Abstract
Aim: Irisin, an exercise-induced myokine, is a potential neurotrophic factor; however, its relationship with cerebral small vessel disease (CSVD) remains unknown. Therefore, we investigated whether serum irisin levels are associated with CSVD in healthy Japanese men.
Methods: We analyzed data from 720 men free of stroke and participated in this observational study. Serum irisin levels were measured by enzyme-linked immunosorbent assay. CSVD was assessed on deep and subcortical white matter hyperintensities (DSWMHs), periventricular hyperintensities (PVHs), lacunar infarcts (LIs), and cerebral microbleeds (CMBs) on brain magnetic resonance imaging. We calculated the total CSVD score (ranges 0–4) to express the total CSVD burden. We computed the adjusted odds ratios (ORs), with 95% confidence intervals (CIs), of the total CSVD score and individual CSVD features using logistic regression models according to the quartiles of irisin (reference: Q1).
Results: Serum irisin levels were associated with lower ORs of higher (vs. zero or lower score) total CSVD score, with the lowest risk (OR, 0.63; 95% CI, 0.41–0.97) being observed in Q3 compared to Q1 after adjustment of potential covariates. Similar results were obtained for younger adults (＜65 years). Among individual CSVD features, irisin was associated with a reduced risk of LIs in the total sample and PVHs, LIs, and CMBs in younger adults. No relationship was observed in older adults (≥ 65 years).
Conclusions: Serum irisin levels were associated with less burden of total CSVD in healthy Japanese men. Serum irisin levels were also related with a reduced risk of PVHs, LIs, and CMBs, but not DSWMHs.
Introduction
Cerebral small vessel disease (CSVD) is a common, subclinical, and progressive health issue in the general population with aging. CSVD causes severe clinical consequences, such as stroke, cognitive decline, depression, dementia, including Alzheimer’s disease, and death1-5); however, there are currently no established therapeutic measures to treat or prevent CSVD1, 4, 6, 7). Emerging evidence shows that physical activity (exercise) is neuroprotective for brain health and has become a promising and viable lifestyle intervention8-10). We recently demonstrated for the first time that step counts—a direct measure of physical activity—were associated with a reduced risk of CSVD in healthy Japanese men11). These findings may be of clinical importance for the attenuation of CSVD. However, the mechanisms by which physical activity reduces the risk of CSVD remain unclear.
Irisin, a newly discovered exercise-induced myokine secreted into the circulation mainly from the skeletal muscle, has been implicated in exercise-mediated decreases in the risk of CSVD. In 2012, a study showed that irisin ameliorated obesity and metabolic disorders by facilitating energy expenditure through the browning of beige fat cells in white adipose tissue12). Its role as a neurotrophic factor has recently been attracting increasing attention, and it was shown to exert neuroprotective effects against the pathogenesis of brain diseases, including neurodegenerative diseases such as Alzheimer’s disease and stroke8, 13-19). However, to the best of our knowledge, the relationship between irisin and the magnetic resonance imaging (MRI)-measured CSVD burden remains uninvestigated.
Therefore, the present study investigated the relationship between serum irisin levels at baseline and the total CSVD burden in a 5-year follow-up in Japanese men and whether this relationship is independent of lifestyle and cardiovascular risk factors. The relationships between serum irisin levels and individual CSVD features (deep and subcortical white matter hyperintensities [DSWMHs], periventricular hyperintensities [PVHs], lacunar infarcts [LIs], and cerebral microbleeds [CMBs]) were also examined.
Methods
We analyzed data from the Shiga Epidemiological Study of Subclinical Atherosclerosis (SESSA) in the present study. This ongoing population-based observational study conducted in Kusatsu City, Shiga Prefecture, Japan, examines factors affecting subclinical atherosclerosis. Its design and enrollment methods have already been reported20-23). The study protocol was approved by the Institutional Review Board of Shiga University of Medical Science (Otsu, Japan). Briefly, at baseline (2006–2008), 2379 general Japanese men aged 40–79 years were randomly selected based on age strata from the Kusatsu City Basic Residents’ Register. Among them, 1094 men participated in baseline medical examinations with written consent; 853 subsequently underwent the SESSA follow-up examination (2010–2014). They were reinvited for brain MRI, and 740 underwent 1.5 T MRI on average 2 years later (between 2012 and 2015). We excluded 20 participants with a history of stroke (n=19) at baseline and missing information on serum irisin levels (n=1), leaving 720 men for the final analyses.
To measure serum irisin levels and other laboratory parameters, including lipid and glucose concentrations, we collected blood samples by venipuncture in baseline clinical examinations after an overnight fast of at least 12 h. We separated serum by centrifugation and sent samples to the laboratory for routine tests. Other samples were immediately frozen at −70℃ for later analyses. In 2016, we thawed specimens and measured serum irisin levels in µg/mL at the clinical chemistry laboratory of FALCO Biosystems (Kyoto, Japan) using a commercially available enzyme-linked immunosorbent assay kit (AG-45A-0046EK-k101; Adipogen, Liestal, Switzerland). Detailed procedures were previously reported21, 22).
We assessed CSVD in participants at the follow-up (2012–2015) using the following neuroimaging markers on brain MRI: DSWMHs, PVHs, LIs, and CMBs24). MRI was performed using a 1.5-Tesla MR scanner (Signa HDxt 1.5T ver. 16; GE Healthcare, Milwaukee, WI, USA). Two neurosurgeons, certified by the Japan Neurosurgery Society, independently assessed all MRI images and rated them in duplicate. They were blinded to the clinical information of participants during their evaluations. We dichotomized each form of the four CSVD features as follows: DSWMHs (grade ≥ 3)25, 26), PVHs (grade ≥ 2)25), LIs (defined as a lesion number ≥ 1), and CMBs (defined as a lesion number ≥ 1). We also created a “total CSVD score” to express the total CSVD burden27, 28). To achieve this, we awarded 1 point to the presence of each MRI CSVD feature and summed them in an ordinal scale ranging between 0 (no outcome) and 4 (all outcomes), which indicated the severity of CSVD27). The frequency of participants with each value of the total CSVD score is presented in Supplemental Table 1. The assessment of CSVD was described in detail in previous studies11, 28).
Supplemental Table 1.
Frequency of Participants with Each Value of the Total Cerebral Small Vessel Disease Score (
n = 720)
| TCSVD score |
Overall (n = 720)
|
＜65 years (n = 386)
|
≥ 65 years (n = 334)
|
| 0 |
395 (54.9) |
257 (66.6) |
138 (41.3) |
| 1 |
158 (21.9) |
71 (18.4) |
87 (26.1) |
| 2 |
98 (13.6) |
36 (9.3) |
62 (18.6) |
| 3 |
50 (6.9) |
15 (3.9) |
35 (10.5) |
| 4 |
19 (2.6) |
7 (1.8) |
12 (3.6) |
Values are presented as numbers (percentages).
TCSVD indicates total cerebral small vessel disease.
Data on demographic characteristics, lifestyle factors, including the smoking and drinking status, medical history, and the use of medications for hypertension, diabetes, and dyslipidemia were collected from each participant using a self-administered questionnaire at the baseline. We objectively measured step counts for baseline participants on 7 consecutive days using a pedometer (DIGI-WALKER DW-200; Yamasa Tokei Keiki, Tokyo, Japan) and calculated the 7-day average step counts. Body mass index was defined as weight (kg) divided by height squared (m2). Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were measured using an automated sphygmomanometer. Fasting blood samples were obtained and centrifuged soon after collection. Total cholesterol and triglyceride concentrations were measured enzymatically and high-density lipoprotein (HDL) cholesterol concentration by a direct method. Low-density lipoprotein (LDL) cholesterol concentration was calculated using the Friedewald equation. Glycated hemoglobin concentration was measured using a latex agglutination immunoassay.
We presented descriptive statistics on participants as percentages for categorical variables and as means (standard deviations) or medians (interquartile ranges) for continuous variables. We performed one-way ANOVA (analysis of variance) to compare continuous variables with normal distribution and the Kruskal–Wallis test for continuous variables that do not follow the normal distribution among the four groups of participants based on the quartile of serum irisin levels. We used the chi-squared test for categorical variables.
We used multivariable ordinal logistic regression analyses to assess the relationship between serum irisin levels and the total CSVD burden because our outcome variable (the total CSVD score) was with five groups on an ordinal scale ranging between 0 (no outcome) and 4 (maximum outcome), and it met the proportional odds assumption. Odds ratios (ORs) were assumed to be the same at each cumulative split of the ordinal outcome variable. Therefore, there was only one OR. We categorized participants into four groups based on the quartiles of serum irisin levels (µg/mL): Q1 (2.67–5.75), Q2 (5.76–7.01), Q3 (7.02–8.33), and Q4 (8.34–24.52). The distribution of serum irisin levels is presented in Supplemental Table 2 and Supplemental Fig.1.
Supplemental Table 2.
The Distribution of Serum Irisin Levels (
n = 720): SESSA, Men Aged 40–79 Years (2006–2008)
|
|
Quartiles of serum irisin levels in μg/mL (n = 180 in each quartile)
|
|
| Parameter |
Overall (n = 720)
|
Q1 (2.67–5.75) |
Q2 (5.76–7.01) |
Q3 (7.02–8.33) |
Q4 (8.34–24.52) |
p-value
|
| Serum irisin levels, μg/mL |
7.0 (5.8–8.3) |
5.0 (4.5 –5.4) |
6.4 (6.1–6.7) |
7.6 (7.3–7.9) |
9.7 (8.9–10.9) |
＜0.001 |
Results are presented as median (interquartile range), and the p-value is calculated by the Kruskal-Wallis test.
We computed ORs and the corresponding 95% confidence intervals (CIs) for the presence of a higher total CSVD score (vs. zero or lower score) according to the quartiles of serum irisin levels using the lowest quartile (Q1) as the reference group. We ran three models in these analyses: model 1 was adjusted for age (years), model 2 for age and behavioral risk factors (smoking [pack-years] and drinking [ethanol, g/week]), and model 3 for variables in model 2 plus conventional vascular risk factors (body mass index [kg/m2] and the status of diabetes, dyslipidemia, and hypertension [yes/no]). Adjustment of step counts in the model may affect the precision of estimates and lead to over-adjustment bias because step counts could be an upstream factor in the causal pathway between serum irisin levels and atherosclerosis burden. Therefore, our analyses did not control for daily step counts. Diabetes was defined as a fasting blood glucose level ≥ 126 mg/dL or glycated hemoglobin (National Glycohemoglobin Standardization Program) value ≥ 6.5%, or the use of antidiabetic medication. Dyslipidemia was defined as a triglyceride level ≥ 150 mg/dL or LDL cholesterol level ≥ 140 mg/dL or HDL cholesterol level ≤ 40 mg/dL or the use of lipid-lowering drugs. Hypertension was defined as average SBP ≥ 140 mmHg or DBP ≥ 90 mmHg or the use of antihypertensive medication.
We used multivariable binary logistic regression analyses to assess the relationships between serum irisin levels and individual MRI CSVD features (DSWMHs, PVHs, LIs, and CMBs [yes/no]). We reproduced the same models of the ordinal logistic regression and computed adjusted ORs with CIs for the presence of each form of CSVD according to the quartiles of serum irisin levels using the lowest quartile (Q1) as the reference group.
Statistical analyses were performed using SAS software version 9.4 (SAS Institute, Cary, NC, USA). We presented results for overall and age-stratified groups (＜65 and ≥ 65 years). A p-value less than the two-sided significance level of 0.05 was significant for all analyses.
Results
Table 1 shows the participant characteristics. The mean (SD) age and educational attainment of all participants were 63.6 (9.4) and 12.7 (3.0) years, respectively. Participants in the lower two quartiles of serum irisin levels were younger, had higher education, and consumed less alcohol. Participants in the upper two quartiles had lower body mass index, a better lipid profile (higher HDL cholesterol and lower LDL cholesterol and triglycerides, whereas total cholesterol did not differ among the quartiles), and required less medication for dyslipidemia.
Table 1.
Participant Characteristics at Baseline (
n = 720): SESSA, Men Aged 40–79 Years (2006–2008)
|
|
Quartiles of serum irisin levels in μg/mL (n = 180 in each quartile)
|
|
| Characteristics |
Overall (n = 720)
|
Q1 (2.67–5.75) |
Q2 (5.76–7.01) |
Q3 (7.02–8.33) |
Q4 (8.34–24.52) |
p-value
|
| Age, y |
63.6 (9.4) |
60.4 (9.9) |
63.6 (9.2) |
64.7 (9.3) |
65.7 (8.3) |
＜0.001 |
| Years of education |
12.6 (3.0) |
12.9 (3.0) |
12.8 (3.0) |
12.6 (2.6) |
12.3 (3.1) |
0.042 |
| 7-d average step counts |
8516.4 (3540.1) |
8885.6 (3619.8) |
8761.4 (3453.0) |
8090.6 (3564.6) |
8339.6 (3496.5) |
0.246 |
| Smoking status, % |
|
|
|
|
|
0.445 |
| Current |
29.4 |
32.2 |
27.8 |
31.1 |
26.7 |
|
| Past |
51.0 |
47.8 |
50.5 |
47.8 |
57.8 |
|
| Never |
19.6 |
20 |
21.7 |
21.1 |
15.5 |
|
| Smoking, pack-years |
23.1 (4.0–42.0) |
21.0 (2.2–38.7) |
24.0 (2.5–44.3) |
23.0 (1.3–45.3) |
24.2 (8.5–42.0) |
0.308＊
|
| Drinking status, % |
|
|
|
|
|
0.596 |
| Current |
79.3 |
75.0 |
79.4 |
83.3 |
79.4 |
|
| Past |
5.1 |
6.1 |
6.1 |
3.3 |
5.0 |
|
| Never |
15.6 |
18.9 |
14.5 |
13.4 |
15.6 |
|
| Alcohol consumption (ethanol), g/wk |
98.0 (9.6–259.3) |
80.6 (0.0–235.2) |
93.3 (7.0–259.3) |
106.0 (14.0–268.8) |
153.1 (12.3–301.3) |
0.032＊
|
| Body mass index, kg/m2
|
23.6 (2.9) |
24.0 (2.7) |
23.7 (2.8) |
23.6 (3.0) |
23.1 (3.0) |
0.006 |
| HbA1c, % |
5.6 (0.8) |
5.6 (0.7) |
5.7 (0.8) |
5.7 (0.8) |
5.6 (0.7) |
0.808 |
| Medication for diabetes mellitus, % |
10.0 |
10.6 |
12.8 |
8.3 |
8.3 |
0.438 |
| Diabetes mellitus, % |
22.4 |
21.1 |
26.7 |
22.8 |
18.9 |
0.341 |
| Total cholesterol, mg/dL |
210.3 (32.5) |
213.8 (31.4) |
207.0 (36.6) |
211.2 (31.2) |
209.4 (30.2) |
0.403 |
| HDL cholesterol, mg/dL |
59.3 (17.0) |
55.2 (13.9) |
55.8 (15.8) |
60.5 (16.5) |
65.9 (19.2) |
＜0.001 |
| LDL cholesterol, mg/dL |
126.2 (30.7) |
132.1 (29.2) |
126.5 (34.4) |
124.6 (31.0) |
121.6 (27.0) |
0.001 |
| Triglycerides, mg/dL |
105.0 (76.0–150.0) |
119.5 (85.5–159.0) |
104.0 (75.5–149.5) |
104.0 (78.0–157.5) |
90.0 (71.5–128.0) |
0.001＊
|
| Medication for dyslipidemia, % |
14.4 |
15.0 |
15.6 |
12.8 |
14.4 |
0.890 |
| Dyslipidemia, % |
55.8 |
66.7 |
57.2 |
52.8 |
46.7 |
0.001 |
| Systolic blood pressure, mmHg |
135.8 (18.0) |
135.7 (17.7) |
135.6 (17.9) |
133.7 (18.1) |
137.4 (18.2) |
0.612 |
| Medication for hypertension, % |
29.2 |
27.2 |
32.2 |
26.1 |
31.1 |
0.515 |
| Hypertension, % |
53.2 |
47.2 |
56.7 |
50.0 |
58.9 |
0.088 |
All continuous variables are presented as means (SDs), except for pack-years smoking, alcohol consumption, and triglycerides (median [interquartile range]); other values are percentages. P values are calculated by one-way ANOVA (analysis of variance) for all continuous variables, unless otherwise indicated, and the chi-squared test for categorical variables. ＊The Kruskal-Wallis test. Diabetes was defined as a fasting blood glucose level ≥ 126 mg/dL or HbA1c (National Glycohemoglobin Standardization Program) value ≥ 6.5% or the use of antidiabetic medication. Dyslipidemia was defined as a triglyceride level ≥ 150 mg/dL or LDL cholesterol level ≥ 140 mg/dL or HDL cholesterol level ≤ 40 mg/dL or the use of lipid-lowering drugs. Hypertension was defined as average systolic blood pressure ≥ 140 mm Hg or diastolic blood pressure ≥ 90 mm Hg or the use of antihypertensive medication. HbA1c, glycated hemoglobin; HDL, high-density lipoprotein; LDL, low-density lipoprotein; and SESSA, Shiga Epidemiological Study of Subclinical Atherosclerosis.
Table 2 shows the multivariable-adjusted ORs, using ordinal logistic regression, for the total CSVD score by the quartiles of serum irisin levels, with Q1 as the reference group. The number and percentage of participants with each CSVD score are shown in Supplemental Table 1. Serum irisin levels were associated with lower ORs of higher (vs. zero or lower score) total CSVD score, with the lowest risk (OR, 0.63; 95% CI, 0.41–0.97) being observed at Q3 (7.02–8.33 µg/mL) relative to Q1 (2.67–5.75 µg/mL) in model 3, which was adjusted for age, behavioral factors (alcohol consumption and smoking), and vascular risk factors (body mass index, diabetes, dyslipidemia, and hypertension). Since SBP is a strong risk factor for CSVD, instead of hypertension in model 3, we also adjusted for SBP and medication use for hypertension, and the results were similar (results are not shown). Similar results were obtained for younger adults (＜65 years) as shown in Table 3.
Table 2.
Multivariable Ordinal Logistic Regression Models for Relationships (Odds Ratio with 95% CI) between Serum Irisin Levels at Baseline and the Total Cerebral Small Vessel Disease Score in the 5-year Follow-up (
n = 720)
|
Quartiles of serum irisin levels in μg/mL (n = 180 in each quartile)
|
|
Q1 (2.67–5.75) |
Q2 (5.76–7.01) |
Q3 (7.02–8.33) |
Q4 (8.34–24.52) |
| TCSVD Score |
|
|
|
|
| Model 1 |
1.00 |
0.70 (0.46–1.05) |
0.65 (0.43–0.99) |
0.71 (0.47–1.07) |
| Model 2 |
1.00 |
0.69 (0.45–1.05) |
0.63 (0.41–0.96) |
0.67 (0.44–1.02) |
| Model 3 |
1.00 |
0.66 (0.43–1.00) |
0.63 (0.41–0.97) |
0.65 (0.42–0.99) |
Model 1 was adjusted for age.
Model 2 was adjusted for variables in model 1 plus alcohol intake (ethanol) and smoking (pack-years).
Model 3 was adjusted for variables in model 2 plus body mass index, diabetes, dyslipidemia, and hypertension.
CI indicates confidence interval; and TCSVD, total cerebral small vessel disease.
Table 3.
Multivariable Ordinal Logistic Regression Models for Relationships (Odds Ratio with 95% CI) between Serum Irisin Levels at Baseline and the Total Cerebral Small Vessel Disease Score in the 5-year Follow-up (
n = 720)
|
Quartiles of serum irisin levels in μg/mL |
|
＜65 years (n = 386)
|
≥ 65 years (n = 334)
|
|
Q1 (2.67–5.74)
n = 119
|
Q2 (5.75–7.01)
n = 97
|
Q3 (7.02–8.33)
n = 88
|
Q4 (8.34–19.67)
n = 82
|
Q1 (3.69–5.75)
n = 61
|
Q2 (5.76–7.01)
n = 83
|
Q3 (7.02–8.30)
n = 92
|
Q4 (8.31–24.52)
n = 98
|
| TCSVD Score |
|
|
|
|
|
|
|
|
| Model 1 |
1.00 |
0.56 (0.31–1.00) |
0.48 (0.26–0.89) |
0.71 (0.39–1.27) |
1.00 |
0.87 (0.48–1.59) |
0.89 (0.50–1.61) |
0.75 (0.42–1.35) |
| Model 2 |
1.00 |
0.53 (0.29–0.97) |
0.45 (0.24–0.85) |
0.69 (0.38–1.27) |
1.00 |
0.86 (0.47–1.57) |
0.87 (0.48–1.57) |
0.69 (0.38–1.25) |
| Model 3 |
1.00 |
0.47 (0.25–0.86) |
0.42 (0.22–0.80) |
0.67 (0.36–1.23) |
1.00 |
0.86 (0.47–1.58) |
0.88 (0.48–1.62) |
0.68 (0.37–1.26) |
Model 1 was adjusted for age.
Model 2 was adjusted for variables in model 1 plus alcohol intake (ethanol) and smoking (pack-years).
Model 3 was adjusted for variables in model 2 plus body mass index, diabetes, dyslipidemia, and hypertension.
CI indicates confidence interval; and TCSVD, total cerebral small vessel disease.
Table 4 shows the multivariable-adjusted ORs, using binary logistic regression, for the prevalent individual CSVD features by quartiles of serum irisin levels, with Q1 as the reference group. The prevalence of CSVD features (DSWMHs, PVHs, LIs, and CMBs) decreased in Q2 and Q3 of serum irisin levels, except for DSWMHs in Q3 and PVHs in Q2, in which their prevalence was slightly higher than in Q1 and/or Q4. Approximately 25% of participants had DSWMHs and PVHs, approximately 20% had LIs, and approximately 15.0% had CMBs across the quartiles of serum irisin levels. Serum irisin levels generally had slightly lower OR for each form of CSVD (DSWMHs, PVHs, LIs, and CMBs). However, serum irisin levels in Q2 and Q3 were significantly associated with a reduced risk of LIs (OR, 0.53; 95% CI, 0.30–0.93 and OR, 0.55; 95% CI, 0.31–0.98, respectively) in the total sample, those in Q3 with PVHs (OR, 0.27; 95% CI, 0.10–0.74), and those in Q2 with LIs (OR, 0.37; 95% CI, 0.15–0.92) and CMBs (OR, 0.25; 95% CI, 0.08–0.83) in younger adults (Table 4 and Supplemental Table 3). No relationships were observed in older adults (≥ 65 years).
Table 4.
Multivariable Logistic Regression Models for Relationships (Odds Ratio with 95% CI) between Serum Irisin Levels at Baseline and Cerebral Small Vessel Disease in the 5-year Follow-up (
n = 720)
|
Quartiles of serum irisin levels in μg/mL (n = 180 in each quartile)
|
| Q1 (2.67–5.75) |
Q2 (5.76–7.01) |
Q3 (7.02–8.33) |
Q4 (8.34–24.52) |
| DSWMHs, n (%)
|
40 (22.2) |
37 (20.6) |
42 (23.3) |
37 (20.6) |
| Model 1 |
1.00 |
0.82 (0.49–1.36) |
0.92 (0.56–1.53) |
0.76 (0.45–1.28) |
| Model 2 |
1.00 |
0.83 (0.49–1.39) |
0.92 (0.55–1.54) |
0.70 (0.41–1.20) |
| Model 3 |
1.00 |
0.80 (0.47–1.36) |
0.94 (0.56–1.59) |
0.70 (0.40–1.20) |
| PVHs, n (%)
|
45 (25.0) |
47 (26.1) |
43 (23.9) |
48 (26.7) |
| Model 1 |
1.00 |
0.84 (0.51–1.39) |
0.67 (0.40–1.12) |
0.74 (0.45–1.23) |
| Model 2 |
1.00 |
0.87 (0.53–1.45) |
0.67 (0.40–1.13) |
0.76 (0.46–1.27) |
| Model 3 |
1.00 |
0.85 (0.51–1.41) |
0.67 (0.39–1.13) |
0.73 (0.43–1.23) |
| LIs, n (%)
|
38 (21.1) |
31 (17.2) |
32 (17.8) |
42 (23.3) |
| Model 1 |
1.00 |
0.61 (0.35–1.06) |
0.58 (0.33–1.00) |
0.79 (0.47–1.33) |
| Model 2 |
1.00 |
0.55 (0.32–0.96) |
0.53 (0.30–0.93) |
0.69 (0.40–1.18) |
| Model 3 |
1.00 |
0.53 (0.30–0.93) |
0.55 (0.31–0.98) |
0.69 (0.39–1.20) |
| CMBs, n (%)
|
27 (15.0) |
18 (10.0) |
26 (14.4) |
27 (15.0) |
| Model 1 |
1.00 |
0.54 (0.28–1.04) |
0.78 (0.43–1.43) |
0.79 (0.44–1.44) |
| Model 2 |
1.00 |
0.54 (0.28–1.03) |
0.75 (0.41–1.38) |
0.77 (0.42–1.42) |
| Model 3 |
1.00 |
0.54 (0.28–1.04) |
0.77 (0.42–1.43) |
0.77 (0.41–1.43) |
Model 1 was adjusted for age.
Model 2 was adjusted for variables in model 1 plus alcohol intake (ethanol) and smoking (pack-years).
Model 3 was adjusted for variables in model 2 plus body mass index, diabetes, dyslipidemia, and hypertension.
CI indicates confidence interval; CMBs, cerebral microbleeds; DSWMHs, deep and subcortical white matter hyperintensities; LIs, lacunar infarcts; and PVHs, periventricular hyperintensities.
Supplemental Table 3.
Multivariable Logistic Regression Models for Relationships (Odds Ratio with 95% CI) between Serum Irisin Levels at Baseline and Cerebral Small Vessel Disease in the 5-year Follow-up (
n = 720)
|
Quartiles of serum irisin levels in μg/mL |
| ＜65 years (n=386)
|
≥ 65 years (n=334)
|
|
Q1 (2.67–5.74)
n=119
|
Q2 (5.75–7.01)
n=97
|
Q3 (7.02–8.33)
n=88
|
Q4 (8.34–19.67)
n=82
|
Q1 (3.69–5.75)
n=61
|
Q2 (5.76–7.01)
n=83
|
Q3 (7.02–8.30)
n=92
|
Q4 (8.31–24.52)
n=98
|
| DSWMHs, n (%)
|
24 (20.2) |
14 (14.4) |
14 (15.9) |
19 (23.2) |
16 (26.2) |
23 (27.7) |
28 (30.6) |
18 (18.4) |
| Model 1 |
1.00 |
0.59 (0.29–1.24) |
0.66 (0.31–1.38) |
0.99 (0.49–1.98) |
1.00 |
1.09 (0.52–2.31) |
1.26 (0.61–2.62) |
0.65 (0.30–1.40) |
| Model 2 |
1.00 |
0.58 (0.27–1.24) |
0.66 (0.31–1.41) |
0.94 (0.46–1.95) |
1.00 |
1.08 (0.51–2.27) |
1.24 (0.59–2.58) |
0.56 (0.25–1.25) |
| Model 3 |
1.00 |
0.50 (0.23–1.10) |
0.61 (0.28–1.33) |
0.97 (0.46–2.04) |
1.00 |
1.08 (0.50–2.30) |
1.24 (0.59–2.62) |
0.55 (0.24–1.24) |
| PVHs, n (%)
|
21 (17.7) |
15 (15.5) |
7 (8.0) |
14 (17.1) |
24 (39.3) |
32 (38.6) |
36 (39.1) |
34 (34.7) |
| Model 1 |
1.00 |
0.70 (0.33–1.49) |
0.31 (0.12–0.78) |
0.68 (0.31–1.47) |
1.00 |
0.99 (0.50–1.96) |
1.04 (0.53–2.03) |
0.85 (0.44–1.66) |
| Model 2 |
1.00 |
0.75 (0.35–1.64) |
0.29 (0.11–0.78) |
0.71 (0.32–1.58) |
1.00 |
0.98 (0.50–1.94) |
1.04 (0.53–2.03) |
0.84 (0.43–1.65) |
| Model 3 |
1.00 |
0.69 (0.31–1.54) |
0.27 (0.10–0.74) |
0.69 (0.31–1.56) |
1.00 |
0.98 (0.49–1.95) |
1.05 (0.53–2.08) |
0.81 (0.41–1.64) |
| LIs, n (%)
|
18 (15.1) |
10 (10.3) |
9 (10.2) |
11 (13.4) |
20 (32.8) |
21 (25.3) |
23 (25.0) |
31 (31.6) |
| Model 1 |
1.00 |
0.56 (0.24–1.29) |
0.54 (0.23–1.29) |
0.68 (0.30–1.56) |
1.00 |
0.66 (0.32–1.38) |
0.62 (0.30–1.28) |
0.88 (0.44–1.75) |
| Model 2 |
1.00 |
0.42 (0.17–1.03) |
0.46 (0.19–1.12) |
0.60 (0.26–1.41) |
1.00 |
0.66 (0.32–1.38) |
0.58 (0.28–1.22) |
0.79 (0.39–1.61) |
| Model 3 |
1.00 |
0.37 (0.15–0.92) |
0.48 (0.19–1.18) |
0.55 (0.23–1.32) |
1.00 |
0.67 (0.31–1.42) |
0.61 (0.29–1.31) |
0.84 (0.40–1.76) |
| CMBs, n (%)
|
15 (12.6) |
4 (4.1) |
9 (10.2) |
12 (14.6) |
12 (19.7) |
14 (16.9) |
17 (18.5) |
15 (15.3) |
| Model 1 |
1.00 |
0.25 (0.08–0.80) |
0.67 (0.27–1.63) |
0.93 (0.40–2.15) |
1.00 |
0.83 (0.35–1.95) |
0.98 (0.41–2.13) |
0.74 (0.32–1.72) |
| Model 2 |
1.00 |
0.25 (0.08–0.79) |
0.59 (0.23–1.52) |
0.87 (0.37–2.06) |
1.00 |
0.81 (0.34–1.91) |
0.95 (0.41–2.19) |
0.73 (0.31–1.70) |
| Model 3 |
1.00 |
0.25 (0.08–0.83) |
0.60 (0.23–1.55) |
0.89 (0.37–2.13) |
1.00 |
0.81 (0.34–1.93) |
0.97 (0.42–2.27) |
0.70 (0.29–1.68) |
Model 1 was adjusted for age.
Model 2 was adjusted for variables in model 1 plus alcohol intake (ethanol) and smoking (pack-years).
Model 3 was adjusted for variables in model 2 plus body mass index, diabetes, dyslipidemia, and hypertension.
CI indicates confidence interval; CMBs, cerebral microbleeds; DSWMHs, deep and subcortical white matter hyperintensities; LIs, lacunar infarcts; and PVHs, periventricular hyperintensities.
Discussion
In this population-based observational study on Japanese men who were free from stroke and cognitive decline, serum irisin levels at baseline were associated with lower total CSVD score, that is, burden in the 5-year follow-up. Baseline serum irisin levels were also associated with a reduced risk of individual MRI CSVD features (PVHs, LIs, and CMBs), except for DSWMHs. These associations were independent of age, behavioral (the smoking and drinking status), and vascular (body mass index, diabetes, dyslipidemia, and hypertension) risk factors. To the best of our knowledge, this is the first study to investigate the relationship between serum irisin levels and the MRI-measured CSVD burden. Therefore, we were unable to compare the results obtained with previous findings; more evidence is awaited in this field. However, since some studies used physical activity (typically measured with a questionnaire), it was reasonable to compare our results with their findings.
In the present study, baseline serum irisin levels were associated with lower total CSVD score (burden) in the 5-year follow-up. A single study recently investigated cross-sectional and longitudinal relationships between self-reported physical activity and the total CSVD score (ranging between 0 and 3, generated from white matter hyperintensities, LIs, and CMBs) among 503 patients with CSVD29). Cross-sectionally, the active physical activity group had a significantly lower CSVD score in a zero-order relationship. However, it was no longer significant after adjustments for confounders, and no relationship was observed between physical activity and the CSVD score in longitudinal analyses29). It is important to note that they used patients with CSVD who were older than those in the present study and analyzed self-reported physical activity, which may explain the discrepancy between the present results and their findings.
Of note, we found a J-shaped relationship between serum irisin levels and the total CSVD burden; Q3 (serum irisin levels of 7.02–8.33 µg/mL) had lower OR than Q1 (Tables 2 and 3). We previously reported a similar relationship between step counts and CSVD11). However, the underlying mechanisms have not yet been elucidated. Recent systematic reviews and meta-analyses confirmed that relationships between physical activity/step counts and health outcomes, including all-cause mortality, are generally curvilinear or nonlinear30, 31). The threshold effects of physical activity/irisin levels may have contributed to these relationships. The beneficial effects of physical activity on CSVD warrant further studies using exercise-induced biomarkers, such as irisin, including randomized controlled trials to eliminate alternative causal mechanisms.
In the present study, serum irisin levels generally tended to have lower OR for prevalent subclinical PVHs; however, a significant association was observed in younger adults. We previously reported, using the same samples, that step counts significantly reduced the risk of prevalent PVHs11). This result is also consistent with the findings of other studies in which increasing self-reported physical activity levels at baseline were associated with a combined32) and separate33) rating of deep and periventricular white matter hyperintensities in 3- and 5-year follow-ups, respectively. In contrast, we did not find statistically significant relationships between serum irisin levels and DSWMHs in overall and age-stratified analyses. This discrepancy may be explained by differences in the functional, microstructural, and etiological features between DSWMHs and PVHs34). However, the underlying mechanisms of such differential relationships of serum irisin levels with location-specific white matter hyperintensities (PVHs and DSWMHs) are unclear and unknown. Additionally, the number of prevalent PVHs and DSWMHs in the present study, especially in younger adults, may not be sufficient to detect small differences or find statistically significant results in DSWMHs, or the observed significant results in PHVs might have occurred by chance. These relationships need to be examined in more detail in longitudinal studies and clinical trials.
In the present study, serum irisin levels were associated with a reduced risk of LIs, reaffirming the relationship between step counts and the prevalent subclinical LIs observed in our previous study11). Two other studies reported similar findings showing that higher self-reported physical activity was associated with a reduced risk of subclinical cerebral infarct on MRI35, 36). A recent study also found a similar cross-sectional relationship between baseline physical activity (self-reported) and LIs among patients with CSVD29). However, this relationship disappeared in the longitudinal analysis because patients with CSVD who were older and medically less fit were more likely to drop out29).
We also observed a significant association of serum irisin levels with prevalent CMBs. In a single cohort study targeting healthy adults, the weekly frequency of physical activity significantly mitigated the risk of CMBs37). Another cohort study that targeted a patient population (mild cognitive impairment with CSVD) reported that low physical activity (self-reported) was associated with lobar CMBs38). In contrast, a recent study with a similar patient population reported a null association29). In the present study, we used irisin, an exercise-induced myokine/biomarker, which provided a more reliable relationship.
We reported no significant associations (for all analyses) in older adults aged ≥ 65 years in our sample. Although serum irisin levels are associated with age, the co-existence of higher comorbidities, such as diabetes and hypertension, in older adults than in younger adults (data not shown) may have contributed to this result even after appropriate adjustments. It is also possible that age is not only a strong confounder but also an effect modifier to these relationships. Evidence shows that age is strongly associated with CSVD in Asian populations5). Our study also showed no association between serum irisin levels and step counts (Supplemental Fig.2). Various factors, including age, cardiometabolic state, and exercise habits/types or duration, may influence the regulation of irisin in humans. For instance, a recent review report suggests that acute exercise can promote irisin secretion in humans. However, the effects of chronic exercise remain under debate39), which may support our study finding on “no association” between serum irisin levels and step counts. Research in this field is still in the infancy stage; therefore, more evidence is needed from clinical trials to clarify these relationships.
The mechanisms underlying the protective effects of serum irisin levels on CSVD remain unclear. However, mounting evidence revealed that irisin crosses the blood-brain barrier through a peroxisome proliferator-activated receptor gamma coactivator 1-alpha/fibronectin type III domain-containing protein 5 pathway to exert its neuroprotective effects in animal models and humans13, 17-19). It has been shown to upregulate brain-derived neurotrophic factor, a neurotrophin, in the brain, which protects neurons from injury. Irisin also influenced mitochondrial dynamics, which are essential for normal neuronal functions13, 17-19). Notably, serum irisin was recently posited as a potential biomarker/mediator for cognitive decline and neurodegenerative diseases, including Alzheimer’s disease18, 19, 40, 41), and a new protagonist to treat and prevent acute brain injury, including stroke42).
The use of serum irisin levels is one of the strengths of the present study, which may help clarify, for the first time, the mechanisms underlying the relationship between physical activity and CSVD, thereby supporting physical activity, including step counts, as a neuroprotective lifestyle factor. A better understanding of how physical activity exerts neuroprotective effects may be a potential motivational factor for individuals and facilitate behavioral changes. The present results provide novel insights that may provide the general population with a better understanding of the brain health implications of physical activity. Moreover, serum irisin levels were independently related with the reduced risk of CSVD and, thus, have potential as a new biomarker and therapeutic target for CSVD.
There are several limitations that need to be addressed. CSVD features on MRI (DSWMHs, PVHs, LIs, and CMBs) were not assessed at baseline; therefore, we were unable to confirm the onset of CSVD during the follow-up. Furthermore, these features may have been misclassified because of their visual ratings, which cannot be overruled for certainty. In addition, the present study only included Japanese men; therefore, the results obtained cannot be applied to other populations. Another limitation is the cross-sectional nature of the study, which prevents a cause–effect relationship being established. The relationships of serum irisin levels with individual CSVD features in younger adults should be interpreted cautiously considering the number of prevalent cases. Furthermore, other unknown or unmeasured factors that influence serum irisin levels and CSVD may have been residual confounders. Due to the paucity of information on the role of irisin in subclinical CSVD, further studies need to be conducted on different populations to elucidate the mechanisms by which irisin mediates the neuroprotective effects of physical activity (exercise) on CSVD.
Conclusions
In general Japanese men, serum irisin levels were associated with less burden of total CSVD, with the lowest risk being observed among participants with serum irisin levels of approximately 7.0–8.5 µg/mL. Serum irisin levels were also related to a reduced risk of subclinical PVHs, LIs, and CMBs, but not DSWMHs. Physical activity (exercise) as a neuroprotective lifestyle factor may contribute to the prevention of CSVD.
Acknowledgements
The authors sincerely appreciate the members of the SESSA Research Group, who approved the current study.
Notice of Grant Support
The SESSA has been supported by the JSPS KAKENHI Grant Numbers (A) 13307016, (A) 17209023, (A) 21249043, (A) 23249036, (A) 25253046, (A) 15H02528, (A) 18H04074, (B) 26293140 (B) 24790616, (B) 21790579, (B) 18H03048, (B) 15K19224, (B) 15K19225, and (C) 23590790 from the Ministry of Education, Culture, Sports, Science, and Technology Japan, and by the grant R01HL068200 from Glaxo-Smith Kline GB. JSPS has also supported the present study for Postdoctoral Fellowships for Research in Japan. The funding agencies listed above had no role in the study design, data collection, analyses, interpretation of results, or writing of the manuscript.
Conflicts of Interest
The authors do not have any conflicts of interest to disclose.
Author Contributions
MM: conceptualized and designed the study, analyzed data, interpreted results, and wrote the manuscript. AK and YY: conceptualized and designed the study, guided the manuscript writing, critically interpreted results, and revised the manuscript. YW, AS, KN, and AF: acquisition of MRI-related data, and reviewed the manuscript. NM and TH: coordinated and supervised data collection, controlled quality at all stages, and reviewed the manuscript. HS, KK, and ST: collected and managed data, and reviewed the manuscript. HU and KM: oversaw the quality control measures, provided overall guidance for manuscript writing, and critically reviewed the manuscript. All authors read and approved the final manuscript.
Members of the SESSA Research Group:
Co-principal investigators: Hirotsugu Ueshima, Katsuyuki Miura (Shiga University of Medical Science, Otsu, Shiga).
Research members: Yoshihisa Nakagawa, Yasutaka Nakano, Emiko Ogawa, Hiroshi Maegawa, Katsutaro Morino, Itsuko Miyazawa, Yoshiyuki Watanabe, Kazuhiko Nozaki, Ikuo Tooyama, Akihiko Shiino, Akira Andoh, Teruhiko Tsuru, Hisakazu Ogita, Naomi Miyamatsu, Yasuyuki Nakamura, Yuichiro Yano, Aya Kadota, Keiko Kondo, Sayuki Torii, Takashi Kadowaki, Sayaka Kadowaki, Takahiro Ito, Ayako Kunimura, Hiroyoshi Segawa, Yukiko Okami (Shiga University of Medical Science, Otsu, Shiga), Akira Fujiyoshi, Aya Higashiyama (Wakayama Medical University, Wakayama), Tomonori Okamura, Naoko Miyagawa (Keio University, Tokyo), Tatsuya Sawamura (Shinshu University, Matsumoto, Nagano), Michiya Igase (Ehime University, Toon, Ehime), Yasuharu Tabara (Shizuoka Graduate University of Public Health, Shizuoka), Akira Sekikawa, Emma JM Barinas-Mitchell (University of Pittsburgh, Pittsburgh, PA, USA),Daniel Edmundowicz (Temple University, Philadelphia, PA, USA), Takayoshi Ohkubo (Teikyo University, Tokyo), Atsushi Hozawa (Tohoku University, Sendai, Miyagi), Yoshitaka Murakami (Toho University, Tokyo), Nagako Okuda (Kyoto Prefectural University), Hisatomi Arima, Atsushi Satoh (Fukuoka University, Fukuoka), Yoshikuni Kita (Tsuruga Nursing University, Tsuruga, Fukui), Takashi Hisamatsu (Okayama University, Okayama), Masahiko Yanagita (Doshisha University, Kyotanabe, Kyoto), Seiko Ohno (National Cerebral and Cardiovascular Center, Suita, Osaka), Naoyuki Takashima (Kindai University, Osakasayama, Osaka), Takashi Yamamoto (Kohka Public Hospital, Shiga), Koichiro Azuma (Nerima General Hospital, Tokyo), Maryam Zaid (Fudan University Shanghai, China), Yoshino Saito (Aino University, Ibaraki, Osaka).
References
- 1) Chojdak-Łukasiewicz J, Dziadkowiak E, Zimny A and Paradowski B: Cerebral small vessel disease: A review. Adv Clin Exp Med, 2021; 30: 349-356
- 2) Chen X, Wang J, Shan Y, Cai W, Liu S, Hu M, Liao S, Huang X, Zhang B, Wang Y and Lu Z: Cerebral small vessel disease: neuroimaging markers and clinical implication. J Neurol, 2019; 266: 2347-2362
- 3) Rensma SP, van Sloten TT, Launer LJ and Stehouwer CDA: Cerebral small vessel disease and risk of incident stroke, dementia and depression, and all-cause mortality: A systematic review and meta-analysis. Neurosci Biobehav R, 2018; 90: 164-173
- 4) Shi Y and Wardlaw JM: Update on cerebral small vessel disease: A dynamic whole-brain disease. Stroke Vasc Neurol, 2016; 1: 83-92
- 5) Das AS, Regenhardt RW, Vernooij MW, Blacker D, Charidimou A and Viswanathan A: Asymptomatic cerebral small vessel disease: Insights from population-based studies. J Stroke, 2019; 21: 121-138
- 6) Li Q, Yang Y, Reis C, Tao T, Li W, Li X and Zhang JH: Cerebral small vessel disease. Cell Transplant, 2018; 27: 1711-1722
- 7) Smith EE and Markus HS: New treatment approaches to modify the course of cerebral small vessel diseases. Stroke, 2020; 51: 38-46
- 8) Nay K, Smiles WJ, Kaiser J, McAloon LM, Loh K, Galic S, Oakhill JS, Gundlach AL and Scott JW: Molecular mechanisms underlying the beneficial effects of exercise on brain function and neurological disorders. Int J Mol Sci, 2021; 22: 4052
- 9) Di Liegro CM, Schiera G, Proia P and Di Liegro I: Physical activity and brain health. Genes (Basel), 2019; 10: 720
- 10) Umegaki H, Sakurai T and Arai H: Active Life for Brain Health: A narrative review of the mechanism underlying the protective effects of physical activity on the brain. Front Aging Neurosci, 2021; 13: 761674
- 11) Moniruzzaman M, Kadota A, Segawa H, Kondo K, Torii S, Miyagawa N, Fujiyoshi A, Hisamatsu T, Watanabe Y, Shiino A, Nozaki K, Ueshima H and Miura K: Relationship between step counts and cerebral small vessel disease in Japanese men. Stroke, 2020; 51: 3584-3591
- 12) Bostrom P, Wu J, Jedrychowski MP, Korde A, Ye L, Lo JC, Rasbach KA, Bostrom EA, Choi JH, Long JZ, Kajimura S, Zingaretti MC, Vind BF, Tu H, Cinti S, Hojlund K, Gygi SP and Spiegelman BM: A PGC1-alpha-dependent myokine that drives brown-fat-like development of white fat and thermogenesis. Nature, 2012; 481: 463-468
- 13) Young MF, Valaris S and Wrann CD: A role for FNDC5/Irisin in the beneficial effects of exercise on the brain and in neurodegenerative diseases. Prog Cardiovasc Dis, 2019; 62: 172-178
- 14) Pedersen BK: Physical activity and muscle–brain crosstalk. Nat Rev Endocrinol, 2019; 15: 383-392
- 15) de Freitas GB, Lourenco MV and De Felice FG: Protective actions of exercise-related FNDC5/irisin in memory and Alzheimer’s disease. J Neurochem, 2020; 155: 602-611
- 16) Severinsen MCK and Pedersen BK: Muscle-organ crosstalk: The emerging roles of myokines. Endocr Rev, 2020; 41: 594-609
- 17) Liu Y, Zhu C, Guo J, Chen Y and Meng C: The neuroprotective effect of irisin in ischemic stroke. Front Aging Neurosci, 2020; 12: 588958
- 18) Pignataro P, Dicarlo M, Zerlotin R, Zecca C, Dell’Abate MT, Buccoliero C, Logroscino G, Colucci S and Grano M: FNDC5/irisin system in neuroinflammation and neurodegenerative diseases: Update and novel perspective. Int Jo Mol Sci, 2021; 22: 1605
- 19) Jodeiri Farshbaf M and Alviña K: Multiple roles in neuroprotection for the exercise derived myokine irisin. Front Aging Neurosci, 2021; 13: 649929
- 20) Kadota A, Miura K, Okamura T, Fujiyoshi A, Ohkubo T, Kadowaki T, Takashima N, Hisamatsu T, Nakamura Y, Kasagi F, Maegawa H, Kashiwagi A and Ueshima H: Carotid intima-media thickness and plaque in apparently healthy Japanese individuals with an estimated 10-year absolute risk of CAD death according to the Japan Atherosclerosis Society (JAS) guidelines 2012: The Shiga Epidemiological Study of Subclinical Atherosclerosis (SESSA). J Atheroscler Thromb, 2013; 20: 755-766
- 21) Moniruzzaman M, Kadota A, Shiino A, Fujiyoshi A, Ito T, Syaifullah AH, Miyagawa N, Kondo K, Hisamatsu T, Segawa H, Tooyama I, Ueshima H and Miura K: Seven-day pedometer-assessed step counts and brain volume: A population-based obervational study. J Phys Act Health, 2021; 18: 157-164
- 22) Hisamatsu T, Miura K, Arima H, Fujiyoshi A, Kadota A, Kadowaki S, Zaid M, Miyagawa N, Satoh A, Kunimura A, Horie M and Ueshima H: Relationship of serum irisin levels to prevalence and progression of coronary artery calcification: A prospective, population-based study. Int J Cardiol, 2018; 267: 177-182
- 23) Ganbaatar N, Kadota A, Hisamatsu T, Araki S, Kume S, Fujiyoshi A, Kadowaki S, Torii S, Kondo K, Segawa H, Salman E, Miyazawa I, Yamamoto T, Nakagawa Y, Maegawa H, Miura K, Ueshima H and for the SRG: Relationship between kidney function and subclinical atherosclerosis progression evaluated by coronary artery calcification. J Atheroscler Thromb, 2022; 29: 1359-1371
- 24) Wardlaw JM, Smith EE, Biessels GJ, Cordonnier C, Fazekas F, Frayne R, Lindley RI, O’Brien JT, Barkhof F, Benavente OR, Black SE, Brayne C, Breteler M, Chabriat H, Decarli C, de Leeuw FE, Doubal F, Duering M, Fox NC, Greenberg S, Hachinski V, Kilimann I, Mok V, Oostenbrugge R, Pantoni L, Speck O, Stephan BC, Teipel S, Viswanathan A, Werring D, Chen C, Smith C, van Buchem M, Norrving B, Gorelick PB and Dichgans M: Neuroimaging standards for research into small vessel disease and its contribution to ageing and neurodegeneration. Lancet Neurol, 2013; 12: 822-838
- 25) Kohara K, Okada Y, Ochi M, Ohara M, Nagai T, Tabara Y and Igase M: Muscle mass decline, arterial stiffness, white matter hyperintensity, and cognitive impairment: Japan Shimanami Health Promoting Program study. J Cachexia Sarcopenia Muscle, 2017; 8: 557-566
- 26) Okada Y, Kohara K, Ochi M, Nagai T, Tabara Y, Igase M and Miki T: Mechanical stresses, arterial stiffness, and brain small vessel diseases. Stroke, 2014; 45: 3287-3292
- 27) Staals J, Makin SDJ, Doubal FN, Dennis MS and Wardlaw JM: Stroke subtype, vascular risk factors, and total MRI brain small-vessel disease burden. Neurology, 2014; 83: 1228-1234
- 28) Salman E, Kadota A, Hisamatsu T, Segawa H, Torii S, Fujiyoshi A, Kondo K, Watanabe Y, Arima H, Shiino A, Nozaki K, Ueshima H, and Miura K: Relationship of four blood pressure indexes to subclinical cerebrovascular diseases assessed by brain MRI in general Japanese men. J Atheroscler Thromb, 2022; 29: 174-187
- 29) Landman TR, Thijssen DH, Tuladhar AM and de Leeuw FE: Relation between physical activity and cerebral small vessel disease: A nine-year prospective cohort study. Int J Stroke, 2021; 16: 962-971
- 30) Warburton DER and Bredin SSD: Health benefits of physical activity: A systematic review of current systematic reviews. Curr Opin Cardiol, 2017; 32: 541-556
- 31) Sheng M, Yang J, Bao M, Chen T, Cai R, Zhang N, Chen H, Liu M, Wu X, Zhang B, Liu Y and Chao J: The relationships between step count and all-cause mortality and cardiovascular events: A dose-response meta-analysis. J Sport Health Sci, 2021; 10: 620-628
- 32) Gow AJ, Bastin ME, Munoz Maniega S, Valdes Hernandez MC, Morris Z, Murray C, Royle NA, Starr JM, Deary IJ and Wardlaw JM: Neuroprotective lifestyles and the aging brain: activity, atrophy, and white matter integrity. Neurology, 2012; 79: 1802-1808
- 33) Podewils LJ, Guallar E, Beauchamp N, Lyketsos CG, Kuller LH and Scheltens P: Physical activity and white matter lesion progression: Assessment using MRI. Neurology, 2007; 68: 1223-1226
- 34) Griffanti L, Jenkinson M, Suri S, Zsoldos E, Mahmood A, Filippini N, Sexton CE, Topiwala A, Allan C, Kivimäki M, Singh-Manoux A, Ebmeier KP, Mackay CE and Zamboni G: Classification and characterization of periventricular and deep white matter hyperintensities on MRI: A study in older adults. Neuroimage, 2018; 170: 174-181
- 35) Dubbert PM, Penman AD, Evenson KR, Reeves RR and Mosley TH, Jr.: Physical activity and subclinical MRI cerebral infarcts: The ARIC Study. J Neurol Sci, 2009; 284: 135-139
- 36) Willey JZ, Moon YP, Paik MC, Yoshita M, Decarli C, Sacco RL, Elkind MS and Wright CB: Lower prevalence of silent brain infarcts in the physically active: The Northern Manhattan Study. Neurology, 2011; 76: 2112-2118
- 37) Daugherty AM and Raz N: Incident risk and progression of cerebral microbleeds in healthy adults: A multi-occasion longitudinal study. Neurobiol Aging, 2017; 59: 22-29
- 38) Valenti R, Del Bene A, Poggesi A, Ginestroni A, Salvadori E, Pracucci G, Ciolli L, Marini S, Nannucci S, Pasi M, Pescini F, Diciotti S, Orlandi G, Cosottini M, Chiti A, Mascalchi M, Bonuccelli U, Inzitari D and Pantoni L: Cerebral microbleeds in patients with mild cognitive impairment and small vessel disease: The Vascular Mild Cognitive Impairment (VMCI)-Tuscany study. J Neurol Sci, 2016; 368: 195-202
- 39) Bao JF, She QY, Hu PP, Jia N and Li A: Irisin, a fascinating field in our times. Trends Endocrinol Metab, 2022; 33: 601-613
- 40) Zhang F, Hou G, Hou G, Wang C, Shi B and Zheng Y: Serum irisin as a potential biomarker for cognitive decline in vascular dementia. Front Neurol, 2021; 12: 755046
- 41) Kim OY and Song J: The role of irisin in Alzheimer’s disease. J Clin Med, 2018; 7: 407
- 42) Tu T, Peng J and Jiang Y: FNDC5/irisin: A new protagonist in acute brain injury. Stem Cells Dev, 2020; 29: 533-543
