Do Cardiometabolic Risk Factors Influence Amyloid,Tau,and Neuronal Function in APOE4 Carriers and Non-Carriers in Alzheimer’s Disease Trajectory?

Abstract

Cardiovascular risk could be calculated using Qrisk2. It is suggested that cardiovascular risk factors influence the progression of Alzheimer’s disease (AD). However, studies have not specifically evaluated the influence of cardiovascular risk using Qrisk2 on neuropathological progression and AD biomarkers. The aim of the study was to evaluate the influence of cardiovascular risk factors using Qrisk2 on CSF amyloid-β (Aβ) and tau, ¹⁸F-AV45-PET, ¹⁸F-FDG-PET, MRI, and cognitive measures in APOE4 negative cognitively normal and mild cognitive impairment (MCI) subjects. 614 cognitively normal, early, and late MCI subjects were selected from the ADNI cohort with a 2-year follow-up. CSF Aβ and tau, ¹⁸F-AV45-PET, ¹⁸F-FDG-PET, MRI, and cognitive measures along with modified Qrisk2 were evaluated. APOE4 non-carrier, high cardiovascular risk sub-group of early and late MCI and cognitively normal subjects, demonstrated worse biomarker and cognitive profile at baseline and during follow up compared to low cardiovascular risk group. Additionally, similar pattern was also observed in APOE4 carriers. We demonstrated that Qrisk2 and APOE4 were independent predictors of biomarker and clinical progression in AD trajectory. High cardiovascular risk is associated with biomarker changes in APOE4 non-carriers in prodromal AD, which may suggest that treatment of cardiovascular risk is an effective prevention strategy even in APOE4 negative subjects and may influence disease progression independent of amyloid pathology. Demonstration of accelerated neuropathological changes in both APOE4 carriers and non-carriers suggest that focusing on modifiable cardiovascular risk factors is an effective preventative strategy while we eagerly waiting for new treatments.

Keywords

Alzheimer’s disease APOE4 biomarkers cardiovascular disease risk factors

INTRODUCTION

Sporadic Alzheimer’s disease (AD) constitutes over 95% of cases and is thought to result from multifactorial interaction of genetic and environmental factors. Apolipoprotein E (APOE) remains the major genetic risk factor for sporadic AD [1], although newly identified genetic risk loci also contributes to AD risk. However, it is suggested that modifiable risk factors like diabetes, midlife hypertension, midlife obesity, physical inactivity, depression, and smoking play significant role in AD [2]; intervening on them is of significant importance to public health. To date there is no drug effectively altering the progression of AD [3]. While the search for novel therapeutic strategies continues, it is important to evaluate the influence of these modifiable risk factors on both AD progression and its neuropathological substrates (amyloid, tau, glucose metabolism, and MRI changes), to promptly pursue effective preventative and therapeutic strategies.

Qrisk2 is a cardiovascular disease (CVD) risk score to calculate an individual’s likelihood of developing CVD over the next ten years [4]. Qrisk2 has been extensively validated and evidence supports the use of this score in improving the 10-year risk prediction for patients suffering from CVD risk [5]. A Qrisk2 threshold greater than 20% to define the high-risk category has a better ability to identify those at high risk of developing CVD, compared to the Framingham risk score [6].

APOE4 is the most important risk factor for AD with a prevalence of up to 60% in people diagnosed with AD. APOE4 is also a risk factor for CVD. Studies have evaluated the relationship between cardiovascular risk factors and presence of APOE genotype on AD progression, often finding a synergistic effect of APOE4 allele and cardiovascular risk factors on AD risk [7 –11]. The exact influence of CVD risk in APOE4 non-carriers is still unknown [12, 13]. Moreover, to date, no studies have evaluated the influence of cardiovascular risk, calculated by Qrisk2, on APOE4 negative subjects on multiple AD biomarkers (amyloid, tau, glucose metabolism, cerebrospinal fluid (CSF) biomarkers) and clinical progression in cognitively normal and early and late mild cognitive impairment (MCI) subjects. Thus, we aimed to investigate the influence of cardiovascular risk defined by Qrisk2, on the progression of AD specific biomarkers in the preclinical stages of the disease on APOE4 non-carriers and compared this with that of APOE4 carriers.

MATERIALS AND METHODS

The Alzheimer’s Disease Neuroimaging Initiative (ADNI) is an ongoing international longitudinal study aimed at the identification of markers for the early detection and monitoring of AD [14, 15]. 614 participants in the diagnostic categories of early MCI (EMCI), late MCI (LMCI), and cognitively normal (CN) subjects with 2-year clinical and biomarker follow up were selected from the ADNI cohort (http://www.loni.ucla.edu/ADNI). EMCI and LMCI were defined, according to the ADNI protocol, based on abnormal memory function documented by scoring within the education adjusted ranges on the Logical Memory II subscale (Delayed Paragraph Recall, Paragraph A only) from the Wechsler Memory Scale – Revised, while all CN subjects had a Mini-Mental State Examination (MMSE) score ≥27. APOE4 positivity was defined as presence of one or more E4 allele.

CSF Aβ:p-tau ratio was calculated, with threshold for positivity set at 6.16 [16]. Quantitative analysis of amyloid PET (¹⁸F-AV45-PET) was performed using target region to cerebellar ratio, and region-of-interest uptake was estimated for posterior cingulate, lateral temporal, frontal and lateral parietal cortical regions [17]. ¹⁸F-FDG PET was used to measure cerebral glucose metabolism and was quantified using target-to-pons RATIO for each subject in the angular gyrus, bilateral posterior cingulate and temporal gyrus [18]. ADAS-Cog13 (Alzheimer’s Disease Assessment Scale 13 items) and CDR-SoB (Clinical Dementia Rating scale – sum of boxes) were used as cognitive measures [19].

Qrisk2

Qrisk2 considers age, sex, ethnicity, smoking status, diagnosis of atrial fibrillation, rheumatoid arthritis, diabetes, chronic kidney disease, whether on blood pressure treatment, and angina or a heart attack in a first degree relative under the age of 60 years. Additionally, it includes body mass index, cholesterol/HDL ratio, and systolic blood pressure. The Qrisk2-2016 web calculator was used to run a modified Qrisk2 calculation. We excluded participants with pre-existing or established CVD, as defined by relevant past medical history.

Statistical analysis

We used IBM SPSS 24.0 to analyze the data. Continuous variables were expressed as mean±SD. One-way ANOVA was used to compare the three groups followed by a Bonferroni post-hoc correction. Categorical variables were compared by χ² test. To evaluate the difference between baseline and follow-up measurements, paired t-tests were carried out for each biomarker within the subgroups and diagnostic categories. Analysis between subgroups were performed using independent samples t-test. To evaluate the influence of Qrisk2 scores on top of APOE4 on neurodegenerative markers (amyloid, glucose metabolism, neurocognitive tests), ANCOVA was performed using APOE4 genetic status as a covariate and adjusted for baseline values of each biomarker.

RESULTS

Baseline characteristics of subjects

The baseline characteristics of CN, EMCI, and LMCI subjects are shown in Table 1. 250 CN, 173 EMCI, and 191 LMCI were evaluated in the study. Qrisk2 score was significantly lower in EMCI subjects compared to the other two groups. 28.7% of CN subjects were found to be amyloid positive at baseline, consistent with other studies [20, 21]. CSF Aβ was significantly lower and CSF t-tau and p-tau were significantly higher in LMCI compared to EMCI and CN subjects. ¹⁸F-FDG uptake was significantly lower in LMCI compared to the other two groups in all the predefined regions.

Table 1

Baseline characteristics of subjects in CN, EMCI, and LMCI

	CN	EMCI	LMCI	ALL
Sample size (male)	250 (118)	173 (93)	191 (115 * )	614 (326)
Age, y (SD)	74.08 (5.16)	69.94 (6.65) *	71.94 (7.21) *#	72.25 (4.8)
range	56–84	55–84	55–84	55–84
MMSE score (SD)	29.3 (0.8)	28.4 (1.5) *	27.5 (1.7) *#	28.5 (1.6)
range	27–30	23–30	24–30	23–30
CDR-SoB (SD)	0.03 (0.11)	1.19 (0.75) *	1.70 (0.96) *#	0.87 (0.99)
range	0–0.5	0.5–4	0.5–5.5	0–5.5
ADAS-Cog13 (SD)	9.08 (4.35)	11.78 (4.79) *	17.78 (6.56) *#	12.53 (6.4)
range	1–24	2–27	3–32	1–32
GDS (SD)	0.77 (1.16)	1.80 (1.53) *	1.57 (1.34) *	1.31 (1.4)
range	0–6	0–6	0–5	0–6
Education, y (SD)	15.22 (3.13)	16.41 (2.62) *	16.43 (2.63) *	15.93 (2.9)
Smokers (%)	8 (3.2)	2 (1.2)	7 (4.6)	17 (2.8)
Ex-smokers (%)	72 (28.8)	33 (19.1)	46 (24.1)	151 (24.6)
Diabetes (%)	19 (7.6)	14 (8.1)	16 (8.4)	49 (8.0)
Hypertension (%)	139 (55.6)	72 (41.6)*	83 (43.5) *	294 (47.9)
ApoE4+ (%)	75 (30.0)	68 (39.3)	103 (53.9) *#	246 (40.1)
Qrisk2 (SD)	24.12 (10.23)	19.31 (9.52) *	22.72(10.49) #	22.33 (10.29)
range	3–68	3–55	3–46	3–68
Amyloid positive (%)	28/93 (30.1)	72/173 (41.6)	53/84 (63.1) *#	153/350 (24.9)
BMI (SD)	28.00 (4.70)	28.64 (5.35)	27.75 (4.37)	28.1 (4.8)
range	19–54	19–54	19–50	19–54
CSF Aβ	201.83 (49.23)	191.18(49.91)	166.38 (50.70) *#	187.32 (51.67)
	N = 107	N = 162	N = 104	N = 373
CSF tau	71.16 (35.10)	73.88 (44.99)	95.21 (55.31) *#	79.03 (46.64)
	N = 105	N = 160	N = 102	N = 367
CSF ptau	31.84 (15.42)	35.05 (19.91)	45.64 (30.88) *#	37.08 (23.1)
	N = 107	N = 162	N = 104	N = 373
FDG posterior cingulate, bilateral	1.39 (0.14)	1.41 (0.15)	1.32 (0.16) *#	1.38 (0.15)
	N = 162	N = 169	N = 141	N = 472
FDG Angular gyrus, left	1.30 (0.13)	1.32 (0.14)	1.22 (0.16) *#	1.28 (0.15)
	N = 162	N = 169	N = 141	N = 472
FDG Angular gyrus, right	1.31 (0.13)	1.32 (0.13)	1.23 (0.15) *#	1.29 (0.14)
	N = 162	N = 169	N = 141	N = 472
FDG Temporal gyrus, left	1.26 (0.13)	1.26 (0.12)	1.18 (0.15) *#	1.23 (0.14)
	N = 162	N = 169	N = 141	N = 472
FDG Temporal gyrus, right	1.23 (0.12)	1.25 (0.12)	1.18 (0.13) *#	1.22 (0.13)
	N = 162	N = 169	N = 141	N = 472

* p < 0.05 EMCI or LMCI versus CN; # p < 0.05 EMCI versus LMCI; CN, cognitively normal subjects; EMCI, early mild cognitive impairment subjects; LMCI, late mild cognitive impairment subjects; ALL, entire study group.

We stratified the three groups based on APOE4 carrier status and Qrisk2 categories, with high-risk defined as 20% or greater and low-risk below 20% [5]. In the whole cohort, the Qrisk2 was not significantly different between APOE4 non-carriers (Qrisk2 = 21.42±9.99) and APOE4 carriers (Qrisk2 = 22.94±10.47) (p = 0.07). Within each diagnostic group, prevalence of APOE4 carriers was not significantly different between high and low cardiovascular risk groups (Supplementary Table 1).

Pathological hallmark in high and low cardiovascular risk subgroup with and without the presence of APOE4

To evaluate the influence of cardiovascular risk on AD pathological substrate, each of the three diagnostic groups was divided into APOE4 non-carriers and carriers, then further divided into high and low cardiovascular risk based on Qrisk2 score.

Cognitively normal subjects

Amyloid deposition

In APOE4 non-carriers, there was no increase in amyloid load, while in APOE4 carriers amyloid deposition significantly increased during follow up in high cardiovascular risk CN compared to the low risk group (Fig. 1B).

Fig. 1

Biomarkers in high cardiovascular risk (HR) and low cardiovascular risk (LR) cognitively normal subjects. Panel A and B show amyloid deposition in the parietal lobe in APOE4 non-carriers and carriers, respectively. Glucose metabolism in the different subgroups is shown in panels C and D for the posterior cingulate and panels E and F for the left angular gyrus. The CDR-SoB scores are shown in panel G for APOE4 non-carriers and panel H for APOE4 carriers, respectively (* indicates a significant difference set at p < 0.05 between HR and LR at either baseline or follow up; # and indicate a significant difference set at p < 0.05 between baseline and follow up within either the HR or LR subgroup).

Glucose metabolism

In CN APOE4 non-carriers, high cardiovascular risk group showed significantly lower glucose metabolism in all the predefined regions at baseline and follow up, compared to the low risk group (Fig. 1C, E and Supplementary Figure 1). A significant reduction in glucose metabolism at follow up was observed in high cardiovascular risk group in the left angular gyrus (Fig. 1E).

In CN APOE4 carriers, there was a significant longitudinal reduction from baseline to follow up in the posterior cingulate and right temporal gyrus in the high cardiovascular risk group but not in the low risk group (Fig. 1D and Supplementary Figure 1). Baseline and longitudinal characteristics are shown in Fig. 1D, F and Supplementary Figure 1.

CSF Aβ:p-tau ratio

In CN APOE4 non-carriers, there was no change in the CSF Aβ:p-tau ratio in high or low cardiovascular risk group during baseline and follow up. The same was true in CN APOE4 carriers with high or low cardiovascular risk.

Neurocognitive tests

In APOE4 non-carriers, high cardiovascular risk group demonstrated significantly worse scores on CDR-SoB and ADAS Cog-13 than low risk group at baseline and during follow up (Fig. 1G and Supplementary Figure 1). Moreover, scores for CDR-SoB worsened in high cardiovascular risk group from baseline to follow up (Fig. 1G).

EMCI subjects

Amyloid deposition

In APOE4 non-carrier EMCI subjects, amyloid load was significantly higher in high cardiovascular risk group compared to low risk group in all regions at baseline. During follow up, high cardiovascular risk group demonstrated significantly higher amyloid load in the posterior cingulate gyrus and in frontal cortices (Fig. 2A, C and Supplementary Figure 2). In APOE4 carriers, amyloid load was significantly higher at baseline in high cardiovascular risk group compared to low risk group in all predefined regions (Fig. 2B, D and Supplementary Figure 2). While amyloid load did not change significantly in APOE4 positive high cardiovascular risk groups longitudinally as amyloid deposition occurs early on in the disease trajectory, we demonstrated a significant ongoing increase in amyloid deposition in the low risk group in all regions (Fig. 2B, D and Supplementary Figure 2), perhaps suggesting that low cardiovascular risk group demonstrates a delayed deposition.

Fig. 2

Amyloid deposition and CSF Aβ/pTau ratio in high cardiovascular risk (HR) and low cardiovascular risk (LR) early mild cognitive impairment subjects. Amyloid deposition in the parietal lobe is shown in panels A and B in APOE4 non-carriers and carriers, respectively. Panels C and D indicate posterior cingulate amyloid load in the four subgroups. CSF Aβ/pTau ratio is shown in panel E for APOE4 non-carriers and in panel F for APOE4 carriers. A dotted line is at the value of 6.16, which is a predictive threshold for cognitive decline (* indicates a significant difference set at p < 0.05 between HR and LR at either baseline or follow up; indicates a significant difference set at p < 0.05 between baseline and follow up within either the HR or LR subgroup).

CSF Aβ:p-tau ratio

APOE4 non-carrier EMCIs with high cardiovascular risk had significant reduction in CSF Aβ:p-tau ratio compared to the low cardiovascular risk group. This reduction was also present in APOE4 carriers with high cardiovascular risk (Fig. 2E, F).

Glucose metabolism

In APOE4 non-carrier EMCIs, high cardiovascular risk group had significantly lower cerebral glucose metabolism than low risk group at both baseline and follow up in all predefined regions. Moreover, cerebral glucose metabolism significantly decreased during follow up in high cardiovascular risk group in almost all regions (Fig. 3A, C, E and Supplementary Figure 2), while it remained stable in low cardiovascular risk group.

Fig. 3

Brain glucose metabolism and neurocognitive measures in high cardiovascular risk (HR) and low cardiovascular risk (LR) early mild cognitive impairment subjects. Glucose metabolism in the different subgroups is shown in panels A and B for the posterior cingulate, C and D for the left angular gyrus and E and F for the right temporal gyrus. The scores for ADAS-Cog13 are reported in panel G and H for APOE4 non-carriers and carriers, respectively. Panel I and J show the CDR-SoB scores in the different subgroups (* indicates a significant difference set at p < 0.05 between HR and LR at either baseline or follow up; # and indicate a significant difference set at p < 0.05 between baseline and follow up within either the HR or LR subgroup).

APOE4 carrier high cardiovascular risk EMCI subjects had significantly lower glucose metabolism at baseline in cingulate and angular gyri (where glucose metabolism is reduced normally in AD) compared to the low cardiovascular risk group (Fig. 3B, D). Low cardiovascular risk group demonstrated a higher baseline glucose metabolism in the posterior cingulate and in the left angular gyri. Longitudinally, the glucose metabolism did not change significantly (Fig. 3B, D).

Neurocognitive tests

APOE4 non-carrier high cardiovascular risk group had significantly worse cognitive performance during follow up compared to low cardiovascular risk EMCI subjects during follow up. Interestingly, there was a significant improvement in ADAS Cog-13 and CDR-SoB during follow up (Fig. 3G, I) in APOE4 non-carrier low cardiovascular risk group.

APOE4 carriers EMCI subjects with high cardiovascular risk had significantly worse ADAS Cog-13 compared to low cardiovascular risk at baseline; CDR-SoB significantly worsened in high cardiovascular risk from baseline to follow up (Fig. 3H, J).

LMCI subjects

Amyloid deposition

APOE4 non-carriers with either low or high cardiovascular risk did not show significant changes in amyloid deposition.

APOE4 carriers with high cardiovascular risk LMCIs demonstrated significantly higher amyloid deposition in the parietal lobe compared to low cardiovascular risk group at baseline (Fig. 4B), without significant change in amyloid load during follow up.

Glucose metabolism

Fig. 4

Biomarkers in high cardiovascular risk (HR) and low cardiovascular risk (LR) late mild cognitive impairment subjects. Amyloid deposition in the parietal lobe is shown in panels A and B in APOE4 non-carriers and carriers, respectively. Glucose metabolism in the different subgroups is shown in panels C and D for the posterior cingulate. The scores for ADAS-Cog13 are reported in panel E and F for APOE4 non-carriers and carriers, respectively, whereas panels G and H show the CDR-SoB scores in the different subgroups (* indicates a significant difference set at p < 0.05 between HR and LR at either baseline or follow up; # and indicate a significant difference set at p < 0.05 between baseline and follow up within either the HR or LR subgroup).

APOE4 non-carrier high cardiovascular risk LMCIs demonstrated significant deterioration in glucose metabolism from baseline to follow up in almost all the predefined regions. During follow up, high cardiovascular risk group showed significantly lower glucose metabolism compared to the low risk group (Fig. 4C and Supplementary Figure 3).

APOE4 carrier LMCI subjects with low and high cardiovascular risk groups showed similar levels of glucose metabolism at baseline and during follow up. In both high and low cardiovascular risk groups, a significant reduction in glucose metabolism was observed in almost all the predefined regions from baseline to follow up (Fig. 4D and Supplementary Figure 3).

CSF Aβ:ptau-ratio

CSF Aβ:ptau ratios demonstrated a trend towards reduction in both high and low cardiovascular risk LMCI subjects.

Neurocognitive tests

In APOE4 non-carrier low cardiovascular risk LMCI subjects, there was no significant change in ADAS Cog-13 and CDR-SoB scores between baseline and follow-up (Fig. 4E, G). In APOE4 carrier LMCIs, ADAS Cog-13 scores and CDR-SoB scores significantly worsened during follow up in both low and high cardiovascular risk groups (Fig. 4E, H).

Covariance analysis

Covariance analysis between Qrisk2 and biomarkers and cognitive measures at follow up was performed in the whole group, with APOE4 status as a covariate and adjusting for baseline levels of each parameter. In the whole study population, Qrisk2 was significantly related to glucose metabolism in posterior cingulate, F(1, 350) = 8.61, p = 0.00, r = 0.15. There was also a significant effect of APOE4 status, F(1, 350) = 14.37, p = 0.00, r = 0.20; similarly, glucose metabolism in the right temporal gyrus was associated with Qrisk2, F(1, 350) = 8.60, p = 0.00, r = 0.15 and APOE4 status, F(1, 350) = 12.03, p = 0.00, r = 0.18. Qrisk2 was significantly related to glucose metabolism in the left angular gyrus, F(1, 350) = 5.09, p = 0.02, r = 0.12, with a significant effect also of APOE4 status, F(1, 350) = 5.04, p = 0.02, r = 0.20. Qrisk2 scores were not related to amyloid uptake in temporal and parietal lobes; a significant association with APOE4 status was found with amyloid levels in both temporal, F(1, 293) = 5.20, p = 0.02, r = 0.13, and parietal lobes, F(1, 293) = 6.81, p = 0.01, r = 0.15. Qrisk2 was significantly related to CDR-SoB, F(1, 600) = 11.35, p = 0.00, r = 0.14, with a significant association also with APOE4 status, F(1, 600) = 38.78, p = 0.00, r = 0.25. ADAS Cog-13 scores were significantly associated with APOE4 status, F(1, 594) = 11.34, p = 0.00, r = 0.14, but not with Qrisk2 scores.

DISCUSSION

In this study we have demonstrated that in APOE4 non-carriers, amyloid deposition, glucose metabolism, CSF Aβ:tau, and neurocognitive measures were significantly influenced by cardiovascular risk in cognitively normal and MCI subjects. The same effect was also demonstrated in APOE4 carriers; and the influence of cardiovascular risk factors on amyloid, glucose metabolism, and tau deposition was independent of APOE4 carrier status. We have demonstrated profound influence on the biomarkers in the early stages of the disease compared to the later stages in the AD trajectory. High cardiovascular risk score is associated with reduced brain glucose metabolism, increased amyloid deposition, and accelerated deterioration of neurocognitive function in both APOE4 carriers and non-carriers in cognitively normal older subjects along with early and late MCI subjects. We have also noticed that cardiovascular risk factors have predominant influence on glucose metabolism, while the APOE4 predominantly influenced amyloid deposition. While it is demonstrated that there is a synergistic effect of APOE4 and cardiovascular risk factors on AD progression [7 –11]; our data indicate that in APOE4 non-carriers, high cardiovascular risk is associated with worse biomarker and cognitive profile compared to low risk group, which indicates that cardiovascular risk factors exert their influence independent of APOE4. This study highlights the importance of addressing modifiable cardiovascular risk factors like smoking, high blood pressure, diabetes, high cholesterol, in prodromal AD subjects and in high risk cognitively normal subjects independent of genetic status. This study further demonstrates that high cardiovascular risk is associated with progressive neuronal damage as evidenced by reduction in both glucose metabolism and CSF Aβ:p-tau ratio, even before the appearance of cognitive symptoms.

Amyloid levels increase over several decades during progression from CN to MCI stage; then the deposition plateaus in AD [22]. APOE4 positivity decreases amyloid clearance in animal models, and subjects who are APOE4 carriers, typically have higher amyloid burden than non-carriers [23]. While we have shown significantly increased amyloid burden in CN APOE4 carriers with high cardiovascular risk, we have also demonstrated that APOE4 non-carriers with high cardiovascular risk also demonstrated a trend toward increased amyloid deposition, although without reaching significance suggesting that even in the very early stages of the disease, cardiovascular risk exerts its influence and accelerates AD related neuropathological changes.

Cerebral glucose metabolism is a marker of neuronal function. Demonstration of lower glucose metabolism in the APOE4 non-carrier high cardiovascular risk group compared to the low risk CN group at baseline and follow up highlights the influence of cardiovascular risk factors at an early stage of the AD trajectory. Moreover, in APOE4 carrier CN high and low cardiovascular risk group had similar cerebral glucose metabolism at baseline; however, only the high cardiovascular risk subgroup showed a longitudinal decline in glucose metabolism in the posterior cingulate and in the temporal gyrus. Demonstration of significant decline in glucose metabolism in EMCI subjects with high cardiovascular risk highlights the role of cardiovascular risk factors along the disease trajectory. Similarly, in LMCI subjects, all subgroups except APOE4 non-carriers with low cardiovascular risk show a longitudinal decline in cerebral glucose metabolism from baseline to follow up. Thus, low cardiovascular risk seems to have a protective role even in LMCI subjects. Moreover, the analysis of covariance shows that Qrisk2 is an independent predictor of brain hypometabolism, on top of APOE4. Overall, our results show that the APOE4 non-carrier low cardiovascular risk group does not decline and the APOE4 carrier low cardiovascular risk group is associated with slower progression.

In EMCI subjects, amyloid deposition at baseline was significantly higher in high cardiovascular risk groups in both APOE4 carriers and non-carriers. Amyloid deposition remained at a higher stable level during follow up in APOE4 carriers with high cardiovascular risk, while it significantly increased in APOE4 carriers with low risk. This may suggest that in high cardiovascular risk group, amyloid deposition reached a plateau earlier than in APOE4 low risk groups, while in the low risk group amyloid load is continuing to rise until it plateaus. Overall, a low cardiovascular risk is protective, and slows down amyloid deposition associated with APOE4, suggesting that even in the presence of APOE4, modifiable cardiovascular risk factors have significant influence on amyloid burden.

Cognitive changes in EMCI subjects as measured with ADAS-Cog13 and CDR-SoB globally indicate that high cardiovascular risk subgroups have worse scores compared to the low risk subgroups. Cognitive scores of APOE4 non-carriers with low cardiovascular risk in EMCI and LMCI did not deteriorate during follow up, which was consistent with the changes seen in cerebral glucose metabolism and amyloid deposition, suggesting that this subgroup consists of those subjects not clinically progressing or there is a delay to convert to AD, which may be influenced by the low cardiovascular risk.

Previous studies have demonstrated that CSF Aβ:p-tau ratio independently predicts functional and cognitive decline even in those with APOE4, indicating that CSF changes is independently driven by disease processes in the brain. In our EMCI subjects, only high cardiovascular risk groups demonstrated a significant decline in CSF ratio whereas the low risk groups remained stable, regardless of APOE4. This suggests that low cardiovascular risk offers a protective mechanism even in the presence of APOE4.

Around a third of AD cases can be attributable to seven potentially modifiable risk factors (diabetes mellitus, midlife hypertension, midlife obesity, physical inactivity, depression, smoking, and low education). Reducing the prevalence of each of the risk factors by 10% or 20% per decade would potentially reduce the worldwide prevalence of AD in 2050 by between 8% and 15% [2]. The Qrisk2 score used in our study considers several of these risk factors and has been extensively validated at identifying subjects with increased cardiovascular risk. While our results are in line with the findings from the FINGER study, where there is overall 25% improvement in Neuropsychological Test Battery test scores in the cardiovascular risk intervention group compared to the control group [24], we were able to demonstrate the neuropathological changes underpinning the cognitive function.

There are a few possible underlying mechanisms by which the cardiovascular risk factors exert their influence on cognitive function. Many studies have confirmed that the risk of dementia and AD is higher in individuals with diabetes mellitus and brains of diabetic patients show increased microvascular damage and atrophy [25]. The exact mechanism of how obesity is related to AD pathology is still yet to be fully understood. Obesity in mid-life is associated with an increased risk of dementia and AD. It is suggested that visceral obesity and insulin resistance may arise from low grade inflammation of the adipose tissue, which may interact with the impaired central inflammatory response, leading to neurodegeneration [26 –28]. Overall our findings further demonstrate that high cardiovascular risk factors exert their effect on AD neuropathological substrates early on in the disease process and support the theory that these factors influence the disease process in midlife when amyloid deposition begins to occur, as shown in Fig. 5.

Fig. 5

Influence of cardiovascular risk factors on dynamic biomarkers in AD trajectory. This figure schematically represents the influence of some of the major cardiovascular risk factors on biomarkers changes (including microglial activation) in the AD trajectory. The shaded area could represent the time when cardiovascular interventions might be most effective on progression (adapted from [35]).

This study demonstrates that we could classify subjects based on APOE4 and cardiovascular risk as “high-risk Alzheimer subjects” who are APOE4 carriers with high cardiometabolic risk, and “low-risk Alzheimer subjects” who are APOE4 non-carriers with low cardiometabolic risk. In between these two categories, we identify a “medium-risk” group, consisting of both APOE4 non-carriers with high cardiovascular risk and APOE4 carriers with low cardiovascular risk. This categorization could be used to reduce the CVD burden, aiming at preventing the progression of neuropathological changes in the general population. Primary prevention when risk rises above the high risk cut-off should be considered not just to prevent cardiovascular events, but also to slow the progression of dementia, in line with evidences from other studies [2 , 30].

One of the limitations of this study is that the information about heart attack in first-degree relatives and cholesterol values were not available, therefore the risk we calculated may have been underestimated in some participants; the prevalence of high cardiovascular risk is also underestimated in our cohort, suggesting that the differences we observed in low risk versus high risk patients could be even larger. It should also be noted that the prevalence of cardiovascular risk factors in the ADNI cohort was lower than in the general population of the same age [31 –33]. Due to the nature of the subgroup analysis, cohort size in each arm was small with minimal difference in age group. However, we were able to demonstrate significant changes in biomarkers even with this small numbers. Alongside this, Qrisk2 is developed for UK population, and caution should be exercised in translating this score to other populations [5, 34].

Conclusion

We have demonstrated that high cardiovascular risk is associated with a more rapid progression as indicated by several AD biomarkers in APOE4 non-carriers and APOE4 carriers independently, and this is seen across prodromal AD. While modification of genetic influence is difficult to achieve, modification of cardiovascular risk factors to lower risk can delay disease progression, in APOE4 negative and APOE4 positive subjects. Given that treatments to reduce cardiovascular risk are readily available and safe, the modification of cardiometabolic risk factors should be a priority to prevent AD, while we eagerly await for disease modifying treatment.

Footnotes

ACKNOWLEDGMENTS

Data collection and sharing for this project was funded by the Alzheimer’s Disease Neuroimaging Initiative (ADNI) (National Institutes of Health Grant U01 AG024904) and DOD ADNI (Department of Defense award number W81XWH-12-2-0012). ADNI is funded by the National Institute on Aging, the National Institute of Biomedical Imaging and Bioengineering, and through generous contributions from the following: AbbVie, Alzheimer’s Association; Alzheimer’s Drug Discovery Foundation; Araclon Biotech; BioClinica, Inc.; Biogen; Bristol-Myers Squibb Company; CereSpir, Inc.; Cogstate; Eisai Inc.; Elan Pharmaceuticals, Inc.; Eli Lilly and Company; EuroImmun; F. Hoffmann-La Roche Ltd and its affiliated company Genentech, Inc.; Fujirebio; GE Healthcare; IXICO Ltd.; Janssen Alzheimer Immunotherapy Research & Development, LLC.; Johnson & Johnson Pharmaceutical Research & Development LLC.; Lumosity; Lundbeck; Merck & Co., Inc.; Meso Scale Diagnostics, LLC.; NeuroRx Research; Neurotrack Technologies; Novartis Pharmaceuticals Corporation; Pfizer Inc.; Piramal Imaging; Servier; Takeda Pharmaceutical Company; and Transition Therapeutics. The Canadian Institutes of Health Research is providing funds to support ADNI clinical sites in Canada. Private sector contributions are facilitated by the Foundation for the National Institutes of Health (). The grantee organization is the Northern California Institute for Research and Education, and the study is coordinated by the Alzheimer’s Therapeutic Research Institute at the University of Southern California. ADNI data are disseminated by the Laboratory for Neuro Imaging at the University of Southern California.

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Chouraki

, Seshadri

(2014) Genetics of Alzheimer’s disease, Adv Genet 87, 245-294.

Norton

, Matthews

, Barnes

, Yaffe

, Brayne

(2014) Potentialfor primary prevention of Alzheimer’s disease: An analysis ofpopulation-based data. Lancet Neurol 13, 788-794.

Cummings

, Morstorf

, Zhong

(2014) Alzheimer’s diseasedrug-development pipeline: Few candidates, frequent failures. Alzheimers Res Ther 6, 37.

Hippisley-Cox

, Coupland

, Vinogradova

, Robson

, May

, Brindle

(2007) Derivation and validation of QRISK, a new cardiovasculardisease risk score for the United Kingdom: Prospective open cohortstudy. BMJ 335, 136.

Collins

, Altman

(2010) An independent and external validation of QRISK2 cardiovascular disease risk score: A prospective opencohort study, BMJ 340, c2442.

Collins

, Altman

(2012) Predicting the 10 year risk of cardiovascular disease in the United Kingdom: Independent and external validation of an updated version of QRISK2, BMJ 344, e4181.

Mielke

, Leoutsakos

, Tschanz

, Green

, Tripodis

, Corcoran

, Norton

, Lyketsos

(2011) Interaction between vascularfactors and the APOE epsilon4 allele in predicting rate of progression in Alzheimer’s disease. J Alzheimers Dis 26, 127-134.

Eriksson

, Bennet

, Gatz

, Dickman

, Pedersen

(2010) Nonstroke cardiovascular disease and risk of Alzheimer disease anddementia. Alzheimer Dis Assoc Disord 24, 213-219.

Kivipelto

, Rovio

, Ngandu

, Kareholt

, Eskelinen

, Winblad

, Hachinski

, Cedazo-Minguez

, Soininen

, Tuomilehto

, Nissinen

(2008) Apolipoprotein E epsilon4 magnifies lifestyle risks fordementia: A population-based study. J Cell Mol Med 12, 2762-2771.

10.

Beeri

, Rapp

, Silverman

, Schmeidler

, Grossman

, Fallon

, Purohit

, Perl

, Siddiqui

, Lesser

, Rosendorff

, Haroutunian

(2006) Coronary artery disease is associated withAlzheimer disease neuropathology in APOE4 carriers. Neurology 66, 1399-1404.

11.

Rusanen

, Rovio

, Ngandu

, Nissinen

, Tuomilehto

, Soininen

, Kivipelto

(2010) Midlife smoking, apolipoprotein E and risk of dementia and Alzheimer’s disease: A population-based cardiovascularrisk factors, aging and dementia study. Dement Geriatr CognDisord 30, 277-284.

12.

Rusanen

, Kivipelto

, Levalahti

, Laatikainen

, Tuomilehto

, Soininen

, Ngandu

(2014) Heart diseases and long-term risk of dementia and Alzheimer’s disease: A population-based CAIDE study. J Alzheimers Dis 42, 183-191.

13.

Donix

, Scharf

, Marschner

, Werner

, Sauer

, Gerner

, Nees

, Meyer

, Donix

, Von Kummer

, Holthoff

(2013) Cardiovascular risk and hippocampal thickness in Alzheimer’sdisease. Int J Alzheimers Dis 2013, 108021.

14.

Aisen

, Petersen

, Donohue

, Gamst

, Raman

, Thomas

, Walter

, Trojanowski

, Shaw

, Beckett

, Jack

Jr , Jagust

, Toga

, Saykin

, Morris

, Green

, Weiner

, Alzheimer’sDisease

Neuroimaging Initiative

(2010) Clinical core of the Alzheimer’s Disease Neuroimaging Initiative: Progress and plans. Alzheimers Dement 6, 239-246.

15.

Weiner

, Veitch

, Aisen

, Beckett

, Cairns

, Cedarbaum

, Green

, Harvey

, Jack

, Jagust

, Luthman

, Morris

, Petersen

, Saykin

, Shaw

, Shen

, Schwarz

, Toga

, Trojanowski

, Alzheimer’s

Disease Neuroimaging Initiative

(2015) 2014 Update of the Alzheimer’s Disease Neuroimaging Initiative: Areview of papers published since its inception, AlzheimersDement 11, e1-120.

16.

Buchhave

, Minthon

, Zetterberg

, Wallin

, Blennow

, Hansson

(2012) Cerebrospinal fluid levels of beta-amyloid 1-42, but not of tau, are fully changed already 5 to 10 years before the onset of Alzheimer dementia. Arch Gen Psychiatry 69, 98-106.

17.

Landau

, Jagust

(2015) Florbetapir processing methods. Alzheimer’s Disease Neuroimaging Initiative. https://adni.bitbucket.io/reference/docs/UCBERKELEYAV45/ADNI_AV45_Methods_JagustLab_06.25.15.pdf

18.

Reiman

, Caselli

, Chen

, Alexander

, Bandy

, Frost

(2001) Declining brain activity in cognitively normal apolipoprotein Eepsilon 4 heterozygotes: A foundation for using positron emissiontomography to efficiently test treatments to prevent Alzheimer’sdisease. Proc Natl Acad Sci U S A 98, 3334-3339.

19.

Zhou

, Nakatani

, Teramukai

, Nagai

, Fukushima

, Alzheimer’s Disease Neuroimaging Initiative (2012) Risk classification in mildcognitive impairment patients for developing Alzheimer’s disease. J Alzheimers Dis 30, 367-375.

20.

JackCR, Jr., Lowe

, Weigand

, Wiste

, Senjem

, Knopman

, Shiung

, Gunter

, Boeve

, Kemp

, Weiner

, Petersen

, Alzheimer’s Disease Neuroimaging Initiative (2009) Serial PIB and MRI in normal, mild cognitive impairment and Alzheimer’s disease:Implications for sequence of pathological events in Alzheimer’s disease. Brain 132, 1355-1365.

21.

Ossenkoppele

, Jansen

, Rabinovici

, Knol

, van der Flier

, van Berckel

, Scheltens

, Visser

, Amyloid

PETSG

, VerfaillieSC , Zwan

, Adriaanse

, Lammertsma

, Barkhof

, Jagust

, Miller

, Rosen

, Landau

, Villemagne

, Rowe

, Lee

, NaDL , Seo

, Sarazin

, Roe

, Sabri

, Barthel

, Koglin

, Hodges

, Leyton

, Vandenberghe

, van Laere

, Drzezga

, Forster

, Grimmer

, Sanchez-Juan

, Carril

, Mok

, Camus

, Klunk

, Cohen

, Meyer

, Hellwig

, Newberg

, Frederiksen

, Fleisher

, Mintun

, Wolk

, Nordberg

, Rinne

, Chetelat

, Lleo

, Blesa

, Fortea

, Madsen

, Rodrigue

, Brooks

(2015) Prevalence of amyloid PET positivity in dementia syndromes: Ameta-analysis. JAMA 313, 1939-1949.

22.

Villemagne

, Burnham

, Bourgeat

, Brown

, Ellis

, Salvado

, Szoeke

, Macaulay

, Martins

, Maruff

, Ames

, Rowe

, Masters

, Australian

Imaging Biomarkers and Lifestyle (AIBL) Research Group

(2013) Amyloid beta deposition,neurodegeneration, and cognitive decline in sporadic Alzheimer’sdisease: A prospective cohort study. Lancet Neurol 12, 357-367.

23.

Castellano

, Kim

, Stewart

, Jiang

, DeMattos

, Patterson

, Fagan

, Morris

, Mawuenyega

, Cruchaga

, Goate

, Bales

, Paul

, Bateman

, Holtzman

(2011) Human apoE isoformsdifferentially regulate brain amyloid-beta peptide clearance, Sci Transl Med 3, 89ra57.

24.

Ngandu

, Lehtisalo

, Solomon

, Levalahti

, Ahtiluoto

, Antikainen

, Backman

, Hanninen

, Jula

, Laatikainen

, Lindstrom

, Mangialasche

, Paajanen

, Pajala

, Peltonen

, Rauramaa

, Stigsdotter-Neely

, Strandberg

, Tuomilehto

, Soininen

, Kivipelto

(2015) A 2 year multidomain intervention of diet, exercise, cognitive training, and vascular risk monitoringversus control to prevent cognitive decline in at-risk elderlypeople (FINGER): A randomised controlled trial. Lancet 385, 2255-2263.

25.

de Bruijn

, Bos

, Portegies

, Hofman

, Franco

, KoudstaalPJ , Ikram

(2015) The potential for prevention of dementia acrosstwo decades: The prospective, population-based Rotterdam Study. BMC Med 13, 132.

26.

Chiu

, Su

, Cheng

, Liu

, Chang

, Dewey

, Stewart

, Huang

(2008) The effects of omega-3 fatty acids monotherapy inAlzheimer’s disease and mild cognitive impairment: A preliminaryrandomized double-blind placebo-controlled study. ProgNeuropsychopharmacol Biol Psychiatry 32, 1538-1544.

27.

van der Heide

, Ramakers

, Smidt

(2006) Insulin signaling inthe central nervous system: Learning to survive. ProgNeurobiol 79, 205-221.

28.

Bettcher

, Kramer

(2014) Longitudinal inflammation, cognitivedecline, and Alzheimer’s disease: A mini-review. Clin PharmacolTher 96, 464-469.

29.

Barnes

, Yaffe

(2011) The projected effect of risk factorreduction on Alzheimer’s disease prevalence. Lancet Neurol 10, 819-828.

30.

Reed

, Marchant

, Jagust

, DeCarli

, Mack

, Chui

(2012) Coronary risk correlates with cerebral amyloid deposition. Neurobiol Aging 33, 1979-1987.

31.

Blackwell

, Lucas

, Clarke

(2014) Summary health statisticsfor US adults: National health interview survey, 2012, Vitaland health statistics. Series 10, Data from the National HealthSurvey 1-161.

32.

Ogden

, Carroll

, Fryar

, Flegal

(2015) Prevalence of obesity among adults and youth: United States, 2011–2014. NCHS Data Brief 219, 1-8.

33.

Centers for Disease Control and Prevention (2013) The state of aging and health in America 2013. Centers for Disease Control and Prevention, US Department of Health and Human Services, Atlanta, GA.

34.

Collins

, Altman

(2009) An independent external validation and evaluation of QRISK cardiovascular risk prediction: A prospectiveopen cohort study, BMJ 339, b2584.

35.

Calsolaro

, Edison

(2016) Neuroinflammation in Alzheimer’sdisease: Current evidence and future directions. Alzheimers Dement 12, 719-732.