Abstract
OBJECTIVE:
High plasma ceramide levels and ratios are associated with poor outcomes in individuals with cardiovascular disease; less is known about their relation to cerebral small vessel disease. We examined whether high plasma ceramide levels or ratios were associated with cerebral microbleeds (CMBs) and lacunes, and whether associations differ by sex.
APPROACH AND RESULTS:
We included 548 participants enrolled in the Mayo Clinic Study of Aging with concurrent plasma ceramide assays and MRI. CMBs were quantified on T2* MRI, and lacunes on T2 FLAIR MRI. Fasting plasma ceramides were assayed using liquid chromatography electrospray ionization tandem mass spectrometry. We used logistic regression models adjusting for age, sex, hypertension, and diabetes to examine the relationship between ceramides and presence of a lacune; hurdle models were used for presence and number of CMBs. Each standard deviation increase in the log ceramide C16:0/24:0 ratio was associated with greater odds of a CMB (odds ratio, 1.28 [95% Cl, 1.01–1.64]). There was an interaction between sex and the ceramide C16:0/24:0 ratio (P=0.049). The association between this ratio and presence of a CMB was stronger for women (odds ratio, 1.87 [95% CI, 1.20–3.00]) than men (odds ratio, 1.09 [95% CI, 0.80–1.46]). Several ceramides and all ceramide ratios were associated with number of CMBs. We did not find associations between plasma ceramides and lacunes.
CONCLUSIONS:
In a population based sample, the plasma ceramide C16:0/24:0 ratio was associated with CMBs and was stronger for women. Plasma ceramides are differentially associated with cerebral small vessel pathologies.
Keywords: ceramides, cerebral microbleeds, lacunes
Subject Codes: Lipids and Cholesterol, Biomarkers, Clinical Studies, Vascular Biology
INTRODUCTION
Cerebral microbleeds (CMBs) and lacunes are common lesions of cerebral small vessel disease that are detected on MRI in non-demented elderly populations. CMBs are detected on hemosiderin sensitive MRI sequences and it is hypothesized that they may represent an earlier process in the development of intracerebral hemorrhage.1 CMBs increase with advanced age, are more frequent in men,2,3 and are leading causes of morbidity and mortality worldwide.4 Developing biomarkers related to the upstream pathogenesis of CMBs and ischemic stroke has the potential to lead to earlier identification, risk stratification, improved outcomes, and prevention. Ceramide levels and ceramide ratios are candidate biomarkers.
Ceramides are bioactive lipids that induce proliferation in many cell types, including vascular cells. Ceramide levels increase with advanced age and are higher in females.5 Higher ceramide levels and ratios are associated with poor outcomes in individuals with cardiovascular disease, such as stable coronary artery disease and acute coronary artery disease.6,7 Recently, it has been shown that ceramide levels can predict major adverse cardiovascular events in healthy individuals.8 However, little research has examined the association between plasma ceramides or ceramide ratios with cerebral small vessel disease. Elevated long-chain plasma ceramide levels have been associated with lower fractional anisotrophy, measured by diffusion tensor imaging, in multiple white matter regions including total cerebral white matter, the corona radiata, and the cingulate gyrus.9 More recent studies suggest that the ceramide C24:0/C16:0 ratio is associated with white matter hyperintensity (WMH) volumes.10,11 The identification of risk factors for cerebral small vessel disease is important to understand possible underlying mechanisms and to develop clinical biomarkers. Therefore, the goal of the present study was to determine whether plasma ceramide levels and ceramide ratios would also be associated with other lesions of cerebral small vessel disease, specifically the presence and number of CMBs and the presence of lacunes, after adjustment for major risk factors including age, sex, diabetes, and hypertension.12–16 In addition, determined whether any relationship between plasma ceramide levels or ratios and odds of CMBs or lacunes differed by sex.
MATERIALS AND METHODS
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Study Design and Participants
The participants in this study were enrolled in the Mayo Clinic Study of Aging (MCSA), a population-based study of residents living in Olmsted County, Minnesota.17 The Olmsted County population was enumerated using the Rochester Epidemiology Project medical records-linkage system.18 MCSA participants were randomly selected in an age- and sex-stratified study design. A detailed description of the MCSA study has been described previously.17,19 Participants underwent detailed clinical visits including neuropsychological testing, a physician examination, and blood draws. Eligible participants (eg, no pacemaker or metal in the body) were invited to participate in a brain MRI. The current analysis included 548 participants, aged 60–92, who had MRI and plasma ceramide levels concurrently measured at the same study visit. Within the overall MCSA cohorts, individuals with MRI imaging were younger and healthier compared to individuals without MRI imaging.10 The study was approved by Mayo Clinic and Olmsted Medical Center Institutional Review Boards and written informed consent was obtained from all participants.
Ceramide Assays
A detailed description of the ceramide assay methods has been previously described.10 Blood samples from participants were collected during clinical examination in the fasting state and stored at −80°C. Liquid chromatography electrospray ionization tandem mass spectrometry (LC/ESI/MS/MS) analysis of ceramides was performed using an AB Sciex quadrupole mass spectrometer 6500 (Sciex, Framingham, MA) equipped with an ESI probe and interfaced with the Agilent 1290 infinity liquid chromatography system (Agilent, Palo Alto, CA). The ultra-performance liquid chromatography (UPLC) system consisted of an Agilent 1290 binary pump, thermostat, thermostatted column compartment (TCC), and sampler. The injection volume was 10 μL for extracted sample. Ceramides were separated with a Poroshell 120 EC- C8 column, 2.1×50 mm, 2.7 μm (Agilent, Palo Alto, CA). Mobile phase A was water:methanol:formic acid:ammonium formate (45/55/0.5%/5 mM by v/v). Mobile phase B was acetonitrile:methanol:formic acid:ammonium formate (50/50/0.5%/5 mM by v/v). The valve, sample loop, and needle were washed with acetonitrile:methanol (50/50 by v/v) for 20 sec. Mass spectrometric analyses were performed online using electrospray ionization tandem mass spectrometry in the positive mode.
Samples were prepared using Biomek FX (Beckman Coulter, Brea, CA). A small amount of plasma sample was added to a 2-mL 96-well plate, followed by an internal standard mixture. Ceramides were extracted using 1 phase extraction with methanol-dichloromethane. Lipid levels were quantified by the ratio of the ceramides to internal standard. Calibration curves were obtained by serial dilution of a mixture of lipid standards and expressed in ng/ml. Pure synthetic ceramide standards were purchased from Avanti Lipids. Isotope labeling synthetic standards were synthesized internally at Eli Lilly and Company.
Calculation of the Ceramide Ratios and Score
Based on previous literature in the cardiology field, we incorporated the established ceramide ratios and the calculation of the ceramide score in the current analysis.6–8 Using three ceramide levels (16:0, 18:0, 24:1) and three ceramide ratios (C16:0/24:0, C18:0/24:0, C24:1/C24:0) a ceramide score was calculated where individuals received 1 point for being in the third quartile of each of the ceramides or ceramide ratios and 2 points for being in the fourth quartile; scores ranged from 0–12.
Vascular Neuroimaging Measures
All images were obtained using 3-tesla General Electric (Milwaukee, WI) MRI scanners. A detailed description of the methods used to quantify CMBs has been described.12 CMBs were evaluated on a T2* GRE (TR/TE = 200/20 ms; flip angle = 12°; in-plane matrix = 256 × 224; phase FOV = 1.00; slice thickness = 3.3 mm). Acquisition time was 5 minutes. CMBs were defined according to consensus criteria as homogeneous hypointense lesions in the gray or white matter, which are distinct from iron or calcium deposits and vessel flow voids on T2* GRE images.20 All CMBs were initially identified by trained image analysts, then subsequently confirmed by a vascular neurologist (J.G.R) who was blinded to the clinical information and was experienced in grading CMBs. The location of the CMBs was recorded and they were categorized as either deep or lobar.12 However, given the small number of deep CMBs, the focus of the current analysis is on any CMBs and lobar CMBs. In the present study, outcomes included both the presence of a CMB (yes/no) and the number of CMBs.
The full details of lacune grading have been published.21 Cortical lacunes were characterized as hyperintense T2 FLAIR lesions (gliosis) involving cortical gray matter that extended to the cortical edge. These lacunes were identified on the T2 FLAIR sequence, with a corresponding T1 hypointensity required for confirmation. Subcortical lacunes were defined as hyperintense T2 FLAIR lesions, 3 to 15 mm in diameter, with a dark center seen in the white matter, infratentorial, or central gray-capsular regions in accordance with the Standards for Reporting Vascular changes in neuroimaging (STRIVE) criteria.22
Covariates
Covariates were chosen based on our previous analyses determining risk factors for both prevalent and incident CMBs and for lacunes in the MCSA.12,14,23 Age and sex were self-reported at the in-person clinical examination. Medical conditions (eg, hypertension, diabetes) were determined for each participant by medical record abstraction via trained nurses using the REP medical records-linkage system, which is more accurate than self-report.18
Statistical Analyses
Differences in baseline ceramide levels and ratios, and participant characteristics by sex were assessed using Student’s t-tests for continuous variables and chi-square tests for categorical variables. The plasma ceramide levels and ratios were log-transformed and z-scored prior to regression analyses in order to more directly compare each level and ratio. Given the small number of deep CMBs, we were not able to examine deep CMBs as an outcome, so analyses with CMBs focused on all CMBs and lobar CMBs only. We used logistic regression to examine the plasma ceramides and ceramide ratios with odds of the presence of a lacune (any, cortical, subcortical). We additionally ran hurdle models to examine the plasma ceramides and ceramide ratios in relation to the number of CMBs (any or lobar).24 The hurdle models consisted of two components: 1) a logistic regression model for predicting no CMBs vs. having one or more CMBs, which constitutes the hurdle, and 2) a truncated negative binomial model for predicting number of CMBs amongst those with CMBs. For all logistic regression models we computed odds ratios (OR), and for the negative binomial portion of the hurdle models we computed incidence rate ratios (IRR). We constructed univariable logistic models and multivariable logistic and hurdle models adjusting for age, sex, diabetes, and hypertension. In additional analyses, we examined the interaction between plasma ceramide levels and ratios with sex by adding interaction terms to the models. Statistical significance was defined as P < 0.05. Statistical analyses were completed using SAS version 9.4 (SAS Institute Inc., Cary, NC) and R version 3.6.2 (R Foundation for Statistical Computing, Vienna, Austria), with the hurdle models run using the pscl package.25,26 Data from the MCSA and this study are available on request.
RESULTS
The demographic characteristics, imaging summaries, and ceramide levels or ratios of the 548 participants, by sex, are shown in Table 1. Of the 548 participants, 281 (51.3%) were male. There were no sex differences with regards to age or a diagnosis of hypertension or diabetes, but men had more years of education. Men had a higher frequency of cortical and subcortical lacunes, any CMB, and lobar CMB compared to women. Most of the ceramide levels and ratios, and the ceramide score, were higher among women than men.
Table 1.
Baseline Characteristics, Lacunes, Cerebral Microbleeds, and Plasma Lipid Levels (N=548)
| Women (N=267) | Men (N=281) | P | |
|---|---|---|---|
| Characteristics | Mean(range)/N(%) | Mean(range)/N(%) | Value |
| Age (years) | 68.4 (60.4, 92.2) | 67.4 (60.0, 92.3) | 0.537 |
| Education (years) | 14.5 (0, 20) | 15.3 (6, 20) | <0.001 |
| APOE (≥1 allele) | 91 (34.1%) | 77 (27.4%) | 0.090 |
| Hypertension | 138 (51.7%) | 167 (59.4%) | 0.068 |
| Diabetes | 33 (12.4%) | 47 (16.7%) | 0.148 |
| Lacune type | |||
| Any | 20 (7.5%) | 48 (17.1%) | <0.001 |
| Cortical | 6 (2.3%) | 15 (5.3%) | 0.062 |
| Subcortical | 13 (4.9%) | 38 (13.5%) | <0.001 |
| CMBs type* | |||
| Total | 0.43 (0, 74) | 0.51 (0, 22) | 0.004 |
| Any | 28 (10.5%) | 53 (18.9%) | 0.006 |
| Lobar | 19 (7.1%) | 45 (16.0%) | 0.001 |
| Deep | 10 (3.7%) | 8 (2.8%) | 0.555 |
| Ceramides, ng/ml | |||
| C14:0 | 12.0 (3.0, 65.7) | 9.0 (6.7, 12.4) | <0.001 |
| C16:0 | 97.3 (47.7, 204.3) | 86.1 (24.7, 175.7) | <0.001 |
| C18:0 | 89.5 (23.8, 246.2) | 74.0 (13.3, 212.2) | <0.001 |
| C18:1 | 3.0 (0.06, 14.05) | 2.7 (0.4, 15.8) | 0.057 |
| C20:0 | 254.3 (111.4, 728.6) | 222.3 (87.6, 486.4) | <0.001 |
| C22:0 | 766.7 (341.9, 1487.6) | 705.3 (298.8, 1488.0) | <0.001 |
| C23:0 | 456.3 (72.3, 919.3) | 407.6 (37.6, 953.0) | <0.001 |
| C24:0 | 1261.0 (448.7, 2431.1) | 1222.8 (102.0, 2270.9) | 0.194 |
| C24:1 | 336.1 (151.9, 711.6) | 306.5 (108.4, 625.3) | <0.001 |
| Ceramide ratios | |||
| C16:0_24:0 | 0.08 (0.04, 0.20) | 0.08 (0.03, 0.89) | 0.222 |
| C18:0_24:0 | 0.07 (0.02, 0.26) | 0.06 (0.01, 0.56) | 0.002 |
| C24:1_24:0 | 0.28 (0.13, 0.75) | 0.27 (0.09, 2.55) | 0.314 |
| Ceramide score | 5.3 (0, 12) | 3.8 (0, 12) | <0.001 |
CMBs indicates cerebral microbleeds. P-values are from Student’s t-test for all continuous variables, except total CMBs which required a Wilcoxon rank sum test due to the skewed distribution, and are from chi-square tests for categorical variables.
The median number of total, lobar, and deep CMBs was 0 for both women and men.
The univariable and multivariable logistic regression models for the association between plasma ceramides (ng/ml) or ceramide ratios with presence of any CMB (including both lobar and deep) or with lobar CMB only are shown in Table 2. In univariable models, each one standard deviation increase in the log of ceramide C24:0 was associated with reduced odds of any CMB. Similarly, increases in the log of ceramides C14:0 and C24:0 were associated with reduced odds of a lobar CMB. In contrast, the ratios of ceramides C16:0/C24:0 and C24:1/24:0 were associated with increased odds of any CMB and of a lobar CMB. After adjusting for age, sex, hypertension, and diabetes in multivariable analyses, only increases in the ceramide C16:0/24:0 ratio remained significantly associated with greater odds of any CMB (OR, 1.28 [95% Cl, 1.01–1.64]) or of a lobar CMB (OR, 1.31 [95% Cl, 1.01–1.71]). The ceramide score was not associated with odds of any CMB or a lobar CMB.
Table 2.
Associations of Plasma Ceramides and Ceramide Ratios with Presence of and Number of Total and Lobar Cerebral Microbleeds
| Univariate | Multivariable* | Multivariable IRR* | ||||
|---|---|---|---|---|---|---|
| z-Log Lipid (ng/ml) | OR (95% CI) | P Value | OR (95% CI) | P Value | IRR (95% CI) | P Value |
| Any CMB | ||||||
| Ceramides | ||||||
| C14:0 | 0.90 (0.70, 1.14) | 0.386 | 1.00 (0.77, 1.28) | 0.974 | 1.53 (0.84, 2.79) | 0.168 |
| C16:0 | 1.13 (0.89, 1.44) | 0.318 | 1.19 (0.92, 1.56) | 0.192 | 1.68 (0.89, 3.19) | 0.112 |
| C18:0 | 0.94 (0.74, 1.19) | 0.586 | 1.05 (0.81, 1.35) | 0.731 | 1.50 (0.79, 2.85) | 0.218 |
| C18:1 | 0.89 (0.70, 1.13) | 0.333 | 0.91 (0.70, 1.17) | 0.462 | 2.54 (1.32, 4.89) | 0.005 |
| C20:0 | 1.08 (0.85, 1.36) | 0.543 | 1.20 (0.92, 1.56) | 0.171 | 1.97 (0.98, 3.94) | 0.057 |
| C22:0 | 0.90 (0.71, 1.14) | 0.394 | 1.01 (0.79, 1.30) | 0.927 | 0.69 (0.35, 1.35) | 0.279 |
| C23:0 | 0.95 (0.76, 1.20) | 0.641 | 1.02 (0.80, 1.32) | 0.899 | 1.01 (0.53, 1.96) | 0.966 |
| C24:0 | 0.78 (0.62, 0.98) | 0.03 | 0.89 (0.70, 1.13) | 0.325 | 0.52 (0.27, 1.02) | 0.056 |
| C24:1 | 1.04 (0.82, 1.31) | 0.774 | 1.10 (0.86, 1.42) | 0.454 | 1.36 (0.69, 2.68) | 0.369 |
| Ceramide ratios | ||||||
| C16:0_24:0 | 1.39 (1.11, 1.77) | 0.004 | 1.28 (1.01, 1.64) | 0.044 | 1.89 (1.16, 3.06) | 0.010 |
| C18:0_24:0 | 1.14 (0.90, 1.44) | 0.293 | 1.14 (0.89, 1.48) | 0.299 | 2.39 (1.27, 4.53) | 0.007 |
| C24:1_24:0 | 1.30 (1.03, 1.64) | 0.025 | 1.21 (0.95, 1.55) | 0.117 | 3.93 (1.73, 8.96) | 0.001 |
| Ceramide score | 1.17 (0.92, 1.48) | 0.189 | 1.16 (0.90, 1.51) | 0.255 | 2.65 (1.42, 4.95) | 0.002 |
| Lobar CMB | ||||||
| Ceramides | ||||||
| C14:0 | 0.75 (0.56, 0.98) | 0.037 | 0.85 (0.63, 1.13) | 0.280 | 2.07 (1.00, 4.27) | 0.050 |
| C16:0 | 1.05 (0.81, 1.36) | 0.735 | 1.07 (0.81, 1.41) | 0.630 | 2.55 (1.10, 5.92) | 0.030 |
| C18:0 | 0.90 (0.69, 1.16) | 0.402 | 1.02 (0.77, 1.35) | 0.916 | 1.71 (0.85, 3.45) | 0.135 |
| C18:1 | 0.77 (0.59, 1.00) | 0.052 | 0.78 (0.59, 1.04) | 0.089 | 4.28 (1.92, 9.52) | <.001 |
| C20:0 | 1.03 (0.80, 1.34) | 0.802 | 1.18 (0.88, 1.57) | 0.271 | 2.26 (1.03, 4.93) | 0.041 |
| C22:0 | 0.86 (0.66, 1.11) | 0.239 | 0.98 (0.75, 1.29) | 0.903 | 0.57 (0.23, 1.38) | 0.210 |
| C23:0 | 0.88 (0.69, 1.14) | 0.326 | 0.95 (0.73, 1.27) | 0.736 | 1.27 (0.61, 2.67) | 0.523 |
| C24:0 | 0.73 (0.57, 0.93) | 0.011 | 0.84 (0.65, 1.09) | 0.165 | 0.54 (0.24, 1.21) | 0.134 |
| C24:1 | 0.98 (0.76, 1.27) | 0.879 | 1.07 (0.81, 1.41) | 0.630 | 1.25 (0.55, 2.84) | 0.588 |
| Ceramide ratios | ||||||
| C16:0_24:0 | 1.41 (1.10, 1.81) | 0.007 | 1.31 (1.01, 1.71) | 0.043 | 2.25 (1.18, 4.28) | 0.014 |
| C18:0_24:0 | 1.16 (0.89, 1.50) | 0.277 | 1.17 (0.89, 1.55) | 0.267 | 2.76 (1.31, 5.84) | 0.008 |
| C24:1_24:0 | 1.33 (1.03, 1.71) | 0.024 | 1.26 (0.96, 1.64) | 0.087 | 3.15 (1.25, 7.94) | 0.015 |
| Ceramide score | 1.13 (0.88, 1.47) | 0.336 | 1.14 (0.85, 1.52) | 0.365 | 3.14 (1.48, 6.66) | 0.003 |
CMBs indicates cerebral microbleeds. Multivariable models adjusted for age, sex, diabetes and hypertension. The Multivariable odds ratio (OR) and Multivariable Incidence Rate Ratio (IRR) constitute the hurdle model. The multivariable models provide the OR for a CMB for each standard deviation increase in each ceramide or ratio. The Multivariable IRR can be interpreted as, among those with a CMB, the IRR for each one standard deviation increase in each ceramide or ceramide ratio.
The truncated negative binomial portion of the multivariable hurdle models showed that the plasma ceramides and ceramide ratios were associated with number of CMBs and number of lobar CMBs (Table 2). Notably, a higher ceramide C16:0/24:0 ratio was associated with a greater number of CMBs (IRR, 1.89 [95% CI, 1.16–3.06]). This can be more precisely interpreted as, for each one standard deviation increase in the log of ceramide C16:0/24:0 ratio amongst those with at least one CMB, the expected number of CMBs goes up by a multiplicative factor of 1.89. Although the ceramide C16:0/24:0 ratio was the only significant predictor of the presence of a CMB in the multivariable logistic models, the hurdle models revealed that several ceramides and ceramide ratios were associated with the number of CMBs. These include C18:1 (IRR, 2.54 [95% CI, 1.32–4.89]), C18:0/24:0 ratio (IRR, 2.39 [95% CI, 1.27–4.53]), C24:1/24:0 ratio (IRR, 3.93 [95% CI, 1.73–8.96]), and the ceramide score (IRR, 2.65 [95% CI, 1.42–4.95]). In addition, several ceramide levels (C14:0, C16:0, C18:1, C20:0), all ceramide ratios (C16:0/24:0, C18:0/24:0, C24:1/24:0), and the ceramide score were associated with number of lobar CMBs (Table 2) amongst those with a CMB.
We repeated the multivariable analyses including a sex interaction term (Table 3). We observed an interaction between sex and the ceramide C16:0/24:0 ratio for any CMB (P=0.049) and for lobar CMB (P=0.045). The association between each one standard deviation increase in the ceramide C16:0/24:0 ratio and odds of any CMB was stronger for women (OR, 1.87 [95% CI, 1.20–3.00]) than for men (OR, 1.09 [95% CI, 0.80–1.46]). Similarly, the relationship between a higher ceramide C16:0/24:0 ratio and odds of a lobar CMB was stronger for women (OR, 2.10 [95% CI,1.25–3.71]) than men (OR, 1.12 [95% CI, 0.81–1.51]). The truncated negative binomial portion of the hurdle models revealed multiple interactions between ceramides and sex for predicting number of CMBs (any or lobar) amongst those with CMBs. For example, the ceramide C16:0/24:0 ratio was strongly associated with the number of CMBs among women (IRR, 12.10 [95% CI, 3.11–47.04]), but not men (IRR, 0.74 [95% CI, 0.36–1.52]). All sex interaction results for number of total CMBs and lobar CMBs are shown in Table 3.
Table 3.
Multivariate Association of Plasma Ceramides and Ceramide Ratios with the Presence of and Number of Cerebral Microbleeds, Including Sex Interaction Terms
| Female | Male | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| z-Log Lipid (ng/ml) | OR (95% CI) | P Value IRR (95% CI) | P Value | OR (95% CI) | P Value | IRR (95% CI) | P Value | OR Interaction P Value | IRR Interaction P Value | |
| Any CMB | ||||||||||
| Ceramides | ||||||||||
| C14:0 | 0.77 (0.48, 1.20) | 0.265 | 2.53 (1.29, 4.95) | 0.007 | 1.13 (0.82, 1.54) | 0.447 | 0.45 (0.20, 1.05) | 0.066 | 0.175 | 0.001 |
| C16:0 | 1.25 (0.80, 1.93) | 0.326 | 5.25 (1.78, 15.50) | 0.003 | 1.17 (0.85, 1.62) | 0.357 | 0.66 (0.31, 1.42) | 0.290 | 0.809 | 0.003 |
| C18:0 | 0.95 (0.61, 1.50) | 0.833 | 2.85 (1.08, 7.53) | 0.034 | 1.09 (0.80, 1.50) | 0.575 | 0.96 (0.40, 2.32) | 0.936 | 0.262 | 0.097 |
| C18:1 | 0.83 (0.57, 1.23) | 0.353 | 4.90 (1.86, 12.87) | 0.001 | 0.97 (0.69, 1.35) | 0.854 | 1.38 (0.61, 3.12) | 0.439 | 0.562 | 0.041 |
| C20:0 | 0.94 (0.60, 1.48) | 0.806 | 4.73 (1.76, 12.68) | 0.002 | 1.36 (0.98, 1.91) | 0.066 | 0.77 (0.30, 1.97) | 0.587 | 0.203 | 0.006 |
| C22:0 | 0.74 (0.48, 1.14) | 0.177 | 0.80 (0.21, 2.97) | 0.734 | 1.18 (0.87, 1.60) | 0.290 | 0.66 (0.31, 1.42) | 0.287 | 0.086 | 0.805 |
| C23:0 | 0.92 (0.60, 1.44) | 0.690 | 0.13 (0.01, 1.39) | 0.092 | 1.07 (0.79, 1.48) | 0.670 | 1.53 (0.72, 3.24) | 0.266 | 0.567 | 0.048 |
| C24:0 | 0.62 (0.39, 0.97) | 0.037 | 0.19 (0.03, 1.33) | 0.095 | 1.03 (0.78, 1.41) | 0.842 | 0.66 (0.32, 1.38) | 0.269 | 0.062 | 0.262 |
| C24:1 | 0.98 (0.60, 1.60) | 0.941 | 6.51 (1.77, 23.89) | 0.005 | 1.15 (0.86, 1.55) | 0.359 | 0.90 (0.43, 1.88) | 0.773 | 0.591 | 0.005 |
| Ceramide ratios | ||||||||||
| C16:0_24:0 | 1.87 (1.20, 3.00) | 0.007 | 12.10 (3.11, 47.04) | <0.001 | 1.09 (0.80, 1.46) | 0.573 | 0.74 (0.36, 1.52) | 0.414 | 0.049 | 0.002 |
| C18:0_24:0 | 1.34 (0.86, 2.14) | 0.199 | 5.28 (1.84, 15.10) | 0.002 | 1.06 (0.78, 1.45) | 0.712 | 1.41 (0.61, 3.27) | 0.420 | 0.393 | 0.062 |
| C24:1_24:0 | 1.52 (0.98, 2.41) | 0.066 | 19.73 (4.04, 96.43) | <0.001 | 1.10 (0.81, 1.47) | 0.528 | 1.26 (0.44, 3.59) | 0.668 | 0.234 | 0.005 |
| Ceramide score | 1.25 (0.83, 1.91) | 0.288 | 4.88 (2.13, 11.17) | <0.001 | 1.11 (0.79, 1.54) | 0.544 | 0.75 (0.27, 2.14) | 0.596 | 0.651 | 0.007 |
| Lobar CMB | ||||||||||
| Ceramides | ||||||||||
| C14:0 | 0.59 (0.32, 1.06) | 0.092 | 3.85 (1.61, 9.17) | 0.002 | 0.97 (0.68, 1.35) | 0.841 | 0.28 (0.07, 1.12) | 0.071 | 0.085 | 0.001 |
| C16:0 | 1.22 (0.70, 2.11) | 0.472 | 27.60 (4.98, 152.97) | <0.001 | 1.11 (0.79, 1.57) | 0.554 | 0.17 (0.04, 0.74) | 0.019 | 0.757 | <0.001 |
| C18:0 | 0.86 (0.49, 1.53) | 0.595 | 5.31 (1.52, 18.51) | 0.009 | 1.06 (0.77, 1.47) | 0.74 | 1.01 (0.42, 2.41) | 0.986 | 0.546 | 0.032 |
| C18:1 | 0.74 (0.46, 1.19) | 0.199 | 8.79 (2.44, 31.68) | 0.001 | 0.79 (0.55, 1.13) | 0.203 | 2.39 (0.86, 6.64) | 0.094 | 0.847 | 0.113 |
| C20:0 | 1.01 (0.56, 1.79) | 0.971 | 9.72 (2.62, 36.07) | 0.001 | 1.23 (0.88, 1.74) | 0.232 | 0.49 (0.12, 2.03) | 0.325 | 0.501 | 0.003 |
| C22:0 | 0.65 (0.37, 1.12) | 0.125 | 1.44 (0.33, 6.19) | 0.625 | 1.12 (0.81, 1.54) | 0.501 | 0.37 (0.11, 1.26) | 0.111 | 0.075 | 0.134 |
| C23:0 | 0.75 (0.46, 1.30) | 0.278 | 0.52 (0.03, 9.10) | 0.653 | 1.05 (0.77, 1.48) | 0.764 | 1.42 (0.64, 3.18) | 0.392 | 0.262 | 0.518 |
| C24:0 | 0.57 (0.32, 0.98) | 0.047 | 0.69 (0.12, 4.00) | 0.675 | 0.94 (0.70, 1.29) | 0.675 | 0.50 (0.19, 1.33) | 0.164 | 0.060 | 0.758 |
| C24:1 | 1.09 (0.58, 1.99) | 0.792 | 8.13 (1.60, 41.20) | 0.011 | 1.03 (0.76, 1.41) | 0.835 | 0.85 (0.35, 2.04) | 0.713 | 0.788 | 0.007 |
| Ceramide ratios | ||||||||||
| C16:0_24:0 | 2.00 (1.16, 3.63) | 0.016 | 24.25 (4.09, 143.89) | <0.001 | 1.15 (0.84, 1.55) | 0.365 | 0.68 (0.29, 1.58) | 0.369 | 0.045 | 0.002 |
| C18:0_24:0 | 1.34 (0.76, 2.45) | 0.331 | 8.72 (2.00, 38.04) | 0.004 | 1.11 (0.80, 1.53) | 0.537 | 1.50 (0.60, 3.77) | 0.383 | 0.443 | 0.058 |
| C24:1_24:0 | 1.81 (1.03, 3.27) | 0.041 | 15.58 (2.66, 91.32) | 0.002 | 1.10 (0.80, 1.48) | 0.562 | 1.33 (0.44, 4.02) | 0.617 | 0.062 | 0.021 |
| Ceramide score | 1.33 (0.78, 2.31) | 0.294 | 6.83 (2.41, 19.35) | <0.001 | 1.05 (0.74, 1.48) | 0.773 | 0.72 (0.20, 2.54) | 0.608 | 0.310 | 0.010 |
CMBs indicates cerebral microbleeds. Multivariable models adjusted for age, sex, diabetes and hypertension and have interaction terms (sex*ceramide level/ratio). The Multivariable odds ratio (OR) and Multivariable Incidence Rate Ratio (IRR) constitute the hurdle model. The multivariable models provide the OR for a CMB for each standard deviation increase in each ceramide or ratio. The Multivariable IRR can be interpreted as, among those with a CMB, the IRR for each one standard deviation increase in each ceramide or ceramide ratio.
There were no associations between plasma ceramides or ceramide ratios with the odds of any lacune in either univariate or multivariate analyses (Table 4). In univariable analyses examining cortical lacunes as the outcome, ceramide C18:1 was associated with lower odds (OR, 0.64 [95% CI, 0.42–0.98]). In addition, ceramide ratios C16:0/24:0 (OR, 1.45 [95% CI, 1.00–2.06]) and C24:1/24:0 (OR, 1.60 [95% CI, 1.09–2.32]) were associated with increased odds. However, the number of cortical lacunes was small so multivariable analyses were not possible. There were no significant interactions between each ceramide level or ceramide ratio and sex for the presence of any lacune, a cortical lacune, or a subcortical lacune (data not shown).
Table 4.
Logistic Regression Models for Associations Between Plasma Ceramides Levels and Ceramide Ratios with the Presence of a Lacune
| Univariate | Multivariate | |||
|---|---|---|---|---|
| z-Log Lipid (ng/ml) | OR (95% CI) | P Value | OR (95% CI) | P Value |
| Any lacune | ||||
| Ceramides | ||||
| C14:0 | 0.91 (0.70, 1.18) | 0.484 | 1.04 (0.79, 1.36) | 0.775 |
| C16:0 | 1.05 (0.82, 1.36) | 0.699 | 1.12 (0.85, 1.49) | 0.429 |
| C18:0 | 0.97 (0.76, 1.26) | 0.842 | 1.13 (0.86, 1.49) | 0.388 |
| C18:1 | 0.91 (0.71, 0.18) | 0.483 | 0.95 (0.72, 1.25) | 0.697 |
| C20:0 | 0.93 (0.72, 1.21) | 0.604 | 1.04 (0.79, 1.37) | 0.806 |
| C22:0 | 0.78 (0.61, 1.01) | 0.058 | 0.87 (0.67, 1.14) | 0.314 |
| C23:0 | 0.87 (0.69, 1.11) | 0.250 | 0.94 (0.73, 1.22) | 0.621 |
| C24:0 | 0.81 (0.64, 1.04) | 0.086 | 0.92 (0.72, 1.19) | 0.495 |
| C24:1 | 1.00 (0.78, 1.29) | 0.995 | 1.08 (0.83, 1.41) | 0.568 |
| Ceramide ratios | ||||
| C16:0_24:0 | 1.27 (1.00, 1.61) | 0.051 | 1.19 (0.91, 1.53) | 0.187 |
| C18:0_24:0 | 1.14 (0.89, 1.48) | 0.301 | 1.20 (0.91, 1.57) | 0.192 |
| C24:1_24:0 | 1.22 (0.95, 1.55) | 0.118 | 1.16 (0.89, 1.50) | 0.255 |
| Ceramide score | 1.16 (0.90, 1.48) | 0.259 | 1.20 (0.91, 1.58) | 0.195 |
| Subcortical lacune | ||||
| Ceramides | ||||
| C14:0 | 0.88 (0.65, 1.18) | 0.399 | 1.03 (0.75, 1.38) | 0.865 |
| C16:0 | 1.10 (0.82, 1.47) | 0.533 | 1.20 (0.88, 1.66) | 0.256 |
| C18:0 | 0.97 (0.73, 1.30) | 0.843 | 1.14 (0.85, 1.56) | 0.392 |
| C18:1 | 0.98 (0.74, 1.31) | 0.909 | 1.04 (0.76, 1.42) | 0.788 |
| C20:0 | 0.96 (0.72, 1.28) | 0.781 | 1.09 (0.80, 1.49) | 0.587 |
| C22:0 | 0.83 (0.62, 1.10) | 0.191 | 0.94 (0.70, 1.26) | 0.688 |
| C23:0 | 0.93 (0.71, 1.24) | 0.584 | 1.02 (0.77, 1.40) | 0.876 |
| C24:0 | 0.87 (0.67, 1.16) | 0.319 | 1.00 (0.76, 1.35) | 0.984 |
| C24:1 | 1.01 (0.76, 1.35) | 0.947 | 1.11 (0.83, 1.50) | 0.482 |
| Ceramide ratios | ||||
| C16:0_24:0 | 1.22 (0.93, 1.59) | 0.141 | 1.15 (0.85, 1.52) | 0.341 |
| C18:0_24:0 | 1.08 (0.81, 1.44) | 0.606 | 1.13 (0.84, 1.53) | 0.414 |
| C24:1_24:0 | 1.15 (0.86, 1.51) | 0.335 | 1.10 (0.82, 1.46) | 0.520 |
| Ceramide score | 1.13 (0.85, 1.51) | 0.383 | 1.20 (0.88, 1.63) | 0.253 |
| Cortical lacune | ||||
| Ceramides | ||||
| C14:0 | 0.92 (0.59, 1.42) | 0.724 | ||
| C16:0 | 1.11 (0.72, 1.73) | 0.640 | ||
| C18:0 | 1.16 (0.75, 1.82) | 0.520 | ||
| C18:1 | 0.64 (0.42, 0.98) | 0.038 | ||
| C20:0 | 1.06 (0.68, 1.63) | 0.804 | ||
| C22:0 | 0.77 (0.50, 1.19) | 0.233 | ||
| C23:0 | 0.92 (0.63, 1.432) | 0.679 | ||
| C24:0 | 0.73 (0.52, 1.06) | 0.073 | ||
| C24:1 | 1.27 (0.82, 2.00) | 0.288 | ||
| Ceramide ratios | ||||
| C16:0_24:0 | 1.45 (1.00, 2.06) | 0.037 | ||
| C18:0_24:0 | 1.51 (0.98, 2.29) | 0.059 | ||
| C24:1_24:0 | 1.60 (1.09, 2.32) | 0.012 | ||
| Ceramide score | 1.52 (0.99, 2.34) | 0.053 | ||
Multivariate models adjusted for age, sex, diabetes and hypertension. Multivariate analysis for cortical lacunes could not be performed due to low numbers.
DISCUSSION
We assessed the relationship between plasma ceramide levels and ratios with two lesions of cerebral small vessel disease, lacunes and cerebral microbleeds, in a community-based study while adjusting for known demographic and vascular risk factors. We found that the plasma ceramide C16:0/24:0 ratio was associated with higher odds of any CMB and of a lobar CMB in multivariate analyses. In addition, many of the plasma ceramide levels, ceramide ratios, and the ceramide risk score were associated with the number of CMBs amongst those with at least one CMB. All of these associations were significantly stronger in women than men. In contrast, there were no associations between plasma ceramide levels or ratios with subcortical lacunes. We did observe associations between plasma ceramide levels and ratios with cortical lacunes in univariate analyses but we had too few outcomes to conduct multivariable analyses. Together, these results suggest that the plasma ceramide C16:0/24:0 ratio is related to a specific pathology of cerebral small vessel disease.
Plasma ceramide ratios have been associated with an increased risk of cardiovascular disease, subsequent cardiovascular events, and mortality.6–8,27 In fact, these ceramide ratios have been found to outperform conventional cardiovascular risk markers for identifying at-risk patients.7 These findings led to the development of a ceramide risk score for clinical use that is comprised of ceramide C16:0/C24:0, C18:0/C24:0, and C24:1/C24:0 ratios. It has been established that cardiovascular pathologies and cerebral small vessel pathologies share similar risk factors (ie, hypertension, diabetes, and advanced age).2,13 Thus, it is important to determine whether plasma ceramide levels or ceramide ratios are also associated with cerebral small vessel disease, and to determine whether the association varies by type of pathology.
We initially examined the association between plasma ceramides and ceramide ratios with white matter hyperintensity (WMH) volumes using the same cohort as in the current analyses. We found that a higher ceramide C16:0/C24:0 ratio was associated with greater WMH volumes.10,11 Subsequently, results from the Framingham Offspring Study cohort reported a similar association between ceramide C16:0/C24:0 ratio and WMH volumes.10,11 In the current study, we have expanded this work and now report that a higher ceramide C16:0/C24:0 ratio is also associated with any CMB and a lobar CMB, even after adjusting for age, sex, hypertension and diabetes. These findings suggest that the ceramide C16:0/C24:0 ratio, may play a role in the pathogenesis of both WMHs and CMBs, two different types of cerebral small vessel disease. Importantly, although only the ceramide C16:0/C24:0 ratio was associated with the presence of a CMB (any or lobar), several ceramides and ceramide ratios were associated with the number of CMBs. This result indicates that plasma ceramide levels and ratios may be markers of disease severity, defined as the number of CMBs. Additional studies are needed to determine whether plasma ceramide levels or ceramide ratios can be used as prognostic markers for risk of WMHs or CMBs, including an incident initial CMB and increasing number of CMBs.
We did not find an association between ceramide levels or ratios with odds of any lacune or a subcortical lacune. Although higher ceramide ratios (C16:0/C24:0 and C24:1/24:0) were associated with increased odds of cortical lacunes in univariable analyses, there were not enough cortical lacunes to run multivariable analyses. Thus, it is possible that ceramide ratios are more specific to cortical lacunes than subcortical lacunes. One previous study comparing plasma ceramides in 207 stroke patients and age-matched controls reported higher plasma ceramide levels (C16:0, C22:0, and C24:0) in stroke patients.28 Furthermore, ceramide levels were much higher for those with moderate-severe stroke compared to minor stroke. The study did not examine ceramide ratios. Given that our participants are community-based, it is possible the infarcts were not severe enough to observe an association and that if we repeated the study in a clinical cohort of symptomatic stroke patients we would observe an association. Alternatively, it is also possible that there is an acute ceramide response that subsides over time after the infarct. Future longitudinal studies of community-based samples are needed to assess ceramide levels and ratios after strokes, and also to determine whether plasma ceramides are associated with risk of incident stroke.
Since lobar CMBs, WMH, and cortical lacunes have been associated with cerebral amyloid angiopathy (CAA),14,29,30 it is possible that the mechanism underlying the association between the ceramide C16:0/C24:0 ratio and these outcomes is related to CAA. Alternatively, in the general population lobar CMBs are also associated with hypertension, particularly when brain amyloid levels are low.12 Future studies investigating the relationship between lobar CMBs and ceramide levels that include biomarkers of brain amyloid will help determine whether the mechanism underlying our observed relationship is secondary to CAA or hypertensive disease.
We found that the association between the ceramide ratio, C16:0/C24:0, and CMBs was stronger for women compared to men. While some studies have not observed an association between CMBs and sex,2,13 other studies have shown that CMBs are more frequent in men.31 Plasma ceramides are higher in women, and there is an age-related greater increase in plasma ceramide levels for women compared to men, particularly after menopause.5,32 Thus, our finding of a stronger association between plasma ceramides and CMBs among women raises the possibility that the pathways which lead to CMB pathogenesis differ by sex. These findings need to be replicated and considered further as a possible means of identifying sex-specific risk factors.
Strengths of this study include the robust clinical and neuroimaging data from a long-standing and well-established cohort in Olmsted County, MN. Limitations of the study warrant consideration. First, is the cross-sectional study design. Longitudinal studies are needed to assess plasma ceramides levels and ratios as prognostic markers of cerebral small vessel disease and to better understand the temporality and causal role of plasma ceramides. Further, it remains possible that plasma ceramide ratios may be associated with risk of future ischemic stroke and longitudinal studies will be necessary to evaluate this risk. Second, given the small number of deep CMBs, we were not able to examine this association. Last, the cohort is primarily of European decent and the generalizability to other races or ethnicities is not known. However, associations between ceramides and cardiovascular disease have been observed across races including white Caucasian, African-American, and Chinese individuals.
In conclusion, results from this study suggest that the plasma ceramide ratio, C16:0/24:0, is associated with CMBs and that the association is modified by sex. In addition, several ceramides and ceramide ratios are associated with number of CMBs. Further work is needed to understand the mechanism and feasibility of ceramides in clinical practice and to determine the prognostic value of this ratio.
Highlights.
Plasma ceramides have been associated with risk of cardiovascular events and death, but have not been examined in relation to cerebral microbleeds (CMBs) or lacunes.
Higher levels of specific plasma ceramide ratios are associated with presence of a CMB.
Higher levels of plasma ceramides and ceramide ratios were associated with number of CMBs.
This association was stronger in women than men.
There were no associations between plasma ceramides and lacunes.
Sources of Funding
This work was supported by the National Institutes of Health/National Institute on Aging grants U01 AG006786, R01 AG011378, R01 AG049704, R01 NS097495, U54 AG44170, and RF1 AG55151, the GHR Foundation, and was made possible by the Rochester Epidemiology Project (R01 AG034676).
Disclosures
Dr. Mielke receives unrestricted research grants from Biogen and consults for Brain Protection Company. Dr. Bai is an employee of Eli Lilly. Dr. Knopman serves on a Data Safety Monitoring Board for Lundbeck Pharmaceuticals and for the DIAN study, and is an investigator in clinical trials sponsored by Biogen, Lilly Pharmaceuticals, and the Alzheimer’s Disease Cooperative Study. Dr. Jack has provided consulting services for Eli Lilly and serves on a DSMB for Roche but he receives no personal compensation from any commercial entity. Dr. Petersen is a consultant for Roche, Inc., Merck, Inc., Biogen, Inc., Genentech, Inc., Eisai, Inc., and has given educational lectures for GE Healthcare. All other authors have nothing to disclose.
Nonstandard Abbreviations and Acronyms
- CMBs
cerebral microbleeds
- eNOS
endothelial nitric oxide synthase
- FLAIR
fluid-attenuated inversion recovery
- GRE
gradient recalled echo
- LC/ESI/MS/MS
liquid chromatography electrospray ionization tandem mass spectrometry
- MCSA
Mayo Clinic Study of Aging
- MRI
magnetic resonance imaging
- REP
Rochester Epidemiology Project
- S1P
sphingosine-1-phosphate
- SD
standard deviation
- TCC
thermostatted column compartment
- UPLC
ultra performance liquid chromatography
- WMH
white matter hyperintensity
REFERENCES
- 1.Akoudad S, Portegies ML, Koudstaal PJ, Hofman A, van der Lugt A, Ikram MA, Vernooij MW. Cerebral microbleeds are associated with an increased risk of stroke: the Rotterdam Study. Circulation. 2015;132:509–516. doi: 10.1161/CIRCULATIONAHA.115.016261 [DOI] [PubMed] [Google Scholar]
- 2.Poels MM, Vernooij MW, Ikram MA, Hofman A, Krestin GP, van der Lugt A, Breteler MM. Prevalence and risk factors of cerebral microbleeds: an update of the Rotterdam scan study. Stroke. 2010;41:S103–106. doi: 10.1161/STROKEAHA.110.595181 [DOI] [PubMed] [Google Scholar]
- 3.Yubi T, Hata J, Ohara T, Mukai N, Hirakawa Y, Yoshida D, Gotoh S, Hirabayashi N, Furuta Y, Ago T, Kitazono T, Kiyohara Y, Ninomiya T. Prevalence of and risk factors for cerebral microbleeds in a general Japanese elderly community. Neurol Clin Pract. 2018;8:223–231. doi: 10.1212/CPJ.0000000000000464 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Lozano R, Naghavi M, Foreman K, et al. Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380:2095–2128. doi: 10.1016/S0140-6736(12)61728-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Mielke MM, Bandaru VV, Han D, An Y, Resnick SM, Ferrucci L, Haughey NJ. Demographic and clinical variables affecting mid- to late-life trajectories of plasma ceramide and dihydroceramide species. Aging Cell. 2015;14:1014–1023. doi: 10.1111/acel.12369 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Cheng JM, Suoniemi M, Kardys I, et al. Plasma concentrations of molecular lipid species in relation to coronary plaque characteristics and cardiovascular outcome: results of the ATHEROREMO-IVUS study. Atherosclerosis. 2015;243:560–566. doi: 10.1016/j.atherosclerosis.2015.10.022 [DOI] [PubMed] [Google Scholar]
- 7.Laaksonen R, Ekroos K, Sysi-Aho M, et al. Plasma ceramides predict cardiovascular death in patients with stable coronary artery disease and acute coronary syndromes beyond LDL-cholesterol. Eur Heart J. 2016;37:1967–1976. doi: 10.1093/eurheartj/ehw148 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Havulinna AS, Sysi-Aho M, Hilvo M, Kauhanen D, Hurme R, Ekroos K, Salomaa V, Laaksonen R. Circulating ceramides predict cardiovascular outcomes in the population-based FINRISK 2002 Cohort. Arterioscler Thromb Vasc Biol. 2016;36:2424–2430. doi: 10.1161/ATVBAHA.116.307497 [DOI] [PubMed] [Google Scholar]
- 9.Gonzalez CE, Venkatraman VK, An Y, Landman BA, Davatzikos C, Ratnam Bandaru VV, Haughey NJ, Ferrucci L, Mielke MM, Resnick SM. Peripheral sphingolipids are associated with variation in white matter microstructure in older adults. Neurobiol Aging. 2016;43:156–163. doi: 10.1016/j.neurobiolaging.2016.04.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Mielke MM, Syrjanen JA, Bui HH, Petersen RC, Knopman DS, Jack CR, Jr., Graff-Radford J, Vemuri P. Elevated plasma ceramides are associated with higher white matter hyperintensity volume-Brief Report. Arterioscler Thromb Vasc Biol. 2019;39:2431–2436. doi: 10.1161/ATVBAHA.119.313099 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.McGrath ER, Himali JJ, Xanthakis V, Duncan MS, Schaffer JE, Ory DS, Peterson LR, DeCarli C, Pase MP, Satizabal CL, Vasan RS, Beiser AS, Seshadri S. Circulating ceramide ratios and risk of vascular brain aging and dementia. Ann Clin Transl Neurol. 2020;7:160–168. doi: 10.1002/acn3.50973 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Graff-Radford J, Botha H, Rabinstein AA, et al. Cerebral microbleeds: prevalence and relationship to amyloid burden. Neurology. 2019;92:e253–e262. doi: 10.1212/WNL.0000000000006780 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Vernooij MW, van der Lugt A, Ikram MA, Wielopolski PA, Niessen WJ, Hofman A, Krestin GP, Breteler MM. Prevalence and risk factors of cerebral microbleeds: the Rotterdam Scan Study. Neurology. 2008;70:1208–1214. doi: 10.1212/01.wnl.0000307750.41970.d9 [DOI] [PubMed] [Google Scholar]
- 14.Graff-Radford J, Lesnick T, Rabinstein AA, et al. Cerebral microbleed incidence, relationship to amyloid burden: The Mayo Clinic Study of Aging. Neurology. 2020;94:e190–e199. doi: 10.1212/WNL.0000000000008735 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.You R, McNeil JJ, O’Malley HM, Davis SM, Donnan GA. Risk factors for lacunar infarction syndromes. Neurology. 1995;45:1483–1487. doi: 10.1212/wnl.45.8.1483 [DOI] [PubMed] [Google Scholar]
- 16.Fisher CM. Lacunes: small, deep cerebral infarcts. Neurology. 1965;15:774–784. doi: 10.1212/wnl.15.8.774 [DOI] [PubMed] [Google Scholar]
- 17.Roberts RO, Geda YE, Knopman DS, Cha RH, Pankratz VS, Boeve BF, Ivnik RJ, Tangalos EG, Petersen RC, Rocca WA. The Mayo Clinic Study of Aging: design and sampling, participation, baseline measures and sample characteristics. Neuroepidemiology. 2008;30:58–69. doi: 10.1159/000115751 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.St. Sauver JL, Grossardt BR, Yawn BP, Melton LJ 3rd, Pankratz JJ, Brue SM, Rocca WA. Data resource profile: the Rochester Epidemiology Project (REP) medical records-linkage system. Int J Epidemiol. 2012;41:1614–1624. doi: 10.1093/ije/dys195 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Petersen RC. Mild cognitive impairment as a diagnostic entity. J Intern Med. 2004;256:183–194. doi: 10.1111/j.1365-2796.2004.01388 [DOI] [PubMed] [Google Scholar]
- 20.Greenberg SM, Vernooij MW, Cordonnier C, Viswanathan A, Al-Shahi Salman R, Warach S, Launer LJ, Van Buchem MA, Breteler MM, Microbleed Study Group. Cerebral microbleeds: a guide to detection and interpretation. Lancet Neurol. 2009;8:165–174. doi: 10.1016/S1474-4422(09)70013-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Graff-Radford J, Aakre JA, Knopman DS, et al. Prevalence and heterogeneity of cerebrovascular disease imaging lesions. Mayo Clin Proc. 2020;95:1195–1205. doi: 10.1016/j.mayocp.2020.01.028 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Wardlaw JM, Smith EE, Biessels GJ, et al. Neuroimaging standards for research into small vessel disease and its contribution to ageing and neurodegeneration. Lancet Neurol. 2013;12:822–838. doi: 10.1016/S1474-4422(13)70124-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Fatemi F, Kantarci K, Graff-Radford J, Preboske GM, Weigand SD, Przybelski SA, Knopman DS, Machulda MM, Roberts RO, Mielke MM, Petersen RC, Jack CR, Jr., Vemuri P. Sex differences in cerebrovascular pathologies on FLAIR in cognitively unimpaired elderly. Neurology. 2018;90:e466–e473. doi: 10.1212/WNL.0000000000004913 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Mullahy J Specification and testing of some modified count data models. Journal of Econometrics. 1986;33:341–365. doi: 10.1016/0304-4076(86)90002-3 [DOI] [Google Scholar]
- 25.Jackman S (2020). pscl: Classes and Methods for R Developed in the Political Science Computational Laboratory. United States Studies Centre, University of Sydney, Sydney, New South Wales, Australia: R package version 1.5.5, https://github.com/atahk/pscl/ (accessed 08/10/2020). [Google Scholar]
- 26.Zeileis A, Kleiber C, Jackman S. Regression Models for Count Data in R. Journal of Statistical Software, Articles. 2008;27 https://www.jstatsoft.org/v027/i008 (accessed 08/10/2020). [Google Scholar]
- 27.Wang DD, Toledo E, Hruby A, et al. Plasma ceramides, Mediterranean diet, and incident cardiovascular disease in the PREDIMED Trial (Prevencion con Dieta Mediterranea). Circulation. 2017;135:2028–2040. doi: 10.1161/CIRCULATIONAHA.116.024261 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Gui YK, Li Q, Liu L, Zeng P, Ren RF, Guo ZF, Wang GH, Song JG, Zhang P. Plasma levels of ceramides relate to ischemic stroke risk and clinical severity. Brain Res Bull. 2020;158:122–127. doi: 10.1016/j.brainresbull.2020.03.009 [DOI] [PubMed] [Google Scholar]
- 29.Pasi M, Boulouis G, Fotiadis P, Auriel E, Charidimou A, Haley K, Ayres A, Schwab KM, Goldstein JN, Rosand J, Viswanathan A, Pantoni L, Greenberg SM, Gurol ME. Distribution of lacunes in cerebral amyloid angiopathy and hypertensive small vessel disease. Neurology. 2017;88:2162–2168. doi: 10.1212/WNL.0000000000004007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Graff-Radford J, Arenaza-Urquijo EM, Knopman DS, et al. White matter hyperintensities: relationship to amyloid and tau burden. Brain. 2019;142:2483–2491. doi: 10.1093/brain/awz162 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Graff-Radford J, Simino J, Kantarci K, Mosley TH Jr., Griswold ME, Windham BG, Sharrett AR, Albert MS, Gottesman RF, Jack CR Jr., Vemuri P, Knopman DS. Neuroimaging correlates of cerebral microbleeds: the ARIC Study (Atherosclerosis Risk in Communities). Stroke. 2017;48:2964–2972. doi: 10.1161/STROKEAHA.117.018336 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Vozella V, Basit A, Piras F, Realini N, Armirotti A, Bossu P, Assogna F, Sensi SL, Spalletta G, Piomelli D. Elevated plasma ceramide levels in post-menopausal women: a cross-sectional study. Aging (Albany NY). 2019;11:73–88. doi: 10.18632/aging.101719 [DOI] [PMC free article] [PubMed] [Google Scholar]