Author + information
- Received November 27, 2019
- Revision received March 11, 2020
- Accepted March 17, 2020
- Published online May 18, 2020.
- Ruixue Song, MSca,b,c,
- Hui Xu, MSca,b,c,
- Christina S. Dintica, MScd,
- Kuan-Yu Pan, PhDe,
- Xiuying Qi, PhDa,b,c,∗∗ (, )
- Aron S. Buchman, MDf,
- David A. Bennett, MDf,∗ and
- Weili Xu, PhDa,b,c,d,∗∗ (, )@weili_xu
- aDepartment of Epidemiology and Biostatistics, School of Public Health, Tianjin Medical University, Tianjin, China
- bTianjin Key Laboratory of Environment, Nutrition and Public Health, Tianjin, China
- cCenter for International Collaborative Research on Environment, Nutrition and Public Health, Tianjin, China
- dAging Research Center, Department of Neurobiology, Care Sciences and Society, Karolinska Institutet, Stockholm, Sweden
- eAmsterdam UMC, Vrije Universiteit, Psychiatry, Amsterdam Public Health Research Institute, Amsterdam, the Netherlands
- fRush Alzheimer’s Disease Center, Rush University Medical Center, Chicago, Illinois
Background The impact of cardiovascular risk burden on cognitive trajectories and brain structure changes remains unclear.
Objectives This study aimed to examine whether cardiovascular risk burden assessed by the Framingham General Cardiovascular Risk Score (FGCRS) is associated with cognitive decline and structural brain differences.
Methods Within the Rush Memory and Aging Project, 1,588 dementia-free participants (mean age: 79.5 years) were followed for up to 21 years. FGCRS was assessed at baseline and categorized into tertiles (lowest, middle, and highest). Episodic memory, semantic memory, working memory, visuospatial ability, and perceptual speed were assessed annually with a battery of 19 tests, from which composite scores were derived. A subsample (n = 378) of participants underwent magnetic resonance imaging. Structural total and regional brain volumes were estimated. Data were analyzed using linear mixed-effects models and linear regression models.
Results In all participants, FGCRS ranged from 4 to 28 (mean score: 15.6 ± 3.7). Compared with the lowest tertile of FGCRS, the highest tertile was associated with faster decline in global cognition (β = −0.019; 95% confidence interval [CI]: −0.035 to −0.003), episodic memory (β = −0.023; 95% CI: −0.041 to −0.004), working memory (β = −0.021; 95% CI: −0.035 to −0.007), and perceptual speed (β = −0.027; 95% CI: −0.042 to −0.011) over the follow-up. In magnetic resonance imaging data analyses, higher FGCRS was related to smaller volumes of the hippocampus (β = −0.021; 95% CI: −0.042 to −0.000), gray matter (β = −1.569; 95% CI: −2.757 to −0.382), and total brain (β = −1.588; 95% CI: −2.832 to −0.344), and greater volume of white matter hyperintensities (β = 0.035; 95% CI: 0.001 to 0.069).
Conclusions Higher cardiovascular risk burden may predict decline in episodic memory, working memory, and perceptual speed and is associated with neurodegeneration and vascular lesions in the brain.
- cognitive decline
- cohort study
- Framingham General Cardiovascular Risk Score
- magnetic resonance imaging
- vascular lesions
Dementia is among the greatest public health challenges to modern societies worldwide (1). In 2017, almost 50 million people had dementia, and the number is projected to reach 82 million by 2030, according to the World Health Organization (2). Considering the lack of an effective treatment for dementia, identifying modifiable risk factors for cognitive decline, which may lead to dementia, has become a prominent strategy to prevent or delay dementia onset (3).
Traditional cardiovascular risk factors, including obesity, diabetes, smoking, hypertension, and hypercholesterolemia, have been individually associated with faster cognitive decline (4–6). It is well known that cardiovascular risk factors are correlated with each other, making it difficult to isolate their individual effects on cognitive decline. The Framingham General Cardiovascular Risk Score (FGCRS), which incorporates demographics (i.e., age and sex) with traditional cardiovascular risk factors, was developed to assess cardiovascular risk burden and the risk of developing cardiovascular disease (CVD) in individuals (7). To date, a number of studies have investigated the association between FGCRS and cognitive decline (8–14), but with inconsistent findings. About one-half of these studies focused only on global cognitive function evaluated by the Mini-Mental State Examination, Modified Mini-Mental State Examination, or Preclinical Alzheimer Cognitive Composite (10,12,13). Other studies examining specific cognitive domains showed an association between higher FGCRS and faster decline in memory, executive function, and verbal fluency (8,9,11,14). However, 1 study found that FGCRS was not associated with episodic memory, semantic memory, or visuospatial ability (14). The assessments of cognitive domains in these studies were limited to 1 or 2 tests.
Previous studies have reported that cardiovascular risk factors were related to smaller volumes of specific brain regions, such as white matter, gray matter, and hippocampus, but with some inconsistent findings (12,15–19). Furthermore, few studies have examined the relationship between cumulative measurements of combined vascular risk factors with cognitive function and brain MRI measures in a cohort with long-term follow-up. In the present study, we systematically examined the longitudinal association between cardiovascular risk burden assessed by FGCRS and decline in global and domain-specific cognitive function and further investigated whether cardiovascular risk burden is associated with total and regional structural brain volumes.
The Rush Memory and Aging Project (MAP) is an ongoing, longitudinal study on aging and dementia (20). Details regarding the MAP study design and the evaluation protocol have been provided previously (20). In brief, participants were recruited from continuous care retirement communities, senior and subsidized housing, church groups, and social service agencies in northeastern Illinois. From 1997, a total of 2,155 participants were annually followed for up to 21 years until 2019. Starting in 2009, 401 participants were invited to undergo MRI scans. Because MAP is an open cohort, the MRI scans could be performed at study entry or any follow-up examinations. Of the total population, 567 persons with prevalent dementia (n = 115), missing FGCRS at study entry (n = 311), and missing data on cognitive function during the follow-up (n = 282) were excluded, leaving 1,588 participants in the study sample. Among them, 378 participants had data on MRI (Supplemental Figure 1).
Written informed consent was obtained from all participants at baseline as well as a repository consent, which allowed the data to be shared. The study was approved by an institutional review board of Rush University Medical Center and was performed in accordance with the ethical standards laid out in the 1964 Declaration of Helsinki and its later amendments.
At study entry and thereafter, all participants underwent a comprehensive clinical assessment, neurologic examination, and extensive cognitive testing conducted by trained staff (21). Information on demographic characteristics, socioeconomic status, and lifestyle factors was also collected (21).
Education was recorded according to years of schooling. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared. Alcohol consumption was measured by the average amount (in grams) consumed per day during the past year. Physical activity was assessed by the total hours of participation per week, based on the National Health Interview Survey. Smoking was categorized as never, former, or current smoker.
Blood pressure was measured twice in the sitting position with a 5-min interval using a mercury sphygmomanometer. The mean of the 2 values was recorded. Hypertension was defined as systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg or the use of antihypertensive drugs. Diabetes was defined by any of the following criteria: hemoglobin A1c ≥6.5%, fasting plasma glucose ≥126 mg/dl, random blood glucose ≥200 mg/dl, history of diabetes, or the use of diabetes medication (22). Heart disease was ascertained based on self-report. Clinical stroke diagnosis was made by clinicians through review of self-report and neurological examination and was dichotomized as probable versus not present. Blood samples were taken, and total cholesterol and high-density lipoprotein (HDL) cholesterol were measured with a lipid panel. In addition, apolipoprotein E (APOE) allele genotype was assessed by Polymorphic DNA Technologies (Alameda, California) and was categorized as ε4 carrier or noncarrier. Additional details about the data collection can be found online on the Rush Alzheimer's Disease Center Resource Sharing Hub.
Framingham General Cardiovascular Risk Score
We calculated FGCRS based on age, sex, smoking, systolic blood pressure, medication for hypertension, total cholesterol, HDL cholesterol, and diabetes for each participant at baseline (as shown in Supplemental Tables 1 and 2) (7). FGCRS is obtained by summing up the points from all of these risk factors, and the score was further categorized into tertiles as the lowest, middle, and highest. A higher FGCRS indicates a greater risk of future cardiovascular events. Missing data on total cholesterol and HDL cholesterol at baseline (n = 340) were imputed by using data within 5 years if the participants had no dementia.
Assessment of cognitive domains, mild cognitive impairment, and dementia
A cognitive battery of 21 tests was administrated at baseline and at each follow-up. One test, the Mini-Mental State Examination, was used to describe the cohort but not in analyses. A second test was used only for diagnostic classification. Episodic memory was evaluated using Word List Memory, Word List Recall, Word List Recognition, as well as immediate and delayed recall of the East Boston Story and of Story A from Logical Memory of the Wechsler Memory Scale–Revised. Semantic memory was assessed by the Boston Naming Test, Verbal Fluency, and the National Adult Reading Test. Working memory was evaluated with 3 measures: Digit Span Forward and Digit Span Backward from the Wechsler Memory Scale–Revised and Digit Ordering. Perceptual speed was tested with 4 measures: the oral version of the Symbol Digit Modalities Test, Number Comparison, and 2 indices from a modified Stroop Neuropsychological Screening Test. Two measures of visuospatial ability were administered: a 15-item form of the Judgment of Line Orientation and a 16-item form of the Standard Progressive Matrices. Scores of all of the tests were converted to z-scores, which were further averaged to yield a composite score for global cognitive functioning, as reported in detail in a previous study (23,24). Higher scores indicate better cognitive function.
Ascertainment of dementia and mild cognitive impairment (MCI) was conducted by following a standard procedure, including computing scores of cognitive tests, clinical judgment by a neuropsychologist, and diagnosis made by a clinician (25,26). Clinical diagnosis of dementia was based on criteria of the joint working group of the National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer’s Disease and Related Disorders Association (26). MCI was diagnosed by experienced neuropsychologists and clinicians and was defined as presence of cognitive impairment without meeting the criteria for dementia (25).
Brain markers on MRI
High-resolution T1-weighted anatomic data were obtained from a 1.5-T GE MRI scanner (General Electric, Waukesha, Wisconsin) using a 3-dimensional magnetization-prepared rapid acquisition gradient echo sequence with the following parameters: time of echo, 2.8 ms; time of repetition, 6.3 ms; preparation time, 1,000 ms; flip angle, 8°; field of view, 24 × 24 cm; 160 slices; 1-mm slice thickness; 224 × 192 image matrix reconstructed to 256 × 256; 2 repetitions (27). MRI data were automatically segmented with FreeSurfer (an open source software suite for processing and analyzing [human] brain MRI images). When necessary, manual intervention was used to increase the accuracy of labeling. Total volumes (adding left and right sides) were calculated and converted from cubic millimeters to tenths of percentage of intracranial volume (using the estimate of intracranial volume from FreeSurfer, version 5.0). In this study, the volumes of hippocampus, white matter hyperintensities (WMH), cerebellum white matter, cortical white matter, total white matter, cortical gray matter, cerebellum gray matter, subcortical gray matter structures, total gray matter, and total brain were obtained. The units for the volume were cubic centimeters. For WMH, the volume was transformed by taking the logarithm because of its skewed distribution.
Baseline characteristics of the study population by FGCRS tertiles were compared using chi-square tests for categorical variables and 1-way analysis of variance or Wilcoxon rank sum tests for continuous variables.
Linear mixed-effects models were used to estimate β-coefficients and 95% confidence intervals (CIs) for the associations between cardiovascular risk burden (i.e., continuous and categorical FGCRS) and annual change in global cognitive function and 5 cognitive domains, with follow-up time (in years) as the time scale. The fixed effect included cardiovascular risk burden, follow-up time, and their interaction. The random effect included random intercept and slope, allowing the individual differences at baseline and across follow-up. Age, sex, education, BMI, stroke, heart disease, alcohol consumption, physical activity engagement, and APOE ε4 were adjusted for as potential confounders in the multiadjusted model. To further explore the role of APOE ε4 in the association between FGCRS and cognitive function, an interaction term between FGCRS categories and APOE ε4 status was also included in the models first; then, stratified analysis by APOE ε4 was performed.
In MRI analyses, linear regression was used to estimate β-coefficients and 95% CIs for the relationship between cardiovascular risk burden using continuous FGCRS and brain volumes. Mixed-effects models were used to estimate the association between brain volumes and cognitive functions. All models were adjusted for age, sex, education, BMI, stroke, heart disease, physical activity engagement, alcohol consumption, global cognitive function, and APOE ε4 as potential confounders. In the sensitivity analysis, we excluded 385 individuals with MCI at baseline and reran the linear mixed-effects model. All statistical analyses were performed using Stata SE 15.0 for Windows (StataCorp, College Station, Texas). To broadly investigate the association between cardiovascular burden and cognitive decline, multiple comparisons were not mathematically corrected for to reduce the chance of type II error. A 2-tailed p value <0.05 was considered to be statistically significant for all tests.
Characteristics of the study population at baseline
Among the 1,588 dementia-free participants (mean age: 79.5 ± 7.5 years; 75.8% female), FGCRS ranged from 4 to 28 (mean score: 15.6 ± 3.7). Considering the sample size in each group, the FGCRS was tertiled as lowest, middle, and highest. Of all participants during the follow-up (median duration: 5.8 years; range 1 to 21 years), 454 (28.6%) had the lowest cardiovascular risk burden, 475 (29.9%) were in the middle tertile, and 659 (41.5%) belonged to the highest tertile. The median (interquartile range [IQR]) of follow-up time among the lowest, middle, and highest risk groups were 5.99 (2.82 to 9.05), 6.00 (2.84 to 9.06), and 5.02 (2.83 to 8.31), respectively. There was no significant difference among these 3 groups (p = 0.114). Participants with the highest burden had poorer global cognitive function and lower scores of episodic memory, semantic memory, and perceptual speed. No significant difference was observed in education, alcohol consumption, total cholesterol, smoking, APOE ε4, and stroke (Table 1).
Relationship between FGCRS and cognitive decline
In the mixed-effects model where FGCRS was treated as a continuous variable, higher score was associated with worse baseline global cognitive function, semantic memory, visuospatial ability, and perceptual speed, and a faster rate of decline in global cognition, episodic memory, working memory, and perceptual speed (Table 2). When FGCRS was used as tertiles, the highest FGCRS was associated with faster global cognitive decline compared with the lowest FGCRS. Moreover, participants with the highest FGCRS had accelerated decline in episodic memory, working memory, and perceptual speed over the follow-up, compared with those with the lowest FGCRS (Table 2, Central Illustration).
We found no statistical interaction between FGCRS and APOE ε4 with respect to cognitive decline (all p values >0.05). However, in the stratified analysis by APOE ε4, the associations between higher FGCRS and faster decline in global cognition and different cognitive domains were present mainly among APOE ε4 noncarriers (Figure 1, Supplemental Tables 3 and 4).
Cardiovascular risk burden, brain MRI markers, and cognitive function
In linear regression analysis for MRI data, higher FGCRS was associated with smaller volumes of the hippocampus, total gray matter, cerebellar gray matter, cortical gray matter, subcortical gray matter, total brain, and larger volume of WMH after adjustment for potential confounders (Table 3). In addition, we found that episodic memory (β = 0.080; 95% CI: 0.060 to 0.101) and working memory (β = 0.054; 95% CI: 0.039 to 0.069) were related to hippocampal volume, whereas perceptual speed was associated with WMH (β = −0.026; 95% CI: −0.038 to −0.014).
Further exclusion of incident MCI cases (n = 385) at baseline produced similar results to those from the main analyses (Supplemental Table 5). Considering the possibility of nonlinear association between FGCRS and cognitive trajectories, we repeated the analysis using a quadratic mixed-effects model and mixed-effects model with splines, which showed no evidence of quadratic curve and splines (data not shown).
In this community-based prospective study, we found that: 1) increased cardiovascular risk burden assessed by FGCRS was associated with accelerated decline during 20 years of follow-up in global cognition and specific cognitive domains, including episodic memory, working memory, and perceptual speed; and 2) higher FGCRS was related to smaller volumes of hippocampus, gray matter, and total brain, as well as larger volume of WMH (Central Illustration).
FGCRS is an aggregated measure of multiple cardiovascular risk factors and a validated predictor of future CVD (7). In the present study, we found that FGCRS was associated with a faster rate of decline in global cognitive function, in line with findings from previous studies (8–13). Furthermore, several studies have shown an association between higher FGCRS and faster cognitive decline in memory, executive function, and verbal fluency (8,9,11). Another study indicated that higher FGCRS was related to lower performance in executive function, but there was no significant association of FGCRS with episodic memory, semantic memory, or visuospatial ability (14). Because detailed measures of cognitive function in specific domains were not widely available, most studies assessed them with only 1 or 2 tests, and there are methodological discrepancies due to difficulties in measuring cognitive function of the specific domains in detail. In this study, we created composite measures of 5 different cognitive domains, including episodic memory, semantic memory, working memory, perceptual speed, and visuospatial ability, by converting the raw scores of 19 tests, which can reduce random variability and floor and ceiling artifacts (23). We found that higher cardiovascular risk burden was associated with faster decline in episodic memory, working memory, and perceptual speed. Cognitive performance in different domains may indicate specific structural brain changes. Episodic memory and working memory decline may reflect hippocampal atrophy (28–30). Poor performance in processing speed was linked to white matter lesions, such as WMH (31,32). In addition, working memory function may also be reflected by WMH in the frontal deep white matter regions (33).
Decreased hippocampal and gray matter volumes are typical markers of Alzheimer dementia-related neurodegeneration (34). The presence of WMH may indicate microvascular lesions in cerebral white matter (35). Smaller total brain volume is one of the major contributors to vascular cognitive decline and may reflect vascular brain alterations (35). Thus, the findings on the association of higher FGCRS with cognitive decline in episodic memory, working memory, and perceptual speed suggest that higher cardiovascular burden might have effects on both neurodegenerative and vascular changes.
To support these results, in the second part of this study on MRI data, we found that higher FGCRS was associated with smaller volumes of hippocampus, cortical gray matter, and total brain but with a greater WMH volume. Indeed, we also found that episodic memory and working memory were related to hippocampal volume, whereas perceptual speed was associated with WMH in our study. Thus, the 2 parts of this study are complementary to each other and provide stronger evidence to support our conclusions.
Neuroimaging studies have shown that specific cardiovascular risk factors, such as alcohol use, diabetes, and cardiac left ventricular hypertrophy, were associated with an increased volume loss in hippocampus, total grey matter, and total brain, respectively (15–17). One study reported that higher FGCRS was related to smaller hippocampus (12). However, studies focusing on the association between individual cardiovascular risk factors and white matter volume are scarce and found inconsistent results. Another study found that dysglycemia was related to smaller white matter volume (18), whereas another study showed that fasting and 2-h post-load plasma glucose parameters were not related to white matter volume (19). To our knowledge, only 1 study has examined the association of cardiovascular risk burden with both cognitive function and brain MRI measures (14). That study found that higher FGCRS was related to lower executive function performance and greater WMH, which would result in white matter tissue loss and atrophy due to demyelination and axonal loss (35).
It is well known that APOE ε4 is a strong predictor of cognitive decline and dementia. Former studies have reported that the APOE ε4 allele was significantly associated with faster cognitive decline (4,13). In the current study, we found that the association between FGCRS and cognitive decline differed between APOE ε4 carriers and noncarriers. Specifically, the relationship between increased FGCRS and faster cognitive decline was present mainly among APOE ε4 noncarriers. However, the interaction between FGCRS and APOE ε4 status on cognitive decline was not significant. There might be several explanations for this finding. First, the smaller sample size of APOE ε4 carriers may lead to a lack of power. Second, because APOE ε4 carriers are more likely to develop dementia, they might have been excluded at baseline.
There are several pathophysiological pathways that may explain the observed associations here. Studies showed that high blood pressure or hypertension, obesity, and diabetes may be associated with global and regional brain atrophy (36–38). High cholesterol has been associated with increased production of β-amyloid, and the presence of the APOE ε4 allele may lead to hypercholesterolemia (39). Furthermore, smoking is correlated with inflammation and oxidative stress (40). In addition, cardiovascular risk factors have divergent effects on blood flow velocity, which may contribute to cerebral pathology (41). Exposure to cardiovascular risk factors might accelerate cognitive decline by causing cerebral hypoperfusion, hypoxia, emboli, or infarcts, which lead to vascular and degenerative brain lesions (42,43). Neuropathological studies have also shown that CVD and dementia share similar pathogenetic processes related to cardiovascular risk factors, and CVD may lower the threshold for cognitive impairment (44–46).
Because of the lack of effective treatments for dementia, modifiable factors have drawn special attention for dementia prevention. Indeed, it has been shown that treating and controlling chronic diseases, such as diabetes, hypertension, and hyperlipidemia, are effective for the prevention of cognitive impairment (47–49). Our study highlights the importance of control vascular risk burden in healthy brain aging. Given the progressive increase in the number of dementia cases worldwide, our findings have both clinical and public health relevance.
Study strengths and limitations
There are several strengths of this study. First, this was a community-based, longitudinal study with a relatively large sample and long follow-up time. Second, data on cardiovascular risk factors were integrated to generate a composite score, and comprehensive assessments on specific cognitive domains were carried out by using a battery of 19 tests. In addition, we explored potential mechanisms linking cardiovascular risk burden to cognitive function using MRI data.
However, the study also has limitations. First, participants were volunteers from the community, which may limit the generalizability of our findings. Second, participants were generally well educated and performed relatively well on cognitive tests; thus, the observed association may have been an underestimation. Third, the association between FGCRS and MRI measures was examined based on cross-sectional data; thus, the temporality for such association is unclear. Finally, potential confounding due to unmeasured factors cannot be completely ruled out.
In this population-based cohort study of dementia-free older adults, we found that higher cardiovascular risk burden assessed by FGCRS accelerates cognitive decline in episodic memory, working memory, and perceptual speed. Moreover, higher cardiovascular risk burden is associated with markers of neurodegeneration and vascular lesions in the brain. Our findings highlight the need to monitor and control cardiovascular burden to maintain cognitive health in late life.
COMPETENCY IN MEDICAL KNOWLEDGE: The Framingham General Cardiovascular Risk Score, which incorporates demographics and cardiovascular risk factors to assess cardiovascular risk, is also associated with vascular lesions in the brain and markers of neurodegenerative disease.
TRANSLATIONAL OUTLOOK: Future studies should explore the mechanisms linking cardiovascular risk, neurodegenerative disease, and cognitive decline.
The authors express their gratitude to the participants and staff involved in data collection and management in the Rush Memory and Aging Project.
↵∗ Drs. Bennett and Xu contributed equally to this work.
This project is part of CoSTREAM and received funding from the European Union’s Horizon 2020 research and innovation program under grant agreement no. 667375. The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. Dr. Bennett has received grants from the National Institutes of Health (R01AG17917 and UH2NS100599). Dr. Xu has received grants from the Swedish Research Council (no. 2017-00981), the National Natural Science Foundation of China (no. 81771519), Demensfonden, the Konung Gustaf V:s och Drottning Victorias Frimurare Foundation (no. 2016-2017), and Alzheimerfonden (2017-2019). All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
The authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC author instructions page.
- Abbreviations and Acronyms
- apolipoprotein E
- body mass index
- confidence interval
- cardiovascular disease
- Framingham General Cardiovascular Risk Score
- high-density lipoprotein
- Rush Memory and Aging Project
- mild cognitive impairment
- magnetic resonance imaging
- white matter hyperintensities
- Received November 27, 2019.
- Revision received March 11, 2020.
- Accepted March 17, 2020.
- 2020 The Authors
- ↵(2018) Response to the growing dementia burden must be faster. Lancet Neurol 17:651.
- Irimata K.E.,
- Dugger B.N.,
- Wilson J.R.
- D’Agostino R.B. Sr..,
- Vasan R.S.,
- Pencina M.J.,
- et al.
- Laughlin G.A.,
- McEvoy L.K.,
- von Muhlen D.,
- et al.
- Zeki Al Hazzouri A.,
- Haan M.N.,
- Neuhaus J.M.,
- et al.
- Wang R.,
- Fratiglioni L.,
- Kalpouzos G.,
- et al.
- Rabin J.S.,
- Schultz A.P.,
- Hedden T.,
- et al.
- Aljondi R.,
- Szoeke C.,
- Steward C.,
- Gorelik A.,
- Desmond P.
- Kharabian Masouleh S.,
- Beyer F.,
- Lampe L.,
- et al.
- Pase M.P.,
- Davis-Plourde K.,
- Himali J.J.,
- et al.
- Markus M.R.P.,
- Ittermann T.,
- Wittfeld K.,
- et al.
- Bennett D.A.,
- Buchman A.S.,
- Boyle P.A.,
- Barnes L.L.,
- Wilson R.S.,
- Schneider J.A.
- American Diabetes Association
- van Leijsen E.M.C.,
- Tay J.,
- van Uden I.W.M.,
- et al.
- Debette S.,
- Markus H.S.
- Mahley R.W.
- Pase M.P.,
- Grima N.A.,
- Stough C.K.,
- Scholey A.,
- Pipingas A.
- Boccardi V.,
- Murasecco I.,
- Mecocci P.
- Wahidi N.,
- Lerner A.J.
- Theobald R.J. Jr..