Abstract
Aim: White matter lesions (WML) are associated with dementia and they are influenced by cardiovascular disease (CVD) risk factors. Managing these risk factors may prevent WML progression. However, few longitudinal studies have examined the association between CVD risk factors and changes in the WML volume. This study aimed to investigate this association across a broad age range, including younger individuals.
Methods: This longitudinal study included 4,595 participants (age range, 21–90 years; women, 61.7%) who underwent brain magnetic resonance imaging. WML was defined on T1-weighted images. The associations between each CVD risk factor (hypertension, diabetes, dyslipidemia, and current smoking) and WML volume changes (per 4 years) were analyzed using a generalized linear model with estimated regression coefficients (β) and 95% confidence intervals (CI). The analyses were stratified by age group (＜50, 50–59, 60–69, and ≥ 70 years).
Results: Hypertension was significantly associated with increased WML volume change in those aged ＜50 and 50–59 years (β [95%CI] = 57.1 [8.0–106.2] and β [95%CI] = 77.9 [6.8–149.0], respectively). For both diabetes and dyslipidemia, the WML volume increased in those aged ＜50 years (β [95%CI] = 376.5 [209.3–543.8] and β [95%CI] = 154.0 [93.0–215.0], respectively). Current smoking showed an increase in those aged 50–59 and ≥ 60 years (β [95%CI] = 104.5 [2.1–206.8] and β [95%CI] = 305.9 [133.4–478.5], respectively). No significant associations were observed for any CVD risk factors in the other age groups.
Conclusions: The WML volume changes were larger in younger age groups for most CVD risk factors, suggesting that early management of these factors may help prevent WML progression.
Keywords: Cardiovascular disease, White matter lesions, Cross-sectional study, Longitudinal study, T1-weighted imaging
Introduction
The number of people with dementia is increasing worldwide and it is estimated to increase from 57 million in 2019 to 153 million by 2050 1) . Addressing the increasing prevalence of dementia is crucial to alleviate the strain on public health. Various factors, including vascular events, have been implicated 2 , 3) . Microvascular disease impacts the cognitive function 4 - 8) , highlighting the importance of maintaining cerebrovascular health. White matter lesions (WML), which are recognized as markers of cerebral small vessel disease (SVD) on magnetic resonance imaging (MRI) 6 , 9) , affect intracerebral hemorrhage, ischemic stroke, cognitive function, and dementia 10 - 15) . Additionally, an increased WML volume has been associated with cognitive decline 16) . Thus, preventing an increase in the WML volume may contribute to preventing cognitive decline and reducing the incidence of dementia, including vascular dementia and Alzheimer’s disease.
Previous studies have shown that cardiovascular disease (CVD) risk factors, such as aging, hypertension, diabetes, atherosclerosis, smoking, and cholesterol, are associated with WML 17 - 22) . As lifestyle-related aspects influence CVD risk factors, each of these risk factors can be improved through lifestyle changes, and interventions for CVD risk factors could help mitigate the increase in WML. Preventing or managing CVD risk factors may contribute to reducing the risk of dementia or SVD due to WML 10 , 23) . The 2024 Lancet Commission reported that the impact of modifiable dementia risk factors varied across different age groups throughout the lifespan, and that addressing CVD risk factors in midlife can reduce dementia incidence 24) . They also emphasized that dementia prevention strategies were more effective when initiated early and that longer durations of maintaining low levels of dementia risk factors are more beneficial than later initiation 24) .
WML are present even in younger populations and naturally increase with aging 25) . Therefore, examining the relationship between CVD risk factors and changes in WML is important in younger individuals, and investigating changes in WML, rather than cross-sectional studies investigating WML values at a single time point, is necessary. However, many previous studies were cross-sectional 17 , 19 , 21 , 22) , and some were longitudinal studies targeting middle-aged or older adults rather than younger populations 18 , 20) . To our knowledge, no longitudinal study has investigated this association across a wide age range in the general population, including younger populations. If our study reveals an association between CVD risk factors and WML changes, it could provide insight into whether lifestyle improvements contribute to the prevention of WML progression in all adults, including young adults. Additionally, considering WML as an intermediate phenotype of dementia, investigating the association could help clarify the underlying mechanisms of dementia.
Aim
This study aimed to investigate the longitudinal effects of CVD risk factors on WML progression across different age groups using a large-scale brain MRI cohort study that targeted the widest age range of the general population. Based on the importance of preventive interventions for CVD risk factors, starting at least by midlife 24) , this study can inform whether the prevention of WML progression should commence earlier in life. In addition to the main longitudinal analysis, we conducted a cross-sectional analysis to explore potential differences between cross-sectional and longitudinal results, highlighting the value of longitudinal studies for understanding WML progression.
Methods
Study Design and Participants
We conducted both cross-sectional and longitudinal studies using cohort survey data from the Tohoku Medical Megabank Community-Based Cohort Study (TMM CommCohort Study) 26) and the Tohoku Medical Megabank Brain Magnetic Resonance Imaging Study (TMM Brain MRI Study) 27) , which mainly collected MRI and neuropsychological assessment data. The participants in the TMM Brain Study were recruited from two prospective cohort studies of the Tohoku Medical Megabank Project 28) : the TMM CommCohort Study 26) and the TMM Birth and Three-Generation Cohort Study (TMM BirThree Cohort Study) 29) . The TMM CommCohort Study participants were recruited using two major approaches: the Type 1 survey, which included basic data such as sociodemographic, blood, urine, and questionnaire; and the Type 2 survey, which additionally included detailed physical assessments such as muscle strength and body composition 26) . The baseline survey of the TMM Brain Study (first MRI) was conducted between July 2014 and October 2019. The follow-up survey (second MRI) was conducted between November 2019 and March 2024 ( Supplementary Fig.1 ) .
Supplementary Fig.1. Timeline of the cohort studies related to the TMM Brain MRI Study.
TMM Brain MRI Study, Tohoku Medical Megabank Brain Magnetic Resonance Imaging Study; TMM CommCohort Study, Tohoku Medical Megabank Community-Based Cohort Study; TMM BirThree Cohort Study, TMM Birth and Three-Generation Cohort Study.
A total of 12,164 participants underwent the first MRI. We excluded eight who withdrew consent (as of March 2024) and 58 with intracranial space-occupying lesions ( Supplementary Table 1 ) , leaving 12,098 participants. Of these, 7,189 were from the TMM CommCohort Study. However, 2,520 did not complete the second MRI (66.3% follow-up rate). We excluded 74 participants with missing smoking data, leaving 4,595 for analysis. Because low-density lipoprotein cholesterol (LDL-C) measurement methods differed by survey (direct in Type 1 survey and calculated using Friedewald’s formula in Type 2 survey), we excluded 3,104 participants in the Type 2 survey when analyzing dyslipidemia exposure ( Fig.1 ) .
Supplementary Table 1. Number of cases with Intracranial space-occupying lesions on the first MRI.
| Intracranial space-occupying lesions | Number of cases |
|---|---|
| Arachnoid cyst | 36 |
| Meningioma | 14 |
| Lacunar infarction / Chronic ischemic change | 9 |
| Arteriovenous malformation | 3 |
| Extracranial brain tumor | 2 |
| Intracranial brain tumor | 1 |
| Pituitary adenoma | 1 |
| Fibrous dysplasia | 1 |
| Saccular aneurysm | 1 |
| Cerebral infarction | 1 |
Excluded 58 participants had one or more intracranial space-occupying lesions on first MRI. MRI, Magnetic Resonance Imaging.
Fig.1.
Flowchart of the selection of the study participants
All participants provided their written informed consent to participate in this study. The study was approved by the Institutional Review Board of the Tohoku Medical Megabank Organization (approval numbers: 2024-4-101 and 2024-4-108).
Exposure
The exposures were four CVD risk factors: hypertension, diabetes, dyslipidemia, and current smoking, as these factors have been reported to be associated with WML in previous studies 17 - 22) . Data on CVD risk factors were obtained from the baseline survey of the TMM CommCohort Study conducted from July 2013 to March 2016. Hypertension was defined as systolic blood pressure (SBP) ≥ 140 mmHg, diastolic blood pressure (DBP) ≥ 90 mmHg, or use of medical treatment for hypertension. Diabetes was defined as hemoglobin A1c (HbA1c) ≥ 6.5%, random blood glucose level ≥ 200 mg/dL, or use of medical treatment for diabetes. Dyslipidemia was defined as LDL-C ≥ 140 mg/dL, or under medical treatment for dyslipidemia. SBP, DBP, HbA1c, random blood glucose level, and LDL-C were measured directly using standardized instruments 26) , whereas medical treatment status was based on responses to a self-report questionnaire. Current smoking was defined based on self-reported smoking status (having smoked ≥ 100 cigarettes during their lifetime and still smoking).
Brain MRI Scanning and Volume Calculation Processing
Two MRI scanners (Philips Medical Systems Ingenia 3.0 Tesla), each equipped with a 32-channel head coil, were used to acquire brain MR images for both the first and second MRI 27) . In this study, we used three-dimensional (3D) T1-weighted imaging (T1WI) and 3D fluid-attenuated inversion recovery (FLAIR) images. The MRI acquisition parameters for the T1WI with MP-RAGE were as follows: repetition time (TR), 11 ms; echo time (TE), 5.2 ms; inversion time, 1068.3 ms; field of view (FOV), 256 mm; matrix, 368×368; slice thickness, 0.7 mm; flip angle, 8°; and acquisition time, 319 sec. The FLAIR imaging parameters were as follows: TR, 4800 ms; TE, 276 ms; inversion time, 1650 ms; FOV, 256 mm; matrix, 212×212; slice thickness, 1.2 mm; flip angle, 90°; and acquisition time, 264 sec 27) . White matter (WM) hypointensity and intracranial volume (ICV) were calculated from T1WI by FreeSurfer 7.4.1 with the Aseg atlas 30 - 33) . The WM hypointensity volume and ICV were processed using a longitudinal pipeline 34) , which enhances the sensitivity to subtle changes in brain structure over time by ensuring consistency within-subject templates. To calculate the Total Lesion Volume (TLV) values, T1WI and FLAIR images were combined using the Lesion Growth Algorithm (LGA) 35) , with a kappa value of 0.3 as the optimal default parameter for LGA. Lesions were segmented by LGA, as implemented in the Lesion Segmentation Toolbox version 3.0.0 (www.statisticalmodelling.de/lst.html) for SPM12.
Outcome
The outcome was cerebral WML, defined as WM hypointensity volume on T1WI, obtained from the TMM Brain MRI Study. A cross-sectional analysis used WML volume [mm3] from the first MRI, whereas a longitudinal analysis used the WML volume changes. Because the follow-up intervals varied (median, 4.3 years; range, 3.7–9.4 years), the WML volume changes were standardized to a 4-year interval using the following formula: (WML volume at the second MRI – WML volume at the first MRI) / interval [days] × 365.25 × 4.
Statistical Analyses
The age at the first MRI was grouped (＜50, 50–59 [50s], 60–69 [60s], ≥ 70), and basic statistics were calculated for each group. Continuous variables were presented as mean (standard deviation: SD) or median (interquartile range: IQR) and analyzed using analysis of variance (ANOVA) or the Kruskal–Wallis test to assess group differences, respectively. Categorical variables were presented as numbers (%) and compared using the chi-squared test.
To examine the association between CVD risk factors and WML, both cross-sectional and longitudinal analyses were performed using generalized linear models (GLM). GLM estimated regression coefficients (β) along with 95% confidence intervals (CI), reflecting the strength of the relationships between CVD risk factors and WML. In the cross-sectional study, for the analysis of the WML volume, we specified a log-link function with a Gaussian family, as the WML volume was not normally distributed. In the longitudinal study, for the analysis of the WML volume change, we assumed a normal distribution and conducted GLM with an identity link function. For the cross-sectional and longitudinal studies, we also conducted stratification analyses separately for each CVD risk factor (hypertension, diabetes, dyslipidemia, and current smoking).
The adjusted models included age (at the time of the first MRI), age squared, sex (male or female), ICV, hypertension, diabetes, dyslipidemia, and current smoking. When hypertension was the outcome, it was excluded from the model as a covariate. Similarly, for diabetes, dyslipidemia, and current smoking, these variables were removed as covariates, respectively. In the longitudinal analysis, the WML volume at the first MRI was added as a covariate to each model.
In addition to the primary analyses focusing on individual CVD risk factors (hypertension, diabetes, dyslipidemia, and current smoking), we conducted supplementary analyses to examine the association between the number of these factors and WML. Because the number of participants with all four risk factors was very small, participants with three or four risk factors were combined into a single category (0, 1, 2, ≥ 3) for these analyses. The models included age, age squared, sex, and ICV as covariates. The WML volume at the first MRI was additionally included as a covariate in the longitudinal analyses. In these supplementary analyses, trend tests were also performed to evaluate whether the WML volume or WML volume change increased with the number of CVD risk factors.
We conducted a validity assessment to determine whether defining the WM hypointensity on T1WI as WML, the outcome of this study, was appropriate. We used LGA, incorporating both T1WI and FLAIR images, to quantify the volume of WML from the obtained TLV values. However, FLAIR imaging was conducted only in a subset of participants in the TMM Brain MRI Study: 2,423 participants for the first MRI and 7,363 participants for the second MRI. We examined the correlation between WM hypointensity volume and TLV values in the first (n = 2,412) and second MRI (n = 7,324), excluding participants who withdrew consent or for whom FreeSurfer or an LGA analysis was not feasible. Pearson correlation coefficient (r) and 95%CI were calculated. Furthermore, we calculated the difference between the WM hypointensity volume and TLV values and examined the distribution of this difference using histograms and quantile-quantile (QQ) plots.
We also performed three sensitivity analyses. First, to address the differences in the dates of the first MRI and baseline survey of the TMM CommCohort Study ( Supplementary Fig.1 ) , we similarly performed GLM analysis using Model 1, which added the difference period [day] as a covariate to the main analytical model. Second, to account for the potential influence of additional CVD risk factors, we analyzed using Model 2, which included the following covariates: the difference period, education level (≤ 12 years, >12 years), drinking status (never-drinker, former-drinker, current-drinker), obesity (Body Mass Index [BMI] ＜25 kg/m2, BMI ≥ 25 kg/m2), and physical activity (moderate-to-vigorous physical activity [MVPA] ＜150 min/week, MVPA ≥ 150 min/week). Third, to assess the potential influence of atrial fibrillation (AF) on WML, we conducted analyses excluding participants with a self-reported history of AF (n = 58) using the main analytical model. In the longitudinal analysis, the WML volume at the first MRI was included as an additional covariate in both Model 1 and Model 2.
All analyses were performed using R version 4.1.2 (Vienna, Austria). Statistical significance was set at P＜0.05.
Results
Characteristics of Study Participants
Table 1 presents the characteristics of the study participants in each age group. The overall participants’ age at the first MRI ranged from 21 to 90 years, with a mean (SD) age of 60.3 (12.0) years. The proportion of women was lower in the older age groups. The prevalence of hypertension, diabetes, and dyslipidemia was higher in older age groups, whereas current smoking was higher in younger age groups. The mean WML volume was larger at the second MRI than that at the first MRI in all age groups and increased with increasing age group. The mean WML volume change was also larger with increasing age group. The distributions of the WML volumes at the first and second MRI, as well as WML volume changes, across the entire cohort and by age group are shown in Fig.2 .
Table 1. Characteristics of the study participants.
| Overall N = 4595 | ＜50 N = 934 (13.8%) | 50–59 N = 928 (20.2%) | 60–69 N = 1,610 (35.0%) | ≥ 70 N = 1,123 (23.5%) | P-value | |
|---|---|---|---|---|---|---|
| Women (n, %) | 2837 (61.7%) | 689 (73.8%) | 707 (76.2%) | 965 (59.9%) | 476 (42.4%) | ＜0.001 |
| Age [years] (mean, SD) | ||||||
| 1st MRI | 60.33 (12.04) | 41.32 (5.92) | 54.83 (2.88) | 65.30 (2.75) | 73.57 (3.15) | ＜0.001 |
| 2nd MRI | 64.91 (11.82) | 46.29 (5.70) | 59.43 (3.03) | 69.83 (2.76) | 77.89 (3.21) | ＜0.001 |
| From Type 1 survey (n, %) | 1491 (32.4%) | 181 (19.4%) | 245 (26.4%) | 540 (33.5%) | 525 (46.7%) | ＜0.001 |
| Hypertension (n, %) | 1419 (30.9%) | 76 (8.1%) | 222 (23.9%) | 614 (38.1%) | 507 (45.1%) | ＜0.001 |
| Under treatment (n, %) | 855 (60.3%) | 23 (30.3%) | 91 (41.1%) | 361 (58.8%) | 380 (75.0%) | ＜0.001 |
| Diabetes (n, %) | 268 (5.8%) | 6 (0.6%) | 37 (4.0%) | 113 (7.0%) | 112 (10.0%) | ＜0.001 |
| Under treatment (n, %) | 202 (75.4%) | 5 (83.3%) | 23 (62.2%) | 83 (73.5%) | 91 (81.3%) | ＜0.001 |
| Dyslipidemia (n, %) | 502 (33.7%) | 22 (12.2%) | 70 (28.6%) | 217 (40.2%) | 193 (36.8%) | ＜0.001 |
| Under treatment (n, %) | 183 (36.5%) | 1 (4.5%) | 20 (2.9%) | 84 (38.7%) | 78 (40.4%) | ＜0.001 |
| Smoking status (n, %) | ＜0.001 | |||||
| Current-smoker (n, %) | 401 (8.7%) | 101 (10.8%) | 95 (10.2%) | 136 (8.4%) | 69 (6.1%) | |
| 1–9 cigarettes/day | 70 (1.5%) | 25 (2.7%) | 18 (1.9%) | 21 (1.3%) | 6 (0.5%) | |
| 10–19 cigarettes/day | 147 (3.2%) | 43 (4.6%) | 30 (3.2%) | 49 (3.0%) | 25 (2.2%) | |
| ＞20 cigarettes/day | 184 (4.0%) | 33 (3.5%) | 47 (5.1%) | 66 (4.1%) | 38 (3.4%) | |
| Former-smoker | 1338 (29.1%) | 246 (26.3%) | 220 (23.7%) | 474 (29.4%) | 398 (35.4%) | |
| Never-smoker | 2856 (62.2%) | 587 (62.8%) | 613 (66.1%) | 1,000 (62.1%) | 656 (58.4%) | |
| Pack-year (mean, SD) | 0.17 (2.63) | 0.04 (0.74) | 0.07 (1.28) | 0.25 (3.31) | 0.24 (3.28) | 0.103 |
| Education level (n, %) | ＜0.001 | |||||
| ≤ 12 years | 2287 (49.8%) | 307 (32.9%) | 432 (46.6%) | 821 (51.0%) | 727 (64.7%) | |
| >12 years | 2257 (49.1%) | 620 (66.4%) | 489 (52.7%) | 771 (47.9%) | 377 (33.6%) | |
| Others, Unknown | 51 (1.1%) | 4 (0.4%) | 4 (0.4%) | 11 (0.7%) | 7 (0.6%) | |
| Drinking status (n, %) | 0.423 | |||||
| Current-drinker | 2788 (60.7%) | 548 (58.7%) | 575 (62.0%) | 977 (60.7%) | 688 (61.8%) | |
| Former-drinker | 89 (1.9%) | 20 (2.1%) | 15 (1.6%) | 31 (1.9%) | 23 (2.0%) | |
| Never-drinker | 1716 (37.3%) | 366 (39.2%) | 338 (36.4%) | 602 (37.4%) | 410 (36.5%) | |
| Unknown | 2 (0.1%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 2 (0.2%) | |
| Obesity (n, %) | 998 (22.0%) | 186 (20.1%) | 187 (20.4%) | 366 (23.1%) | 259 (23.4%) | 0.127 |
| Physical activity (n, %) | ＜0.001 | |||||
| MVPA ＜150 min/week | 3070 (66.8%) | 796 (85.2%) | 735 (79.2%) | 987 (61.3%) | 552 (49.2%) | |
| MVPA ≥ 150 min/week | 1525 (33.2%) | 138 (14.8%) | 193 (20.8%) | 623 (38.7%) | 571 (50.8%) | |
| Atrial fibrillation (n, %) | 58 (1.4%) | 1 (0.1%) | 2 (0.2%) | 22 (1.5%) | 33 (3.1%) | ＜0.001 |
| SBP [mmHg] (mean, SD) | 126.96 (17.59) | 115.26 (12.96) | 124.96 (16.45) | 131.79 (16.55) | 135.54 (17.46) | ＜0.001 |
| DBP [mmHg] (mean, SD) | 77.46 (10.74) | 74.19 (10.30) | 78.29 (11.11) | 79.57 (10.58) | 76.91 (10.09) | ＜0.001 |
| HbA1c (mean, SD) | 5.51 (0.52) | 5.27 (0.29) | 5.46 (0.51) | 5.59 (0.55) | 5.64 (0.57) | ＜0.001 |
| LDL-Cholesterol [mg/dL] (mean, SD) | 123.08 (30.11) | 112.78 (24.87) | 126.02 (30.59) | 127.36 (30.89) | 120.92 (29.70) | ＜0.001 |
| WML volume [mm3] (mean, SD) | ||||||
| 1st MRI | 2264.58 (3700.90) | 620.57 (373.11) | 962.81 (909.15) | 2331.76 (2812.88) | 4611.30 (5896.22) | ＜0.001 |
| 2nd MRI | 3102.06 (4934.95) | 725.61 (540.60) | 1301.41 (1585.07) | 3325.34 (4160.21) | 6246.43 (7431.49) | ＜0.001 |
| WML volume change [mm3] (mean, SD) | ||||||
| 2nd MRI – 1st MRI | 837.48 (1570.73) | 105.05 (267.49) | 338.60 (802.83) | 993.57 (1,651.18) | 1635.12 (2067.82) | ＜0.001 |
| (2nd MRI – 1st MRI) /4-year | 744.88 (1403.00) | 85.62 (224.96) | 286.32 (667.68) | 869.10 (1,425.66) | 1494.05 (1901.30) | ＜0.001 |
| MRI interval [day] (median, interquartile range) | 1567 (1499–1761) | 1672 (1540–2034) | 1594 (1511–1762) | 1556 (1498–1743) | 1514 (1487–1610) | ＜0.001 |
Data are presented -as number (%), mean (SD), or median (interquartile range). Continuous variables are analyzed using analysis of variance or the Kruskal Wallis test. Categorical variables are analyzed using Chi-squared test. The variables dyslipidemia and LDL-Cholesterol represent the characteristics of participants in the Type 1 survey only. Under treatment, n (%) refers to the total number of individuals with hypertension, diabetes, or dyslipidemia.
MRI, Magnetic Resonance Imaging; MVPA, moderate-to-vigorous physical activity; SBP, systolic blood pressure; DBP, diastolic blood pressure; HbA1c, glycated hemoglobin; LDL, low-density lipoprotein; WML, white matter lesions; SD, standard deviation.
Fig.2. Distributions of the WML volume.
Violin plots showing the crude distributions of the WML volumes, stratified by age group and for the entire cohort, as follows: (1) log (WML1), (2) log (WML2), and (3) log (WML2) – log (WML1).
WML, white matter lesions; WML1, WML volumes at the first MRI; WML2, WML volumes at the second MRI; MRI, magnetic resonance imaging.
Cross-sectional Analysis of CVD Risk Factors and WML
Table 2 presents the results of the cross-sectional analysis of the CVD risk factors and WML volume. Participants with hypertension had a significantly larger WML volume than those without hypertension in the overall sample (β [95%CI] = 0.330 [0.267–0.394]). By age group, the ＜50, 60s, and ≥ 70 age groups had significantly larger WML volumes (＜50: β [95%CI] = 0.201 [0.097–0.300], 60s: β [95%CI] = 0.357 [0.256–0.457], ≥ 70: β [95%CI] = 0.334 [0.202–0.469]). Participants with diabetes had a significantly larger WML volume than those without diabetes in the ＜50 age group and the 50s age group (＜50: β [95%CI] = 0.596 [0.377–0.790], 50s: β [95%CI] = 0.494 [0.313–0.656]). No significant associations with the WML volume were observed in participants with dyslipidemia, both in the overall and each age group. Participants who were current smokers had a significantly larger WML volume in the overall sample and in the ≥ 70 age group (overall: β [95%CI] = 0.150 [0.032–0.260], ≥ 70: β [95%CI] = 0.290 [0.030–0.511]).
Table 2. Cross-sectional study of the association between CVD risk factors and WML volume on the first MRI.
| Hypertension | Diabetes | Dyslipidemia | Current-smoking | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| β | (95% CI) | P-value | β | (95% CI) | P-value | β | (95% CI) | P-value | β | (95% CI) | P-value | |
| Overall | 0.330 | (0.267, 0.394) | ＜0.001 | 0.020 | (-0.085, 0.118) | 0.694 | -0.013 | (-0.134, 0.103) | 0.822 | 0.150 | (0.032, 0.260) | 0.007 |
| ＜50 | 0.201 | (0.097, 0.300) | ＜0.001 | 0.596 | (0.377, 0.790) | ＜0.001 | 0.043 | (-0.093, 0.168) | 0.525 | 0.023 | (-0.084, 0.123) | 0.659 |
| 50–59 | 0.068 | (-0.049, 0.181) | 0.250 | 0.494 | (0.313, 0.656) | ＜0.001 | 0.001 | (-0.146, 0.135) | 0.995 | -0.169 | (-0.361, 0.004) | 0.057 |
| 60–69 | 0.357 | (0.256, 0.457) | ＜0.001 | 0.148 | (-0.011, 0.292) | 0.049 | 0.005 | (-0.107, 0.111) | 0.931 | -0.048 | (-0.233, 0.115) | 0.584 |
| ≥ 70 | 0.334 | (0.202, 0.469) | ＜0.001 | -0.055 | (-0.295, 0.147) | 0.624 | 0.075 | (-0.067, 0.211) | 0.284 | 0.290 | (-0.030, 0.511) | 0.012 |
The variables in the adjusted model are age, age-squared, sex, ICV, hypertension, diabetes, dyslipidemia, and current-smoking. When hypertension was the outcome, we excluded the hypertension variables from the model. Similarly, in cases of diabetes, dyslipidemia, and current-smoking, we excluded these variables.
CVD, Cardiovascular Disease; WML, white matter lesions; MRI, Magnetic Resonance Imaging; CI, confidence interval; ICV, intracranial volume.
Longitudinal Analysis of CVD Risk Factors and WML Volume Change
Fig.3 shows the results of the longitudinal analysis of CVD risk factors and changes in the WML volume. For hypertension, no significant association was observed between hypertension and WML volume change in the overall participants. However, in the ＜50 and 50s age groups, the WML volume was significantly increased in participants with hypertension (＜50: β [95%CI] = 57.110 [8.013–106.204], 50s: β [95%CI] = 77.930 [6.823–149.045]). For diabetes, participants with diabetes had a significantly increased the WML volume compared to those without diabetes, which was observed only in the ＜50 age group (β [95%CI] = 376.500 [209.272–543.806]). A similar result was observed in participants with dyslipidemia (β [95%CI] = 154.000 [93.035–215.045]). Current-smoking showed a significant increase in the WML volume compared to non-current-smoking in overall participants (β [95%CI] = 113.400 [12.902–213.908]), with significant increases observed in the 50s and 60s age groups (50s: β [95%CI] = 104.500 [2.105–206.803], 60s: β [95%CI] = 305.900 [133.369–478.526]).
Fig.3. Longitudinal study of the association between CVD risk factors and changes in the WML volume.
The variables in the adjusted model were age, age-squared, sex, ICV, WML volume at first MRI, hypertension, diabetes, dyslipidemia, and current smoking. When hypertension was the outcome, we excluded the hypertension variables from the model. Similarly, in cases of diabetes, dyslipidemia, and current smoking, we excluded these variables.
CVD, cardiovascular disease; WML, white matter lesions; ICV, intracranial volume; MRI, magnetic resonance imaging; CI, confidence interval.
Supplementary Analyses to Explore the Number of CVD Risk Factors
We examined the association between the number of CVD risk factors and WML ( Supplementary Figs.2 and 3 ) . In the cross-sectional analyses, a greater number of CVD risk factors was associated with a larger WML volume overall (p for trend ＜0.001). Similar increasing trends were observed in the ＜50, 60s, and ≥ 70 age groups (all p for trend ＜0.001), whereas no significant trend was found in the 50s age group (p for trend = 0.060). In the longitudinal analyses, significant increasing trends were observed in the ＜50 (p for trend ＜0.001) and 50s (p for trend = 0.008) age groups. No significant trends were detected in the overall, 60s, and ≥ 70 age groups (p for trend = 0.113, 0.124, and 0.073, respectively).
Supplementary Fig.2. Association between the number of CVD risk factors and WML volume in cross-sectional.
The model was adjusted for age, age-squared, sex, and ICV.
CVD, cardio vascular diseases; WML, white matter lesions; ICV, intracranial volume; N, number of participants; CI, confidence interval. analyses
Supplementary Fig.3. Association between the number of CVD risk factors and WML volume change in longitudinal analyses.
The model was adjusted for age, age-squared, sex, ICV, and WML volume at the first MRI.
CVD, cardio vascular diseases; WML, white matter lesions; ICV, intracranial volume; MRI, Magnetic Resonance Imaging; N, number of participants; CI, confidence interval.
Validity Assessment of WM Hypointensity as WML
We conducted a validity assessment to examine the correlation between WM hypointensity volume and TLV values ( Supplementary Fig.4 ) . The analysis for the first MRI included 2,412 participants (mean age: 63.27, age range: 49–90 years), and the correlation was statistically significant (r = 0.959, 95%CI: 0.956–0.962, P＜0.001). The analysis for the second MRI included 7,324 participants (mean age: 59.53, age range: 26–94 years), and the correlation was also statistically significant (r = 0.959, 95%CI: 0.958–0.961, P＜0.001).
Supplementary Fig.4. Correlation between WM hypointensity volume and TLV values on the first and second MRI.
r represents the Pearson correlation coefficient with values in parentheses indicating 95% confidence intervals.
WM, white matter; TLV, Total Lesion Volume; MRI, Magnetic Resonance Imaging.
Sensitivity Analyses
We conducted three sensitivity analyses. First, to address the differences in the dates of the first MRI and baseline survey of TMM CommCohort Study, we used Model 1, in which the difference period was added as a covariate to the main analytical model ( Supplementary Figs.5 and 6 ) . Second, to account for the influence of other CVD risk factors, we used Model 2, which included education level, drinking status, obesity, and physical activity as additional covariates to Model 1 ( Supplementary Figs.7 and 8 ) . Third, to assess the influence of AF on WML, we conducted analyses excluding participants with AF (n = 58) using the main analytical model ( Supplementary Figs.9 and 10 ) . For the longitudinal analysis, WML volume at the first MRI was included as an additional covariate in all models. For both the cross-sectional and longitudinal analyses, these sensitivity analyses yielded findings similar to those of the main analysis results. Detailed β coefficients and 95% CIs for all covariates, including age-stratified estimates, are presented in Supplementary Tables 2 and 3 .
Supplementary Fig.5. Sensitivity analysis using Model 1 in the cross-sectional study.
The cross-sectional analysis was conducted in a similar manner to the main analysis, adding the period (the differences in the dates of the first MRI and the baseline survey of the TMM CommCohort Study) as an adjustment variable to the model. Specifically, Model 1 was adjusted for age, age-squared, sex, ICV, hypertension, diabetes, dyslipidemia, current-smoking, and period.
MRI, Magnetic Resonance Imaging; TMM CommCohort Study, Tohoku Medical Megabank Community-Based Cohort Study; ICV, intracranial volume; CI, confidence interval.
Supplementary Fig.6. Sensitivity analysis using Model 1 in the longitudinal study.
The longitudinal analysis was conducted in a similar manner to the main analysis, adding the period (the differences in the dates of the first MRI and the baseline survey of the TMM CommCohort Study) as an adjustment variable to the model. Specifically, Model 1 was adjusted for age, age-squared, sex, ICV, WML volume at the first MRI, hypertension, diabetes, dyslipidemia, current-smoking, and period.
MRI, Magnetic Resonance Imaging; TMM CommCohort Study, Tohoku Medical Megabank Community-Based Cohort Study; ICV, intracranial volume; WML, white matter lesions; CI, confidence interval.
Supplementary Fig.7. Sensitivity analysis using Model 2 in the cross-sectional study.
The cross-sectional analysis was conducted in a similar manner to the main analysis, with education level, drinking status, obesity, and physical activity as additional covariates to Model 1. Specifically, Model 2 was adjusted for age, age-squared, sex, ICV, hypertension, diabetes, dyslipidemia, current-smoking, period, education level, drinking status, obesity, and physical activity.
ICV, intracranial volume; CI, confidence interval.
Supplementary Fig.8. Sensitivity analysis using Model 2 in the longitudinal study.
The longitudinal analysis was conducted in a similar manner to the main analysis, with education level, drinking status, obesity, and physical activity as additional covariates to Model 1. Specifically, Model 2 was adjusted for age, age-squared, sex, ICV, WML volume at first MRI, hypertension, diabetes, dyslipidemia, current-smoking, period, education level, drinking status, obesity, and physical activity.
ICV, intracranial volume; WML, white matter lesions; MRI, Magnetic Resonance Imaging; CI, confidence interval.
Supplementary Fig.9. Sensitivity analysis using Model 2 in the cross-sectional study excluding participants with AF.
The cross-sectional analysis was conducted in a similar manner to the main analysis using Model 2, excluded participants with AF (n=58). Model 2 was adjusted for age, age-squared, sex, ICV, hypertension, diabetes, dyslipidemia, current-smoking, period, education level, drinking status, obesity, and physical activity.
AF, atrial fibrillation; ICV, intracranial volume; CI, confidence interval.
Supplementary Fig.10. Sensitivity analysis using Model 2 in the longitudinal study excluding participants with AF.
The longitudinal analysis was conducted in a similar manner to the main analysis using Model 2, excluded participants with AF (n = 58). Model 2 was adjusted for age, age-squared, sex, ICV, WML volume at first MRI, hypertension, diabetes, dyslipidemia, current-smoking, period, education level, drinking status, obesity, and physical activity. AF, atrial fibrillation; ICV, intracranial volume; WML, white matter lesions; MRI, Magnetic Resonance Imaging; CI, confidence interval.
Supplementary Table 2. Estimates for all covariates in the cross-sectional analyses (overall and age-stratified).
| Covariates | Overall | ＜50 | 50-59 | 60-69 | ≥ 70 | |||||
|---|---|---|---|---|---|---|---|---|---|---|
| β | (95%CI) | β | (95%CI) | β | (95%CI) | β | (95%CI) | β | (95%CI) | |
| Age | 0.109 | (0.033, 0.206)＊＊ | -0.087 | (-0.149, -0.019)＊＊ | 0.201 | (-0.561, 0.993) | 0.729 | (-0.438, 1.980) | 1.046 | (0.316, 1.922)＊＊ |
| Age2 | 0.000 | (-0.001, 0.000) | 0.001 | (0.000, 0.002)＊＊ | -0.002 | (-0.009, 0.005) | -0.005 | (-0.014, 0.004) | -0.006 | (-0.012, -0.002)＊＊ |
| Female | 0.292 | (0.185, 0.399)＊＊＊ | -0.205 | (-0.298, -0.110)＊＊＊ | -0.147 | (-0.297, 0.005)＊ | 0.116 | (-0.050, 0.283) | 0.475 | (0.243, 0.710)＊＊＊ |
| Hypertension | 0.363 | (0.295, 0.432)＊＊＊ | 0.192 | (0.085, 0.294)＊＊＊ | 0.078 | (-0.033, 0.186) | 0.411 | (0.302, 0.524)＊＊＊ | 0.398 | (0.251, 0.551)＊＊＊ |
| Diabetes | 0.037 | (-0.071, 0.137) | 0.539 | (0.312, 0.741)＊＊＊ | 0.529 | (0.369, 0.677)＊＊ | 0.088 | (-0.089, 0.524) | -0.005 | (-0.249, 0.201) |
| Dyslipidemia | 0.035 | (-0.036, 0.105) | 0.022 | (-0.120, 0.152) | -0.064 | (-0.212, -0.072) | -0.050 | (-0.174, 0.249) | 0.106 | (-0.044, 0.201) |
| Smoking status | ||||||||||
| Former | 0.095 | (0.014, 0.176)＊ | 0.028 | (-0.058, 0.112) | -0.087 | (-0.223, 0.047) | -0.091 | (-0.222, 0.068) | 0.205 | (0.027, 0.392)＊ |
| Current | 0.253 | (0.117, 0.382)＊＊＊ | 0.022 | (-0.095, 0.134) | -0.147 | (-0.341, 0.032) | -0.174 | (-0.396, 0.023) | 0.648 | (0.357, 0.912)＊＊＊ |
| ICV | 0.000 | (0.000, 0.000)＊＊＊ | 0.000 | (0.000, 0.000)＊＊＊ | 0.000 | (0.000, 0.000)＊＊＊ | 0.000 | (0.000, 0.000)＊＊＊ | 0.000 | (0.000, 0.000)＊＊＊ |
| Obesity | -0.008 | (-0.084, 0.065) | 0.080 | (-0.005, 0.163) | 0.163 | (0.050, 0.272)＊＊ | -0.033 | (-0.151, 0.082) | -0.046 | (-0.214, 0.110) |
| MVPA ＜150min/week | -0.170 | (-0.234, -0.105)＊＊＊ | 0.057 | (-0.056, 0.180) | 0.091 | (-0.040, 0.234) | 0.158 | (0.046, 0.275)＊＊ | -0.300 | (-0.440, -0.162)＊＊＊ |
| Education level | ||||||||||
| 12year＜ | 0.019 | (-0.046, 0.084) | 0.002 | (-0.075, 0.081) | -0.064 | (-0.165, 0.038) | -0.050 | (-0.155, 0.054) | 0.110 | (-0.029, 0.246) |
| Other, unknown | -0.346 | (-0.742, -0.058)＊ | -0.404 | (-1.460, 0.120) | -0.436 | (-2.014, 0.159) | -0.452 | (-1.551, 0.071) | -0.254 | (-1.467, 0.295) |
| Drinking status | ||||||||||
| Former | 0.085 | (-0.159, 0.289) | -0.127 | (-0.470, 0.136) | 0.003 | (-0.419, 0.319) | 0.204 | (-0.146, 0.486) | 0.074 | (-0.545, 0.485) |
| Current | 0.065 | (-0.012, 0.145) | -0.023 | (-0.100, 0.056) | -0.043 | (-0.154, 0.071) | -0.056 | (-0.181, 0.073) | 0.100 | (-0.064, 0.272) |
| Unknown | 0.932 | (0.376, 1.491)＊ | ˘ | ˘ | ˘ | 0.873 | (NA, 1.732) | |||
| Period | 0.000 | (0.000, 0.000)＊＊＊ | 0.000 | (0.000, 0.000) | 0.000 | (0.000, 0.000) | 0.000 | (0.000, 0.000) | 0.000 | (0.000, 0.000) |
Smoking status was categorized into three groups (never, former, current). The model was adjusted for age (at the first MRI), age squared, sex, hypertension, diabetes, dyslipidemia, smoking status, ICV, obesity, physical activity, education level, drinking status, and period. “―” indicates that the estimate was not calculated because no participants were present in that subgroup. “NA” indicates that the 95%CI could not be estimated due to sparse data. p-value shows as following: ＊p＜0.05, ＊＊p＜0.01, ＊＊＊p＜0.001.
CI, confidence intervals; MVPA, moderate-to-vigorous physical activity; MRI, Magnetic Resonance Imaging; ICV, intracranial volume
Supplementary Table 3. Estimates for all covariates in the longitudinal analyses (overall and age-stratified).
| Covariates | Overall | ＜50 | 50-59 | 60-69 | ≥70 | |||||
|---|---|---|---|---|---|---|---|---|---|---|
| β | (95%CI) | β | (95%CI) | β | (95%CI) | β | (95%CI) | β | (95%CI) | |
| Age | 12.66 | (-6.62, 31.95) | 12.60 | (-10.30, 35.51) | -297.20 | (-734.29, 139.91) | 711.60 | (-190.42, 1613.71) | -906.50 | (-1774.40, -38.58)＊ |
| Age2 | -0.02 | (-0.20, 0.15) | -0.14 | (-0.43, 0.16) | 2.78 | (-1.22, 6.78) | -5.41 | (-12.37, 1.55) | 5.85 | (0.08, 11.61)＊ |
| Female | 164.70 | (83.12, 246.31)＊＊＊ | 9.22 | (-29.04, 47.48) | 106.70 | (15.11, 198.21)＊ | 269.40 | (120.75, 418.10)＊＊＊ | 414.40 | (142.59, 686.17)＊＊ |
| Hypertension | 21.53 | (-43.79, 86.85) | 45.06 | (-6.09, 96.20) | 76.28 | (50.23, 147.53)＊ | -27.41 | (-131.93, 77.12) | -10.22 | (-189.16, 168.72) |
| Diabetes | 74.37 | (-47.96, 196.70) | 360.90 | (190.86, 531.01)＊＊＊ | 70.02 | (-84.08, 224.12) | 114.30 | (-81.32, 309.93) | -129.80 | (-416.25, 156.71) |
| Dyslipidemia | -10.56 | (-82.84, 61.71) | 149.20 | (84.01, 214.39)＊＊＊ | 11.30 | (-73.70, 96.29) | 56.84 | (-54.98, 168.65) | -190.90 | (-383.77, 2.06) |
| Smoking status | ||||||||||
| Former | -71.54 | (-142.88, -0.19)＊ | -11.49 | (-44.30, 21.32) | -58.53 | (-136.88, 19.82) | -74.16 | (-205.25, 56.93) | -34.68 | (-259.62, 190.27) |
| Current | 47.45 | (-59.47, 154.37) | -42.20 | (-88.47, 4.08) | 78.62 | (-29.14, 186.38) | 211.40 | (17.41, 405.41)＊ | -129.90 | (-527.32, 267.49) |
| ICV | 0.00 | (0.00, 0.00)＊＊＊ | 0.00 | (0.00, 0.00) | 0.00 | (0.00, 0.00) | 0.00 | (0.00, 0.00) | 0.00 | (0.00, 0.00)＊＊＊ |
| Obesity | 13.84 | (-54.55, 82.24) | 33.75 | (-1.28, 68.78) | -13.22 | (-88.41, 61.96) | -21.84 | (-139.09, 96.03) | -14.91 | (-220.37, 190.56) |
| MVPA ＜150min/week | 37.93 | (-25.63, 101.48) | 18.28 | (-23.26, 59.82) | 74.15 | (-1.34, 149.65) | 37.56 | (-65.37, 14.05) | 13.89 | (-171.79, 174.57) |
| Education level | ||||||||||
| 12year＜ | -3.02 | (-59.83, 53.80) | -3.82 | (-33.07, 25.44) | 23.66 | (-36.04, 83.36) | -40.75 | (-137.57, 56.07) | 109.40 | (-70.74, 289.55) |
| Other, unknown | 21.29 | (-252.37, 294.95) | 106.40 | (-61.09, 27.39) | 65.65 | (-296.22, 427.51) | -197.50 | (-650.07, 255.16) | 242.80 | (-468.32, 953.89) |
| Drinking status | ||||||||||
| Former | 46.44 | (-158.30, 251.18) | -20.47 | (-113.26, 72.31) | 8.85 | (-240.64, 258.34) | -151.40 | (-512.04, 209.21) | 360.40 | (-246.25, 967.03) |
| Current | -27.55 | (-88.73, 33.64) | 13.89 | (-15.15, 42.92) | -1.12 | (-63.84, 61.61) | -23.31 | (-132.38, 85.76) | -5.49 | (-206.44, 195.46) |
| Unknown | 2427.00 | (634.52, 4219.42)＊＊ | 2660.00 | (7.09, 5312.32)＊ | ||||||
| Period | 0.02 | (-0.03, 0.08) | 0.02 | (0.00, 0.05) | -0.02 | (-0.07, 0.04) | 0.02 | (-0.07, 0.11) | 0.06 | (-0.11, 0.23) |
| WML volume | 0.27 | (0.26, 0.28)＊＊＊ | 0.21 | (0.17, 0.25)＊＊＊ | 0.57 | (0.54, 0.61)＊＊＊ | 0.37 | (0.35, 0.39)＊＊＊ | 0.23 | (0.21, )0.25＊＊＊ |
Smoking status was categorized into three groups (never, former, current). The model was adjusted for age (at the first MRI), age squared, sex, hypertension, diabetes, dyslipidemia, smoking status, ICV, obesity, physical activity, education level, drinking status, period, and WML volume (at the first MRI). “―” indicates that the estimate was not calculated because no participants were present in that subgroup. “NA” indicates that the 95%CI could not be estimated due to sparse data. p-value shows as following: ＊p＜0.05, ＊＊p＜0.01, ＊＊＊p＜0.001.
CI, confidence intervals; MVPA, moderate-to-vigorous physical activity; MRI, Magnetic Resonance Imaging; WML, white matter lesions; ICV, intracranial volume
Discussion
This cohort study was the first to investigate the association between CVD risk factors and WML across different age groups, including younger individuals. In the longitudinal analysis, hypertension, diabetes, and dyslipidemia showed a significant increase in the WML volume in younger age groups (＜50 years). Although a significant increase in the WML volume was observed with hypertension in the 50s age group, no significant associations were found with hypertension, diabetes, or dyslipidemia in the 60s and ≥ 70 age groups. Regarding current smoking, a significant increase in the WML volume was observed in the 50s and 60s age groups. None of the risk factors showed a significant association with the WML volume change in the oldest age group (≥ 70 years). In the cross-sectional analysis, patients with hypertension exhibited significantly larger WML volumes than those without hypertension across most age groups. For diabetes, the WML volumes were significantly larger in the ＜50 and 50s age groups, with point estimates decreasing as the age group increased. For dyslipidemia, no significant association was observed in any age group. For current smoking, the WML volumes were significantly larger in only the ≥ 70 age group. Supplementary analyses examining the number of CVD risk factors showed that participants with more risk factors had larger WML volumes across most age groups in the cross-sectional analyses and greater WML volume changes in the ＜50 and 50s age groups in the longitudinal analyses. Furthermore, sensitivity analyses showed that the results remained consistent across both the cross-sectional and longitudinal analyses.
The discrepancy between the cross-sectional and longitudinal findings observed in our study likely reflects the difference between cumulative lesion burden and recent progression. Cross-sectional analyses capture the total WML accumulated over a lifetime, thus showing significant associations with CVD risk factors, even in older adults. In contrast, longitudinal analyses measure recent changes in the WML volume over the follow-up period. In older adults, a pre-existing lesion burden and the possible effects of age-related long-term vascular risk may attenuate detectable WML volume changes during follow-up, whereas in younger individuals, new WML formation may be more apparent relative to the baseline burden. In addition, selection bias related to longitudinal follow-up may have contributed to the attenuated associations in older adults. Participants who underwent follow-up MRI examinations were likely to be healthier and have better vascular risk control than those who did not participate. This effect may have resulted in an underestimation of WML progression, particularly in older age groups.
A study on individuals undergoing brain check-ups reported that significant risk factors for WML include aging and hypertension, that WML can onset as early as in those in their 30s or 40s, and that the prevalence of WML and the progression of WML grades increase with age 36) . In our study, an increase in the WML volume was observed in younger participants. Although the assessment methods differed — prior studies evaluated WML using grading scales 36) , our study used WML volumetric measures, the results showed similar trends. This may suggest that WML begins in early adulthood and tends to increase with age. In contrast, this study did not observe an increase in the WML volume associated with CVD risk factors in older adults. One possible explanation is that the WML volume was already substantial at baseline, limiting the extent of further detectable progression over the follow-up period. In this study, WML were defined based on T1WI; therefore, the observed volume changes reflected macroscopic structural alterations rather than microstructural integrity. Nevertheless, age-related degenerative processes, such as chronic demyelination and axonal loss 37 , 38) , may lead to sustained macroscopic white matter damage and a plateau in WML progression. Consequently, although vascular risk factors may continue to exert effects on cerebral microvasculature, further volumetric increases in WML may be less apparent in older adults.
Previous studies in the general population have shown that both SBP and DBP are significantly associated with an increased WML volume 39) , and antihypertensive treatment has been reported to be beneficial in preventing an increase in the WML volume 39 - 41) . The study participants showed that the proportion of individuals under hypertension treatment was the lowest in the ＜50 age group, with the percentage increasing as the age group advanced. Therefore, this may explain the findings of the study, in which the change in the WML volume was significantly greater in the ＜50 and 50s age groups, while no significant changes were observed in the group aged 60s and above.
In our cross-sectional study, we observed a trend indicating that the association between diabetes and WML appears to be weaker in older age groups. Furthermore, in the longitudinal study, a significant association was observed only in the ＜50 age group, while no such association was found in the other age groups. Based on these findings, we considered that diabetes treatment might also affect the WML volume. However, previous studies have suggested that further research is needed to determine the effectiveness of diabetes treatment in preventing WML 42) .
Dyslipidemia has not been consistently associated with WML in the existing literature. While cross-sectional studies have reported no significant association 43) , some have even suggested that high levels of cholesterol and triglycerides may be linked to a lower risk of WML and may have a protective effect 44 , 45) , longitudinal studies in older adults have shown the opposite trend—namely, that increased high-density lipoprotein cholesterol and decreased LDL-C are associated with the worsening of WM grades 46) . A recent review concluded that there is no consensus on the relationship between dyslipidemia and WML 42) . Inconsistencies across studies, including the present one, may be attributable to differences in lipid subtypes, study populations, and study designs. Thus, the role of dyslipidemia in the development or progression of WML remains unclear and warrants further investigation.
Smoking was associated with the WML volume in overall sample and ≥ 70 age groups in this cross-sectional analysis. In contrast, the longitudinal analysis revealed a significant increase in the WML volume in the 50s and 60s age groups. These results suggest that the WML volume may begin to increase in the 50s and 60s individuals, and become substantially larger in ≥ 70 individuals. One possible explanation for the lack of a significant association between smoking and WML increase in younger individuals (＜50 age group; 21–49 years) may be due to the legal smoking age in Japan being 20 years, and smoking may not immediately affect WML progression. Since a higher pack-year can negatively affect the WML volume and WM microstructure 13) , we explored cumulative smoking exposure using pack-year as a continuous measure in the main models ( Supplementary Table 4 ) . These results suggested that greater cumulative exposure may be associated with larger WML volumes in some age groups, and with WML progression particularly in younger adults. Therefore, prolonged smoking might contribute to WM damage. In this study, smoking status was defined based on current smoking or not; former smokers were classified as non-current smokers. This classification may have led to an underestimation of the association between smoking and WML. The ≥ 70 age group had the highest proportion of former smokers, which raises the possibility that smoking cessation due to underlying health conditions may have contributed to the suppression of WML progression. Nonetheless, our findings indicated a significant association between smoking and WML. Several studies have reported that Asians, including the Japanese, did not find an association between smoking and WML, whereas some studies in European and American populations have reported a significant association 13 , 36 , 47 - 49) . This suggests that regional or ethnic differences may influence the association, and our results diverge from those reported in previous research. However, the biological mechanisms by which smoking influences WML remain unclear 42) , and further research is needed to clarify this relationship.
Supplementary Table 4. The dose-response relationship between smoking (pack-year) and WML.
| Cross-sectional | Longitudinal | |||||
|---|---|---|---|---|---|---|
| β | (95% CI) | P-value | β | (95%CI) | P-value | |
| Overall | 0.008 | (0.005, 0.011) | ＜0.001 | -10.160 | (-20.740, 0.415) | 0.060 |
| ＜50 | 0.040 | (0.020, 0.055) | ＜0.001 | 26.350 | (8.716, 43.990) | 0.003 |
| 50–59 | 0.019 | (-0.001, 0.031) | 0.010 | 81.760 | (58.856, 104.669) | ＜0.001 |
| 60–69 | 0.002 | (-0.006, 0.007) | 0.568 | -25.510 | (-39.675, -11.344) | ＜0.001 |
| ≥ 70 | 0.011 | (0.004, 0.016) | ＜0.001 | -5.336 | (-30.234, 19.563) | 0.675 |
We conducted GLM analyses using pack-year as a continuous measure in the main analytical models. WML, white matter lesions; CI, confidence interval.
This study has several strengths. First, it involved a population of community residents with a wide age range (21–90 years at the time of the first MRI). Moreover, we stratified the participants by age group to reveal differences in the associations between CVD risk factors and WML across various age groups. Second, we conducted both cross-sectional and longitudinal analyses within the same cohort and found that the association between CVD risk factors and WML differed between the cross-sectional and longitudinal analyses. In the cross-sectional analysis stratified by age group, hypertension was particularly associated with the WML volume, whereas in the longitudinal analysis, increases in the WML volume were observed for most of the examined CVD risk factors, especially in younger adults. These differences underscore the need to examine WML volume changes over time.
Our study has some limitations. First, the baseline survey of the TMM CommCohort Study and the first MRI scan were not conducted the same date ( Supplementary Fig.1 ) . All participants in this study underwent the baseline survey of the TMM CommCohort Study, the first MRI, and the second MRI in this order. For the analysis, we assumed that the baseline survey of the TMM CommCohort Study and the first MRI scan were conducted on the same date. To address this limitation, we conducted sensitivity analyses by adjusting for the date difference (period). The results remained unchanged in both cross-sectional and longitudinal analyses ( Supplementary Figs.5 and 6 ) . Second, this study may have been affected by selection bias. Participants in the TMM Brain MRI Study were recruited from both the TMM CommCohort Study and the TMM BirThree Cohort Study. Therefore, the study participants may have included individuals with a high health awareness and motivation. Furthermore, of the 7,189 participants from the TMM CommCohort Study, 2,520 were lost to follow-up, resulting in an attrition rate of approximately 35%. As those who remained were sufficiently healthy and active to visit the study site, this attrition could have influenced the results of this study, particularly in the older age group. However, the baseline characteristics were similar between the included and excluded (lost to follow-up) participants, suggesting that the influence of selection bias is likely minimal ( Supplementary Table 5 ) . While the results of this study may have been underestimated due to the selection bias, it is unlikely that the findings would be completely reversed in a general population. Therefore, it would not be appropriate to entirely dismiss the generalizability of these findings. Third, WML was evaluated only using T1WI. WML is usually assessed using FLAIR, T2-weighted imaging (T2WI), and proton density (PD) 50) . Although the use of FLAIR and T2WI/PD provides more accurate estimates of WML than using T1WI alone, strong correlations with these WML were observed, even in the absence of T2WI/PD or FLAIR, and valuable estimates of WML can still be obtained using T1WI 51) . However, FLAIR imaging was acquired only in a subset of participants: 2,423 participants for the first MRI, and 7,363 for the second MRI. To ensure an adequate sample size for the present analysis, we used T1WI as the outcome measures. Therefore, in this study, we examined the correlation between WM hypointensity on T1WI and TLV values on T1WI/FLAIR to assess the validity of using T1WI alone for WML evaluation. We confirmed the correlation between these values ( Supplementary Fig.2 ) . Furthermore, to assess the potential measurement bias between the two imaging modalities, we calculated the difference between WM hypointensity and TLV values, and examined the distribution of the differences. The distribution of the differences was centered around zero, with the histogram and QQ plots visually confirming normality, indicating no significant bias in the measurements.
Supplementary Table 5. Comparison of baseline characteristics between included participants and those lost to follow-up.
| Overall | ＜50 | 50–59 | 60–69 | ≥ 70 | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Included | Not followed | Included | Not followed | Included | Not followed | Included | Not followed | Included | Not followed | |
| N = 4,595 | N = 2,520 | N = 934 | N = 467 | N = 928 | N = 356 | N = 1,610 | N = 839 | N = 1,123 | N = 858 | |
| Women | 2837 (61.7%) | 1602 (63.6%) | 689 (73.8%) | 359 (76.9%) | 707 (76.2%) | 268 (75.3%) | 965 (59.9%) | 558 (66.5%)＊＊ | 476 (42.4%) | 417 (48.6)＊＊ |
| Age at 1st MRI [years] | 60.33 (12.04) | 62.16 (13.01)＊＊＊ | 41.32 (5.92) | 39.77 (6.59)＊＊＊ | 54.83 (2.88) | 54.00 (2.82) | 65.30 (2.75) | 65.40 (2.67) | 73.57 (3.15) | 74.23 (3.39)＊＊＊ |
| From Type 1 survey | 1491 (32.4%) | 1093 (43.4%)＊＊＊ | 181 (19.4%) | 125 (26.8%)＊＊ | 245 (26.4%) | 120 (33.7%)＊ | 540 (33.5%) | 372 (44.3%)＊＊＊ | 525 (46.7%) | 476 (55.5%)＊＊＊ |
| Hypertension | 1419 (30.9%) | 870 (34.5%)＊ | 76 (8.1%) | 41 (8.8%) | 222 (23.9%) | 100 (28.1%) | 614 (38.1%) | 322 (38.4%) | 507 (45.1%) | 407 (47.4%) |
| Under treatment | 855 (60.3%) | 594 (68.3%) | 23 (30.3%) | 9 (22.0%) | 91 (41.1%) | 51 (51.0%) | 361 (58.8%) | 214 (66.5%) | 380 (75.0%) | 320 (78.6%) |
| Diabetes | 268 (5.8%) | 185 (7.3%)＊ | 6 (0.6%) | 4 (0.9%) | 37 (4.0%) | 16 (4.5%) | 113 (7.0%) | 71 (33.1%) | 112 (10.0%) | 94 (11.0%) |
| Under treatment | 202 (75.4%) | 138 (74.6%) | 5 (83.3%) | 1 (25.0%) | 23 (62.2%) | 10 (62.5%) | 83 (73.5%) | 51 (71.8%) | 91 (81.3%) | 76 (80.9%) |
| Dyslipidemia | 502 (33.7%) | 375 (34.3%) | 22 (12.2%) | 12 (9.6%) | 70 (28.6%) | 32 (26.7%) | 217 (40.2%) | 141 (37.9%) | 193 (36.8%) | 190 (39.9%) |
| Under treatment | 183 (36.5%) | 148 (39.5%) | 1 (4.5%) | 2 (16.7%) | 20 (2.9%) | 6 (18.8%) | 84 (38.7%) | 57 (40.4%) | 78 (40.4%) | 83 (43.7%) |
| Smoking status | ＊ | |||||||||
| Current-smoker | 401 (8.7%) | 266 (10.6%) | 101 (10.8%) | 76 (16.3%)＊＊ | 95 (10.2%) | 44 (12.4%) | 136 (8.4%) | 83 (9.9%) | 69 (6.1%) | 63 (7.3%) |
| 1–9cigarettes/day | 70 (1.5%) | 38 (1.5%) | 25 (2.7%) | 16 (3.4%) | 18 (1.9%) | 4 (1.1%) | 21 (1.3%) | 9 (1.1%) | 6 (0.5%) | 9 (1.0%) |
| 10–19cigarettes/day | 147 (3.2%) | 114 (4.5%) | 43 (4.6%) | 33 (7.1%) | 30 (3.2%) | 17 (4.8%) | 49 (3.0%) | 35 (4.2%) | 25 (2.2%) | 29 (3.4%) |
| ＞20cigarettes/day | 184 (4.0%) | 114 (4.5%) | 33 (3.5%) | 27 (5.8%) | 47 (5.1%) | 23 (6.5%) | 66 (4.1%) | 39 (4.6%) | 38 (3.4%) | 25 (2.9%) |
| Former-smoker | 1338 (29.1%) | 681 (27.0%) | 246 (26.3%) | 100 (21.4%) | 220 (23.7%) | 95 (26.7%) | 474 (29.4%) | 194 (23.1%) | 398 (35.4%) | 292 (34.0%) |
| Never-smoker | 2856 (62.2%) | 1573 (62.4%) | 587 (62.8%) | 291 (62.3%) | 613 (66.1%) | 217 (61.0%) | 1,000 (62.1%) | 562 (67.0%) | 656 (58.4%) | 503 (58.6%) |
| Pack-year | 0.17 (2.63) | 0.23 (3.41) | 0.04 (0.74) | 0.01 (0.23) | 0.07 (1.28) | 0.05 (0.49) | 0.25 (3.31) | 0.14 (3.05) | 0.24 (3.28) | 0.52 (4.98) |
| Education level | ＊＊＊ | ＊＊＊ | ＊ | |||||||
| ≤ 12 years | 2287 (49.8%) | 1485 (58.9%) | 307 (32.9%) | 174 (37.3%) | 432 (46.6%) | 185 (52.0%) | 821 (51.0%) | 518 (61.7%) | 727 (64.7%) | 608 (70.9%) |
| ＞12 years | 2257 (49.1%) | 1006 (39.9%) | 620 (66.4%) | 289 (61.9%) | 489 (52.7%) | 169 (47.5%) | 771 (47.9%) | 310 (36.9%) | 377 (33.6%) | 238 (27.7%) |
| Others, Unknown | 51 (1.1%) | 29 (1.2%) | 4 (0.4%) | 4 (0.9%) | 4 (0.4%) | 2 (0.6%) | 11 (0.7%) | 11 (1.3%) | 7 (0.6%) | 12 (1.4%) |
| Drinking status | ＊ | |||||||||
| Current-drinker | 2788 (60.7%) | 1449 (57.5%) | 548 (58.7%) | 265 (56.7%) | 575 (62.0%) | 222 (62.4%) | 977 (60.7%) | 341 (40.6%) | 688 (61.8%) | 364 (42.4%) |
| Former-drinker | 89 (1.9%) | 49 (1.9%) | 20 (2.1%) | 13 (2.8%) | 15 (1.6%) | 7 (2.0%) | 31 (1.9%) | 10 (1.2%) | 23 (2.0%) | 19 (2.2%) |
| Never-drinker | 1716 (37.3%) | 1020 (40.5%) | 366 (39.2%) | 188 (40.3%) | 338 (36.4%) | 127 (35.7%) | 602 (37.4%) | 487 (58.0%) | 410 (36.5%) | 475 (55.4%) |
| Unknown | 2 (0.1%) | 2 (0.1%) | 0 (0.0%) | 1 (0.2%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 1 (0.1%) | 2 (0.2%) | 0 (0.0%) |
| Obesity | 998 (22.0%) | 594 (23.8%) | 186 (20.1%) | 76 (16.5%) | 187 (20.4%) | 92 (26.0%)＊ | 366 (23.1%) | 204 (24.5%) | 259 (23.4%) | 222 (26.1%) |
| Physical activity | ||||||||||
| MVPA ＜150 min/week | 3070 (66.8%) | 1609 (69.8%) | 796 (85.2%) | 397 (88.0%) | 735 (79.2%) | 272 (80.2%) | 987 (61.3%) | 523 (62.3%) | 552 (49.2%) | 417 (48.6%) |
| MVPA ≥ 150 min/week | 1525 (33.2%) | 911 (36.2%) | 138 (14.8%) | 70 (15.0%) | 193 (20.8%) | 84 (19.8%) | 623 (38.7%) | 316 (37.7%) | 571 (50.8%) | 441 (51.4%) |
| Atrial fibrillation | 58 (1.4%) | 37 (1.7%) | 1 (0.1%) | 0 (0.0%) | 2 (0.2%) | 2 (0.6%) | 22 (1.5%) | 16 (2.3%) | 33 (3.1%) | 19 (2.5%) |
| SBP [mmHg] | 126.96 (17.59) | 129.24 (18.13)＊＊＊ | 115.26 (12.96) | 116.50 (13.98) | 124.96 (16.45) | 127.17 (16.85) | 131.79 (16.55) | 133.39 (17.16) | 135.54 (17.46) | 136.83 (17.13) |
| DBP [mmHg] | 77.46 (10.74) | 78.17 (10.67)＊ | 74.19 (10.30) | 74.95 (10.52) | 78.29 (11.11) | 80.48 (11.41)＊ | 79.57 (10.58) | 80.01 (10.12) | 76.91 (10.09) | 77.38 (10.20) |
| HbA1c [%] | 5.51 (0.52) | 5.55 (0.52)＊ | 5.27 (0.29) | 5.25 (0.32) | 5.46 (0.51) | 5.47 (0.58) | 5.59 (0.55) | 5.61 (0.52) | 5.64 (0.57) | 5.68 (0.51) |
| LDL-C [mg/dL] | 123.08 (30.11) | 123.88 (29.61) | 112.78(24.87) | 110.49 (25.45) | 126.02(30.59) | 130.38 (31.41) | 127.36(30.89) | 126.60 (29.93) | 120.92(29.70) | 123.01 (29.26) |
| WML volume [mm3] | 2264.58 (3700.90) | 2974.59 (4408.00) | 620.57 (373.11) | 657.69 (620.18) | 962.81 (909.15) | 1117.52 (1249.94)＊ | 2331.76 (2812.88) | 2749.95 (3962.70)＊＊ | 4611.30 (5896.22) | 5225.60 (5605.07)＊ |
Continuous variables are presented as number (%), and are analyzed using analysis of variance. Categorical variables are presented as mean (SD), and are analyzed using Chi-squared test. The variables dyslipidemia and LDL-Cholesterol represent the characteristics of participants in the Type 1 survey only. Under treatment, n (%) refers to the total number of individuals with hypertension, diabetes, or dyslipidemia. p-value shows as following: ＊p＜0.05, ＊＊p＜0.01, ＊＊＊p＜0.001.
MRI, Magnetic Resonance Imaging; MVPA, moderate-to-vigorous physical activity; SBP, systolic blood pressure; DBP, diastolic blood pressure; HbA1c, glycated hemoglobin; LDL-C, low-density lipoprotein cholesterol; WML, white matter lesions; SD, standard deviation.
Conclusion
In conclusion, CVD risk factors made a particularly significant contribution to WML in the younger age group. Although the cross-sectional analysis indicated that factors other than dyslipidemia were associated with WML, the longitudinal analysis examining changes over time revealed that hypertension, diabetes, and dyslipidemia were particularly related to WML progression in the younger age group. This suggests that early preventive interventions may be crucial for reducing the progression of WML and they may help prevent cerebrovascular diseases or dementia. By incorporating a longitudinal design, which is scarce in previous studies, we were able to describe WML progression across age groups, offering insights that cannot be obtained from cross-sectional analyses alone. Further cohort and mechanistic studies are warranted to better understand these findings.
Conflict of Interest
The authors declare that they have no conflict of interest.
Acknowledgements
The authors are deeply grateful to the members of ToMMo, including the Genome Medical Research Coordinators, office, and administrative personnel, for their assistance. The complete list of members is available at: https://www.megabank.tohoku.ac.jp/english/a240901/.
Notice of Grant Support
This study was supported by the Reconstruction Agency, the Ministry of Education, Culture, Sports, Science and Technology (MEXT), and the Japan Agency for Medical Research and Development (AMED, JP 21 km0105001, JP 21 km0105002).
Author Contributions
Megumi Satake: Methodology, Formal analysis, Writing- Original draft preparation. Ippei Chiba: Methodology, Writing- Original draft preparation. Mana Kogure: Reviewing. Rieko Hatanaka: Reviewing. Kumi Nakaya: Reviewing. Masato Takase: Reviewing. Sayuri Tokioka: Reviewing. Naoki Nakaya: Reviewing. Naoko Mori: Reviewing. Takuya Koyama: Brain MRI data preprocessing. Yuto Abe: Brain MRI scanning and preprocessing. Yasuyuki Taki: Reviewing. Nobuo Fuse: Reviewing. Kengo Kinoshita: Reviewing. Yoko Izumi: Reviewing. Shunji Mugikura: Methodology, Reviewing. Atsushi Hozawa: Supervision, Methodology, Writing- Reviewing and Editing.
References
- 1).GBD 2019 Dementia Forecasting Collaborators: Estimation of the Global Prevalence of Dementia in 2019 and Forecasted Prevalence in 2050: An Analysis for the Global Burden of Disease Study 2019. Lancet Public Health, 2022; 7: e105-e125 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2).Mustapha M, Nassir C M N C M, Aminuddin N, Safri A A, Ghazali M M: Cerebral Small Vessel Disease (CSVD) - Lessons From the Animal Models. Front. Physiol., 2019; 10: 1317 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3).Vermeer S E, Prins N D, den Heijer T, Hofman A, Koudstaal P J, Breteler M M B: Silent Brain Infarcts and the Risk of Dementia and Cognitive Decline. N. Engl. J. Med., 2003; 348: 1215-1222 [DOI] [PubMed] [Google Scholar]
- 4).Charidimou A, Werring D J: Cerebral Microbleeds and Cognition in Cerebrovascular Disease: An Update. J. Neurol. Sci., 2012; 322: 50-55 [DOI] [PubMed] [Google Scholar]
- 5).Werring D J, Gregoire S M, Cipolotti L: Cerebral Microbleeds and Vascular Cognitive Impairment. J. Neurol. Sci., 2010; 299: 131-135 [DOI] [PubMed] [Google Scholar]
- 6).Debette S, Schilling S, Duperron M-G, Larsson S C, Markus H S: Clinical Significance of Magnetic Resonance Imaging Markers of Vascular Brain Injury: A Systematic Review and Meta-Analysis. JAMA Neurol., 2019; 76: 81-94 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7).Bos D, Wolters F J, Darweesh S K L, Vernooij M W, de Wolf F, Ikram M A, Hofman A: Cerebral Small Vessel Disease and the Risk of Dementia: A Systematic Review and Meta-Analysis of Population-Based Evidence. Alzheimers Dement. J. Alzheimers Assoc., 2018; 14: 1482-1492 [DOI] [PubMed] [Google Scholar]
- 8).Wardlaw J M, Smith E E, Biessels G J, Cordonnier C, Fazekas F, Frayne R, Lindley R I, O’Brien J T, Barkhof F, Benavente O R, Black S E, Brayne C, Breteler M, Chabriat H, Decarli C, de Leeuw F-E, Doubal F, Duering M, Fox N C, Greenberg S, Hachinski V, Kilimann I, Mok V, Oostenbrugge R van, Pantoni L, Speck O, Stephan B C M, Teipel S, Viswanathan A, Werring D, Chen C, Smith C, van Buchem M, Norrving B, Gorelick P B, Dichgans M, STandards for ReportIng Vascular changes on nEuroimaging (STRIVE v1): Neuroimaging Standards for Research into Small Vessel Disease and Its Contribution to Ageing and Neurodegeneration. Lancet Neurol., 2013; 12: 822-838 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9).Pasi M, Cordonnier C: Clinical Relevance of Cerebral Small Vessel Diseases. Stroke, 2020; 51: 47-53 [DOI] [PubMed] [Google Scholar]
- 10).Nakazawa T, Ohara T, Hirabayashi N, Furuta Y, Hata J, Shibata M, Honda T, Kitazono T, Nakao T, Ninomiya T: Association of White Matter Lesions and Brain Atrophy with the Development of Dementia in a Community: The Hisayama Study. Psychiatry Clin. Neurosci., 2023; 77: 330-337 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11).Scheltens P, Barkhof F, Valk J, Algra P R, van der Hoop R G, Nauta J, Wolters E C: White Matter Lesions on Magnetic Resonance Imaging in Clinically Diagnosed Alzheimer’s Disease. Evidence for Heterogeneity. Brain J. Neurol., 1992; 115 ( Pt 3): 735-748 [DOI] [PubMed] [Google Scholar]
- 12).Au R, Massaro J M, Wolf P A, Young M E, Beiser A, Seshadri S, D’Agostino R B, DeCarli C: Association of White Matter Hyperintensity Volume With Decreased Cognitive Functioning: The Framingham Heart Study. Arch. Neurol., 2006; 63: 246-250 [DOI] [PubMed] [Google Scholar]
- 13).Cox S R, Lyall D M, Ritchie S J, Bastin M E, Harris M A, Buchanan C R, Fawns-Ritchie C, Barbu M C, de Nooij L, Reus L M, Alloza C, Shen X, Neilson E, Alderson H L, Hunter S, Liewald D C, Whalley H C, McIntosh A M, Lawrie S M, Pell J P, Tucker-Drob E M, Wardlaw J M, Gale C R, Deary I J: Associations between Vascular Risk Factors and Brain MRI Indices in UK Biobank. Eur. Heart J., 2019; 40: 2290-2300 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14).Garnier-Crussard A, Bougacha S, Wirth M, Dautricourt S, Sherif S, Landeau B, Gonneaud J, De Flores R, de la Sayette V, Vivien D, Krolak-Salmon P, Chételat G: White Matter Hyperintensity Topography in Alzheimer’s Disease and Links to Cognition. Alzheimers Dement. J. Alzheimers Assoc., 2022; 18: 422-433 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15).Hu H-Y, Ou Y-N, Shen X-N, Qu Y, Ma Y-H, Wang Z-T, Dong Q, Tan L, Yu J-T: White Matter Hyperintensities and Risks of Cognitive Impairment and Dementia: A Systematic Review and Meta-Analysis of 36 Prospective Studies. Neurosci. Biobehav. Rev., 2021; 120: 16-27 [DOI] [PubMed] [Google Scholar]
- 16).Wang Y-L, Chen W, Cai W-J, Hu H, Xu W, Wang Z-T, Cao X-P, Tan L, Yu J-T, Alzheimer’s Disease Neuroimaging Initiative: Associations of White Matter Hyperintensities with Cognitive Decline: A Longitudinal Study. J. Alzheimers Dis. JAD, 2020; 73: 759-768 [DOI] [PubMed] [Google Scholar]
- 17).Koohi F, Harshfield E L, Markus H S: Contribution of Conventional Cardiovascular Risk Factors to Brain White Matter Hyperintensities. J. Am. Heart Assoc., 2023; 12: e030676 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18).Power M C, Tingle J V, Reid R I, Huang J, Sharrett A R, Coresh J, Griswold M, Kantarci K, Jack C R, Knopman D, Gottesman R F, Mosley T H: Midlife and Late-Life Vascular Risk Factors and White Matter Microstructural Integrity: The Atherosclerosis Risk in Communities Neurocognitive Study. J. Am. Heart Assoc., 2017; 6: e005608 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19).Keller J A, Kant I M J, Slooter A J C, van Montfort S J T, van Buchem M A, van Osch M J P, Hendrikse J, de Bresser J: Different Cardiovascular Risk Factors Are Related to Distinct White Matter Hyperintensity MRI Phenotypes in Older Adults. NeuroImage Clin., 2022; 35: 103131 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20).de Leeuw F-E, de Groot J C, Oudkerk M, Witteman J C M, Hofman A, van Gijn J, Breteler M M B: Hypertension and Cerebral White Matter Lesions in a Prospective Cohort Study. Brain J. Neurol., 2002; 125: 765-772 [DOI] [PubMed] [Google Scholar]
- 21).Pase M P, Himali J J, Mitchell G F, Beiser A, Maillard P, Tsao C, Larson M G, DeCarli C, Vasan R S, Seshadri S: Association of Aortic Stiffness With Cognition and Brain Aging in Young and Middle-Aged Adults: The Framingham Third Generation Cohort Study. Hypertens. Dallas Tex 1979, 2016; 67: 513-519 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22).Breteler M M, van Swieten J C, Bots M L, Grobbee D E, Claus J J, van den Hout J H, van Harskamp F, Tanghe H L, de Jong P T, van Gijn J: Cerebral White Matter Lesions, Vascular Risk Factors, and Cognitive Function in a Population-Based Study: The Rotterdam Study. Neurology, 1994; 44: 1246-1252 [DOI] [PubMed] [Google Scholar]
- 23).Wang Z, Chen Q, Chen J, Yang N, Zheng K: Risk Factors of Cerebral Small Vessel Disease: A Systematic Review and Meta-Analysis. Medicine (Baltimore), 2021; 100: e28229 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24).Livingston G, Huntley J, Liu K Y, Costafreda S G, Selbæk G, Alladi S, Ames D, Banerjee S, Burns A, Brayne C, Fox N C, Ferri C P, Gitlin L N, Howard R, Kales H C, Kivimäki M, Larson E B, Nakasujja N, Rockwood K, Samus Q, Shirai K, Singh-Manoux A, Schneider L S, Walsh S, Yao Y, Sommerlad A, Mukadam N: Dementia Prevention, Intervention, and Care: 2024 Report of the Lancet Standing Commission. The Lancet, 2024; 404: 572-628 [DOI] [PubMed] [Google Scholar]
- 25).Groh J, Simons M: White Matter Aging and Its Impact on Brain Function. Neuron, 2025; 113: 127-139 [DOI] [PubMed] [Google Scholar]
- 26).Hozawa A, Tanno K, Nakaya N, Nakamura T, Tsuchiya N, Hirata T, Narita A, Kogure M, Nochioka K, Sasaki R, Takanashi N, Otsuka K, Sakata K, Kuriyama S, Kikuya M, Tanabe O, Sugawara J, Suzuki K, Suzuki Y, Kodama E N, Fuse N, Kiyomoto H, Tomita H, Uruno A, Hamanaka Y, Metoki H, Ishikuro M, Obara T, Kobayashi T, Kitatani K, Takai-Igarashi T, Ogishima S, Satoh M, Ohmomo H, Tsuboi A, Egawa S, Ishii T, Ito K, Ito S, Taki Y, Minegishi N, Ishii N, Nagasaki M, Igarashi K, Koshiba S, Shimizu R, Tamiya G, Nakayama K, Motohashi H, Yasuda J, Shimizu A, Hachiya T, Shiwa Y, Tominaga T, Tanaka H, Oyama K, Tanaka R, Kawame H, Fukushima A, Ishigaki Y, Tokutomi T, Osumi N, Kobayashi T, Nagami F, Hashizume H, Arai T, Kawaguchi Y, Higuchi S, Sakaida M, Endo R, Nishizuka S, Tsuji I, Hitomi J, Nakamura M, Ogasawara K, Yaegashi N, Kinoshita K, Kure S, Sakai A, Kobayashi S, Sobue K, Sasaki M, Yamamoto M: Study Profile of the Tohoku Medical Megabank Community-Based Cohort Study. J. Epidemiol., 2021; 31: 65-76 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27).Taira M, Mugikura S, Mori N, Hozawa A, Saito T, Nakamura T, Kiyomoto H, Kobayashi T, Ogishima S, Nagami F, Uruno A, Shimizu R, Kobayashi T, Yasuda J, Kure S, Sakurai M, Motoike I N, Kumada K, Nakaya N, Obara T, Oba K, Sekiguchi A, Thyreau B, Mutoh T, Takano Y, Abe M, Maikusa N, Tatewaki Y, Taki Y, Yaegashi N, Tomita H, Kinoshita K, Kuriyama S, Fuse N, Yamamoto M: Tohoku Medical Megabank Brain Magnetic Resonance Imaging Study: Rationale, Design, and Background. JMA J., 2023; 6: 246-264 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28).Kuriyama S, Yaegashi N, Nagami F, Arai T, Kawaguchi Y, Osumi N, Sakaida M, Suzuki Y, Nakayama K, Hashizume H, Tamiya G, Kawame H, Suzuki K, Hozawa A, Nakaya N, Kikuya M, Metoki H, Tsuji I, Fuse N, Kiyomoto H, Sugawara J, Tsuboi A, Egawa S, Ito K, Chida K, Ishii T, Tomita H, Taki Y, Minegishi N, Ishii N, Yasuda J, Igarashi K, Shimizu R, Nagasaki M, Koshiba S, Kinoshita K, Ogishima S, Takai-Igarashi T, Tominaga T, Tanabe O, Ohuchi N, Shimosegawa T, Kure S, Tanaka H, Ito S, Hitomi J, Tanno K, Nakamura M, Ogasawara K, Kobayashi S, Sakata K, Satoh M, Shimizu A, Sasaki M, Endo R, Sobue K, Group the T M M P S, Yamamoto M: The Tohoku Medical Megabank Project: Design and Mission. J. Epidemiol., 2016; 26: 493 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29).Kuriyama S, Metoki H, Kikuya M, Obara T, Ishikuro M, Yamanaka C, Nagai M, Matsubara H, Kobayashi T, Sugawara J, Tamiya G, Hozawa A, Nakaya N, Tsuchiya N, Nakamura T, Narita A, Kogure M, Hirata T, Tsuji I, Nagami F, Fuse N, Arai T, Kawaguchi Y, Higuchi S, Sakaida M, Suzuki Y, Osumi N, Nakayama K, Ito K, Egawa S, Chida K, Kodama E, Kiyomoto H, Ishii T, Tsuboi A, Tomita H, Taki Y, Kawame H, Suzuki K, Ishii N, Ogishima S, Mizuno S, Takai-Igarashi T, Minegishi N, Yasuda J, Igarashi K, Shimizu R, Nagasaki M, Tanabe O, Koshiba S, Hashizume H, Motohashi H, Tominaga T, Ito S, Tanno K, Sakata K, Shimizu A, Hitomi J, Sasaki M, Kinoshita K, Tanaka H, Kobayashi T, Tohoku Medical Megabank Project Study Group, Kure S, Yaegashi N, Yamamoto M: Cohort Profile: Tohoku Medical Megabank Project Birth and Three-Generation Cohort Study (TMM BirThree Cohort Study): Rationale, Progress and Perspective. Int. J. Epidemiol., 2020; 49: 18-19m [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30).Desikan R S, Ségonne F, Fischl B, Quinn B T, Dickerson B C, Blacker D, Buckner R L, Dale A M, Maguire R P, Hyman B T, Albert M S, Killiany R J: An Automated Labeling System for Subdividing the Human Cerebral Cortex on MRI Scans into Gyral Based Regions of Interest. NeuroImage, 2006; 31: 968-980 [DOI] [PubMed] [Google Scholar]
- 31).Fischl B, Salat D H, Busa E, Albert M, Dieterich M, Haselgrove C, van der Kouwe A, Killiany R, Kennedy D, Klaveness S, Montillo A, Makris N, Rosen B, Dale A M: Whole Brain Segmentation: Automated Labeling of Neuroanatomical Structures in the Human Brain. Neuron, 2002; 33: 341-355 [DOI] [PubMed] [Google Scholar]
- 32).Fischl B, Salat D H, van der Kouwe A J W, Makris N, Ségonne F, Quinn B T, Dale A M: Sequence-Independent Segmentation of Magnetic Resonance Images. NeuroImage, 2004; 23 Suppl 1: S69-84 [DOI] [PubMed] [Google Scholar]
- 33).Fischl B: FreeSurfer. NeuroImage, 2012; 62: 774-781 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34).Reuter M, Schmansky N J, Rosas H D, Fischl B: Within-Subject Template Estimation for Unbiased Longitudinal Image Analysis. NeuroImage, 2012; 61: 1402-1418 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35).Schmidt P, Gaser C, Arsic M, Buck D, Förschler A, Berthele A, Hoshi M, Ilg R, Schmid V J, Zimmer C, Hemmer B, Mühlau M: An Automated Tool for Detection of FLAIR-Hyperintense White-Matter Lesions in Multiple Sclerosis. NeuroImage, 2012; 59: 3774-3783 [DOI] [PubMed] [Google Scholar]
- 36).Masana Y, Motozaki T: Emergence and Progress of White Matter Lesion in Brain Check-Up. Acta Neurol. Scand., 2003; 107: 187-194 [DOI] [PubMed] [Google Scholar]
- 37).Peters A: The Effects of Normal Aging on Myelin and Nerve Fibers: A Review. J. Neurocytol., 2002; 31: 581-593 [DOI] [PubMed] [Google Scholar]
- 38).Sandell J H, Peters A: Disrupted Myelin and Axon Loss in the Anterior Commissure of the Aged Rhesus Monkey. J. Comp. Neurol., 2003; 466: 14-30 [DOI] [PubMed] [Google Scholar]
- 39).Verhaaren B F J, Vernooij M W, de Boer R, Hofman A, Niessen W J, van der Lugt A, Ikram M A: High Blood Pressure and Cerebral White Matter Lesion Progression in the General Population. Hypertens. Dallas Tex 1979, 2013; 61: 1354-1359 [DOI] [PubMed] [Google Scholar]
- 40).Godin O, Tzourio C, Maillard P, Mazoyer B, Dufouil C: Antihypertensive Treatment and Change in Blood Pressure Are Associated with the Progression of White Matter Lesion Volumes: The Three-City (3C)-Dijon Magnetic Resonance Imaging Study. Circulation, 2011; 123: 266-273 [DOI] [PubMed] [Google Scholar]
- 41).SPRINT MIND Investigators for the SPRINT Research Group, Nasrallah I M, Pajewski N M, Auchus A P, Chelune G, Cheung A K, Cleveland M L, Coker L H, Crowe M G, Cushman W C, Cutler J A, Davatzikos C, Desiderio L, Doshi J, Erus G, Fine L J, Gaussoin S A, Harris D, Johnson K C, Kimmel P L, Kurella Tamura M, Launer L J, Lerner A J, Lewis C E, Martindale-Adams J, Moy C S, Nichols L O, Oparil S, Ogrocki P K, Rahman M, Rapp S R, Reboussin D M, Rocco M V, Sachs B C, Sink K M, Still C H, Supiano M A, Snyder J K, Wadley V G, Walker J, Weiner D E, Whelton P K, Wilson V M, Woolard N, Wright J T, Wright C B, Williamson J D, Bryan R N: Association of Intensive vs Standard Blood Pressure Control With Cerebral White Matter Lesions. JAMA, 2019; 322: 524-534 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42).Huang W-Q, Lin Q, Tzeng C-M: Leukoaraiosis: Epidemiology, Imaging, Risk Factors, and Management of Age-Related Cerebral White Matter Hyperintensities. J. Stroke, 2024; 26: 131-163 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43).Han F, Zhai F-F, Wang Q, Zhou L-X, Ni J, Yao M, Li M-L, Zhang S-Y, Cui L-Y, Jin Z-Y, Zhu Y-C: Prevalence and Risk Factors of Cerebral Small Vessel Disease in a Chinese Population-Based Sample. J. Stroke, 2018; 20: 239-246 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44).Ke D, Zhou F, Liang H, Xu Y, Lou H: Hypertriglyceridemia Is Associated with Reduced Leukoaraiosis Severity in Patients with a Small Vessel Stroke. Behav. Neurol., 2018; 2018: 1361780 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45).Ohwaki K, Yano E, Tamura A, Inoue T, Saito I: Hypercholesterolemia Is Associated with a Lower Risk of Cerebral Ischemic Small Vessel Disease Detected on Brain Checkups. Clin. Neurol. Neurosurg., 2013; 115: 669-672 [DOI] [PubMed] [Google Scholar]
- 46).Longstreth W T, Arnold A M, Beauchamp N J, Manolio T A, Lefkowitz D, Jungreis C, Hirsch C H, O’Leary D H, Furberg C D: Incidence, Manifestations, and Predictors of Worsening White Matter on Serial Cranial Magnetic Resonance Imaging in the Elderly: The Cardiovascular Health Study. Stroke, 2005; 36: 56-61 [DOI] [PubMed] [Google Scholar]
- 47).Sudre C H, Smith L, Atkinson D, Chaturvedi N, Ourselin S, Barkhof F, Hughes A D, Jäger H R, Cardoso M J: Cardiovascular Risk Factors and White Matter Hyperintensities: Difference in Susceptibility in South Asians Compared With Europeans. J. Am. Heart Assoc., 2018; 7: e010533 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48).Hilal S, Mok V, Youn Y C, Wong A, Ikram M K, Chen C L-H: Prevalence, Risk Factors and Consequences of Cerebral Small Vessel Diseases: Data from Three Asian Countries. J. Neurol. Neurosurg. Psychiatry, 2017; 88: 669-674 [DOI] [PubMed] [Google Scholar]
- 49).Habes M, Erus G, Toledo J B, Zhang T, Bryan N, Launer L J, Rosseel Y, Janowitz D, Doshi J, Van der Auwera S, von Sarnowski B, Hegenscheid K, Hosten N, Homuth G, Völzke H, Schminke U, Hoffmann W, Grabe H J, Davatzikos C: White Matter Hyperintensities and Imaging Patterns of Brain Ageing in the General Population. Brain J. Neurol., 2016; 139: 1164-1179 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50).Caligiuri M E, Perrotta P, Augimeri A, Rocca F, Quattrone A, Cherubini A: Automatic Detection of White Matter Hyperintensities in Healthy Aging and Pathology Using Magnetic Resonance Imaging: A Review. Neuroinformatics, 2015; 13: 261-276 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51).Dadar M, Maranzano J, Ducharme S, Carmichael O T, Decarli C, Collins D L, Alzheimer’s Disease Neuroimaging Initiative: Validation of T1w-Based Segmentations of White Matter Hyperintensity Volumes in Large-Scale Datasets of Aging. Hum. Brain Mapp., 2018; 39: 1093-1110 [DOI] [PMC free article] [PubMed] [Google Scholar]