ELEVATED CEREBRAL BLOOD FLOW IN PATIENTS WITH PURE AUTONOMIC FAILURE

Abstract

Purpose:

Pure autonomic failure (PAF) results from an impaired peripheral autonomic nervous system, and clinical symptoms present with orthostatic hypotension. While the impact on cardiovascular indices of orthostatic intolerance are well-characterized, more limited information is available regarding cerebral hemodynamic dysfunction in PAF. The objective of this study was to test the hypothesis that cerebral blood flow (CBF) is reduced in PAF, and to quantify the relationship between CBF and clinical indicators of disease severity, including peripheral supine arterial blood pressure.

Methods:

Participants with PAF (n=17) and age and sex matched normotensive healthy controls (n=17) were examined using established clinical rating scales, cardiovascular autonomic function tests, and 3T MRI measurements of CBF. CBF-weighted images were also used to determine the prevalence of venous hyperintensities from the major dural sinuses as evidence of abnormal capillary flow. Non-parametric tests and general linear models were used to evaluate differences and correlations between study variables.

Results:

Gray matter CBF was higher in PAF (51.1±13.4 ml/100g/min) compared to controls (42.9±6.5 ml/100g/min, p=0.007). Venous hyperintensities were more prevalent in PAF relative to controls, and the presence and degree of venous hyperintensities was associated with higher mean CBF (p=0.027). In PAF participants, CBF and supine systolic blood pressure were inversely related (Spearman’s-rho=−0.545, p=0.024).

Conclusions:

Findings suggest that PAF patients may exhibit elevated CBF and provide evidence that this condition exerts a hemodynamic impact in the central nervous system.

INTRODUCTION

Pure autonomic failure (PAF) presents in middle-to-late life with symptomatic orthostatic hypotension, and frequently with supine hypertension. PAF pathology is characterized by sympathetic peripheral postganglionic autonomic denervation, with neuronal cytoplasmic alpha-synuclein inclusions in the peripheral autonomic small nerve fibers, and low circulating norepinephrine levels [1]. Patients presenting with PAF are at high risk of phenoconverting to a manifest central nervous system (CNS) alpha-synucleinopathy, including Parkinson’s disease (PD), Dementia with Lewy bodies (DLB), and Multiple Systems Atrophy (MSA) [2]. Importantly, patients can have symptoms of autonomic dysfunction decades before clinical conversion [2], and therefore investigating central changes in PAF provides a potential opportunity to understand prodromal stages of central synucleinopathies and to identify markers that can inform phenoconversion. While the impact of PAF on peripheral blood pressure has been well-characterized, more limited information is available regarding alterations to cerebral hemodynamic function. Investigating the central effects of PAF could have important implications for patient management, including targeted therapies that modulate cerebral hemodynamics. Moreover, biomarkers that can detect early brain changes and predict phenoconversion from peripheral to central synucleinopathies would be an impactful diagnostic tool and could be incorporated into quantitative outcomes of medication efficacy, early-phase drug discovery, and longitudinal assessments of disease progression.

Previous studies report reduced middle cerebral artery velocities in patients with orthostatic hypotension relative to healthy controls [3, 4]; such a reduction in macrovascular flow velocities could be explained by similar cardiac output in the presence of autoregulatory vasodilation or reduced hemodynamic demand by the tissue. Importantly, both effects could have relevance to oxygen and energy substrate delivery to brain tissue and may help to inform clinical symptoms and progression.

PAF patients live with chronic orthostatic hypotension, where alterations to CBF, and potential oxygen delivery, are likely secondary physiological consequences of the condition. While a small amount of oxygen is dissolved in plasma, oxygen delivery to the brain is primarily a product of the CBF and the hemoglobin-bound blood oxygen content, which is the product of hemoglobin concentration, arterial oxygen saturation fraction, and hemoglobin oxygen carrying capacity. The oxygen extraction fraction (OEF), or ratio of the cerebral metabolic rate of oxygen (CMRO₂) to oxygen delivered, is generally 30–40% in healthy brain tissue [5] and can be maintained over a wide range of cerebral perfusion pressures. More specifically, in the presence of reduced oxygen delivery, from arterial steno-occlusion, anemia, or reduced oxygen saturation, smooth muscle lining arterioles will relax to facilitate arteriolar vasodilation, leading to local increases in CBF. While such autoregulatory increases in CBF occur to increase overall oxygen delivery, they may also impact blood transit times, and evidence of accelerated capillary transit time, or capillary transit time heterogeneity, have been reported in high flow velocity scenarios such as anemia [6], as well as in neurodegenerative [7], and ischemic cerebrovascular conditions [8]. Importantly, accelerated blood transit through the capillary bed can led to less time for oxygen to diffuse from the capillary compartment to tissue, which can reduce oxygen delivery even in the presence of high CBF. Given the large range of orthostatic hypotension, and supine hypertension, present in patients with PAF, it is logical that the sequelae of these vascular effects could manifest as cerebral hemodynamic aberrations.

The goal of this pilot study was to test the hypothesis that CBF is reduced in participants with PAF. For this purpose, we used arterial spin labeling (ASL) MRI, which provides a non-invasive reproducible measurement of CBF comparable to CBF obtained with oxygen-15 positron emission tomography (PET) reference standard [9]. Secondary goals were to characterize the relationship between CBF and peripheral supine blood pressure measures, and to evaluate whether evidence of abnormal capillary flow was present in participants with PAF relative to healthy adults.

METHODS

Participants

The study was performed in accordance with The Declaration of Helsinki, and all participants provided written, informed consent before participating in the study as regulated and approved by the Vanderbilt University Medical Center Institutional Review Board.

Adults with a clinical diagnosis of PAF were recruited at Vanderbilt University Medical Center though the Autonomic Rare Diseases Clinical Research Consortium, a multicenter collaborative research group supported by the NIH Rare Disease Clinical Research Network (RDCRN). Normotensive age and sex matched healthy control participants without evidence of brain injury confirmed by neuroimaging were recruited from the local community. Healthy control participants were also required to have normal blood pressure (<130/80 systolic/diastolic mmHg), and did not have a history of anemia, psychiatric illness, head trauma, substance abuse, cerebrovascular disease, orthostatic hypotension, or diabetes, as confirmed by a board-certified neurologist.

Participants with PAF underwent clinical evaluations including sleep questionnaires [10, 11], University of Pennsylvania Smell Identification Test (UPSIT) [12], and sympathetic and parasympathetic cardiovascular autonomic function tests as described in [2]. Blood pressure was recorded during the tilt table test; beat-to-beat blood pressure was measured with finger plethysmography with the hand supported at heart level, and blood pressure was also measured at 1-minute intervals with a validated automated cuff sphygmomanometer over the brachial artery. Supine blood pressure measures were obtained during an overnight stay at the clinical research center using automated blood pressure equipment (IntelliVue MX400, Philips), with at least a 30 minute rest period prior to the supine measurement. During the night (8 PM - 8 AM), supine blood pressure was measured twice, two minutes apart, every two hours (12 total measurements). Participants with PAF were required to have neurogenic orthostatic hypotension defined as a decrease in blood pressure of 20 mmHg systolic or 10 mmHg diastolic within the first three minutes of upright posture, and an absence of phase IV blood pressure overshoot after release of the Valsalva strain, consistent with sympathetic autonomic failure and a neurogenic cause. PAF participants were also characterized by no increase in heart rate in the upright position in response to the drop in blood pressure [13, 14]. For all autonomic function tests, patients were off medications (e.g., midodrine, fludrocortisone, and droxidopa) to raise blood pressure.

PAF participants also underwent neurological examinations including the Movement Disorders Society United Parkinson’s Disease Rating Scale (UPDRS) [15], the Montreal Cognitive Assessment (MoCA) [16], and the Unified Multiple System Atrophy Rating Scale (UMSARS) [17], and were excluded if they (i) fulfilled diagnostic clinical criteria for central alpha-synucleinopathy (i.e. PD [18], MSA [19], or DLB [20]), or (ii) demonstrated clinical symptoms of dementia, depression, cerebrovascular disease, or other CNS disorders.

MRI acquisition

All participants were scanned using a 3T MRI scanner (Philips Healthcare, Best, The Netherlands). PAF participants were scanned off medications. Anatomical MRI scans, including (i) T₁-weighted (MPRAGE; spatial resolution=1×1×1 mm³; TR/TE=8.9/4.6 ms) and (ii) T₂-weighted FLAIR (spatial resolution=1×1×3 mm³; TR/TE=11000/120ms; TI=2800 ms) were obtained to exclude coexisting CNS disorders, as well as for image co-registration and tissue atrophy characterization. Pseudo-continuous ASL data were acquired using 2D single shot echo planar imaging (field of view = 220×220×119 mm³, slices=20; spatial resolution=3.5×3.5×5 mm³; gap = 1 mm, TR/TE=4000/12 ms with post-labeling delay and labeling pulse train length of 1500 ms).

MRI analysis

Anatomical scans were reviewed for evidence of overt stroke, tumor, moderate-to-severe white matter ischemic injury, or other anatomical abnormalities inconsistent with age.

All analyses for the ASL images were performed in the participant’s native space and CBF maps were obtained as described in [21]. T₁-weighted images were brain extracted and and the total gray matter mask was obtained by merging all the cortical and subcortical structures from FSL FIRST [22] and FAST [23]. Finally, the CBF maps were co-registered to the T₁-weighted images using FSL’s FLIRT with 6 degrees of freedom, and the mean gray matter CBF was recorded.

Previous studies have suggested that ASL images can be used to detect evidence of abnormal water exchange between capillaries and tissue [6]. In particular, the presence of venous sinus hyperintensity in ASL images that contain major draining veins (e.g., superior sagittal sinus, straight sinus, and transverse sinus) has been associated with altered capillary/tissue water exchange, which may be indicative of suboptimal tissue oxygen delivery [6], or blood-brain barrier (BBB) disruption [24]. Here, in a supplementary investigation of the CBF-weighted ASL data we used the CBF maps to determine the prevalence of venous hyperintensities from these major dural sinuses, and if present, how this imaging finding covaries with CBF. Using a previously reported categorical scoring system [6], two raters independently assessed the CBF maps for venous hyperintensity, blinded to the clinical status of the participant, and assigned categorical venous hyperintensity scores of 0=no venous hyperintensity, 1=focal hyperintensity of the superior sagittal sinus or straight sinus, or 2=diffuse hyperintensity of the superior sagittal sinus extending from the crown to approximately the level of the transverse sinuses (Fig. 4). When an extensive venous hyperintensity was present (i.e., score=2), the hyperintensity could be tracked along the length of the superior sagittal sinus into the transverse sinuses. Interobserver agreement was assessed using Cohen’s κ, and a consensus hyperintensity score was determined by discussion-consensus if there was disagreement; the consensus hyperintensity score was used in all subsequent analyses.

Fig. 4.

Representative examples of participants across the spectrum of hyperintensity scores on three different axial slices. For (a) a pure autonomic failure (PAF) participant with hyperintensity score=0, no hyperintensities were observed along any portion of the superior sagittal sinus or straight sinus. (b) For the PAF participant with score=1, a region of focal hyperintensity (arrows) was observed along the superior sagittal sinus, although not extending to the confluence of the superior sagittal sinus and transverse sinuses. (c) For the PAF participant with score=2, hyperintensities (arrows) were observed along a large portion of the superior sagittal sinus as well as in the straight sinus and transverse sinuses and could be tracked on multiple axial slices. (d) Boxplots showing the CBF and the venous hyperintensity scores for controls and PAF participants. Central solid lines denote median, upper and lower lines denote 75^th and 25^th percentiles respectively, and whiskers extend to all data points

Statistical analysis and hypothesis testing

First, group differences in age and sex were evaluated using a Wilcoxon rank-sum test and Fisher’s Exact test, respectively. The primary hypothesis of this study was that CBF is lower in a cohort of PAF participants compared normotensive healthy controls. We applied a general linear model (GLM) regression analysis using total gray matter CBF as the dependent variable, and group (PAF or control) and age as independent variables. The secondary hypothesis was that lower CBF is associated with lower peripheral blood pressure. To evaluate this hypothesis, we calculated the Spearman’s rho correlation coefficient to evaluate the relationships between total gray matter CBF and supine blood pressure (i.e., systolic blood pressure (SPB), diastolic blood pressure (DBP), mean arterial pressure (MAP)=SBP+2DBP/3), and heart rate.

To test the supplementary hypothesis that PAF participants may have abnormal capillary flow dynamics relative to healthy adults, we used a Fisher’s Exact test to evaluate the difference in prevalence of the venous hyperintensities between controls and PAF participants. To understand the relationship between the degree of venous hyperintensity and total gray matter CBF, we used a GLM analysis in which total gray matter CBF was used as the dependent variable, while degree of venous hyperintensity was the independent variable. Finally, to understand if there were clinical differences between PAF participants with (score>1) and without (score=0) venous hyperintensities, we conducted a post hoc Wilcoxon rank-sum tests on the neurological scores (MoCA, UPDRS and UMSARS).

All analyses were performed using R version 4.0.2 (R Foundation for Statistical Computing, Vienna). To account for multiple comparisons, the results were considered significant at the level of false discovery rate (FDR) of 0.1 [25].

RESULTS

Demographics

The recruited cohort consisted of 18 PAF and 17 control participants; one PAF participant was excluded due to discovery of normal pressure hydrocephalus. Demographic and clinical features for the remaining 17 PAF participants meeting the inclusion criteria are summarized in Table 1. There was no significant difference in distribution of males/females between PAF (male/female=12/5) and control (male/female=9/8) (p=0.296) cohorts. Mean age was not significantly different between PAF (70.1 ± 5.6 years) compared to control (66.8 ± 5.5 years) participants (p=0.166). In PAF participants, plasma norepinephrine levels in the supine position were 115 ± 56 pg/ml, consistent with reduced sympathetic activity [26]. Four PAF participants were unable to stand for three minutes due to a drastic drop in blood pressure. No PAF participants had evidence of anemia (hemoglobin < 10 g/dL) on complete blood count. Two PAF participants had evidence of mild white matter changes as observed on FLAIR MRI; no participants had evidence of prior overt stroke, moderate-to-severe ischemic white matter disease, or an independent major neurological condition on neuroimaging or neurological evaluation. The most commonly prescribed medications were midodrine (41%), fludrocortisone (29%), and droxidopa (29%).

Table 1.

Demographic and clinical data of PAF patients

Sex [male/female]	12/5
Age [years]	70.2 ± 5.6
Disease Duration [years]	7.7 ± 3.3
Montreal Cognitive Assessment (MoCA) [min=0, max=30] a	25.6 ± 3.7
University of Pennsylvania Smell Identification Test (UPSIT) [min=0, max=40] a	21.3 ± 7.7
Epworth Sleepiness Scale (ESS) [min=0, max=24] b	6.1 ± 4.2
Unified Multiple System Atrophy Rating Scale (UMSARS) [min=0, max=104] b	8.3 ± 5.6
Movement Disorders Society-United Parkinson’s Disease Rating Scale (MDS-UPDRS)
Part I [min=0, max=16] b	12.5 ± 7.4
Part II [min=0, max=52] b	5.0 ± 4.9
Part III [min=0, max=108] b	9.9 ± 8.4
Hemoglobin [grams per deciliter]	13.2 ± 1.2
Hematocrit [%]	39.9 ± 3.4
Supine norepinephrine [pg/ml]	114 ± 56
Head-up tilt norepinephrine [pg/ml]	200 ± 124
Supine [n=17]
Systolic Blood Pressure (SBP) [millimeters of mercury]	147.8 ± 23.1
Diastolic Blood Pressure (DBP) [millimeters of mercury]	79.4 ± 13.9
Mean Arterial Blood Pressure (MAP) [millimeters of mercury]	103.5 ± 12.0
Heart Rate (HR) [beats per minute]	61.2 ± 6.6
3 minutes of upright posture [n=13]
Systolic Blood Pressure (SPB) [millimeters of mercury]	89.3 ± 26.6
Diastolic Blood Pressure (DBP) [millimeters of mercury]	60.0 ± 17.5
Mean Arterial Blood Pressure (MAP) [millimeters of mercury]	69.8 ± 19.9
Heart rate (HR) [beats per minute]	70.9 ± 11.6
Delta SBP (3 minutes of upright SBP – Supine SBP) [millimeters of mercury]	−63.9 ± 30.1
Delta DBP (3 minutes of upright DBP – Supine DBP) [millimeters of mercury]	−22.8 ± 17.8
Delta HR (3 minutes of upright HR – Supine HR) [beats per minute]	10.9 ± 12.6

Data are shown as mean ± standard deviation

^aLower scores reflect greater disease burden

^bHigher scores reflect greater disease burden

CBF in PAF vs. control participants

Fig. 1 shows representative examples of anatomical images and CBF maps for a control and a PAF participant. Fig. 2a–b shows the mean CBF maps for PAF and control participants. We observed that gray matter CBF was statistically higher in PAF (51.1 ± 13.4 ml/100g/min) compared to control (42.9 ± 6.5 ml/100g/min) participants while accounting for age (estimate=10.44, CI=[−1.32,−0.03], p=0.007) (Fig. 2c).

Fig. 1.

Case examples of T₁-weighted, FLuid Attenuated Inversion Recovery (FLAIR), and cerebral blood flow (CBF) images from (a) a control (sex=female, age=67 years) and (b) pure autonomic failure (PAF) (sex=female, age=63 years) participant. Note that the PAF participant has unremarkable anatomical imaging for age, but elevated CBF, a finding that was observed at the group level (Fig. 2)

Fig. 2.

(a)-(b) Orthogonal representation of the mean quantitative cerebral blood flow (CBF) maps (ml/100 g/min) across (a) controls and (b) pure autonomic failure (PAF) participants. (c) Boxplots showing total gray matter CBF for the control and PAF participants (general linear model p=0.007). Central solid lines denote median, upper and lower lines denote 75^th and 25^th percentiles respectively, and whiskers extend to all data points

CBF and blood pressure

In PAF participants, there was a significant inverse relationship between CBF and supine SBP (Spearman’s-rho=−0.545, p=0.024) (Fig. 3b). No significant relationship was observed between CBF and heart rate (Spearman’s-rho=−0.035, p=0.893) (Fig. 3d).

Fig. 3.

Scatter plots showing the relationships between gray matter cerebral blood flow (CBF) and supine (a) mean arterial blood pressure (MABP), (b) systolic blood pressure (BP), (c) diastolic blood pressure (BP), and (d) heart rate in pure autonomic failure (PAF) participants

Venous hyperintensities

There was good inter-observer agreement in determining hyperintensity scores (overall agreement=76.5%, Cohen’s κ = 0.539, 95% CI=[0.273, 0.805]). The prevalence of any venous hyperintensity in PAF participants was observed to be higher relative to control participants, but did not meet criteria for statistical significance (p=0.057). Venous hyperintensities with a score of 1 were observed in 12% of the controls (2/17); no controls had score=2. For PAF participants, 47% (8/17) demonstrated venous hyperintensities, with only two participants having score=1 but six participants having score=2. Representative examples of PAF participants with hyperintensity scores of 0, 1 and 2 are shown in Fig. 4. For participants with score=1, the hyperintensities were primarily localized to the superior sagittal sinus, while for participants with score=2, hyperintensities were observed along the superior sagittal sinus as well as in the straight and transverse sinuses and could be tracked on multiple axial slices. PAF participants with score=2 also had elevated gray matter CBF (Fig. 4d) (estimate=10.94, CI=[1.31–20.57], p=0.027). Post hoc analyses showed that PAF participants with venous hyperintensities (score>1) had significantly higher UPDRS Part I (p=0.043), UPDRS Part II (p=0.014), and UMSARS (p=0.027) scores compared to PAF participants without the hyperintensities.

DISCUSSION

We examined CBF, blood pressure, and presence of venous hyperintensities in CBF-weighted MRI, in a cohort of PAF participants. Contrary to our primary hypothesis, the main finding was that participants with PAF have slightly but significantly higher supine CBF compared to age and sex matched healthy control participants on average. In PAF participants, CBF was inversely related to supine systolic blood pressure. The prevalence of venous hyperintense signal in CBF-weighted images was higher in PAF participants relative to controls, and these hyperintensities were noted in participants with higher mean CBF. These results collectively provide evidence that PAF participants exhibit different cerebral hemodynamics relative to control participants.

Conventionally, PAF is considered a peripheral autonomic nervous system disorder, and while there is evidence showing that the majority of patients with PAF have anosmia and REM sleep behavior disorders [2] indicating an involvement of the CNS, previous studies have focused almost exclusively on cardiovascular indices of orthostatic intolerance, and assessments on brain function are limited. In PAF, orthostatic symptoms are likely a manifestation of impaired cerebral perfusion, and therefore, it is necessary to understand the possible compensatory mechanisms that may account for this change. Higher supine CBF in PAF as compared to controls suggests that there exist compensatory changes to cerebral hemodynamic function in this population. One mode of compensation is through increased CBF, which can be achieved through changes in microvascular resistance and corresponding increases in cerebral blood volume. Previous studies suggest that the cerebral autoregulation curve widens over time in response to chronic orthostasis, such as in PAF [27] and MSA [28], which is thought to explain orthostatic tolerance in these patient populations [29]. The cerebral vasculature possibly shifts toward a more vasodilated state at baseline, as a way to maintain adequate CBF during orthostatic hypotension. Therefore, it is possible that the CBF also becomes elevated even in the supine state, because the curve is shifted toward a higher state of vasodilation to better maintain CBF during low blood pressure, resulting in the observed inverse relationship between supine CBF and blood pressure. The higher CBF in PAF participants compared to control participants in all subcortical regions suggests that the elevated CBF is not regional, but rather a global effect possibly associated with changes in vasodilatory mediators.

A previous study using transcranial Doppler showed that patients with orthostatic hypotension had higher resting supine heart rate, SBP, DBP, and MAP compared to controls, while blood velocities, cardiovascular indices (stroke volume and cardiac output), total peripheral resistance and cerebrovascular resistance did not differ [4]. If blood velocities in the supine position are not different between PAF participants and controls, then the differences that we observe in CBF could be partly related to underlying changes in the flow and distribution of blood across arterioles and capillaries. More specifically, oxygen transfer from arterial blood to tissue depends on the distribution and velocity of blood at the level of the capillaries. Several studies have investigated the relationship between blood flow and oxygen availability in tissue, taking into account the distribution of blood across the capillary bed [7], and indicate that higher shunting (i.e., lower capillary transit time and higher capillary flow heterogeneity) is consistent with tissue hypoxia, despite an elevated CBF in many cases [6, 7]. Here, we observed that PAF participants had higher CBF compared to normotensive control participants, as well as more prevalent venous hyperintensities, indicating that capillary shunting or abnormal water exchange at the capillary level, associated with BBB dysfunction [24], may be present in these participants.

In the context of chronic orthostatic hypotension, increases in CBF may be necessary to compensate for reductions in oxygen extraction efficiency which could result from changes in capillary function. Such shunting may be indicative of suboptimal tissue oxygen delivery, even in the presence of high CBF, and would manifest as reduced oxygen delivery to tissue per unit of blood. In the presence of capillary shunting, a larger amount of labeled blood water may traverse the capillaries without exchanging with tissue water, leading to hyperintensities in the draining veins (e.g. dural venous sinuses) in the CBF map [6]. Our results reveal that venous hyperintensities were more prevalent in participants with PAF compared to controls, and that the presence and degree of hyperintensities were associated with higher mean CBF, providing evidence that some PAF participants exhibit impaired cerebral hemo-metabolic relationships. Moreover, we find that PAF participants with venous hyperintensities had higher UPDRS Part I and Part II, and higher UMSARS compared to PAF participants without the hyperintensities, suggesting that the presence of venous hyperintensities corresponds with a greater disease burden and may herald the conversion from a prodromal to manifest neurodegenerative disorder.

It is worth noting that several participants in this study (n=7) exhibited supine hypertension (SBP>150 or DBP>90 mmHg), which is a common, but as yet unexplained phenomenon in PAF. The fact that CBF is decreased with higher supine blood pressure suggests that those participants may actually have less cerebral vasodilation compared to those with lower supine blood pressure. Interestingly, a study by Goldstein and colleagues [30] found that supine hypertension was correlated with the magnitude of orthostatic hypotension, suggesting that these participants also exhibit less cerebral vasodilation during orthostasis. Our findings indicate that participants without supine hypertension are the ones with higher supine CBF, and it is possible that PAF participants with and without supine hypertension may have different central compensatory mechanisms, which should be investigated in future studies.

Our findings should be interpreted in the context of several limitations. First, the sample size, while large for neuroimaging studies of PAF, did not permit for multiple covariates to be considered in analysis. Second, for the purpose of the autonomic function tests and MRI scanning the PAF participants were off medications that interfere with supine blood pressures (e.g., midodrine, fludrocortisone, and droxidopa), we cannot rule out longer-term effects from these therapies. Third, while ASL MRI is a promising non-invasive alternative for obtaining accurate and reproducible CBF measurements, it presents numerous acquisition and quantification choices that vary across studies and can impact CBF quantification. While within a physiological range, we obtained slightly lower CBF values for controls compared to previously published data in healthy controls of the same age range [31]. The absolute CBF discrepancy elucidates potential sources of acquisition and post-processing variability in the ASL technique, however, an identical procedure was used in both cohorts in this study, and therefore any residual bias in CBF quantification will be present to a similar extend in both cohorts. Compared to previous studies, our CBF estimates remain in largely a healthy range for age for most participants in both cohorts, but are higher on average in PAF participants compared to age and sex matched healthy control participants. It is worth noting that while mean CBF for PAF participants remained in the healthy range for age, 30% of the PAF participants had markedly elevated gray matter CBF of > 60 ml/100g/min, and these participants were not anemic. This finding emphasizes the heterogeneity of CBF values in the PAF cohort. It is possible that PAF participants with such markedly elevated CBF may exhibit abnormal compensatory elevations in response to orthostatic hypotension, and these changes may contribute to the clinical course. Additional efforts to recruit a larger longitudinal population is required to better define abnormal ranges of CBF values for PAF patients, and to understand if the changes in CBF can inform the clinical progression. Finally, given the inherent nature of the MRI procedure, we CBF was assessed in the supine position, and therefore we were not able to assess CBF changes in the standing posture. Future studies should investigate CBF while supine and during simulated orthostasis, which would probably involve (i) modulating peripheral pressures with medications [32], or pressure chambers [33], and/or (ii) modulating intracranial vascular pressures with graded hypercapnia [34]. Such experiments should clarify if the changes in CBF are related to orthostatic hypotension or supine hypertension.

In conclusion, our results suggest that PAF participants exhibit globally elevated CBF relative to age and sex matched controls, and exhibit an inverse relationship between CBF and peripheral arterial blood pressure. Furthermore, more subtle changes on imaging are consistent with either capillary shunting or BBB permeability elevation, and as such could provide a basis for future studies to evaluate whether these characteristics underlie conversion to a neurodegenerative disorders in these participants.