ARTICLE OPEN ACCESS

Circulating inflammatory biomarkers are relatedto cerebrovascular disease in older adultsYian Gu PhD Jose Gutierrez MD MPH Irene B Meier PhD Vanessa A Guzman MS Jennifer J Manly PhD

Nicole Schupf PhD Adam M Brickman PhD and Richard Mayeux MD

Neurol Neuroimmunol Neuroinflamm 20196e521 doi101212NXI0000000000000521

Correspondence

Dr Gu

yg2121columbiaedu

AbstractObjectiveThis investigation aimed at examining whether circulating inflammatory biomarkers C-reactiveprotein (CRP) interleukin-6 (IL6) and alpha 1-antichymotrypsin (ACT) were related tocerebrovascular disease (CVD) assessed by MRI

MethodsThe study included nondemented elderly participants of a community-based multiethniccohort who received baseline MRI scans and had CRP (n = 508) ACT (435) and IL6 (N =357) measured by ELISA Silent brain infarcts and white matter hyperintensities (WMH) werederived from all available MRI scans at baseline approximately 44 years after blood samplecollection for inflammatory biomarkers Repeated assessments of infarcts and WMH as well asmicrobleeds assessment were performed at follow-up MRI visits around 45 years later Cross-sectional and longitudinal relationship between inflammatory biomarkers and CVD were an-alyzed using appropriate logistic regression models generalized linear models or COXmodels

ResultsAfter adjusting for age sex ethnicity education APOE genotype and intracranial volume 1 SDincrease in log10IL6 was associated with infarcts on MRI odds ratio [OR] (95 confidenceinterval [CI]) = 128 [102ndash160] p = 0033 and 1 SD increase in log10CRP and log10ACTwasassociated with microbleeds (OR [95 CI] = 146 [102ndash209] p = 0041 and 165[111ndash246] p = 0013 respectively) One SD increase in log10ACT was also associated withlarger WMH at the follow-upMRI (b = 0103 p = 0012) and increased accumulation of WMHvolume (b = 0062 p = 0041) during follow-up The associations remained significant afteradditional adjustment of vascular risk factors and excluding participants with clinical stroke

ConclusionsAmong older adults increased circulating inflammatory biomarkers were associated with thepresence of infarcts and microbleeds WMH burden and progression of WMH

From the Taub Institute for Research in Alzheimerrsquos Disease and the Aging Brain (YG IBM VAG JJM NS AMB RM) Columbia University Department of Neurology (YG JGJJM NS AMB RM) Columbia University and the New York Presbyterian Hospital Department of Epidemiology (YG NS RM) Joseph P Mailman School of Public HealthColumbia University and the Gertrude H Sergievsky Center (YG JJM AMB RM) Columbia University New York

Funding information and disclosures are provided at the end of the article Full disclosure form information provided by the authors is available with the full text of this article atNeurologyorgNN

The Article Processing Charge was funded by NIH

This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 40 (CC BY-NC-ND) which permits downloadingand sharing the work provided it is properly cited The work cannot be changed in any way or used commercially without permission from the journal

Copyright copy 2018 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the American Academy of Neurology 1

Subclinical cerebrovascular diseases (CVDs) including pri-marily lacunar infarcts or silent brain infarcts white matterhyperintensities (WMH) and cerebral microbleeds are fre-quent findings on MRI scans of normal elderly people1ndash3

ldquoSilentrdquo changes in the brain reflect individuals carrying sub-clinical CVD without clinical symptoms but they are notbenign because of the strong association with increased risk ofrecurrent strokes4 dementia5 and mortality

Strong evidence suggests that chronic inflammation plays a cru-cial role in the development of stroke cardiovascular diseasedementia6 and atherosclerosis7 Therefore it is reasonable tohypothesize that chronic inflammation might also be involved inarteriolosclerosis damage of brain vessels manifested radiologi-cally as localized ischemic infarcts diffuseWMH ormicrobleeds8

However the association between inflammatory biomarkers andCVD remains unclear9ndash27 Some existing studies were based onclinical convenience samples some included middle-aged par-ticipants and few studies included Hispanics or African-Americans10111819 who have a higher frequency of vascular riskfactors28 Therefore the results may not be applied to clinicallynormal ethnically diverse elderly populations who could be thekey target population for potential intervention In addition thereare few longitudinal studies9152223 and few studies1319ndash21 thatexploredmicrobleedsMost studiesmeasured 2 key inflammatorymarkers c-reactive protein (CRP) and interleukin-6 (IL6) al-though alpha 1-antichymotrypsin (ACT) which has long beenimplicated in AD29 has never been examined

We examined whether circulating levels of inflammatorymarkers (CRP IL6 and ACT) were associated with thepresence and progression of CVD (ie infarcts WMH andmicrobleeds) among nondemented elderly participants ofa community-based multiethnic cohort

MethodsStudy participantsThe current study included participants from an ongoingprospective study of aging and dementia the WashingtonHeights-Hamilton Height-Inwood and Columbia AgingProject (WHICAP) who were elders identified from Medi-care beneficiaries residing in northern Manhattan30 Theoriginal cohort for this study included 2776 participants Atbaseline participants were asked about medical and neuro-logic history received assessments of health and function andwere assessed using a neuropsychological battery Participantswere followed approximately for 18 months repeating the

baseline examinations The diagnosis of dementia or its ab-sence was based on standard research criteria using all avail-able information at a consensus conference

A total of 769 WHICAP participants received MRI scans in-cluding T1-weighted images fluid-attenuated inversion re-covery (FLAIR)-weighted images and proton density andT2-weighted double-echo images31 Among the 680 non-demented (n = 45 demented) participants who had baselineMRI data available and analyzed by FreeSurfer (n = 44 withoutFreeSurefer processing)30 a total of 508 (75) participantshad CRP 435 (64) had ACT and 357 (53) had IL6detected (figure 1) Participants who had measured biomarkerswere younger had fewer African-Americans had fewer vascularrisk factors than those who did not have biomarker data sam-ple30 RepeatedMRI scans were completed for 310 participants45 (SD = 09) years after their baseline scans and 230 of themalso received gradient echo (GRE) scans for microbleeds as-sessment Compared with the neuroimaging project in whichparticipants did not have repeated scans those with repeatedscans were younger at baseline (793 vs 807 p = 0002) andhad more hypertension (90 vs 83 p = 008)30

Standard protocol approvals registrationsand patient consentsThe Columbia University Institutional Review Board hasreviewed and approved this project All individuals providedwritten informed consent

MRI protocolAll scans were acquired on a 15T Philips Intera scanner atColumbia University T1-weighted images were acquired witha repetition time of 20 milliseconds (ms) echo time 21 msfield of view (FOV) 240 cm 256times 160ndashpixelmatrix with 13-mmsection thickness and voxel size 1 times 1 times 13 mm Freesurfer(V51) (surfernmrmghharvardedu) was used to analyze T1images with visual quality control and manual correctionWMH quantification was based on T2-weighted FLAIR MRIscans (repetition time 11000ms echo time 144ms inversiontime 2800 and FOV 25 cm 2 nex 256 times 192 matrix with3-mm slice thickness) An optimized high-resolution three-dimensional T2-weighted GRE image (repetition time 45msecho time 31 ms flip angle 13 and slice thickness 2 mm) wasacquired at the follow-up MRI visit for microbleeds visualiza-tion and quantification The imaging processing was blinded toany information about participants

We used all the available images to determine whether a par-ticipant had brain infarction on MRI which was defined as

GlossaryACT = alpha 1-antichymotrypsin APOE = apolipoprotein BMI = body mass index CAA = cerebral amyloid angiopathyCVD = cerebrovascular disease CRP = C-reactive protein CV = coefficient of variation FOV = field of view GRE = gradientecho ICV = intracranial volume IL6 = interleukin-6WHICAP = Washington HeightsHamilton Heights Inwood ColumbiaAging Project WMH = white matter hyperintensity sICAM-1 = intercellular adhesion molecule-1

2 Neurology Neuroimmunology amp Neuroinflammation | Volume 6 Number 1 | January 2019 NeurologyorgNN

lesions 3 mm or larger Signal void seen on the T2-weightedimages was used to indicate a vessel We also consider othernecessary imaging characteristics such as CSF density on theT1-weighted image and distinct separation from the circle ofWillis vessels and perivascular spaces

Total WMH volume was determined following the proce-dures as previously described32 First FLAIR images wereskull stripped and then a threshold and seed-growing algo-rithm was applied to identify voxels that fell within ana priorindashdetermined distribution of hyperintense signal TotalWMH volume was then calculated by multiplying labeledvoxels by voxel dimensions We calculated the ratio of WMHover total intracranial volume (ICV) to adjust the head sizedifferences We log-transformedWMHICV ratio value in theanalyses

Microbleeds were visually inspected and rated according tothe research criteria2 We classified microbleeds into ldquolobarrdquo(in the frontal temporal parietal and occipital lobes) andldquodeeprdquo (in basal ganglia thalamus and infratentorial regions)

The number of microbleeds was also recorded A dichotomousmicrobleeds status presence (at least one microbleeds) vs ab-sence (as reference group) was used in the main analysis

Biomarker measurementThe circulating levels of CRP IL6 and ACT were measuredin previous studies of WHICAP Nonfasting plasma (CRP)and serum (IL6 and ACT) samples collected 44 years beforethe first MRI scan (figure 1) were used in all analyses Brieflyblood samples were collected in BD Vacutainer Ethyl-enediaminetetraacetic acid (EDTA) tubes and centrifuged at2000g for 15 minutes at room temperature and then stored atminus80degC in polypropylene cryotubes until analysis For serumsamples collected blood samples were allowed to sit for 30minutes before they were centrifuged aliquoted and stored atminus80degC The CRP levels were measured using a high-sensitivity(16 ngmL) ELISA kit (Diagnostic systems laboratories IncWebster TX with the intra-assay and interassay coefficient ofvariations (CVs) of 46 and 117 respectively The IL-6 levelswere measured using a high-sensitivity (011 pgmL) ELISA kit(RampD Systems Minneapolis MN intra-assay CV = 74 and

Figure 1 Selection of the study population

Selection of study participants from the neuroimaging study of the Washington HeightsHamilton Heights Inwood Columbia Aging Project (WHICAP) Thestudy participants were those participants of the WHICAP imaging substudy started in 2004 A total of 769 WHICAP participants received MRI scans includingT1-weighted images FLAIR-weighted images and proton density and T2-weighted double-echo images whichmeasured whitematter hyperintensity (WMH)and infarcts Among the 680 nondemented (n = 45 demented) participants who had baseline MRI data available and analyzed by FreeSurfer (n = 44 withoutFreeSurefer processing) a total of 508 (75) participants had CRP and 435 (64) had ACTIL6 detected (figure 1) For the analysis for IL6 357 participants wereincluded after excluding 78 participants whose IL6 levels were out of the measurement range A total of 310 (46 of 680) participants completed repeat MRIscans approximately 45 (SD = 09) years after their baseline scans and 230 (74 of 310) of them also received gradient echo (GRE) scans for microbleedsassessment In total 235 (76) had CRP and 204 (66) had IL6ACTmeasured among the 310 participants who received second structural MRI scans and 171(74) had CRP and 145 (63) had IL6ACT measured among those who received GRE scans (figure 1) ACT = alpha 1-antichymotrypsin CRP = C-reactiveprotein IL6 = interleukin-6

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 6 Number 1 | January 2019 3

inter-assay CV = 78) The ACT levels were measured using animmunoperoxidase assay kit (Immunology Consultants Labora-tory Inc Newberg sensitivity of 2518 ngmL both intra-assayand interassayCVs lt10) Laboratory personnel were blinded asto the demographic and clinical status of the study participants

CovariatesEthnicity including Hispanic non-Hispanic black non-Hispanic white or Other based on self-report using the for-mat of the 2000 US census was used as a dummy variable withnon-Hispanic white or other as the reference Sex (male as thereference) smoking status (nonsmoker [as the reference] orever smoked [including current smokers and past smokerswho had ge1 cigaretteday for ge1 year]) and apolipoprotein(APOE) status (presence of either 1 or 2 vs absence of laquo4alleles [as the reference]) were used as dichotomous varia-bles Body mass index (BMI kgm2) was calculated as weightover height squared The presence of depression was definedas ge4 on the 10-item version of the Center for EpidemiologicStudies Depression scale Presence or absence of heart dis-ease diabetes mellitus and hypertension as well as the use ofmedication for the conditions were based on self-reportedinformation and clinical stroke was determined by self-reportneurologic examination or a review of medical records Avascular comorbidity burden score (range 0ndash4) was calculatedby summing these 4 dichotomized vascular comorbidity var-iables Fasting plasma lipids were determined on baselineblood samples using standard techniques

Statistical analyses

Characteristic variables of the participantsCirculating levels of inflammatory biomarkers were trans-formed using logarithm and then standardized ([log10Bio-marker minus mean of log10Biomarker]SD of log10Biomarker)The resulting standardized values were approximately normallydistributed Demographics clinical characteristics vascular andlifestyle factors and brain imaging findings of participants bybiomarker tertiles were compared using analysis of variance forcontinuous variables and the χ2 test for categorical variables

Cross-sectional associations between theinflammatory biomarkers and cerebrovasculardiseaseGeneralized linear models were used to test the associationsbetween the inflammatory biomarkers and baseline WMHand logistic regression models were used for the inflammatorybiomarkers and infarcts or microbleeds status All the modelswere adjusted for age at time of scan sex education ethnicityand APOE laquo4 status In the fully adjusted models we addi-tionally adjusted for health and vascular risk factors includingBMI depression plasma lipids level smoking status andvascular comorbidity burden All covariates were treated astime-independent variables

Longitudinal data analysisThe changes of WMH were calculated as [log10(follow-upWMHICV) minus log10(baseline WMHICV)] Multiple re-gression models were used to test whether inflammatory

biomarkers were associated with differential WMH changeadjusted for time from baseline scan to the follow-up scan inaddition to all the other covariates Cox proportional hazardmodels were used to examine whether baseline inflammatorybiomarkers were associated with the risk of developing in-cident infarcts among those without infarcts at baseline

Supplementary analysesWe excluded 102 (201 of 508) participants with clinicalstroke and repeated the analyses only among participantswithout clinical stroke We also excluded 24 participants whodeveloped incident dementia during follow-up and repeatedthe analysis for microbleeds and infarcts at follow-up In ad-dition we examined whether certain regions were particularlyvulnerable to inflammatory effects by examining WMH in thefrontal parietal occipital and temporal regions separately andby examining lobar and deep microbleeds separately Finallywe examined the relationship between inflammatory bio-markers and CVD by raceethnic groups sex and APOEstatus

All analyses were performed using PASW Statistics 170(formerly SPSS Inc Chicago IL) All p values were based ontwo-sided tests with the statistical significance level setat 005

Data availabilityThe data that support the findings of this study are availableon request from the corresponding author yg2121colum-biaedu Some access restrictions apply to the data used for thecurrent study because of privacy concerns including the fol-lowing (1) the data contain elements that are consideredProtected Health Information under HIPAA regulations and(2) there is a potential risk for linkage of data made publiclyavailable in conjunction with multiple studies based on thesame cohort Although the data on which the manuscript isbased are not publicly available a limited data set is availableunder a standard HIPAA Data Use Agreement subject toreview and approval by the Columbia University PrivacyOfficer

ResultsDemographic and cerebrovascular diseasecharacteristics according to levels ofinflammatory biomarkersThe 3 inflammatory markers positively correlated with eachother30 with correlation coefficients between CRP and IL6between CRP and ACT and between IL6 and ACT being051 039 and 021 respectively all p lt 00001

Those with a higher level of inflammatory biomarkers aremore likely to be women and more likely to be African-Americans (table 1) Those in the highest tertiles were lesslikely to be APOE laquo4 positive (CRP) had higher BMI (CRPand IL6) had worse lipids profiles (CRP and IL6) had morevascular comorbidities (CRP) and had higher education

4 Neurology Neuroimmunology amp Neuroinflammation | Volume 6 Number 1 | January 2019 NeurologyorgNN

Table 1 Demographic clinical and dietary characteristics in relation to CRP IL6 and ACT

Lowest Middle Highest Total p Value p (H vs L)

CRP tertile

Number 169 170 169 508

CRP mgL median (IQR) 203 (122ndash289) 555 (444ndash751) 223 (157ndash373) 559 (287ndash158)

Age at blood collection mean (SD) 7295 (587) 7205 (583) 7253 (542) 7251 (571) 0435 0546

Age at baseline MRI scan mean (SD) 8049 (584) 7941 (56) 7987 (557) 7992 (568) 0199 0305

Age at follow-up mean (SD) 8455 (554) 8362 (511) 8405 (536) 8407 (533) 0559 0576

Education mean (SD) 1092 (48) 1084 (488) 106 (455) 1078 (474) 0810 0533

Female N () 102 (60) 113 (67) 119 (70) 334 (66) 0162 0062

Whites N () 64 (38) 50 (30) 42 (25) 156 (31) 0080 0025

African-American N () 43 (254) 53 (314) 63 (371) 159 (313)

Hispanics N () 61 (361) 61 (361) 62 (365) 184 (362)

Others N () 1 (06) 5 (3) 3 (18) 9 (18)

APOE laquo4 positive N () 54 (32) 40 (24) 33 (19) 127 (25) 0025 0008

BMI kgm2 mean (SD) 2565 (46) 277 (523) 2989 (645) 2777 (574) lt00001 lt00001

Triglycerides (TG) mgdL mean (SD) 1345 (756) 144 (686) 1697 (1112) 1493 (881) 0001 lt00001

Total cholesterol mgdL mean (SD) 1992 (387) 2008 (37) 1981 (414) 1993 (39) 0815 0794

HDL-cholesterol mgdL mean (SD) 518 (141) 505 (159) 446 (126) 49 (146) 0000 lt00001

LDL-cholesterol mgdL mean (SD) 1212 (337) 1215 (319) 1198 (33) 1208 (328) 0877 0704

Ratio of TG to HDL-cholesterol mean (SD) 303 (247) 34 (257) 44 (412) 361 (319) lt00001 lt00001

Heart disease N () 61 (36) 95 (55) 74 (43) 230 (45) 0001 0149

Hypertension N () 135 (79) 145 (85) 149 (88) 429 (84) 0103 0038

Diabetes mellitus N () 28 (16) 45 (26) 62 (36) 135 (26) lt00001 lt00001

Clinical stroke N () 31 (18) 33 (19) 38 (22) 102 (20) 0614 0345

Vascular score mean (SD) 151 (096) 187 (098) 191 (101) 176 (100) lt00001 lt00001

Infarct presence at baseline N () 54 (32) 49 (29) 66 (39) 169 (33) 0123 0173

Infarct presence at follow-up N () 37 (52) 37 (49) 38 (59) 112 (53) 0481 0396

WMHdagger at baseline mean (SD) minus602 (071) minus605 (064) minus597 (067) minus602 (067) 0523 0497

WMHdagger at follow-up mean (SD) 2553 (061) minus548 (045) 2534 (056) minus545 (055) 0088 0032

Increased WMHDagger mean (SD) 063 (037) 060 (045) 064 (038) 062 (04) 0847 0939

Microbleeds at follow-up N () 13 (20) 18 (30) 15 (33) 46 (27) 0225 0104

Lobar microbleeds at follow-up N () 9 (14) 12 (20) 14 (31) 35 (20) 0081 0026

Deep microbleeds at follow-up N () 4 (6) 7 (12) 4 (9) 15 (9) 0539 0572

No of microbleeds mean (SD) 023 (052) 042 (079) 082 (195) 045 (116) 0028 0008

No of lobar microbleeds mean (SD) 017 (048) 03 (074) 073 (18) 036 (108) 0021 0007

No of deep microbleeds mean (SD) 006 (024) 012 (032) 009 (029) 009 (028) 0544 0608

IL6 tertile

Number 119 119 119 357

IL6 pgmL median (IQR) 141 (118ndash165) 243 (204ndash287) 455 (385ndash619) 244 (165ndash391)

Continued

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 6 Number 1 | January 2019 5

Table 1 Demographic clinical and dietary characteristics in relation to CRP IL6 and ACT (continued)

Lowest Middle Highest Total p Value p (H vs L)

Age at blood collection mean (SD) 7238 (571) 7219 (596) 7312 (536) 7256 (568) 0408 0312

Age at baseline MRI scan mean (SD) 7974 (566) 7933 (567) 8011 (512) 7973 (548) 0552 0603

Age at follow-up mean (SD) 8383 (545) 8366 (546) 8425 (474) 839 (522) 0835 0685

Education mean (SD) 1108 (495) 1119 (452) 1079 (458) 1102 (468) 0790 0629

Female N () 73 (62) 75 (63) 89 (74) 237 (66) 0085 0042

Whites N () 46 (39) 38 (319) 29 (242) 113 (317) 0346 0083

African-American N () 32 (271) 37 (311) 44 (367) 113 (317)

Hispanics N () 36 (305) 40 (336) 44 (367) 120 (336)

Others N () 4 (34) 4 (34) 3 (25) 11 (31)

APOE laquo4 positive N () 25 (33) 33 (28) 28 (23) 86 (24) 0486 0691

BMI kgm2 mean (SD) 264 (449) 2766 (491) 2854 (564) 2754 (51) 0006 0001

Triglycerides (TG) mgdL mean (SD) 1367 (733) 157 (782) 1555 (1116) 1496 (896) 0159 0110

Total cholesterol mgdL mean (SD) 2055 (38) 2008 (423) 1997 (366) 202 (39) 0485 0256

HDL-cholesterol mgdL mean (SD) 536 (166) 50 (145) 463 (138) 50 (153) 0001 lt00001

LDL-cholesterol mgdL mean (SD) 125 (33) 1182 (312) 1233 (326) 1222 (323) 0260 0696

Ratio of TG to HDL-cholesterol mean (SD) 308 (258) 365 (251) 402 (409) 358 (316) 0076 0175

Heart disease N () 52 (43) 51 (42) 49 (41) 152 (42) 0923 0694

Hypertension N () 99 (83) 102 (85) 99 (83) 300 (84) 0829 0999

Diabetes mellitus N () 31 (26) 30 (25) 33 (27) 94 (26) 0904 0770

Clinical stroke N () 18 (15) 20 (16) 32 (26) 70 (19) 0047 0026

Vascular score mean (SD) 168 (100) 171 (096) 179 (100) 176 (100) 0670 0393

Infarct presence at baseline N () 33 (28) 40 (34) 47 (40) 120 (34) 0158 0055

Infarct presence at follow-up N () 26 (52) 31 (55) 28 (60) 85 (56) 0754 0453

WMHdagger at baseline mean (SD) minus605 (068) minus602 (072) minus595 (072) minus601 (071) 0513 0265

WMHdagger at follow-up mean (SD) 2559 (062) minus541 (054) 2534 (058) minus545 (059) 0072 0028

Increased WMHDagger mean (SD) 056 (039) 066 (044) 059 (04) 061 (041) 0416 0670

Microbleeds at follow-up N () 12 (31) 13 (28) 10 (26) 35 (28) 0905 0665

Lobar microbleeds at follow-up N () 10 (26) 10 (21) 8 (21) 28 (23) 0858 0634

Deep microbleeds at follow-up N () 4 (10) 4 (9) 5 (13) 13 (10) 0784 0692

No of microbleeds mean (SD) 054 (100) 043 (083) 074 (205) 056 (136) 0576 0524

No of lobar microbleeds mean (SD) 044 (091) 034 (076) 061 (188) 045 (125) 0623 0554

No of deep microbleeds mean (SD) 010 (031) 009 (028) 013 (034) 010 (031) 0788 0682

ACT tertile

Number 145 145 145 435

ACT mgL median (IQR) 158 (146ndash165) 188 (181ndash195) 232 (216ndash272) 188 (165ndash216)

Age at blood collection mean (SD) 7255 (627) 7272 (54) 7356 (578) 7295 (583) 0266 0128

Age at baseline MRI scan mean (SD) 7986 (625) 7984 (533) 8028 (544) 7999 (568) 0747 0522

Age at follow-up mean (SD) 8412 (587) 8464 (528) 8371 (489) 8416 (537) 0632 0666

Continued

6 Neurology Neuroimmunology amp Neuroinflammation | Volume 6 Number 1 | January 2019 NeurologyorgNN

(ACT) compared with those in the lowest tertile of the cor-responding biomarker (table 1)

Those with a higher tertile of any of the inflammatory bio-marker tended to have more CVD (table 1) Comparing theinflammatory marker levels according to the CVD status

revealed a similar relationship ie participants with infarcts atbaseline had higher IL6 levels than those without those withhigher WMH at follow-up had higher ACT levels than thosewith lower WMH and those with microbleeds had higherACT levels than those without microbleeds (table 2 andfigure 2)

Table 1 Demographic clinical and dietary characteristics in relation to CRP IL6 and ACT (continued)

Lowest Middle Highest Total p Value p (H vs L)

Education mean (SD) 1025 (471) 1053 (489) 1137 (436) 1072 (467) 0096 0038

Female N () 84 (58) 105 (71) 101 (71) 290 (67) 0035 0030

Whites N () 42 (292) 41 (277) 52 (364) 135 (31) 0046 0048

African-American N () 46 (319) 42 (284) 53 (371) 141 (324)

Hispanics N () 54 (375) 61 (412) 33 (231) 148 (34)

Others N () 2 (14) 4 (27) 5 (35) 11 (25)

APOE laquo4 positive N () 38 (26) 25 (17) 38 (27) 101 (23) 0081 0972

BMI kgm2 mean (SD) 2768 (565) 2695 (495) 2752 (62) 2738 (563) 0501 0815

Triglycerides (TG) mgdL mean (SD) 1453 (823) 1443 (611) 151 (1087) 1469 (864) 0789 0583

Total cholesterol mgdL mean (SD) 1967 (408) 2011 (405) 2008 (371) 1995 (395) 0581 0380

HDL-cholesterol mgdL mean (SD) 502 (167) 502 (143) 494 (14) 499 (15) 0861 0627

LDL-cholesterol mgdL mean (SD) 1164 (304) 122 (359) 1224 (328) 1203 (331) 0253 0138

Ratio of TG to HDL-cholesterol mean (SD) 352 (293) 331 (212) 366 (384) 35 (305) 0644 0715

Heart disease N () 56 (38) 64 (44) 64 (44) 184 (42) 0547 0340

Hypertension N () 115 (79) 124 (85) 127 (87) 366 (84) 0133 0058

Diabetes mellitus N () 31 (21) 46 (31) 41 (28) 118 (27) 0131 0174

Clinical stroke N () 29 (20) 28 (19) 32 (22) 89 (20) 0832 0666

Vascular score mean (SD) 159 (104) 181 (100) 182 (096) 176 (100) 0095 0053

Infarct presence at baseline N () 47 (33) 46 (32) 50 (35) 143 (33) 0873 0709

Infarct presence at follow-up N () 32 (46) 28 (54) 36 (62) 96 (54) 0210 0077

WMHdagger at baseline mean (SD) minus604 (07) minus595 (067) minus595 (073) minus598 (07) 0476 0288

WMHdagger at follow-up mean (SD) 256 (062) 254 (053) 2534 (058) 2546 (059) 0028 0013

Increased WMHDagger mean (SD) 054 (039) 062 (038) 070 (042) 061 (04) 0085 0027

Microbleeds at follow-up N () 10 (20) 7 (15) 23 (49) 40 (28) lt00001 0003

Lobar microbleeds at follow-up N () 8 (16) 7 (15) 18 (38) 33 (23) 0008 0013

Deep microbleeds at follow-up N () 2 (4) 2 (4) 10 (21) 14 (10) 0005 0010

No of microbleeds mean (SD) 030 (076) 021 (054) 113 (197) 054 (131) 0001 0001

No of lobar microbleeds mean (SD) 026 (075) 017 (043) 091 (184) 044 (120) 0004 0002

No of deep microbleeds mean (SD) 004 (020) 004 (020) 021 (041) 010 (030) 0004 0004

Abbreviations ACT = alpha 1-antichymotrypsin ANOVA = analysis of variance BMI = body mass index CRP = C-reactive protein ICV = total intracranialvolume IL6 = interleukin-6 WMH = white matter hyperintensitiesp Values were from the χ2 test for categorical variables or the one-way ANOVA test for continuous variables p (H vs L) is the p value corresponding to thecomparison of the highest (H) to the lowest (L) tertiledagger ldquoWMHrdquo was calculated as log10(WMH volumeintracranial volume)Dagger ldquoIncreased WMHrdquo was calculated as the increase from the first to the second WMH value [log10(follow-up WMHICV) minus s(baseline WMHICV)]Bold values indicate statistical significant difference with p lt 005

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 6 Number 1 | January 2019 7

Table 2 Average CRP IL6 and ACT levels according to the cerebrovascular disease status

CRP IL6 ACT

N

z-scoremean(SD)

Geometricmean (SD)

pValue N

z-scoremean(SD)

Geometricmean (SD)

pValue N

z-scoremean(SD)

Geometricmean (SD)

pValue

Infarct atbaseline

No 339 minus002(098)

606 (33) 0388 237 2008(098)

241 (179) 0036 292 minus004(099)

1907 (125) 0357

Yes 169 006(106)

669 (36) 120 015(103)

277 (184) 143 006(102)

1948 (126)

Total 508 0 (101) 626 (34) 357 0 (1) 253 (181) 435 0 (1) 1920 (125)

Infarct atfollow-up

No 98 minus011(095)

548 (32) 0556 68 minus012(093)

235 (174) 0236 83 minus022(094)

1829 (124) 0168

Yes 112 minus003(099)

604 (33) 85 007(102)

264 (183) 96 minus001(104)

1916 (127)

Total 210 minus007(098)

577 (33) 153 minus002(099)

251 (179) 179 minus011 (1) 1875 (125)

Totalmicrobleeds

No 125 minus024(103)

466 (35) 0063 89 minus005(096)

245 (176) 0709 105 2023(097)

1823 (125) 0019

Yes 46 009(097)

692 (32) 35 002(098)

256 (179) 40 020(103)

2012 (126)

Total 171 minus015(102)

519 (34) 124 minus003(096)

248 (177) 145 2011(100)

1873 (126)

WMH atbaseline

Low 256 minus005(103)

586 (35) 0104 170 minus003 (1) 248 (181) 0581 208 minus008(09)

1887 (123) 0084

High 234 01 (099) 700 (33) 175 003(098)

257 (179) 214 009 (11) 1961 (128)

Total 490 002(101)

638 (34) 345 0 (099) 252 (18) 422 0 (101) 1924 (126)

WMH atfollow-up

Low 115 minus015(098)

520 (33) 0189 83 minus015(098)

232 (179) 0166 98 2031(086)

1792 (122) 0005

High 111 002(097)

639 (33) 80 006(093)

263 (174) 95 008(105)

1959 (127)

Total 226 minus007(098)

575 (33) 163 minus004(096)

246 (177) 193 2012(098)

1873 (125)

Infarctduringfollow-updagger

No 98 minus011(095)

548 (32) 088 68 minus012(093)

235 (174) 0674 83 minus022(094)

1829 (124) 0995

Yes 42 minus008(089)

566 (29) 32 minus004(093)

248 (174) 37 minus001(104)

1828 (128)

Continued

8 Neurology Neuroimmunology amp Neuroinflammation | Volume 6 Number 1 | January 2019 NeurologyorgNN

Cross-sectional association betweeninflammatory biomarkers andcerebrovascular diseaseIn adjusted models 1 SD increase in circulating log-transformed CRP level was associated with increased oddsof having microbleeds (table 3) and the association remainedsignificant in the fully adjusted model (odds ratio [OR] =180 95 confidence interval [CI] = 107ndash303 p = 0026)

One SD increase in circulating log-transformed IL6 level wasassociated with increased odds of having infarcts at initial(baseline) scan visit (table 3) and the association remainedsignificant in the fully adjusted model (OR = 130 95 CI =101ndash166 p = 0039)

ACT seemed to be associated with increased odds of infarctsat the follow-up visit (table 3) but the association was nolonger significant in the fully adjusted model An increasedACT level was associated with increased odds of havingmicrobleeds (table 3) and remained so in the fully adjustedmodel (OR comparing 3rd to 1st tertile = 374 95 CI =128ndash1090 p = 0016 p-trend = 0011) ACT was also as-sociated with larger WMH at the follow-up visit (table 3) andremained associated in the fully adjusted model (b = 0087p = 0041)

Longitudinal association betweeninflammatory biomarkers andcerebrovascular diseaseA higher ACT level was associated with greater accumulationof WMH during follow-up (table 3) In the fully adjustedmodel a 1-SD increase in log10aACT was associated with0083 (p = 0011) increase in log10WMH during follow-up(ie approximately 121 times increase in WMH volume)None of the inflammatory biomarkers was associated with riskof incident infarcts (table 3)

Supplementary analysesWhen participants with clinically diagnosed stroke were ex-cluded log10CRP remained associated with microbleeds(OR = 177 95 CI = 108ndash292 p = 0024) log10IL6 withbaseline infarcts (OR = 140 95 CI = 105ndash187 p = 0022)and log10ACTwithmicrobleeds (OR= 201 95CI = 115ndash350p = 0015) larger WMH at follow-up visit (b = 0101 p = 0044)

and greater accumulation of WMH (b = 0119 p = 0001) all inthe fully adjusted models

Excluding 24 participants who developed dementia during thefollow-up adjusted models showed that higher log10CRP andlog10ACT remained significantly associated with increasedodds of microbleeds with OR (95 CI) = 149 (103ndash214)p = 003 and 162 (108ndash242) p = 002 respectively andlog10ACTwas also associated withWMHburden at follow-up(b = 0116 p = 0006) and accumulation (b = 0063 p =0048)

In the fully adjusted model higher log10ACT was associatedwith larger WMH at the follow-up visit in the frontal (b =0114 p = 0017) parietal (b = 0127 p = 0028) and temporal(b = 0140 p = 0029) regions and was associated with greateraccumulation of WMH during follow-up in the frontal (b =0102 p = 0037) parietal (b = 0117 p = 0052) and temporal(b = 0133 p = 0047) regions

One SD increase in log10CRP level was associated with in-creased odds of lobar microbleeds (OR = 181 95 CI =112ndash294 p = 0016 p-trend = 0019 for tertiles) whereasACT was associated with deep microbleeds (OR = 272 95CI = 111ndash666 p = 0028 p-trend = 0007 for tertiles) in thefully adjusted model

The associations of inflammatory biomarkers with CVD ingeneral did not vary by ethnic groups sex or APOE statusThe only exception was that the association between ACTand WMH accumulation during the follow-up (b for in-teraction between ACT and Hispanic ethnicity = minus0143 p =0068) was significant for whites (b = 0124 p = 0042) andAfrican-Americans (b = 0131 p = 0037) but not for His-panics (b = minus0018 p = 0669)

DiscussionIn this multiethnic nondemented elderly population wefound that higher circulating proinflammatory biomarkerswere associated with more CVD Our finding that IL6 wasassociated with the presence of silent brain infarcts is con-sistent with some previous studies1012 but not others1422

Table 2 Average CRP IL6 and ACT levels according to the cerebrovascular disease status (continued)

CRP IL6 ACT

N

z-scoremean(SD)

Geometricmean (SD)

pValue N

z-scoremean(SD)

Geometricmean (SD)

pValue N

z-scoremean(SD)

Geometricmean (SD)

pValue

Total 140 minus010(093)

553 (31) 100 minus009(093)

239 (173) 120 minus011 (1) 1828 (125)

Abbreviations ACT = alpha 1-antichymotrypsin CRP = C-reactive protein IL6 = interleukin-6 WMH = white matter hyperintensity Low and high WMHs were based on the median split p Values from t-testsdagger Among those without infarct at baseline onlyBold values indicate statistical significant difference with p lt 005

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 6 Number 1 | January 2019 9

Although some studies found CRP to be associated withinfarcts101213 our study along with a few otherstudies914ndash1619 did not replicate this finding Studies alsofound that CRP or intercellular adhesion molecule-1(sICAM-1) levels were associated with incident lacunarinfarcts923 Overall there is some evidence showing a cross-sectional relationship between inflammatory biomarkers and

CVD but longitudinal evidence is limited The potential roleof inflammatory markers other than CRP or IL6 in CVD isworth further exploration

A few studies found that higher CRP91019 IL61026 or both22

were associated with more white matter lesions Two otherstudies found that the association between CRP and WMH

Figure 2 Association of circulating concentrations of inflammatory biomarkers and cerebrovascular disease

Association of circulating concentrations of C-reactive protein (CRP) interleukin-6 (IL6) and alpha 1-antichymotrypsin (ACT) with infarct and microbleedsstatus (AandashAc) and white matter hyperintensity (WMH) level (BandashBc) p Values in the figure were from unadjusted t-test (for infarcts and microbleeds) orfrom Pearson correlation (for WMH) WMH = white matter hyperintensity p lt 005

10 Neurology Neuroimmunology amp Neuroinflammation | Volume 6 Number 1 | January 2019 NeurologyorgNN

Table 3 Association between inflammatory biomarkers and brain structural MRI measures

Logistic regression models

CRP IL6 ACT

OR (95 CI) p Value OR (95 CI) p Value OR (95 CI) p Value

Infarcts at baseline

1 SD increase 110 (089ndash130) 0333 128 (102ndash160) 0033 112 (091ndash138) 0290

Middle tertile 090 (056ndash144) 0657 130 (074ndash227) 0361 100 (061ndash166) 0995

Highest tertile 142 (090ndash224) 0136 174 (100ndash303) 0052 112 (068ndash185) 0646

p-trend 0130 0051 0646

Infarcts at follow-up

1 SD increase 107 (080ndash144) 0642 117 (083ndash165) 0370 132 (096ndash181) 0089

Middle tertile 097 (050ndash189) 0932 107 (049ndash235) 0860 144 (068ndash307) 0341

Highest tertile 133 (065ndash270) 0432 122 (053ndash283) 0643 217 (103ndash457) 0042

p-trend 0446 0645 0042

Total microbleeds

1 SD increase 146 (102ndash209) 0041 118 (076ndash181) 0468 165 (111ndash246) 0013

Middle tertile 186 (081ndash432) 0146 099 (037ndash262) 0983 087 (029ndash259) 0798

Highest tertile 233 (093ndash585) 0071 100 (030ndash257) 0996 462 (178ndash1199) 0002

p-trend 0108 0996 0092

Linear regression models

CRP IL6 ACT

b p Value b p Value b p Value

WMH at baseline

1 SD increase 0026 0368 minus0002 0962 0060 0072

Middle tertile minus0032 0654 0022 0798 0108 0179

Highest tertile 0004 0950 0028 0753 0090 0262

p-trend 0945 0752 0263

WMH at follow-up

1 SD increase 0052 0151 0054 0245 0103 0012

Middle tertile 0050 0545 0131 0204 0184 0057

Highest tertile 0116 0190 0158 0157 0218 0025

p-trend 0191 0150 0022

WMH changedagger

1 SD increase 0002 0959 0041 0248 0062 0041

Middle tertile minus0031 0639 0086 0282 0081 0260

Highest tertile 0028 0697 0053 0538 0148 0041

p-trend 0728 0519 0040

Cox proportional hazard models

CRP IL6 ACT

HR (95 CI) p Value HR (95 CI) p Value HR (95 CI) p Value

Incident infarctsDagger

1 SD increase 089 (064ndash124) 0491 082 (049ndash137) 0446 134 (098ndash183) 0070

Continued

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 6 Number 1 | January 2019 11

was attenuated by the adjustment of cardiovascular13 orother factors11 Our study and several others did not findCRP15162427 or other cytokines142527 to be associated withthe severity of WMH volume One previous study9 foundthat CRP was associated with more progression of whitematter lesions However other studies152223 did not findsuch a relationship We found that ACT but not CRP orIL6 was associated with WMH severity cross-sectionally atthe follow-up MRI scan visit and with progression of WMHlongitudinally Two recent studies found that sICAM-1 butnot CRP was associated with the presence21 or progressionof WMH23 Therefore inflammatory biomarkers otherthan CRP or IL6 may be relevant for WMH Interestinglyboth ACT and sICAM-1 play a role in cell adhesion andendothelial dysfunction which may contribute to thepathogenesis of WMH33 Similar to our study other bi-racial studies1019 found that there was no evidence for ef-fect of inflammatory biomarkers on WMH to be differentbetween white and black adults More studies are needed toexamine the effect of inflammatory biomarkers on WMHamong Hispanics

Only a few studies have examined the relationship betweencirculating inflammatory biomarkers with microbleeds1319ndash21

but the results are inconsistent Two studies found that CRPwas not associated with microbleeds1319 but a Japanese studyfound that CRP IL6 and IL18 were all associated with thepresence of microbleeds20 The Framingham offspring studyfound that tumor necrosis factor receptor 2 and myeloper-oxidase were associated with increased odds of havingmicrobleeds especially deep microbleeds21 We found thatCRP was associated with increased odds of lobar but not deepmicrobleeds whereas ACT was associated with increasedodds of deep but not lobar microbleeds The distribution ofmicrobleeds is believed to reflect 2 distinct underlying types ofmicroangiopathy34 lobar microbleeds are considered likely tobe attributable to cerebral amyloid angiopathy (CAA)whereas deep microbleeds are considered to be due to hy-pertensive arteriopathy Therefore CRP and ACT may rep-resent 2 different pathways toward small hemorrhagic lesionsin the brain This is in line with our previous findings that

increased CRP but not ACT was associated with smaller graymatter volume30 as it has been shown that CAA-related butnot CAA-unrelated microbleeds are associated gray matteratrophy35 Nevertheless such differential roles of CRP ACTor other inflammatory biomarkers in microbleeds need to beconfirmed in future studies

Growing evidence suggest that inflammatory cytokines areassociated with increased risk of developing stroke cardio-vascular disease and dementia6 In addition our study andother longitudinal studies923 found inflammatory biomarkerswere associated with progression of CVD (particularlyWMH) suggesting cytokines acute-phase proteins endo-thelial cell adhesive molecules and other immune-relatedproteins may play an active role in building up or contributingto the vascular injuries in the brain However current evi-dence has been inconsistent The exact reason for this is un-known but it might be due to differences in age prevalence ofvascular risk factors ethnicity MRI acquisition techniquesand analytic strategies In addition there are few longitudinalstudies and some existing ones that failed to establish thislongitudinal relationship Therefore an alternative explana-tion of the results might be that elevated inflammatory bio-markers may be a marker of the inflammation as a result ofvascular or other pathologic injury Generation of cytokinescan be upregulated in the brains of patients with stroke36 andCAA37 probably representing an immune response stimu-lated by the vascular deposits of β-amyloid and other injuriesIn addition increased circulating levels of peripheral proin-flammatory cytokines are found in patients with AD com-pared with controls38 Overall it is possible that bothdirections exist and create a vicious circle for a progressiveaccumulation of vascular and neurodegenerative damages aswell as elevated inflammatory responses Either way furtherunderstanding the role of inflammatory biomarkers in thepresence and progression of CVD is important because it mayhelp elucidate the pathogenesis of CVD and allow the de-velopment of immune-modulating intervention measures forCVD and related neurologic outcomes or it may help de-termine how inflammatory markers can be used to monitordisease progression

Table 3 Association between inflammatory biomarkers and brain structural MRI measures (continued)

Cox proportional hazard models

CRP IL6 ACT

HR (95 CI) p Value HR (95 CI) p Value HR (95 CI) p Value

Middle tertile 092 (043ndash200) 0838 072 (030ndash176) 0473 184 (076ndash442) 0176

Highest tertile 060 (026ndash137) 0224 065 (024ndash176) 0394 217 (091ndash519) 0080

p-trend 0222 0400 0071

Abbreviations ACT = alpha 1-antichymotrypsin CI = confidence interval CRP = C-reactive protein ICV = intracranial volume IL6 = interleukin-6 OR = oddsratio WMH = white matter hyperintensity All models adjusted for age at time of scan sex ethnicityrace APOE status and for WMH only also ICVdagger Additionally adjusted for duration between scansDagger Limit to 158 participants without infarcts at baseline onlyBold values indicate statistical significant difference with p lt 005

12 Neurology Neuroimmunology amp Neuroinflammation | Volume 6 Number 1 | January 2019 NeurologyorgNN

Although we adjusted for several key factors we cannotcompletely exclude the possibility that our results are subjectto residual confounding due to unmeasured factors We onlyexamined 3 of the many inflammatory markers in our studywhereas the role of many other inflammatory markers is alsoworthy of examination We only measured the inflammatorymarkers once and they may not represent the long-term av-erage level of the subject Although WMH overall is stronglyassociated with vascular factors and viewed as a reflection ofsmall vessel CVD39 the pathogenesis of WMHs is not fullyunderstood and could be multifactorial40 Thus anotherlimitation of our study is that we did not subclassifyWMHs bytheir putative etiologies Our studymight be underpowered toexamine interactions stratified analyses in subgroups and thelongitudinal relationship We did not correct for multiplecomparison as we consider our study to be exploratory innature and need replication in an external sample

Our study has many strengths We added new evidence to thisline of research by including several key inflammatory mark-ers both cross-sectional and longitudinal analyses an elderlypopulation with advanced age and different cerebrovascularfeatures In addition our study population comprised multi-ple ethnic groups making the results more generalizable tothe increasingly diverse US population Finally several keypotential confounding factors were adjusted in the analyses

Among the old participants of our study higher levels ofperipheral CRP and IL6 were associated with increased oddsof microbleeds and infarcts respectively and increased ACTwas associated with cerebrovascular events including in-creased odds of infarcts and microbleeds and larger burdenand more accumulation of WMH More studies especiallylongitudinal studies with a wide range of inflammatory bio-markers are needed to clarify whether inflammatory bio-markers play an etiologic role in CVD or represent anepiphenomenon of CVD

Author contributionsY Gu design and conceptualized the study analysis and in-terpretation and drafted the manuscript for intellectualcontent J Gutierrez-Contreras IB Meier VA Guzman JJManly and N Schupf acquisition of data and revised themanuscript for intellectual content AM Brickman majorrole in the acquisition of data interpreted the data and re-vised the manuscript for intellectual content R Mayeuxdesign and conceptualized the study acquisition of datastudy supervision and revised the manuscript for intellectualcontent

AcknowledgmentThis manuscript has been reviewed byWHICAP investigatorsfor scientific content and consistency of data interpretationwith previous WHICAP Study publications The authorsacknowledge the WHICAP study participants and theWHICAP research and support staff for their contributionsto this study

Study fundingSupported by NIH PO1AG007232 R01AG037212RF1AG054023 UL1TR001873 AG042483 and AG034189

DisclosureY Gu received travel funding from California Walnut Com-mission and received research support from NIH J Gutierrezserved as a review editor for PLoS One and Frontiers in Neu-rology provided medico-legal consultancies and receivedresearch support from the NIH and NINDS IB Meier re-ceived research support from the Synapsis Foundation andVELUX Stiftung VA Guzman received research supportfrom the NIH and NIMHD JJ Manly received travel fundingfrom the Alzheimerrsquos Association served as an associate ed-itor for the Journal of the International Neuropsychological So-ciety consulted for National Academies of Medicine andNational Academy of Sciences and received research supportfrom NIA NIDDK and NINDS N Schupf received researchsupport from the NIH and Alzheimerrsquos Association AMBrickman served on the scientific advisory board of KeystoneHeart an NIA-supported study served on the editorial boardof the Journal of the International Neuropsychological Societyand Neuropsychology Review served as an associate editor ofNeurodegenerative Diseases holds a patent for technologies forwhite matter hyperintensity quantification consulted forProPhase Keystone Heart and Cognition Therapeutics andreceived research support from NIH Columbia UniversityAlzheimerrsquos Association and Mary E Groff Surgical MedicalResearch and Education Charitable Trust R Mayeux receivedresearch support from NIA Full disclosure form informationprovided by the authors is available with the full text of thisarticle at NeurologyorgNN

Received March 28 2018 Accepted in final form October 12 2018

References1 de Leeuw FE de Groot JC Achten E et al Prevalence of cerebral white matter lesions

in elderly people a population based magnetic resonance imaging study The Rot-terdam Scan Study J Neurol Neurosurg Psychiatry 2001709ndash14

2 Greenberg SM Vernooij MW Cordonnier C et al Cerebral microbleeds a guide todetection and interpretation Lancet Neurol 20098165ndash174

3 Fisher CM Lacunes small deep cerebral infarcts Neurology 196515774ndash7844 Vermeer SE Hollander M van Dijk EJ Hofman A Koudstaal PJ Breteler MM Silent

brain infarcts and white matter lesions increase stroke risk in the general populationthe Rotterdam Scan Study Stroke 2003341126ndash1129

5 Brickman AM Zahodne LB Guzman VA et al Reconsidering harbingers ofdementia progression of parietal lobe white matter hyperintensities predictsAlzheimerrsquos disease incidence Neurobiol Aging 20153627ndash32

6 Holmes C Review systemic inflammation and Alzheimerrsquos disease NeuropatholAppl Neurobiol 20133951ndash68

7 Ross R Atherosclerosismdashan inflammatory disease N Engl J Med 1999340115ndash1268 Saji N Toba K Sakurai T Cerebral small vessel disease and arterial stiffness tsunami

effect in the brain Pulse (Basel) 20163182ndash1899 van Dijk EJ Prins ND Vermeer SE et al C-reactive protein and cerebral small-vessel

disease the Rotterdam Scan Study Circulation 2005112900ndash90510 Fornage M Chiang YA OrsquoMeara ES et al Biomarkers of inflammation and MRI-

defined small vessel disease of the brain the cardiovascular health study Stroke 2008391952ndash1959

11 Wright CB Moon Y Paik MC et al Inflammatory biomarkers of vascular risk ascorrelates of leukoaraiosis Stroke 2009403466ndash3471

12 Hoshi T Kitagawa K Yamagami H Furukado S Hougaku H Hori M Relations ofserum high-sensitivity C-reactive protein and interleukin-6 levels with silent braininfarction Stroke 200536768ndash772

13 Mitaki S Nagai A Oguro H Yamaguchi S C-reactive protein levels are associatedwith cerebral small vessel-related lesions Acta Neurol Scand 201613368ndash74

14 Baune BT Ponath G Rothermundt M Roesler A Berger K Association betweencytokines and cerebral MRI changes in the aging brain J Geriatr Psychiatry Neurol20092223ndash34

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 6 Number 1 | January 2019 13

15 Schmidt R Schmidt H Pichler M et al C-reactive protein carotid atherosclerosisand cerebral small-vessel disease results of the Austrian Stroke Prevention StudyStroke 2006372910ndash2916

16 Wada M Nagasawa H Kurita K et al Cerebral small vessel disease and C-reactiveprotein results of a cross-sectional study in community-based Japanese elderlyJ Neurol Sci 200826443ndash49

17 Jefferson AL Massaro JM Wolf PA et al Inflammatory biomarkers are associatedwith total brain volume the Framingham Heart Study Neurology 2007681032ndash1038

18 Windham BG Wilkening SR Lirette ST et al Associations between inflammationand physical function in African Americans and European Americans with prevalentcardiovascular risk factors J Am Geriatr Soc 2016641448ndash1455

19 Walker KA Power MC Hoogeveen RC et al Midlife systemic inflammation late-lifewhite matter integrity and cerebral small vessel disease the atherosclerosis risk incommunities study Stroke 2017483196ndash3202

20 Miwa K Tanaka M Okazaki S Furukado S Sakaguchi M Kitagawa K Relations ofblood inflammatory marker levels with cerebral microbleeds Stroke 2011423202ndash3206

21 Shoamanesh A Preis SR Beiser AS et al Inflammatory biomarkers cerebralmicrobleeds and small vessel disease Framingham Heart Study Neurology 201584825ndash832

22 Satizabal CL Zhu YC Mazoyer B Dufouil C Tzourio C Circulating IL-6 and CRPare associated with MRI findings in the elderly the 3C-Dijon Study Neurology 201278720ndash727

23 Umemura T Kawamura T Umegaki H et al Endothelial and inflammatory markersin relation to progression of ischaemic cerebral small-vessel disease and cognitiveimpairment a 6-year longitudinal study in patients with type 2 diabetes mellitusJ Neurol Neurosurg Psychiatry 2011821186ndash1194

24 Avci AY Lakadamyali H Arikan S Benli US Kilinc M High sensitivity C-reactiveprotein and cerebral white matter hyperintensities on magnetic resonance imaging inmigraine patients J Headache Pain 2015169

25 Aribisala BS Wiseman S Morris Z et al Circulating inflammatory markers are as-sociated with magnetic resonance imaging-visible perivascular spaces but not directlywith white matter hyperintensities Stroke 201445605ndash607

26 Nagai K Kozaki K Sonohara K Akishita M Toba K Relationship betweeninterleukin-6 and cerebral deep white matter and periventricular hyperintensity inelderly women Geriatr Gerontol Int 201111328ndash332

27 Silbert LC Lahna D Promjunyakul NO et al Risk factors associated with corticalthickness and white matter hyperintensities in dementia free Okinawan elderlyJ Alzheimers Dis 201863365ndash372

28 Sacco RL Boden-Albala B Gan R et al Stroke incidence among white black andHispanic residents of an urban community the Northern Manhattan Stroke StudyAm J Epidemiol 1998147259ndash268

29 Licastro F Parnetti L Morini MC et al Acute phase reactant alpha 1-anti-chymotrypsin is increased in cerebrospinal fluid and serum of patients with probableAlzheimer disease Alzheimer Dis Assoc Disord 19959112ndash118

30 Gu Y Vorburger R Scarmeas N et al Circulating inflammatory biomarkers in relationto brain structural measurements in a non-demented elderly population Brain BehavImmun 201765150ndash160

31 Brickman AM Schupf N Manly JJ et al Brain morphology in older African Ameri-cans Caribbean Hispanics and Whites from northern Manhattan Arch Neurol 2008651053ndash1061

32 Brickman AM Sneed JR Provenzano FA et al Quantitative approaches for assessment ofwhite matter hyperintensities in elderly populations Psychiatry Res 2011193101ndash106

33 Huang Y Zhang W Lin L et al Is endothelial dysfunction of cerebral small vesselresponsible for white matter lesions after chronic cerebral hypoperfusion in ratsJ Neurol Sci 201029972ndash80

34 Kakar P Charidimou A Werring DJ Cerebral microbleeds a new dilemma in strokemedicine JRSM Cardiovasc Dis 201212048004012474754

35 Samuraki M Matsunari I Yoshita M et al Cerebral amyloid angiopathy-relatedmicrobleeds correlate with glucose Metabolism and brain volume in Alzheimerrsquosdisease J Alzheimers Dis 201548517ndash528

36 Tarkowski E Rosengren L Blomstrand C et al Intrathecal release of pro- and anti-inflammatory cytokines during stroke Clin Exp Immunol 1997110492ndash499

37 Eng JA Frosch MP Choi K Rebeck GW Greenberg SM Clinical manifestations ofcerebral amyloid angiopathy-related inflammation Ann Neurol 200455250ndash256

38 Fillit H Ding WH Buee L et al Elevated circulating tumor necrosis factor levels inAlzheimerrsquos disease Neurosci Lett 1991129318ndash320

39 Wardlaw JM Smith EE Biessels GJ et al Neuroimaging standards for research intosmall vessel disease and its contribution to ageing and neurodegeneration LancetNeurol 201312822ndash838

40 Gouw AA Seewann A van der Flier WM et al Heterogeneity of small vessel diseasea systematic review of MRI and histopathology correlations J Neurol NeurosurgPsychiatry 201182126ndash135

14 Neurology Neuroimmunology amp Neuroinflammation | Volume 6 Number 1 | January 2019 NeurologyorgNN

DOI 101212NXI000000000000052120196 Neurol Neuroimmunol Neuroinflamm

Yian Gu Jose Gutierrez Irene B Meier et al adults

Circulating inflammatory biomarkers are related to cerebrovascular disease in older

This information is current as of November 14 2018

Academy of Neurology All rights reserved Online ISSN 2332-7812Copyright copy 2018 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the AmericanPublished since April 2014 it is an open-access online-only continuous publication journal Copyright

is an official journal of the American Academy of NeurologyNeurol Neuroimmunol Neuroinflamm

ServicesUpdated Information amp

httpnnneurologyorgcontent61e521fullhtmlincluding high resolution figures can be found at

References httpnnneurologyorgcontent61e521fullhtmlref-list-1

This article cites 40 articles 12 of which you can access for free at

Subspecialty Collections

httpnnneurologyorgcgicollectionall_immunologyAll Immunology

httpnnneurologyorgcgicollectionall_imagingAll Imaging

httpnnneurologyorgcgicollectionall_epidemiologyAll epidemiology a

httpnnneurologyorgcgicollectionall_cognitive_disorders_dementiAll Cognitive DisordersDementia

okehttpnnneurologyorgcgicollectionall_cerebrovascular_disease_strAll Cerebrovascular diseaseStrokefollowing collection(s) This article along with others on similar topics appears in the

Permissions amp Licensing

httpnnneurologyorgmiscaboutxhtmlpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in

Reprints

httpnnneurologyorgmiscaddirxhtmlreprintsusInformation about ordering reprints can be found online

Academy of Neurology All rights reserved Online ISSN 2332-7812Copyright copy 2018 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the AmericanPublished since April 2014 it is an open-access online-only continuous publication journal Copyright

is an official journal of the American Academy of NeurologyNeurol Neuroimmunol Neuroinflamm

ServicesUpdated Information amp

httpnnneurologyorgcontent61e521fullhtmlincluding high resolution figures can be found at

References httpnnneurologyorgcontent61e521fullhtmlref-list-1

This article cites 40 articles 12 of which you can access for free at

Subspecialty Collections

httpnnneurologyorgcgicollectionall_immunologyAll Immunology

httpnnneurologyorgcgicollectionall_imagingAll Imaging

httpnnneurologyorgcgicollectionall_epidemiologyAll epidemiology a

httpnnneurologyorgcgicollectionall_cognitive_disorders_dementiAll Cognitive DisordersDementia

okehttpnnneurologyorgcgicollectionall_cerebrovascular_disease_strAll Cerebrovascular diseaseStrokefollowing collection(s) This article along with others on similar topics appears in the

Permissions amp Licensing

httpnnneurologyorgmiscaboutxhtmlpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in

Reprints

httpnnneurologyorgmiscaddirxhtmlreprintsusInformation about ordering reprints can be found online

Academy of Neurology All rights reserved Online ISSN 2332-7812Copyright copy 2018 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the AmericanPublished since April 2014 it is an open-access online-only continuous publication journal Copyright

is an official journal of the American Academy of NeurologyNeurol Neuroimmunol Neuroinflamm