Trajectory of Cognitive Decline After Incident Heart Failure Hospitalization: Findings From the REGARDS Study

Abstract

Background: Cognitive impairment is common among adults with heart failure (HF) and associated with poor outcomes. However, less is known about the trajectory of cognitive decline after a first HF hospitalization. We examined the rate of cognitive decline among adults with incident HF hospitalization compared with those without HF hospitalization.

Methods and results: The REGARDS (Reasons for Geographic and Racial Differences in Stroke) study is a prospective longitudinal study of 23 894 participants aged ≥45 years free of HF at baseline. HF hospitalization was expert adjudicated. Changes in global cognitive function (primary outcome) were assessed with the Six-Item Screener (range, 0-6). Secondary outcomes included change in Word List Learning (range, 0-30), Word List Delayed Recall (WLD; range, 0-10), and Animal Fluency Test (range, 0+). Segmented linear mixed-effects regression models were used. Over 5 years, mean scores across all 4 cognitive tests declined for all participants regardless of HF status. Those with incident HF hospitalization experienced faster declines in the Six-Item Screener versus those who were HF free (difference, -0.031 [95% CI, -0.047 to -0.016]; P<0.001), a finding that persisted in fully adjusted models. Those with incident HF hospitalization did not experience faster declines in Word List Learning, Word List Delayed Recall, or Animal Fluency Test scores compared with those without HF hospitalization. Participants with hospitalization for HF with preserved, compared with reduced, ejection fraction had faster decline in Animal Fluency Test.

Conclusions: Global cognitive decline occurred faster among adults with incident HF hospitalization compared with those who remained free of HF hospitalization. This pattern was not seen for the other cognitive domains.

Keywords: cognition; cognitive impairment; cohort study; heart failure.

Figure 1. Minimally adjusted graphs comparing the cognitive trajectory of REGARDS participants after incident HF hospitalization event compared with those with no incident HF hospitalization event.

Minimally adjusted^# graphs comparing the cognitive trajectory for SIS (A), WLL (B), WLD (C), and AFT (D). Among REGARDS participants after incident HF hospitalization event (black line) compared with those with no incident HF hospitalization event (gray line). AFT indicates Animal Fluency Test; DIF, difference; HF, heart failure; REGARDS, Reasons for Geographic and Racial Differences in Strokes; SIS, Six‐Item Screener; WLD, Word List Delayed Recall; WLL, Word List Learning. * SIS scores range from 0 to 6; mean (SD) number of assessments over follow‐up for no HF group was 8 (4) and for HF group was 9 (3). ^†WLL scores range from 0 to 30; mean (SD) number of assessments over follow‐up for no HF group was 3 (1) and for HF group was 3 (1). ^‡WLD scores range from 0 to 10; mean (SD) number of assessments over follow‐up for no HF group was 3 (1) and for HF group was 3 (1). ^§ AFT) scores ranged from 0 to greater; mean (SD) number of assessments over follow‐up for no HF group was 3 (2) and for HF group was 4 (2). ^||Models used age as the time scale and include random intercepts for each participant and random slopes for age at the time of cognitive assessment. ^#Time 0 refers to the time of incident HF hospitalization event for those in the HF group. **Bayesian information criterion values for models: SIS, 414466.7; WLL, 282008.1; WLD, 198225.7; AFT, 308 231.

Figure 2. Fully adjusted graphs comparing the cognitive trajectory of REGARDS participants after incident HF hospitalization event compared with those with no incident HF hospitalization event.

Fully adjusted^# graphs comparing the cognitive trajectory for SIS (A), WLL (B), WLD (C), and AFT (D). Among REGARDS participants after incident HF hospitalization event (black line) compared with those with no incident HF hospitalization event (gray line). AFT indicates Animal Fluency Test; DIF, difference; HF, heart failure; REGARDS, Reasons for Geographic and Racial Differences in Strokes; SIS, Six‐Item Screener; WLD, Word List Delayed Recall; and WLL, Word List Learning. *SIS scores range from 0 to 6; mean (SD) number of assessments over follow‐up for no HF group was 8 (4) and for HF group was 9 (3). ^†WLL scores range from 0 to 30; mean (SD) number of assessments over follow‐up for no HF group was 3 (1) and for HF group was 3 (1). ^‡WLD scores range from 0 to 10; mean (SD) number of assessments over follow‐up for no HF group was 3 (1) and for HF group was 3 (1). ^§AFT scores ranged from 0 to greater; mean (SD) number of assessments over follow‐up for no HF group was 3 (2) and for HF group was 4 (2). ^||Time 0 refers to the time of incident HF hospitalization event for those in the HF group. ^#Adjusted for sociodemographics (sex, race, education, income, region of residence, social network, social isolation), comorbidities (hypertension, history of heart disease, dyslipidemia, diabetes, body mass index, albumin–creatinine ratio, atrial fibrillation, depression, physical component summary score, mental component summary score) and health behaviors (smoking, alcohol use, and exercise). **Bayesian information criterion values for models: SIS, 290853.4; WLL, 202568.3; WLD, 141749.4; AFT, 221682.4.

Figure 3. Cognitive trajectory of REGARDS participants after incident HF hospitalization, by HF subtype.

Cognitive trajectory for SIS (A), WLL (B), WLD (C), and AFT (D). Among REGARDS participants after incident HF hospitalization, by HF subtype (HFpEF=black line; HFrEF=gray line). ^||AFT indicates Animal Fluency Test; DIF, difference; HF, heart failure; REGARDS, Reasons for Geographic and Racial Differences in Strokes; SIS, Six‐Item Screener; WLD, Word List Delayed Recall; and WLL, Word List Learning. *SIS scores range from 0 to 6; mean (SD) number of assessments over follow‐up for no HF group was 8 (4) and for HF group was 9 (3). ^†WLL scores range from 0 to 30; mean (SD) number of assessments over follow‐up for no HF group was 3 (1) and for HF group was 3 (1). ^‡WLD scores range from 0 to 10; mean (SD) number of assessments over follow‐up for no HF group was 3 (1) and for HF group was 3 (1). ^§AFT scores ranged from 0 to greater; mean (SD) number of assessments over follow‐up for no HF group was 3 (2) and for HF group was 4 (2). ^||Out of incident HF sample (n=584). ^#Difference in 1‐year change in AFT after HF with preserved ejection fraction versus HF with reduced ejection fraction is statistically significant. Cochrane's Q test used to determine significance. **P value remains significant after Benjamini–Hochberg correction for type 1 error with multiple tests. ^¥Time 0 refers to the time of incident HF hospitalization event for those in the HF group.