Download Brain Imaging Findings Predict Blood Pressure

Brain
Brain Imaging Findings Predict Blood Pressure Response to
Pharmacological Treatment
J. Richard Jennings, Matthew F. Muldoon, Ellen M. Whyte, Joelle Scanlon,
Julie Price, Carolyn C. Meltzer
Downloaded from http://hyper.ahajournals.org/ by guest on May 6, 2017
Abstract—Hypertension appears to alter brain morphology, as well as the cerebrovascular support for information
processing. Because these effects might reflect progressive effects of essential hypertension on the brain, we asked
whether structural and functional brain indices would predict the success of pharmacological treatment of hypertension
among 45 previously unmedicated individuals. After initial structural MRI and functional positron emission tomography
imaging, subjects were randomly assigned in a double-blind fashion for treatment for 1 year with either lisinopril or
atenolol. Systolic and diastolic blood pressure decreases after treatment stabilization were correlated to a pretreatment
index of brain aging (combined ratings of ventricle and sulcal size and white matter hyperintensities) and the
pretreatment change in regional cerebral blood flow during working memory in the thalamus and posterior parietal
regions of interest. In multiple regression analyses, the structural brain index and the blood flow response in the thalamus
predicted 20% of the variance in the systolic blood pressure response to treatment controlling for pretreatment blood
pressure, age, gender, and type and dose of medication. Alcohol use influenced the thalamic response measure, but
covariates did not alter the relation between greater indices of brain aging and less successful blood pressure response
to treatment. The state of the brain may be an important factor in the remediation of blood pressure. (Hypertension.
2008;52:1113-1119.)
Key Words: hypertension 䡲 MRI 䡲 positron emission imaging 䡲 atenolol 䡲 lisinopril 䡲 cerebral blood flow
䡲 blood pressure treatment
A
lthough characteristics placing a person at risk for
essential hypertension are well known, predictors of
blood pressure (BP) response to pharmacological treatment,
other than current BP, have received less attention.1,2 Once
hypertension management has begun, appropriate clinical
management and compliance with the medical regimen are
clearly factors in treatment outcome.3 In addition, different
treatment approaches may be more effective with, eg, diastolic hypertension versus isolated systolic hypertension; but
clear clinical evidence for this has yet to be established.4
Cerebrovascular factors are rarely considered as central to
disease diagnosis or treatment but are considered only if frank
cerebrovascular disease co-occurs with hypertension.5 Increasing evidence of central nervous system involvement as
either a cause or consequence of essential hypertension,6 –12
however, led us to ask whether brain morphological indices
of brain aging or regional cerebral blood flow responsivity
would predict BP decline during pharmacological treatment
of essential hypertension.
A number of factors suggest that cerebral structure and
function might relate to success of BP treatment. First,
hypertension has been characterized as accelerated aging, and
aging changes in the brain are well established.13–16 Second,
hypertension has been related to changes in both brain
structure and cerebrovascular responsivity during information
processing.6,9,17–19 Third, hypertension in mid-to-late life is
associated with mild cognitive deficits.7,11,20 These observations, which distinguish hypertensive patients from similarly
aged normotensive individuals, suggest that the brain is a
target organ for essential hypertension. The impact of the
disease on the brain may also interact with treatment efficacy.21 Treatment-induced reductions in BP require adaptation
of central nervous system blood flow regulation, an adjustment that may be more efficient in a “younger” brain.22
Indirect evidence for this suggestion is difficulty in controlling hypertension in the elderly, the increasing use of multiple
drugs, and the high prevalence of difficult-to-control hypertension.23–25 Taken together, brain indices were hypothesized
to predict the degree of success of BP lowering, because the
state of the brain may indicate both severity and progression
of hypertensive disease and capacity for return to normal
regulation of blood flow and pressure.
Based on this rationale, we examined the success of
pharmacological treatment of BP as a function of brain
Received July 21, 2008; first decision August 2, 2008; revision accepted October 9, 2008.
From the Departments of Psychiatry and Psychology (J.R.J.), Department of Medicine (M.F.M.), Department of Psychiatry (E.M.W.), and Department
of Radiology (J.S., J.P.), University of Pittsburgh, Pa; and Departments of Radiology, Neurology, and Psychiatry (C.C.M.), Emory University, Atlanta,
Ga.
Correspondence to J. Richard Jennings, E1329 WPIC, 3811 O’Hara St, Pittsburgh, PA 15213. E-mail [email protected]
© 2008 American Heart Association, Inc.
Hypertension is available at http://hyper.ahajournals.org
DOI: 10.1161/HYPERTENSIONAHA.108.120196
1113
1114
Hypertension
December 2008
structure and responsivity. Relevant results were available
from a project comparing the effects of an angiotensinconverting enzyme inhibitor and a ␤-blocker on regional
cerebral blood flow (CBF; rCBF).26 We used structural
indices of brain aging (sulcal and ventricular size, as well as
white matter hyperintensities) and used a measure of functional cerebrovascular responsivity, the rCBF response to
information processing in 2 areas previously shown sensitive
to both aging and hypertension, the posterior parietal cortex
and thalamus.9,27–29 Responsivity was examined, because
overall CBF does not seem altered by hypertension in midlife
samples.9,18 In sum, using data collected in a 1-year clinical
trial involving 45 hypertensive individuals, we tested the
hypothesis that magnetic resonance (MRI)– derived indices of
brain aging and positron emission tomography (PET) indices
of functional response to working memory predict the magnitude of BP lowering with standard pharmacological
treatment.
Downloaded from http://hyper.ahajournals.org/ by guest on May 6, 2017
Methods
Participants
Participants were a community sample recruited from the Pittsburgh
area via radio, television, newspaper, and health fairs. Participants
were between 35 and 65 years of age and had an average diastolic
(fifth phase) BP of 90 to 109 mm Hg, systolic BP of 140 to
179 mm Hg, or both. Nonelderly individuals with mild-to-moderate
hypertension were selected to enhance the likelihood that treatment
would facilitate the reversibility of any cerebrovascular dysfunction.
Seated BP was assessed after ⱖ5 minutes of rest using the ausculatory technique with a mercury manometer. Baseline BP was calculated from the average of the last 2 of 3 readings done on 2
occasions. Participants were required to have either no previous
pharmacological treatment for hypertension or ⱕ6 months of BP
medication within the past 5 years, with no BP medication taken in
the 6 months preceding enrollment. Detailed inclusion/exclusion
criteria for this study have been presented previously.26 Screening
was designed to exclude secondary hypertension, use of drugs/
substances interfering with accurate and safe treatment/assessment,
and the presence of other serious disease, notably coronary or
cerebrovascular disease. The University of Pittsburgh Institutional
Review Board approved all of the procedures as consistent with
ethical principles, and subjects provided informed consent.
Design
Initial BP assessment, medical screening and history, physical
examination, administration of self-report questionnaires, and detailed consent followed telephone screening. The next session
included the second BP assessment and neuropsychological examination. Separate sessions followed for brachial artery ultrasound,
MRI, and PET examinations. All of the examinations were repeated
after 1 year, except for the screening and self-report instruments.
Details of psychological and neuropsychological testing and results
will be presented elsewhere. This report focused on 43 of the 45
participants who completed the entire study. Participants were
similar to noncompleting individuals in age, education, and personality factors; they differed in that continuing participants were
significantly (␹2 P⬍0.05) more likely to be male, white, and married.
Medication Procedures
Patients were treated for 1 year within a randomized, double-blind
design using either lisinopril or atenolol, a choice of drugs based on
the initial hypothesis that lisinopril would alter brain function more
than atenolol.9 Medications were contained in unembossed capsules
to ensure blinding. During a 6-week titration phase, drug dosage was
gradually increased across 4 dosage levels: 10, 20, 30, and 40 mg for
lisinopril and 25, 50, 75, and 100 mg for atenolol. During this phase,
participants were seen every 2 weeks for dosage adjustment and
determination of BP, compliance, and adverse effects. Upward
titration stopped if a participant’s BP had fallen to ⱕ135/85 mm Hg
or if resting pulse fell to ⬍50 bpm. If a participant’s BP remained
⬎140/90 mm Hg on the full-dose treatment, 12.5 mg of hydrochlorothiazide was added (4 in lisinopril group and 7 in ␤-blocker group).
Participants were withdrawn from the study if BP averaged ⬎160/
95 mm Hg on 2 consecutive visits during the maintenance phase
(week 6 through the final visit at week 52). During visits, physical
symptoms and impact of hypertension on quality of life were
assessed with the Bulpitt Hypertension Questionnaire,30,31 and adherence was estimated based on returned pill counts. The BP
treatment effect was calculated from standardized BP assessment
performed during the last regular clinical visit (approximately week
40). Estimates from final readings during the PET session (week 52)
were deemed unrepresentative given the stress of that session.
Structural MRI Measures
All of the subjects underwent a structural MRI using a GE Signa 1.5
T scanner to provide a detailed image for mapping PET results and
for assessment of white matter hyperintensities. A brief scout
T1-weighted image was followed by an axial series oriented to the
plane along anterior and posterior commissures: fast spin echo T2weighted (effective echo time/repetition time: 102/2500 milliseconds, 1
excitation), proton density–weighted (effective echo time/repetition
time: 17/2000 milliseconds, 1 excitation), and fast fluid-attenuated
inversion recovery (FLAIR; effective echo time/repetition time: 56/
9002 milliseconds, time to inversion: 2200 milliseconds, 1 excitation).32 Section thickness was 5 mm with a 1-mm intersection gap. A
field of view of 24 cm and image matrix of 256⫻192 pixels were
used for all of the axial magnetic resonance series. A volumetric
spoiled gradient-recalled sequence with parameters optimized for
maximal contrast among gray matter, white matter, and spinal fluid
was acquired for purposes of region of interest (ROI) placement in
the coronal plane (echo time/repetition time: 5/25 milliseconds, flip
angle 40°, 1 excitation, section thickness 1.5 mm, 0-mm intersection). Severity of white matter hyperintensities, ventricle size, and
sulcal size were graded using a 10-point scale (range of these
summary measures: 0 to 27) by 2 trained raters using standards
developed for the Cardiovascular Health Study.33 Percentage agreement on the ratings done from T1, T2, and proton density images was
93% and for FLAIR images was 94%. Participants showing evidence
of significant lacunar or other infarcts (n⫽2) were excluded from
participation. These were identified from the MRI films by a
board-certified neuroradiologist. Significant infarcts were those
judged to potentially influence functional images and/or to influence
cognitive processing.
Two summary measures of brain aging were created. Based on
ratings of T2/T1/proton measures, we combined ventricle size, sulcal
size, and white matter load (correlations between individual scores
and index, r⫽0.69 to 0.84). Based on the FLAIR images, we
combined overall brain ratings of ventricle size, of sulcal size, of
periventricular white matter load, and of subcortical white matter
load (correlations between individual scores and index, r⫽0.67 to
0.86). Ratings were scaled identically (0 to 9), so ratings from each
category were simply added together (combined with equal weighting). Because of technical issues, 1 fewer patient had FLAIR indices
available relative to the T2/T1 index. We only present FLAIR-based
rating data, however, because this modality has greater specificity for
white matter abnormalities. Both summary indices, however, related
very similarly to BP results.
Functional PET Measures
Four quantitative (arterial sampling) PET scans were performed that
tested rCBF responses during information processing (working
memory tasks). Two scans examined a control state and 2 the active
memory state; task order was randomized. Each task used the same
display and 2 button responses. The control memory task required
the subject to respond with the thumb to any letter appearing on the
left of the screen and with the index finger for a letter appearing on
the right. The 2-back memory task required the subject to respond
Jennings et al
Downloaded from http://hyper.ahajournals.org/ by guest on May 6, 2017
with the thumb if spatial position of the letter currently appearing
matched the spatial position, which appeared 2 times back, and the
index finger if it did not. Each task lasted 5 minutes and started 2
minutes before tracer injection. An intertask interval of ⬇5 minutes
rested the subject and permitted time for preparation and delivery of
the tracer. See our previous article for further description.26 PET
scans were acquired, in 3D imaging mode, using a Siemens/CTI
ECAT HR⫹PET scanner (63 transaxial planes, 2.4-mm section
thickness). PET data were corrected for radioactive decay, photon
attenuation, and scatter.34 Data were reconstructed using filtered
backprojection. The final image resolution was ⬇6 mm (transverse
and axial). Head motion was minimized using a thermoplastic mask.
A peripheral vein catheter was inserted for tracer injections and a
radial artery catheter for arterial blood sampling. An initial 10minute transmission scan was performed using rotating 68Ge/68Ga
rods (attenuation correction). Nine millicuries of [15O]water were
injected as a rapid bolus using an automated injector system. A
180-second emission scan was acquired in 20 sequential frames
(10⫻3 seconds, 3⫻10 seconds, 4⫻15 seconds, and 3⫻20 seconds),
followed by a 5-minute rest period for a total of 8 minutes between
injections. Through calibration measurements, both blood and PET
brain time-activity data were converted to microcuries per milliliter.
All of the PET images were corrected for small head movements
using Automated Image Registration,35 which was also used in a
modified form to align and center the PET data.36
Quantitative CBF
The [15O]water data were analyzed using a 1-tissue compartment
model.37,38 Model parameters corresponded with the clearance of
water from blood to brain (K1; mL/min per milliliter), brain-to-blood
transfer (k2; min⫺1), and an arterial input function timing delay (⌬t).
The 3 parameters were simultaneously determined using an iterative
least-squares curve fitting on a regional basis. K1 is a clearance
parameter that is directly proportional to flow, and K1 was used as a
blood flow measure (mL/min per milliliter).
ROIs were defined based on areas of activation that differed
between patients and normotensive controls in a previous/separate
study.9 The ROIs were regionally defined on the spoiled gradientrecalled MRI for each participant and manually drawn using a
version of the Imagetool software (CTI PET Systems). ROI placement is illustrated in a previous online supplement.26 For current
purposes, we examined the response to 2-back memory relative to
control in the thalamic and posterior parietal ROIs, because these
differed between hypertensive patients and controls in our previous
work.9
Analysis
We tested the hypothesis that the brain aging index and rCBF
response to working memory in thalamic or posterior parietal ROIs
would predict treatment outcome using ordinary least-squares multiple regression (StatSoft, Inc). A hierarchical regression analysis
was performed. In step 1 (basic model), age, drug assignment
(lisinopril versus atenolol), and pretreatment systolic BP were used
to predict change in systolic BP with treatment. Step 2 added
variables to see whether they further explained variance in treatment
response after adjustment for the basic model. Separate step 2
models were tested for parietal response during working memory,
thalamic response during working memory, and brain aging index.
Brain factors that significantly improved on the variance explained
by the basic model were then entered together in a separate step 2
model to provide an estimate of overall prediction of BP treatment
success using brain indices. Finally, covariates (see Table 1) were
added to this model to see whether they altered the predictivity of
brain factors. A covariate for dosage levels was used with a score
corresponding to the dose level achieved at the end of treatment (1
to 4 for lisinopril or atenolol dose level and a 5 if hydrochlorothiazide was added). Because of the relatively small sample size,
covariates were not added concurrently. The same analytic approach
was repeated with diastolic BP and pulse pressure change during
treatment as the dependent variable.
Brain Aging and Blood Pressure Lowering
1115
Results
Participant Characteristics
Figure 1 diagrams selection and retention characteristics of
the sample. Medication was well tolerated, but a substantial
number of patients were excluded after initial consent because of our medical exclusions (see above) and because of
their loss of interest before treatment but after further consideration of the demands of the study.
Table 1 provides details of subject characteristics. To
provide evidence of the variability in treatment response and
to allow examination of factors related to treatment response,
the sample was divided by median systolic BP response. Most
baseline characteristics and medication type (lisinopril or
atenolol) did not significantly influence BP response; however, higher initial systolic BP was associated with a greater
decrease in systolic BP with treatment.
Overall, patients showed values of white matter and rCBF
response indices comparable to previous work. Ratings from
the FLAIR images showed mean values (SEs) for ventricle
size of 2.30 (0.14), sulcal size of 2.70 (0.15), periventricular
white matter of 1.40 (0.13), and subcortical white matter of
1.50 (0.27). These values are comparable to those from our
previous study9 and appear appropriate for age in comparison
with epidemiological data.33 Similarly, values observed were
somewhat higher but comparable to ours and others for mean
rCBF in the thalamic area (63.8 [(2.1] mL/min per milliliter)
and response in this area to the 2-back memory task (3.9 [0.9]
mL/min per milliliter).9,39,40
Association With Treatment Response
MRI indices of brain aging, as well as the responsivity of the
thalamic ROI during PET scanning, were associated with a
subsequent degree of BP lowering with treatment. Table 2
shows the results of the hierarchical multiple regression
analyses. Step 1 in the analysis tested a basic model that
included age, medication type, and initial BP (factors typically assumed to alter treatment response). The table shows
the standardized ␤ weights for the predictors as related to the
change in BP in millimeters of mercury. For example,
pretreatment BP shows a negative ␤ weight, indicating that
the higher initial pressure was related to a greater decrease in
BP with treatment. Step 2 analyses then asked if further
variance in treatment outcome could be explained by brain
variables. Step 2a added the brain aging index. Step 2b added
only the responsivity in the thalamic ROI. Step 2c added the
brain aging index and the thalamic responsivity concurrently.
The increment in variance explained associated with each
step 2 analysis is shown in the far right column. Addition of
the parietal ROI is not shown, because responsivity in this
area failed to relate to BP treatment response. Note that the
“n” with different analyses differs somewhat, eg, the “n” for
the PET response is decreased because of refusal of the
arterial catheter by 4 participants. The step 2c model with
both the FLAIR index and thalamic responsivity predicted
60% of the variance in SBP treatment outcome. Brain indices
added 20% to the prediction of the basic model. The regression results in Table 2 are supplemented by Figure 2, which
1116
Hypertension
December 2008
Table 1. Comparison of the Characteristics of Hypertensive Patients Who
Showed Above and Below Median BP Changes With Complete Treatment:
Treatment Response of Systolic BP
Above Median Decrease
(n⫽22)
Below Median Decrease
(n⫽21)
Age, y
53.7 (4.8)
51.3 (8.2)
BMI
29.9 (5.4)
30.1 (4.2)
Heart rate, bpm
73.0 (9.8)
69.5 (7.0)
Qualifying systolic, mm Hg
150.0 (7.2)
145.0 (2.9)
Qualifying diastolic, mm Hg
94.0 (9.6)
97.0 (6.6)
Gender, proportion male
0.68
0.86
Race, proportion white
0.90
0.86
Subject Characteristic
Nicotine use, mean No. of cigarettes
per week
Downloaded from http://hyper.ahajournals.org/ by guest on May 6, 2017
0.09 (0.42)
0.43 (0.84)
Alcohol use, mean No. of drinks per week
2.0 (0.98)
2.2 (1.5)
Dosage level, low⫽1, high plus
hydrochlorothiazide⫽5
2.8 (1.5)
2.1 (1.3)
Percent compliance*
98 (3)
99 (2)
SBP change in ACE-I treated, mm Hg
⫺25.0 (3.9)
⫺4.0 (10.4)†
SBP change in ␤-blocker treated, mm Hg
⫺26.0 (9.3)
⫺8.0 (8.7)†
Total cholesterol, mg/dL
214 (34)
209 (33)
High-density lipoprotein, mg/dL
60 (43)
56 (31)
Low-density lipoprotein, mg/dL
135 (32)
129 (33)
Triglycerides, mg/dL
152 (138)
168 (126)
Data are mean (SE) unless otherwise specified. No significant difference (P⬍0.05) was found
between groups by t test. No differences were statistically significant for proportions. Median
posttreatment systolic BP was 123 mm Hg, with patients separated by the median change in systolic
BP of ⫺18 mm Hg. BMI indicates body mass index; SBP, systolic BP.
*Compliance was assessed by pill counts at the time of clinical visits.
†Group differences were necessarily significant by definition, but no differences were found
between medication groups by t test or 2-way ANOVA.
illustrates the basic bivariate relationship between the systolic
BP treatment outcome and brain aging.
The combined model, ie, brain aging and thalamic responsivity, together with the factors in the basic model, was
further tested using diastolic BP and pulse pressure rather
than systolic BP as the dependent variable. The results for
diastolic BP completely paralleled the results for systolic BP
shown in Table 2. Statistical significance at P⬍0.05 or better
was consistent with the systolic BP results throughout. The
combined model showed significant standardized ␤s for
initial diastolic pressure (⫺0.429, t⫽⫺2.98, P⬍0.001),
FLAIR aging index (0.374, t⫽2.19, P⫽0.03), and rCBF
change in the thalamic ROI (⫺0.351, t⫽2.42, P⫽02). Results
for pulse pressure remained significant only for the FLAIR
Consented
N=81
Withdrew N=36
Randomize
N=45
N=14
N=18
No Longer
Interested
Medical
Exclusion
Lisinopril
N=21
N=4
Poor BP Control N=1
Claustrophobia
Atenolol
N=24
N=1
N=43
Completed 1
year of
treatment
Figure 1. Patient retention and study
design.
Jennings et al
Brain Aging and Blood Pressure Lowering
1117
Table 2. Multiple Regression Results for Base Model and for Additions to the Model of Brain Aging Indices and rCBF Response to
Working Memory in the Thalamic ROI
␤ Weight (Standardized)
Factor
Standard Error of ␤
t Value
P
r 2 Increment
Base model (step 1) N⫽43 (r 2⫽0.39)
Age
0.240
0.131
1.84
Drug group
0.068
0.123
0.24
⫺0.660
0.130
⫺5.09
⬍0.001
0.428
0.143
3.00
0.005
0.11
⫺0.303
0.128
⫺2.36
0.02
0.09
Initial systolic BP
0.07
Ns
Base model plus FLAIR aging (step 2a) N⫽43 (r 2⫽0.50)
FLAIR aging
Base model plus PET rCBF ROI response (step 2b) N⫽37 (r 2⫽0.49)
Thalamic ROI
Base model plus FLAIR aging and thalamic ROI (step 2c) N⫽37 (r ⫽0.60)
2
FLAIR aging
Thalamic ROI
0.396
0.137
2.90
⫺0.276
0.116
⫺2.38
0.007
0.02
0.20
2
aging index: initial pulse pressure at ⫺.0736 (t⫽⫺5.67,
P⬍0.001), FLAIR aging index at 0.318 (t⫽2.28, P⫽0.03),
and rCBF change in the thalamic ROI at ⫺0.143 (t⫽⫺1.17,
P value not significant).
Covariates were added to the combined model to check for
potential confounding. Race, body mass index, educational
level, gender, number of cigarettes, dosage level, percent
compliance, resting CBF, and estimated intelligence quotient
all failed to make any significant change in the associations
among the FLAIR brain aging index, thalamic ROI rCBF
responsivity, and systolic BP treatment outcome (P⬍.05 in
all cases). Number of alcoholic drinks per week did not alter
the influence of the brain aging index but did reduce the
variance attributed to thalamic responsivity (t⫽⫺1.52, P
value not significant).
Discussion
A substantial fraction of the variability between patients in
the success of 2 standard pharmacological treatments of BP
Post-PreTreatment SBP (mmHg)
Downloaded from http://hyper.ahajournals.org/ by guest on May 6, 2017
Greater systolic BP treatment response is scored as a more negative difference from initial systolic BP. The r values associated with models define the proportion
of variance explained by that model; those in the increment column define the change in the proportion of variance explained added by the variable(s) tabled. Base
model ␤ weights and contributions to variance shift slightly as different variables are added. These minor changes are not presented for economy of presentation.
All of the r 2 increments shown are statistically significant, at P⬍0.05. The parietal ROI was also added as a separate step 2 variable but was not significant (t⫽⫺0.92).
20
10
0
-10
-20
-30
-40
-50
0
3
6
9
Index of Structural Brain Aging
12
Figure 2. Change in systolic BP over 40 weeks (posttreatment⫺
pretreatment) with angiotensin-converting enzyme I or ␤-blocker
shown as a function of pretreatment ratings of the presence of
white matter lesions and ventricle and sulcal size derived from
FLAIR MRIs (r⫽0.30, n⫽43).
was accounted for by a combination of structural and functional indices of the brain. Structural and functional indices
were independently related to treatment outcome and were
also independent of the influences of chronological age,
medication type, and initial BP. Furthermore, a substantial
number of covariates, which potentially could have accounted
for our findings, failed to alter the results (body mass index,
gender, race, drug dosage level, smoking, estimated intelligence, and percent compliance). Level of alcohol use did alter
the strength of the functional, but not the structural, index’s
relationship to treatment outcome.
Ratings based on FLAIR images of the brain were combined to provide a general index of brain aging. This index
was related to BP treatment outcomes. Scores going into the
index were morphological changes in the ventricles and sulci,
as well as the presence of small lesions in white matter.
Subsequent checks on the contribution of these individual
scores to the treatment effects showed that each score
demonstrated the same direction of relationship with BP
change, albeit with some variation in the strength of the
relationship. The FLAIR index was recomputed without the
subcortical white matter index, which was not an original
scoring dimension in the Cardiovascular Health Study.33
Results remained essentially the same. Given this, the combined indices should provide more reliable and replicable
measures than the individual scores.41
Our results show poorer responses to treatment among
hypertensive individuals with greater brain aging relative to
their chronological age. This may imply that hypertension
accelerates brain aging or that a third factor influences both
hypertension and brain aging. Interestingly, accelerated aging
has been postulated previously as an effect of hypertension
based on physiological data unrelated to brain function or
structure.13,42– 44 Alternatively, Diz and Lewis21 recently reviewed results suggesting that the brain renin-angiotensin
system may have parallel effects on aging and BP. Given the
evidence that BP reduction can be resistant to treatment in a
significant portion of cases, particularly among the elder-
1118
Hypertension
December 2008
Downloaded from http://hyper.ahajournals.org/ by guest on May 6, 2017
ly,23–25 our results suggest that an aging brain may impede
successful therapy.
The thalamic rCBF response to working memory tasks but
not to the posterior parietal rCBF response related to the
success of BP treatment, although this effect was not significant for pulse pressure or when alcohol use was controlled.
Among the ROIs studied in our previous work, only the
thalamic and posterior parietal showed a diminished rCBF
response to working memory relative to our normotensive
controls.9 Given this, we examined these 2 ROIs to see if the
degree of diminished perfusion would relate to success of
treatment. The thalamus shows reduced metabolic activity
with aging27,45 and less gray matter volume with age in some
but not all of the studies.29,46 Changes with age in thalamic
activity during working memory tasks have been less consistently identified.28,47 Overall, however, the thalamus may be
more susceptible to age-related, and possibly alcohol-related,
changes than the posterior parietal regions. This susceptibility
appears to extend to the influence of hypertension and
potentially accounts for the relationship of thalamic responsivity, but not posterior parietal responsivity, to the efficacy
of BP treatment.
Except for the influence of alcohol use on the thalamic
relationship just noted, neither basic demographic factors nor
health -relevant factors modified the relationship between
brain indices and treatment success. Compliance to treatment
was uniformly high in this sample and unrelated to either BP
or brain measures. There was minimal dropout after treatment
was initiated. Overall, however, no covariate of our brain
measures provided a convincing alternative explanation of
the link between brain indices and treatment outcome.
Our interpretations are limited by our data. Our sample size
and representativeness were constrained by the requirement
for initially nonmedicated patients and the general acceptability to potential participants of imaging procedures. This limits
the generalizability of our results. We did not use a placebotreated control group for ethical and cost reasons and, thus,
cannot control for the effect of time and spontaneous changes
in BP. Duration of heightened BP, indices of vascular aging,
and related longitudinal data were not available but would be
very important for refining our interpretation. Finally, we
compared only 2 common BP medications, and the results
may be specific to these medications.
Overall, however, we found a surprisingly strong relationship between brain indices and the success of standard
pharmacological treatment of hypertension. We do not know
of any physiological mechanism that would provide a convincing, definitive explanation of how brain aging or responsivity would influence treatment outcome. Two speculative
implications may be worth noting, however. First, the pathology of essential hypertension may not solely influence BP but
may have systemic effects that have suggested to many that it
hastens biological aging. If so, then an early target of these
processes may be the brain. A close examination of factors
such as the actions of the angiotensin system in the brain and
interactions with brain stem BP control areas would seem
warranted based on this speculation.21 Second, the systemic
reaction to BP lowering, per se, may involve adjustments in
the brain. Autonomic nervous system regulation of vascular
tone, sensitivity of baroreceptors, and other regulatory systems should adjust to altered pressure profiles. Such adjustment may be more effective in those with “younger,” more
responsive brains. Studies throughout the course of treatment
could assess whether such changes occur and whether they
directly contribute to treatment success.
Perspectives
The correlation between brain indices and treatment outcome
would suggest that more aggressive treatment early in the
course of hypertensive disease may be beneficial for brain
structure and function. At present, the expense of brain
imaging precludes diagnostic use, but technical advances may
permit evaluation of the use of early treatment for brain
structure and function in the foreseeable future.
Acknowledgments
We thank Dr Christopher Ryan and Michelle Geckle for their
assistance with the neuropsychological testing and Michael Eddy and
Carl Becker for their technical assistance.
Sources of Funding
The National Heart, Lung, and Blood Institutes grant 057529
supported this research. E.M.W. was supported by a K-Award
(MH067710). Assistance from the Pittsburgh Mind-Body Center
(National Institutes of Health grants HL076852/076858) is also
acknowledged.
Disclosures
None.
References
1. Izzo JI Jr, Black HR, Sica DA. Office management of hypertension. In:
Izzo JI Jr, Sica DA, Black HR, eds. Hypertension Primer, Fourth Edition.
Philadelphia, PA: Lippincott, Williams, & Wilkins; 2008:401– 406.
2. Rosendorff C, Black HR, Cannon CP, Gersh BJ, Gore J, Izzo JL Jr,
Kaplan NM, O’Connor CM, O’Gara PT, Oparil S. American Heart
Association Council for High Blood Pressure Research, American Heart
Association Council on Epidemiology and Prevention. Treatment of
hypertension in the prevention and management of ischemic heart
disease: a scientific statement from the American Heart Association
Council for High Blood Pressure Research and the Councils on Clinical
Cardiology and Epidemiology and Prevention. Circulation. 2007;115:
2761–2788.
3. Burke LE, Dunbar-Jacob JE, Hill MN. Compliance with cardiovascular
disease prevention strategies: a review of the research. Ann Behav Med.
1997;19:239 –263.
4. Izzo JL Jr, Giles TD, Materson BJ. Defining the syndrome of hypertension. In: Izzo JL Jr, Sica DA, Black HR, eds. Hypertension Primer,
Fourth Edition. Philadelphia, PA: Lippincott, Williams, & Wilkins; 2008:
351–355.
5. Phillips SJ. Neurologic evaluation in hypertension. In: Izzo JL Jr, Sica
DA, Black HR, eds. Hypertension Primer, Fourth Edition. Philadelphia,
PA: Lippincott, Williams, & Wilkins; 2008:379 –382.
6. Beason-Held LL, Moghekar A, Zonderman AB, Kraut MA, Resnick SM.
Longitudinal changes in cerebral blood flow in the older hypertensive
brain. Stroke. 2007;38:1766 –1773.
7. Elias MF, Sullivan LM, Elias PK, D’Agostino RB Sr., Wolf PA, Seshadri
SR, Av R, Benjamin ES, Vasan RS. Left ventricular mass, blood pressure,
and lowered cognitive performance in the Framingham offspring. Hypertension. 2007;49:439 – 445.
8. Ficzere A, Valikovics A, Fulesdi B, Juhasz A, Czuriga I, Csiba L.
Cerebrovascular reactivity in hypertensive patients: a transcranial Doppler study. J Clin Ultrasound. 1997;25:383–389.
9. Jennings JR, Muldoon MF, Ryan C, Price JC, Greer P, Sutton-Tyrrell K,
van der veen FM, Meltzer CC. Reduced cerebral blood flow response and
compensation among patients with untreated hypertension. Neurology.
2005;64:1358 –1365.
Jennings et al
Downloaded from http://hyper.ahajournals.org/ by guest on May 6, 2017
10. Robbins MA, Elias MF, Elias PK, Budge MM. Blood pressure and
cognitive function in an African-American and a Caucasian-American
sample: the Maine-Syracuse Study. Psychosom Med. 2005;67:707–714.
11. Waldstein SR, Jennings JR, Ryan CM, Muldoon MF, Shapiro AP,
Polefrone JM, Fazzari TV, Manuck SB. Hypertension and neuropsychological performance in men: interactive effects of age. Health Psychol.
1996;15:102–109.
12. Waldstein SR, Manuck SB, Ryan CM, Muldoon MF. Neuropsychological
correlates of hypertension: review and methodologic considerations.
Psychol Bull. 1991;110:451– 468.
13. Lakatta EG. Similar myocardial effects of aging and hypertension. Eur
Heart J. 1990;11(suppl G):29 –38.
14. Lakatta EG. Arterial pressure and aging. Int J Cardiol. 1989;25(suppl 1):
81–89.
15. Raz N. The aging brain observed in vivo: differential changes and their
modifiers. In: Cabez R, Nyberg, L, Park D, ed. Cognitive Neuroscience of
Aging. Oxford, United Kingdom: Oxford University Press; 2005:19 –57.
16. Raz N, Rodrigue KM, Kennedy KM, Acker JD. Vascular health and
longitudinal changes in brain and cognition in middle-aged and older
adults. Neuropsychology. 2007;21:149 –157.
17. Gianaros PJ, Greer PJ, Ryan CM, Jennings JR. Higher blood pressure
predicts lower regional grey matter volume: consequences on short-term
information processing. Neuroimage. 2006;31:754 –765.
18. Jennings JR, Muldoon MF, Ryan CM, Mintun MA, Meltzer CC,
Townsend DW, Sutton-Tyrrell K, Shapiro AP, Manuck SB. Cerebral
blood flow in hypertensive patients: an initial report of reduced and
compensatory blood flow responses during performance of two cognitive
tasks. Hypertension. 1998;31:1216 –1222.
19. Raz N, Rodrigue KM, Acker JD, Raz N, Rodrigue KM, Acker JD.
Hypertension and the brain: vulnerability of the prefrontal regions and
executive functions. Behav Neurosci. 2003;117:1169 –1180.
20. Waldstein SR, Giggey PP, Thayer JF, Zonderman AB. Nonlinear
relations of blood pressure to cognitive function: the Baltimore Longitudinal Study of Aging. Hypertension. 2005;45:374 –379.
21. Diz DI, Lewis K. Dahl memorial lecture: the renin-angiotensin system
and aging. Hypertension. 2008;52:37– 43.
22. Benarroch E. Central Autonomic Network: Functional Organization and
Clinical Correlations Armonk, NY: Futura Publishing Company, Inc;
1997.
23. Leotta G, Rabbia F, Canade A, Testa E, Papotti G, Mulatero P, Veglio F.
Characteristics of the patients referred to a Hypertension Unit between
1989 and 2003. J Hum Hypertens. 2008;22:119 –121.
24. Moser M, Setaro JF. Clinical practice. Resistant or difficult-to-control
hypertension. N Engl J Med. 2006;355:385–392.
25. Trewet CLB, Ernst ME. Resistant hypertension: identifying causes and
optimizing treatment regimens. South Med J. 2008;101:166 –173.
26. Jennings JR, Muldoon MF, Price JC, Christie IC, Meltzer CC. Cerebrovascular support for cognitive processing in hypertensive patients is
altered by blood pressure treatment. Hypertension. 2008;52:65–71.
27. Pardo JV, Lee JT, Sheikh SA, Surerus-Johnson C, Shah H, Munch KR,
Carlis JV, Lewis SM, Kuskowski MA, Dysken MW. Where the brain
grows old: decline in anterior cingulate and medial prefrontal function
with normal aging. Neuroimage. 2007;35:1231–1237.
28. Stern Y, Habeck C, Moeller J, Scarmeas N, Anderson KE, Hilton HJ,
Flynn J, Sackheim H, van Heertum R. Brain networks associated with
cognitive reserve in healthy young and old adults. Cereb Cortex. 2005;
15:394 – 402.
29. Tisserand DJ, Jolles J. On the involvement of prefrontal networks in
cognitive ageing. Cortex. 2003;39:1107–1128.
Brain Aging and Blood Pressure Lowering
1119
30. Croog SH, Levine S, Testa MA, Brown B, Bulpitt CJ, Jenkins CD,
Klerman GL, Williams GH. The effects of antihypertensive therapy on
the quality of life. N Engl J Med. 1986;314:1657–1664.
31. Bulpitt CJ, Dollery CT, Bulpitt CJ, Dollery CT. Side effects of hypotensive agents evaluated by a self-administered questionnaire. BMJ. 1973;
3:485– 490.
32. Bastianello S, Bozzao A, Paolillo A, Giugni E, Gasperini C, Koudriavtseva T, Millefiorini E, Horsfield MA, Colonnese C, Toni D, Fiorelli M,
Pozzilli C, Bozzao L. Fast spin-echo and fast fluid-attenuated inversionrecovery versus conventional spin-echo sequences for MR quantification
of multiple sclerosis lesions. Am J Neuroradiol. 1997;18:699 –704.
33. Longstreth WT Jr, Manolio TA, Arnold A, Burke GL, Bryan N, Jungreis
CA, Enright PL, O’Leary D, Fried L. Clinical correlates of white matter
findings on cranial magnetic resonance imaging of 3301 elderly people.
The Cardiovascular Health Study. Stroke. 1996;27:1274 –1282.
34. Watson CC, Newport D, Casey ME, DeKemp RA, Beanlands RS. Evaluation of simulation-based scatter correction for 3D PET cardiac imaging.
IEEE Trans Nucl Sci. 1997;44:90 –97.
35. Woods RP, Mazziotta JC, Cherry SR. MRI-PET registration with
automated algorithm. J Comput Assist Tomogr. 1993;17:536 –546.
36. Minoshima S, Berger K, Lee K, Mintun MA. An automated method for
rotational correction and centering of three-dimensional functional brain
images. J Nucl Med. 1992;33:1579 –1585.
37. Iida H, Higano S, Tomura N, Shishido F, Kanno I, Miura S, Murakami M,
Takahashi K, Sasaki H, Uemura K. Evaluation of regional differences of
tracer appearance in cerebral tissues using [15O]water and dynamic
positron emission tomography. J Cereb Blood Flow Metab. 1988;8:
285–288.
38. Price JC, Drevets WC, Ruszkiewicz J, Greer PJ, Villemagne VL, Xu L,
Mazumdar S, Cantwell MN, Mathis CA. Sequential [15O]water PET
studies in baboons: pre- and post-amphetamine. J Nucl Med. 2002;43:
1090 –1100.
39. Dai W, Lopez OL, Carmichael OT, Becker JT, Kuller LH, Gach HM.
Abnormal regional cerebral blood flow in cognitively normal elderly
subjects with hypertension. Stroke. 2008;39:349 –354.
40. van Laar PJ, van der Graaf Y, Mali WPTM, van der Grond J, Hendrikse
J, Group SS. Effect of cerebrovascular risk factors on regional cerebral
blood flow. Radiology. 2008;246:198 –204.
41. Strube MJ, Newman LC. Psychometrics. In: Cacioppo JT, Tassinary LG,
Berntson GG, eds. Handbook of Psychophysiology (III Edition). New
York, NY: Cambridge University Press; 2007:789 – 811.
42. Hamet P, Thorin-Trescases N, Moreau P, Dumas P, Tea BS, deBlois D,
Kren V, Pravence M, Kunes J, Sun Y, Tremblay J. Workshop: excess
growth and apoptosis: Is hypertension a case of accelerated aging of
cardiovascular cells? Hypertension. 2001;37:760 –766.
43. Swan GE, DeCarli C, Miller BL, Reed T, Wolf PA, Carmelli D. Biobehavioral characteristics of nondemented older adults with subclinical
brain atrophy. Neurology. 2000;54:2108 –2114.
44. Lakatta EG. Similar myocardial effects of aging and hypertension.
European Heart J. 1990;11(suppl G):29 –38.
45. Eustache F, Rioux P, Desgranges B, Marchal G, Petit-Taboue MC, Dary
M, Lechavalier B, Baron JC. Healthy aging, memory systems and
regional cerebral oxygen consumption. Neuropsychologia. 1995;33:
867– 887.
46. Grieve S, Clark C, Williams L, Peduto A, Gordon E. Preservation of
limbic and paralimbic structures in aging. Hum Brain Map. 2005;25:
391– 401.
47. Jennings JR, van der Veen FM, Meltzer CC. Verbal and spatial working
memory in older individuals: a positron emission tomography study.
Brain Res. 2006;1092:177–189.
Brain Imaging Findings Predict Blood Pressure Response to Pharmacological Treatment
J. Richard Jennings, Matthew F. Muldoon, Ellen M. Whyte, Joelle Scanlon, Julie Price and
Carolyn C. Meltzer
Downloaded from http://hyper.ahajournals.org/ by guest on May 6, 2017
Hypertension. 2008;52:1113-1119; originally published online November 3, 2008;
doi: 10.1161/HYPERTENSIONAHA.108.120196
Hypertension is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2008 American Heart Association, Inc. All rights reserved.
Print ISSN: 0194-911X. Online ISSN: 1524-4563
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://hyper.ahajournals.org/content/52/6/1113
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Hypertension can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial
Office. Once the online version of the published article for which permission is being requested is located,
click Request Permissions in the middle column of the Web page under Services. Further information about
this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Hypertension is online at:
http://hyper.ahajournals.org//subscriptions/