004–0.05), and AFT (β=0.05; 95% CI, 0.02–0.07). Compared with participants with the lowest quintile of LS7 score, those who were in the highest quintile were associated with higher Z scores of global cognition (β=0.29; 95% CI, 0.15–0.43), DSST (β=0.29, 95% CI, 0.16–0.42), CERAD‐WL immediate memory (β=0.16; 95% CI, 0.03–0.28), and AFT (β=0.27; 95% CI, 0.13–0.41) (Table 3). Stratified by sex, higher LS7 score was associated with higher Z scores of global cognition, DSST, and AFT among men, as well as higher Z scores of global cognition and DSST among women. The highest quintile of LS7 was associated with higher Z scores of DSST and AFT among men, as well as higher Z scores of global cognition, DSST, CERAD‐WL delayed memory, and AFT among women (Table S3). Stratified by race and ethnicity, higher LS7 scores were associated with higher Z scores of global cognition, DSST, CERAD‐WL immediate memory, and AFT among non‐Hispanic White participants. The highest quintile of LS7 was associated with higher Z scores of global cognition, DSST, CERAD‐WL immediate memory, and AFT among non‐Hispanic White participants. Higher LS7 was associated with a lower Z score of AFT among participants other than Hispanics, non‐Hispanic White participants, and Non‐Hispanic Black participants (Table S4).

Table 3.

Multivariable Linear Regression for the Associations of Total Life’s Simple 7 Score With Test‐Specific and Global Cognition^*

	Quintile 1 Total score<6 (n=609)	Quintile 2 Total score=6 (n=512)		Quintile 3 Total score=7 (n=473)		Quintile 4 Total score=8 (n=425)		Quintile 5 Total score>8 (n=566)		Continuous score
	β (95% CI)	β (95% CI)	P value	β (95% CI)	P value	β (95% CI)	P value	β (95% CI)	P value	β (95% CI)	P value
Global cognition	Reference	0.15 (−0.01 to 0.31)	0.0681	0.15 (0.04 to 0.27) †	0.0108 †	0.12 (−0.02 to 0.27)	0.0969	0.29 (0.15 to 0.43) †	0.0001 †	0.05 (0.02 to 0.07) †	0.0002 †
DSST	Reference	0.10 (−0.02 to 0.23)	0.1027	0.16 (0.02 to 0.30) †	0.0255 †	0.26 (0.12 to 0.40) †	0.0007 †	0.29 (0.16 to 0.42) †	<0.0001 †	0.05 (0.03 to 0.07) †	<0.0001
CERAD‐WL delayed memory	Reference	0.11 (−0.07 to 0.28)	0.2190	0.09 (−0.05 to 0.23)	0.2196	‐0.03 (−0.20 to 0.14)	0.7266	0.13 (−0.02 to 0.28)	0.0862	0.02 (−0.01 to 0.04)	0.1254
CERAD‐WL immediate memory	Reference	0.15 (−0.02 to 0.31)	0.0757	0.04 (−0.07 to 0.14)	0.4668	0.02 (−0.15 to 0.19)	0.8102	0.16 (0.03 to 0.28) †	0.0201 †	0.03 (0.004 to 0.05) †	0.0218 †
AFT	Reference	0.09 (−0.08 to 0.26)	0.2785	0.16 (−0.01 to 0.32)	0.0595	0.08 (−0.07 to 0.24)	0.2798	0.27 (0.13 to 0.41) †	0.0006 †	0.05 (0.02 to 0.07) †	0.0004 †

AFT indicates Animal Fluency Test; CERAD‐WL, Consortium to Establish a Registry for Alzheimer’s Disease Word Learning subtest; and DSST, Digit Symbol Substitution Test.

^*Models were adjusted for age, sex, race and ethnicity, education, other Life’s Simple 7 components, coronary heart disease.

^† P<0.05.

Each increment on ideal components was associated with higher Z scores of global cognition (β=0.07; 95% CI, 0.02–0.11), DSST (β=0.06; 95% CI, 0.02–0.10), and AFT (β=0.07; 95% CI, 0.02–0.11). In contrast, each increment on poor items was associated with lower Z scores of global cognition (β=−0.07; 95% CI, −0.11 to −0.04), DSST (β=−0.09; 95% CI, −0.13 to −0.06), CERAD‐WL immediate memory (β=−0.04; 95% CI, −0.08 to −0.01), and AFT (β=−0.07; 95% CI, −0.10 to −0.03) (Table S5).

For clusters created on the basis of behavioral factors (diet, physical activity, smoking, BMI) and biological factors (blood pressure, FPG, TC), combinations of better biological factors were associated with better cognitive function, while combinations of better behavioral factors were associated with better or poorer cognitive function, although the sample size was small (Tables S6 and S7).

Discussion

In this nationally representative sample of older adults, the adherence to LS7 was not adequate among older adults in the United States, particularly in diet, physical activity, FPG, and blood pressure. We found that better adherence to LS7 was associated with better global and test‐specific cognitive function.

The associations between adherence to LS7 and cognitive outcomes have been examined in previous studies. González et al ¹⁴ found that better LS7 metrics were associated with higher scores in verbal learning and memory, verbal fluency, and psychomotor processing speed in a cohort of 9623 middle‐aged and older Hispanics/Latinos (aged 45–74 years). Thacker et al ¹⁶ found that compared with low adherence to LS7, intermediate and high levels of adherence were both associated with lower incidence of cognitive impairment in 17 761 middle‐aged and older adults with normal global cognitive status at baseline. Samieri et al ¹⁷ found that a higher number of LS7 scores was associated with a lower risk of dementia and lower rates of cognitive decline over a mean follow‐up of 8.5 years among 6626 older adults. González et al found that higher midlife LS7 score was associated with lower cognitive decline over 20 years among 13 270 White and Black participants. ¹⁵ All of these studies pointed to the associations between higher adherence to LS7 and better cognitive function. The results of our study are consistent with these previously published studies. Given the fact that our study is based on a large and nationally representative sample of the US population, it provides stronger evidence of importance of the association between LS7 and cognitive function. A previous study indicated that the rate of cognitive decline among older adults with normal cognitive aging is 0.04 SDs per year, ²⁹ and our results suggested that healthy lifestyle may preserve cognition for a few more years.

The associations between LS7 and cognitive function among older adults may be explained by the associations of individual components of LS7 with cognitive function, and each of the associations has different mechanisms. Physical inactivity and smoking are associated with higher risk of hypertension, diabetes, and cardiovascular disease, which are highly related to cognitive impairment. ³⁰ , ³¹ , ³² , ³³ , ³⁴ , ³⁵ Physical activity may also enhance social engagement and reduce depressive symptoms, which could improve cognitive function. ³⁶ , ³⁷ Physical activity, blood pressure, fasting glucose, and smoking are associated with cerebral small‐vessel disease, which is a predictor of cognitive impairment. ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ In addition, inflammation may also play a role between lifestyle behaviors, such as diet and physical activity with cognitive health. ⁴² , ⁴³

Although the importance of lifestyle in the maintenance and improvement of cardiovascular and cognitive health has been widely acknowledged, the adherence to some of the LS7 components has been poor. Particularly, the prevalence of ideal diet is close to 0, which is consistent with previous report in the United States from 2003 to 2016. ⁴⁴ This suggested that health‐related behaviors, such as diet and physical activity are unlikely to be changed in a simple manner. Therefore, a joint effort of public health professionals, experts of medicine, and policy makers is needed to make changes to lifestyle behaviors. Also, given the low prevalence of ideal level in physical activity and diet, the goal for diet and physical activity may be set to a more realistic and feasible level, so that older adults may be more motivated to engage in physical activity and pursue a healthy diet.

This study provided important results to guide future public health endeavors. Dementia has been associated with multiple risk factors, and it may be reasonable to prevent dementia through targeting a multitude of risk factors simultaneously to make the prevention more scalable. As shown in our results, better levels of physiological factors (such as blood pressure) and greater adherence to healthy lifestyle factors were associated with better cognitive function. Also, small changes in lifestyle factors and behaviors could potentially be beneficial for improving cognition. Given the fact that other prospective cohort studies have also found the associations between adherence to LS7 and beneficial cognitive outcomes, multidomain lifestyle intervention should be considered. A few multidomain lifestyle trials have been conducted for prevention of cognitive outcomes, such as the FINGER (Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability), ⁴⁵ MAPT (Multidomain Alzheimer Preventive Trial), ⁴⁶ and preDIVA (Prevention of Dementia by Intensive Vascular Care) trials. ⁴⁷ Some of the results were mixed, and it may be likely from the limitations of the fact that participants were in their late life and had a short period of intervention, while the process of cognitive aging starts from an early stage in life and lasts for a long time. However, it is challenging to conduct these intervention trials among midlife adults for several decades. Therefore, alternative epidemiological methods may be applied to estimate the long‐term impact of multidomain lifestyle intervention in preventing cognitive outcomes among older adults in population. ⁴⁸

In our study, the association between adherence to LS7 and cognitive function was linear, which suggests that there is no threshold for the LS7 score in its relationship with cognition, and any small increment in the score is helpful for improving cognitive performance. Therefore, “population strategy”, ⁴⁹ which aims for modest improvement of risk factors in the whole population, is critical for reducing the burden of dementia.

One major limitation of our study is the cross‐sectional design, which precludes causal inferences for LS7 and cognitive function among older adults. We cannot exclude the possibility that poor cognitive function can potentially reduce adherence to LS7. Also, some components of LS7 (eg, physical activity, smoking, coronary heart disease) were obtained by self‐report. This may contribute to misclassification, as participants may overreport their physical activity while underreporting their unhealthy diet. We also lacked information on the participants’ places of residence as neighborhood socioeconomic status may be an important covariate for our analyses. Selection bias is also possible, with different characteristics of participants included and not included in the analysis. It is possible that our findings were more conservative, as participants who were excluded were older, reported a lower level of education, and had worse health status than those included in the analysis. Finally, there are only 3 sets of cognitive tests, which may not be sufficient to capture the overall cognitive function. Despite these limitations, we had access to validated assessments of cognitive function and all components of LS7 (including biological factors) that strengthened our analyses. Furthermore, NHANES has a large and diverse representative sample, which makes the results generalizable to the US population.

Conclusions

In conclusion, greater adherence to LS7 metrics is associated with better cognitive function among older adults. Future research is expected to learn how improvement in lifestyle and physiological factors can affect cognitive performance and understanding the mechanistic pathways.

Sources of Funding

This work was supported by National Institute on Minority Health and Health Disparities #U54MD012530 (to Dr Xu). Dr Kulshreshtha was supported by a National Institutes of Health grant (K23 AG066931‐01A1). The funding agencies had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. No financial disclosures were reported by the authors of this paper.

Disclosures

None.

Supporting information

Tables S1–S7

Figure S1