Circulating Inflammatory Markers Are Associated With Magnetic Resonance Imaging-Visible Perivascular Spaces But Not Directly With White Matter Hyperintensities
Background and Purpose—White matter hyperintensities (WMH) and perivascular spaces (PVS) are features of small vessel disease, found jointly on MRI of older people. Inflammation is a prominent pathological feature of small vessel disease. We examined the association between inflammation, PVS, and WMH in the Lothian Birth Cohort 1936 (N=634).
Methods—We measured plasma fibrinogen, C-reactive protein, and interleukin-6 and rated PVS in 3 brain regions. We measured WMH volumetrically and visually using the Fazekas scale. We derived latent variables for PVS, WMH, and Inflammation from measured PVS, WMH, and inflammation markers and modelled associations using structural equation modelling.
Results—After accounting for age, sex, stroke, and vascular risk factors, PVS were significantly associated with WMH (β=0.47; P<0.0001); Inflammation was weakly but significantly associated with PVS (β=0.12; P=0.048), but not with WMH (β=0.02; P=NS).
Conclusions—Circulating inflammatory markers are weakly associated with MR-visible PVS, but not directly with WMH. Longitudinal studies should examine whether visible PVS predate WMH progression and whether inflammation modulators can prevent small vessel disease.
The pathogenesis of small vessel disease (SVD) is poorly understood. It is thought to result from arteriolosclerosis in the penetrating arterioles leading to ischemia with diffuse rarefaction, necrosis, and cavitation in the subcortical tissues seen as white matter hyperintensities (WMH) and lacunes on MRI.1 Although perivascular inflammation is a prominent well-established pathological feature in WMH2 and lacunar stroke,3 the nature of inflammation and its role in the pathogenesis of SVD is uncertain.
Perivascular spaces (PVS), another marker of SVD, are thought to be associated with elevated plasma inflammatory markers in older subjects4 and in patients with small subcortical stroke.5 The association between plasma markers of inflammation and WMH is less clear.6 We tested if inflammatory markers had a direct and potentially causal relationship with WMH, or if any relationship was via an association with PVS.
Participants are members of the Lothian Birth Cohort 1936 (LBC1936),7 all born in 1936. At mean age of 73 years, inflammatory markers (N=866) were measured and brain MRI was performed (N=700). Participants provided demographic information, medical history, and informed consent.
Measurement of Inflammatory Markers
C-reactive protein (CRP) and interleukin-6 (IL-6) were analyzed using high-sensitivity ELISA (R&D Systems, Oxford, United Kingdom). Fibrinogen was measured using an automated Clauss assay (TOPS coagulometer; Instumentation Laboratory, Warrington, United Kingdom). See the online-only Data Supplement for details.
Brain MRI data were acquired on a 1.5-T GE Signa Horizon HDx scanner (GE, Milwaukee, WI) and included8 T1-W, T2-W, T2*-W, and fluid attenuated inversion recovery brain imaging.
WMH were segmented using MCMxxxVI9 and visually rated by a neuroradiologist on fluid attenuated inversion recovery images using the Fazekas scale.10 Another neuroradiologist crosschecked 20%. The intraclass correlation coefficient was 0.96. Intracranial volume was extracted using Analyze 9.0.
PVS were rated in the hippocampus, basal ganglia, and centrium semiovale. Twenty percent were crosschecked by another neuroradiologist. Intra- and interrater κ statistics ranged from 0.68 to 0.90. See the online-only Data Supplement for details.
Bivariate associations were assessed between markers of inflammation, WMH, PVS, and covariates using Pearson correlation implemented in IBM SPSS version 19.0 (New York, NY). Covariates were sex, age at scanning, and history of cardiovascular disease, diabetes mellitus, hypertension, smoking, hypercholesterolemia, and stroke.
Multivariate associations were investigated using structural equation modelling,11 implemented in Amos 18.0.0 (Amos Development Corporation, FL; see the online-only Data Supplement for details). This allowed derivation of latent variables for WMH, PVS, and inflammation (WMH, PVS, and Inflammation, respectively) and also the assessment of relationships between latent variables while including covariates.
Of the 700 subjects who underwent brain MRI, 66 had incomplete data, reducing the final sample to 634 (Table I in the online-only Data Supplement). Almost half of the participants had a history of hypertension, smoking, or hypercholesterolemia, whereas 11% had diabetes mellitus and 6.9% had a history of stroke.
In bivariate correlation analysis (Table II in the online-only Data Supplement), all measures of WMH (ie, %WMH volume in intracranial volume, Fazekas periventricular, and Fazekas deep WMH scores) were significantly associated with PVS (in the hippocampus, basal ganglia, and centrium semiovale; r range, 0.11–0.32; P<0.001). No significant association was found between WMH measures and inflammation markers. Centrium semiovale PVS were significantly associated with CRP (r=0.10; P=0.010) but not fibrinogen or IL-6. All P values were Bonferroni corrected to account for multiple comparisons.
Multivariate analysis using structural equation modelling showed that PVS was significantly associated with WMH (β=0.47; P<0.0001), accounting for ≈22% variation in WMH (Figure 1). Inflammation was weakly but significantly associated with PVS (β=0.13; P=0.048), explaining 1.6% of the variation in PVS (Figure 2). There was no significant association between Inflammation and WMH (Figure 3).
We demonstrate a strong association between increased numbers of visible PVS and increased amounts of WMH in adults aged 71 to 74 years. We demonstrate that variation in PVS accounted for ≈22% of the variation in WMH. However, despite previous reports of raised plasma inflammatory markers in subjects with SVD,5,12 we found only a weak association between inflammatory markers and PVS and no association between inflammatory markers and WMH. The narrow age cohort may have allowed us to unmask several relationships that are actually coassociations rather than direct associations as suggested in wider age range cohorts.
The weak association between Inflammation and PVS is consistent with the hypothesis that inflammation influences SVD through effects on the small perforating arterioles, which in turn precipitate WMH, but substantially more work is required to determine the direction and strength of the association.
The lack of association between inflammatory markers and WMH contrasts with some previous studies that found associations between CRP (n=6518),4,12 or IL-6 (n=3644),12 and WMH. However, it agrees with other studies (n=1699),5,13–15 a total of 2333 subjects, including the present study, not showing an association between CRP and WMH. Fewer studies have addressed associations between IL-6 or fibrinogen and WMH or PVS. Studies of inflammation and WMH are dominated by the Cardiovascular Health Study,12 which contributes two thirds of the data on CRP and all of the data on IL-6 and WMH. The study12 used MR images acquired from 1992 to 1994, which may have been less sensitive to WMH compared with current scanners; the wider age cohort may have suggested an inflammation–WMH association that was, in part, a residual age coassociation despite correction for age.
The strengths of the present work include the use of both volumetric and visual WMH scores, the use of structural equation modelling for a robust multivariate analysis, the cohort with a narrow age range, and little ethnic diversity to reduce confounding. Weaknesses include the relative health of the participants (about half were hypertensive) and the limited number of inflammatory variables examined (but the 3 used are well understood and key to pathological processes).
Future studies should consider examining inflammatory markers in younger subjects with SVD, because they may show a stronger differential inflammatory plasma profile. Studies should also examine direct evidence of inflammation in the brain.
Image acquisition and analysis was performed at the Brain Research Imaging Centre, University of Edinburgh (http://www.bric.ed.ac.uk). The work was undertaken as part of the Cross Council and University of Edinburgh Centre for Cognitive Ageing and Cognitive Epidemiology (http://www.ccace.ed.ac.uk).
Sources of Funding
Supported by Research Into Ageing, Age UK, Medical Research Council, the Scottish Funding Council and Scottish Imaging Network: A Platform for Scientific Excellence collaboration (http://www.sinapse.ac.uk), and the Row Fogo Charitable Trust.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.113.004059/-/DC1.
- Received November 11, 2013.
- Accepted December 3, 2013.
- © 2014 American Heart Association, Inc.
- Gouw AA,
- Seewann A,
- van der Flier WM,
- Barkhof F,
- Rozemuller AM,
- Scheltens P,
- et al
- Hooper D,
- Coughlan J,
- Mullen M
- Fornage M,
- Chiang YA,
- O’Meara ES,
- Psaty BM,
- Reiner AP,
- Siscovick DS,
- et al
- Schmidt R,
- Schmidt H,
- Pichler M,
- Enzinger C,
- Petrovic K,
- Niederkorn K,
- et al
- Wright CB,
- Moon Y,
- Paik MC,
- Brown TR,
- Rabbani L,
- Yoshita M,
- et al