PDF
ePub
Citation
Print
INTRODUCTION
An interarm blood pressure difference (IABD) is frequently observed clinically. The prevalence of IABD ≥10 mm Hg is 4.4% in the general population without vascular disease, but this is higher in patients with cardiovascular risk factors including diabetes mellitus, hypertension, and stroke.12 The IABD was reported to be independently related to cardiovascular and all-cause mortality,3 and this association has been demonstrated in cohorts without pre-existing cardiovascular disease.4
Cerebral small-vessel diseases (SVDs) represent ischemic or hemorrhagic damage in cerebral small arteries that appear as white-matter hyperintensities (WMHs), cerebral microbleeds (CMBs), perivascular spaces (PVSs), and asymptomatic lacunar infarctions (ALIs) in brain MRI.56 Although each SVD has diverse implications (e.g., WMHs, PVSs, ALIs for cerebral ischemia, and CMBs for cerebral hemorrhage), all cerebral SVDs have similar vascular risk factors or pathogenic mechanisms.78 Because these cerebral SVDs are independently associated with cognitive dysfunction, future stroke, and the prognosis,56910 it is important to identify the factor associated with cerebral SVDs in order to prevent and treat these neurological diseases.
The IABD may be due to stenosis or stiffness of the subclavian artery or aorta, and it may induce cerebral hypoperfusion that results in brain damage.11 Moreover, IABD is associated with elevated arterial stiffness in the elderly.12 Because cerebral hypoperfusion and arterial stiffness are closely associated with cerebral SVDs, IABD may be related to the presence and burden of cerebral SVDs. However, few studies have attempted to confirm this association. We therefore aimed to determine that whether IABD is related to cerebral SVDs.
METHODS
Subjects
Between January 2012 and June 2016, 1,958 consecutive patients who developed first-ever acute stroke within 7 days after developing neurological symptoms were included from our prospective stroke registry.13 During admission, demographics, past medical and medication history, clinical and neurological presentation, and classical risk factors for cerebrovascular disease were registered. Based on the protocol of our department, brain CT and/or MRI, imaging of the cerebral vasculature, chest X-rays, 12-lead cardiac electrocardiography, and blood laboratory tests were routinely performed. For investigating concomitant systemic atherosclerosis, the ankle-brachial index (ABI) examination was also routinely performed during admission using an automated device (VP-1000, Colin Medical Technology Corporation, Komaki, Japan).1314
These 1,958 subjects did not include patients with a potential source of cardioembolism [n=301; atrial fibrillation (n=270), sick sinus syndrome (n=5), and other cardioembolic source (n=26)], rare causes such as dissection or venous thromboembolism (n=45), transient ischemic attack (n=205), or undetermined etiology due to an incomplete investigation (n=17). Because arrhythmia could prohibit the accurate investigation of arterial stiffness [brachial-ankle pulse wave velocity (baPWV)] and blood pressure, patients with cardioembolism stroke subtype were excluded,215 as were patients in whom brain MRI was not performed (n=25), images of the cerebral intra- and extravasculature were not available (n=12), gradient recalled echo (GRE) images were not available (n=8), and the available MRI images were of poor quality (n=6). Additionally, subjects in whom the ABI examination was not performed (n=72) or ABI <0.9 (n=62), which may be related to inaccurate measurements of the arterial stiffness and blood pressure,16 were also excluded. Ultimately 1,205 patients were finally included in our study (Fig. 1). The stroke classification was determined based on the Trial of Org 10,172 in Acute Stroke Treatment classification system.17 Our study was approved by the Institutional Review Board of our hospital (IRB No. 2017-04-017-001), and the requirement to obtain informed consent from patients was waived because of the retrospective, cross-sectional, and observational design of the study.
Measurement of blood pressures in both arms and IABDs
The systolic and diastolic blood pressures were investigated in both arms with the subject in a supine position using an automated device designed primarily to measure ABI. The ABI test was performed by a well-trained examiner with more than 5 years of experience after the subject had rested in a quiet room for at least 5 minutes. The blood pressure was measured bilaterally in the brachial and posterior tibial arteries automatically and simultaneously using the oscillometric method.18 These blood pressures in both arms were checked once. Blood pressure differences in the lower limbs were not used in our study because such evaluations were outside the scope of this study. The presence of a significant IABD [interarm systolic blood pressure difference (IASBD) and/or interarm diastolic blood pressure difference (IADBD)] was considered as an absolute IASBD ≥10 mm Hg or an absolute IADBD ≥10 mm Hg.19
Protocol of brain MRI and definition of cerebral SVDs
The protocol of brain MRI used in this study was described in detail previously.2021 All brain MRI was performed using a 3-T scanner (Philips Achieva version 2.6, Best, the Netherlands). Brain MRI slices were acquired parallel to the orbitomeatal line using the following parameters: TR/TE=12,000/120 ms, pixel spacing=0.449/0.449 mm, field of view (FOV)=183×230 mm, and slice thickness=5 mm for FLAIR images; TR/TE=15,000/90 ms, pixel spacing=0.240/0.240 mm, FOV=176×220 mm, and slice thickness=5 mm for T2-weighted images; and TR/TE=571/21.9 ms, pixel spacing=0.449/0.449 mm, FOV=145×250 mm, and slice thickness=5 mm for GRE images.2021
The degree of WMHs was decided based on the deep or periventricular white matter in FLAIR images according to the Fazekas grading methods.15 A Fazekas grade of ≥2 in the deep or periventricular white matter was defined as HWH. The presence of CMBs was indicated by punctate hypointense lesions <10 mm on GRE images.22 PVSs were defined as punctate and/or linear hyperintense lesions <3 mm in the basal ganglia and centrum semiovale on T2-weighted images.23 High-grade perivascular spaces (HPVSs) were considered to be present if there were PVSs of grade 2–4 in the basal ganglia and centrum semiovale, based on a previous report.8 ALIs were considered as circular and/or cavitary lesions (signals similar to cerebrospinal fluid) with hyperintensities ≥3 mm and <15 mm on T2-weighted images with decreased signal intensity on T1-weighted images, with no relevant history of neurological signs or symptoms (Fig. 2). The degree of HWHs, CMBs, HPVSs, and ALIs were defined outside the area of acute cerebral infarction (based on diffusion-weighted images). The brain MRI lesions were independently measured by two neurologists (Y.C. and T.J.S.) who were blinded to the clinical information.
The total SVD score was determined as the summation of all cerebral SVDs present. One point was given for each of the following parameters: existence of HWHs, CMBs, HPVSs, or ALIs.24 The coefficients for the interobserver agreements on the existence of HWHs, CMBs, HPVSs, and ALIs were 0.912, 0.956, 0.938, and 0.888, respectively. If there was disagreement over the existence of cerebral SVDs, a final decision was reached by consensus.
Clinical and laboratory variables
The risk factors for hypertension, diabetes mellitus, hyperlipidemia, smoking, coronary artery disease, metabolic syndrome, and alcohol intake were defined in detail in a previous study25 and Supplementary Material (in the online-only Data Supplement). Antihypertensive treatment after admission was defined as treatment with intravenous or oral antihypertensive agents within 7 days of admission. Our target level of antihypertensive therapy was defined as a systolic blood pressure of up to 200–220 mm Hg or a diastolic blood pressure of 120 mm Hg. However, in patients receiving thrombolytic treatment, the target level of antihypertensive therapy was a systolic blood pressure of up to 185 mm Hg or a diastolic blood pressure of 110 mm Hg. Left ventricular hypertrophy was diagnosed when electrocardiography findings were matched with at least one of the relevant voltage criteria.26 The baPWV was defined as the mean baPWV value bilaterally.
Statistical analysis
The Windows SPSS software package (version 21.0, IBM Corp., Armonk, NY, USA) was used for statistical analysis. Categorical variables are expressed as frequency and percentage values, and continuous variables are expressed as mean±SD values. Differences in demographic characteristics, presence of vascular risk factors, and brain MRI findings were compared between patients with IASBD and IADBD using the chi-square test, Fisher's exact test, and the independent t-test.
Receiver operating characteristic (ROC) curves were investigated in terms of the area under the curve (AUC), standard error, p value, sensitivity, and specificity according to IASBD ≥10 mm Hg or IADBD ≥10 mm Hg.
The associations between the presence of IABD and cerebral SVDs and the total SVD score were checked using multivariate binary logistic regression (for the presence of each cerebral SVD) and linear regression (for the total SVD score as the dependent variable) after entering sex, age, and variables with p<0.1 in the univariate analysis. Due to multicollinearity, the presence of each cerebral SVD was analyzed separately. A two-tailed p value of <0.05 was defined as statistically significant. There was no statistical interaction for the existence of cerebral SVDs between IABD (IASBD and IADBD) and baPWV. Among the demographics and risk factors, there was no multicollinearity for the presence of each cerebral SVD and the total SVD score as the dependent variable (variance inflation factor <2.0). For the sensitivity analysis, we performed further multivariate analyses for IABD ≥5 mm Hg, IABD ≥15 mm Hg, per 1-mm Hg change in IABD, both IASBD ≥5 mm Hg and IADBD ≥5 mm Hg, and both IASBD ≥10 mm Hg and IADBD ≥10 mm Hg as dependent variables.
RESULTS
Demographics and comparisons between patients with IABD ≥10 mm Hg and <10 mm Hg
The demographics and the frequency of vascular risk factors did not differ between the patients included in and excluded from this study, except for age (Supplementary Table 1 in the online-only Data Supplement). Among all included patients, 61.6% (742/1,205) were men and they were aged 64.6±11.3 years. The following characteristics were more common in the IASBD ≥10 mm Hg group (n=126, 10.4%) than in the IASBD <10 mm Hg group (n=1,079, 89.6%): male sex, older age, hypertension, smoking, metabolic syndrome, left ventricular hypertrophy, increased baPWV, presence of high-grade white-matter hyperintensities (HWHs), HPVSs, ALIs, and total SVD score of 1–4. Previous antithrombotic medication prior to admission was less common in the IASBD ≥10 mm Hg group (Table 1).
Table 1
Clinical characteristics and comparison of study patients according to different values of the IASBD and the IADBD
Data are n (%) or mean±SD values.
ABI: ankle-brachial index, ALIs: asymptomatic lacunar infarctions, baPWV: brachial-ankle pulse wave velocity, CMBs: cerebral microbleeds, DBP: diastolic blood pressure, HPVSs: high-grade perivascular spaces, HWHs: high-grade white-matter hyperintensities, IADBD: interarm diastolic blood pressure difference, IASBD: interarm systolic blood pressure difference, NIHSS: National Institutes of Health Stroke Scale, SBP: systolic blood pressure, SVD: small-vessel disease.
The following characteristics were more common in the IADBD ≥10 mm Hg group (n=62, 5.1%) than in the IADBD <10 mm Hg group (n=1,143, 94.9%): left ventricular hypertrophy, presence of HWHs and ALIs, and total SVD score of 1–3. Previous antithrombotic and antihypertensive medication before admission were less common in the IADBD ≥10 mm Hg group (Table 1).
Association between IABD and presence of cerebral SVDs
HWHs, high-grade deep WMHs, and high-grade periventricular WMHs were present in 378 (31.4%), 288 (23.9%), and 350 (29.0%) of the 1,205 subjects, respectively. CMBs were found in 395 (32.8%) of the patients: 29.1% in the mixed (lobar+nonlobar) area and 3.7% in the lobar area only. HPVSs and ALIs were evident in 244 (20.2%) and 376 (31.2%) of the patients, respectively. The total SVD scores were 0, 1, 2, 3, and 4 in 514 (42.7%), 274 (22.7%), 217 (18.0%), 115 (9.5%), and 85 (7.1%) of the patients, respectively (Table 2). The comparison results including demographics, risk factors, stroke subtypes, ABI parameters, IABD (≥5 mm Hg, ≥10 mm Hg, ≥15 mm Hg), and both IASBD and IASBD ≥5 mm Hg and ≥10 mm Hg according to the presence or absence of each cerebral SVD are presented in Table 2 and Supplementary Table 2 (in the online-only Data Supplement).
Table 2
Clinical characteristics and comparison of the study patients according to the presence of different types of cerebral SVDs
Data are n (%) or mean±SD values.
*p<0.05, †p<0.1.
ABI: ankle-brachial index, ALIs: asymptomatic lacunar infarctions, baPWV: brachial-ankle pulse wave velocity, CMBs: cerebral microbleeds, DBP: diastolic blood pressure, HPVSs: high-grade perivascular spaces, HWHs: high-grade white-matter hyperintensities, IABD: interarm blood pressure difference, IADBD: interarm diastolic blood pressure difference, IASBD: interarm systolic blood pressure difference, NIHSS: National Institutes of Health Stroke Scale, SBP: systolic blood pressure, SVDs: small-vessel diseases.
After adjusting for sex, age, and variables with p<0.1 in the univariate analysis, IASBD ≥10 mm Hg was independently related to the existence of HWHs [odds ratio (OR)= 1.94, 95% CI=1.32–2.84, p=0.011) and had a tendency to be associated with the presence of HPVSs (OR=1.45, 95% CI=0.49–2.23, p=0.089) and ALIs (OR=1.42, 95% CI=0.96–2.11, p=0.052). However, IASBD ≥10 mm Hg was not associated with the presence of CMBs (OR=1.09, 95% CI=0.73–1.61, p=0.634) (Table 3). The results of the sensitivity analysis of different IABD cutoff values (≥5 mm Hg or ≥15 mm Hg) are presented in Table 3 and Supplementary Tables 3 and 4 (in the online-only Data Supplement).
Table 3
Results of the multivariate analysis for the presence of cerebral SVDs according to IASBD and IADBD
Data are odds ratio (95% CI) values. Adjusted for sex, age, hypertension, diabetes mellitus, metabolic syndrome, and baPWV (a), adjusted for sex, age, hypercholesterolemia, left ventricular hypertrophy, antihypertensive medication after admission, and baPWV (b), adjusted for sex, age, hypertension, left ventricular hypertrophy, and baPWV (c), and adjusted for sex, age, hypertension, diabetes mellitus, metabolic syndrome, alcohol intake, and baPWV (d).
*p<0.05, †p<0.1.
ALIs: asymptomatic lacunar infarctions, baPWV: brachial-ankle pulse wave velocity, CMBs: cerebral microbleeds, HPVSs: high-grade perivascular spaces, HWHs: high-grade white-matter hyperintensities, IABD: interarm blood pressure difference, IADBD: interarm diastolic blood pressure difference, IASBD: interarm systolic blood pressure difference, SVDs: small-vessel diseases.
IADBD ≥10 mm Hg was independently related to the existence of HWHs (OR=2.23, 95% CI=1.32–3.76, p=0.012) and ALIs (OR=1.92, 95% CI=1.13–3.25, p=0.044), but not to CMBs and HPVSs. Moreover, both IASBD ≥10 mm Hg and IADBD ≥10 mm Hg were also associated with the presence of HWHs and ALIs (Table 3).
Association between IABD and burden of cerebral SVDs
In multivariate linear regression with the total SVD score as the dependent variable and with adjustment for sex, age, and variables with p<0.1 in the univariate analysis (hypertension, diabetes mellitus, hypercholesterolemia, metabolic syndrome, alcohol intake, left ventricular hypertrophy, and baPWV), IASBD ≥10 mm Hg (β=0.080, p=0.006) and IADBD ≥10 mm Hg (β=0.065, p=0.023) were significantly and positively correlated with the total SVD score (Table 4).
Table 4
Association of IABD with the total SVD score
*Results from multivariate linear regression with total SVD score as the dependent variable and with adjustment for sex, age and variables with p<0.1 in the univariate analysis (hypertension, diabetes mellitus, hypercholesterolemia, metabolic syndrome, alcohol intake, left ventricular hypertrophy, and baPWV).
baPWV: brachial-ankle pulse wave velocity, IABD: interarm blood pressure difference, IADBD: interarm diastolic blood pressure difference, IASBD: interarm systolic blood pressure difference, SE: standard error, SVDs: small-vessel diseases.
Predictability of IABD for presence of cerebral SVDs
In ROC curve analyses, the AUC for IABD ≥10 mm Hg was significant for the presence of HWHs (AUC=0.777 for IASBD ≥10 mm Hg and 0.736 for IADBD ≥10 mm Hg) and ALIs (AUC=0.745 for IASBD ≥10 mm Hg and 0.751 for IADBD ≥10 mm Hg), but not for CMBs and HPVSs (Supplementary Table 5 in the online-only Data Supplement).
DISCUSSION
This study found that IASBD ≥10 mm Hg was independently related to the existence of HWHs and had a tendency to be related to the presence of HPVSs and ALIs. IADBD ≥10 mm Hg was also associated with the existence of HWHs and ALIs, but not with CMBs and HPVSs. Furthermore, IASBD ≥10 mm Hg and/or IADBD ≥10 mm Hg were positively correlated with the burden of cerebral SVDs. Thus, our study has revealed that IASBD and IADBD are variously related to the presence and burden of cerebral SVDs. In the recent Framingham Heart Study, a high IABD was associated with an increased risk of Alzheimer's disease and subclinical brain injury.27 The results of the present study are in line with those of that previous population-based study, and they provide additional information on stroke subjects. In clinical practice it is easier to measure IABD than to perform brain CT or MRI, and so our findings suggest the usefulness of IABD as one of the screening tools or criterion for investigating the presence of cerebral SVDs.
The mechanisms underlying the present results are unclear, but following hypotheses can be proposed. First, a large IABD can result from stenosis of the proximal aorta, brachiocephalic artery, and subclavian artery.3 This would mean that the large IABD is associated with poor cerebral perfusion, which results in damage to the brain parenchyma.11 A previous study found cerebral hypoperfusion to be associated with an increased cerebral SVD burden.28 Second, increased arterial stiffness would be an associated factor. IASBD >10 mm Hg was previously found to be related to increased arterial stiffness,12 and it is well known that this is associated with the presence of cerebral SVDs.15 In the present study, baPWV was higher in patients with IASBD ≥10 mm Hg. Third, the well-known cardiovascular risk factors of hypertension and diabetes mellitus are also associated with cerebral SVDs. Therefore, the relationship between IABD and cerebral SVDs elucidated in the present study might be an epiphenomenon elicited by similar associated factors.
Our study revealed that IABD ≥10 mm Hg was related to the existence of HWHs, but was not associated with the presence of CMBs and HPVSs. Both CMBs and HPVSs are significantly associated with arterial stiffness,15 impaired permeability of the blood–brain barrier, and vascular inflammation, all of which are factors closely related to HWHs.589 Accordingly, because IABD is also related to increased arterial stiffness, it is likely that not only HMHs but also CMBs and HPVSs may be associated with IABD, which contrasts with the results of the present study. In contrast to cerebral hypoperfusion resulting from an IABD, CMBs can be induced by hypertension-related mechanical damage that ruptures tight junctions.9 Therefore, this difference in the mechanism resulting in the development of CMBs can explain the discrepancy of the nonsignificant relationship between IABD and the presence of CMBs. Also, the reasons for the lack of an association between IABD and HPVSs in this study remain unknown. Because IASBD ≥15 mm Hg was found to be associated with HPVSs, our results might have been due to a weak statistically relationship between IABD ≥10 mm Hg and HPVSs or decreased statistical power resulting from controlling for other strongly associated factors such as age and hypertension.
The present study further found that IASBD was significantly associated with deep WMHs, while not being associated with periventricular WMHs. A previous study investigating the correlation between brain MRI and brain histopathology found that periventricular WMHs were mainly associated with myelin loss or subependymal gliosis, which represent senile changes, whereas deep WMHs were associated with the loss of myelin or subependymal gliosis as well as ischemic damage of vascular origin.29 Although the present study did not reveal the exact mechanism, our findings suggest that IABD is at least partially associated with or contributes to cerebral SVDs via cerebral ischemia (HWHs, HPVSs, and ALIs) rather than cerebral hemorrhage (CMBs).
This study was subject to some limitations. First, although ABI is routinely measured in almost all consecutive patients, selection bias was possible due to the retrospective design of our study. Second, our study population was limited to noncardioembolic stroke patients. Since all stroke patients receive brain CT and/or MRI, which can give information about cerebral SVDs, adding IABD has little value when screening SVDs. However, our study is significant in that it showed a correlation between cerebral SVDs and IABD, which is easy to measure in clinical practice. Third, the characteristics of our study population make it difficult to generalize our findings to another population or cohort. Fourth, multiple, automatic, and simultaneous assessments are recommended for accurate IABD measurements, rather than one-time, manual, and sequential evaluation methods. We used an automatic and simultaneous device, but IABD was measured only once when checking ABI, and so the consistency of the measured IABD values is uncertain. Finally, because our study had a cross-sectional design, further long-term follow-up research is needed into the association of IABD with cerebral SVDs.
In conclusion, our study suggests that IABD ≥10 mm Hg could be a useful indicator of the presence and burden of cerebral SVDs in noncardioembolic stroke patients.
XML Download