Abstract
Objective To evaluate the association between brain amyloid β (Aβ) and cerebral small vessel disease (CSVD) markers, as well as their joint effect on cognition, in a memory clinic study.
Methods A total of 186 individuals visiting a memory clinic, diagnosed with no cognitive impairment, cognitive impairment no dementia (CIND), Alzheimer dementia (AD), or vascular dementia were included. Brain Aβ was measured by [11C] Pittsburgh compound B–PET global standardized uptake value ratio (SUVR). CSVD markers including white matter hyperintensities (WMH), lacunes, and cerebral microbleeds (CMBs) were graded on MRI. Cognition was assessed by neuropsychological testing.
Results An increase in global SUVR is associated with a decrease in Mini-Mental State Examination (MMSE) in CIND and AD, as well as a decrease in global cognition Z score in AD, independent of age, education, hippocampal volume, and markers of CSVD. A significant interaction between global SUVR and WMH was found in relation to MMSE in CIND (P for interaction: 0.009), with an increase of the effect size of Aβ (β = −6.57 [−9.62 to −3.54], p < 0.001) compared to the model without the interaction term (β = −2.91 [−4.54 to −1.29], p = 0.001).
Conclusion Higher global SUVR was associated with worse cognition in CIND and AD, but was augmented by an interaction between global SUVR and WMH only in CIND. This suggests that Aβ and CSVD are independent processes with a possible synergistic effect between Aβ and WMH in individuals with CIND. There was no interaction effect between Aβ and lacunes or CMBs. Therefore, in preclinical phases of AD, WMH should be targeted as a potentially modifiable factor to prevent worsening of cognitive dysfunction.
Glossary
- Aβ=
- amyloid β;
- AD=
- Alzheimer dementia;
- CI=
- confidence interval;
- CIND=
- cognitive impairment no dementia;
- CMB=
- cerebral microbleed;
- CSVD=
- cerebral small vessel disease;
- DSM-IV=
- Diagnostic and Statistical Manual for Mental Disorders, fourth edition;
- FLAIR=
- fluid-attenuated inversion recovery;
- MCI=
- mild cognitive impairment;
- MMSE=
- Mini-Mental State Examination;
- NCI=
- no cognitive impairment;
- PCA=
- principal component analysis;
- PiB=
- Pittsburgh compound B;
- ROI=
- region of interest;
- RR=
- rate ratio;
- SUVR=
- standardized uptake value ratio;
- svMCI=
- subcortical vascular MCI;
- SWI=
- susceptibility-weighted image;
- VaD=
- vascular dementia;
- WMH=
- white matter hyperintensities
Brain amyloid β (Aβ) and cerebral small vessel disease (CSVD) are major causes of cognitive impairment and dementia. Autopsy studies have shown that brain Aβ frequently co-occurs with CSVD among patients with clinically diagnosed Alzheimer dementia (AD).1,–,3 Moreover, co-occurrence increases the risk of dementia, suggesting a possible synergistic effect.4,5 Previous studies evaluating the relationship between brain Aβ (utilizing PET imaging), CSVD, and cognition have shown conflicting results. A systematic review concluded that most studies failed to find a relationship between brain Aβ and white matter hyperintensities (WMH) in groups including cognitively normal, mild cognitive impairment (MCI), subcortical vascular MCI (svMCI), AD, and vascular dementia (VaD),6 while a few studies reported that higher brain Aβ was correlated with higher WMH volumes.7,8 Only one study showed an interaction between brain Aβ and WMH on cognition in individuals with svMCI.9 Studies evaluating the relationship between brain Aβ, lacunes, and cognition have been inconsistent and largely limited to individuals with subcortical vascular cognitive impairment.9,–,12 Lastly, the correlation between brain Aβ and cerebral microbleeds (CMBs) is difficult to interpret since studies were performed in different subgroups, ranging from cognitively normal to mild cognitive impairment, AD,13,14 and Aβ-negative subcortical vascular cognitive impairment,10 while no studies have evaluated their joint effect on cognition.
Limitations of current literature include underrepresentation of participants with mixed pathology, since participants were often categorized into pure brain Aβ and pure vascular groups.6 This categorization into pure groups might result in masking of possible interaction effects between Aβ and CSVD. Moreover, pure groups are not representative of the actual population of individuals with cognitive impairment, since it has been shown that up to 47% of individuals with dementia have mixed pathologies rather than pure disease.3 In this study, we used a memory clinic population with mixed pathology and a wide spectrum of cognitive impairment. We aimed to evaluate the association between brain Aβ as identified by PET imaging and individual markers of CSVD (WMH, lacunes, and CMBs) as well as their joint effect on cognition.
Methods
Study population
We employed a case–control design. From April 2016 to September 2018, we recruited 186 individuals with [11C]-Pittsburgh compound B ([11C]-PiB) PET scans from an ongoing memory clinic study at the National University Hospital, Singapore. Controls were individuals classified as no cognitive impairment (NCI), who had no objective evidence of neuropsychological deficits on a locally validated standard neuropsychological battery15,16 or any functional loss. Cases were individuals who were either diagnosed with cognitive impairment no dementia (CIND) or dementia. The diagnosis of CIND was given if an individual had no loss of independence in daily activities and impairment in at least one domain of the standard neuropsychological battery. When individuals scored 1.5 SD below education-adjusted cutoff values on each individual test, they were considered to have a failed test.15,16 Impairment in a domain was defined by failure in at least half of the tests in that domain. The diagnosis of dementia was made according to DSM-IV criteria. The etiologic diagnoses of dementia were based on the following criteria: AD was diagnosed using the National Institute of Neurologic and Communicative Disorders and Stroke–Alzheimer’s Disease and Related Disorders Association (NINCDS-ADRDA) criteria.17 VaD was defined using the National Institute of Neurologic Disorders and Stroke–Association Internationale pour la Recherche et l' Enseignement en Neurosciences (NINDS-AIREN) criteria.18 All individuals underwent physical, clinical, and neuropsychological assessments as well as neuroimaging at the National University Hospital, Singapore.
Standard protocol approvals, registrations, and patient consents
Ethics approval was obtained from the National Healthcare Group Domain-Specific Review Board. The study was conducted in accordance with the Declaration of Helsinki. All participants provided written informed consent in the preferred language of the participants by bilingual study coordinators prior to their recruitment into the study.
Brain imaging
Image acquisition
All participants underwent structural MRI and PET-MRI. Structural MRI was performed on a 3T Siemens Magnetom Trio Tim scanner, using a 32-channel head coil, at the Clinical Imaging Research Centre of the National University of Singapore. The standardized neuroimaging protocol included a 3D T1-weighted, T2-weighted, fluid-attenuated inversion recovery (FLAIR) and susceptibility-weighted image (SWI) sequence. PET-MRI was done on an mMR synchronous PET/magnetic resonance scanner (Siemens Healthcare GmbH).19 All individuals underwent a 30-minute brain PET scan 40 minutes after injection of 370 (±10%) MBq of [11C]-PiB. Each list-mode datum was rebinned before tomographic reconstruction into a single static frame, where motion correction was applied using an in-house developed rebinner.20 Every 20 seconds, motion parameters were estimated from an initial dynamic reconstruction. Motion and corrected scans were then reconstructed into a 344 × 344 × 344 voxel volume with a voxel size of 2.09 × 2.09 × 2.03 mm3 using 3D ordinary Poisson ordered-subsets expectation maximization21 with all corrections applied including resolution modeling and using 3 iterations and 21 subsets.
Image analysis
Markers of CSVD were defined based on the Standards for Reporting Vascular Changes on Neuroimaging (STRIVE) criteria. The following MRI markers were analyzed for each participant:
Lacunes were classified as round or ovoid lesions involving the subcortical regions, 3–15 mm in diameter, with a low signal on T1-weighted images and FLAIR, a high signal on T2-weighted images, and a hyperintense rim with a center following the CSF intensity.22
CMBs were graded using the Brain Observer Microbleed Scale and were defined as focal, rounded areas of hypointensity, 2–10 mm in diameter, with blooming on SWI.23
Total intracranial volume and WMH volume were quantified by automatic segmentation at the Erasmus MC, University Medical Center Rotterdam, the Netherlands. The proton density–weighted T1 sequence and T2-weighted images were utilized to quantify brain tissue segmentation, while WMH volume was segmented using the FLAIR sequence, as described previously.24,–,27 Hippocampus volume was derived using a model-based automated procedure (FreeSurfer, v.5.1.0) on T1-weighted MRI.28
For the brain Aβ quantification measured by standardized uptake value ratio (SUVR) (figure 1), [11C]-PiB PET images were processed and analyzed using our automated pipeline, where each PET image was first linearly registered to the patient's T1-weighted MRI (step 1), and then spatially normalized in Montreal Neurological Institute 152 space (step 2). The SUVR image was then computed using a generic region of interest (ROI) for cerebellar gray matter that was generated from the ICBM152 T1 template. The cerebellar gray matter was chosen as a reference region as it is relatively devoid of senile brain Aβ plaques and shows low [11C]-PiB binding.29 The global SUVR was finally computed from the SUVR volume using a total brain Aβ-specific target ROI (step 3). With the aim of maximizing sensitivity, the total target was defined from the Aβ deposition spatial pattern, which was directly derived from PiB data instead of being defined a priori, using structural information. For this, the map of the brain Aβ deposition pattern was generated using principal component analysis (PCA) on the SUVR [11C]-PiB images. It reported that the brain Aβ deposition pattern is mainly contained in the first component of PCA.30,31 In this work, the first component explained 74% of the variability while only 1.6% of the variability was explained by the second component. The global brain Aβ-specific region was then derived by thresholding the generated brain Aβ deposition map and included frontal, parietal, temporal, and occipital regions as well as the nucleus accumbens and the thalamus. Significant global SUVR was defined as a cutoff ≥1.5, such that global SUVR values <1.5 were considered Aβ negative, while global SUVR values ≥1.5 were considered Aβ positive. The average interval between MRI and PET imaging was a mean (SD) of 7.6 months (6.9).
AD = Alzheimer disease; MNI = Montreal Neurological Institute; MPRAGE = magnetization-prepared radiofrequency pulses and rapid gradient-echo; MR = magnetic resonance; SUVR = standardized uptake value ratio; VOI = volume of interest.
Cognitive assessment
Cognitive function was assessed using the Mini-Mental State Examination (MMSE) as well as a composite Z score from a previously described detailed neuropsychological battery.32 This composite Z score reflected global cognitive functioning.
Covariates
All individuals were administered a standardized demographic questionnaire during the baseline visit and data on age, sex, and years of formal education were recorded. In addition, the above described markers of CSVD, including WMH, lacunes, and CMBs, were utilized as covariates.
Statistical analysis
To compare the baseline characteristics between the different diagnostic groups, χ2 test was used for categorical variables (or exact test33,34 when applicable) and analysis of variance for continuous variables. In case of skewed distribution of continuous variables, differences between groups were determined using the Kruskal-Wallis test. WMH volume was logarithmically transformed due to the skewed distribution for further analysis. For our analysis, continuous measurements of WMH and global SUVR were used, whereas counts were used for lacunes and CMBs. After performing regression analysis to show the overall difference for the diagnostic groups (significant p value <0.05), linear regression models were utilized to evaluate the association between the global SUVR and WMH volume (log-transformed), with the respective βs and 95% confidence intervals (CIs). Due to the count data of the markers of CSVD, negative binomial regression was used to evaluate the association between global SUVR and lacunes or CMBs, with respective rate ratios (RRs) and 95% CIs. The models were adjusted for age and sex. For the negative binomial regression model, interpretation of the effect sizes was as follows: a patient with 1 SD increment in Aβ level will have on average RR times as many lacunes or CMBs on MRI compared to a person with lower Aβ levels.
To explore the effect of global SUVR on MMSE and global composite Z score, linear regression models were constructed to analyze the effect size (β) and standard error. We first performed univariate analysis with cognition as the outcome. A selection criterion, p value <0.05 was used to select variables for the model for multivariate analysis. In addition, we have also added age and sex for the model of multivariate analysis, since age has a known effect on cognition and sex was significantly different between groups. For the multivariate analysis, a forward selection approach was used to decide the final model. Models were thus adjusted for age, sex, education, and hippocampal volume (model I) and individual markers of CSVD, including WMH, lacunes, and CMBs (model II). In order to determine the possible interaction between brain Aβ and CSVD markers on MMSE and global cognition Z score, we added the cross-product term of SUVR × WMH, SUVR × lacunes, or SUVR × CMBs to the regression models described above. For all models, p value <0.05 was considered statistically significant. Statistical analysis was performed using standard statistical software (Statistical Package for Social Science, SPSS V25, SPSS, Inc).
Data availability
Data can be obtained upon request. Requests should be directed towards the data manager (phcshy{at}nus.edu.sg), which has a protocol for approving data requests. Because of restrictions based on privacy regulations and informed consent of the participants, data cannot be made freely available in a public repository.
Results
Clinical and demographic data are reported in table 1. Of the 186 individuals, 29 (15.5%) were diagnosed with NCI, 101 (54.3%) with CIND, 36 (19.3%) with AD, and 20 (10.7%) with VaD. The diagnostic groups did not differ in age, with a global mean age (SD) of 75.6 (7.2). Of the total cohort, 53.8% was female, and more women were diagnosed with AD (81%) compared to the other diagnostic groups. Individuals with AD and VaD had lower years of education compared to individuals with NCI or CIND (p < 0.001). There was a higher percentage of APOE4 carriers (47%) among individuals with AD, although this was not significantly different from individuals with NCI (28%), CIND (26%), or VaD (30%). There was no significant difference between the diagnostic groups for vascular risk factors including hypertension, hyperlipidemia, and diabetes mellitus type II. Individuals with AD had a higher global [11C]-PiB SUVR compared to individuals with NCI, CIND, and VaD (p < 0.001). Moreover, individuals with CIND or VaD had lower hippocampal volume compared to individuals with NCI, while individuals with AD had lower hippocampal volume compared to individuals with NCI, CIND, or VaD (p < 0.001). Individuals with VAD had higher numbers of cortical infarcts (p = 0.002) and lacunes (p < 0.001) compared to individuals with NCI, CIND, or AD, while they had more WMH compared to individuals with NCI (p = 0.010).
Characteristics of study participants (n = 186)
Figure 2 shows the prevalence of brain Aβ-negative and brain Aβ-positive individuals in different diagnostic groups. Of the individuals with NCI, 5 of the 29 (17.2%) were brain Aβ-positive, while 24 of the 101 (23.8%) individuals with CIND were brain Aβ-positive. In the AD group, 27 of the 36 (75%) were brain Aβ-positive compared to 3 out of 17 (15%) with VaD.
PiB PET images of one representative individual from each diagnostic group. The percentages represent the prevalence of brain Aβ-negative and brain Aβ-positive individuals in different diagnostic groups. AD = Alzheimer dementia; CIND = cognitive impairment, no dementia; NCI = no cognitive impairment; VaD = vascular dementia.
We performed regression analysis to show the overall difference for the diagnostic groups (data not shown). Upon adding the SUVR as exposure, CSVD as outcome, age and sex as covariates, and the interaction term between SUVR and diagnosis, the overall p value for WMH was borderline (p = 0.08); for lacunes, the difference was nonsignificant (p = 0.57), and for CMB, the overall difference was significant (p < 0.001). On observing the significant p values for interaction for WMH and CMB, we further performed a subgroup analysis. The associations between global SUVR and different markers of CSVD across the diagnostic groups are shown in table 2. Higher levels of global SUVR were associated with higher WMH volumes in individuals with VaD (β [SE], 2.75 [1.10–4.40], p = 0.003) after adjusting for age and sex. Higher global SUVR values were not associated with increasing lacunar or CMB counts in any of the diagnostic groups.
Association between global standardized uptake value ratio (SUVR) and cerebral small vessel disease markers
The associations between global SUVR and cognitive performance are listed in table 3. An increase in global SUVR was associated with a decrease in MMSE in individuals with CIND or AD as well as a decrease in global cognition Z score in individuals with AD, independent of age, sex, education, hippocampal volume, and markers of CSVD (including WMH, lacunes, and CMBs). When the cross-product term of global SUVR and the individual markers for CSVD was included in the regression model, adjusted for all previously described possible risk factors, an interaction was found between global SUVR burden and WMH on MMSE in individuals with CIND only (p for interaction 0.009) (table 4). To further understand this interaction, we examined the association between Aβ and WMH on MMSE within the tertiles of WMH adjusting for similar covariates as mentioned previously. Among individuals in the highest tertile of WMH, an interaction SUVR × WMH, p = 0.019 was observed (data not shown). In the majority of the models with the cross-product term SUVR × lacunes or SUVR × CMBs, SUVR had a significant effect on MMSE and global cognition Z score, but there was no significant interaction term.
Association of brain amyloid β (Aβ) burden measured as global standardized uptake value ratio (SUVR) with cognition among different diagnostic groups
Association of brain amyloid β (Aβ) burden measured as global standardized uptake value ratio (SUVR) with cognition in a model with the interaction term global SUVR × individual marker of cerebral small vessel disease (CSVD)
Discussion
This study evaluated the association between brain Aβ and CSVD in an Asian memory clinic population with mixed pathologies and a wide range of cognitive impairment. The major findings of our study are as follows. First, brain Aβ has an effect on cognition independent of age, sex, education, hippocampal volume, and individual markers of CSVD in CIND and AD. Second, there was a significant interaction between global SUVR and higher burden of WMH among individuals who were diagnosed with CIND.
Few studies investigating the effect of brain Aβ measured by PET imaging on cognition have controlled for the possible effects of CSVD. Some of these studies have suggested that an increase in brain Aβ is associated with lower cognition in individuals with subcortical vascular MCI or subcortical VaD,9,12 while others showed no effect of brain Aβ on cognition in the presence of CSVD.11 Our study confirms that an increase of brain Aβ is associated with a decrease in cognitive scores, independent from age, sex, education, hippocampal volume, and markers of CSVD, but only in individuals with CIND and AD and not in individuals with VaD. In individuals with VaD, cognitive dysfunction might be predominantly driven by a high burden of CSVD.35
Although the majority of the current literature failed to find an association between brain Aβ and WMH, a few studies have shown that there is a positive association between the 2 biomarkers in patients without dementia as well as in a cohort including individuals with AD.7,8,36 These results suggest a possible link between brain Aβ pathology and WMH. The present study shows that an increase in WMH is associated with an increase in brain Aβ in individuals with VaD, but not in individuals with NCI, CIND, or AD. The observation that the association between WMH and an increased Aβ burden is only seen in VaD might be explained by the possible heterogeneity in the mechanisms underlying WMH, which can include cerebral ischemia, neuroinflammation, failure of the glymphatic system, and venous collagenosis.37 Previous studies have shown that in individuals with recent ischemic stroke, increased PiB retention was found in the peri-infarct region compared to the contralateral region, suggesting an impaired Aβ clearance pathway caused by infarcted tissue.38 We suggest that in individuals with VaD, WMH represents cerebral ischemia related to small vessel disease. This may lead to a reduced clearance of brain Aβ through impairment of the glymphatic system,39 which normally allows clearance of soluble proteins and metabolites via para-intra-arterial influx and para-venular efflux of CSF.40 Alternatively, WMH may result in a sustained activation of the immune response, in particular of cytokine-producing microglia, leading to neuronal loss and Aβ accumulation.41 In addition, sustained microglial activation may induce a reduction in microglia efficiency for binding and phagocytosing Aβ.41 However, a biological link between brain Aβ pathology and cerebral microvascular damage cannot be ruled out.
The current literature shows no consensus regarding the association between brain Aβ and lacunes,9,10,12 while the majority of previous studies have found that higher levels of brain Aβ are associated with a higher number of CMBs in a wide variety of study populations.10,13,14 The present study found no association between brain Aβ and lacunes or global CMBs. For CMBs, the discrepancy between our results and previous literature may be due to a very strict selection in previous literature of either brain Aβ-negative individuals with subcortical vascular cognitive impairment10 or healthy controls.14 Moreover, it is reported that although individuals with MCI or AD had higher brain Aβ burden compared to the healthy controls, all groups had similar prevalence of lobar CMB.14 This study also did not report an association between brain Aβ and lobar CMBs in individuals with MCI or AD.
Of the studies that evaluated the joint effect of brain Aβ and CSVD on cognition, only one study has shown a significant interaction effect between brain Aβ and WMH on visuospatial functioning,9 while the remaining studies suggested an independent and additive effect of brain Aβ and WMH or lacunes.11,12 So far, no studies have evaluated the joint effect of brain Aβ and CMBs on cognition. Regarding brain Aβ and lacunes or CMBs, this study supports the current literature that suggests an independent effect on cognition.9,10,12 Overall, we also did not find a significant interaction effect between brain Aβ and WMH, except on MMSE specifically in individuals with CIND. We propose that the negative effect of brain Aβ on MMSE when co-occurring with severe WMH might be caused by damage to functional or structural connectivity network in the subcortical white matter. This hypothesis is further supported by our recent study demonstrating different functional and structural network changes between amnestic MCI and AD with and without cerebrovascular disease.42
The strength of our study is that we have investigated a memory clinic population with a wide range of cognitive impairment, which enabled the analyses of different diagnostic subgroups. In addition, our study utilized a novel Aβ PET quantification method, which does not require the parcellation of patient MRI. This enabled us to obtain global SUVR values for individuals with large strokes, where the conventional parcellation methods often fail. The limitations include a cross-sectional design. Moreover, the sample size per diagnosis group was not equally distributed and relatively small in certain groups. Finally, the results were obtained from a memory clinic cohort, which limits the generalizability of the data to the general population. Therefore, large population-based and memory clinic–based, longitudinal studies with mixed pathology are needed to further investigate the interaction effect of Aβ on cognition when co-occurring with markers of CSVD. Future studies should include other markers of CSVD including brain atrophy and perivascular spaces. This may have direct clinical relevance by intensifying cardiovascular risk management.
Study funding
Supported by the Yong Loo Lin School of Medicine Aspiration Fund (grant number NUHSRO/2014/102AF-WORLD CLASS/01) and the National Medical Research Council of Singapore (NMRC/CG/M009/2017_ NUH/NUHS and NMRC/CIRG/1485/2018). The sponsor has no role in study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the article for publication.
Disclosure
F. Saridin, S. Hilal, S. Villaraza, A. Reilhac, B. Gyanwali, T. Tanaka, M. Stephenson, S. Ng, and H. Vrooman report no disclosures. W.M. van der Flier holds the Pasman chair and is recipient of a grant for the project titled Heart Brain Connection (CVON 2018–28 & 2012-06 Heart Brain Connection). C. Chen reports no disclosures. Go to Neurology.org/N for full disclosures.
Acknowledgment
The authors thank the research coordinators of the Memory Ageing and Cognition Center of the National University Hospital Singapore, in particular Jhun Ng, as well as the neuropsychology raters-team.
Appendix Authors
Footnotes
Go to Neurology.org/N for full disclosures. Funding information and disclosures deemed relevant by the authors, if any, are provided at the end of the article.
Patient page e2951
- Received November 21, 2019.
- Accepted in final form June 15, 2020.
- © 2020 American Academy of Neurology
References
Disputes & Debates: Rapid online correspondence
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.