Abstract
Findings from recent Alzheimer’s disease prevention trials have shown subjects with increased dementia score based upon mid-life cardiovascular risk factors, to benefit from multi-domain intervention strategies to some extent. The effects of such interventions on cognitive functions remains yet to be well-established. This study is a secondary analysis of the MAPT study, 1,293 older subjects (mean age 75 years) with high CAIDE score (i.e., ≥6) were classified according to the four intervention groups: 1) multi-domain intervention plus placebo, 2) isolated supplementation with Omega-3 polyunsaturated fatty acid (n-3 PUFA), 3) combination of the two interventions, and 4) placebo alone. Linear mixed-model repeated-measures analyses were used to assess the cognitive changes according to various neuropsychological test scores between intervention groups compared to the placebo at 36 months from baseline. Compared to the placebo, group with multi-domain intervention in combination withn-3PUFA was found to show significant improvement in the delayed total recall test of the free and cued selective reminding test (FCSRT) (mean±standard error(SE) = 0.20±0.10) and MMSE orientation test (mean±SE = 0.15±0.06) at 36 months. Isolated multi-domain intervention group showed significant less decline in the MMSE orientation test (mean±SE = 0.12±0.06) compared to the placebo. There was significant less improvement (mean±SE = – 1.01±0.46) in the FCSRT free recall test in the n-3 PUFA intervention group compared to the placebo at 36 months. Our findings show high-risk subjects for dementia screened with CAIDE dementia score might benefit from multi-domain intervention strategies as in the MAPT study, particularly in the orientation and delayed recall domain.
Keywords
INTRODUCTION
Reducing the global burden of cognitive disorders such as dementia is a major challenge for the scientific community today [1]. Based on the evidence from past dementia clinical trials (in particular Alzheimer’s disease, AD), a consensus has been reached that interventions targeting the pre-clinical stage of dementia might be more beneficial in preventing cognitive deterioration than at the later stage [2, 3]. Moreover, implementation of such disease modifying preventive measures requires identification of the at-risk population at the earliest. Several modifiable cardiovascular risk factors such as hypertension, physical inactivity, obesity, hyperlipidemia, and diabetes have been identified as major contributor to age associated cognitive decline and dementia [4, 5]. Given the close association of cardiovascular disease (CVDs) and dementia, a risk score known as Cardiovascular risk factors, Aging and Incidence of Dementia (CAIDE) score was developed [6]. The CAIDE dementia score consists of multifactorial risk assessment using midlife risk factors and has been validated to predict the risk of late-life dementia [7]. Furthermore, the effectiveness of the CAIDE dementia score has been tested with well-established cardiovascular risk score such as Framingham cardiovascular risk algorithms to predict future cognitive decline [8].
Past studies have effectively used the CAIDE dementia score to screen high-risk population [6–9]. In addition, a large scale randomized control trial (RCT), The Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER) study, suggested that multi-domain intervention (consisting of diet, exercise, cognitive training, and vascular monitoring) techniques might improve or maintain cognitive functions in a high-risk population screened by the CAIDE dementia score [10]. Significant beneficial multi-domain intervention effect for the cognitive outcomes of executive functioning and processing speed was observed in the study. The results of the FINGER study has been further supported by our findings from the Multi-domain Alzheimer Preventive Trial (MAPT) [11]. Result of the MAPT study showed multi-domain intervention (that included nutritional counseling, physical exercise advice and cognitive stimulation) alone or in combination with Omega-3 polyunsaturated fatty acid (n-3PUFA) may slow cognitive decline in the subjects with high CAIDE dementia score (i.e., ≥6) compared to the placebo. However, the primary outcome of the MAPT study was the three year change in a pre-clinical Alzheimer cognitive composite score [12], a Z score of four separate neuropsychological tests, and intervention effects on the separate cognitive domains were not explored in the subgroups. Further identification of the effect of interventions on particular cognitive functions could facilitate in designing domain specific (i.e., more personalized) preventive studies for preventing dementia.
Therefore, in this study we further elaborate in detail the long-term efficacy of a multi-domain intervention, with or without n-3 PUFA supplementation, on changes in cognitive function, as measured by various neuropsychological tests, in an older population with CAIDE dementia risk score≥6 in the MAPT study.
MATERIALS AND METHODS
Multi-domain Alzheimer Disease Preventive Trial
Details of the MAPT study have been described before [11, 13]. In brief, MAPT was a four arm RCT that investigated the efficacy of isolated supplementation with n-3 PUFA, multi-domain intervention (consisting of nutritional counseling, physical exercise advice, cognitive stimulation) plus placebo, or multi-domain in combination with n-3 PUFA, versus control on cognitive function of community dwelling older adults aged 70 or above. The primary outcome was assessed by the thirty-six months change in cognitive composite score from baseline in the four intervention groups. The composite score included four cognitive tests: Free and total recall of the Free and Cued Selective Reminding test (FCSRT) [14], Mini-Mental State Examination (MMSE) [15] Orientation subset, Digit Symbol Substitution Test (DSST) [16], and Category Naming Test (CNT) [17].
The MAPT study is registered on a public-access clinical trial database (www.clinicaltrials.gov, Number: NCT01513252). Written consent was obtained from all participants. The trial protocol was approved by the French Ethical Committee located in Toulouse (CPP SOOM II) and was authorized by the French Ministry of Health.
Study population
A total of 1680 subjects were randomized and followed for 3 years in the MAPT study. The inclusion criteria of the MAPT population were any of the following: 1) subjective memory complaint, 2) limitation in at least one instrumental activity of daily living (IADL) [18], 3) slow gait speed (i.e.,≤0.8 m/s or 5 s to walk 4 meters). Exclusion criteria included were: 1) subjects with dementia and MMSE [15] score less than 24, 2) limitation in basic activities of daily living (ADL score less than 6) [19], 3) with severe depression (assessed by Geriatric depression scale or GDS) [20], 4) intake of omega-3 supplementation within 6 months. For this current study, 1,293 participants with high cardiovascular risk factors, aging and incidence dementia score (i.e., CAIDE dementia risk score≥6) were selected from the modified ITT population (i.e., with at least one composite score after baseline).
Randomization and masking
Subjects were randomized to one of the four MAPT intervention groups (1:1:1:1 ratio) using computer-generated randomization procedures stratified by centers. Allocation concealment was guaranteed. The subject-level data was un-blinded before the analyses, at the end of the data management.
Assessment of Cardiovascular risk factors, Aging and Incidence of Dementia (CAIDE) score
The CAIDE dementia risk score was based on age, sex, education, systolic blood pressure, body mass index (BMI), hypercholesterolemia, and physical activity [6, 7]. Points were given according to the standard operating protocols to measure the risk scores assessed at baseline. The participants with CAIDE dementia risk score of≥6 were taken as high-risk subjects for future dementia.
Cognitive tests
The following cognitive tests were included to assess the efficacy of intervention from baseline to thirty-six months: 1) FCSRT [14] for verbal episodic memory, (use of two different word lists in alternative visits to avoid learning effects): further stratified into- free recall(score range 0–48), total recall(score range 0–48), free and total recall (score range 0–96), delayed free recall(score range 0–16), delayed total recall (score range 0–16), 2) MMSE [15] for global cognition: stratified into MMSE total (score range 0–30), and MMSE orientation (score range 0–10), 3) DSST of the Wechsler Adult Intelligence Scale-Revised (WAIS-R) (number of symbols/90 seconds) [16] for attention and processing, 4) CNT (number of words/2 minutes) [17] for categorical fluency, 5) Trail Making Test (TMT) (time to complete in seconds) [21]: stratified into TMT A for attention, visual tracking and processing, and TMT B for attention, executive function, 6) Controlled Oral Word Association Test (COWAT) [17] (number of words/2 minutes) for attention, verbal fluency and executive function. Results from above cognitive tests at baseline 6, 12, 24, and 36 months were included in this study.
Statistical analysis
For the baseline characteristics, data is expressed as mean (±standard deviation) or as frequencies/percentages. We used linear mixed-model repeated-measures (MMRM) analyses to assess within group (i.e., 4 intervention groups) changes in various cognitive test scores and the differences between intervention groups compared to placebo from baseline to 36 months. These linear mixed-models included a center-specific random intercept (when this parameter was significant) to take into account the intra-center correlation and a subject-specific random intercept and slopes to take into account the intra-subject correlation. For each model we included the following fixed effects: intervention group, time, interaction between intervention group and time. Time was treated as a continuous variable and we included in the model time² and time3 terms when the trajectory forms were not linear. For each model, we presented means estimates and standard errors (SE) and statistical significance was set at a p value < 0.05. No adjustment for potential confounders was made as there were no differences between the four intervention groups at baseline. Since the study was an exploratory analysis we did not perform statistical correction for multiple testing. All analyses were performed using SAS software version 9.4 (SAS Institute Inc, Cary, NC).
RESULTS
Baseline characteristics of the 1,293 subjects with CAIDE dementia score≥6 are shown in Table 1. The mean age at baseline was 75.39 years (sd 4.32), more than 60% of the participants were female, less than 25% had low education, the subjects had a mean supine systolic blood pressure of approximately 143.12 (sd 19.94) mmHg and diastolic blood pressure of 79.37 (sd 11.75) mmHg, the mean BMI of 26.40 (sd 4.13)Kg/m2, almost 40% had hypercholesterolemia (using clinical cut-off) and approximately 24% were physically inactive (<twice per week, 20–30 min. per time [6]). No significant differences were observed between groups by any demographic or clinical characteristics.
Baseline characteristics of subjects with CAIDE dementia score≥6
N, total number = frequency; SD, standard deviation; APOE-4, apolipoprotein E-4 allele; IADL, Instrumental Activities of Daily Livings; SPPB, Short Physical Performance Battery; MMSE, Mini-Mental State Examination; GDS, Geriatric Depression Scale; ADCS-ADL, Alzheimer’s Disease Cooperative Study Alzheimer’s Disease Cooperative Study- Activities of Daily Living; FCSRT, Free and Cued Selective Reminding Test; DSST, Digit Symbol Substitution Test; WAIS-R, Wechsler Adult Intelligence Scale; COWAT, Controlled Oral Word Association Test; CAIDE, Cardiovascular Risk Factors, Aging, and Incidence of Dementia; Physically inactive: Leisure time physical activity <twice/week, lasting minimum of 20–30 min/time.
Three years within and between group changes in cognitive function
**p < 0.001, *p < 0.05, Abbreviations: SE, Standard Error; FCSRT, Free and Cued Selective Reminding Test; MMSE, Mini-Mental State Examination; GDS, Geriatric Depression Scale; DSST, Digit Symbol Substitution Test; WAIS-R, Wechsler Adult Intelligence Scale; CNT, Category Naming Test; COWAT, Controlled Oral Word Association Test; TMT, Trail Making Test.
Thirty-six months within group and between group changes in cognitive functions according to various neuropsychological tests are shown in Table 2.
While comparing the within group changes: significant improvement was observed in the Free recall test of the FCSRT within all intervention groups. All intervention groups except n-3PUFA were found to significantly improve in free and total recall test and the delayed free recall test of the FCSRT. Multi-domain intervention alone was found to have significant improvement in the DSST and COWAT tests. Multi-domain combined with n-3PUFA intervention groups showed significant improvement in the TMT-A.
Compared to the placebo, there was significantly less improvement in the FCSRT free recall test at 36 months in the n-3 PUFA intervention group (mean±SE = – 1.01±0.46), showing a significant improvement in the placebo group and not in the n-3 PUFA intervention group. The mean difference at 36 months between the placebo group and multi-domain combined with n-3 PUFA intervention group was found to be significant in the delayed total recall test (mean±SE = 0.20±0.10) and MMSE orientation test (mean±SE = 0.15±0.06). There was significantly less decline in the MMSE orientation test (mean±SE = – 0.12±0.06) in the isolated multi-domain intervention group compared to the placebo at 36 months.
DISCUSSION
Dementia risk scores based on cardiovascular risk factors have been well documented to predict future dementia and could facilitate in designing preventive studies [7, 10]. Effect of interventions on separate cognitive domain of such high-risk population should be well understood for designing more effective intervention studies. Past clinical trials have shown some single-domain interventions with physical exercise or cognitive training to have beneficial effects on cognition in older adults [22–25]. Similarly, some studies show supplementation with n-3PUFA [26, 27] and intervention with suitable diet pattern [28, 29] to have positive effect in cognition.
In the MAPT study, subjects with high CAIDE dementia score of≥6 at baseline were found to demonstrate lower cognitive decline when treated with multi-domain intervention alone or a combined with n-3PUFA for 3 years compared to the placebo [11]. In this study, we further elaborated the findings of the MAPT trial stratifying the cognitive composite-score into multiple neuropsychological tests. Compared to the placebo, group with isolated multi-domain intervention was found to decline less in the MMSE orientation. Group with multi-domain intervention in combination with n-3PUFA was found to have improvement in the delayed total recall of the FCSRT along with the MMSE orientation. Besides, evidence from past studies have shown the effects of n-3PUFA on cognitive decline to be mixed, some RCTs showing no effects on cognitive decline [30, 31], while other have shown some beneficial effects [26, 27], including improved immediate and delayed verbal recognition memory scores, but not in the working memory or executive function tests [27].
In the FINGER study the authors demonstrated the beneficial effects of multi-domain intervention (i.e., combination of diet, exercise, cognitive training and vascular risk monitoring) in older population with CAIDE dementia score≥6 [10]. Significant beneficial effect of the multi-domain intervention was reported for the primary outcome of the study (i.e., neuropsychological test battery (NTB) total Z score), as well as NTB executive functioning and NTB processing speed, compared to the control group at 24 months. Compared to the MAPT study, the study population of the FINGER study was relatively younger (mean age 69 versus 75 years), the multi-domain intervention was more intensive, although the intervention period was only two years, and the inclusion of vascular monitoring in the multi-domain intervention might have potentially added to the beneficial effects of interventions. However, no direct significant effect of the multi-domain intervention was observed in the memory domain (besides reporting of post-hoc analyses showing significant effect on much complex memory– abbreviated memory score) [10]. Besides, in our study no significant effect of interventions was observed on tests of processing speed and executive function domain compared to the placebo group. One of the reason could be our study population was comparatively older than the FINGER study [11, 32], who tend to have decreased deficits of these functions [33, 34]. Nevertheless, our findings show multi-domain intervention combined with n-3PUFA for 36 months to have improvement in the delayed total recall of the FCSRT and the MMSE orientation compared to the placebo group. Isolated multi-domain intervention group was found to have improvement in the MMSE orientation. Orientation as assessed by MMSE has been long considered as the earliest function to be impaired in older subjects with AD [35, 36]. Similarly, impairment in episodic memory is also known to be one of the early signs in AD and delayed recall has been implied to be a most sensitive measure of the memory impairment in early AD [37, 38]. The findings from our study suggest that n-3PUFA in combination with multi-domain intervention might have comparatively positive effect in the early AD cognitive deficits such as delayed recall and orientation than the multi-domain or n-3 PUFA as separate interventions. Future intervention studies are warranted to confirm with our findings.
In this study, we found a 36-month intervention with multi-domain alone to improve orientation, and multi-domain intervention in combination with n-3PUFA to improve the orientation and episodic memory domain in subjects with high CAIDE dementia risk score. However, the differences in cognitive tests were quite minimal compared to the placebo group, which at a glance might seem clinically negligible. However, the fact that such changes or intervention effects in preventive studies are mostly minimal should be considered. Furthermore, the long-term positive effects of preventive measures although very minimal, could have high relevance in the context of public health. A very minimal decrease in cognitive deterioration in large population might potentially play a significant role in reducing the global burden of dementia. Furthermore, cognitive disorders are largely impacted by cardiovascular risk factor such as physical inactivity, hence similar interventions with physical exercise not limited to advice-but include monitored physical training might further strengthen the effects. Additionally, inclusion of regular monitoring of vascular risks along with other multi-domain interventions might potentially be more effective to reduce the risks of dementia in older adults. Besides, such interventions strategies could be more effective in delaying cognitive decline, if applied to a comparatively younger old population with high cardiovascular risk factors, hence should be further investigated. Furthermore, outcomes of such studies would enable us to design more personalized intervention strategies according to the cognitive deficits, hence slowing cognitive deterioration and maintain continuous cognitive functionality in the older population at high risk of dementia.
One of the major strength of our study is the comparatively longer duration of follow-up and the wider range of intervention groups than the FINGER study. The key limitation of this study is the fact that it is a post-hoc analysis of a larger RCT.
The focus of this study, was to investigate the effect of a multi-domain intervention with or without Omega-3 polyunsaturated fatty acid on cognitive functions in high risk older population for dementia using CAIDE dementia score. Cognitive impairments in the orientation and delayed recall domain are known to be some of the early cognitive deficits of AD, and such population might benefit from intervention strategies as in the MAPT study.
