Association of self-measured home, ambulatory, and strictly measured office blood pressure and their variability with intracranial arterial stenosis

Abstract

Objective:

Hypertension and intracranial artery stenosis (ICAS) are closely related; however, few studies have compared the strength of the relationship between strictly measured office and out-of-office blood pressure (BP) measurements. The relationship of day-by-day or short-term variability in BP to asymptomatic ICAS also remains unclear.

Methods:

In apparently healthy 677 men (mean age, 70.0 years) from a population-based cohort, we examined the association of strictly measured office BP and 7-day home BP with ICAS on magnetic resonance angiography. We conducted 24-hour ambulatory BP monitoring in 468 of the men. Variability indices included day-by-day, daytime, and nighttime variability, nocturnal decline, and morning pressor surge. Any ICAS was defined as either mild (1%–49%) or severe (≥50%) stenosis.

Results:

We observed mild and severe ICAS in 153 (22.6%) and 36 (5.3%) participants, respectively. In multivariable-adjusted Poisson regression with robust error variance, higher systolic BP in office, home, or ambulatory BP monitoring was associated with the presence of any or severe ICAS. The associations with ICAS were comparable between office, home, and ambulatory systolic BP (all heterogeneity P-values >0.1). Independent of mean systolic BP, greater nocturnal decline or morning pressor surge, but not day-by-day, daytime, or nighttime variability, in systolic BP was associated with higher burden of any or severe ICAS.

Conclusions:

The magnitude of association of strictly measured office BP for asymptomatic ICAS was comparable with that of BP measured at home or in ambulatory BP monitoring. Circadian BP variation based on ambulatory BP monitoring was positively associated with asymptomatic ICAS burden.

INTRODUCTION

There is increasing interest in intracranial artery stenosis (ICAS) as a crucial subclinical disease related to stroke and to dementia.[1,2] ICAS is estimated to be responsible for up to 8% of ischemic stroke and 34% of dementia in the United States.[1,2] The epidemiology of ICAS and its determinants may be more important in East Asia than in the United States/Western countries[3] because stroke has been more prevalent than coronary heart disease in East Asia.[4,5]

Elevated blood pressure (BP) is an established risk factor for stroke and for ICAS,[3,6] and the prevalence of elevated BP remains high, with the greatest absolute burden of elevated BP particularly in the East Asian and Pacific regions.[7] Recent international hypertension management guidelines confer increasing weight to methods of measuring BP outside the medical office: self-measurements at home and/or 24-hour ambulatory BP measurements, to assess cardiovascular risk.[8-10] Indeed, numerous studies have demonstrated their prognostic value for cardiovascular disease.[11-13] However, few studies have investigated the association of self-measured or ambulatory BP outside the medical office with asymptomatic ICAS. There are also still insufficient data directly comparing strictly measured office, self-measured home, and ambulatory BP in the association with target organ damage, especially with subclinical cerebrovascular diseases such as asymptomatic ICAS. Further, independent of the mean BP, the day-by-day variability derived from self-measured home BP or daytime/nighttime variability, nocturnal decline, or morning pressor surge based on ambulatory BP is associated with risk of cardiovascular disease or extracranial target organ damage,[14-18] although their association with subclinical intracranial diseases remains unclear.

As the Shiga Epidemiological Study of Subclinical Atherosclerosis (SESSA) in a population-based sample of Japanese men assessed strictly measured office, self-measured home, and ambulatory BP as well as asymptomatic ICAS observed on magnetic resonance angiography (MRA), the SESSA provides a unique opportunity to investigate these issues. Using the SESSA data, we examined the association of strictly measured office, self-measured home (mean values and day-by-day variability), and ambulatory BP indices (24-hour/daytime/nighttime mean values, short-term variability, nocturnal decline, and morning pressor surge) with asymptomatic ICAS on MRA. Asymptomatic ICAS is a common cause of stroke and dementia,[1,2] and thus it is of great importance for their prevention to examine the association between various BP indices and asymptomatic ICAS on MRA.

METHODS

Study population

The SESSA design and objectives have been described previously.[3,19] In brief, from May 2006 through March 2008, we randomly selected residents of Kusatsu City, Shiga, based on the Basic Resident Registry of the city. The city, located in central Japan, has an industrial structure similar to the average of Japan according to the Ministry of Health, Labour and Welfare. We invited 2,379 Japanese men aged 40 to 79 years to participate in our study. A total of 1,094 males agreed to a baseline examination. In a follow-up examination from October 2010 through August 2014, 853 of the original participants were reassessed. Beginning in January 2012 through February 2015, SESSA participants were invited to participate in a brain magnetic resonance imaging (MRI) substudy, and 740 participants underwent 1.5-tesla MRI. We excluded 44 with a history of stroke or myocardial infarction; 6 with current or past atrial fibrillation diagnosed by medical history, serial 12-lead electrocardiography at baseline and follow-up examinations, or a 24-hour Holter recording; and 13 who had inadequate home BP data or missing covariates. The remaining 677 participants [mean age, 70.0; standard deviation (SD), 8.8] years were analyzed for the association of office or home BP with ICAS. From October 2014 through September 2015, MRI substudy participants were invited for a more extensive cardiovascular evaluation, including ambulatory BP measurement. After excluding 209 who declined to participate and 3 who had inadequate ambulatory BP data, 468 participants were included in the ambulatory BP analysis. All participants provided written informed consent. The study was approved by the Institutional Review Board of Shiga University of Medical Science.

MRI protocol and image analysis

Details for the assessment of ICAS have been previously described.[3,19] All MRI and MRA studies were obtained on an MRI scanner (Signa HDxt 1.5T, version. 16; GE Healthcare, Milwaukee, WI, USA). Three-dimensional T1-weighted spoiled gradient recalled, T2- and T2*-weighted, fluid-attenuated inversion-recovery (FLAIR), and time-of-flight MRA were done to diagnose cerebral artery stenosis and other vessel diseases. The T2-, T2*-weighted, and FLAIR images were obtained with 4-mm thickness and no inter-slice gaps. Parameters for the time-of-flight MRA sequence were: repetition time, 28 msec; echo time, 3.6 msec; flip angle, 20 degrees; field of view, 180 mm; receiver bandwidth, 17.9 kHz; frequency 320 x phase 192; and number of excitations, 1. Two neurosurgeons (KN, AS), certified by the Japan Neurosurgery Society, independently assessed all MRA/MRI images in duplicate without knowledge of the participants’ characteristics. Disagreement was resolved by consensus between them. In assessing ICAS, 11 intracranial arteries (basilar artery plus the following 5 vessels bilaterally: intracranial segments of the internal carotid artery, middle cerebral artery (M1–M3 segments), anterior cerebral artery (A1–A3 segments), intracranial segments of the vertebral artery, and posterior cerebral artery (P1–P3 segments) were evaluated. For each artery, the ordinal degree of narrowing was graded as no detectable stenosis, 1% to <50% stenosis, and 50%–100% stenosis using criteria established in the Warfarin–Aspirin Symptomatic Intracranial Disease trial.[20] Mild ICAS was defined as 1% to <50% stenosis, severe ICAS as ≥50% stenosis, and any ICAS as either mild or severe ICAS.[3,19] The most stenotic segment of each vessel was analyzed and represented the degree of stenosis in each participant.

Office BP measurement

In the morning, a trained physician placed an appropriately sized cuff on the right arm of each participant. Participants were instructed to rest alone for 5 min, in a sitting position, in a silent room, without crossing the legs or speaking. The physician then returned and turned on an automated device (BP-8800SF, Omron Healthcare Co. Ltd., Kyoto, Japan)[21] to begin measuring BP. The physician was completely out of the room before office BP measurement to minimize the white-coat effect. Office BP was measured twice consecutively with an interval of 30 seconds; the mean of the 2 BP readings was used. Indeed, we previously demonstrated in the SESSA cohort[22] high correlation coefficients for correlation between office and home BP (r=0.74; 95% CI, 0.71-0.77 for systolic BP; r=0.74; 95% CI, 0.71-0.77 for diastolic BP); and we found overall agreement from low to high BP levels with no evidence of systematic difference (mean difference [standard deviation] between office and home BP [home minus office BP], 0.47 [13.48] mmHg for systolic BP; 0.77 [7.88] mmHg for diastolic BP).

Home BP measurement

Participants were asked to measure BP themselves at home using another validated automated device (HEM-705 IT Fuzzy Cuff, Omron Healthcare Co. Ltd.)[23] once in the morning during 7 consecutive days. The instructions for measuring home BP with the device were given to the participants by trained research staff on the same day as the office BP measurements. Home BP was measured in a seated position after a 2-min rest, within an hour after waking up, after urination, and before breakfast. We included participants who conducted home BP measurement for at least 5 consecutive days.[24]

The mean of self-measured BP at home was calculated by averaging all readings for each participant. We used the coefficient of variation (CV) as an index of day-by-day BP variability. CV was calculated by dividing the within-individual standard deviation by the within-individual mean. Consequently, CV is less influenced by BP level and is therefore considered an applicable index in variability studies.[17,18] In the present study, we therefore used CV as our main exposure variable because it can be calculated relatively easily in clinical practice, and a universal reference frame can be defined.[17]

Ambulatory BP measurement

Ambulatory BP monitoring was performed with an appropriately sized cuff on the non-dominant arm using a fully automatic cuff-oscillometric method device (FM 800; Fukuda Denshi, Tokyo, Japan). The device has been validated and meets the criteria of the Association for the Advancement of Medical Instrumentation or the European Society of Hypertension.[25] The device was set to obtain BP readings every 30 min during the day and every 60 min during the night. BP was measured every 60 min throughout an entire day when arm pain or numbness developed because of frequent measurement. Artifactual readings on ambulatory BP were defined according to the criteria previously published from the Ohasama Study[26] and omitted from the analysis: systolic BP <60 mmHg and mean BP <40 mmHg; systolic BP >250 mmHg and/or mean BP >200 mmHg with no similar preceding or subsequent respective value; pulse pressure ≤10 mmHg; and abrupt increase or decrease in systolic and/or mean BP, pulse pressure, and/or heart rate by ≥50% from the value immediately before or after respective readings. No participants were excluded because of the criteria; 1.6 readings on average were excluded for each participant. We analyzed ambulatory BP readings only if, after editing for artefacts, the valid readings were at least 70% of the expected readings.[27]

The mean BP obtained by ambulatory BP assessment was calculated for a 24-hour period and separately for daytime and nighttime periods, which were determined using individual diary reports of actual awake and sleep times. We also used the CV as an index of ambulatory daytime or nighttime BP variability. The percentage of nocturnal decline in BP was calculated as follows: (daytime BP [minus] nighttime BP) ×100/daytime BP.[14] The amplitude of the morning pressor surge was calculated as follows: 2-hour average of BP after waking [minus] 2-hour average of BP before waking.[14]

Covariate assessment

Blood specimens were obtained early in the clinic visit after a 12-hour fast and used for laboratory testing, including lipid and glucose concentrations, at a single laboratory (Shiga Laboratory; MEDIC, Shiga, Japan). Lipid measurements were standardized annually according to the protocol of the Centers for Disease Control and Prevention/Cholesterol Reference Method Laboratory Network. Total cholesterol and triglycerides levels were measured using enzymatic assays, and high-density lipoprotein cholesterol levels were measured using a direct method. The Friedewald formula was used to calculate low-density lipoprotein cholesterol when triglycerides concentration was less than 400 mg/dL. Plasma glucose levels were determined by sodium fluoride-treated plasma using the hexokinase glucose-6 phosphate-dehydrogenase enzymatic assay. Serum creatinine levels were measured using an enzymatic assay (Espa CRE-liquid II; NIPRO, Osaka, Japan). The estimated glomerular filtration rate (mL/min/1.73 m²) was calculated using serum creatinine levels for Japanese men:[28] 194 × serum creatinine (mg/dL)^−1.094 × age^−0.287.

A self-administered questionnaire was used to obtain information on demographics, smoking habits, alcohol drinking, physical activity, and medication use and history. After the participants had completed the questionnaires, trained nurses confirmed all responses with the participants. Body mass index was calculated as weight (kg) divided by height squared (m²). The number of exercise days per week was calculated based on how many days per week participants regularly engaged in either brisk walking or more active exercise for ≥1 hour.[29] Dyslipidemia was defined as low-density lipoprotein cholesterol ≥140 mg/dL, triglycerides ≥150 mg/dL, high-density lipoprotein cholesterol <40 mg/dL, or lipid-lowering medication use.[30] Diabetes mellitus was defined as fasting blood glucose ≥126 mg/dL or antidiabetic medication use. Chronic kidney disease was defined as estimated glomerular filtration rate <60 mL/min/1.73 m².

Statistical analysis

The differences in characteristics across ICAS categories (none, mild, severe) were compared using linear regression for continuous variables or χ² test for categorical variables. We used a Poisson regression with robust error variance[31] to estimate relative risks and 95% confidence intervals per 1-SD increment of each BP index for the presence of any or severe ICAS. Because the prevalence of any ICAS was >10% in the cohort, odds ratios could not be interpreted as relative risks. For the direct comparison of office, home, and ambulatory BP, heterogeneity of the associations was examined by adding interaction terms (BP values × type of BP measurements [office, home, or 24-hour/daytime/nighttime ambulatory BP]) to the statistical models.[22] We repeated the comparison analysis using the same number of measurements (2 measurements): we used the mean of home BP values in the morning during the first 2 days and that of ambulatory BP values during the first 2 measurements after waking up. We also compared relative risks and 95% confidence intervals per 10 mmHg increment of office, home, and ambulatory BP for the presence of any or severe ICAS. A multivariable model was adjusted for age, smoking status (current, former, or never), alcohol drinking (yes/no), exercise (yes/no), body mass index, dyslipidemia (yes/no), diabetes mellitus (yes/no), estimated glomerular filtration rate, and antihypertensive medication use (yes/no). We further adjusted for home or daytime/nighttime mean BP levels in the analysis involving day-by-day or ambulatory daytime/nighttime BP variability, respectively, and for 24-hour mean BP in the analysis involving nocturnal decline or morning pressor surge in BP. These covariates were selected a priori because they correlate with BP indices and are risk factors for ischemic stroke or ICAS. We repeated analyses to assess the associations of BP indices with the severity (0, 1% to <50%, ≥50% stenosis) or number of lesions (0, 1, ≥2 lesions) of ICAS using the ordinal logistic regression. For a sensitivity analysis, we conducted a stratified analysis by antihypertensive medication use; we further tested for multiplicative interactions between antihypertensive medication use and BP indices in relation to ICAS. Analyses were performed using a statistical program (STATA, version 16.0; StataCorp LP, College Station, TX, USA). Two-tailed P-values <0.05 were considered statistically significant.

RESULTS

We observed mild and severe ICAS in 153 (22.6%) and 36 (5.3%) participants, respectively. We also observed single and multiple lesions of ICAS in 90 (13.3%) and 99 (14.6%) participants, respectively. The characteristics of participants with no, mild, and severe ICAS are shown in Table 1. Those with a higher degree of ICAS were older and had higher prevalence of dyslipidemia, diabetes mellitus, and antihypertensive medication use and lower estimated glomerular filtration rate. For BP indices, those with a higher degree of ICAS had higher levels of systolic BP at office and home, and during 24-hour, daytime, and nighttime ambulatory BP monitoring and higher diastolic morning pressor surge.

Table 1.

Characteristics of participants according to the ICAS status, SESSA, Shiga, Japan 2010-2015

	ICAS			P
	No n=488	Mild n=153	Severe n=36
Age, years	68.7 (9.0)	71.1 (8.1)	75.3 (6.4)	<0.001
Smoking status, %
Current	20.7	17.0	19.4	0.591
Former	57.6	64.7	55.6
Alcohol drinker, %	83.0	75.8	75.0	0.093
Exercise, %a	56.6	54.9	75.0	0.079
Body mass index, kg/m2	23.2 (2.9)	23.6 (2.8)	23.4 (2.9)	0.341
Dyslipidemia, %b	51.6	61.4	75.0	0.005
Diabetes mellitus, %c	16.0	22.2	36.1	0.004
Estimated glomerular filtration rate, mL/min/1.73m2	70.3 (13.4)	68.7 (15.0)	63.5 (11.3)	0.011
Chronic kidney disease, %d	21.3	26.2	36.1	0.077
Antihypertensive medication, %	32.2	47.7	55.6	<0.001
Office BP, mmHg
Systolic	129.5 (16.0)	136.4 (18.3)	141.5 (15.4)	<0.001
Diastolic	76.9 (10.0)	78.0 (12.1)	74.8 (9.8)	0.219
Home BP, mmHg
Systolic	132.7 (15.9)	137.9 (16.4)	144.0 (19.8)	<0.001
Diastolic	77.9 (9.9)	78.3 (10.1)	74.6 (12.6)	0.132
Ambulatory BP, mmHg
24-hour systolic	121.3 (12.9)	127.0 (12.9)	130.4 (16.4)	<0.001
24-hour diastolic	75.9 (7.9)	77.4 (7.9)	75.0 (9.9)	0.210
Daytime systolic	125.4 (13.4)	131.4 (13.6)	135.9 (16.3)	<0.001
Daytime diastolic	78.6 (8.4)	80.2 (8.5)	77.8 (9.1)	0.203
Nighttime systolic	113.6 (14.6)	117.7 (15.6)	118.6 (16.7)	0.030
Nighttime diastolic	70.6 (8.9)	71.6 (9.3)	68.9 (12.8)	0.430
BP variability, %e
Home day-by-day systolic	5.9 (2.4)	6.2 (2.3)	6.5 (2.2)	0.538
Home day-by-day diastolic	5.8 (2.7)	5.8 (3.0)	5.9 (3.0)	0.996
Ambulatory daytime systolic	10.1 (3.1)	10.4 (3.4)	10.6 (3.5)	0.676
Ambulatory daytime diastolic	12.1 (3.7)	12.1 (3.7)	12.5 (4.2)	0.939
Ambulatory nighttime systolic	8.9 (2.9)	9.0 (3.1)	10.4 (4.6)	0.131
Ambulatory nighttime diastolic	11.2 (4.3)	11.8 (4.8)	13.2 (5.1)	0.172
Nocturnal decline, %f
Systolic	9.3 (8.0)	10.2 (10.1)	12.6 (8.7)	0.221
Diastolic	10.0 (8.9)	10.5 (9.9)	11.8 (10.5)	0.672
Morning pressor surge, mmHgg
Systolic	9.2 (12.7)	11.2 (16.1)	13.6 (12.0)	0.232
Diastolic	6.3 (14.6)	9.9 (15.4)	14.7 (9.0)	0.012

Data are presented as mean (standard deviation) unless otherwise specified. No, mild, and severe ICAS were observed in 353 (75.4%), 99 (21.2%), and 16 (3.4%) participants, respectively, who underwent ambulatory BP measurement.

^aParticipants who exercised were defined as those who regularly engaged in either brisk walking or more active exercise for ≥1 hour/week.

^bDyslipidemia was defined as low-density lipoprotein cholesterol ≥140 mg/dL, triglycerides ≥150 mg/dL, high-density lipoprotein cholesterol <40 mg/dL, or lipid-lowering medication use.

^cDiabetes mellitus was defined as fasting blood glucose ≥126 mg/dL or antidiabetic medication use.

^dChronic kidney disease was defined as estimated glomerular filtration rate <60 mL/min/1.73 m².

^eAn index of day-by-day or short-term BP variability was defined as a coefficient of variation that was calculated by dividing the within-individual standard deviation by the within-individual mean.

^fThe percentage of nocturnal decline in BP was calculated as follows: (daytime BP [minus] nighttime BP) ×100/daytime BP.

^gThe amplitude of the morning pressor surge was calculated as follows: 2-hour average of BP after waking [minus] 2-hour average of BP before waking.

BP, blood pressure; ICAS, intracranial artery stenosis; SESSA, Shiga Epidemiological Study of Subclinical Atherosclerosis.

Comparison between strictly measured office, self-measured home, and ambulatory BP

The association of strictly measured office, self-measured home, or ambulatory systolic BP levels with the presence of any or severe ICAS is shown in Figure 1. In Poisson regression with robust error variance after adjustment for potential confounding covariates, including age and other cardiovascular risk factors, those with higher levels of systolic BP at office or home, or during 24-hour, daytime, or nighttime ambulatory BP monitoring had greater prevalence of any ICAS. Higher levels of office, home, or daytime ambulatory systolic BP were also independently associated with the presence of severe ICAS. The associations with any or severe ICAS were comparable between office, home, and ambulatory systolic BP (all P-values for heterogeneity >0.1). When using mean values of the first 2 days of home BP and of the first 2 ambulatory BP measurements after waking up, we found almost the same associations with ICAS between strictly measured office, self-measured home, or ambulatory BP indices (all P-values for heterogeneity >0.3) (Supplementary Figure S1). We also found similar findings in the comparison of relative risks and 95% confidence intervals per 10 mmHg increment of office, home, or ambulatory BP for the presence of any or severe ICAS (Supplementary Figure S2). For diastolic BP, only daytime ambulatory BP levels were associated with the presence of any ICAS, and no significant differences in relation to the presence of ICAS were found between office, home, and ambulatory BP (all P-values for heterogeneity >0.4) (Supplementary Figure S3).

Figure 1. The association of strictly measured office, self-measured home, or ambulatory systolic BP with the presence of (A) any ICAS or (B) severe ICAS.

Solid squares and horizontal lines indicate point estimates and 95% confidence intervals for the presence of ICAS, respectively. Relative risks are shown with a 1-standard deviation increment in office (16.9 mmHg), home (16.5 mmHg), and ambulatory 24-hour (13.3 mmHg), daytime (13.9 mmHg), and nighttime (15.0 mmHg) systolic BP. Data are adjusted for age, smoking status, alcohol drinking, exercise, body mass index, dyslipidemia, diabetes mellitus, estimated glomerular filtration rate, and antihypertensive medication use.

BP, blood pressure; ICAS, intracranial artery stenosis.

BP variability indices and ICAS

The association of BP variability indices with the presence of any or severe ICAS is shown in Figure 2. After adjustment for possible confounding factors and mean home systolic BP levels, there was no significant association between day-by-day BP variability and ICAS. Whereas, after adjustment for mean 24-hour ambulatory systolic BP in addition to possible confounders, those with greater magnitude of nocturnal decline or morning pressor surge in systolic BP had higher burden of any or severe ICAS. For diastolic BP, neither day-by-day nor short-term variability was independently associated with the presence of ICAS (Supplementary Figure S4). We observed similar results when we assessed the associations with the severity or number of lesions of ICAS using the ordinal logistic regression (Supplementary Figures S5 and S6). The associations with ICAS of mean systolic BP levels were similar between participants with and without antihypertensive medication (all P-values for heterogeneity >0.2) (Supplementary Figure S7). Indeed, the results adjusted with antihypertensive medication were similar to those performed in participants without antihypertensive medication. Whereas, the association with ICAS of ambulatory nighttime variability and nocturnal decline in systolic BP were more pronounced among participants not taking antihypertensive medication (P-values for heterogeneity, 0.043 and 0.023, respectively).

Figure 2. The association of day-by-day or short-term variability, nocturnal decline, or morning pressor surge in systolic BP with the presence of (A) any ICAS or (B) severe ICAS.

Solid squares and horizontal lines indicate point estimates and 95% confidence intervals for the presence of ICAS, respectively. Relative risks are shown with a 1-standard deviation increment in coefficient of variation of day-by-day home (2.4%) and of ambulatory daytime (3.2%) and nighttime (3.0%) systolic BP, and with that in nocturnal decline (8.5%) and morning pressor surge (13.5 mmHg) of systolic BP. Data are adjusted for age, smoking status, alcohol drinking, exercise, body mass index, dyslipidemia, diabetes mellitus, estimated glomerular filtration rate, and antihypertensive medication use. Analysis involving BP variability was further adjusted for home/daytime/nighttime mean systolic BP. Analysis involving nocturnal decline or morning pressor surge was further adjusted for 24-hour mean systolic BP.

BP, blood pressure; ICAS, intracranial artery stenosis.

DISCUSSION

In this population-based study of apparently healthy middle-aged to older men, higher levels of strictly measured office, self-measured home, or ambulatory systolic BP were associated with the presence of any or severe ICAS on MRA, after adjustment for other cardiovascular risk factors. The degrees of the associations of office, home, and ambulatory BP with ICAS were almost comparable. Additionally, independent of mean systolic BP, greater magnitude of nocturnal decline or morning pressor surge in systolic BP was associated with higher burden of ICAS.

Elevated BP and asymptomatic ICAS in the context of current literature

Our findings support previous observation demonstrating the association between elevated BP levels at a medical office and subclinical ICAS.[6,32-35] However, the majority of those studies were based on ICAS assessed by transcranial Doppler ultrasonography, which has limited accuracy and reproducibility because of its heavy reliance on operator skill and the difficulty evaluating vessels other than the middle cerebral artery.[36] MRA, in contrast, has better accuracy than transcranial Doppler for the evaluation of ICAS.[36] Furthermore, few studies have investigated the relationship between BP levels outside the medical office and asymptomatic ICAS. In a patient-based study of 757 patients with hypertension who had at least 2 cardiovascular risk factors (mean age, 60 years) by Chen et al.,[35] systolic BP in daytime or nighttime derived from 24-hour ambulatory BP monitoring was associated with the presence of ICAS diagnosed with computerized tomographic angiography.

Direct comparison between office, home, and ambulatory BP for target organ damage

A key finding in our study is that office BP measured in an optimal setting with an automated device, as recommended by clinical guidelines,[8-10] had an almost similar association with asymptomatic ICAS as that of self-measured home or ambulatory BP. This is in line with our prior report comparing office and home BP related to coronary artery calcium.[22] The major advantage of out-of-office BP measurement is that it provides a large number of BP measurements and minimization of the white-coat effect and observer bias, representing a highly reliable assessment of actual BP.[11,12] In fact, a number of studies have shown that home or ambulatory BP is more strongly associated with cardiovascular disease than BP measured in the office setting.[11-13] However, our serial findings suggest that the usefulness of office BP measured under a strict protocol may be comparable with that of home or ambulatory BP in the detection of target organ damage in coronary or intracranial arteries. Additionally, further studies are warranted to investigate whether strictly measured office BP is superior to usually measured office BP in the association with target organ damage or cardiovascular disease risk.

There have been a few direct comparisons between ambulatory BP monitoring and self-measured home BP measurements with respect to the association with target organ damage, but the results remained controversial. In middle-aged to older Japanese (mean age, 66 years) from the Ohasama Study,[37] the association with silent lacunar infarction or white matter hyperintensity of ambulatory BP was stronger than that of home BP while the association of carotid atherosclerosis of ambulatory BP was weaker than that of home BP. For the magnitude of the association with left ventricular hypertrophy, ambulatory BP was comparable to home BP,[38] superior to home BP,[39] or vice versa.[40] For different markers of target organ damage, including urinary albumin excretion[38] and carotid intima-media thickening,[38] it showed a comparable degree of association for ambulatory and home BP. To date, few data have been available with respect to the association of ambulatory or home BP values with asymptomatic ICAS on MRA. We found the associations with asymptomatic ICAS were almost comparable between ambulatory and self-measured home BP.

Day-by-day or short-term BP variability and asymptomatic ICAS

Our study provides the first evaluation of the relationship of the day-by-day or short-term BP variability to subclinical intracranial diseases such as asymptomatic ICAS. Our findings are in line with a series of studies examining the association of home or ambulatory BP variability with extracranial target organ damage such as carotid atherosclerosis,[15,16,18] left ventricular hypertrophy,[16] microalbuminuria,[16] aortic calcification, and ankle–brachial index.[18] In our study, greater nocturnal dipping and/or morning pressor surge in BP were associated with higher burden of ICAS. The finding could be explained by the direct association between the BP dip from day to night and the BP surge: the greater the day-night dip, the greater the morning BP surge and vice versa.[41] It is reasonable to suppose that an excessive BP dip and/or BP surge might develop atherosclerotic plaques in intracranial arteries through some hemodynamic mechanism, such as increased shear stress, that, in turn, may activate the molecular pathway favoring plaque instability, such as increased oxidative stress, ubiquitin proteasome activity, and inflammation.[42,43] Interestingly, our findings in individuals with a mean age of 70 years are partly in line with a recent report demonstrating an association of extreme nocturnal BP dipping with cardiovascular disease risk among older people aged ≥70 years[44] because of subclinical ICAS predisposing to clinical stroke later in life.

We found that the association for ICAS of mean BP was similar but that of short-term BP variability was possibly modified by antihypertensive medication use: those without antihypertensives were particularly susceptible to the impact of nighttime variability or nocturnal decline in BP. Similar findings were reported in other studies in which the relation between BP variability and cardiovascular diseases was stronger in those with relatively low cardiovascular risks (e.g., young age or low BP levels).[45,46] In our study, in fact, participants without antihypertensive medication had better cardiovascular profiles including age, body mass index, dyslipidemia, diabetes mellitus, chronic kidney disease, and BP than those on antihypertensives (Supplementary Table S1). Therefore, short-term BP variability would be an important determinant, especially in individuals with low cardiovascular risk profiles, because of fewer competing causes of or greater susceptibility to target organ damage.[45,46]

Study limitations

Several limitations of the present study warrant consideration. First, we studied only middle-aged to older men and the sample was obtained from a single area in Japan, both of which may limit the generalizability of the present results. Additionally, individuals who did not undergo ambulatory BP monitoring were older and had worse profiles of cardiovascular risk factors (e.g., smoking, antihypertensive medication use, or BP levels) at the baseline examination than those who underwent this monitoring (Supplementary Table S2). Individuals who did not undergo ambulatory BP monitoring may have been at higher risk for developing asymptomatic ICAS. Therefore, we may have underestimated effects and associations. Second, the numbers of home (median [25, 75%tiles], 7 [7, 7]) or ambulatory BP (25 [24, 26]) measurements are relatively smaller than those used in previous studies. It is possible that large number of measurements may be responsible for the more pronounced association with ICAS. However, in our study, results were almost unchanged when we compared office, home, and ambulatory BP values using the same number of measurements (2 measurements) in the morning. Furthermore, although there are no standard guidelines to determine the minimum number of BP readings as well as the optimal intervals between them to reliably estimate BP variability,[47] more frequent measurement may be related to the better assessment, which could lead to the more precise association with asymptomatic ICAS. Third, because of the small sample sizes and low prevalence of severe ICAS, the present study may have been underpowered to detect significant associations especially for the analyses of ambulatory BP and severe ICAS. Finally, the difference in the period between brain MRI and ambulatory BP monitoring (median [25, 75%tiles], 11 [10, 12] months) may affect our results. However, given the increased burden of ICAS with aging, we could have confirmed a more solid association between BP indices and ICAS if the brain MRI were done in the same period as ambulatory BP monitoring.

Conclusions

The findings in the present study suggest 2 important implications. First, contrary to overwhelming evidence, for asymptomatic ICAS on MRA crucially resulting in clinical stroke or dementia, the magnitude of association of BP strictly measured at the medical office was comparable with that of BP measured outside the medical office, that is, self-measured home BP or ambulatory BP during daily life. Further prospective studies are warranted to investigate whether office BP under a strict protocol has a similar predictive value for other target organ damages or clinical cardiovascular diseases compared with self-measured home or ambulatory BP. Second, greater levels of nocturnal decline or morning pressor surge in systolic BP was independently associated with higher burden of asymptomatic ICAS. Therefore, in addition to mean BP levels, circadian BP variation based on ambulatory BP monitoring may be considered when assessing the risk of ICAS in apparently healthy individuals especially in those not taking antihypertensive medication. Therapeutic interventions are needed to investigate whether and how a higher magnitude of circadian BP variation should be treated with the purpose of reducing cardiovascular disease risks.