2025 Volume 32 Issue 10 Pages 1304-1315
Aim: We investigated the association of obesity and metabolic health status with cerebral small-vessel disease (SVD), a predictor of stroke, in stroke-free participants during brain health checkups.
Methods: An observational cross-sectional study was conducted on 6,088 stroke-free participants who underwent brain magnetic resonance imaging (MRI). Abdominal obesity was defined as a waist circumference ≥ 90 cm for men and ≥ 80 cm for women. A metabolically healthy status was defined as having none of the three components of metabolic syndrome, except abdominal obesity. The total SVD scores were derived from four MRI markers: silent lacunar infarcts, cerebral microbleeds, moderate-to-severe white-matter hyperintensity, and enlarged perivascular spaces.
Results: The mean age of participants was 55±12 years old. Obesity was prevalent in 50% of the patients. The prevalence of a total SVD score ≥ 2 (moderate-to-severe SVD) was 348 (6%), which was elevated in metabolically unhealthy individuals regardless of obesity status. Compared with the metabolically healthy non-obese group, the metabolically unhealthy non-obese (odds ratio [OR] 2.08, [95% confidence interval {CI}, 1.33–3.27]) and metabolically unhealthy obese (OR 2.62, [95% CI, 1.70–4.04]) groups had a higher multivariable-adjusted risk for a total SVD score ≥ 2. Similar results were obtained for obesity defined as a body mass index ≥ 25 kg/m2 instead of abdominal obesity.
Conclusions: Abdominal and general obesity alone were not associated with high total SVD scores in stroke-free individuals. Metabolically unhealthy status, especially high blood pressure and hyperglycemia, are significant risk factors for moderate-to-severe SVD.
Cerebral small-vessel disease (SVD) is a subclinical disorder of the cerebral microcirculation that manifests as white matter hyperintensities (WMHs), cerebral microbleeds (CMBs), silent lacunar infarcts, and enlarged perivascular spaces (PVSs) on brain magnetic resonance imaging (MRI)1). The incidence of SVD continues to increase because of the aging population2). SVD is a predictive factor for symptomatic stroke, cognitive decline, dementia, depression, age-related disabilities, and increased mortality3, 4).
Okinawa is known as a “Blue Zone,” and residents of Okinawa Prefecture have the longest life expectancy in the world, partly due to their traditional diet5-7). The traditional Okinawa diet is low in calories and fat and high in carbohydrates, vegetables, and soy products. After World War II, Okinawa was occupied and governed by the United States from 1945 until being returned to Japan in 1972. The rapid nutrition transition to a Western-style diet characterized by high saturated fat, protein, and processed foods has been widely adopted by Okinawa’s post-war generation. Combined with reduced physical activity, this shift has led to an increased incidence of obesity and its metabolic sequelae6). A similar nutritional transition has occurred worldwide, especially in East Asia (Japan, China, and South Korea) and Africa, leading to a global obesity epidemic8).
Obesity is associated with an increased risk of cardiovascular diseases, including stroke8-10), and metabolic disorders, such as high blood pressure (BP), insulin resistance, diabetes, and dyslipidemia. However, a subset of obese individuals termed metabolically healthy obese (MHO), who do not exhibit overt cardiometabolic abnormalities, has a prevalence ranging between one-sixth and two-thirds of the obese population11, 12). In addition, whether or not obesity or obesity-related metabolic abnormalities are associated with stroke remains unclear13-15).
Metabolically healthy obesity often represents a transitional stage to an unhealthy phenotype, eventually resulting in metabolically unhealthy obesity, which is associated with an increased risk of developing cardiovascular disease11). However, whether or not the risk of stroke increases over time in patients with persistent MHO is unclear.
Previous prospective studies have demonstrated that MHO at baseline was not associated with an increased risk of ischemic stroke13, 16), whereas women who remained MHO during a follow-up period of 24 years had a higher risk of stroke than those with stable metabolic healthy non-obese individuals17).
We hypothesized that subclinical SVD observed on brain MRI, an early predictor of stroke, is more prevalent in individuals with MHO than in metabolically healthy non-obese individuals. In this study, we aimed to investigate the combined impact of obesity and metabolic abnormalities on SVD in stroke-free individuals.
This observational cross-sectional study included participants recruited from a brain health checkup program (Brain Dock) conducted among the general population by the Okinawa Health Promotion Foundation. A total of 8,187 participants underwent brain MRI between April 2013 and March 2020. Some individuals (n = 1,992) participated multiple times during this period; however, only their initial data points were analyzed. Participants with a history of stroke (n = 107) were excluded from this study. Ultimately, 6,088 stroke-free participants were enrolled in the study (Fig.1).
MRI, magnetic resonance imaging
This study was conducted in accordance with the revised Declaration of Helsinki and approved by the Ethics Committee of the University of the Ryukyus (#1674). Individual data were anonymized and routinely collected for brain docking, posing no risk to participants. Thus, the requirement for individual informed consent was waived, and an opt-out method was employed to obtain informed consent in this study.
Data CollectionTrained staff members performed the clinical examinations and blood sampling. Body height and weight were measured with the participants wearing light clothing without shoes. Body mass index (BMI) was calculated as the weight in kilograms divided by the height in meters squared. Waist circumference (WC) was measured at the umbilical level with participants in the standing position. After resting for a few minutes in a seated position with the legs uncrossed, BP was measured using a validated oscillometric device (HBP-1600; Omron Healthcare Co., Ltd., Kyoto, Japan) with an appropriately sized cuff. The average of two BP readings was used in subsequent analyses. Participants’ medical histories were collected using self-administered questionnaires and confirmed through interviews with the physicians. Blood samples were collected after an overnight fast. The estimated glomerular filtration rate (eGFR) was calculated using the Japanese Society of Nephrology equation18). Current smoking status (never, former, or current) and alcohol intake (currently habitual or not) were determined through interviews. Abdominal obesity was defined as WC ≥ 90 cm for men and ≥ 80 cm for women, according to modified Japanese criteria19). The elevated BP component of metabolic syndrome (MetS) (MetS-BP) was defined as systolic BP ≥ 130 mmHg, diastolic BP ≥ 85 mmHg, or the current use of antihypertensive agents. The hyperglycemic component of MetS (MetS-Glu) was defined as a fasting blood glucose level ≥ 110 mg/dL and/or the current use of antidiabetic agents. The dyslipidemia component of MetS (MetS-Lipid) was defined as triglyceride levels ≥ 150 mg/dL and/or high-density lipoprotein cholesterol ＜40 mg/dL. MetS was defined as the presence of abdominal obesity plus any two of MetS-BP, MetS-Glu, or MetS-Lipid20). A metabolically healthy status was defined as strictly having none of the three components of MetS, except abdominal obesity12). Conversely, a metabolically unhealthy status was defined as the presence of one or more of the three components of MetS (excluding abdominal obesity). We did not use lipid-lowering agents for the diagnosis of MetS-Lipid, as information on the breakdown of agents for dyslipidemia was not available.
MRI AssessmentMRI was conducted using 3.0-T (Magnetom Spectra; Siemens AG, Erlangen, Germany) superconducting magnets including T1-weighted imaging (T1W1; to repeat [TR] = 600 ms; time to echo [TE] = 9.5 ms), T2-weighted imaging (T2WI; TR = 5,000 ms; TE = 96 ms), fluid attenuated inversion recovery (FLAIR) imaging (TR = 12,000 ms; TE = 83 ms), time-of-flight images of the intracranial artery, and susceptibility-weighted imaging (SWI; TR = 28 ms; TE = 20 ms)21). A lacunar infarct was defined as a small subcortical infarct, 3–15 mm in diameter, with hyperintensity on T2WI and hypointensity on FLAIR imaging. CMBs were defined as small (2–5 mm diameter) areas of signal voids with associated blooming, as observed on susceptibility-weighted imaging (SWI). WMH is a descriptive term for an area of increased signal intensity in the cerebral white matter, as detected using T2WI or FLAIR imaging. These areas can be divided based on the location of the WMH into lateral PVHs and deep and subcortical WMHs (DSWMHs). The WMHs were graded according to the 2008 Japanese Brain Dock guidelines22). This classification is similar to that of the Fazekas Scale. PVHs were graded as follows: grade 0, absent or periventricular rims only; grade 1, localized lesions, such as periventricular caps; grade 2, lesions extending along the entire periventricular area; grade 3, irregular periventricular hyperintensity extending into the deep white matter; and grade 4, periventricular hyperintensity extending throughout the deep and subcortical white matter. PVH grades 3 and 4 correspond to Fazekas grade 3. DSWMHs were graded as follows: grade 0, absent; grade 1, punctate foci ＜3 mm in diameter or extended periventricular foci; grade 2, punctate or discrete foci ≥ 3 mm in diameter in the subcortical and deep white matter; grade 3, confluent foci with indistinct boundaries in the subcortical and deep white matter; and grade 4, widely distributed confluence in most of the white matter. DSWMH grade 2 corresponds to Fazekas grade 2, whereas DSWMH grades 3 and 4 correspond to Fazekas grade 3. We defined WMHs as PVHs ≥ 3 (equivalent to Fazekas grade 3) and/or DSWMH grade ≥ 2 (equivalent to Fazekas grade ≥ 2) as these are considered pathological conditions. If the WMH scores differed between the left and right cerebral hemispheres, the higher score was used in the analysis. An enlarged PVS is a point or line ＜3 mm in diameter that is hyperintense on T2WI and hypointense on T1WI and FLAIR images. Enlarged PVS was categorized as grades 0–3 based on the number of lesions (i.e., grade 0, absent; grade 1, 1–5; grade 2, 6–10; grade 3, ≥ 11). WMHs, lacunar infarcts, CMB, and enlarged PVS on MRI were defined as SVD. Imaging analyses were performed by two physicians who were blinded to the participants’ clinical information. One radiologist rated the images and one neurologist/neurosurgeon reviewed the images and made the final decision. The total SVD score was calculated on an ordinal scale from 0 to 4 by allocating one point for the presence of each of the following: lacunar infarct, CMB, enlarged PVS (≥ 11), and moderate-to-severe WMH (PVH ≥ 3 and/or DSWMH ≥ 2).
Statistical AnalysesVariables were categorized as parametric or nonparametric based on their distribution. Continuous variables with normal distribution are presented as mean (±standard deviation). Non-normally distributed variables were presented as medians (interquartile ranges). Categorical variables were presented as absolute numbers (percentages). Participants were categorized into the following four groups according to their abdominal obesity and metabolic health status: metabolically healthy non-obese (MHN), metabolically unhealthy non-obese (MUN), MHO, or metabolically unhealthy obese (MUO). Differences in the mean values between groups were tested using an analysis of variance and post-hoc Tukey-Kramer honest significant difference tests. Non-normally distributed variables were compared using the Kruskal-Wallis test. The chi-squared test was used to compare categorical variables. Multivariate logistic regression analyses were used to examine the effects of abdominal obesity and metabolic health status on moderate-to-severe total SVD scores (≥ 2) after adjusting for the age, sex, current smoking status, alcohol intake, history of heart disease, low-density lipoprotein cholesterol level, and use of lipid-lowering agents. Variance inflation factors were computed to evaluate multicollinearity among the independent variables. All values were below 2.0, indicating no multicollinearity between variables. Furthermore, BP, fasting blood glucose, triglycerides, high-density lipoprotein cholesterol, and the use of any agent for hypertension or diabetes were not included as independent variables because they had already been incorporated into the definition of metabolic health status.
Potential statistical interactions between abdominal obesity and a metabolically unhealthy status were tested by entering these variables as interaction terms in a multivariate logistic regression model (adjusted for the covariates noted above). We then evaluated the associations between the 4 groups defined above and a total SVD score of ≥ 2 and each SVD component using a multivariable-adjusted model. Given that the prevalence of enlarged PVS (≥ 11) was very low, a multivariate logistic regression analysis for enlarged PVS was not performed. In the sensitivity analysis, we defined abdominal obesity as a WC ≥ 85 cm for men and ≥ 90 cm for women, as proposed by the Japan Society for Obesity20). We also evaluated BMI (≥ 25 kg/m2)-defined obesity rather than abdominal obesity. Odds ratios (ORs) are presented with 95% confidence intervals (CIs). P＜0.05 was considered to indicate statistical significance. All analyses were performed using the JMP Pro software program, ver. 15 (SAS Institute, Cary, NC, USA).
Among the 6,088 participants, the prevalence of abdominal obesity was 50%, and 59% were women. MHO participants accounted for 33% of the population with abdominal obesity. MUN participants accounted for 48% of the non-obese population. The most represented group was MUO (34%), followed by MHN (26%), MUN (24%), and MHO (16%) (Table 1). Compared to the participants in the metabolically healthy groups (MHN and MHO), those in the metabolically unhealthy groups (MUN and MUO) were older and had higher systolic and diastolic BP, pulse pressure, fasting blood glucose, triglycerides, lower high-density lipoprotein-cholesterol, and eGFR values. Furthermore, the prevalence of SVD was significantly higher in the two metabolically unhealthy groups than in the two metabolically healthy groups (Table 2). The prevalence of a total SVD score ≥ 2 was highest in participants with MUO (9.0%), followed by those with MUN (7.9%), MHO (2.1%), and MHN (1.7%). A total of 1,929 (32%) participants were overweight (25 ≤ BMI ＜30 kg/m2), and 505 (8%) were obese (BMI ≥ 30 kg/m2). When BMI (≥ 25 kg/m2)-defined obesity was used for the MHO definition instead of abdominal obesity, the overall prevalence of obesity was 40%, and MHO was found in 25% of the obese population.
| Variables |
MHN (n = 1,573, 26%) |
MUN (n = 1,476, 24%) |
MHO (n = 997, 16%) |
MUO (n = 2,042, 34%) |
P |
|---|---|---|---|---|---|
| Age, y | 49.1 (11.5) | 57.4 (12.0)＊ | 52.5 (11.6)＊† | 58.0 (11.8)＊‡ | ＜0.001 |
| Women, n (%) | 761 (49.7) | 242 (16.4) | 770 (77.2) | 1008 (49.4) | ＜0.001 |
| Body mass index, kg/m2 | 21.3 (2.3) | 23.0 (2.3) | 25.2 (3.1) | 27.5 (3.9) | ＜0.001 |
| Waist circumference, cm | 77.1 (6.2) | 82.2 (5.5) | 89.3 (7.1) | 94.6 (8.7) | ＜0.001 |
| Current smoking, n (%) | 249 (15.8) | 296 (20.1) | 88 (8.8) | 295 (14.5) | ＜0.001 |
| Alcohol intake, n (%) | 282 (17.9) | 458 (31.0) | 133 (13.3) | 340 (16.7) | ＜0.001 |
| Systolic BP, mmHg | 109.5 (10.6) | 126.6 (14.4)＊ | 111.7 (10.6)† | 128.2 (14.5)＊†‡ | ＜0.001 |
| Diastolic BP, mmHg | 68.5 (8.0) | 78.7 (10.2)＊ | 69.4 (7.9)† | 78.9 (9.9)＊‡ | ＜0.001 |
| Pulse pressure, mmHg | 41.0 (7.3) | 47.9 (10.9)＊ | 42.3 (7.6)＊† | 49.3 (11.0)＊†‡ | ＜0.001 |
| Heart rate, bpm | 60.8 (8.3) | 62.9 (9.6)＊ | 62.9 (7.9)＊ | 64.8 (9.2)＊†‡ | ＜0.001 |
| Fasting blood glucose, mg/dL | 92 (88–97) | 99 (93–108)＊ | 94 (90–99)＊† | 101 (94–112)＊†‡ | ＜0.001 |
| LDL-cholesterol, mg/dL | 112.5 (28.6) | 116.4 (30.2)＊ | 123.2 (28.7)＊† | 124.7 (29.9)＊† | ＜0.001 |
| HDL-cholesterol, mg/dL | 66.5 (14.5) | 58.4 (16.1)＊ | 62.9 (13.3)＊† | 54.8 (13.7)＊†‡ | ＜0.001 |
| Triglycerides, mg/dL | 72 (56–97) | 115 (77–175)＊ | 86 (66–110)＊† | 130 (91–183)＊†‡ | ＜0.001 |
| eGFR, ml/min per 1.73m2 | 74.8 (13.0) | 70.7 (13.7)＊ | 73.9 (12.9)† | 70.8 (14.3)＊‡ | ＜0.001 |
| C-reactive protein, mg/dL | 0.05 (0.03–0.09) | 0.07 (0.04–0.14)＊ | 0.09 (0.05–0.16)＊† | 0.12 (0.07–0.23)＊†‡ | ＜0.001 |
| History of heart disease, n (%) | 35 (2.2) | 91 (6.2) | 30 (3.0) | 137 (6.7) | ＜0.001 |
| Antihypertensive agent, n (%) | 0 (0) | 537 (36.4) | 0 (0) | 928 (45.5) | ＜0.001 |
| Anti-diabetic agent, n (%) | 0 (0) | 123 (8.3) | 0 (0) | 226 (11.1) | ＜0.001 |
| Lipid-lowering agent, n (%) | 76 (4.8) | 272 (18.4) | 73 (7.3) | 477 (23.4) | ＜0.001 |
Values are expressed as mean (SD), median (interquartile range), or number (%). BP indicates blood pressure; eGFR, estimated glomerular filtration rate; HDL, high-density lipoprotein; LDL, low-density lipoprotein; MHN, metabolically healthy non-obese; MUN, metabolically unhealthy non-obese; MHO, metabolically healthy obese; MUO, metabolically unhealthy obese. ＊Significant vs MHN. †Significant vs MUN. ‡Significant vs MHO.
| MHN | MUN | MHO | MUO | P | |
|---|---|---|---|---|---|
| SVD, % | 310 (20) | 612 (41) | 257 (26) | 861 (42) | ＜0.001 |
| Cerebral microbleeds, % | 29 (1.8) | 73 (5.0) | 15 (1.5) | 114 (5.6) | ＜0.001 |
| Silent lacunar infarcts, % | 21 (1.3) | 116 (7.9) | 13 (1.3) | 122 (6.0) | ＜0.001 |
| White matter hyperintensities, % | 281 (18) | 536 (36) | 244 (25) | 800 (39) | ＜0.001 |
| Enlarged perivascular spaces, % | 11 (0.7) | 16 (1.1) | 6 (0.6) | 20 (1.0) | 0.49 |
| Total SVD score, % | |||||
| Score 0 | 1,268 (81) | 870 (59) | 743 (75) | 1,192 (58) | ＜0.001 |
| Score 1 | 278 (18) | 489 (33) | 233 (23) | 667 (33) | ＜0.001 |
| Score 2 | 23 (1.5) | 90 (6.1) | 20 (2.0) | 148 (7.3) | ＜0.001 |
| Score 3 | 4 (0.3) | 24 (1.6) | 1 (0.1) | 29 (1.4) | ＜0.001 |
| Score 4 | 0 (0) | 3 (0.2) | 0 (0) | 6 (0.3) | 0.017 |
Values are expressed as number (%). MHN, metabolically healthy non-obese; MUN, metabolically unhealthy non-obese; MHO, metabolically healthy obese; MUO, metabolically unhealthy obese; SVD, cerebral small vessel disease.
A multivariate logistic analysis demonstrated that the OR for a total SVD score ≥ 2 was significantly increased in participants with a metabolically unhealthy status (OR, 2.37 [95% CI, 1.70–3.30]) but not in those with abdominal obesity (OR, 1.21 [95% CI, 0.93–1.56]). For a total SVD score ≥ 2, no interaction was identified between a metabolically unhealthy status and the abdominal obesity status (P = 0.43). The OR for a total SVD score ≥ 2 was significantly higher in the MUN (OR, 2.08 [95% CI, 1.33–3.27]) and MUO groups (OR, 2.62 [95% CI, 1.70–4.04]) than in the MHN group (Table 3). Similar results were obtained for the multivariate adjusted model including C-reactive protein (CRP) as an independent variable (Supplementary Table 1). A sensitivity analysis using the original Japanese criteria for abdominal obesity demonstrated similar results to those of multivariate logistic analysis (Supplementary Table 2, model 1). Another sensitivity analysis using BMI (≥ 25 kg/m2)-defined obesity demonstrated similar results (Supplementary Table 2, model 2). Multivariate logistic analyses for each SVD component, including WMHs, silent lacunar infarcts, and CMB as dependent variables, demonstrated similar results to those with a total SVD score of ≥ 2 as the dependent variable (Supplementary Table 3).
| Variables | OR (95% CI) | P |
|---|---|---|
| Age, y | 1.12 (1.10–1.13) | ＜0.001 |
| Women | 0.86 (0.66–1.14 | 0.29 |
| LDL-cholesterol, mg/dL | 1.00 (0.99–1.00) | 0.054 |
| Current smoking | 1.20 (0.81–1.79) | 0.37 |
| Alcohol intake | 1.14 (0.84–1.57) | 0.40 |
| History of cardiac disease | 0.99 (0.66–1.47) | 0.94 |
| Lipid-lowering agent | 0.98 (0.74–1.29) | 0.87 |
| MHN | 1 (reference) | - |
| MUN | 2.08 (1.33–3.27) | 0.001 |
| MHO | 0.97 (0.53–1.77) | 0.92 |
| MUO | 2.62 (1.70–4.04) | ＜0.001 |
CI indicates confidence interval; LDL, low-density lipoprotein; MHN, metabolically healthy non-obese; MUN, metabolically unhealthy non- obese; MHO, metabolically healthy obese; MUO, metabolically unhealthy obese; OR, odds ratio.
| Variables | OR (95% CI) | P |
|---|---|---|
| Age, y | 1.11 (1.10–1.13) | ＜0.001 |
| Women | 0.89 (0.67–1.18) | 0.40 |
| LDL-cholesterol, mg/dl | 1.00 (0.99–1.00) | 0.062 |
| Current smoking | 1.24 (0.82–1.88) | 0.31 |
| Alcohol intake | 1.10 (0.79–1.53) | 0.57 |
| History of cardiac disease | 1.01 (0.66–1.53) | 0.97 |
| Lipid-lowering agent | 1.01 (0.76–1.36) | 0.92 |
| C-reactive protein, mg/dl | 1.07 (0.85–1.35) | 0.56 |
| MHN | 1 (reference) | - |
| MUN | 2.00 (1.26–3.18) | 0.003 |
| MHO | 1.01 (0.55–1.85) | 0.97 |
| MUO | 2.51 (1.62–3.91) | ＜0.001 |
Adjusted for age, sex, current smoking, alcohol intake, history of heart disease, low-density lipoprotein cholesterol, lipid-lowering agent use, and C-reactive protein level. The total sample size was 5,358 because of missing C-reactive protein data.
LDL indicates low-density lipoprotein; MHN, metabolically healthy non-obese; MUN, metabolically unhealthy non-obese; MHO, metabolically healthy obese; MUO, metabolically unhealthy obese.
| Model 1 | Model 2 | |||
|---|---|---|---|---|
| OR (95% CI) | P | OR (95% CI) | P | |
| MHN | 1 (reference) | - | 1.0 (reference) | - |
| MUN | 2.40 (1.57–3.67) | ＜0.001 | 2.04 (1.37–3.03) | ＜0.001 |
| MHO | 1.39 (0.75–2.56) | 0.30 | 1.00 (0.51–1.96) | 0.99 |
| MUO | 3.09 (2.03–4.71) | ＜0.001 | 2.92 (1.98–4.30) | ＜0.001 |
Adjusted for age, sex, current smoking, alcohol intake, history of heart disease, low-density lipoprotein cholesterol, and lipid-lowering agent use.
CI indicates confidence interval; MHN, metabolically healthy non-obese; MUN, metabolically unhealthy non-obese; MHO, metabolically healthy obese; MUO, metabolically unhealthy obese; OR, odds ratio.
Model 1: Obesity was defined as a waist circumference ≥ 85 cm for men and ≥ 90 cm for women. Model 2: Obesity was defined as body mass index ≥ 25.0 kg/m2 instead of abdominal obesity.
| WMH OR (95% CI) | Silent lacunar infarcts OR (95% CI) | CMB OR (95% CI) | |
|---|---|---|---|
| MHN | 1.00 | 1.00 | 1.00 |
| MUN | 1.53 (1.26–1.86) | 3.01 (1.84–4.92) | 1.43 (0.90–2.27) |
| MHO | 1.05 (0.84–1.32) | 0.80 (0.39–1.62) | 0.74 (0.39–1.41) |
| MUO | 1.50 (1.25–1.80) | 2.28 (1.40–3.72) | 1.86 (1.21–2.87) |
Adjusted for age, sex, current smoking, alcohol intake, history of heart disease, low-density lipoprotein cholesterol, and lipid-lowering agent use. CMB, cerebral microbleeds; MHN, metabolically healthy non-obese; MUN, metabolically unhealthy non-obese; MHO, metabolically healthy obese; MUO, metabolically unhealthy obese; SVD, cerebral small vessel disease; WMH, white matter hyperintensities.
A total of 869 (14.2%) participants were diagnosed with MetS according to the modified Japanese criteria and 958 (15.7%) participants had MetS according to the original Japanese criteria. The OR for a total SVD score ≥ 2 was significantly high in participants with MetS than non-MetS (OR, 1.73 [95% CI, 1.32–2.28]) (Table 4, model 1). When MetS components were entered separately as independent variables, MetS-BP (OR, 2.92 [95% CI, 2.18–3.91]) and MetS-Glu (OR, 1.51 [95% CI, 1.16–1.97]) had significantly high ORs for a total SVD score ≥ 2 than non-MetS-BP and non-MetS-Glu, respectively. However, abdominal obesity (OR, 1.10 [95% CI, 0.84–1.44]) and MetS-Lipid (OR, 0.93 [95% CI, 0.70–1.24]) did not exhibit significant associations (Table 4, model 2). Similar results were obtained in a multivariate logistic analysis using the original Japanese criteria for abdominal obesity (Supplementary Table 4).
| Model 1 | Model 2 | |||
|---|---|---|---|---|
| Variables | OR (95% CI) | P | OR (95% CI) | P |
| MetS | 1.73 (1.32–2.28) | ＜0.001 | - | - |
| Abdominal obesity | - | - | 1.10 (0.84–1.44) | 0.48 |
| MetS-BP | - | - | 2.92 (2.18–3.91) | ＜0.001 |
| MetS-Glu | - | - | 1.51 (1.16–1.97) | 0.002 |
| MetS-Lipid | - | - | 0.93 (0.70–1.24) | 0.61 |
Adjusted for age, sex, current smoking, alcohol intake, history of heart disease, low-density lipoprotein cholesterol, and lipid-lowering agent use. MetS indicates metabolic syndrome; MetS-BP, high blood pressure component of MetS; MetS-Glu, hyperglycemic component of MetS; MetS- Lipid, dyslipidemia component of MetS.
| Model 1 | Model 2 | |||
|---|---|---|---|---|
| OR (95% CI) | P | OR (95% CI) | P | |
| MetS | 1.81 (1.38–2.38) | ＜0.001 | - | - |
| Abdominal obesity | - | - | 1.19 (0.92–1.53) | 0.189 |
| MetS-BP | - | - | 2.89 (2.16–3.86) | ＜0.001 |
| MetS-Glu | - | - | 1.49 (1.14–1.95) | 0.004 |
| MetS-Lipid | - | - | 0.92 (0.69–1.22) | 0.56 |
Adjusted for age, sex, current smoking, alcohol intake, history of heart disease, low-density lipoprotein cholesterol, and lipid-lowering agent use. MetS indicates metabolic syndrome; MetS-BP, high blood pressure component of MetS; MetS-Glu, hyperglycemic component of MetS; MetS- Lipid, dyslipidemia component of MetS. Abdominal obesity was defined as a waist circumference ≥ 85 cm for men and ≥ 90 cm for women.
In this cross-sectional observational study, MHO was not associated with a high prevalence of moderate to severe SVD in stroke-free individuals. Conversely, a metabolically unhealthy status (defined as having one or more of the three MetS components, except for abdominal obesity) was associated with subclinical structural abnormalities of the brain, regardless of obesity status. A metabolically unhealthy status, especially a high BP and hyperglycemia, was associated with moderate-to-severe SVD, regardless of abdominal obesity.
The absence of an established definition of MHO complicates the comparison of existing data to understand its prevalence and clinical outcomes23). Many previous studies have defined metabolic health status as not meeting any of the MetS criteria or having only one MetS criterion excluding abdominal obesity11, 24). We found that MetS-BP and MetS-Glu were independently associated with total SVD scores ≥ 2; therefore, participants with any component of MetS (except abdominal obesity) were not considered metabolically healthy. Consistent with this, individuals with at least one MetS component have poor outcomes in terms of cardiovascular disease events, including coronary heart disease and cerebrovascular disease24). Therefore, we believe that a strict definition should be used to indicate a metabolically healthy status.
BMI ≥ 30 kg/m2 or ≥ 25 kg/m2 is the most commonly used measure of obesity, depending on ethnicity. However, the BMI cannot distinguish between fat and muscle mass and does not consider body fat distribution, while the WC is a more accurate indicator of visceral fat volume25, 26). The WC has been reported to be causally associated with cardiometabolic risk independent of the BMI. It can better predict all-cause and cardiovascular disease mortality than the BMI in both women and men11). In a retrospective observational study using the Japanese nationwide epidemiological database, MHO, defined by WC-based obesity but not by BMI-based obesity, was associated with a high risk of myocardial infarction and heart failure during three years of follow-up27). However, stroke was not found to be linked to MHO as determined by the WC or BMI. These results suggest that the impact of MHO on organ damage and cardiovascular disease outcomes may vary depending on the organ involved and clinical outcomes.
MHO often reflects a transitional stage to an unhealthy phenotype, resulting in MUO11, 23). Approximately one-third to half of individuals with MHO develop MUO, which is associated with an increased risk of incident cardiovascular diseases11, 28, 29). It remains unclear whether silent SVD, a predictor of symptomatic stroke and dementia in a healthy asymptomatic population, is more prevalent in individuals with MHO than MHN. In a cross-sectional study of the Japanese population, MHO (defined as BMI-based obesity with strict metabolic criteria) and metabolically unhealthy individuals (MUN and MUO) had a higher risk of leukoaraiosis than the risk observed in individuals with MHN, as assessed using MRI30). However, the Framingham Heart Study conducted in young to middle-aged adults without a history of stroke showed that the WMH volume was not significantly different between MHO (defined as BMI-based obesity with less strict criteria) and MHN31). Consistent with this study, we found that MHO was not associated with moderate-to-severe SVD in stroke-free individuals. Further prospective studies are warranted to elucidate whether SVD progressively increases in individuals with MHO compared with individuals with MHN.
The adjusted OR for a total SVD score of ≥ 2 did not differ markedly between MUN and MUO when MUN was used as the reference (OR, 0.88 [95% CI, 0.75–1.05]). No additive effects of obesity or metabolically unhealthy status on moderate-to-severe SVD were observed. In line with our findings, cardiovascular disease risk in individuals with MUN is similar to that in individuals with MUO17). As highlighted in previous studies, a metabolically unhealthy status defined by strict metabolic criteria, especially high BP and blood glucose values, is a risk factor for moderate-to-severe SVD, regardless of obesity status.
The proportion of women was lower in the MUN group but higher in the MHO group. However, the prevalence of a total SVD score of ≥ 2 was higher in participants with MUN and MUO, both in women and men, as was the case when assessed overall (Supplementary Table 5). Multivariate analysis stratified by sex showed similar results in both women and men (Supplementary Table 6). Therefore, we assumed that the results would be similar if analyzed separately for women and men. These results should be interpreted with caution because of the limited number of SVD total scores ≥ 2 for each of the four groups according to obesity and metabolic health status.
| MHN | MUN | MHO | MUO | P | |
|---|---|---|---|---|---|
| Total | 27 (1.7) | 117 (7.9) | 21 (2.1) | 183 (9.0) | ＜0.001 |
| Women | 10 (1.3) | 20 (8.3) | 14 (1.8) | 104 (10.3) | ＜0.001 |
| Men | 17 (2.2) | 97 (7.9) | 7 (3.1) | 79 (7.6) | ＜0.001 |
Values are expressed as number (%). MHN, metabolically healthy non-obese; MUN, metabolically unhealthy non-obese; MHO, metabolically healthy obese; MUO, metabolically unhealthy obese.
| Women | Men | |||
|---|---|---|---|---|
| Variables | OR (95% CI) | P | OR (95% CI) | P |
| Age, y | 1.12 (1.10–1.14) | ＜0.001 | 1.12 (1.10–1.13) | ＜0.001 |
| LDL-cholesterol, mg/dL | 1.00 (0.99–1.00) | 0.51 | 0.99 (0.99–1.00) | 0.052 |
| Current smoking | 0.49 (0.11–2.13) | 0.29 | 1.35 (0.88–2.07) | 0.164 |
| Alcohol intake | 1.18 (0.49–2.85) | 0.71 | 1.15 (0.82–1.62) | 0.41 |
| History of cardiac disease | 0.86 (0.44–1.68) | 0.66 | 1.08 (0.65–1.80) | 0.77 |
| Lipid-lowering agent | 0.90 (0.59–1.37) | 0.62 | 1.05 (0.72–1.53) | 0.81 |
| MHN | 1 (reference) | - | 1 (reference) | - |
| MUN | 2.51 (1.11–5.64) | 0.026 | 1.95 (1.13–3.37) | 0.017 |
| MHO | 0.80 (0.35–1.87) | 0.61 | 1.51 (0.60–3.80) | 0.38 |
| MUO | 2.61 (1.30–5.24) | 0.007 | 2.57 (1.48–4.47) | ＜0.001 |
CI indicates confidence interval; LDL, low-density lipoprotein; MHN, metabolically healthy non-obese; MUN, metabolically unhealthy non- obese; MHO, metabolically healthy obese; MUO, metabolically unhealthy obese; OR, odds ratio
Systemic and vascular inflammation caused by obesity promotes oxidation of low-density lipoprotein cholesterol and endothelial dysfunction, which in turn promotes atherosclerosis8). Inflammation is also associated with SVD pathogenesis1, 2). In patients with a history of cerebrovascular disease, high levels of inflammatory markers such as CRP and interleukin-6 are associated with CMBs32, 33), although one study did not report a relationship between CRP level and WMH34). We found that the MHO group had significantly higher CRP levels than the non-obese groups (MHN and MUN). However, CRP levels were lower in the MHO group than in the MUO group, as previously reported11). No significant association was observed between CRP level and moderate-to-severe SVD in this study. The effect of cholesterol depends on the type of cerebral infarction, with low-density lipoprotein cholesterol implicated in atherothrombotic cerebral infarction but not in other types of cerebral infarction35). Hypertension is a major risk factor for all cerebral infarctions, including atherothrombotic cerebral infarctions.
Study LimitationsSeveral limitations associated with the present study warrant mention. The study had a single-center retrospective cross-sectional design because a cause-effect relationship could not be determined; however, we only studied associations between the variables. This study used an observational design, and the results may have been affected by residual confounding effects. Furthermore, the limited number of individuals with moderate-to-severe SVD and each SVD component may have affected the statistical significance of the results. Because this study included only Japanese individuals, the results may not be generalizable to other ethnic groups. Ethnic differences in the association between obesity and clinical outcomes have been reported in the general population8). The role of obesity in the pathogenesis of SVD may differ among ethnic groups. Further longitudinal studies are required to examine the effects of obesity on SVD in the Asian population. In this study, the participants were self-selected and may have been more health-conscious than the general population. We did not evaluate other factors associated with SVD, such as blood-brain barrier disruption or genetic predisposition. Increasing evidence suggests that the anatomical location of CMBs varies according to their etiology. Arteriolosclerosis-associated CMBs are located in the deep cerebral area, whereas cerebral amyloid angiopathy-associated CMBs are located in the lobar areas. Due to a lack of data, it was not possible to assess the anatomical distribution of CMBs. This study lacked intra- and inter-rater reliability information concerning MRI assessments because SVD ratings are highly rater-dependent. The visual ratings of SVD can introduce potential errors, such as misclassification. Individuals with MHO tend to be physically active and closely adhere to healthy diets11, 12). We were unable to evaluate the effects of the diet on cardiorespiratory fitness.
Abdominal obesity and general obesity alone were not associated with moderate-to-severe SVD in stroke-free individuals. A metabolically unhealthy status, defined as having at least one MetS component other than abdominal obesity, especially high BP and hyperglycemia, was an important risk factor for moderate-to-severe SVD regardless of obesity.
We are grateful to all participants of the study and the staff of the Okinawa Health Promotion Foundation. We thank Ms. Mahiro Miyagi at the University of the Ryukyus for organizing the data. We would like to thank Editage (www.editage.jp) for the English language editing.
This work was partly supported by the Japan Society for the Promotion of Science (JSPS) KAKENHI, Grant Number 21K08281.
The authors declare no competing interests.
