This cohort study uses pooled data from 6 US population-based cohort studies to assess whether incident myocardial infarction is associated with changes in cognition immediately after in the years following the event.
Key Points
Question
Is myocardial infarction (MI) associated with cognition acutely after MI or in the years following MI?
Findings
In this cohort study of 30 465 adults without MI, stroke, or dementia, overall, incident MI was not associated with an acute decrease in global cognition, memory, or executive function at the time of the event compared with no MI. The rate of decline in global cognition, memory, and executive function was significantly faster over the years for adults with an MI event compared with those without an MI.
Meaning
These findings suggest that prevention of MI may be important for long-term brain health.
Abstract
Importance
The magnitude of cognitive change after incident myocardial infarction (MI) is unclear.
Objective
To assess whether incident MI is associated with changes in cognitive function after adjusting for pre-MI cognitive trajectories.
Design, Setting, and Participants
This cohort study included adults without MI, dementia, or stroke and with complete covariates from the following US population-based cohort studies conducted from 1971 to 2019: Atherosclerosis Risk in Communities Study, Coronary Artery Risk Development in Young Adults Study, Cardiovascular Health Study, Framingham Offspring Study, Multi-Ethnic Study of Atherosclerosis, and Northern Manhattan Study. Data were analyzed from July 2021 to January 2022.
Exposures
Incident MI.
Main Outcomes and Measures
The main outcome was change in global cognition. Secondary outcomes were changes in memory and executive function. Outcomes were standardized as mean (SD) T scores of 50 (10); a 1-point difference represented a 0.1-SD difference in cognition. Linear mixed-effects models estimated changes in cognition at the time of MI (change in the intercept) and the rate of cognitive change over the years after MI (change in the slope), controlling for pre-MI cognitive trajectories and participant factors, with interaction terms for race and sex.
Results
The study included 30 465 adults (mean [SD] age, 64 [10] years; 56% female), of whom 1033 had 1 or more MI event, and 29 432 did not have an MI event. Median follow-up was 6.4 years (IQR, 4.9-19.7 years). Overall, incident MI was not associated with an acute decrease in global cognition (−0.18 points; 95% CI, −0.52 to 0.17 points), executive function (−0.17 points; 95% CI, −0.53 to 0.18 points), or memory (0.62 points; 95% CI, −0.07 to 1.31 points). However, individuals with incident MI vs those without MI demonstrated faster declines in global cognition (−0.15 points per year; 95% CI, −0.21 to −0.10 points per year), memory (−0.13 points per year; 95% CI, −0.22 to −0.04 points per year), and executive function (−0.14 points per year; 95% CI, −0.20 to −0.08 points per year) over the years after MI compared with pre-MI slopes. The interaction analysis suggested that race and sex modified the degree of change in the decline in global cognition after MI (race × post-MI slope interaction term, P = .02; sex × post-MI slope interaction term, P = .04), with a smaller change in the decline over the years after MI in Black individuals than in White individuals (difference in slope change, 0.22 points per year; 95% CI, 0.04-0.40 points per year) and in females than in males (difference in slope change, 0.12 points per year; 95% CI, 0.01-0.23 points per year).
Conclusions
This cohort study using pooled data from 6 cohort studies found that incident MI was not associated with a decrease in global cognition, memory, or executive function at the time of the event compared with no MI but was associated with faster declines in global cognition, memory, and executive function over time. These findings suggest that prevention of MI may be important for long-term brain health.
Introduction
Dementia is common, costly, highly morbid, and lacking treatments.1 An understanding of the vascular contributions to cognitive decline could identify potential targets for interventions to slow or prevent dementia.2 Acute myocardial infarction (MI) is a severe manifestation of coronary artery disease and affects more than 7 million people worldwide.3,4 While MI has previously been associated with incident dementia and cognitive decline,5,6,7 meta-analyses have been limited by heterogeneity among studies,5 and other prospective cohorts suggest no association.8 Importantly, the magnitude of changes in cognition around the time of the event and the long-term post-MI cognitive trajectory controlling for the individual’s pre-MI cognitive trajectory have not been well described. Prior studies have lacked repeated, standardized measures of cognition before and after the MI event or to have combined MI with another cardiovascular outcome.5,6,7,8 It is also unclear whether the association of incident MI with cognition varies by race and sex or if a second MI event might be more detrimental to brain health than 1 MI event, similar to what a previous study has shown for stroke.9 We conducted a pooled analysis of 6 population-based cohort studies to assess whether incident MI is associated with faster long-term cognitive decline compared with pre-MI cognitive trajectories, accounting for the acute change in cognition at the time of an MI event. If an association was found, we also aimed to assess whether the association between incident MI and cognitive decline was different among individuals with more than 1 MI event or by race or sex and whether these associations persisted among those without atrial fibrillation (AF), as AF is associated with cognitive decline and can occur after MI.10
Methods
Study Population and Ethics
This cohort study included individuals from the following 6, well-characterized, US prospective cohort studies conducted from 1971 to 2019, with physician-adjudicated MI cases and repeated measures of cognition and blood pressure (BP): Atherosclerosis Risk in Communities Study, Coronary Artery Risk Development in Young Adults Study, Cardiovascular Health Study, Framingham Offspring Study, Multi-Ethnic Study of Atherosclerosis, and Northern Manhattan Study. Individuals were excluded from the study if they had a history of MI, dementia, or stroke at cohort baseline; incident MI, dementia, or stroke before the first cognitive assessment (except for the Northern Manhattan Study participants, for whom this information was not available); no BP measurement before the first cognitive assessment; and lack of complete data on covariates of interest. The analysis represents a complete case analysis, as only 3.5% of the total population had missing covariates (eFigure 1 in Supplement 1). Individuals had 1 or more cognitive assessments, while individuals with an incident MI had 1 or more cognitive assessments before and 1 or more assessments after MI. Individuals’ cognitive observations were censored after incident stroke if it occurred during follow-up. This study was approved by each study’s institutional review board, and all participants gave written informed consent. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline was followed.
Measures of Cognition
The primary outcome measure was change in global cognition. Secondary outcomes were change in memory and change in executive function. Details about the harmonization of cognitive measures across the 6 cohorts have been previously described.11 To make inferences about cognitive domains instead of individual cognitive test items and to account for different cognitive tests administered across cohorts, we cocalibrated available cognitive test items into factors representing global cognition (executive function or processing speed; learning or memory; general mental status; and language, motor, and visuospatial ability), memory (learning or delayed recall), and executive function (complex or speeded cognitive function) using item response theory methods that leverage all available cognitive information in common across cohorts and test items unique to particular cohorts.12 In a prestatistical harmonization phase, we identified 153 test items from 34 cognitive instruments across the cohorts and determined shared items among cohorts. Expert neuropsychologists assigned each test item to a cognitive domain. In item response theory, each test item is weighted based on its correlation with other items and empirically assigned a relative location along the latent trait (eg, global cognition) corresponding to its estimated difficulty. We computed factor scores for each domain. Cognitive outcomes were set to a mean (SD) T score metric of 50 (10) at an individual’s first cognitive assessment with a 1-point difference representing a 0.1-SD difference in the distribution of cognition across the 6 cohorts. Higher cognitive scores indicate better cognitive performance.
Incident MI
Incident MI was measured during cohort follow-up. Incident MIs were adjudicated by clinical experts who used published guidelines and reviewed medical records according to each cohort’s standardized protocol. For fatal MIs, clinical experts reviewed the participant’s medical history, hospital records, interviews with next of kin or proxies, and death certificate or National Death Index data to adjudicate the cause of death.
Covariates
Covariates at or before the first cognitive assessment were chosen based on the potential to confound the association between MI and cognition and were available in all cohorts (Table 1 and Table 2). Covariates included age, sex, educational level, self-reported race and ethnicity (Black, Hispanic [any race], and White), waist circumference, body mass index, fasting glucose level, low-density lipoprotein cholesterol level, cumulative mean systolic BP, use of any antihypertensive medication, current smoking, number of alcoholic drinks per week, and any physical activity. Race included only Black and White individuals because the number of individuals in other racial groups in any 1 cohort was small. Atrial fibrillation was defined in 2 ways: (1) prevalent AF (any AF up to the time of the first cognitive assessment) or (2) any AF (any AF at any time during study follow-up, including after the MI event).
Table 1. Characteristics of Participants Who Did and Did Not Have an Incident MI During Study Follow-up.
| Characteristic | Study participants (N = 30 465)a | P valueb | |
|---|---|---|---|
| With at least 1 incident MI event (n = 1033) | Without an MI event (n = 29 432) | ||
| Age at cohort baseline, median (IQR), y | 59 (49-70) | 54 (46-64) | <.001 |
| Age at first cognitive assessment, median (IQR), y | 64 (55-71) | 60 (52-68) | <.001 |
| Sex | |||
| Female | 478 (46) | 16 484 (56) | <.001 |
| Male | 555 (54) | 12 948 (44) | |
| Race and ethnicity | |||
| Black | 131 (13) | 8649 (23) | <.001 |
| Hispanic (any race) | 37 (4) | 2399 (8) | |
| White | 865 (83.7) | 20 184 (68.6) | |
| Educational level | |||
| Eighth grade or lower | 95 (9) | 2874 (10) | .002 |
| Grades 9-11 or some high school | 126 (12) | 2980 (10) | |
| Completed high school or GED | 326 (32) | 8147 (28) | |
| Some college but no degree | 184 (18) | 5418 (18) | |
| College degree or higher | 302 (29) | 10 013 (34) | |
| Current smoker | 218 (21) | 5118 (18) | .002 |
| Alcoholic drinks, No./wk | |||
| 0 | 542 (52) | 15 687 (53) | .22 |
| 1-6 | 325 (31) | 8534 (29) | |
| 7-13 | 86 (8) | 2873 (10) | |
| ≥14 | 80 (8) | 2338 (8) | |
| Any physical activity | 818 (79) | 23 542 (80) | .53 |
| Waist circumference, median (IQR), cm | 98 (90-105) | 96 (86-105) | <.001 |
| Fasting glucose level, median (IQR), mg/dL | 100 (93-111) | 97 (90-106) | <.001 |
| Cumulative mean systolic blood pressure, median (IQR), mm Hg | 142 (130-155) | 134 (123-147) | <.001 |
| LDL cholesterol level, median (IQR), mg/dL | 135 (112-161) | 123 (100-147) | <.001 |
| Body mass index, median (IQR)c | 27 (25-30) | 27 (24-31) | .38 |
| Antihypertensive medication use | 419 (41) | 10 085 (34) | <.001 |
| History of atrial fibrillation | 24 (2) | 514 (2) | .17 |
| Cognitive scores at first cohort assessment, median (IQR) | |||
| Global cognitive performance | 52 (47-57) | 52 (46-57) | .63 |
| Memory | 50 (46-56) | 51 (46-56) | .39 |
| Executive function | 50 (46-56) | 51 (46-57) | .70 |
| Study cohort | |||
| ARIC | 388 (38) | 11 422 (39) | <.001 |
| CARDIA | 13 (1) | 3646 (12) | |
| CHS | 396 (38) | 4427 (15) | |
| FOS | 172 (16) | 3511 (12) | |
| MESA | 12 (1) | 4102 (14) | |
| NOMAS | 52 (5) | 2324 (8) | |
| Cognitive measurements censored after stroked | 121 (12) | 1077 (4) | <.001 |
Abbreviations: ARIC, Atherosclerosis Risk in Communities Study; CARDIA, Coronary Artery Risk Development in Young Adults Study; CHS, Cardiovascular Health Study; FOS, Framingham Offspring Study; GED, General Educational Development; LDL, low-density lipoprotein; MESA, Multi-Ethnic Study of Atherosclerosis; MI, myocardial infarction; NOMAS, Northern Manhattan Study.
SI conversion: To convert fasting glucose level to millimoles per liter, multiply by 0.0555; LDL cholesterol level to millimoles per liter, multiply by 0.0259.
Data are presented as number (percentage) of participants unless otherwise indicated.
Calculated using t tests for continuous variables and χ2 tests for categorical variables.
Calculated as weight in kilograms divided by height in meters squared.
Censoring refers to individuals who contributed time in a study but were excluded during follow-up at the time of incident stroke.
Table 2. Adjusted Changes in Global Cognition, Executive Function, and Memory Acutely and in the Years Following Incident MI Among All 30 465 Individualsa.
| Measure | Coefficient (95% CI)b | |||||
|---|---|---|---|---|---|---|
| Global cognition (n = 30 465) | Executive function (n = 28 001) | Memory (n = 20 061) | ||||
| Model A | Model B | Model A | Model B | Model A | Model B | |
| Baseline cognitive score | 55.5 (55.3 to 55.8) | 55.5 (55.3 to 55.7) | 53.0 (52.8 to 53.3) | 53.0 (52.8 to 53.3) | 54.7 (54.5 to 55.0) | 54.7 (54.5 to 55.0) |
| Baseline slope without incident MI, per y | −0.26 (−0.27 to −0.25) | −0.25 (−0.27 to −0.24) | −0.29 (−0.30 to −0.27) | −0.31 (−0.32 to −0.30) | −0.31 (−0.32 to −0.30) | −0.28 (−0.30 to −0.27) |
| Difference in baseline cognitive scores for every 10 y older age | −2.19 (−2.29 to −2.09) | −2.19 (−2.29 to −2.09) | −2.84 (−2.95 to −2.74) | −2.84 (−2.95 to −2.73) | −1.72 (−1.85 to −1.58) | −1.72 (−1.85 to −1.58) |
| Change in cognition slope for every 10 y older age | −0.16 (−0.16 to −0.15) | −0.16 (−0.16 to −0.15) | −0.12 (−0.13 to −0.11) | −0.12 (−0.13 to −0.11) | −0.18 (−0.20 to −0.17) | −0.18 (−0.20 to −0.17) |
| Acute change in cognition after incident MI | −0.70 (−0.99 to −0.40) | −0.18 (−0.52 to 0.17) | −0.68 (−0.97 to −0.39) | −0.17 (−0.53 to 0.18) | −0.01 (−0.55 to 0.52) | 0.62 (−0.07 to 1.31) |
| Change in cognition slope after incident MI, per y | NA | −0.15 (−0.21 to −0.10) | NA | −0.14 (−0.20 to −0.08) | NA | −0.13 (−0.22 to −0.04) |
Abbreviations: AF, atrial fibrillation; MI, myocardial infarction; NA, not applicable.
For linear mixed-effects models, model A included a time-varying incident MI variable to estimate the association of incident MI with decline in cognitive function (intercept) at the time of the event (value changed from 0 to 1 at the time of incident MI). Model B included covariates in model A and added a time-after-MI covariate to estimate the association of incident MI with decline in cognitive function (slope) over the years following the event. Covariates in model A included time since the first cognitive assessment, time-varying incident MI, and baseline values (measured before or at the time of the first cognitive assessment) of sex, race and ethnicity, age, cohort study, years of education, current smoking, number of alcoholic drinks per week, body mass index, waist circumference, physical activity, cumulative systolic mean blood pressure, antihypertension medication use, fasting glucose level, low-density lipoprotein cholesterol level, history of AF (defined as history of AF at cohort baseline or incident AF before incident MI), age × follow-up time, race and ethnicity × follow-up time, sex × follow-up time. Covariates in model B included covariates in model A plus time after MI.
All cognitive measures were set to a mean (SD) T score metric (50 [10]) at a participant’s first cognitive assessment; a 1-point difference represented a 0.1-SD difference in the distribution of the specified cognitive domain across the 6 cohorts. Higher cognitive scores indicate better performance. Cognitive observations were censored after stroke.
Statistical Analysis
Characteristics between individuals with and without MI were compared using a t test or a χ2 test as appropriate. We used 2 sequential linear mixed-effects models to estimate the association of MI with cognition trajectories, adjusting for potential confounders. Follow-up time was defined as years since the first measurement of each cognitive outcome. For all 3 cognitive domains, we found no evidence of nonlinear associations of time with cognition and therefore assumed that the cognitive score for individuals with no MI followed a linear time function. Model A included a time-varying incident MI variable (changed from 0 to 1 on the date of the first incident MI) to estimate the association of incident MI with cognitive function acutely or the acute change as estimated from the model (not observed) after the first MI event (mixed linear-effects model intercept).
In model B, we added a time-after-MI covariate to model A to assess whether incident MI was associated with a faster rate of cognitive decline in the years following the MI event. The coefficient associated with this covariate quantified the change in slope and represented the association of incident MI with the rate of decline in cognitive function. Therefore, for each individual who had an incident MI, the time-after-MI variable would be 0 before their first MI event time and 1 after the MI event and would remain at 0 for all other individuals. In both models A and B, we included the interaction terms race and ethnicity × time and sex × time to allow different rates of decline before the MI event for different racial and ethnic groups and by sex.
To explore differences by race and ethnicity and sex in the association of MI with cognitive decline, we added interaction terms for sex and race and ethnicity and the incident MI and the time-after-MI variables to model B. To assess whether the association differed among those with more than 1 MI event, we added a second MI time-varying variable (changed from 0 to 1 on the date of the second MI) and a time-from-the-second-MI variable (representing slope) to model B. This was performed for global cognition and executive function but not memory due to insufficient measurements to reliably calculate these estimates. Finally, to assess whether these associations persisted among individuals without AF, we repeated the analysis with model B excluding first those with prevalent AF and then excluding those with any AF.
Data were analyzed from July 2021 to January 2022. Study indicators were included in all models to account for study heterogeneity. Statistical analyses were performed using SAS, version 9.4 (SAS Institute Inc). Two-sided P < .05 was considered statistically significant.
Results
Demographics and Cognitive Testing
The total study cohort included 40 016 adults from 6 studies. After excluding 8139 individuals who did not meet the inclusion and exclusion criteria and 1412 with missing covariate data, the study sample included 30 465 individuals (mean [SD] age, 64 [10] years) without MI, stroke, or dementia at the time of the first cognitive assessment (eFigure 1 in Supplement 1). Of these individuals, 29% were Black, 8% were Hispanic, 69% were White, 56% were female, and 44% were male; 1033 individuals had 1 or more incident MI event, and 137 had 2 MI events (Table 1).
The median follow-up time was 6.4 years (IQR, 4.9-19.7 years). The median time between the last pre-MI cognitive assessment and incident MI was 1.88 years (IQR, 0.65-4.22 years) for global cognition, 2.17 years (IQR, 0.69-4.52 years) for executive function, and 3.67 years (IQR, 2.13-6.15 years) for memory. The median time between the first pre-MI cognitive assessment and incident MI event for each cognitive outcome was 7.10 years (IQR, 3.43-13.62 years) for global cognition, 5.71 years (IQR, 2.87-11.98 years) for executive function, and 8.31 years (IQR, 3.84-16.64 years) for memory. The median time between incident MI and the first post-MI assessment was 1.85 years (IQR, 0.59-4.45 years) for global cognition, 2.21 years (IQR, 0.62-4.93 years) for executive function, and 3.71 years (IQR, 1.85-6.49 years) for memory, with 398 individuals having their first cognitive test after MI within 1 year. The mean (SD) number of cognitive assessments after incident MI for global cognition, executive function, and memory was 2.63 (2.23), 2.06 (1.68), and 1.41 (0.68), respectively. (eTable 1 in Supplement 1 provides the number of cognitive tests after the first and second MI for each cognitive domain.)
Association Between Incident MI and Cognitive Decline
Overall, incident MI was associated with an acute decrease in global cognition (−0.70 points; 95% CI, −0.99 to −0.40 points) and executive function (−0.68 points; 95% CI, −0.97 to −0.39 points) but not memory (−0.01 points; 95% CI, −0.55 to 0.52 points) acutely after the MI (model A). However, after adjusting for the decline in cognition in the years following the MI event, incident MI was not associated with acute decreases in global cognition (−0.18 points; 95% CI, −0.52 to 0.17 points) and executive function (−0.17 points; 95% CI, −0.53 to 0.18 points) or with memory (0.62 points; 95% CI, −0.07 to 1.31 points) (model B) (Table 2).
Over long-term follow-up (model B), individuals with 1 or more MI event demonstrated faster declines over the years after MI in all 3 measures of cognition: global cognition (−0.15 points per year; 95% CI, −0.21 to −0.10 points per year), memory (−0.13 points per year; 95% CI, −0.22 to −0.04 points per year), and executive function (−0.14 points per year; 95% CI, −0.20 to −0.08 points per year) (Table 2). Figure 1 presents the estimated cognitive scores for an exemplar patient based on estimated parameters of model B.
Figure 1. Estimated Values of Cognition for the Association Between Incident MI and Change in Global Cognition, Executive Function, and Memory Acutely and in the Years Following the Myocardial Infarction (MI) Event Among 30 465 Participants.
Global cognition measures global cognitive performance. All cognitive measures were set to a mean (SD) T score metric (50 [10]) at a participant’s first cognitive assessment; a 1-point difference represented a 0.1-SD difference in the distribution of the specified cognitive domain across the 6 cohorts. Higher cognitive scores indicate better performance. Cognitive observations were censored after stroke. Graphs represent a patient who is aged 70 years with a high school education and is a nonsmoker with no alcohol use, no history of atrial fibrillation, no physical activity, no antihypertensive medication use, a body mass index of 27 (calculated as weight in kilograms divided by height in meters squared), a waist circumference of 96 cm, a fasting blood glucose level of 97 mg/dL (to convert to millimoles per liter, multiply by 0.0555), a low-density lipoprotein cholesterol level of 123 mg/dL (to convert to millimoles per liter, multiply by 0.0259), and a cumulative systolic blood pressure of 135 mm Hg.
Association Between More Than 1 MI Event and Cognitive Decline
After adjusting for a second MI event during study follow-up, there was still no difference in cognition (global cognition or executive function) acutely after the first MI event. There remained a significant difference in the long-term change after the first MI in global cognition (−0.13 points per year; 95% CI, −0.19 to −0.08 points per year) and executive function (−0.13 points per year; 95% CI, −0.19 to −0.07 points per year). Among individuals with a second MI event, there was no acute decrease in global cognition after the second MI but there was an acute decrease in executive function (−1.55 points per year; 95% CI, −2.63 to −0.48 points per year), with no significant difference in the decline in either cognitive slope after the second MI (Table 3).
Table 3. Adjusted Change in Global Cognition and Executive Function Acutely After and in the Years Following Incident MI Accounting for a Potential Second MI Event Among 30 465 Individualsa.
| Measure | Coefficient (95% CI)b | |
|---|---|---|
| Global cognition (n = 30 465) | Executive function (n = 28 001) | |
| Baseline cognitive score | 55.49 (55.25 to 55.74) | 54.71 (54.46 to 54.97) |
| Baseline slope without incident MI, per y | −0.25 (−0.27 to −0.24) | −0.31 (−0.32 to −0.30) |
| Difference in baseline cognitive scores for every 10 y older age | −2.19 (−2.29 to −2.09) | −2.84 (−2.95 to −2.74) |
| Acute change in cognition after first incident MI | −0.20 (−0.55 to 0.16) | −0.08 (−0.45 to 0.28) |
| Change in cognition slope after first incident MI, per y | −0.13 (−0.19 to −0.08) | −0.13 (−0.19 to −0.07) |
| Acute change in cognition after second incident MI | −0.05 (−1.06 to 0.95) | −1.55 (−2.63 to −0.48) |
| Change in cognition slope after second incident MI, per y | −0.15 (−0.35 to 0.05) | 0.02 (−0.20 to 0.23) |
Abbreviations: AF, atrial fibrillation; MI, myocardial infarction.
Covariates in the model included time since the first cognitive assessment, time-varying incident MI, and baseline values (measured before or at the time of the first cognitive assessment) of sex, race and ethnicity, age, cohort study, years of education, current smoking, number of alcoholic drinks per week, body mass index, waist circumference, physical activity, cumulative systolic mean blood pressure, antihypertension medication use, fasting glucose level, low-density lipoprotein cholesterol level, history of AF (defined as history of AF at cohort baseline or incident AF before incident MI), age × follow-up time, race and ethnicity × follow-up time, sex × follow-up time, time-varying incident MI × follow-up time, time-varying second MI, and time-varying second MI × follow-up time.
Global cognition measures global cognitive performance. All cognitive measures were set to a mean (SD) T score metric (50 [10]) at a participant’s first cognitive assessment; a 1-point difference represented a 0.1-SD difference in the distribution of the specified cognitive domain across the 6 cohorts. Higher cognitive scores indicate better performance. Cognitive observations were censored after stroke.
Association Between Incident MI and Cognitive Decline by Race, Ethnicity, and Sex and Sensitivity Analyses Excluding Individuals With AF
Black individuals had lower baseline cognitive scores in global cognition (−5.90 points; 95% CI, −6.09 to −5.71 points), executive function (−5.72 points; 95% CI, −5.91 to −5.52 points), and memory (−3.00 points; 95% CI, −3.22 to −2.78 points) compared with White individuals. Results suggested that the association between incident MI and cognition differed between Black and White individuals when holding sex constant (eTable 2 in Supplement 1). Global cognition did not decrease acutely after MI in White individuals, but in Black individuals (Figure 2 and eTable 2 in Supplement 1), the difference in the acute change was significant (−1.27 points; 95% CI, −2.37 to −0.17 points; race × MI interaction term, P = .02).
Figure 2. Estimated Values of Global Cognition for Differences by Race or by Sex in the Association Between Incident Myocardial Infarction (MI) and Change in Global Cognition Among 30 465 Participants.
A, Sex was held constant (P = .02 for differences between race in acute decline after MI; P = .02 for difference between race in long-term cognitive slope change after MI). B, P = .02 for difference between race in acute decline after MI; P = .02 for differences between race in long-term cognitive slope change after MI). C, Race was held constant (P = .27 for differences in sex in acute decline after MI; P = .04 for differences in sex in long-term cognitive slope change after MI). D, P = .27 for difference in sex in acute decline after MI; P = .04 for differences in sex in long-term cognitive slope change after MI). Global cognition measures global cognitive performance. All cognitive measures were set to a mean (SD) T score metric (50 [10]) at a participant’s first cognitive assessment; a 1-point difference represented a 0.1-SD difference in the distribution of the specified cognitive domain across the 6 cohorts. Higher cognitive scores indicate better performance. Cognitive observations were censored after stroke. Graphs represent a patient who is aged 70 years with a high school education and is a nonsmoker with no alcohol use, no history of atrial fibrillation, no physical activity, no antihypertensive medication use, a body mass index of 27 (calculated as weight in kilograms divided by height in meters squared), a waist circumference of 96 cm, a fasting blood glucose level of 97 mg/dL (to convert to millimoles per liter, multiply by 0.0555), a low-density lipoprotein cholesterol level of 123 mg/dL (to convert to millimoles per liter, multiply by 0.0259), and a cumulative systolic blood pressure of 135 mm Hg.
Compared with pre-MI cognitive trajectories, there was a significantly faster post-MI, long-term decline in global cognition among White individuals. In contrast, there was no change among Black individuals (difference in change in slope, 0.22 points per year; 95% CI, 0.40 to 0.04 points per year; race × post-MI slope interaction term, P = .02) (Figure 2). There was no evidence of effect modification by Black race for executive function (eFigure 2 in Supplement 1) or memory either acutely after or in the years following MI. We found no difference in the association between incident MI and any cognitive domain either acutely or in the rate of long-term decline after MI between Hispanic and White individuals (eTable 2 in Supplement 1).
Females had higher baseline cognitive scores in global cognition (1.88 points; 95% CI, 1.73-2.02 points), executive function (1.73 points; 95% CI, 1.58-1.89 points), and memory (1.84 points; 95% CI, 1.66-2.02 points) compared with males. There was no evidence that the association between incident MI and an acute decrease in cognition differed by sex in any domain (eTable 2 in Supplement 1 and Figure 2). There was evidence of modification by sex in the change in cognitive slope over the years after MI, with females having less of an acceleration in decline in global cognition (difference in change in slope, 0.12 points per year 95% CI, 0.01-0.23 points per year; sex × post-MI slope interaction term, P = .04) than males but more of an acceleration of decline in executive function (difference in slope change, −0.15 points per year; 95% CI, −0.26 to −0.03 points per year; sex × post-MI slope interaction term, P = .01) than males (eFigure 2 in Supplement 1). There was no difference by sex in the association between incident MI and memory (eTable 2 in Supplement 1). Results were similar in analyses excluding those with prevalent AF (n = 536) and after excluding those with any AF ever, including during study follow-up (n = 2919) (eTables 3 and 4 in Supplement 1).
Discussion
Among the 30 465 community-dwelling individuals in this cohort study, incident MI was associated with long-term decline in global cognition, executive function, and memory after MI. In the overall population, we did not find a significant acute decrease in cognition at the time of the MI after accounting for individuals’ pre-MI and post-MI cognitive changes over time. However, we found that incident MI was associated with accelerated cognitive decline over the years after MI.
We added to prior work investigating disability and dementia after MI13,14,15,16,17 by providing evidence that a decrease in cognition after incident MI was not sudden but happened more slowly over the years following the event and may have differed by race. Myocardial infarction has previously been associated with dementia, but that finding was attributable to post-MI stroke.6 We demonstrated an association independent of clinical stroke and AF and not explained by a second MI event. A meta-analysis of 24 studies suggested that participants with coronary heart disease, including MI, have an increased risk of cognitive impairment, consistent with our results.8 Our findings strengthened those by showing that acute MI was associated with declines in executive function and memory, which results in impairment in decision-making and recall. We also quantified the magnitude of post-MI cognitive change; the decline in global cognition after incident MI was equivalent to 6 to 13 years of cognitive aging, representing an important public health problem.
Incident MI may have an association with adults’ cognitive trajectories through several mechanisms. An MI event could exacerbate preexisting cerebral sequelae of long-standing cerebrovascular disease, such as white matter disease from hypertension. It may initiate a process of vascular dementia through systemic inflammation leading to oxidative stress, chronic hypoperfusion secondary to impaired left ventricular ejection fraction, development of AF, or subclinical ischemic stroke.18,19 Notably, our results remained robust when removing individuals with AF and censoring for stroke.
The lack of a significant cognitive change in the short term in the overall study population suggests that cognitive change is not secondary to acute illness or delirium but rather a sustained impairment in cognition. This finding would support the importance of shared vascular risk factors (MI and vascular dementia) contributing over time to this trajectory.15,20,21 We carefully accounted for known vascular risk factors, but subclinical disease or a predisease state may play a role.
We found that, compared with White individuals, Black individuals had lower cognitive scores at baseline and an acute decrease in global cognition after incident MI but less of a decline in global cognition over the years after MI. The results emphasize the need for effective interventions to reduce cognitive racial disparities22 and to improve control of vascular risk factors, such as hypertension, which are associated with cognition, among Black individuals and may have led to the lower baseline scores in Black individuals than in White individuals in the present study. Of interest, in the acute period after MI, Black individuals had a significant decrease in global cognition. They may have lower cognitive reserve due to lower quality education23 or may experience decreased access to care, support systems, or protective medications in the short term after MI compared with White individuals. Black individuals may have higher rates of preexisting heart failure or more severe MI events compared with White individuals due to genetic variation.24
We also found that sex modified the association between incident MI and the rate of cognitive decline over the years after MI. Females had less of a decline in global cognition than males but more of a decline in executive function than males. There has been other evidence of differential decline across cognitive domains by sex,25 but our results may also suggest a nonbiological reason that may be related to the cognitive measures themselves, and therefore these findings merit confirmation.
Limitations
Our study has limitations. Measures of post-MI cognition may have occurred months after the event, which could have led to underestimation of the acute change in cognition immediately after the event. Data on incident dementia were unavailable in all cohorts. Additional socioeconomic factors (eg, occupation) and clinical factors (depressive symptoms) that may influence the association between MI and cognition were either unavailable for all cohorts or occurred after the first cognitive assessment. We had a small number of Hispanic individuals in our analysis and may not have had adequate power to detect ethnic differences. We did not consider cardiac procedures, such as a coronary artery bypass graft or MI severity or complications in our analysis because not all cohorts had this information. However, prior research has suggested that long-term cognitive decline in patients with MI who undergo a coronary artery bypass graft is similar to that among those who do not have the procedure.26,27,28 We did not account for multiple comparisons, so the findings of effect modification by race and sex should be interpreted cautiously. The relatively limited number of post-MI cognitive assessments might have decreased our ability to detect an association between MI and cognitive changes, but to our knowledge, our study represents 1 of the largest studies to assess this association.
Conclusions
In this cohort study, incident MI was not associated with an acute decrease in global cognition, memory, and executive function at the time of the event but was associated with a faster and persistent decrease in cognition in the years following the MI event. Our results may have important public health implications. Discussion of the potential cognitive ramifications of MI should be considered as a potential motivator when counseling patients at risk for MI. Additionally, individuals who have experienced an MI should be followed up for accelerated cognitive decline in the years after MI. The findings also suggest that prevention of MI may be a strategy to preserve brain health in older adults.
eTable 1. Number of Cognitive Tests That Was Obtained for Participants After Their First or Second Myocardial Infarction Across Cohorts Results of Quality Assessment per Study
eTable 2. Adjusted Change in Global Cognition, Executive Function, and Memory Acutely and the Years Following Incident MI With Interaction by Race and Ethnicity and Sex Among All Individuals (N = 30 465)
eTable 3. Adjusted Change in Global Cognition, Executive Function, and Memory in the Short-term and the Years Following (Model B) Incident MI Among Participants Without Prevalent Atrial Fibrillation (N = 29 929)
eTable 4. Adjusted Change in Global Cognition, Executive Function, and Memory in the Short-term and the Years Following Incident MI Among Participants Without Atrial Fibrillation Either at Cohort Baseline or Incident Atrial Fibrillation During Any Time on Study, Including After MI Event (N = 27 546)
eFigure 1. Participant Flow Diagram
eFigure 2. Predicted Values of Executive Function Demonstrating Differences by Race or by Sex in the Association Between Incident MI and Change in Executive Function (N = 30 465)
Data Sharing Statement
References
- 1.Wimo A, Guerchet M, Ali GC, et al. The worldwide costs of dementia 2015 and comparisons with 2010. Alzheimers Dement. 2017;13(1):1-7. doi: 10.1016/j.jalz.2016.07.150 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Snyder HM, Corriveau RA, Craft S, et al. Vascular contributions to cognitive impairment and dementia including Alzheimer’s disease. Alzheimers Dement. 2015;11(6):710-717. doi: 10.1016/j.jalz.2014.10.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Reed GW, Menon V. Reducing the incidence and mortality from myocardial infarction. Lancet Public Health. 2022;7(3):e202-e203. doi: 10.1016/S2468-2667(22)00027-5 [DOI] [PubMed] [Google Scholar]
- 4.Reed GW, Rossi JE, Cannon CP. Acute myocardial infarction. Lancet. 2017;389(10065):197-210. doi: 10.1016/S0140-6736(16)30677-8 [DOI] [PubMed] [Google Scholar]
- 5.Wolters FJ, Segufa RA, Darweesh SKL, et al. Coronary heart disease, heart failure, and the risk of dementia: a systematic review and meta-analysis. Alzheimers Dement. 2018;14(11):1493-1504. doi: 10.1016/j.jalz.2018.01.007 [DOI] [PubMed] [Google Scholar]
- 6.Sundbøll J, Horváth-Puhó E, Adelborg K, et al. Higher risk of vascular dementia in myocardial infarction survivors. Circulation. 2018;137(6):567-577. doi: 10.1161/CIRCULATIONAHA.117.029127 [DOI] [PubMed] [Google Scholar]
- 7.Xie W, Zheng F, Yan L, Zhong B. Cognitive decline before and after incident coronary events. J Am Coll Cardiol. 2019;73(24):3041-3050. doi: 10.1016/j.jacc.2019.04.019 [DOI] [PubMed] [Google Scholar]
- 8.Deckers K, Schievink SHJ, Rodriquez MMF, et al. Coronary heart disease and risk for cognitive impairment or dementia: systematic review and meta-analysis. PLoS One. 2017;12(9):e0184244. doi: 10.1371/journal.pone.0184244 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Koton S, Pike JR, Johansen M, et al. Association of ischemic stroke incidence, severity, and recurrence with dementia in the Atherosclerosis Risk in Communities Cohort Study. JAMA Neurol. 2022;79(3):271-280. doi: 10.1001/jamaneurol.2021.5080 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Schmitt J, Duray G, Gersh BJ, Hohnloser SH. Atrial fibrillation in acute myocardial infarction: a systematic review of the incidence, clinical features and prognostic implications. Eur Heart J. 2009;30(9):1038-1045. doi: 10.1093/eurheartj/ehn579 [DOI] [PubMed] [Google Scholar]
- 11.Levine DA, Gross AL, Briceño EM, et al. Association between blood pressure and later-life cognition among Black and White individuals. JAMA Neurol. 2020;77(7):810-819. doi: 10.1001/jamaneurol.2020.0568 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Briceño EM, Gross AL, Giordani BJ, et al. Pre-statistical considerations for harmonization of cognitive instruments: harmonization of ARIC, CARDIA, CHS, FHS, MESA, and NOMAS. J Alzheimers Dis. 2021;83(4):1803-1813. doi: 10.3233/JAD-210459 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Dhamoon MS, Longstreth WT Jr, Bartz TM, Kaplan RC, Elkind MSV. Disability trajectories before and after stroke and myocardial infarction: the Cardiovascular Health Study. JAMA Neurol. 2017;74(12):1439-1445. doi: 10.1001/jamaneurol.2017.2802 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Levine DA, Davydow DS, Hough CL, Langa KM, Rogers MAM, Iwashyna TJ. Functional disability and cognitive impairment after hospitalization for myocardial infarction and stroke. Circ Cardiovasc Qual Outcomes. 2014;7(6):863-871. doi: 10.1161/HCQ.0000000000000008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Cermakova P, Eriksdotter M, Lund LH, Winblad B, Religa P, Religa D. Heart failure and Alzheimer’s disease. J Intern Med. 2015;277(4):406-425. doi: 10.1111/joim.12287 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Witt LS, Rotter J, Stearns SC, et al. Heart failure and cognitive impairment in the Atherosclerosis Risk in Communities (ARIC) Study. J Gen Intern Med. 2018;33(10):1721-1728. doi: 10.1007/s11606-018-4556-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Adelborg K, Horváth-Puhó E, Ording A, Pedersen L, Sørensen HT, Henderson VW. Heart failure and risk of dementia: a Danish nationwide population-based cohort study. Eur J Heart Fail. 2017;19(2):253-260. doi: 10.1002/ejhf.631 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Aldrugh S, Sardana M, Henninger N, Saczynski JS, McManus DD. Atrial fibrillation, cognition and dementia: a review. J Cardiovasc Electrophysiol. 2017;28(8):958-965. doi: 10.1111/jce.13261 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Bailey MJ, Soliman EZ, McClure LA, et al. Relation of Atrial Fibrillation to Cognitive Decline (from the REasons for Geographic and Racial Differences in Stroke [REGARDS] Study). Am J Cardiol. 2021;148:60-68. doi: 10.1016/j.amjcard.2021.02.036 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Ferrer I. Cognitive impairment of vascular origin: neuropathology of cognitive impairment of vascular origin. J Neurol Sci. 2010;299(1-2):139-149. doi: 10.1016/j.jns.2010.08.039 [DOI] [PubMed] [Google Scholar]
- 21.Frankish H, Horton R. Prevention and management of dementia: a priority for public health. Lancet. 2017;390(10113):2614-2615. doi: 10.1016/S0140-6736(17)31756-7 [DOI] [PubMed] [Google Scholar]
- 22.Kornblith E, Bahorik A, Boscardin WJ, Xia F, Barnes DE, Yaffe K. Association of race and ethnicity with incidence of dementia among older adults. JAMA. 2022;327(15):1488-1495. doi: 10.1001/jama.2022.3550 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Avila JF, Rentería MA, Jones RN, et al. Education differentially contributes to cognitive reserve across racial/ethnic groups. Alzheimers Dement. 2021;17(1):70-80. doi: 10.1002/alz.12176 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Calhoun DA, Booth JN III, Oparil S, et al. Refractory hypertension: determination of prevalence, risk factors, and comorbidities in a large, population-based cohort. Hypertension. 2014;63(3):451-458. doi: 10.1161/HYPERTENSIONAHA.113.02026 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.McCarrey AC, An Y, Kitner-Triolo MH, Ferrucci L, Resnick SM. Sex differences in cognitive trajectories in clinically normal older adults. Psychol Aging. 2016;31(2):166-175. doi: 10.1037/pag0000070 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Selnes OA, Royall RM, Grega MA, Borowicz LM Jr, Quaskey S, McKhann GM. Cognitive changes 5 years after coronary artery bypass grafting: is there evidence of late decline? Arch Neurol. 2001;58(4):598-604. doi: 10.1001/archneur.58.4.598 [DOI] [PubMed] [Google Scholar]
- 27.Selnes OA, Grega MA, Borowicz LM Jr, et al. Cognitive outcomes three years after coronary artery bypass surgery: a comparison of on-pump coronary artery bypass graft surgery and nonsurgical controls. Ann Thorac Surg. 2005;79(4):1201-1209. doi: 10.1016/j.athoracsur.2004.10.011 [DOI] [PubMed] [Google Scholar]
- 28.Selnes OA, Grega MA, Borowicz LM Jr, Barry S, Zeger S, McKhann GM. Self-reported memory symptoms with coronary artery disease: a prospective study of CABG patients and nonsurgical controls. Cogn Behav Neurol. 2004;17(3):148-156. doi: 10.1097/01.wnn.0000126982.53726.67 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eTable 1. Number of Cognitive Tests That Was Obtained for Participants After Their First or Second Myocardial Infarction Across Cohorts Results of Quality Assessment per Study
eTable 2. Adjusted Change in Global Cognition, Executive Function, and Memory Acutely and the Years Following Incident MI With Interaction by Race and Ethnicity and Sex Among All Individuals (N = 30 465)
eTable 3. Adjusted Change in Global Cognition, Executive Function, and Memory in the Short-term and the Years Following (Model B) Incident MI Among Participants Without Prevalent Atrial Fibrillation (N = 29 929)
eTable 4. Adjusted Change in Global Cognition, Executive Function, and Memory in the Short-term and the Years Following Incident MI Among Participants Without Atrial Fibrillation Either at Cohort Baseline or Incident Atrial Fibrillation During Any Time on Study, Including After MI Event (N = 27 546)
eFigure 1. Participant Flow Diagram
eFigure 2. Predicted Values of Executive Function Demonstrating Differences by Race or by Sex in the Association Between Incident MI and Change in Executive Function (N = 30 465)
Data Sharing Statement