Independent and Interactive Effects of Blood Pressure and Cardiac Function on Brain Volume and White Matter Hyperintensities in Heart Failure

Abstract

Background

Reduced systemic perfusion and comorbid medical conditions are key contributors to adverse brain changes in heart failure (HF). Hypertension, the most common co-occurring condition in HF, accelerates brain atrophy in aging populations. However, the independent and interactive effects of blood pressure and systemic perfusion on brain structure in HF have yet to be investigated.

Methods

Forty-eight older adults with HF underwent impedance cardiography to assess current systolic blood pressure status, and cardiac index to quantify systemic perfusion. All participants underwent brain magnetic resonance imaging to quantify total brain, total and subcortical gray matter volume, and white matter hyperintensities (WMH) volume.

Results

Regression analyses adjusting for medical and demographic factors showed decreased cardiac index was associated with smaller subcortical gray matter volume (p < .01) and higher systolic blood pressure predicted reduced total gray matter volume (p = .03). The combination of higher blood pressure and lower cardiac index exacerbated WMH (p = .048).

Conclusions

Higher blood pressure and systemic hypoperfusion are associated with smaller brain volume and these factors interact to exacerbate WMH in HF. Prospective studies are needed to clarify the effects of blood pressure on the brain in HF, including the role of long-term blood pressure fluctuations.

1. Introduction

Heart failure (HF) is a widely recognized risk factor for cognitive impairment and is associated with greater risk for neurodegenerative disease and cognitive decline (e.g., Alzheimer’s disease and vascular dementia).^{1, 2} Rapidly growing evidence shows that structural brain damage (e.g., reduced total and regional brain volume, white matter hyperintensities, WMH) is common in HF and associated with poorer cognitive performance.^3–5 Reduced cerebral blood flow secondary to decreased cardiac pumping efficiency^{6, 7} is believed to underpin brain atrophy and ischemic changes in HF. For instance, past work has identified indicators of systemic perfusion (e.g., ejection fraction) as important predictors of neuropathological outcomes (e.g., WMH) in HF.⁸

Common comorbid vascular conditions (e.g., diabetes, sleep apnea) are also believed to contribute to reduced brain volume and ischemic changes in HF. Of all vascular risk factors, high blood pressure is the most common comorbid condition in HF. It affects up to 75% of patients with HF prior to HF diagnosis.⁹ Recent work shows that HF patients with hypertension exhibit greater impairments on cognitive testing compared to those without hypertension.¹⁰ These findings may be attributed to the negative impact of high blood pressure on the brain. Indeed, hypertension is a well-documented risk factor for accelerated age-related brain shrinkage even in otherwise healthy adults.^{11, 12} High arterial pressure promotes endothelial dysfunction and atherosclerosis¹³ that may lay the foundation for cerebrovascular disease¹⁴ and ensuing cognitive dysfunction.^{15, 16} Moreover, high blood pressure may produce additive deficits in systemic perfusion¹⁷ that exacerbate atrophy and white matter disease in patients with HF.

To our knowledge no study to date has examined the impact of blood pressure on brain structure in older adults with HF. The purpose of the current study was to examine both the independent and interactive effects of blood pressure and systemic perfusion—as operationalized by cardiac index—on structural brain indices in older adults with HF.

2. Methods

2.1 Participants

Persons with HF (48 consecutive admissions) were recruited for an ongoing study examining neurocognitive outcomes in HF. Inclusion criteria were as follows: ages of 50–85 years, English as a primary language, and a diagnosis of New York Heart Association (NYHA) class II or III at the time of enrollment. Potential participants were excluded for any contraindications to magnetic resonance imaging (MRI) (e.g., pacemaker), history of significant neurological disorder (e.g., dementia), history of head injury with more than10 minutes loss of consciousness, axis I psychiatric disorders (e.g. schizophrenia, bipolar disorder), substance abuse and/or dependence, and renal failure. See Table 1 for demographic and medical information.

Table 1.

Demographic, Medical, and Cognitive Characteristics of 48 Older Adults with Heart Failure

Demographic Characteristics
Age, mean (SD) years	68.04 (7.87)
Education, mean (SD) years	13.96 (2.63)
Female (%)	37.5
Race (% Caucasian)	81.3
Medical Characteristics
Cardiac Index, mean (SD) L/min/m2	2.74 (.84)
Systolic Blood Pressure, mean (SD) mm Hg	118.25 (19.17)
Diastolic Blood Pressure, mean (SD) mm Hg	66.56 (9.55)
Diabetes (% yes)	29.2
Myocardial Infarction (% yes)	50.0
Sleep Apnea (% yes)	22.9
Elevated Total Cholesterol (% yes)	54.2
Cognitive Function
3MS	92.88 (5.68)
Medications, (% yes)
Diuretics	37.5
Beta-blockers	68.8
ACE Inhibitors	37.5
Angiotensin II Receptor Blockers	6.3

3MS = Modified Mini Mental State Examination; medication status for 3 participants was missing

2.2 Measures

2.1 Neuroimaging

Whole-brain, high-resolution 3D T1-weighted images (Magnetization Prepared Rapid Gradient-Echo, MPRAGE) were acquired on a Siemens Symphony 1.5 Tesla magnetic resonance imaging scanner for morphologic analysis. Twenty-six slices were acquired in the sagittal plane with a 230 × 100 mm field of view. The acquisition parameters were as follows: Echo time (TE) = 17, repetition time (TR) = 360, acquisition matrix = 256×100, and slice thickness = 5 mm. Whole-brain FLAIR images were also acquired to quantify WMH. For the FLAIR images, twenty-one 5-mm slices were acquired with TR = 8500, TI = 2500, Flip Angle = 150 degrees, TE = 115, and FOV = 220×75.

Morphometric analysis of brain structure was completed with FreeSurfer Version 5.1 (http://surfer.nmr.mgh.harvard.edu). Detailed methodology for regional and total volume derivation has been described in detail previously.^18–20 FreeSurfer was used to perform image preprocessing (e.g. intensity normalization, skull stripping), then to provide both cortical and subcortical volume measures using the surface stream and the subcortical segmentation stream respectively. Freesurfer performs such parcellations by registering images to a probabilistic brain atlas, built from a manually labeled training set, and then using this probabilistic atlas to assign a neuroanatomical label to each voxel in an MRI volume. Total brain volume, total gray matter volume, and subcortical gray matter volume were automatically derived with the subcortical processing stream (i.e., “aseg.stats”).

Total WMH volume was derived by a three-step operator-driven protocol that has been described in detail previously.²¹ Briefly, in Step 1, a threshold was applied to each FLAIR image to label all voxels that fell within the intensity distribution of hyperintense signal. In Step 2, gross regions-of-interest (ROI) were drawn manually to include WMH but to exclude other regions (e.g., dermal fat) that have similar intensity values. In Step 3, a new image is generated that contains the intersection of voxels labeled in Step 1 and those labeled in Step 2. The resulting image contains labeled voxels that are common in Step 1 and Step 2. The number of resulting voxels is summed and multiplied by voxel dimensions to derive a total volume score. We have shown the validity and reliability of this approach previously.²¹

2.2.2 Cognitive Function

The Modified Mini Mental State Examination (3MS) was administered to assess general cognitive status. The 3MS is a brief screening measure that evaluates aspects of attention, orientation, memory, language and visuospatial abilities.²² Scores range from 0–100 with higher scores reflective of better performance.

2.2.3 Impedance Cardiography

Cardiac Output (CO) from a seated resting baseline was calculated for each patient to estimate preservation of cardiac function. Impedance cardiography signals were recorded via a Hutcheson Impedance Cardiograph (Model HIC-3000, Bio-Impedance Technology, Chapel Hill, NC) using a tetrapolar band-electrode configuration. The electrocardiogram (ECG) was recorded from the Hutcheson Impedance Cardiograph using disposable ECG electrodes. The basal thoracic impedance (Zo), the first derivative of the pulsatile impedance (dZ/dt) and the ECG waveforms were processed using specialized ensemble-averaging software (COP, BIT Inc., Chapel Hill, NC), which was used to derive stroke volume using the Kubicek equation. Following instrumentation, impedance cardiographic signals were recorded for seven 40-second periods during a 10-minute resting baseline. Finally, all CO measurements were divided by body surface area (BSA), yielding cardiac index. Cardiac index operationalized systemic perfusion in this study.

Blood pressure was measured, by a trained research assistant, seven times during the 10 minute resting baseline using an automated oscillometric BP device (Accutor Plus Oscillometric BP Monitor, Datascope Corp, Mahwah, NH) providing systolic and diastolic pressures. Initiating the blood pressure reading triggered a concurrent 40-second impedance cardiography measure. A composite consisting of the average systolic and diastolic blood pressure across the seven trials was computed. For the current analyses, blood pressure was operationalized using participants’ systolic blood pressure. High systolic blood pressure is most prevalent in persons aged 50 or over and is also a better predictor of cognitive function than diastolic blood pressure.^{23, 24}

2.2.4 Demographic and Medical History

Participants’ medical and demographic history was ascertained through self-report and corroborated by medical record review. Current prescribed hypertensive medications were also ascertained through medical record review. Antihypertensive medications were categorized into four categories as according to the American Heart Association, including: diuretics, beta-blockers, angiotensin-converting enzyme (ACE) inhibitors, and angiotensin II receptor blockers. Refer to Table 1.

2.3 Procedures

The local Institutional Review Board (IRB) approved the study procedures and all participants provided written informed consent prior to study enrollment. In addition to medical record review, participants completed demographic and medical history self-report measures. Individuals were then administered the 3MS and completed impedance cardiography that was conducted by a trained research assistant. Finally, all participants underwent MRI as part of the baseline assessment protocol.

2.4 Statistical Methods

A log transformation of WMH was performed to correct for the positively skewed distribution of this variable. A series of multivariable hierarchical regression analyses was performed to examine the independent and interactive effects of systolic blood pressure and cardiac index with total brain volume, gray matter volume, subcortical gray matter volume, and WMH. For all analyses, continuous predictor variables were transformed to z-scores. Intracranial volume and factors known to influence cognitive outcomes in HF were entered in block 1 of the regression model. These covariates included age, sex (0 = female, 1 = male), and diagnostic history of myocardial infarction, diabetes, elevated total cholesterol, and sleep apnea (0 = negative diagnostic history, 1 = positive diagnostic history). To determine the independent effects of cardiac index on brain structure, cardiac index was entered in block 2. Systolic blood pressure was entered in block 3 to determine its independent association with total brain volume, gray matter volume, subcortical gray matter volume, and WMH. The cross product of systolic blood pressure and cardiac index was computed to create an interaction term that was entered in block 4 of the model. The interactive effects of systolic blood pressure and cardiac index from the regression model were then plotted on to the criterion WMH using the simple slopes test. Lines were computed for individuals 1 SD above the mean blood pressure of the sample (i.e., individuals with high blood pressure) and those at the mean (i.e., average blood pressure of the sample). Finally, past work shows diagnostic history of hypertension negatively impacts cognition in HF,¹⁰ thus regression analyses also examined the association between resting systolic blood pressure and 3MS performance.

3. Results

Systolic Blood Pressure and Cardiac Function

When using clinical convention to define pre-hypertension and hypertension, hypertension was common in this sample with 48.3% exhibiting a systolic blood pressure above 120 mm Hg and 14.7% exhibited a blood pressure consistent with hypertension (i.e., > 140 mm Hg). Many participants were also prescribed antihypertensive agents, including diuretics, beta blockers, ACE inhibitors, and antiogensin II receptor blocker. The cardiac index in the sample fell within the average range (mean = 2.74, SD = .84). See Table 1. Bivariate correlations showed no significant relationship between cardiac index and systolic blood pressure (r (46) = 0.23, p = 0.11).

The Association between Systolic Blood Pressure and Brain Structure

Refer to Table 2 for a full summary of hierarchical regression analyses. After controlling for medical and demographic variables, lower cardiac index was associated with smaller subcortical gray matter volume (β = .37, p < .01). This relationship did not emerge for total brain volume, gray matter volume, or WMH (p > .05 for all). Increased systolic blood pressure was associated with lower total gray matter volume (β = −.27, p = .03) above-and-beyond medical, demographic, and cardiac index variables. However, no significant associations emerged for total brain volume, subcortical gray matter volume, or WMH (p > .05). Despite non-significant independent associations, the interaction between higher systolic blood pressure and lower cardiac index produced synergistic effects on greater WMH (β = −.38, p = .048), but this pattern did not emerge for any of the other MRI indices (p > .05 for all). Figure 1 shows that relative to the sample mean, high blood pressure (as defined by 1 SD above the average blood pressure of the sample) interacts with low cardiac index to exacerbate WMH. There was no association between systolic blood pressure or cardiac index with 3MS performance (p > .05).

Table 2.

Predictive Validity of Systolic Blood Pressure and Cardiac Function on Structural Brain Volume in Older Adults with Heart Failure (N = 48)

	WMH	TBV	GM	Subcortical GM
Variable	β (SE b)	β (SE b)	β (SE b)	β (SE b)
Block 1
Age	.21(.06)	−.05(12097.87)	−.32(8347.43)*	−.19(3229.43)
Sex	−.26(.14)	.01(29530.42)	−.07(20375.74)	−.10(7882.92)
Diabetes	.09(.13)	−.08(28411.80)	−.34(19603.90)**	−.19(7584.31)
Sleep Apnea	−.12(.15)	−.01(32073.91)	−.14(22130.73)	.09(8561.88)
Cholesterol	−.26(.12)	.11(24710.53)	.29(17050.05)*	.11(6586.28)
MI	.30(.12)	.12(25455.41)	−.02(17564.02)	−.06(6795.13)
Volume	.08(.06)	.85(.13621.49)**	.55(9398.71)**	.64(3636.15)**
R2	.29	.77	.55	.40
F	1.90	15.92**	5.72**	3.14**
Block 2
CI	−.11(.06)	.15(12587.15)	−.24(8549.20)	.37(3138.55)**
R2	.30	.79	.60	.52
F for ΔR2	.48	2.96	4.08	8.07**
Block 3
Systolic	−.07(.07)	.14(13047.60)	−.27(8492.33)*	−.23(3227.42)
R2	.30	.81	.66	.56
F for ΔR2	.15	2.41	5.38*	2.95
Block 4
CI X Systolic	−.38(.08)*	.05(17331.46)	−.04(11316.19)	−.16(4230.40)
R2	.39	.81	.66	.58
F for ΔR2	4.24*	.27	.08	1.08

Note.

^*p < 0.05;

^**p < .01;

^***p = .055

Abbreviations: β – standardized regression coefficients, SE – standard error; MI = Myocardial infarction; Volume = Intracranial Volume; CI = Cardiac Index; Systolic = Systolic Blood Pressure; WMH = White Matter Hyperintensities; GM = Gray Matter

Figure 1.

The Interactive Effects of High Systolic Blood Pressure and Low Cardiac Index on Greater White Matter Hyperintensities in Older Adults with Heart Failure

Note. The Figure shows that high blood pressure interacts with lower cardiac index to exacerbate white matter hyperintensities. Lines reflect model specifics. Lower scores on the x-axis reflect lower cardiac index and higher scores on the y-axis represent greater white matter hyperintensities (log transformed). Plotted lines are systolic blood pressure values z-scores (distribution with a mean of 0 and a SD of 1). Mean blood pressure representative of the within-sample blood pressure average (z-score = 0) and high blood pressure indicative of those scores 1 SD above the mean. Cardiac index scores on the x-axis are also within-sample z-scores.

4. Discussion

Consistent with past work, the current study shows that reduced systemic perfusion is associated with brain volume in older adults with HF. The current study also extends these findings by showing that systolic blood pressure is independently associated with cerebral structure in HF and interacts with systemic perfusion to exacerbate WMH in this population.

These findings show that higher systolic blood pressure is associated with smaller total gray matter volume in HF above and beyond the effects of systemic perfusion. These findings are consistent with past work that shows high blood pressure is linked with brain atrophy in aging older adults.^{11, 12} In addition to the known effects of reduced systemic perfusion on the brain in HF,⁸ it is possible that higher blood pressure may result in additive amount of brain atrophy in this population through independent mechanisms. For instance, higher blood pressure can lead to endothelial dysfunction¹³ breakdown of the blood brain barrier,²⁵ and/or atherosclerosis in the large cerebral and cervicocerebral arteries²⁶—all identified risk factors for brain atrophy.^{27, 28} Indeed, past work suggests that higher arterial pressure may be uniquely associated with mortality risk in HF compared to cardiac function²⁹ and future work should determine whether such findings involve the effects of high blood pressure on the brain.

We also found an independent association between decreased systemic perfusion and smaller subcortical gray matter volume in persons with HF. Up to 31% of HF patients exhibit reductions in cerebral blood flow and decreased systemic perfusion has been linked with reduced cognitive test performance and cerebrovascular disease in HF.^{8, 30} These effects may reflect ischemic damage secondary to chronic disruptions in cerebral blood flow.⁷ Interestingly, the current association between cardiac index and subcortical gray matter is consistent with the greater susceptibility of subcortical structures to insufficient cerebral circulation.³¹ However, the differential effects of reduced systemic perfusion on subcortical structures remains to be elucidated. Nonetheless, cerebral blood flow is partly modifiable and past work shows that improvements in cardiac function results in better cognitive test performance.³⁰ Future work should examine whether factors that promote cerebral perfusion (e.g., exercise) can attenuate cognitive decline and brain atrophy in patients with HF.

The current study found no independent association between systolic blood pressure and cardiac index with WMH. However, the combination of increased systolic blood pressure and decreased systemic perfusion produced synergistic effects on WMH volume in HF. The exact reason for this pattern is not entirely clear, though given the ischemic nature of WMH these findings likely involve cerebral hypoperfusion.³² Not surprisingly, reduced cardiac dysfunction often corresponds to lower levels of brain perfusion.³³ High blood pressure elevates perfusion pressure in order to maintain autoregulation of cerebral blood flow levels.³⁴ This increase in pressure can cause hypoperfusion due to the formation of atherosclerotic plaques and vascular narrowing.³⁵ In turn, the presence of both high blood pressure and decrease cardiac index may have disrupted cerebral blood flow beyond a critical threshold necessary for ischemic changes (e.g., WMH) to occur. In addition, the lack of interactive effects between blood pressure and cardiac function on brain volume suggests these factors may alter the cerebral structure in HF through mechanisms independent of brain hypoperfusion and white matter changes. Supporting this hypothesis is past work that suggests white matter disease plays a limited role in brain atrophy secondary to the effects of hypertension.³⁶ Prospective studies are needed to clarify underlying mechanisms for the influence of blood pressure and systemic perfusion on the brain in HF.

Systolic blood pressure was not associated with cognitive function in this sample of HF patients, suggesting the effects of resting blood pressure on the brain in this sample may be subclinical. The association between resting blood pressure and cognitive function in cardiovascular disease populations is mixed with some studies even showing a positive relationship between these variables.^{10, 28} Measured blood pressure is modified by the antihypertensive medication that almost all HF patients are currently prescribed. Thus, a diagnostic history of hypertension may be a better predictor of cognitive impairment as it may signify the sustained and cumulative pre-diagnostic effects of uncontrolled hypertension on the brain rather than readings from a single session.¹⁰ For instance, duration of hypertension and uncontrolled blood pressure have been shown to predict cognitive function and the combination of these factors exacerbate susceptibility to cognitive impairment.¹⁶ Consistent with this notion, many of the HF participants in this sample were prescribed anti-hypertensive agents and it has been shown that those treated for high blood pressure exhibit better cognitive function than those untreated.³⁷ Nonetheless, the effects of anti-hypertensive medications on cognition are small and lifelong control of blood pressure is suggested to have the greatest impact on cognition in these patients.^{15, 38} Future work is much needed to clarify the effects of resting vs. lifelong history of high blood pressure on the brain.

Several limitations of the current study warrant brief discussion. The present analyses were conducted using a cross-sectional design and longitudinal studies would provide key insight into the long-term effects of blood pressure on brain structure in HF. In addition, the current study found no association between resting blood pressure and WMH. It is possible that while anti-hypertensive medication exercised beneficial effects on the measured blood pressure, it did not affect WMH thus disengaging the two variables. Indeed, previous studies showed that even an effective antihypertensive therapy has limited if any effect on the brain structure.³⁹ It is possible that blood pressure fluctuations or diagnostic history of hypertension may be more sensitive predictors of brain changes in HF, particularly WMH. For instance, recent work suggests that high blood pressure interacts with long-term blood pressure fluctuations to exacerbate white matter disease among an elderly sample⁴⁰ and future studies should explore this possibility in HF. Specifically, variability in blood pressure across time may produce distinct effects on the brain (e.g., inconsistent perfusion) relative to resting blood pressure level and future studies should differentiate such effects and determine whether stabilization of blood pressure fluctuation can improve neurocognitive outcomes. The lack of association between blood pressure and cognitive function in this sample may also be due to methodological limitations, including the use of a brief cognitive screening measure. Future work should investigate the association among blood pressure, brain structure, and cognition using neuropsychological test batteries that assess multiple domains of cognition. Finally, larger case-controlled studies would help to clarify the underlying mechanisms of blood pressure on the brain in persons with HF.

In summary, the current study shows that higher blood pressure and decreased systemic perfusion are associated with smaller brain volume and in synergy exacerbate WMH in HF. Longitudinal studies are needed to confirm our findings and clarify the differential effects of current blood pressure status and long-term high blood pressure on the brain.