Aging, Cognitive Impairment, Dementia
Conditions
Keywords
Memory, Mild Cognitive Impairment, Alzheimer's Disease, Cognition
Brief summary
The parent trial (COcoa Supplement and Multivitamin Outcomes Study, COSMOS; NCT02422745) is a randomized clinical trial of cocoa extract supplement (containing a total of 500 mg/d flavanols, including 80 mg. (-)-epicatechins), and a standard multivitamin supplement to reduce the risk of cardiovascular disease and cancer among men aged 60 years and older and women aged 65 years and older. A cognitive ancillary study (COSMOS-Mind) is being conducted among men and women, 65 years and older, who are enrolled in COSMOS and will examine whether the cocoa extract supplement or the multivitamin supplement improves cognitive function and reduces risk of cognitive impairment, including Alzheimer's disease (AD) and other related types of dementia. Participants at the Wake Forest site will only be getting a phone call.
Detailed description
Identifying a safe, affordable, and well-tolerated intervention that prevents cognitive decline in older adults is of critical public health importance. There is compelling evidence from basic science, and small clinical studies that cocoa flavanols may protect cognitive function in older adults and reduce risk of Alzheimer's disease (AD) and related dementia. Thus, this intervention could have important downstream benefits for health care utilization and cost, caregiver burden, and overall quality of life for older adults. Cocoa flavanol effects on cognition, however, need to be assessed in a definitive clinical trial. The parent trial (COSMOS) will enroll 21,442 women and men nationwide, drawn from the Women's Health Initiative (WHI) participant cohort and the VITamin D and OmegA-3 TriaL (VITAL) non-randomized respondent cohort, and will be conducted primarily by mail to assess whether high-potency cocoa flavanol extract and a multivitamin - alone or in combination - will reduce risk of cardiovascular disease and cancer over 3 years of follow-up. After the COSMOS trial began, an advanced method to analyze cocoa flavanols was accredited by Association of Official Agricultural Chemists (AOAC) International as a First Action Official Method of Analysis https://doi.org/10.1093/jaoacint/qsaa132 \[secure-web.cisco.com\]). This updated method relies on a reference material (RM8403) recently standardized and made commercially available by the U.S. National Institute of Standards and Technology. While the actual cocoa flavanol content of the COSMOS intervention remained unchanged throughout the trial, the application of this new analytical method led to expected changes in how the total cocoa flavanol content is now reported. Applying AOAC 2020.05/RM8403 to the COSMOS intervention, the total cocoa flavanol content of the COSMOS intervention is now 500 mg/day. Reporting of (-)-epicatechin content remained unaffected. Going forward, this record will therefore apply AOAC 2020.05/RM8403 and report that the COSMOS intervention tested 500 mg/day of cocoa flavanols, including 80 mg of (-)-epicatechin. Once participants are enrolled in COSMOS, the study team at Brigham and Women's Hospital will provide contact information and other parent trial outcomes for COSMOS-Mind participants to the Wake Forest team using a secure web-based data transfer system. This seamless exchange of information between sites will permit timely accounting to ensure that the targeted demographics for enrollment are achieved (e.g. gender, age, race, ethnicity). COSMOS-Mind will only be administered a telephone cognitive assessment to all participants at enrollment (baseline), and annually for 3 years of follow-up by trained and certified staff of the Wake Forest COSMOS-Mind team.
Interventions
None listed
Sponsors
National Institute on Aging (NIA)
Brigham and Women's Hospital
Wake Forest University Health Sciences
Study design
Observational model
COHORT
Time perspective
PROSPECTIVE
Eligibility
Sex/Gender
ALL
Age
65 Years to No maximum
Healthy volunteers
Yes
Inclusion criteria
Participants in COSMOS (NCT02422745) who meet the following criteria are eligible to participate in this ancillary study: Inclusion Criteria - 1. Men and women 65 years and older. 2. Willingness to participate. 3. Ability to answer questions by phone.
Exclusion criteria
- 1. Participants who are co-enrolled in the WHI Memory Study. 2. Participants with insulin-dependent diabetes. 3. Participants who are unable to complete the baseline cognitive assessments.
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Global Cognitive Function: Average Change Over Follow-up From Baseline | From Baseline through Year 3 | A composite cognitive measure that averages individual test z-scores of the Telephone Interview for Cognitive Status-modified (TICSm) Total and Long, the Immediate and Delayed Story Recall tests, the Oral Trail Making A and B tests (log-transformed), the Category Fluency and Verbal Fluency tests, and the Number Span Backward and Forward tests. Scores are ordered so that better performance corresponds to higher positive scores. Longitudinal data are used to create a linear contrast for the primary analysis to compare the difference between the average of scores at Years 1, 2, and 3 and the baseline score, i.e. (z1+z2+z3)/3 - z0, where z's denote the composited measures at baseline and years 1-3. This contrast is fitted using a random effects model for repeated measures. The primary comparison is for the margins of the factorial design: active cocoa flavonal (with or without active multivitamin) versus placebo flavonal (with or without active multivitamin). |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Composite Cognitive Function | From baseline through year 3 | A composite cognitive measure that averages individual test z-scores of the Telephone Interview for Cognitive Status-modified (TICSm) Total and Long, the Immediate and Delayed Story Recall tests, the Oral Trail Making A and B tests (log-transformed), the Category Fluency and Verbal Fluency tests, and the Number Span Backward and Forward tests. Scores are ordered so that better performance corresponds to higher positive scores. Longitudinal data are used to create a linear contrast for the primary analysis to compare the difference between the average of scores at Years 1, 2, and 3 and the baseline score, i.e. (z1+z2+z3)/3 - z0, where z's denote the composited measures at baseline and years 1-3. This contrast is fitted using a random effects model for repeated measures. The comparison for this secondary outcome is based on the margins of the factorial design to compare multivitamin (with or without cocoa flavonal) with multivitamin placebo (with or without cocoa flavonal) |
| Incident Mild Cognitive Impairment | From baseline through year 3 | Mild cognitive impairment based on centralized adjudication of cognitive test scores and input from a proxy. Limited to participants free of mild cognitive impairment at baseline. Incidence over 3 years of follow-up. |
| Incidence of Mild Cognitive Impairment Among Those Free at Baseline | From baseline through year 3 | Mild cognitive impairment based on central adjudication of test scores with input from proxies. Based on test scores administered annually across 3 years. |
| Executive Function | From baseline through year 3 | Average of z-scores for the following tests: Oral Trail Making Test - B (log-transformed), Category Verbal Fluency, Letter Verbal Fluency, Number Span Test (Backwards and Forward). Higher (positive) scores reflect better performance. Linear contrast of scores collected at baseline, year 1, year 2, and year 3: (z1+z2+z3)/3 - z0) |
| Episodic Memory | From baseline through year 3 | Average z-scores from Telephone Interview for Cognitive Status-modified Delayed Word List, Story Immediate Recall, Story Delayed Recall. Higher (positive) scores reflect better performance. Annual test administrations. |
Countries
United States
Participant flow
Participants by arm
| Arm | Count |
|---|---|
| Cocoa Extract + Multivitamin Daily cocoa extract (capsule containing 750 mg/d flavanols, including 75 mg (-)-epicatechin, and 90 mg theobromine) Daily Centrum Silver | 571 |
| Cocoa Extract + Multivitamin Placebo Daily cocoa extract (capsule containing 750 mg/d flavanols, including 75 mg (-)-epicatechin, and 90 mg theobromine) Daily Centrum Silver placebo | 553 |
| Cocoa Extract Placebo + Multivitamin Daily cocoa extract placebo Daily Centrum Silver | 551 |
| Cocoa Extract Placebo + Multivitamin Placebo Daily cocoa extract placebo Daily Centrum Silver placebo | 587 |
| Total | 2,262 |
Withdrawals & dropouts
| Period | Reason | FG000 | FG001 | FG002 | FG003 |
|---|---|---|---|---|---|
| Overall Study | Death | 9 | 12 | 9 | 10 |
| Overall Study | Lost to Follow-up | 67 | 51 | 64 | 69 |
| Overall Study | Withdrawal by Subject | 51 | 49 | 37 | 46 |
Baseline characteristics
| Characteristic | Cocoa Extract + Multivitamin | Total | Cocoa Extract Placebo + Multivitamin Placebo | Cocoa Extract Placebo + Multivitamin | Cocoa Extract + Multivitamin Placebo |
|---|---|---|---|---|---|
| Age, Continuous | 72.7 years STANDARD_DEVIATION 5.5 | 73.0 years STANDARD_DEVIATION 5.6 | 73.1 years STANDARD_DEVIATION 5.7 | 72.9 years STANDARD_DEVIATION 5.6 | 73.2 years STANDARD_DEVIATION 5.6 |
| Global Cognitive Composite | -0.01 z-scores STANDARD_DEVIATION 0.99 | 0.00 z-scores STANDARD_DEVIATION 1 | -0.05 z-scores STANDARD_DEVIATION 0.99 | 0.05 z-scores STANDARD_DEVIATION 1.02 | 0.02 z-scores STANDARD_DEVIATION 1 |
| Race/Ethnicity, Customized American Indian/Alaska Native | 2 Participants | 13 Participants | 2 Participants | 8 Participants | 1 Participants |
| Race/Ethnicity, Customized Asian/Pacific Islander | 9 Participants | 39 Participants | 8 Participants | 10 Participants | 12 Participants |
| Race/Ethnicity, Customized Black/African American | 40 Participants | 131 Participants | 34 Participants | 34 Participants | 23 Participants |
| Race/Ethnicity, Customized Hispanic | 13 Participants | 67 Participants | 21 Participants | 15 Participants | 18 Participants |
| Race/Ethnicity, Customized Multiracial/Unknown | 1 Participants | 6 Participants | 2 Participants | 2 Participants | 1 Participants |
| Race/Ethnicity, Customized Non-Hispanic White | 506 Participants | 2006 Participants | 520 Participants | 482 Participants | 498 Participants |
| Region of Enrollment United States | 571 participants | 2262 participants | 587 participants | 551 participants | 553 participants |
| Sex: Female, Male Female | 339 Participants | 1348 Participants | 360 Participants | 319 Participants | 330 Participants |
| Sex: Female, Male Male | 232 Participants | 914 Participants | 227 Participants | 232 Participants | 223 Participants |
Adverse events
| Event type | EG000 affected / at risk | EG001 affected / at risk | EG002 affected / at risk | EG003 affected / at risk |
|---|---|---|---|---|
| deaths Total, all-cause mortality | 0 / 0 | 0 / 0 | 0 / 0 | 0 / 0 |
| other Total, other adverse events | 0 / 0 | 0 / 0 | 0 / 0 | 0 / 0 |
| serious Total, serious adverse events | 0 / 0 | 0 / 0 | 0 / 0 | 0 / 0 |
Outcome results
Primary
Global Cognitive Function: Average Change Over Follow-up From Baseline
A composite cognitive measure that averages individual test z-scores of the Telephone Interview for Cognitive Status-modified (TICSm) Total and Long, the Immediate and Delayed Story Recall tests, the Oral Trail Making A and B tests (log-transformed), the Category Fluency and Verbal Fluency tests, and the Number Span Backward and Forward tests. Scores are ordered so that better performance corresponds to higher positive scores. Longitudinal data are used to create a linear contrast for the primary analysis to compare the difference between the average of scores at Years 1, 2, and 3 and the baseline score, i.e. (z1+z2+z3)/3 - z0, where z's denote the composited measures at baseline and years 1-3. This contrast is fitted using a random effects model for repeated measures. The primary comparison is for the margins of the factorial design: active cocoa flavonal (with or without active multivitamin) versus placebo flavonal (with or without active multivitamin).
Time frame: From Baseline through Year 3
Population: Intention to treat -- All enrolled participants
| Arm | Measure | Value (MEAN) | Dispersion |
|---|---|---|---|
| Cocoa Extract + Multivitamin | Global Cognitive Function: Average Change Over Follow-up From Baseline | 0.24 score on a scale | Standard Error 0.02 |
| Cocoa Extract + Multivitamin Placebo | Global Cognitive Function: Average Change Over Follow-up From Baseline | 0.15 score on a scale | Standard Error 0.02 |
| Cocoa Extract Placebo + Multivitamin | Global Cognitive Function: Average Change Over Follow-up From Baseline | 0.20 score on a scale | Standard Error 0.02 |
| Cocoa Extract Placebo + Multivitamin Placebo | Global Cognitive Function: Average Change Over Follow-up From Baseline | 0.15 score on a scale | Standard Error 0.02 |
Comparison: Marginal comparisons from a repeated measures model using a linear contrast to estimate the mean differences between baseline versus average values across follow-up.p-value: <0.0595% CI: [-0.02, 0.08]Mixed Models Analysis
Secondary
Composite Cognitive Function
A composite cognitive measure that averages individual test z-scores of the Telephone Interview for Cognitive Status-modified (TICSm) Total and Long, the Immediate and Delayed Story Recall tests, the Oral Trail Making A and B tests (log-transformed), the Category Fluency and Verbal Fluency tests, and the Number Span Backward and Forward tests. Scores are ordered so that better performance corresponds to higher positive scores. Longitudinal data are used to create a linear contrast for the primary analysis to compare the difference between the average of scores at Years 1, 2, and 3 and the baseline score, i.e. (z1+z2+z3)/3 - z0, where z's denote the composited measures at baseline and years 1-3. This contrast is fitted using a random effects model for repeated measures. The comparison for this secondary outcome is based on the margins of the factorial design to compare multivitamin (with or without cocoa flavonal) with multivitamin placebo (with or without cocoa flavonal)
Time frame: From baseline through year 3
Population: Complete randomized cohort
| Arm | Measure | Value (LEAST_SQUARES_MEAN) | Dispersion |
|---|---|---|---|
| Cocoa Extract + Multivitamin | Composite Cognitive Function | 0.24 score on a scale | Standard Error 0.02 |
| Cocoa Extract + Multivitamin Placebo | Composite Cognitive Function | 0.15 score on a scale | Standard Error 0.02 |
| Cocoa Extract Placebo + Multivitamin | Composite Cognitive Function | 0.20 score on a scale | Standard Error 0.02 |
| Cocoa Extract Placebo + Multivitamin Placebo | Composite Cognitive Function | 0.15 score on a scale | Standard Error 0.02 |
Comparison: Marginal comparisons from a repeated measures model using a linear contrast to estimate the mean differences between baseline versus average values across follow-up.p-value: <0.0595% CI: [0.02, 0.12]Wald test
Secondary
Episodic Memory
Average z-scores from Telephone Interview for Cognitive Status-modified Delayed Word List, Story Immediate Recall, Story Delayed Recall. Higher (positive) scores reflect better performance. Annual test administrations.
Time frame: From baseline through year 3
Population: In a factorial design such as this, primary comparisons are cocoa flavanol vs cocoa flavanol placebo (pooling across multivitamin and multivitamin placebo) and multivitamin vs multivitamin placebo (pooling across cocoa flavanol and cocoa flavanol placebo). To provide outcomes for the four cells does not conform with the protocol or the published results and will not be consistent with the statistical analysis plan.
| Arm | Measure | Value (LEAST_SQUARES_MEAN) | Dispersion |
|---|---|---|---|
| Cocoa Extract + Multivitamin | Episodic Memory | 0.29 score on a scale | Standard Error 0.02 |
| Cocoa Extract + Multivitamin Placebo | Episodic Memory | 0.27 score on a scale | Standard Error 0.02 |
Comparison: Unadjusted 2-tailed testp-value: 0.4195% CI: [-0.04, 0.09]Mixed Models Analysis
Secondary
Episodic Memory
Average z-scores from Telephone Interview for Cognitive Status-modified Delayed Word List, Story Immediate Recall, Story Delayed Recall. Higher (positive) scores reflect better performance.
Time frame: From baseline through year 3
Population: All randomized participants. In a factorial design such as this, primary comparisons are cocoa flavanol vs cocoa flavanol placebo (pooling across multivitamin and multivitamin placebo) and multivitamin vs multivitamin placebo (pooling across cocoa flavanol and cocoa flavanol placebo). To provide outcomes for the four cells does not conform with the protocol or the published results and will not be consistent with the statistical analysis plan.
| Arm | Measure | Value (LEAST_SQUARES_MEAN) | Dispersion |
|---|---|---|---|
| Cocoa Extract + Multivitamin | Episodic Memory | 0.31 score on a scale | Standard Error 0.02 |
| Cocoa Extract + Multivitamin Placebo | Episodic Memory | 0.25 score on a scale | Standard Error 0.02 |
Comparison: Multivitamin minus placebop-value: 0.0495% CI: [0.002, 0.126]Mixed Models Analysis
Secondary
Executive Function
Average of z-scores for the following tests: Oral Trail Making Test - B (log-transformed), Category Verbal Fluency, Letter Verbal Fluency, Number Span Test (Backwards and Forward). Higher (positive) scores reflect better performance. Linear contrast of scores collected at baseline, year 1, year 2, and year 3: (z1+z2+z3)/3 - z0)
Time frame: From baseline through year 3
Population: All randomized participants. In a factorial design such as this, primary comparisons are cocoa flavanol vs cocoa flavanol placebo (pooling across multivitamin and multivitamin placebo) and multivitamin vs multivitamin placebo (pooling across cocoa flavanol and cocoa flavanol placebo). To provide outcomes for the four cells does not conform with the protocol or the published results and will not be consistent with the statistical analysis plan.
| Arm | Measure | Value (LEAST_SQUARES_MEAN) | Dispersion |
|---|---|---|---|
| Cocoa Extract + Multivitamin | Executive Function | 0.06 score on a scale | Standard Error 0.02 |
| Cocoa Extract + Multivitamin Placebo | Executive Function | 0.04 score on a scale | Standard Error 0.02 |
p-value: 0.2495% CI: [-0.02, 0.08]Mixed Models Analysis
Secondary
Executive Function
Average of z-scores for the following tests: Oral Trail Making Test - B (log-transformed), Category Verbal Fluency, Letter Verbal Fluency, Number Span Test (Backwards and Forward). Higher (positive) scores reflect better performance. Tests administered annually.
Time frame: From baseline through year 3
Population: All randomized participants. In a factorial design such as this, primary comparisons are cocoa flavanol vs cocoa flavanol placebo (pooling across multivitamin and multivitamin placebo) and multivitamin vs multivitamin placebo (pooling across cocoa flavanol and cocoa flavanol placebo). To provide outcomes for the four cells does not conform with the protocol or the published results and will not be consistent with the statistical analysis plan.
| Arm | Measure | Value (LEAST_SQUARES_MEAN) | Dispersion |
|---|---|---|---|
| Cocoa Extract + Multivitamin | Executive Function | 0.08 score on a scale | Standard Error 0.02 |
| Cocoa Extract + Multivitamin Placebo | Executive Function | 0.02 score on a scale | Standard Error 0.02 |
Comparison: 2-sided test, unadjustedp-value: 0.0295% CI: [0.01, 0.11]Mixed Models Analysis
Secondary
Incidence of Mild Cognitive Impairment Among Those Free at Baseline
Mild cognitive impairment based on central adjudication of test scores with input from proxies. Based on test scores administered annually across 3 years.
Time frame: From baseline through year 3
Population: Participants free of mild cognitive impairment at baseline. In a factorial design such as this, primary comparisons are cocoa flavanol vs cocoa flavanol placebo (pooling across multivitamin and multivitamin placebo) and multivitamin vs multivitamin placebo (pooling across cocoa flavanol and cocoa flavanol placebo). To provide outcomes for the four cells does not conform with the protocol or the published results and will not be consistent with the statistical analysis plan.
| Arm | Measure | Value (COUNT_OF_PARTICIPANTS) |
|---|---|---|
| Cocoa Extract + Multivitamin | Incidence of Mild Cognitive Impairment Among Those Free at Baseline | 53 Participants |
| Cocoa Extract + Multivitamin Placebo | Incidence of Mild Cognitive Impairment Among Those Free at Baseline | 57 Participants |
p-value: 0.6295% CI: [0.63, 1.32]Regression, Cox
Secondary
Incident Mild Cognitive Impairment
Mild cognitive impairment based on centralized adjudication of cognitive test scores and input from a proxy. Limited to participants free of mild cognitive impairment at baseline. Incidence over 3 years of follow-up.
Time frame: From baseline through year 3
Population: Limited to participants free of mild cognitive impairment at baseline. In a factorial design such as this, primary comparisons are cocoa flavanol vs cocoa flavanol placebo (pooling across multivitamin and multivitamin placebo) and multivitamin vs multivitamin placebo (pooling across cocoa flavanol and cocoa flavanol placebo). To provide outcomes for the four cells does not conform with the protocol or the published results and will not be consistent with the statistical analysis plan.
| Arm | Measure | Value (COUNT_OF_PARTICIPANTS) |
|---|---|---|
| Cocoa Extract + Multivitamin | Incident Mild Cognitive Impairment | 47 Participants |
| Cocoa Extract + Multivitamin Placebo | Incident Mild Cognitive Impairment | 63 Participants |
p-value: 0.1595% CI: [0.52, 1.11]Regression, Cox