Microvascular Function and Ambulatory Capacity in Peripheral Artery Disease.
Sep 4, 2025DOI: 10.1161/circinterventions.125.015582Clinical TrialAHA
Written by: Alexander E SullivanAdam BehroozianCrystal CoolbaughEmily ShardelowEmily K SmithQuinn S WellsDaniel G ClairAaron W AdayC Louis GarrardJohn A CurciTara A HolderJoey V BarnettMatthew S FreibergRachelle L CrescenziDenis J WakehamChristopher M HearonManus J DonahueJoshua A Beckman
Peripheral artery disease (PAD), affecting more than 200 million people globally, is among the primary causes of ambulatory dysfunction world wide.1 PAD is defined by atherosclerotic stenoses of the lower extremities sufficient to reduce ankle systolic arterial perfusion pressure below 90% of brachial artery systolic pressure as expressed by the ankle-brachial index (ABI).2Large artery occlusive disease and resultant blood flow restriction during ambulation is thought to cause intermittent claudication through the development of repeated skeletal muscle ischemia with attendant muscle wasting, fibrosis, and mitochondrial injury.3 Recurrent episodes of ambulation-associated ischemia increase oxidative stress in patients with PAD and may mediate these adverse adaptations.4 However, the association between the ABI and ambulatory performance is modest when associated with incline treadmill walking, 6-minute walk testing, and disease specific health status questionnaire evaluation, suggesting that artery occlusive disease and degree of blood flow restriction incompletely explains limitations in ambulation.5–7Lower extremity skeletal musce capillary density is reduced in patients with PAD, but the mechanistic relationship between microvascular dysfunction and ambulatory impairment is unknown.8 In other vascular beds, such as the myocardium, microvascular function plays a critical role in end-organ dysfunction, and better predicts functional capacity than conduit artery occlusive disease.9,10 In patients with PAD, calf conductance after ischemia, an index of microvascular vasodilation, has been shown to associate with walking performance, although the influence of blood flow restriction was not assessed.11Understanding lower limb vascular dysfunction may help to inform the mechanisms of claudication and explain ambulatory improvement seen following two different PAD interventions, exercise and revascularization therapies. Therefore, we measured vascular function at rest and after ischemia across the spectrum of disease, from severe arterial occlusive disease to healthy controls. We hypothesized that microvascular function following ischemia will associate with ambulatory performance in patients with PAD.
Materials and Methods
This was a prospective clinical trial (NCT03490968) that was designed to investigate the underlying mechanisms of ambulatory dysfunction in patients with PAD. A brief description of the trial protocol can but found on clinicaltrials.gov.12 The institutional review board at Vanderbilt University Medical Center approved the study protocol. All participants provided written, informed consent. Enrollment began July 2018 and concluded in August 2023. Data will be made available upon appropriate request to the corresponding author.Patients with PAD and age-matched controls were recruited from vascular medical and surgical clinics, the vascular imaging laboratory, advertising postcards, and ResearchMatch.org. Participants were ineligible to participate if they were age < 50 years, had a recent (≤ 6 months) myocardial infarction, stroke, limb revascularization, ischemic ulceration or rest pain consistent with chronic limb threatening ischemia, severe chronic kidney disease (estimated glomerular filtration rate < 30 mL/min/1.73 m2), cirrhosis or active hepatitis, or exercise limitation aside from claudication. PAD was defined by an ABI ≤ 0.85 in the index leg to increase the specificity for PAD diagnosis.13 As most persons with PAD do not have classic symptoms of intermittent claudication, there was no requirement for specific symptoms.14 Inclusion and exclusion criteria were otherwise the same for healthy control and PAD particpants (Table S1). Only one leg per participant was analyzed in the trial, and, if both legs met inclusion criteria for the trial, the leg that had the lower ABI was included.The study was prospectively structured in two complementary phases. First, every enrolled patient underwent a baseline evaluation that represents the cross-sectional dataset. Second, the subset enrolled for supervised exercise therapy or scheduled for revascularization entered a longitudinal arm, allowing us to reassess vascular function 12 weeks later and quantify within-subject changes relative to baseline.All potential participants presented for a screening visit to ensure eligibility (Figure S1). This visit included a full history and physical exam, laboratory testing, electrocardiogram, and ABI assessment. Those that met all inclusion criteria returned for a baseline testing visit. A testing visit consisted of a history, physical exam, MRI of the index limb, and walk testing, including both objective (6-minute walk test; 6MWT) and subjective (Walking Impairment Questionaire; WIQ) measures. After initial testing, participants with PAD were referred for either supervised exercise therapy (SET) or revascularization as previously determined by their treating physician. Those participants that completed SET or underwent revascularization returned for a second testing visit 12–13 weeks after their initial visit. Follow-up testing occurred according to the same protocol as the initial testing. Individual components of the testing visits are described below.A subgroup of PAD participants were recruited for supervised exercise therapy. A second group underwent revascularization per the treatment plan of their primary provider. Exercise therapy consisted of a one-hour session three times per week for 12 weeks at a certified rehabilitation center in accordance with current standards, as previously described.15 Revascularization, including either endovascular intervention or surgical bypass was performed at the discretion of the treating physician. Follow-up visits occurred within a week of completion of supervised exercise therapy or 12-weeks after definitive revascularization. Due to the pandemic, there were subjecta who could not complete exercise therapy or return after revascularization. Only those who completed therapy and both study visits were included in the post-intervention analysis.A handheld Doppler probe (MediChoice MC150, Owens&Minor, Alpharetta, GA) was used to measure systolic blood pressures in the right and left brachial, dorsalis pedis, and posterior tibial arteries according to standard technique.2Participants walked up and down a 100-foot hallway after receiving instructions to travel as far a distance as possible in six minutes.14 Identical instructions were provided to all participants by a research coordinator using a standard script. After six minutes, the distance completed in meters was recorded.The WIQ, a PAD-specific measure of self-reported limitations with three domains: walking distance, walking speed, and stair climbing was administered to all participants.16To interrogate the relationship between vascular function and walking distance, three measures of flow through the peripheral vasculature were used. First was tibial artery ankle pressure at rest in the supine position, as measured by the ABI.2As resting hemodynamics are not necessarily reflective of physiology during and after walking, hemodynamics were also assessed using magnetic resonance imaging (MRI). To assess the hemodynamic changes, MRI was performed at a field strength 3.0 Tesla (T) (Philips Healthcare, Best, The Netherlands) using a dual-channel body coil transmission and phased array knee coil for reception. Participants were positioned supine, leg-first in the scanner, and leg positioning was maintained such that the calf was equivalent in height as the heart and secured using non-magnetic foam padding and loop-and-hook fasteners.A standard non-contrasted anatomical imaging protocol was applied at baseline consisting of 2D T1-weighted turbo-spin-echo and 2D T2-weighted fat-suppressed turbo-spin-echo for tissue characterization and localization of anatomy for subsequent functional scans. For functional assessments, post-ischemic sphygmomanometric cuff occlusion vascular responses were assessed in both the below-the-knee macrovascular arteries (anterior tibial, posterior tibial, and peroneal artery) and calf skeletal muscle microvasculature using multiple MRI measures. Sphygmomanometric cuff occlusion is a reliable means to reproduce ischemia and the post-ischemic vascular response to ambulation in patients with peripheral artery disease.17–19 Baseline perfusion and hyperemic reactivity data were acquired using a cuff-occlusion protocol applied during a dual-echo arterial spin labeling (ASL) sequence (Figure S2). An MRI-compatible sphygmomanometric cuff was applied to the proximal calf and inflated to a suprasystolic blood pressure of 50 mmHg greater than the participant’s resting index limb systolic blood pressure. MRI data were acquired for one minute prior to cuff occlusion at baseline, five minutes during cuff occlusion, and five minutes following cuff deflation. The dual-echo sequence (TE1=17 ms; TE2=37 ms; inversion time=1050 ms; spatial resolution=3.3 × 3.3 × 7 mm) was applied with a single-shot gradient-echo echo-planar-imaging (EPI) readout at a temporal resolution (repetition time) of 1700 ms. The short TE1 label and control acquisitions enable absolute perfusion quantification in units of ml/100g/min and the control acquisitions for the long TE2=37 ms provide blood oxygenation level-dependent (BOLD) contrast.20 The sequence was applied over five axial slices and a fiducial marker was placed at the most distal aspect of the blood pressure cuff.Anatomical data were used primarily for planning and are presented qualitatively for purposes of localization and visualizing extent of underlying pathology.Functional data were corrected for slice timing and drift; perfusion-weighted images were generated by pair-wise subtraction of the label from control ASL preparations using the short TE acquisition to reduce susceptibility weighting. BOLD images were generated from the long TE acquisition without spin labeling preparation (e.g., control preparation). Perfusion maps were quantified in absolute units upon solving the simplified flow-modified Bloch equation incorporating tissue T1=1000 ms and labeling efficiency=0.95.20A shared region of interest was segmented from calf muscle parenchyma (combined tissue and microvasculature) on all slices, which were centered at the level of approximately the largest girth of the gastrocnemius muscle; voxels with signal from bone were excluded. For macrovascular analysis, endovascular signal localizing primarily to the anterior tibial, posterior tibial, and peroneal arteries was identified and used to create a single macrovascular artery mask. Within this macrovascular region, and separately for the regions within the calf muscle but excluding the macrovascular region, mean signal was recorded separately for perfusion- and blood oxygenation-weighted scans during the initial baseline, cuff-occlusion, and post-occlusion epochs.It should be noted that contrast changes within the macrovascular mask versus surrounding muscle mask derive from different physiological effects. Specifically, signal changes that occur on the BOLD scan within the macrovascular mask derive predominately from blood inflow effects (e.g., change in volume of blood per unit time). This occurs because traditional BOLD effects occur from changes in deoxy-hemoglobin relative to oxy-hemoglobin, which in turn alter the water relaxation time in and around capillaries and veins in a manner that depends on the perfusion (e.g., rate of blood delivery to tissue) relative to oxygen consumption. Within the pre-capillary conduit arteries, arterial blood is nearly fully oxygenated and oxygenation status is largely stable, therefore, changes that occur after cuff-induced hyperemia result from a larger influx of blood water itself, which is a reflection of macrovascular flux. In this study, we define such flux effects within the conduit arteries as macrovascular flow versus the BOLD and ASL changes within the muscle mask as microvascular effects, and it should be noted that such microvascular effects are directly localized to the capillaries for ASL or capillaries and downstream venules for BOLD. As such, macrovessels in this study refer to conduit arteries of several millimeters whereas microvascular effects reflect capillary and venule changes that are typically submillimeter but span a range of vessel sizes. Finally, while ASL and BOLD microvascular contrast localizes to the capillary and venules predominately, the perfusion changes that give rise to these effects are dependent on compliance of the arterioles, which contain smooth muscle and predominately regulate local blood delivery.First, the quantitative perfusion maps were considered. Within the calf muscle parenchyma, we recorded the (i) initial resting perfusion (baseline perfusion; ml/100g/min), (ii) perfusion during the final minute of cuff-occlusion (cuff-occlusion perfusion; ml/100g/min), and (iii) maximal perfusion change from baseline to cuff release (maximum calf skeletal muscle microvascular reactivity; ml/100g/min). Note that as different participants may have different perfusion baselines, the perfusion reactivity is defined as the absolute change from baseline, rather than the absolute change from the occlusion epoch as perfusion during occlusion is approximately null in all individuals and largely insensitive to inter-subject variation in baseline physiology.Next, the BOLD reactivity maps were considered. Within the calf muscle parenchyma and during the post-occlusion epoch, the maximal signal change relative to baseline (BOLD maximum calf skeletal muscle microvascular reactivity; fractional signal) was assessed. Note that for both perfusion and BOLD, we assessed maximal reactivity to reduce confounds from variation in response times, as is common in other applications of MRI-based reactivity.21Physiologic mechanisms by which perfusion and metabolism changes occur are complex and can be mediated by macrovascular arterial flow and changes in arteriolar blood volume. Rather than ignore these complex and dynamic contributions, these methods allowed modeling of both inflow as well as more microvascular perfusion and BOLD reactivity in the calf muscle. Inflow effects in the macrovasculature (i.e., conduit vessels visible at the spatial resolution of the BOLD experiment) were assessed in the short-to-intermediate repetition time BOLD data while the ASL and parenchymal BOLD effects provide an estimate of downstream extravascular or more microvascular effects.The primary statistical objective was to evaluate whether microvascular BOLD reactivity is an independent predictor of function, defined by the 6MWT. A secondary objective was to perform an assessment in a subgroup of participants to understand whether microvascular BOLD reactivity or macrovascular inflow changes paralleled functional improvements resulting from exercise therapy and revascularization.Continuous descriptive and clinical features are presented as median (IQR) and categorical measures as frequency and percent. Normality assessments were performed using Q-Q plots and quantified with a Shapiro-Wilk test (normality criteria: p>0.05). First, univariate linear regression analyses were performed to understand relationships between variables of interest. To confirm perfusion was reduced to null by the conclusion of the occlusion epoch, a t-test was applied to quantitative perfusion values for the final one minute of cuff occlusion. To confirm correspondence between maximal perfusion and BOLD reactivity responses, Pearson correlation testing was applied between each measure across all enrollment scans. To evaluate the primary study hypothesis that quantitative reactivity metrics provide an independent predictor of function, multivariable-adjusted linear regression was applied using the 6MWT as the dependent variable and (i) reactivity, (ii) age, (iii) sex, (iv) diabetes status, (v) group status (i.e., control or PAD), and (vi) ABI as independent variables. To compare the relative strength of each predictor on the same scale, we obtained the standardized regression coefficients (β). These analyses adjusted for population characteristics known to affect walking distance (e.g. age, ABI)22–25 but not for cardiovascular risk factors common in PAD that do not associate with walking distance (e.g., hypertension).26,27 To understand specifically whether findings were driven by group associations, a sensitivity analysis that repeated the above regression analysis using only participants with PAD was performed.Finally, to assess the secondary hypothesis of whether intervention, either exercise or revascularization, could alter reactivity profiles, we assessed the correlation between the percent change in function from the 6MWT and percent change in BOLD reactivity measures and ABI using Pearson correlation tests.In all analyses, significance criteria was defined as two-sided p<0.05. Data were analyzed using R version 4.2.2 (R Foundations for Statistical Computing).A total of 50 participants with PAD and 18 control participants were enrolled, with 28 PAD participants completing imaging and functional assessments before and after either exercise therapy (n=14) or surgical revascularization (n=14, Figure S3). Participant characteristics are presented in Table 1. Across all 68 participants evaluated at enrollment (i.e., prior to any intervention), ABI ranged from 0.23 to 1.4, with a median (IQR) of 0.57 (0.43, 0.65) in participants with PAD and 1.15 (1.09, 1.24) in controls. Participants with PAD were more likely to have a history of hypertension, cardiovascular disease, diabetes, and tobacco use. Use of statins and modulators of the renin angiotensin system were more common in those with PAD. Median 6MWT distance (314.7 vs 579.2, p<0.001) and WIQ total score (20.3 vs 100.0, p<0.001) were significantly different between those with PAD and healthy controls.Across the spectrum of limb perfusion pressure, walking distance diminished as the ABI declined (Figure 1A). The measured ABI correlated with 6MWT distance across the cohort (R=0.74, p<0.001). After controlling for age, sex, PAD diagnosis, and diabetes, ABI remained significantly correlated with 6MWT distance (β=280.1, p=0.013, Table 2). Results were also significant when considering only those with PAD (Table S2).Following the ischemic stimulus, the macrovascular blood inflow (i.e., blood inflow after cuff release) was reduced for patients versus controls (p=0.028) (Figure 1B). There was a significant correlation between post-ischemic macrovascular blood inflow response and 6MWT distance across the entire cohort (R=0.40, p<0.001). In a similar model controlling for age, sex, PAD diagnosis, and diabetes, post-ischemic macrovascular blood inflow significantly correlated with 6MWT distance (β=115.5, p=0.027, Table 2). Results were similar when considering only those with PAD (Table S2).Figure 2 shows group-level results, using enrollment data from 50 PAD and 18 control participants, for microvascular quantitative perfusion and BOLD reactivity responses in the skeletal muscle. During the final minute of cuff occlusion, maximum calf skeletal muscle perfusion was 1.5±3.7 ml/100g/min and was not significantly different from zero (p=0.7) in either cohort, confirming the effectiveness of the cuff occlusion to eliminate arterial blood flow. The maximal calf perfusion reactivity and maximal calf BOLD reactivity were significantly correlated with each other (R=0.52; p<0.001). Microvascular BOLD reactivity within the calf skeletal muscle was therefore used in all subsequent analyses of microvascular function given the higher temporal resolution and signal-to-noise ratio of BOLD compared to ASL and to reduce redundant comparisons.28There was a significant correlation between post-ischemic calf skeletal muscle BOLD microvascular reactivity and six-minute walking distance across the entire cohort (R=0.66, p<0.001). In a regression model controlling for age, sex, PAD diagnosis, and diabetes, post-ischemic calf skeletal muscle microvascular BOLD reactivity significantly correlated with 6MWT distance (β=1172.5, p=0.001, Table 2). Results were similar when considering only those with PAD (Table S2).A schematic of the location of functional imaging and quantified hemodynamic measures is summarized in Figure S2.Results from a multivariable regression model evaluating the association between 6MWT distance and all measures of vascular function (ABI, macrovascular blood inflow, and calf skeletal muscle microvascular BOLD reactivity) controlling for age, sex, PAD diagnosis, and diabetes are shown in Table 3. After controlling for both resting ankle perfusion pressure (ABI) and post-ischemic macrovascular blood inflow, calf skeletal muscle microvascular BOLD reactivity retained the strongest association with 6MWT distance (β=825.3, p=0.023) followed by ABI (β=203.8, p=0.033, Table 3). Macrovascular blood flow no longer associated with 6MWT (β=40.7, p=0.401) when the other two measures were included in the model. Similar findings were noted when considering PAD subjects alone (Table S3).We evaluated whether ABI, macrovascular blood inflow, and calf skeletal muscle microvascular BOLD reactivity improved and correlated with functional measures following clinically indicated therapies intended to improve ambulation. Table S4 shows baseline characteristics of those who returned following a PAD-specific intervention and Table S5 summarizes participant functional values before and after each intervention. The correlation between change in ABI, macrovascular blood inflow, and calf skeletal muscle microvascular BOLD reactivity is shown in Figure 3. There was no significant association between change in ABI (R=0.28, p=0.192) or post-ischemic macrovascular blood inflow (R=0.07, p=0.712) and improvement in 6MWT distance. In contrast, there was a significant association between calf skeletal muscle microvascular BOLD reactivity (R=0.46, p=0.014) and improvement in six minute walking distance.In this prospective mechanistic clinical trial, we sought to understand the association between microvascular and macrovascular function with ambulatory function before and after established therapies that improve walking. Our findings show that microvascular function is a stronger predictor of 6MWT distance than upstream macrovascular dysfunction, the latter measured as post-ischemic arterial inflow. Further, we show that improvements in microvascular function associate with improvements in ambulatory function in response to exercise and revascularization therapies. These data suggest that microvascular function, as opposed to in-line, macrovascular, large artery blood flow, may be a key mechanistic driver of limited limb function and contribute to functional improvement after both supervised exercise training and revascularization.The ABI provides insight into the severity of atherosclerotic disease in conduit arteries, but its relationship to walking distance is modest.7,11,24,29–33 While some have demonstrated a strong association with objective walking distance and patient-reported symptoms, ABI in these studies acts more to differentiate healthy controls from those with PAD.24,25,29,31 We see a similar effect in our cohort, as walking distance varies by study group until adjusting for ABI. ABI captures the disease-state implications of PAD, but the modest relationship with ambulatory function in the PAD subjects alone suggests that it fails to capture the full range of dysfunction in the vascular-skeletal muscle unit and the complex interplay of capillary density, metabolic changes, insulin resistance, and alterations in ATP generation that are characteristic of PAD.34–37 Furthermore, we show that measured peripheral blood flow at rest is similar in both controls and those with PAD regardless of atherosclerotic burden. Moreover with ambulation, the ability to supply blood flow (macrovascular blood inflow) did not associate with 6MWT when microvascular function was accounted.38,39 Supporting our findings, others have shown that neither resting ABI nor post-exercise ABI associate well with leg symptoms in PAD thus pointing to other vascular determinants of ambulation.7We demonstrated that the correlation between walking distance and microvascular reactivity was nearly ten-fold stronger than that of the increase in large artery blood inflow. Therefore, in our hands, microvascular function better predicts walking distance than resting ankle perfusion pressure and ischemia-induced increases in macrovascular blood flow through the major conduit arteries. This association is consistent not only across the entire spectrum of impairment from healthy controls to those with PAD, but also specifically within PAD participants after controlling for group (control vs PAD). This further corroborates the importance of the calf skeletal muscle microvasculature as a mechanistic driver of attenuated limb function by building on prior biopsy studies which demonstrated a similar relationship using different endpoints.34,40While the cardiology community has come to appreciate the importance of coronary microvascular dysfunction in the last two decades, in part due to both invasive and non-invasive diagnostic methods, these data suggest that skeletal muscle microvascular reactivity may be similarly relevant in patients with PAD and claudication. There is a strong relationship between both large artery disease and microvascular disease and subsequent adverse limb outcomes, but the combination of both together is synergistic.41 We show that while both ABI (large vessel occlusive disease) and microvascular reactivity associate with walking distance, only changes in microvascular reactivity associate with improvement in walking after revascularization or exercise therapy. Just as microvascular dysfunction predicts cardiac performance more reliably than epicardial coronary stenosis, skeletal-muscle microvascular function may be a principal driver of walking impairment in patients with PAD.42 Even amongst healthy sedentary persons, microvascular perfusion and density improves with training in paralell with exercise capacity.43Response to revascularization is variable, and microvascular function may influence symptomatic response more so than large artery flow and change in ABI.2 Similarly, semaglutide, a glucagon-like peptide 1 receptor agonist, improves pain free walking distance in patients with symptomatic claudication.44 This has been proposed to be through direct vascular effects and is supported by studies demonstrating improvements in inflammation and microvascular function.45 Furthermore, the implications of microvascular dysfunction extend beyond the specific vascular beds involved and has been shown to predict adverse limb outcomes even when vascular dysfunction is localized to the retinas, kidneys, and peripheral nerves.41 As the role of microvascular function in claudication becomes clearer, modalities such as MRI perfusion imaging, near-infrared spectroscopy, and pressure-wire measurements are poised to evolve from experimental tools into clinically relevant metrics of lower-extremity blood flow and vascular health.10,46Two therapies reliably improve walking capacity in patients with PAD: supervised exercise therapy and revascularization.47,48 The mechanisms by which these treatments improve claudication symptoms and walking distance remain poorly understood, but we provide evidence suggesting improvements in microvascular function are an important component. Interestingly, the increase in ABI and macrovascular blood inflow after revascularization does not well correlate with improvements in tested walking time or distance.49 It is, rather, the improvement in microvascular function that better correlates with walking distance even after a large vessel intervention. In contrast, prior imaging and skeletal muscle biopsies studies have demonstrated increased capillary density following exercise interventions precede and associate with functional improvment.50–54 This is a demonstrated mechanism in healthy persons and endurance athletes undergoing exercise training as well, thus supporting the importance of microvascular density across the range of conduit artery health and disease states.55While others have evaluated microvascular function and density in patients with PAD, we extend prior work by assessing unique relationships between microvascular function and ambulatory function. In a randomized controlled trial, Englund et al demonstrated improvements in both microvascular reactivity and time to onset of claudication in those referred for exercise therapy compared to usual care.51 Interestingly, in their experience, there was no relationship between improvements in functional capacity and MRI measures. One important difference between this study and ours was the functional measure employed. These investigators used treadmill walking time in contrast to our use of the 6MWT. To reach a consistent baseline treadmill walking time requires multiple visits by patients with PAD. Indeed, it has been shown that subjects with PAD will increase total walking time for 6 treadmill tests completed after baseline over 24 weeks without any treatment.56–58 In contrast, we used 6MWT which has not been associated with a significant learning effect generally or in PAD in particular.59,60 With a different functional measure not subject to a learning effect, we were able to show a direct correlation of improvements in ischemia-induced calf skeletal muscle microvascular reactivity after exercise or revascularization associates with improvements in function. While the current sample size precludes evaluating these therapies separately, these findings suggest a common mechanism to ambulation improvement linking the two therapies and serve as a necessary prerequisite to a more definitive clinical trial. Taken together these data suggest that these two interventions may share, in part, a common mechanistic pathway of functional improvement and enhance the link between microvascular function in limb function.There are limitations to the present study. The study was adequately powered to assess a relationship between microvascular function and walking impairment at the initial visit, but the exercise and revascularization arms had fewer subjects than anticipated. This was in part due to limitations imposed during the COVID-19 pandemic and the well reported poor adherence to supervised exercise therapy. This is a well-established issue with exercise therapy and has been extensively reported.61,62 For conventional BOLD, vessels parallel to the main magnetic field exhibit limited BOLD effects, whereas randomly oriented microvascular vessels exhibit more typical BOLD effects arising from reductions in paramagnetic deoxy-hemoglobin in capillaries and post-capillary vasculature, and corresponding intravascular and extravascular water relaxation time lengthening secondary to perfusion and oxygen extraction changes. We therefore quantified macrovascular effects not assuming typical BOLD contrast mechanisms, but in identified conduit arteries and using inflow effects (i.e., signal changes in the major supplying arteries) that arose from the increased flux of blood water after cuff occlusion, a metric that is independent of vessel orientation as it is proton density weighted rather than susceptibility weighted. Finally, the association of microvascular function with function independent of macrovascular blood flow is novel, and future studies are needed to confirm this in a different cohort and to further understand the physiological effects of exercise and revascularization therapies.Patients with PAD have reduced microvascular reactivity compared to controls and exhibit blunted and delayed hyperemic response to ischemia. Calf skeletal muscle microvascular BOLD reactivity associates with ambulatory distance independent of large artery ankle pressure at rest and macrovascular blood inflow post-ischemia. Following exercise or revascularization therapies, increases in calf skeletal muscle microvascular BOLD reactivity directly associate with ambulatory improvements, while ABI and macrovascular blood inflow do not correlate significantly with the change in function. This suggests that microvascular reactivity, more than macrovascular blood inflow, has a key role in both the development of functional impairment and subsequent improvement following exercise or revascularization.Supplemental Materials:Tables S1–S5Figures S1–S3