Abstract
Background
Vascular risk factors have been associated with differences in cognitive performance in epidemiological studies, but evidence in patients with coronary heart disease is more limited.
Methods and Results
The Montreal Cognitive Assessment score obtained 3.2±0.37 years after randomization to darapladib, a reversible inhibitor of lipoprotein phospholipase A2 or placebo was evaluated for 10 634 patients with coronary heart disease from 38 countries in the STABILITY (Stabilization of Atherosclerotic Plaque by Initiation of Darapladib Therapy) trial. The Montreal Cognitive Assessment scores for darapladib and placebo groups were similar (mean±SD, 25.3±3.84 versus 25.4±3.73, respectively; P=0.27) and the adjusted odds ratio (OR) for mild cognitive impairment (Montreal Cognitive Assessment score <26) was 1.00 (95% CI, 0.93–1.09). Mild cognitive impairment was more likely with increasing age (OR, 1.33 [1.27–1.41], +5 years after 65). For other baseline clinical characteristics, the strongest independent predictors of cognitive impairment were education (≤8 years versus college/university, OR, 2.95 [2.60–3.35]; >8 years/trade school versus college/university, OR, 1.38 [1.25–1.52] and geographic grouping). Cardiovascular risk factors independently associated with cognitive impairment were history of stroke (OR, 1.43 [1.20–1.71]); <2.5 hours of moderate or vigorous intensity exercise/week (OR, 1.19 [1.04–1.37]); high‐density lipoprotein cholesterol <1.16 mmol/L (OR, 1.19 [1.04–1.37]); diabetes mellitus requiring treatment (OR, yes versus no: 1.15 [1.05–1.26]); and history of hypertension (OR, 1.12 [1.02–1.23]).
Conclusions
In patients with stable coronary heart disease, cognitive performance was associated with modifiable cardiovascular risk factors, educational level, and global region, but was not influenced by darapladib.
Clinical Trial Registration
URL: http://www.clinicaltrials.gov. Unique identifier: NCT00799903.
Keywords: cognitive impairment, coronary heart disease, Montreal Cognitive Assessment, risk factor
Subject Categories: Cardiovascular Disease, Aging, Epidemiology, Lifestyle, Mental Health
Clinical Perspective
What Is New?
History of stroke, low high‐density lipoprotein cholesterol, low physical activity, diabetes mellitus, and history of hypertension were cardiovascular risk factors independently associated with Montreal Cognitive Assessment scores indicative of cognitive impairment in 10 634 patients with stable coronary heart disease from 38 countries.
The strongest independent predictors of cognitive impairment were educational level, history of stroke, and region of residence.
Inhibition of lipoprotein‐associated phospholipase A2 activity with darapladib was not associated with cognitive impairment.
What Are the Clinical Implications?
In patients with stable coronary heart disease, potentially modifiable risk factors may influence cognitive performance.
Introduction
Dementia is a major cause of disability in aging populations.1, 2 The most common causes are Alzheimer disease and vascular diseases.3 Dementia is usually preceded by milder cognitive impairment that may progress gradually. Cardiovascular risk factors during middle age have been associated with a higher risk of cognitive impairment in later life,4, 5, 6, 7 possibly by increasing the risk of vascular or mixed dementia. Patients with coronary heart disease (CHD) could be at greater risk for vascular dementia because of established atherosclerotic disease and a greater burden of vascular risk factors.8 It is also possible that treatments that lower vascular risk could reduce the risk of cognitive impairment.
Lipoprotein‐associated phospholipase A2 (Lp‐PLA2), which increases production of proinflammatory and proapoptotic mediators, is one of several inflammatory biomarkers that have been associated with the risk of dementia or impaired cognitive function. Lp‐PLA2 activity has been reported to be higher in patients with Alzheimer disease compared with controls,9 and to be associated with dementia in epidemiologic studies.10, 11, 12, 13 However, it is not known whether these associations are causal.
The current study was undertaken as part of the STABILITY (Stabilization of Atherosclerotic Plaque by Initiation of Darapladib Therapy) trial. In this large global trial of patients with stable CHD, darapladib decreased plasma Lp‐PLA2 activity by ≈65% but did not decrease major cardiovascular events.14, 15 The aims of this evaluation were first to determine whether darapladib treatment had any effect on cognitive function, and second to evaluate whether other potentially modifiable risk factors are associated with cognitive function in patients with stable CHD.
Methods
The STABILITY trial was a randomized, double‐blind, placebo‐controlled global cardiovascular outcomes trial. The primary aim was to determine whether darapladib, a specific inhibitor of Lp‐PLA2, reduced the risk of major adverse cardiovascular events (MACE), defined as cardiovascular death, nonfatal myocardial infarction, and nonfatal stroke in patients with CHD. The study recruited patients between December 2008 and April 2010 and was continued until ≈1500 major adverse cardiovascular events had occurred. Assessment of cognitive function was included as an exploratory outcome in a protocol modification in February 2010.14 The study was approved by the institutional review board at all contributing centers, and all participants provided written informed consent for the STABILITY trial and separately for the cognitive function study. Anonymized individual participant data and study documents can be requested.16
Study Population
Participants had stable CHD, defined as prior myocardial infarction, prior coronary revascularization, or multivessel CHD, and also had to meet at least 1 of the following cardiovascular risk enrichment criteria: age ≥60 years; diabetes mellitus requiring pharmacotherapy; high‐density lipoprotein (HDL) cholesterol <1.03 mmol/L; current or previous smoker defined as ≥5 cigarettes per day on average; significant renal dysfunction, defined as estimated glomerular filtration rate ≥30 and <60 mL/min per 1.73 m2 or urine albumin‐to‐creatinine ratio ≥30 mg albumin/g creatinine; or polyvascular disease, defined as CHD and cerebrovascular disease (carotid artery disease, defined as >50% stenosis or previous carotid surgery, or ischemic stroke >3 months) or CHD and peripheral arterial disease. Patients were excluded if there was evidence of dementia. A total of 15 828 subjects from 39 countries were randomized into the study. Use of guideline‐recommended standard of care therapies was encouraged, with 98% taking statins and ≥95% taking aspirin. There was no significant difference in major adverse cardiovascular events for subjects randomized to darapladib compared with placebo after a median follow‐up of 3.7 years. More detailed descriptions of the study design and population have been published previously.14
Baseline Clinical Assessment
In addition to the risk markers listed above, body weight, body mass index, history of hypertension, prior coronary artery bypass grafting and history of paroxysmal or persistent atrial fibrillation were noted. Diabetes mellitus was defined as a physician diagnosis requiring pharmacotherapy. Fasting plasma low‐density lipoprotein and HDL cholesterol, glucose, and Lp‐PLA2 activity were measured in core laboratories. ApoE4ɛ allele status was determined in 9388 participants using the HumanOmniExpressExome‐8 version 1 array by Expression Analysis Inc. (Durham, NC).
Participants were also invited to complete a lifestyle questionnaire. Physical activity was assessed by asking, “How many hours during a typical week do you spend doing the following activities for 10 minutes or more?” (1) Doing MILD physical activity (estimated to be 2 metabolic equivalents) such as easy walking, yoga, tai chi, mild housework? (2) Doing MODERATE (4 metabolic equivalents) physical activity such as fast walking, jogging, aerobics, gardening, bicycling, dancing, swimming, or housecleaning. (3) Doing VIGOROUS (8 metabolic equivalents) physical activity such as running, lifting heavy objects, playing strenuous sport, or doing strenuous work?17 The risk factor of low physical activity was defined as taking <2.5 hours of moderate or greater intensity physical activity recommended in physical activity guidleines.18 A Mediterranean diet score (MDS) was based on self‐reported weekly and daily consumption of foods from a Mediterranean dietary pattern using a food frequency questionnaire.19 A Mediterranean diet score ≥15 was associated with a reduced risk of major adverse cardiovascular events and lower mortality, and therefore defined as “healthy.”19 Years of formal education completed were categorized as “none or 1 to 8 years,” “9 to 12 years,” “trade,” and “college or university.” Depressive symptoms were evaluated by responses (always, often, sometimes, never) to 2 short questions on mood and interest in activities.
Countries were grouped in the following “geographic” regions: North America (United States and Canada), South America (Argentina, Brazil, Chile, and Peru), Western Europe (Belgium, Denmark, France, Germany, Italy, Netherlands, Norway, Spain, Sweden, and United Kingdom) and Australia and New Zealand, Eastern Europe (Bulgaria, Czech Republic, Estonia, Greece, Hungary, Poland, Romania, Russia, Slovakia, Ukraine), and Asia/other (China, Hong Kong, India, Japan, Korea, Pakistan, Philippines, Taiwan, Thailand, and South Africa).
Assessment of Cognitive Function
Cognitive function was assessed once using an appropriate language version of the Montreal Cognitive Assessment (MoCA),20 which is a screening test developed to identify potential cases of mild cognitive impairment. The MoCA was separately validated for each available language version. It was administered at the last routine study visit, conducted 3 to 6 months before study close‐out as specified by the STABILITY coordinating center. Patients were eligible for inclusion in the cognitive function study if they had taken the study medication for at least 1 year and provided written informed consent. They were excluded if there was no suitable validated language version of the MoCA, or if they were unable to complete the MoCA for physical reasons.
The MoCA is sensitive for detection of mild cognitive impairment and can be completed in about 10 minutes. Questions are designed to assess several cognitive domains: memory, visual‐spatial ability, executive function, attention, concentration, language, and orientation to time and place. Points are allocated for correct responses to a total of ≤30 points. An additional point was also allocated for people with <8 years education if the total score was <30. The primary outcome was a MoCA score <26, chosen to indicate cognitive impairment based on published literature.21 Based on the overall distribution of MoCA scores, a score <22 representing the lowest 10% to 15% of MoCA scores among these subjects, was chosen as a secondary outcome to identify participants with “moderate” or “severe” cognitive impairment. A MoCA score <16 was chosen to indicate severe cognitive impairment.
Statistical Analysis
To compare MoCA scores by randomized treatment group, the mean and SD for the overall and components of the MoCA are reported, and the proportion of subjects with a MoCA <26, MoCA <22, and MoCA <16, with statistical testing using the MantelHaenszel test.
Baseline characteristics were then reported for categories of cognitive function classification. Discrete factors were reported as frequencies and percentages. Continuous factors were reported as means and SDs. Mantel‐Haenszel tests were generated for discrete factors, and the Spearman rank sum test for continuous measurements. Three statistical models were generated to assess the MoCA score. Two logistic models were fitted, one for any cognitive impairment (MoCA score <26 versus ≥26) and a second for moderate or severe cognitive impairment (MoCA score <22 versus ≥22). A linear regression model was fitted for the continuous MoCA score. Although the marginal distribution of the MoCA score is left skewed, this modeling approach was chosen so that covariate effects could be quantified in terms of changes in the mean MoCA score. With our large sample size, inference based on asymptotic normality of the parameter estimates is reasonable even when the underlying response variable does not have a normal distribution. Before modeling continuous risk factors, the distribution of measurements was assessed and modeling assumptions verified. For continuous and ordinal covariates, a linearity test was performed by testing whether a model including restricted cubic spline transformations of the covariate improved the fit of the model. For age, we found significant evidence of nonlinearity in the 2 logisitic models. The nonlinear relationship was approximated using a linear spline with a knot at 65 years, that is, by including 2 linear terms in the model. This allows for a different coefficient (eg, odds ratio [OR]) to be estimated for unit increments in age <65 years than increments >65 years.
Covariates were selected in 2 steps in the logistic regression analysis of cognitive impairment (MoCA <26). At the first step, the model included treatment, age, and sex and allowed for the forward selection of region of enrollment, years of education (0–8 years, either 9–12 years or trade school [combined because MoCA scores were similar], and college/university), and physical activity level. Factors were allowed into the model if the P≤0.05. At the second step, factors found to be significant in the first step were included, and markers of disease severity (prior myocardial infarction, prior coronary revascularization, multivessel disease confirmed by angiography, polyvascular disease, and estimated glomerular filtration rate <60 mL/min per m2), prior stroke, current smoking, history of hypertension, diabetes mellitus requiring pharmacotherapy, ApoE4ɛ allele positive, Mediterranean diet score ≥15, alcohol consumption of 6 or more drinks in a single session at least once per week, HDL cholesterol (mmol/L), low‐density lipoprotein cholesterol (mmol/L), and body mass index (kg/m2) were added by forward selection using a P≤0.05. The same set of covariates was used in the models for moderate/severe cognitive impairment (MoCA <22) and continuous MoCA score. For the logistic regression models, the OR, 95% CI, and P values are reported. All modeling assumptions were verified. Transformations were used for continuous measurements when the assumption of linearity was not met. For the generalized linear regression model, the beta coefficients (average difference in MoCA), 95% CI, and P values are reported. Statistical significance was assessed using 2‐sided P values. All statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC).
Results
A total of 5532 of 7904 (70%) of STABILITY participants randomized to placebo and 5191 of 7924 (66%) randomized to darapladib completed the MoCA after a mean (SD) duration of treatment of 3.2±0.37 years. Reasons for not completing the MoCA are displayed by treatment group in Figure 1. There were small differences in some baseline clinical characteristics for subjects who did and did not complete the MoCA, presented in Table S1. Subjects who did not complete the MoCA were slightly older and more likely to be women; have diabetes mellitus; get less exercise; have polyvascular disease, history of stroke, or significant renal disease; have fewer than 8 years of education; or live in North America. However, baseline characteristics for subjects who completed the MoCA were well matched by treatment allocation.
Figure 1.
Study flow diagram. Flow diagram indicating STABILITY study participants who completed the MoCA after a median follow‐up of 3.2±0.4 years, and reasons for noncompletion by treatment allocation. MoCA indicates Montreal Cognitive Assessment. Patients may have more than one reason for exclusion; STABILITY, Stabilization of Atherosclerotic Plaque by Initiation of Darapladib Therapy.
MoCA results are presented by treatment allocation in Table 1. There was no difference in overall MoCA scores between subjects randomized to placebo or darapladib (mean±SD: 25.3±3.82 versus 25.3±3.71, respectively; P=0.27). For each cognitive domain, scores were similar between treatment groups. For mild cognitive impairment, the adjusted OR (95% CI) comparing darapladib to placebo was 1.00 (0.93–1.09) and for moderate/severe cognitive impairment, the adjusted OR (95% CI) was 1.07 (0.95–1.21).
Table 1.
MoCA Results by Treatment Allocation to Darapladib or Placebo
| Possible Score on MoCA | Placebo | Darapladib | |
|---|---|---|---|
| Number of subjects | 5532 | 5191 | |
| Mean±SD | Mean±SD | ||
| Components of MoCA | |||
| Visuospatial/executive | 6 | 4.0±1.19 | 4.0±1.18 |
| Naming | 3 | 2.9±0.41 | 2.9±0.42 |
| Attention: digits | 2 | 1.8±0.50 | 1.7±0.51 |
| Attention: letters | 1 | 0.9±0.29 | 0.9±0.29 |
| Attention: subtraction | 3 | 2.6±0.76 | 2.6±0.78 |
| Language: repeat | 2 | 1.6±0.60 | 1.6±0.62 |
| Language: fluency | 1 | 0.6±0.48 | 0.6±0.48 |
| Abstraction | 2 | 1.7±0.58 | 1.7±0.59 |
| Delayed recall (no cue) | 4 | 2.8±1.63 | 2.8±1.64 |
| Orientation | 6 | 5.9±0.49 | 5.9±0.49 |
| Total score | 30 | 25.4±3.73 | 25.3±3.84 |
| MoCA <26, n (%) | 2347 (42.4) | 2231 (42.0) | |
| MoCA <22, n (%) | 751 (13.7) | 751 (14.6) | |
| MoCA <16, n (%) | 106 (1.9) | 123 (2.4) | |
Results are mean±standard deviation (SD) or n (%). One point is also added for people with <8 years education if total is <30. For all comparisons P>0.10. MoCA indicates Montreal Cognitive Assessment.
Associations With Cognitive Impairment
The baseline characteristics for all participants who completed the MoCA are displayed for all subjects and by MoCA score group in Table 2. The ORs for the multivariate model of any cognitive impairment or moderate/severe cognitive impairment according to prespecified risk factors are shown in Table 3. After adjusting for all other covariates, there were increased odds of cognitive impairment associated with baseline age, geographic regions of the United States/Canada and South America (compared with Western Europe), education less than college/university, diabetes mellitus requiring pharmacotherapy, history of hypertension, low physical activity, low HDL cholesterol, history of stroke, and severe depressive symptoms. In the multivariate model, there was no evidence for an association between cognitive impairment (P>0.05) and current smoking, body mass index, diet score, multivessel CHD, polyvascular disease, significant renal dysfunction, ApoE4ɛ allele, low‐density lipoprotein cholesterol, plasma levels of Lp‐PLA2 activity, prior coronary artery bypass grafting, or history of atrial fibrillation.
Table 2.
Baseline Characteristics of Study Population Overall and by MoCA Score After >1 Year of Randomized Treatment
| Characteristics | All Patients (N=10 634) | Moderately Impaired Cognitive Function (N=1502) | Mildly Impaired Cognitive Function (N=3076) | Normal Cognitive Function (N=6056) | P Value |
|---|---|---|---|---|---|
| MoCA score after mean (SD) 3.2 (±0.37) years of treatment | <22 | 22 to 25 | ≥26 | ||
| Randomized treatment | 0.31 | ||||
| Placebo | 5485 (51.6%) | 751 (50.0%) | 1596 (51.9%) | 3138 (51.8%) | |
| Darapladib | 5149 (48.4%) | 751 (50.0%) | 1480 (48.1%) | 2918 (48.2%) | |
| Baseline characteristics | |||||
| Age, y | 64.0±9.0 | 67.8±8.6 | 64.8±8.7 | 62.6±9.0 | <0.0001 |
| Female sex | 1908 (17.9%) | 402 (26.8%) | 548 (17.8%) | 958 (15.8%) | <0.0001 |
| Cardiovascular risk factors, n (%) | |||||
| Smoker status | <0.0001 | ||||
| Never smoked | 3254 (30.6%) | 570 (37.9%) | 959 (31.2%) | 1725 (28.5%) | |
| Former smoker | 5484 (51.6%) | 733 (48.8%) | 1589 (51.7%) | 3162 (52.2%) | |
| Current smoker | 1895 (17.8%) | 199 (13.2%) | 527 (17.1%) | 1169 (19.3%) | |
| LDL cholesterol, mmol/L | 2.20±0.83 | 2.22±0.82 | 2.19±0.83 | 2.19±0.84 | 0.30 |
| Lp‐PLA2 activity, μmol/min per liter | 174±47 | 172±50 | 174±47 | 175±45 | 0.21 |
| HDL cholesterol, mmol/L | 1.21±0.32 | 1.22±0.31 | 1.21±0.32 | 1.21±0.32 | 0.13 |
| Diabetes mellitus | 3966 (37.3%) | 608 (40.5%) | 1186 (38.6%) | 2172 (35.9%) | 0.0002 |
| Body mass index | 0.0030 | ||||
| <25 kg/m2 | 2138 (20.1%) | 360 (24.0%) | 608 (19.8%) | 1170 (19.4%) | |
| 25 to <30 kg/m2 | 4567 (43.0%) | 646 (43.1%) | 1304 (42.5%) | 2617 (43.3%) | |
| ≥30 kg/m2 | 3910 (36.8%) | 493 (32.9%) | 1158 (37.7%) | 2259 (37.4%) | |
| Hypertension | 7544 (70.9%) | 1146 (76.3%) | 2232 (72.6%) | 4166 (68.8%) | <0.0001 |
| Moderate or vigorous activity <2.5 h/wk | 3786 (35.6%) | 709 (47.2%) | 1117 (36.3%) | 1956 (32.3%) | <0.0001 |
| Excessive alcohol use | 606 (5.8%) | 76 (5.2%) | 183 (6.1%) | 347 (5.8%) | 0.65 |
| Mediterranean diet score ≥15 | 1981 (18.9%) | 245 (16.5%) | 548 (18.0%) | 1188 (19.8%) | 0.0015 |
| Cardiovascular and renal disease | |||||
| Prior myocardial infarction | 6263 (58.9%) | 887 (59.1%) | 1777 (57.8%) | 3599 (59.4%) | 0.42 |
| Prior coronary artery bypass graft | 3505 (33.0%) | 486 (32.4%) | 1045 (34.0%) | 1974 (32.6%) | 0.72 |
| Prior stroke | 581 (5.5%) | 106 (7.1%) | 201 (6.5%) | 274 (4.5%) | <0.0001 |
| Multivessel CHD | 1569 (14.8%) | 257 (17.1%) | 467 (15.2%) | 845 (14.0%) | 0.0016 |
| Polyvascular disease | 1480 (13.9%) | 235 (15.6%) | 459 (14.9%) | 786 (13.0%) | 0.0014 |
| Significant renal disease | 2998 (28.2%) | 539 (35.9%) | 953 (31.0%) | 1506 (24.9%) | <0.0001 |
| Atrial fibrillation | 839 (7.9%) | 132 (8.8%) | 248 (8.1%) | 459 (7.6%) | 0.27 |
| Genetic markers | |||||
| ApoE4ɛ | |||||
| 1 allele | 1967 (20.6%) | 260 (19.5%) | 563 (20.2%) | 1144 (21.0%) | 0.012 |
| 2 alleles | 156 (1.6%) | 24 (1.8%) | 49 (1.8%) | 83 (1.5%) | |
| Geographic region grouping | <0.0001 | ||||
| Western Europe Australia, New Zealand | 2851 (26.8%) | 305 (20.3%) | 785 (25.5%) | 1761 (29.1%) | |
| Asia, South Africa | 1956 (18.4%) | 346 (23.0%) | 570 (18.5%) | 1040 (17.2%) | |
| Eastern Europe | 2614 (24.6%) | 312 (20.8%) | 771 (25.1%) | 1531 (25.3%) | |
| South America | 819 (7.7%) | 273 (18.2%) | 243 (7.9%) | 303 (5.0%) | |
| North America | 2394 (22.5%) | 266 (17.7%) | 707 (23.0%) | 1421 (23.5%) | |
| Education | <0.0001 | ||||
| None | 320 (3.1%) | 132 (9.0%) | 96 (3.2%) | 92 (1.5%) | |
| 1–8 y | 1902 (18.2%) | 542 (36.9%) | 601 (19.8%) | 759 (12.7%) | |
| 9–12 y | 3244 (31.0%) | 396 (26.9%) | 914 (30.2%) | 1934 (32.4%) | |
| Trade school | 1957 (18.7%) | 198 (13.5%) | 591 (19.5%) | 1168 (19.6%) | |
| College/university | 3047 (29.1%) | 202 (13.7%) | 827 (27.3%) | 2018 (33.8%) | |
Results are number (%) or mean± SD. CHD indicates coronary heart disease; HDL, high‐density lipoprotein; LDL, low‐density lipoprotein; Lp‐PLA2, lipoprotein‐associated phospholipase A2; MI, myocardial infarction; MoCA, Montreal Cognitive Assessment.
Table 3.
Independent Predictors of Cognitive Dysfunction
| Any Cognitive Dysfunction: MoCA <26 OR (95% CI) | P Value | Moderate or Severe Cognitive Dysfunction: MoCA <22 OR (95% CI) | P Value | |
|---|---|---|---|---|
| Age ≤65 (+5 y) | 1.17 (1.12, 1.22) | <0.0001 | 1.22 (1.14, 1.31) | <0.0001 |
| Age >65 (+5 y) | 1.33 (1.27, 1.41) | <0.0001 | 1.50 (1.41, 1.61) | <0.0001 |
| Female vs male | 1.09 (0.97, 1.21) | 0.1508 | 1.29 (1.11, 1.49) | 0.0008 |
| Education level (vs college/university) | ||||
| ≤8 y | 2.95 (2.60, 3.35) | <0.0001 | 5.31 (4.40, 6.41) | <0.0001 |
| >8 y/trade | 1.38 (1.25, 1.52) | <0.0001 | 1.89 (1.60, 2.25) | 0.0023 |
| Geographic region (vs Western Europe/Australia/New Zealand) | ||||
| United States/Canada | 1.34 (1.18, 1.51) | 0.021 | 1.49 (1.23, 1.81) | 0.0163 |
| Asia | 1.60 (1.41, 1.82) | 0.069 | 2.06 (1.71, 2.48) | 0.0054 |
| Eastern Europe | 1.45 (1.29, 1.64) | 0.645 | 1.63 (1.35, 1.97) | 0.2663 |
| South America | 2.27 (1.91, 2.70) | <0.0001 | 3.22 (2.61, 3.97) | <0.0001 |
| Other risk factors | ||||
| Low physical activity | 1.19 (1.04, 1.37) | 0.0005 | 1.31 (1.16, 1.49) | <0.0001 |
| Hypertension | 1.12 (1.02, 1.23) | 0.0215 | 1.19 (1.03, 1.38) | 0.0159 |
| Diabetes mellitus | 1.15 (1.05, 1.26) | 0.0027 | 1.15 (1.01, 1.31) | 0.0335 |
| HDL cholesterol ≤1.16 mmol/L | 1.19 (1.04, 1.37) | 0.0134 | 1.24 (1.01, 1.24) | 0.0371 |
| History of stroke | 1.43 (1.20, 1.71) | <0.0001 | 1.27 (1.00, 1.62) | 0.0494 |
The odds of mild (MoCA <26 vs MoCA ≥26) and moderate (MoCA <22 vs MoCA ≥22) cognitive dysfunction are presented as OR and 95% CI from multivariable models including all listed covariates as well as randomized treatment. Nonsignificant variables are not shown. An OR >1 indicates greater odds of cognitive dysfunction.
<2.5 h/wk moderate or vigorous exercise. HDL indicates high‐density lipoprotein; MoCA, Montreal Cognitive Assessment; OR, odds ratio.
Increased ORs for moderate or severe cognitive impairment were observed for the same baseline characteristics as for MoCA <26 versus ≥26, and the strength of association appeared to be stronger for most but not all associations (Table 3).
Multivariate Models and Continuous MoCA Scores
In the multivariable model that included all covariates, education and geographic region were the strongest independent predictors of continuous MoCA scores. Compared with people with a college or university education, MoCA score was on average 2.52 (95% CI, 2.31–2.73) points lower for those with ≤8 years education, and 0.61 (95% CI, 0.45–0.77) points lower for those with >8 years or a trade school education after adjusting for all other covariates. Compared with people living in Western Europe, Australia, or New Zealand, MoCA scores were lower for North America by 0.63 (95% CI, 0.83–0.43) points, Eastern Europe by 0.78 (95% CI, 0.97–0.58) points, Asia/Pacific by 1.08 (95% CI, 1.28–0.87) points, and South America by 2.27 (95% CI, 2.55–2.00) points after adjusting for all other covariates. Age and sex were also independently associated with MoCA score.
In fully adjusted models, MoCA score was lower (−0.59; 95% CI, −0.30 to −0.88 points) for patients with a history of stroke. Lower MoCA scores (95% CI) were also observed for participants reporting <2.5 h/wk of moderate or vigorous exercise (−0.45; 95% CI, −0.31 to −0.59 points), diabetes mellitus on pharmacotherapy (−0.21; 95% CI, −0.07 to −0.36 points), and history of hypertension (−0.17; 95% CI, −0.01 to −0.32 points). There was no evidence for differences (P>0.05) in MoCA score by Mediterranean diet score, ApoE4ɛ, body mass index, current smoking, prior coronary artery bypass grafting, or markers of cardiovascular disease after adjusting for all other covariates.
At baseline, 166 (1.6%) subjects reported feeling down “always,” and these subjects were more likely to have a MoCA score <26 (OR, 1.94; 95% CI, 1.34–2.81) and a MoCA score <22 (OR, 2.66; 95% CI, 1.76–4.02). However, less persistent depressive symptoms and “loss of interest” in activities were not independently associated with MoCA score. Associations of other risk factors with MoCA score were similar before and after also adjusting for depressive symptoms (Table S2).
Interactions Between Age and Other Risk Factors
MoCA score decreased with older age throughout the distribution of scores. The decrease in median MoCA score (Figure 2) was parallel between strata based on physical activity less than versus ≥2.5 hours of moderate or vigorous physical activity (Figure 2A) or education (Figure 2B). There was no evidence for an interaction between MoCA score, age, and physical activity, education, diabetes mellitus requiring pharmacotherapy, hypertension, HDL cholesterol, history of stroke, or geographic region (P>0.05 for all).
Figure 2.
MoCA score by age stratified by (A) physical activity level, and (B) education. Average age‐related change in MoCA score is plotted for subjects by presence or absence of (A) <2.5 h/wk of moderate or vigorous physical activity and (B) education level. MoCA indicates Montreal Cognitive Assessment.
Statistical tests for interaction between sex and MoCA were significant (P<0.0001) for geographic region and years of education but not age and other predictors of MoCA score in the multivariable model (P>0.05; Table S3). For subjects with <8 years’ education, women had lower MoCA scores than men, while for participants with >12 years education, women were less likely to have lower MoCA scores than men. Adjusting for education and other covariates, MoCA scores were lower in women compared with men in Eastern Europe, South America, and Asia, but higher in North America, Western Europe, Australia, and New Zealand (Table S4).
Discussion
This study assessed effects of darapladib compared with placebo, as well as associations of potentially modifiable cardiovascular risk factors, with mild cognitive impairment, in a diverse global population of patients with stable CHD. Mild cognitive impairment is important because it is associated with an elevated risk of developing dementia.22 The MoCA was chosen because it is a short, easily administered, validated test developed to screen patients for minor cognitive impairment, which is available in many languages.20 A MoCA score <26 was the prespecified threshold for mild cognitive impairment,20 but scores <26 have been reported as common in other ethnically diverse populations.23 In this study, which included subjects with a mean age of 64 years, ≈40% of participants had a MoCA score <26. This prevalence of “mild cognitive impairment” was higher than the 4% to 19% reported in some general populations ≥65 years using other criteria.8, 24 It is probable MoCA scores below 26 reflect “usual” cognitive performance for some people, while for others there has been a change from a previously higher level. The lower MoCA score of <22, found in ≈14% of this study population, was included to provide information on associations with a lower level of cognitive function.
Each language version of the MoCA was validated before release. However, socioeconomic, cultural, and other differences between the validation studies and patients in the STABILITY trial, including differences in familiarity with this type of testing, may have been important. The odds of mild and moderate cognitive impairment differed between geographic regions. Because reasons for geographic differences are uncertain and may be multiple, using a single threshold may be misleading when assessing cognitive function in people from diverse cultures and backgrounds, unless effects of these potential covariates are considered.23
The strongest association was between cognitive function and level of education, which was consistent across the age range. Low education could also explain the lower MoCA scores in women compared with men in several geographic locations. These observations are consistent with the “brain reserve hypothesis” in which longer time in education allows a subject to compensate for neuropathological burden in later life.22, 25 However, while college‐ or university‐level education was associated with a higher MoCA score, the decrease in MoCA score with increase in age was similar by education level.
Low physical activity, diabetes mellitus, hypertension, and stroke have each been associated with risk of dementia or cognitive impairment in previous epidemiological studies.4, 5, 6, 7
The current study extends this evidence by demonstrating independent associations between cognitive impairment and these risk factors among patients with CHD. These results support the idea that interventions to reduce cardiovascular risk may have additional benefits of risk reduction for cognitive impairment.
Lp‐PLA2, which increases production of proinflammatory and proapoptotic mediators, is one of several inflammatory biomarkers that have been associated with the risk of dementia or impaired cognitive function.26, 27 Lp‐PLA2 activity has been reported to be higher in patients with Alzheimer disease compared with controls,9 and to be associated with dementia in epidemiologic studies.10, 11, 12, 13 However, these associations may not be causal. In the current study, darapladib, which decreased plasma Lp‐PLA2 activity by ≈65%, was not associated with a difference in cognitive function. Compared with observational data, the randomized comparison of darapladib with placebo provides more reliable evidence regarding the potential role of Lp‐PLA2. However, this result does not exclude the importance of other pathways that could link cardiovascular risk factors to the risk of cognitive impairment.
Study Limitations
The primary aim of the STABILITY trial was to evaluate effects of darapladib on cardiovascular end points. A single MoCA assessment was included to evaluate a possible association between darapladib therapy and cognitive function, but with no attempt to ensure adequate power for this exploratory objective, and change in cognitive performance could not be assessed because a baseline MoCA was not included. Despite this, the 95% CIs suggest that a clinically important effect of darapladib on cognitive function is unlikely. The MoCA assessment was performed at a specified visit near study end. Although the length of treatment varied among patients depending on the timing of randomization, this was balanced by treatment group and is unlikely to bias assessment of the treatment effect.
Observed associations between cardiovascular risk factors and cognitive function are consistent with results from large epidemiologic cohorts,4, 5, 6, 7 but causality cannot be confirmed from observational studies. Also, evaluations of predictors of cognitive function were post hoc. The study did not assess change in MoCA scores over time. The ability to identify some risk factors, such as ApoE4ɛ allele, would be reduced if most of the difference in MoCA scores were not related to Alzheimer disease. A history of paroxysmal or persistent atrial fibrillation was present at baseline in ≈8% of participants, so the study had low statistical power to evaluate a modest association with the risk of cognitive impairment or the impact of preventive treatments such as oral anticoagulants. The study did not include assessment of left ventricular function. In this study, 98% of subjects were taking statins. The lack of an association between low‐density lipoprotein cholesterol, which is influenced by statin dose, and MoCA score is reassuring regarding possible concerns that statins may cause memory impairment.28
Depressive symptoms have been associated with risk of dementia in previous studies,29 and a small proportion of the participants in this study who reported depressive symptoms “always” had lower MoCA scores. It is possible that depression is a marker of the presence of, rather than a risk factor for, cognitive impairment, and for this reason it was not included in the primary analysis. Other study results were similar in a secondary analysis that also included depressive symptoms as a covariate. Activities of daily living and social functioning were not assessed as part of this study.
Conclusions
In patients with stable CHD, inhibition of Lp‐PLA2 with darapladib did not influence cognitive function. However, low physical activity, diabetes mellitus, stroke, education, and geography were associated with differences in cognitive performance. Although causality was not confirmed, these observations are consistent with the hypothesis that long‐term adherence to a healthy lifestyle and optimal management of cardiovascular risk factors may have favorable effects on cognitive function.
Author Contributions
The study was conceived and designed in collaboration with GlaxoSmithKline by Drs Stewart, Krug‐Gourley, Waterworth, Hagstrom, Held, Armstrong, Wallentin, and White. Statistical analysis was undertaken by Ms Stebbins and Dr Chiswell, who had full access to study data and take responsibility for the integrity of the data and the accuracy of the data analysis. The manuscript was drafted by Dr Stewart. All authors contributed to the critical review and final approval of the submitted manuscript.
Sources of Funding
The STABILITY trial and this cognitive function analysis were both sponsored/funded by GlaxoSmithKline. This included funding for conduct of the STABILITY trial, collection of data, interpretation, and statistical analyses. GlaxoSmithKline reviewed the final manuscript and approved the decision to submit for publication.
Disclosures
Drs Stewart, Held, Hagstrom, Armstrong, Wallentin, and White are STABILITY Study Investigators. Dr Stewart reports grants and nonfinancial support from GlaxoSmithKline; Dr Held reports institutional research grant and speaker's bureau from AstraZeneca, and institutional research grants from Bristol‐Myers Squibb Merck & Co, GlaxoSmithKline, and Roche; Dr Krug‐Gourley is an employee of and has stock ownership in GlaxoSmithKline; Dr Waterworth is an employee of and has stock ownership in GlaxoSmithKline; Ms Stebbins has no disclosures to report; Dr Chiswell has no disclosures to report; Dr Hagstrom reports institutional research grants from GlaxoSmithKline, AstraZeneca, Amgen, Sanofi, and Ariad, and expert committee member, lecture fees, and institutional research grant from Sanofi and Amgen; institutional research grants from AstraZeneca, and GlaxoSmitKline; and expert committee member for Ariad and MSD; Dr Armstrong reports grants and personal fees from Merck and Bayer, grants from Sanofi‐Aventis, and personal fees from AstraZeneca, Axio/Orexigen, Eli Lilly, and Mast Therapeutics; Dr Wallentin reports institutional research grants, consultancy fees, lecture fees, and travel support from Bristol‐Myers Squibb/Pfizer, AstraZeneca, GlaxoSmithKline, and Boehringer Ingelheim; institutional research grants from Merck & Co and Roche; consultancy fees from Abbott; and holds 2 patents involving GDF‐15; and Dr White reports research grants from Sanofi‐Aventis, Eli Lilly, National Institute of Health, Omthera Pharmaceuticals, Pfizer New Zealand, Intarcia Therapeutics Inc., Elsai Inc., and Dalcor Pharma UK Inc; honoraria and nonfinancial support from AstraZeneca; and is on the Advisory Boards for Sirtex, Acetilion, CSL Behring LLC, and Luitpold Pharmaceuticals Inc.
Supporting information
Table S1. Baseline Characteristics for STABILITY Trial Patients Included and Not Included in the MoCA Assessment Cohort
Table S2. Independent Predictors of Cognitive Dysfunction After Also Adjusting for Depressive Symptoms at Baseline
Table S3. Statistical Tests for Interaction Between Sex, Other Risk Factors, and MoCA Score
Table S4. Sex Differences in MoCA Scores by Geographic Region of Enrollment and by Level of Education
Acknowledgments
The authors thank the STABILITY study investigators and participants for their contribution to this study.
(J Am Heart Assoc. 2019;8:e010641 DOI: 10.1161/JAHA.118.010641.)
References
- 1. Gill TM, Gahbauer EA, Han L, Allore HG. Trajectories of disability in the last year of life. N Engl J Med. 2010;362:1173–1180. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Ferri CP, Prince M, Brayne C, Brodaty H, Fratiglioni L, Ganguli M, Hall K, Hasegawa K, Hendrie H, Huang Y, Jorm A, Mathers C, Menezes PR, Rimmer E, Scazufca M. Global prevalence of dementia: a Delphi consensus study. Lancet. 2005;366:2112–2117. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Gauthier S, Reisberg B, Zaudig M, Petersen RC, Ritchie K, Broich K, Belleville S, Brodaty H, Bennett D, Chertkow H, Cummings JL, de Leon M, Feldman H, Ganguli M, Hampel H, Scheltens P, Tierney MC, Whitehouse P, Winblad B. Mild cognitive impairment. Lancet. 2006;367 1262–1270. [DOI] [PubMed] [Google Scholar]
- 4. Debette S, Seshadri S, Beiser A, Au R, Himali JJ, Palumbo C, Wolf PA, DeCarli C. Midlife vascular risk factor exposure accelerates structural brain aging and cognitive decline. Neurology. 2011;77:461–468. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Plassman BL, Williams JW Jr, Burke JR, Holsinger T, Benjamin S. Systematic review: factors associated with risk for and possible prevention of cognitive decline in later life. Ann Intern Med. 2010;153:182–193. [DOI] [PubMed] [Google Scholar]
- 6. Gorelick PB, Furie KL, Iadecola C, Smith EE, Waddy SP, Lloyd‐Jones DM, Bae HJ, Bauman MA, Dichgans M, Duncan PW, Girgus M, Howard VJ, Lazar RM, Seshadri S, Testai FD, van Gaal S, Yaffe K, Wasiak H, Zerna C. Defining optimal brain health in adults: a presidential advisory from the American Heart Association/American Stroke Association. Stroke. 2017;48:e284–e303. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Yaffe K, Vittinghoff E, Pletcher MJ, Hoang TD, Launer LJ, Whitmer R, Coker LH, Sidney S. Early adult to midlife cardiovascular risk factors and cognitive function. Circulation. 2014;129:1560–1567. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Ganguli M, Fu B, Snitz BE, Hughes TF, Chang CC. Mild cognitive impairment: incidence and vascular risk factors in a population‐based cohort. Neurology. 2013;80:2112–2120. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Doody RS, Demirovic J, Ballantyne CM, Chan W, Barber R, Powell S, Pavlik V. Lipoprotein‐associated phospholipase A2, homocysteine, and Alzheimer's disease. Alzheimers Dement (Amst). 2015;1:464–471. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. van Oijen M, van der Meer IM, Hofman A, Witteman JC, Koudstaal PJ, Breteler MM. Lipoprotein‐associated phospholipase A2 is associated with risk of dementia. Ann Neurol. 2006;59:139–144. [DOI] [PubMed] [Google Scholar]
- 11. Fitzpatrick AL, Irizarry MC, Cushman M, Jenny NS, Chi GC, Koro C. Lipoprotein‐associated phospholipase A2 and risk of dementia in the Cardiovascular Health Study. Atherosclerosis. 2014;235:384–391. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Cai R, Huang R, Han J, Sun H, Sun J, Xia W, Tian S, Dong X, Shen Y, Wang S. Lipoprotein‐associated phospholipase A2 is associated with risk of mild cognitive impairment in Chinese patients with type 2 diabetes. Sci Rep. 2017;7:12311. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Jiang R, Chen S, Shen Y, Wu J, Chen S, Wang A, Wu S, Zhao X. Higher levels of lipoprotein associated phospholipase A2 is associated with increased prevalence of cognitive impairment: the APAC Study. Sci Rep. 2016;6:33073. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. White HD, Held C, Stewart R, Tarka E, Brown R, Davies RY, Budaj A, Harrington RA, Steg PG, Ardissino D, Armstrong PW, Avezum A, Aylward PE, Bryce A, Chen H, Chen MF, Corbalan R, Dalby AJ, Danchin N, De Winter RJ, Denchev S, Diaz R, Elisaf M, Flather MD, Goudev AR, Granger CB, Grinfeld L, Hochman JS, Husted S, Kim HS, Koenig W, Linhart A, Lonn E, Lopez‐Sendon J, Manolis AJ, Mohler ER III, Nicolau JC, Pais P, Parkhomenko A, Pedersen TR, Pella D, Ramos‐Corrales MA, Ruda M, Sereg M, Siddique S, Sinnaeve P, Smith P, Sritara P, Swart HP, Sy RG, Teramoto T, Tse HF, Watson D, Weaver WD, Weiss R, Viigimaa M, Vinereanu D, Zhu J, Cannon CP, Wallentin L. Darapladib for preventing ischemic events in stable coronary heart disease. N Engl J Med. 2014;370:1702–1711. [DOI] [PubMed] [Google Scholar]
- 15. Wallentin L, Held C, Armstrong PW, Cannon CP, Davies RY, Granger CB, Hagstrom E, Harrington RA, Hochman JS, Koenig W, Krug‐Gourley S, Mohler ER III, Siegbahn A, Tarka E, Steg PG, Stewart RA, Weiss R, Ostlund O, White HD; STABILITY Investigators . Lipoprotein‐associated phospholipase A2 activity is a marker of risk but not a useful target for treatment in patients with stable coronary heart disease. J Am Heart Assoc. 2016;5:e003407 DOI: 10.1161/JAHA.116.003407. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. ClinicalStudyDataRequest.com (CSDR) . ClinicalStudy DataRequest.com. 2019. Available at: https://www.clinicalstudydatarequest.com/Default.aspx. Accessed December 1, 2018.
- 17. Stewart R, Held C, Brown R, Vedin O, Hagstrom E, Lonn E, Armstrong P, Granger CB, Hochman J, Davies R, Soffer J, Wallentin L, White H. Physical activity in patients with stable coronary heart disease: an international perspective. Eur Heart J. 2013;34:3286–3293. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Physical Activity Guidelines Advisory Committee . Physical Activity Guidelines Advisory Committee Report, 2008. Washington, D.C.: US Department of Health and Human Services; 2008. [DOI] [PubMed] [Google Scholar]
- 19. Stewart RA, Wallentin L, Benatar J, Danchin N, Hagstrom E, Held C, Husted S, Lonn E, Stebbins A, Chiswell K, Vedin O, Watson D, White HD. Dietary patterns and the risk of major adverse cardiovascular events in a global study of high‐risk patients with stable coronary heart disease. Eur Heart J. 2016;37:1993–2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Nasreddine ZS, Phillips NA, Bedirian V, Charbonneau S, Whitehead V, Collin I, Cummings JL, Chertkow H. The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc. 2005;53:695–699. [DOI] [PubMed] [Google Scholar]
- 21. Davis DHJ, Creavin ST, Yip JLY, Noel‐Storr AH, Brayne C, Cullum S. Montreal Cognitive Assessment for the diagnosis of Alzheimer's disease and other dementias. Cochrane Database Syst Rev. 2015;29:CD010775. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Livingston G, Sommerlad A, Orgeta V, Costafreda SG, Huntley J, Ames D, Ballard C, Banerjee S, Burns A, Cohen‐Mansfield J, Cooper C, Fox N, Gitlin LN, Howard R, Kales HC, Larson EB, Ritchie K, Rockwood K, Sampson EL, Samus Q, Schneider LS, Selbaek G, Teri L, Mukadam N. Dementia prevention, intervention, and care. Lancet. 2017;390:2673–2734. [DOI] [PubMed] [Google Scholar]
- 23. Rossetti HC, Lacritz LH, Cullum CM, Weiner MF. Normative data for the Montreal Cognitive Assessment (MoCA) in a population‐based sample. Neurology. 2011;77:1272–1275. [DOI] [PubMed] [Google Scholar]
- 24. Lopez OL, Kuller LH, Becker JT, Dulberg C, Sweet RA, Gach HM, Dekosky ST. Incidence of dementia in mild cognitive impairment in the Cardiovascular Health Study cognition study. Arch Neurol. 2007;64:416–420. [DOI] [PubMed] [Google Scholar]
- 25. Meng X, D'Arcy C. Education and dementia in the context of the cognitive reserve hypothesis: a systematic review with meta‐analyses and qualitative analyses. PLoS One. 2012;7:e38268. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Darweesh SKL, Wolters FJ, Ikram MA, de Wolf F, Bos D, Hofman A. Inflammatory markers and the risk of dementia and Alzheimer's disease: a meta‐analysis. Alzheimers Dement. 2018;14:1450–1459. [DOI] [PubMed] [Google Scholar]
- 27. Koyama A, O'Brien J, Weuve J, Blacker D, Metti AL, Yaffe K. The role of peripheral inflammatory markers in dementia and Alzheimer's disease: a meta‐analysis. J Gerontol A Biol Sci Med Sci. 2013;68:433–440. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. U.S. Department of Health and Human Services . FDA Drug Safety Communication: Important Safety Label Changes to Cholesterol‐Lowering Statin Drugs. Silver Spring, MD: US Food and Drug Administration; 2012. [Google Scholar]
- 29. Ismail Z, Elbayoumi H, Fischer CE, Hogan DB, Millikin CP, Schweizer T, Mortby ME, Smith EE, Patten SB, Fiest KM. Prevalence of depression in patients with mild cognitive impairment: a systematic review and meta‐analysis. JAMA Psychiatry. 2017;74:58–67. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1. Baseline Characteristics for STABILITY Trial Patients Included and Not Included in the MoCA Assessment Cohort
Table S2. Independent Predictors of Cognitive Dysfunction After Also Adjusting for Depressive Symptoms at Baseline
Table S3. Statistical Tests for Interaction Between Sex, Other Risk Factors, and MoCA Score
Table S4. Sex Differences in MoCA Scores by Geographic Region of Enrollment and by Level of Education