Microvascular dysfunction and neurovascular uncoupling are exacerbated in peripheral artery disease, increasing the risk of cognitive decline in older adults

Abstract

Peripheral artery disease (PAD) is a vascular pathology with high prevalence among the aging population. PAD is associated with decreased cognitive performance, but the underlying mechanisms remain obscure. Normal brain function critically depends on an adequate adjustment of cerebral blood supply to match the needs of active brain regions via neurovascular coupling (NVC). NVC responses depend on healthy microvascular endothelial function. PAD is associated with significant endothelial dysfunction in peripheral arteries, but its effect on NVC responses has not been investigated. This study was designed to test the hypothesis that NVC and peripheral microvascular endothelial function are impaired in PAD. We enrolled 11 symptomatic patients with PAD and 11 age- and sex-matched controls. Participants were evaluated for cognitive performance using the Cambridge Neuropsychological Test Automated Battery and functional near-infrared spectroscopy to assess NVC responses during the cognitive n-back task. Peripheral microvascular endothelial function was evaluated using laser speckle contrast imaging. We found that cognitive performance was compromised in patients with PAD, evidenced by reduced visual memory, short-term memory, and sustained attention. We found that NVC responses and peripheral microvascular endothelial function were significantly impaired in patients with PAD. A positive correlation was observed between microvascular endothelial function, NVC responses, and cognitive performance in the study participants. Our findings support the concept that microvascular endothelial dysfunction and neurovascular uncoupling contribute to the genesis of cognitive impairment in older PAD patients with claudication. Longitudinal studies are warranted to test whether the targeted improvement of NVC responses can prevent or delay the onset of PAD-associated cognitive decline.

NEW & NOTEWORTHY Peripheral artery disease (PAD) was associated with significantly decreased cognitive performance, impaired neurovascular coupling (NVC) responses in the prefrontal cortex (PFC), left and right dorsolateral prefrontal cortices (LDLPFC and RDLPFC), and impaired peripheral microvascular endothelial function. A positive correlation between microvascular endothelial function, NVC responses, and cognitive performance may suggest that PAD-related cognitive decrement is mechanistically linked, at least in part, to generalized microvascular endothelial dysfunction and subsequent impairment of NVC responses.

INTRODUCTION

Peripheral artery disease (PAD) is a highly prevalent age-related cardiovascular disease that affects 10%–20% of people over 70 yr of age (1–3). PAD represents a considerable global public health problem that imposes an increasing burden on society (4), leads to poor quality of life (5), and is associated with a significantly increased risk of cardiovascular morbidity and mortality (6–10). PAD is characterized by the narrowing of arteries in the lower extremities due to atherosclerosis, which can be accompanied by a variety of symptoms, the most common being leg pain during ambulation (known as intermittent claudication) (11). A growing body of epidemiological and clinical studies suggest that PAD is also associated with pathological alterations in other vascular beds (3). Importantly, clinical evidence demonstrates that PAD is associated with decreased cognitive performance (12–18), increasing the risk for the development of vascular cognitive impairment (VCI) and dementia (19–21).

The potential mechanism contributing to the genesis of cognitive impairment in patients with PAD is not well understood and may involve decreased cerebral perfusion. It is possible that atherosclerotic lesions in larger cerebral arteries are one contributing factor to impaired cerebral perfusion, such as vessels constituting and originating from the circle of Willis at the base of the brain. Functional impairment in the cerebral microcirculation is another contributing factor, which may promote cognitive decline (12, 14, 22–28). Because energetic demands of neurons are high and the brain has very little reserve capacity, maintenance of normal brain function is critically dependent on moment-to-moment adjustment of local cerebral blood flow to match the metabolic needs of activated neurons (23). In response to neuronal activity, signaling between constituents of the corresponding neurovascular unit (pericytes, astrocytes, and endothelial cells adjacent to a given neuron) leads to the dilation of feeding cerebral microvessels (29). These mechanisms, collectively termed as neurovascular coupling (NVC), result in a marked hemodynamic change in response to the release and accumulation of vasoactive mediators. Therefore, NVC responses enable rapid increases in oxygen and glucose delivery to active brain regions via increased perfusion and thus play a vital role in homeostatic support of the neural function. NVC responses are mediated, at least in part, by endothelial production of vasodilator nitric oxide (NO) (30–33), and pathological conditions that promote endothelial dysfunction are known to impair NVC responses. Importantly, patients with PAD exhibit generalized microvascular dysfunction (34–39). Previous studies have demonstrated that atherosclerotic vascular disease is not limited only to large vessels, but extends to other vascular beds, including microvascular circulation (25, 40–43). Based on these findings, we hypothesized that systemic impairment of endothelial function in patients with PAD contributes to the dysregulation of local cerebral blood flow and the genesis of cognitive impairment.

The present study was designed to test the hypothesis that microvascular hemodynamic responses in the microcirculation of the skin and in the brain cortex are impaired in patients with PAD. To test our hypothesis, we enrolled patients with PAD and compared their peripheral microvascular endothelial function [laser speckle contrast imaging (LSCI)] and NVC responses [functional near-infrared spectroscopy (fNIRS)], elicited by the cognitive n-back test, to age- and sex-matched controls. We found significantly impaired NVC responses in the prefrontal cortex (PFC) of patients with PAD and significant peripheral microvascular endothelial dysfunction. The association between microvascular endothelial function, NVC responses, and cognitive performance was analyzed.

METHODS

Study Design and Participant Characteristics

A total of 22 adults [11 participants with PAD, 63.7 ± 5.2 yr of age, 6 males (2 African Americans and 4 Caucasians) and 5 females (2 African Americans and 3 Caucasians); and 11 controls, 64.1 ± 7.6 yr of age, 6 males (all Caucasians) and 5 females (1 African American and 4 Caucasians); P = 0.88, means ± SD] were enrolled in this study. Patients with PAD were recruited through the Vascular Medicine Program, Department of Internal Medicine at the University of Oklahoma Health Sciences Center, and healthy controls were enrolled via advertisement. All study participants obtained at least a high school degree. All participants were assessed for cognitive performance and peripheral microvascular endothelial function. Out of 11 patients enrolled, 5 participants with PAD (62.0 ± 8.6 yr of age, 3 males and 2 females; means ± SD) consented to the NVC assessment with functional near-infrared spectroscopy (fNIRS). In the interest of an unbiased statistical comparison, five age- and sex-matched (P = 0.82) participants were selected from the control group (60.8 ± 6.5 yr of age, 3 males and 2 females; means ± SD) to assess the impact of PAD on NVC responses. Although control participants reported that they had not been smoking ever or in the preceding 20 yr, everyone in the PAD group were smokers with no intention to quit during the study. Participating patients with PAD had an ankle-brachial index (ABI) of 0.9–0.4 and presented with claudication at a >200-m distance. Neither PAD nor controls had a previous diagnosis of cognitive impairment, had a history of stroke or major coronary events in the past 6 mo, or were hypertensive on the day of physiological assessments. Comorbidities of the participants are summarized in Table 1 and were well controlled at the time of visit. All participants were asked to refrain from drinking caffeinated beverages for at least 6 h before the assessments. All participants were right-handed.

Table 1.

Health conditions of the patients with PAD and control participants

Health Conditions	PAD (n = 11)	Control (n = 11)
Hypertension	5/11	7/11
Diabetes	4/11 (3 type 2)	1/11 (type 2)
Hyperlipidemia	7/11	2/11
Other conditions	Osteoarthritis (1), coxarthrosis (1), depression (2), reflux (1)	Osteoporosis (1), osteoarthritis (1), reflux (1)

Comorbidities considered as potential confounders are reported for the peripheral artery disease (PAD) and control group. All participants with reported comorbidities were medically controlled by their physician: β-blockers/diuretics/calcium channel blockers for hypertension, insulin/metformin for diabetes, and statins for hypercholesterinemia.

All procedures and protocols were approved by the Institutional Review Board of the University of Oklahoma Health Sciences Center. Written, informed consent was obtained from each participant before participation.

Cognitive Assessment

Different domains of cognition were assessed using a selection of tests from the Cambridge Neuropsychological Test Automated Battery (CANTAB, Cambridge Cognition). The CANTAB Connect Research tool is a validated and sensitive tool optimized to detect signs of age-related cognitive impairment and mild cognitive decline(25). Testing started with a Motor Screening Task to determine the presence of any sensorimotor deficit, and then continued with the following tests: Reaction Time, Paired Association Learning, Delayed Matching Sample, Spatial Working Memory, Rapid Visual Processing, and Attention. Participants were left alone in a quiet environment during the test with the touchscreen device running the application.

Neurovascular Coupling Protocol with Functional Near-Infrared Spectroscopy

An n-back cognitive test was used to evoke NVC responses, as previously described (40). In brief, participants were seated in front of a computer monitor with their right hand on the mouse, and the n-back tasks were explained before the start of the paradigm, as well as during the task. Participants were instructed to perform to the best of their ability during each task, with 60 letters presented as stimuli during each task. The experimental design consisted of 0-back, 1-back, 0-back, and 2-back sessions during which they had to click only in case of the target stimulus. However, response to a nontarget stimulus (by clicking) was counted as an error (Fig. 1A). During 0-back, participants were asked to identify and respond to the letter “W” when presented on the screen with a left mouse button click. For the 1-back task, participants were asked to respond every time they saw two identical letters in a row (e.g., respond to second A and K sequences in A-A-B-C-K-K). For the 2-back task, the target stimulus was any letter that repeated itself two steps back (e.g., respond to a second letter B in A-A-B-C-B series). The n-back task requires monitoring, updating, and manipulating the information and is therefore assumed to place great demands on several key processes within working memory. The sequence of letters was randomized by the custom software developed in ePrime 3 (Psychology Software Tools, Sharpsburg, PA). The average block length was 134.13 ± 1.37 s.

Figure 1.

Experimental setup for neurovascular coupling (NVC) assessments. A: cognitive n-back task was implemented to evoke NVC responses in study participants. In brief, participants were presented with four tasks in the following order: 0-back, 1-back, 0-back, and 2-back. During 0-back, participants were asked to identify and respond to the letter “W” when presented on the screen with a left mouse button click. For the 1-back task, participants were asked to respond every time they saw two identical letters in a row (e.g., respond to second A in a sequence A-A-B). For the 2-back task, study participants were asked to respond when they saw any letter that repeated itself two letters back (e.g., respond to a second letter A in a sequence A-B-A). The average block length was 134.13 ± 1.37 s. B: NVC responses were assessed using the near-infrared spectroscopy (fNIRS), NIRScout platform (NIRx Medical Technologies LLC, Glen Head, NY) equipped with 16 light sources and 16 detectors. To standardize optode placement, a 128-port Easycap head cap (Easycap GmbH, Woerthsee-Etterschlag, Germany) was positioned on the participant’s head by aligning its Fpz-Iz ports to the sagittal plane of the head and the optode in the Fpz position of the cap with Fpz on the participant. Hemodynamic responses were measured from 48 channels—defined by source-detector pairs with 3 cm separation according to the international 10-20 system—covering the prefrontal and motor cortex regions. C: fNIRS data were analyzed using a pipeline based on General Linear Model (GLM) approach created using the Brain AnalyzIR toolbox. At the individual (1st) level, regression coefficients (β-weights) were determined from the hemoglobin dynamics and the convolution of the design matrix with a canonical hemodynamic response function, indicating brain activation. Group-level (2nd) statistics (β_pd and β_cn; cn: control, pd: peripheral artery disease) were obtained by fitting a mixed-effect model β contrast. The model was defined in a Wilkinson–Rogers formula of “β ∼ −1 + Group +(1| Participant),” and its output was evaluated by a t-contrast of [pd-cn] for all cognitive n-back tests. For further details see main text. fNIRS, functional near-infrared spectroscopy; HbO, oxyhemoglobin.

For the fNIRS examination, we used a NIRScout platform (NIRx Medical Technologies LLC, Glen Head, NY) equipped with 16 light sources and 16 detectors. To standardize optode placement, a 128-port Easycap head cap (Easycap GmbH, Woerthsee-Etterschlag, Germany) was positioned on the participant’s head by aligning its Fpz-Iz ports to the sagittal plane of the head and the optode in the Fpz position of the cap with Fpz on the participant. Hemodynamic responses were measured from 48 channels—defined by source-detector pairs with 3 cm separation according to the international 10-20 system—covering the prefrontal and motor cortex regions (Fig. 1B). Measurements were performed in a quiet, dimly lit room.

Assessment of Peripheral Microvascular Endothelial Function

To measure endothelial function in skin microcirculation, we used a flow-mediated dilation approach (25). In brief, a sphygmomanometer cuff was placed above the left antecubital fossa, and after recording 60 s of stable baseline perfusion, occlusion was performed by inflating the arterial cuff to 50 mmHg above current systolic blood pressure for 5 min. After the release of cuff occlusion, vascular responses were recorded for 3 min using a Perimed PSI System (Perimed, Järfälla, Sweden). Skin temperature was measured with a noncontact thermometer (Thermoworks TW2) from less than 10 mm at the back of the hand and the first and last phalanx of the middle finger. We determined baseline perfusion and maximal postocclusive perfusion from the recordings.

Data Processing and Statistical Analysis

The impact of PAD on cognitive performance was assessed by unpaired t test or Mann–Whitney test after checking the normality of the data set with Shapiro–Wilk test.fNIRS data were analyzed using a pipeline based on General Linear Model (GLM) approach created using the Brain AnalyzIR toolbox (41). After the conversion of optical densities to change of hemoglobin concentration using the modified Beer–Lambert law (42), prewhitening of data with an autoregressive model-based algorithm, and a discrete cosine transformation using a high-pass filter (cutoff frequency: 0.0045 Hz), design matrices were convolved with a canonical hemodynamic response function to predict brain activation. A mixed-effect model was fitted for group-level statistics using parameter estimates (β-weights) and scaling the predictors. The model was defined in a Wilkinson–Rogers formula of “β ∼ −1 + Group +(1 | Participant),” and its output was evaluated by a t-contrast of (PAD-Control) for the entire n-back test (i.e., contrast weights for each session was +1). Statistical significance was assumed for brain regions with a false discovery rate-corrected q < 0.05, according to the Benjamini–Hochberg method. A separate analysis was carried out for the channel between source AF3 and detector F5, which is located in the left dorsolateral prefrontal cortex (a region specifically involved in the n-back task). Averaged change of oxyhemoglobin (HbO) and reduced hemoglobin (HbR) was compared between PAD and control participants by Student’s t test for the duration of the n-back tasks (Fig. 1C).

Assessment of peripheral microvascular endothelial function was based on maximal postocclusion perfusion (normalized to baseline perfusion), which was then analyzed via Student’s t test comparing responses from PAD and control participants after checking the normal distribution of the data set.

Finally, to evaluate the association between cognitive performance, NVC responses, and peripheral microvascular endothelial function, we determined an n-back accuracy index as a measure of working memory function, as it has been used as a cognitive stimulus to evoke NVC responses captured by fNIRS. The n-back accuracy index describes the ability of study participants to discriminate between target and nontarget stimuli. This parameter is defined as the difference between z-transform of hit-rate (correct responses to target stimuli/number of target stimuli) and z-transform of false alarm rate (incorrect responses to nontarget stimuli/number of nontarget stimuli), as previously described (43). In this study, we calculated the average of the n-back accuracy indices obtained for the 1- and 2-back sessions. The association between n-back accuracy index, NVC responses (mean β values across all the channels of the prefrontal cortex for all the n-back tasks), and microvascular endothelial function (maximal postocclusion perfusion, normalized to baseline perfusion) was performed using linear regression analysis, and correlation matrices were built using the Jamovi software (v. 1.6) (44, 45).

Figures were prepared using MATLAB 2019 b (MathWorks, Natick, MA) and GraphPad Prism (v. 9, GraphPad Software, San Diego, CA). Tables were prepared using Jamovi software (46, 47). We report means ± SD for normally distributed variables, medians, and interquartile ranges otherwise.

RESULTS

PAD Alters Cognitive Performance

To assess cognitive performance, we used a battery of cognitive tests that has previously shown to be sensitive to age-related changes in healthy older adults free from dementia. All study participants were evaluated for performance in several domains of cognition, including psychomotor speed, attention, memory, executive function, and specifically, for reaction time, sustained attention, visual memory, short-term recognition, visual matching, and spatial working memory.

We observed significant group differences in the performance of a Delayed Matching to Sample test (Fig. 2A, median: 100, IQR: 80–100), indicating that PAD is associated with poorer short-term recognition performance in older adults, extending the results of previous studies (18). We also observed a significantly worse performance on a Rapid Visual Processing test in participants with PAD (Fig. 2B, means ± SD: 0.910 ± 0.0316), indicating that PAD is associated with lower levels of performance on tests of sustained attention and visual memory in older adults.

Figure 2.

PAD alters cognitive performance in older adults. A: we observed a significant decrease in the performance of delayed matching to sample test in participants with PAD versus controls, evidenced as a significantly lower percentage of correct responses with no delay (Mann–Whitney test, P = 0.004, n = 11 per group). B: we also observed a significantly decreased performance in the rapid visual processing test (0 to 1 from poor to good performance; two-sample unpaired t test, P = 0.038, n = 11 per group) in participants with PAD compared with age- and sex-matched controls. PAD, peripheral artery disease.

PAD Impairs NVC Responses in Prefrontal Cortex during Cognitive n-Back Task

We measured hemodynamic NVC responses in the cerebral cortex during the cognitive n-back paradigm using the fNIRS approach. Our data demonstrate that PAD significantly impairs NVC responses during the n-back task when compared with healthy sex- and age-matched controls. We found that the distribution of oxyhemoglobin was significantly lower in participating patients with PAD when compared with healthy controls in the entire prefrontal cortex (see negative contrast values in blue-colored areas in Fig. 3A). The biggest change was localized to the prefrontal cortex (PFC), including the left and right dorsolateral prefrontal cortices (LDLPFC and RDLPFC, respectively), brain regions involved in higher cognitive function, including attention, working memory, and inhibition of inappropriate responses. The β values from the channels of the PFC—including the left and right DLPFC—were averaged for control and participants with PAD and presented in Fig. 3B. No significant changes were observed for reduced hemoglobin between PAD and control groups during the cognitive task.

Figure 3.

PAD significantly impairs neurovascular coupling (NVC) responses during a cognitive task. A: representative oxyhemoglobin (HbO) curves are displayed for a randomly selected PAD patient (red) and control participant (blue), respectively. The underlying fNIRS recordings were acquired before and during 2-back task from a brain region (F3-FC3 channel) in the prefrontal cortex (PFC). B: group-level statistics β values from the prefrontal cortex (PFC), left and right dorsolateral prefrontal cortices (LDLPFC and RDLPFC, respectively) were averaged for all n-back tasks in control and participants with PAD (n = 5 per group, *P < 0.05 vs. controls, means ± SD). C: fNIRS heatmap demonstrates the effect of PAD on the distribution of HbO/NVC responses measured in the cortical tissues during cognitive n-back task vs. healthy controls. β values of NVC responses of healthy controls during all cognitive tasks were subtracted from corresponding tasks of participants with PAD (contrast PAD-controls). Blue or red shaded areas highlight cortical zones with the significantly lower or higher distribution of HbO/NVC responses (false discovery rate-corrected q < 0.05 for multiple comparisons) and demonstrate that NVC responses in the highlighted regions of participants with PAD were significantly lower than in healthy age- and sex-matched controls (n = 5 per group). No significant difference in the reduced hemoglobin (HbR) was observed between the two groups. fNIRS, functional near-infrared spectroscopy; HbO, oxyhemoglobin; PAD, peripheral artery disease.

NVC responses evaluated separately for 1- and 2-back tasks normalized for baseline 0-back task, separately for PAD and control participants, are presented in Fig. A1 and Fig. A2.

PAD Impairs Peripheral Microvascular Endothelial Function

We measured microvascular endothelial function in the skin microcirculation of the hand in response to 5-min blood flow occlusion in the brachial artery using the LSCI approach (Fig. 4A). We found that patients with PAD had significantly impaired peripheral microvascular endothelial function, evidenced by a significantly reduced postocclusive reperfusion (maximal/baseline perfusion; P = 0.0065, Fig. 4B).

Figure 4.

PAD significantly impairs peripheral microvascular endothelial function. A: representative recordings showing changes in microvascular perfusion in the left hand in response to occlusion and to a release of a 5-min blood flow restriction in the brachial artery of a participant with PAD (61-yr-old) and age- and sex-matched control participants. B: PAD significantly impaired microvascular endothelial function measured by maximal/baseline perfusion in the postocclusion period (Mann–Whitney test, P = 0.003, n = 11 per group). PAD, peripheral artery disease.

Association between Microvascular Endothelial Function, NVC Responses, and Cognitive Load

To study the association between cognitive performance, NVC responses, and peripheral microvascular endothelial function, we calculated the n-back accuracy index (average for 1- and 2-back tests), as it has been used to evoke NVC responses during fNIRS assessments. In control participants, we found a nonsignificant positive strong correlation between n-back accuracy index and NVC responses (r = 0.7), a nonsignificant positive weak correlation between n-back accuracy index and peripheral microvascular endothelial function (r = 0.2), and a nonsignificant positive strong correlation between NVC responses and peripheral microvascular endothelial function (r = 0.8) in control participants (Table 2). In patients with PAD, we observed a nonsignificant moderate positive correlation between n-back accuracy index and NVC responses (r = 0.4), nonsignificant moderate positive correlation between n-back accuracy index and peripheral microvascular endothelial function (r = 0.4), and a significant strong positive correlation between NVC responses and peripheral microvascular endothelial function (r = 0.9, P = 0.004, Table 3). These data suggest the association between NVC responses evoked by cognitive stimulation and peripheral microvascular endothelial function in the PAD group.

Table 2.

Cognitive load is associated with NVC responses and peripheral microvascular endothelial function in control participants

Control Group (n = 5)	n-Back Accuracy Index (Mean)	fNIRS (β, Mean)
fNIRS (β, mean)
Pearson’s r	0.708	-
P value	0.146	-
Peripheral microvascular endothelial function
Pearson’s r	0.233	0.843
P value	0.383	0.078

We found a nonsignificant strong positive correlation between n-back accuracy index and neurovascular coupling (NVC) responses (r = 0.70), nonsignificant positive weak correlation between n-back accuracy index and peripheral microvascular endothelial function (r = 0.23), and nonsignificant strong positive correlation between NVC responses and peripheral microvascular endothelial function (r = 0.84) in control participants (n = 5). fNIRS, functional near-infrared spectroscopy.

Table 3.

Cognitive load is associated with NVC responses and peripheral microvascular endothelial function in participants with PAD

PAD Group (n = 5)	n-Back Accuracy Index (Mean)	fNIRS (β, Mean)
fNIRS (β, mean)
Pearson’s r	0.422	-
P value	0.578	-
Peripheral microvascular endothelial function
Pearson’s r	0.422	0.996
P value	0.578	0.004

We observed a nonsignificant moderate positive correlation between n-back accuracy index and neurovascular coupling (NVC) responses (r = 0.4), nonsignificant moderate positive correlation between n-back accuracy index and peripheral microvascular endothelial function (r = 0.4), and a significant strong positive correlation between NVC responses and peripheral microvascular endothelial function (r = 0.9, P = 0.04) in peripheral artery disease (PAD) participants. These data suggest an association between NVC responses evoked by cognitive stimulation and peripheral microvascular endothelial function in the PAD group. fNIRS, functional near-infrared spectroscopy.

DISCUSSION

Cognitive decline has become a major healthcare burden in aging Western societies. It is recently recognized that vascular pathologies play a major role in the development of age-related cognitive decline and its progression to dementia (22). Relatedly, the National Institutes of Health (NIH) have initiated the consortium aimed to promote scientific efforts in the understanding of vascular contributions to cognitive decline (https://www.nih.gov/news-events/news-releases/nih-consortium-takes-aim-vascular-disease-linked-cognitive-impairment-dementia). However, the association between cerebromicrovascular pathologies with an accompanying neurovascular uncoupling and the increased prevalence of cognitive dysfunction observed in older adults is not fully understood.

With currently >8 million cases in the United States (48) and an exponentially increasing prevalence after the age of 60, peripheral artery disease (PAD) occupies a high rank among age-related morbidities. A growing body of evidence suggests that PAD exacerbates cognitive decline in older adults, increasing the risk for the development of dementia (12–16, 18). In line with these findings, our cohort of patients with PAD showed decreased performance in neuropsychological tests, including short-term memory and sustained attention. Considering that all study participants had at least a high school degree, we speculate that these differences can be attributed to cognitive dysfunction associated with the PAD.

PAD is associated with the damage to vascular endothelium (49). Recent studies suggest that the PAD-associated damage is not limited to large vessels, but that it also extends to microvascular beds, including cerebral microcirculation. Given that the brain is very sensitive to reduced nutrient supply, maintaining the optimal microenvironment of metabolically active brain regions critically depends on NVC. Furthermore, it is now established that microvascular endothelial dysfunction is directly linked to the impairment of NVC responses (23). Here, we establish an association between impaired neurovascular function and diminished cognitive function in older adults with PAD. We demonstrate that NVC responses are significantly impaired in the PFC (including LDLPFC and RDLPFC) during a cognitive task in patients with PAD when compared with age- and sex-matched controls. These brain regions are implicated in higher cognitive functions, including attention, working memory, and inhibition of inappropriate responses (26). Clinical and experimental evidence supports the concept that impaired NVC responses (50–52) may be causally linked to age-related cognitive decline (53, 54). First, clinical studies demonstrate that cognitive performance is strongly correlated with NVC responses (55, 56) and that improvement of NVC responses is associated with improved cognitive performance in older adults (55, 56). Furthermore, short-term disruption of hemodynamic NVC responses, even in healthy young individuals, is associated with deterioration of cognitive performance. For example, short-term sleep deprivation in healthy young adults is associated with reduced NVC responses measured in the cerebral cortex and impaired cognitive performance (57, 58). Second, preclinical studies demonstrate that pharmacological disruption of NVC responses impairs cognitive performance in healthy young mice, mimicking the aging phenotype (31, 59). Third, the rescue of NVC responses in aged mice with pharmacological treatments that confer antiaging endothelial effects improves cognitive function (53, 54). These findings suggest that NVC may be an attractive target for the development of pharmacological interventions to improve cognitive performance in older adults with vascular diseases, such as PAD.

There is ample evidence that pathological conditions that promote endothelial dysfunction [e.g., hypertension (60), diabetes mellitus (61–63), obesity (64–66)] in the peripheral circulation also impair NVC responses. Our studies provide additional support to the concept that PAD is associated with generalized microvascular endothelial dysfunction, which in turn is associated with neurovascular impairment. However, microvascular endothelial dysfunction is not only a direct consequence of atherosclerotic disease. Studies using ex vivo approaches demonstrate that circulating factors present in the sera of patients with PAD elicit proinflammatory and prooxidative phenotypic changes in endothelial cells (67–69). Thus, it is likely that the same factors that promote atherogenesis in the large arteries also lead to the development of generalized endothelial dysfunction and thereby impair NVC responses. The same mechanisms are also likely to be responsible for the association between PAD and coronary endothelial dysfunction (70). The circulating factors that promote generalized endothelial dysfunction in patients with PAD may include factors related to the known dietary and lifestyle risk factors of PAD [e.g., cigarette smoke constituents (71), dyslipidemia, inflammatory cytokines, and other mediators]. Importantly, studies in the field of geroscience have identified interventions that can reverse endothelial dysfunction, rescue NVC responses, and thereby improve cognitive function (53, 72), including mitochondrial antioxidants (54), polyadenosine diphosphate-ribose polymerase (PARP)-1 inhibitors (73), dietary polyphenols (74), and modified dietary eating patterns (75). Although many of these experimental treatments have been tested in clinical studies (76, 77), future studies are warranted to test the efficacy of such interventions in older adults with PAD.

Furthermore, we found that peripheral microvascular endothelial function was significantly impaired in participants with PAD when compared to healthy controls. Although microvascular endothelial impairment is not a direct consequence of atherosclerotic disease, they often accompany each other (49), and recent studies indicate that peripheral microvascular endothelial dysfunction may increase the incidence of dysregulated perfusion to limbs (78).

Although not significant, most likely due to a small sample size, we observed a medium-to-strong association between cognitive load, NVC responses, and peripheral microvascular endothelial function both in participants with PAD and in control participants. We also observed a significant and strong association between NVC responses evoked by cognitive stimulation and peripheral microvascular endothelial function in participants with PAD. Therefore, our findings provide initial evidence to support the theory that systemic endothelial dysfunction may extend from periphery to cerebral circulation to affect NVC responses and, possibly, to promote cognitive deficits in older adults with PAD (Fig. 5) (12, 15, 16, 79, 80). Our findings emphasize a need for further clinical studies to investigate the role of microvascular endothelial function in NVC responses and cognitive performance in aging and age-related diseases.

Figure 5.

Proposed mechanism by which PAD exacerbates the cognitive decline in aging. Based on our data, PAD exacerbates macro- and microvascular endothelial dysfunction in aging, which then extends to cerebral circulation, negatively impacting neurovascular coupling and promoting the cognitive decline in aging.

Conclusions

This study is the first, to our knowledge, to report the impairment of NVC responses in older adults with PAD. Our results suggest that PAD-related cognitive decrement may be mechanistically linked, at least in part, to generalized microvascular endothelial dysfunction and subsequent impairment of NVC responses. The clinical significance is that since endothelial dysfunction can be targeted pharmacologically, future studies are warranted to evaluate whether treatments, which improve NVC responses, can prevent or delay the onset of vascular cognitive impairment and dementia in older adults with PAD.

Limitations of the Study

The major limitation of our study is the small sample size, particularly with the NVC measurements. As with many clinical studies, there may have been a self-selection bias regarding study participation. Furthermore, the results of this study are only generalizable to patients with PAD with claudication. Although none of our study participants have been diagnosed with psychiatric or neurological disease, our study is limited because the familial medical history of psychiatric or neurologic has not been collected, nor it has been indicated in the electronic medical records. Furthermore, our study is limited because we were not equipped to evaluate carotid artery stenosis, which could be one of the contributing factors leading to the reduced cerebral circulation, nor this information was reflected in the electronic medical records. Finally, due to a small sample size, our study does not allow to control for smoking, comorbidities, and medications. However, recent evidence suggests that chronic smoking does not affect microvascular reactivity when measured in calves with fNIRS approach (81), and our participants abstained from smoking for at least 24 h before the assessments.

There are limitations associated with the design of the study and cross-sectional data analyses. Significant group differences and associations found in the variables do not provide evidence of causality. Although the association analysis demonstrates a positive correlation between key parameters measured, follow-up studies with larger sample sizes and longitudinal designs are needed to confirm these associations and potential changes over time. Although our study aimed to investigate hemodynamic changes in the brain during cognitive stimulation, the fNIRS methodology used in our study allows assessment of NVC responses limited to the cerebral cortex. Other methodologies such as transcranial Doppler flowmetry or functional MRI (fMRI) may provide additional insights into PAD-related hemodynamic changes in other deeper brain regions. Dynamic retinal vessel analysis may also provide additional insights into hemodynamic changes due to its ability to directly visualize NVC responses in retinal vasculature (26, 40, 51, 82).

GRANTS

This work was supported by American Heart Association Grants 941290, 966924, and 834339; National Institute on Aging Grants RF1AG072295, R01AG055395, R01AG068295, K01AG073614, and K01AG073613; National Institute of Neurological Disorders and Stroke Grant R01NS100782; National Cancer Institute Grant R01CA255840; Oklahoma Shared Clinical and Translational Resources Grant U54GM104938 with an Institutional Development Award from National Institute of General Medical Sciences; the Presbyterian Health Foundation; the Reynolds Foundation; NIA-supported Geroscience Training Program in Oklahoma Grant T32AG052363; Oklahoma Nathan Shock Center Grant P30AG050911; and Cellular and Molecular GeroScience CoBRE Grant P20GM125528. Project no. TKP2021-NKTA-47 has been implemented with the support provided by the Ministry of Innovation and Technology of Hungary from the National Research, Development and Innovation Fund, financed under the TKP2021-NKTA funding scheme.

DISCLAIMERS

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health, the American Heart Association, the Presbyterian Health Foundation, or the Ministry of Innovation and Technology of Hungary.

DISCLOSURES

A. Yabluchanskiy and Z. Ungvari serve as consulting editors for and A. Csiszar is an Editorial Board member of the American Journal of Physiology-Heart and Circulatory Physiology. None of the other authors has any conflicts of interest, financial or otherwise, to disclose.

AUTHOR CONTRIBUTIONS

C.D.O., P.M., T.C., A.L., F.S.-P., T.W.D., S.T., and A.Y. conceived and designed research; C.D.O., P.M., T.C., A.L., and A.Y. performed experiments; C.D.O., P.M., T.C., A.L., S.T., and A.Y. analyzed data; C.D.O., P.M., T.C., A.L., F.S.-P., T.W.D., S.T., A.W.G., P.S.M., S.R.W., J.M.K., A.N.-T, P.B., P.S., A.C., Z.U., and A.Y. interpreted results of experiments; C.D.O., P.M., T.C., A.L., and A.Y. prepared figures; C.D.O., P.M., T.C., A.L., S.T., and A.Y. drafted manuscript; C.D.O., F.S.-P., T.W.D., S.T., A.W.G., P.S.M., S.R.W., J.M.K., A.N.-T., P.B., P.S., A.C., Z.U., and A.Y. edited and revised manuscript; C.D.O., P.M.,T.C., A.L., F.S.-P., T.W.D., S.T., A.W.G., S.R.W., J.M.K., P.B., P.S., Z.U., and A.Y. approved final version of manuscript.

APPENDIX

Figures A1 and A2 demonstrate NVC responses measured during cognitive 1-back and 2-back stimulations.

Figure A1.

NVC responses measured during cognitive 1-back stimulation (vs. 0-back). Significant increase in oxyhemoglobin (HbO)/NVC responses was observed—shown as t-contrast map—in the right dorsolateral prefrontal cortex of the control group, whereas participants with PAD showed significantly reduced distribution of HbO/NVC responses in the prefrontal cortex including the left and right dorsolateral prefrontal cortices. We also observed a significant increase in reduced hemoglobin (HbR) in the prefrontal cortex including right dorsolateral prefrontal cortex of participating patients with PAD. Significant changes correspond to q < 0.05 (false discovery rate-corrected), n = 5 per group. NVC, neurovascular coupling; PAD, peripheral artery disease.

Figure A2.

NVC responses measured during cognitive 2-back stimulation (vs. 0-back). Significant increase in oxyhemoglobin (HbO)/NVC responses was observed—shown as t-contrast map—in the prefrontal cortex including the left and right dorsolateral prefrontal cortices in the control group, whereas participants with PAD showed significantly reduced distribution of HbO/NVC responses in the prefrontal cortex including the left dorsolateral prefrontal cortex. We also observed a significant increase in reduced hemoglobin (HbR) in the prefrontal cortex of participants with PAD. Significant changes correspond to q < 0.05 (false discovery rate-corrected), n = 5 per group. NVC, neurovascular coupling; PAD, peripheral artery disease.