UBC-Nepal Expedition: imposed oscillatory shear stress does not further attenuate flow-mediated dilation during acute and sustained hypoxia

Abstract

Experimentally induced oscillatory shear stress (OSS) and hypoxia reduce endothelial function in humans. Acute and sustained hypoxia may cause increases in resting OSS; however, whether this influences endothelial susceptibility to further increases in OSS is unknown. Healthy lowlanders (n = 15, 30 ± 6 yr; means ± SD) participated in three OSS interventions: two interventions at sea level [normoxia and after 20 min of normobaric hypoxia (acute hypoxia, 11% O₂)] and one intervention 5–7 days after a 9-day ascent to 5,050 m (sustained hypoxia). OSS was provoked in the brachial artery using a 30-min distal cuff inflation (75 mmHg). Endothelial function was assessed before and after each intervention by reactive hyperemia flow-mediated dilation (FMD). Shear stress magnitude and patterns were obtained via Duplex ultrasound. Baseline retrograde shear stress and OSS were greater in acute hypoxia versus normoxia (P < 0.001), and OSS was elevated in sustained hypoxia versus normoxia (P = 0.011). The intervention further augmented OSS during each condition. Preintervention FMD was decreased by 29 ± 48% in acute hypoxia and by 25 ± 31% in sustained hypoxia compared with normoxia (P = 0.001 and 0.026); these changes correlated with changes in baseline mean and antegrade shear stress. After the intervention, FMD decreased during normoxia (−41 ± 26%, P < 0.001) and was unaltered during acute or sustained hypoxia. Therefore, a 30-min exposure to OSS reduced FMD during normoxia, a condition with an unchallenged, healthy endothelium; however, imposed OSS did not appear to worsen endothelial function during acute or sustained hypoxia. Exposure to an altered magnitude and pattern of shear stress at baseline in hypoxia may contribute to the insensitivity to further acute augmentation of OSS.

NEW & NOTEWORTHY We investigated whether the endothelium remains sensitive to experimental increases in oscillatory shear stress in acute (11% O₂) and sustained (2 wk at 5,050 m) hypoxia. Hypoxia altered baseline shear stress and decreased endothelial function (flow-mediated dilation); however, exposure to experimentally induced oscillatory shear stress only impaired flow-mediated dilation in normoxia.

INTRODUCTION

Shear stress, the frictional force exerted by blood on the arterial wall, plays a critical role in the short- and long-term regulation of endothelial function. Laminar blood flow and associated high, primarily antegrade, wall shear stress preserves or improves endothelial function (20). In contrast, disturbed blood flow and associated oscillatory shear stress (OSS), characterized by increased retrograde shear stress and a low mean shear stress, deteriorate endothelial function (12). Acute (30 min) (28, 47, 54, 55, 57, 58, 60) and sustained (2 wk) (55) exposure to an experimentally imposed OSS pattern via a distal external pressure cuff intervention typically reduces endothelium-dependent reactive hyperemia flow-mediated dilation (FMD) in conduit arteries of young, healthy men during normoxic conditions. However, the baseline flow pattern and level of endothelial function may influence vulnerability to FMD impairment via intervention-imposed OSS. For example, the cuff-OSS intervention does not appear to reduce FMD in elderly men (47, 55), a population that exhibits endothelial dysfunction and a sympathetically mediated enhancement of OSS at baseline (9, 11, 39).

Acute (<6 h) and sustained (7–90 days) hypoxic exposure confers unique hemodynamic challenges to the vascular system that may influence vulnerability to OSS-induced dysfunction. Initially, a reduction in arterial O₂ content provides a stimulus for both local vasodilation in the skeletal muscle (13, 36) and an increase in sympathetic nervous activity (30, 44, 45). Whether the net impact is 1) an increase in vascular resistance, accompanied by low mean and high conduit artery retrograde shear stress, or 2) a decrease in vascular resistance accompanied by an elevated mean and low retrograde shear stress appears to depend, at least in part, on the magnitude and duration of the hypoxic stimulus (24, 29, 32). More severe acute hypoxia stimulates greater sympathoexcitation (44, 45) and may be more likely to result in a disturbed flow pattern (11, 40), although this has not always been observed (25, 26). Furthermore, sustained hypoxia is associated with a chronic elevation in sympathetic nervous system activity and increased vascular resistance (22). The only study to report shear patterns during sustained hypoxia identified a markedly elevated retrograde shear rate at 5,050 m compared with sea level (62). Thus, both acute and sustained hypoxia may expose the endothelium to a lowered mean shear stress and increased OSS at baseline. The presence of disturbed baseline shear stress in acute and sustained hypoxia may reduce FMD and mitigate the responsiveness of peripheral conduit artery endothelial function to an experimentally induced exaggeration of OSS. Elucidating the acute and chronic interaction between hypoxia and shear stress patterns is critical to understanding the impact of environmental and clinical hypoxemia on vascular function in health and disease conditions.

The purpose of this investigation was therefore to determine the impact of intervention induced OSS on endothelial function during exposure to acute [<1 h, 11% fraction of inspired O₂ ($F{I}_{O_2}$)] and sustained (~2 wk, 5,050 m) hypoxia compared with normoxia. We hypothesized that regardless of duration, hypoxia exposure resulting in low mean and increased retrograde shear stress at baseline would attenuate the reduction in FMD following experimentally induced OSS compared with normoxia.

MATERIALS AND METHODS

This study was part of the University of British Columbia-Nepal 2016 expedition. As such, participants took part in multiple studies conducted at the University of British Columbia-Okanagan (Kelowna, BC, Canada; 344 m) and during 3 wk at the Ev-K2-CNR Pyramid Laboratory (Lobuche, Nepal; 5,050 m). However, the a priori, primary research questions addressed in the present report are novel and dealt with exclusively in this study alone; there is no overlap between this investigation and others completed on the research expedition.

Participants

Healthy, nonsmoking men (n = 15, 30 ± 6 yr, body mass index: 23 ± 2 kg/m², means ± SD) free from known cardiovascular disease and risk factors participated in the investigation. All participants were native lowlanders who had not been exposed to elevations >3,000 m for at least 3 mo before the expedition. All participants provided written, informed consent, and protocols were approved by the University of British Columbia Clinical Ethics Review Board and Queen’s University Health Sciences Research Ethics Board in adherence with the principles of the Declaration of Helsinki.

Study Design

The testing protocol was performed on three separate occasions: twice at sea level (University of British Columbia-Okanagan, Kelowna, BC, Canada; 344 m) and once 5–7 days after a 9-day hiking ascent to 5,050 m (Ev-K2-CNR Laboratory, Lobuche, Nepal; 5,050 m). All participants refrained from alcohol, exercise, and caffeine for a minimum of 12 h before testing and were fasted for a minimum of 6 h. All tests were performed in the morning to account for any diurnal variation in vascular function. The study outline is shown in Fig. 1 and was identical for sea level normoxia (room air), acute hypoxia ($F{I}_{O_2}$ = 0.11, similar Po₂ to 5,000 m), and sustained hypoxia (5,050 m). The normoxia and acute hypoxia trials were performed in a hypoxia chamber at the University of British Columbia-Okanagan, and the trial order was randomized, single blinded, and counterbalanced.

Fig. 1.

Schematic of the experimental protocol. The intervention was repeated three times: in normoxia, acute hypoxia, and sustained hypoxia. The illustration of the arm depicts the placement of the pneumatic cuff and its inflation level during the intervention and the illustration of the needle depicts the venous blood sample to measure whole blood viscosity. FMD, flow-mediated dilation.

After participants arrived at the laboratory, a venous blood sample was acquired to measure whole blood viscosity. Subsequently, participants laid supine for 20 min (at $F{I}_{O_2}$ = 0.11 in the acute hypoxia trial); resting brachial artery blood pressure, heart rate (HEM-775CAN, Omron Healthcare), and oxyhemoglobin saturation (Nonin Medical) were recorded, and the average in the last minute of rest is reported. Endothelial function was measured in the brachial artery via reactive hyperemia FMD in accordance with internationally accepted guidelines (53). A 1-min recording of baseline arterial diameter and blood velocity was recorded followed by a 5-min cuff occlusion (250 mmHg, distal to epicondyles). After cuff deflation, recording resumed for 3 min. All measurements were performed by the same experienced sonographer with a 10-MHz multifrequency linear array probe (15L4 Smart Mark, Teratech) attached to a high-resolution ultrasound machine (Terason usmart 3300, Teratech). After FMD, OSS was induced for 30 min as previously described (54, 57). One pneumatic cuff (SC5, Hokanson) was placed distal to the epicondyles and inflated to 75 mmHg to provoke OSS in the brachial artery. One-minute recordings of mean, antegrade, and retrograde shear stress were obtained at 10, 20, and 30 min during the intervention; the average of the three time points is reported as they have previously been reported to be stable during this OSS intervention (60). After the 30-min intervention, the cuff was deflated, arterial blood pressure, heart rate, and oxyhemoglobin saturation were recorded, and a second reactive hyperemia FMD was performed.

Blood Viscosity

Venous blood (5 ml) was drawn into a BD Vacutainer Blood Collection Tube (Becton Dickinson) that contained lithium heparin. Blood viscosity was measured within 15 min of blood sample acquisition at a shear rate of 225 s⁻¹ at 37°C with a cone-and-plate viscometer (Brookfield DV2T, Brookfield AMETEK).

Data Analysis

Baseline.

Pulse pressure was calculated as the difference between systolic and diastolic blood pressures, and mean arterial pressure (MAP) was calculated as (2 × diastolic blood pressure + systolic blood pressure)/3. Forearm blood flow was calculated from the 1-min baseline ultrasound recording as [peak envelope blood velocity (in cm/min)/2]/2 × {π[0.5 × diameter (in cm)]²} (16, 33). Forearm vascular resistance (FVR) was calculated as MAP/forearm blood flow.

Shear stress.

Shear stress was calculated as the product of shear rate (4 × peak envelope blood velocity/arterial diameter) and whole blood viscosity at a shear rate of 225 s⁻¹ (19). Antegrade, retrograde, and mean shear stress are reported. The oscillatory shear index (OSI) was calculated as |retrograde shear stress|/(|antegrade shear stress| + |retrograde shear stress|) (37).

Flow-mediated dilation.

Standardized software approaches to acquire and analyze the Doppler ultrasound recordings were used, as described extensively elsewhere (31, 54, 61). The angle of insonation for the acquisition of velocity was 60°. Screen capture of the ultrasound was saved as an audio video interleave file (Camtasia Studio, Techsmith) for future analysis using edge-detection software (FMD/BloodFlow Software version 5.1, Reed C) (65). A region of interest was placed around the highest quality portion of the B-mode longitudinal image of the artery and a second region of interest surrounded the Doppler strip to record blood velocity. The software automatically and continuously tracks the walls of the vessel and peak envelope velocity trace within the regions of interest at a frequency of 30 Hz (65). Peak diameter was automatically detected using a moving window-smoothing function (smoothed median across time) postcuff deflation. FMD was calculated as the relative and absolute difference between peak and baseline diameter. Previous studies have identified that with adherence to guidelines, relative FMD is reproducible across 1 wk and 1 mo (coefficient of variations of 10.6% and 6.6%, respectively) (14). From a previous investigation in young, healthy men (n = 10), the sonographer on this investigation (J. C. Tremblay) presented a between-day coefficient of variation of 8.3% (59). The FMD stimulus was quantified as the shear stress area under the curve (SSAUC) from cuff deflation to peak diameter (38).

Statistics

All statistical analyses were performed using IBM SPSS 24. Data were compared within participants with significance set at P < 0.05 and are presented as means ± SD. Baseline (preintervention) parameters (blood pressures, heart rate, oxyhemoglobin saturation, and blood viscosity) were assessed using a one-factor mixed linear model (normoxia vs. acute hypoxia vs. sustained hypoxia). To assess the effects of the intervention on shear stress patterns, mean, antegrade, and retrograde shear stress, and OSI were compared using a two-factor mixed linear model with repeated measures of time (preintervention vs. during intervention vs. postintervention) and condition (normoxia vs. acute hypoxia vs. sustained hypoxia). To test the effect of the OSS intervention during normoxia, acute hypoxia, and sustained hypoxia, FMD and SSAUC measurements were compared using a two-factor mixed linear model with repeated measures of time (pre- vs. postintervention) and condition (normoxia vs. acute hypoxia vs. sustained hypoxia).

To account for differences in the FMD stimulus, testing was also performed with SSAUC included as a covariate. When significant effects were present, post hoc tests were performed with Bonferroni adjustment for multiple comparisons. Furthermore, allometric scaling was performed to account for differences in baseline diameter between conditions. Briefly, the diameter change on a logarithmic scale [ln(peak diameter) − ln(baseline diameter)] was assessed as the outcome variable in a linear mixed model with logarithmically transformed baseline diameter included as a covariate (3, 4). Pearson product-moment correlation tests were performed between baseline and intervention shear stress parameters and FMD, and Cook’s distance was used to identify outliers (Cook’s distance > 4/n) (2, 41).

RESULTS

Baseline Measures

Baseline arterial blood pressures, heart rate, oxyhemoglobin saturation, and whole blood viscosity are shown in Table 1. Hypoxic conditions were characterized by a lower arterial O₂ saturation and a higher heart rate compared with normoxia. MAP, blood viscosity, and forearm vascular resistance were elevated during sustained hypoxia compared with normoxia and acute hypoxia (P < 0.01).

Table 1.

Baseline oxyhemoglobin saturation, whole blood viscosity, and hemodynamics

	Normoxia	Acute Hypoxia	Sustained Hypoxia	Main Effect of Condition
Oxyhemoglobin saturation, %	97 ± 1	77 ± 5*	84 ± 2†	P < 0.001
Viscosity225, cP	4.03 ± 0.35	4.03 ± 0.25	5.02 ± 0.38†	P < 0.001
Heart rate, beats/min	51 ± 10	64 ± 14*	58 ± 11†	P < 0.001
Systolic blood pressure, mmHg	114 ± 7	109 ± 8	125 ± 13†	P < 0.001
Diastolic blood pressure, mmHg	60 ± 6	58 ± 8	76 ± 7†	P < 0.001
Mean arterial pressure, mmHg	78 ± 6	75 ± 7	93 ± 8†	P < 0.001
Pulse pressure, mmHg	54 ± 7	51 ± 9	49 ± 9	P = 0.162
Forearm blood flow, ml/min	47 ± 29	43 ± 20	31 ± 15	P = 0.057
Forearm vascular resistance, mmHg·ml−1·min	2.22 ± 1.18	2.24 ± 1.16	3.68 ± 1.77†	P = 0.002

Values are means ± SE.

^*P < 0.05 vs. normoxia;

^†P < 0.05 vs. normoxia and acute hypoxia.

Baseline and intervention shear stress.

P values for main effects and interactions are shown in Fig. 2 and revealed that there were no significant interaction effects. P values in the following text refer to post hoc analysis.

Fig. 2.

Mean (A), antegrade (B), and retrograde shear stress (C) and oscillatory shear index (OSI; D) pre-, during, and postintervention in normoxia, acute, and sustained hypoxia. Symbols on the right in A–C indicate post hoc analysis exploring main effects of condition. Symbols on the graph indicate post hoc analysis exploring main effects of time. *P < 0.05 vs. pre- and during intervention; ‡P < 0.05 vs. pre- and postintervention; †P < 0.05 vs. sustained hypoxia; ^P < 0.05 vs. normoxia.

A main effect of condition was identified for antegrade shear stress, retrograde shear stress, and OSI. Antegrade shear stress was lower in sustained hypoxia versus acute hypoxia (P = 0.019). Retrograde shear stress was elevated during acute hypoxia compared with normoxia and sustained hypoxia (P < 0.001 and 0.001; Fig. 2B), and there was a trend toward greater retrograde shear stress in sustained hypoxia vs. normoxia (P = 0.067). Finally, OSI was elevated in both acute and sustained hypoxia versus normoxia (P < 0.001 and 0.011).

A main effect of time (levels pre-, during, and postintervention) was observed for all shear stress parameters. The intervention decreased mean shear stress, increased retrograde shear stress, and increased OSI without altering antegrade shear stress. Postintervention mean shear stress and antegrade shear stress were lower and OSI greater versus preintervention (P = 0.001, 0.022, and 0.004), whereas retrograde shear stress was similar pre- and postintervention (P = 0.332).

Baseline diameter and the shear stress stimulus for FMD.

All main effect and interaction P values are shown in Table 2. Baseline brachial artery diameter did not change with the intervention; however, it was larger in acute hypoxia compared with normoxia and sustained hypoxia (P = 0.001 and >0.001; Table 2). The SSAUC was smaller postintervention (P = 0.002) and larger during sustained hypoxia compared with acute hypoxia (P = 0.032).

Table 2.

Reactive hyperemia FMD parameters before and after 30 min of imposed oscillatory shear stress (cuff positioned distal to the epicondyles and inflated to 70 mmHg) during normoxia, acute, and sustained hypoxia

	Normoxia		Acute Hypoxia		Sustained Hypoxia
	Pre	Post	Pre	Post	Pre	Post
Baseline diameter, mm	4.36 ± 0.47	4.33 ± 0.41	4.57 ± 0.52*	4.56 ± 0.57*	4.30 ± 0.44	4.22 ± 0.44
	Time, P = 0.358; Condition, P < 0.001; Interaction, P = 0.826
Relative FMD, %	6.2 ± 2.6‡	3.6 ± 2.0†	3.7 ± 1.8	3.7 ± 2.5	4.3 ± 2.0	3.3 ± 1.5
	Time, P = 0.004; Condition, P = 0.027; Interaction, P = 0.033
Absolute FMD, mm	0.27 ± 0.11‡	0.15 ± 0.09†	0.16 ± 0.08	0.16 ± 0.10	0.18 ± 0.08	0.14 ± 0.06
	Time, P = 0.002; Condition, P = 0.021; Interaction, P = 0.026
Time to peak, s	51 ± 11	46 ± 21	54 ± 18	61 ± 25	55 ± 21	49 ± 18
	Time, P = 0.699; Condition, P = 0.179; Interaction, P = 0.324
Shear stress area under the curve, arbitrary units	848 ± 199	749 ± 218	815 ± 352	699 ± 261	1,005 ± 300§	796 ± 260§
	Time, P = 0.002; Condition, P = 0.031; Interaction, P = 0.556

FMD, flow-mediated dilation; Pre, before 30 min of imposed oscillatory shear stress; Post, after 30 min of imposed oscillatory shear stress.

^*P < 0.05 vs. normoxia and sustained hypoxia;

^†P < 0.05 vs. normoxia pre;

^‡P < 0.05 vs. acute hypoxia pre and sustained hypoxia pre;

^§P < 0.05 vs. acute hypoxia.

Flow-mediated dilation.

The P values for main effects and interactions are shown in Table 2. There was a significant interaction between condition and time. Post hoc analysis of the effect of the intervention in each condition identified that absolute and relative FMD was decreased postintervention in normoxia (by 41 ± 27% and 41 ± 26%, both P < 0.001) but not during acute (P = 0.963 and 0.985) or sustained hypoxia (P = 0.109 and 0.145; Table 2 and Fig. 3). These findings persisted after including SSAUC as a covariate (normoxia, P = 0.001; acute hypoxia, P = 0.405; sustained hypoxia, P = 0.856) and allometric scaling (normoxia, P < 0.001; acute hypoxia, P = 0.953; sustained hypoxia, P = 0.067).

Fig. 3.

Mean and individual reactive hyperemia flow-mediated dilation (FMD) pre- and postintervention during normoxia (A), acute (B), and sustained hypoxia (C). *P < 0.05, preintervention vs. postintervention.

Post hoc analysis comparing preintervention FMD between conditions also determined that absolute and relative FMD was greater preintervention in normoxia compared with preintervention acute hypoxia (27 ± 48% and 29 ± 48%, P = 0.002 and 0.001) and preintervention sustained hypoxia (26 ± 30% and 25 ± 31%, P = 0.011 and 0.026; Fig. 4A). These findings persisted when SSAUC was included as a covariate (normoxia vs. acute hypoxia, P = 0.001; normoxia vs. sustained hypoxia, P < 0.001) and allometric scaling (normoxia vs. acute hypoxia, P = 0.005; normoxia vs. sustained hypoxia, P = 0.009). Furthermore, the changes in preintervention FMD in normoxia versus acute hypoxia and in normoxia versus sustained hypoxia were correlated (Fig. 4C; one outlier removed, r = 0.745, P = 0.002).

Fig. 4.

Preintervention flow-mediated dilation (FMD) from normoxia to acute hypoxia (A) and sustained hypoxia (B). The change in preintervention FMD in normoxia vs. acute hypoxia was correlated with the change in preintervention FMD in normoxia vs. sustained hypoxia (C). One outlier, identified as having a Cook’s distance > 4/n (0.267), was excluded from regression analysis (dashed circle). *P < 0.05, preintervention normoxia vs. preintervention acute hypoxia and preintervention sustained hypoxia.

Correlations Between Changes in Shear Stress and FMD

Did preintervention hypoxia-induced changes in baseline shear stress predict hypoxia-induced changes in FMD?

The change in baseline antegrade and mean shear stress from normoxia to acute hypoxia correlated with the change in FMD from normoxia to acute hypoxia (Fig. 5, A and B). In contrast, the change in baseline OSI and retrograde shear stress and the change in FMD were unrelated (Fig. 5, C and D). In sustained hypoxia, with one outlier removed, the change in mean and antegrade shear stress, as well as OSI from normoxia, correlated with the change in FMD from normoxia, whereas the change in retrograde shear stress did not.

Fig. 5.

Change in preintervention baseline mean (A), antegrade (B), and retrograde (C) shear stress and oscillatory shear index (OSI; D) and preintervention flow-mediated dilation (FMD) from normoxia to acute (circle, solid line) and sustained hypoxia (square, dashed line). One outlier, identified as having a Cook’s distance > 4/n (0.267), was excluded from regression analysis in sustained hypoxia for mean and antegrade shear stress and OSI (dashed circle).

Did changes in baseline shear stress with hypoxia predict intervention-induced changes in FMD under hypoxia conditions?

There was a correlation between the changes in OSI and retrograde shear stress from normoxia to sustained hypoxia and the change in FMD evoked by the intervention in sustained hypoxia (r = 0.533, P = 0.041; one outlier removed, r = −0.772, P = 0.001), such that individuals with the greater perturbation in the shear pattern demonstrated smaller responses to the intervention; these correlations were not present in acute hypoxia (r = −0.433, P = 0.107; r = 0.184, P = 0.511).

DISCUSSION

This study sought to determine the impact of hypoxia exposure on vulnerability to acute OSS-induced endothelial dysfunction. The primary novel finding was that imposed OSS decreased brachial artery FMD during normoxia but not acute (<1 h) or sustained (~2 wk) hypoxia. Hypoxia resulted in a decline in FMD that was correlated with the hypoxia-associated change in baseline shear stress. The existence of a low mean, mildly OSS at baseline during hypoxia may limit any further reduction in endothelial function by experimentally imposed OSS (Fig. 6). These findings advance our understanding of the interaction between hypoxia exposure, shear stress patterns, and endothelial function in humans.

Fig. 6.

An illustrative summary of the findings. Experimentally induced oscillatory shear stress (OSS) decreased flow-mediated dilation (FMD) in the brachial artery at sea level but not in acute or sustained hypoxia. The alterations in baseline shear stress observed in acute and sustained hypoxia may have contributed to the decreased FMD observed in acute and sustained hypoxia, leading to an insensitivity to further increases in OSS.

Impact of Imposed OSS in Normoxia

The observation of a 41% decrease in FMD after OSS during normoxia supports the theory that disrupting local shear stress patterns, for just 30 min, incurs an impairment in endothelium-dependent vasodilation. In agreement with our normoxia condition observation, most studies using the forearm cuff inflation intervention to exaggerate OSS have reported transient reductions in brachial artery FMD in healthy young men (28, 46, 54, 55). Postintervention decrements in endothelial function have also been reported in the lower limb (46, 58) and in individuals with spinal cord injury (57) and when assessed as the increase in circulating endothelium-derived microparticles, a distinct marker of endothelial dysfunction (6, 27, 60). In contrast, populations that have high baseline retrograde shear stress display blunted (6) or no impairment (47, 55) in FMD after the OSS intervention, suggesting the necessity of an endothelium that is naïve to disturbed shear stress and/or free from preexisting dysfunction to elicit the disruption.

Impact of Imposed OSS in Hypoxia

In conditions of both acute and sustained hypoxia, FMD was not impaired by the OSS intervention. The reduction in FMD at baseline in acute and sustained hypoxia may have rendered the endothelium insensitive to further OSS-induced impairment. The changes in baseline mean and antegrade shear stress were related to the change in baseline FMD in both acute and sustained hypoxia, suggesting that hypoxia-associated alterations in shear stress may have contributed to the observed reduction in FMD. Furthermore, greater increases in OSI and retrograde shear stress at baseline in the sustained hypoxia condition predicted smaller intervention-induced changes in FMD, suggesting that the presence of relatively elevated OSS at baseline limited a further negative impact of the OSS intervention on FMD in sustained hypoxia. Collectively, these correlations indicate that disturbed baseline shear stress with hypoxia influences the impact of hypoxia on FMD. The presence of disturbed shear stress at baseline, perhaps particularly in sustained hypoxia, may influence whether FMD will be impaired with further exaggeration of OSS in hypoxia.

In a recent expedition, Tremblay et al. (60) performed a similar OSS intervention during a subacute duration hypoxia exposure (24–48 h) after a passive, rapid ascent to 3,800 m. In contrast to the present study, it was reported (60) that 1) preintervention baseline shear rate and FMD were preserved with ascent and 2) there was a reduction in FMD after the OSS intervention at 3,800 m but not at sea level. These observations suggest that this cohort displayed enhanced (rather than reduced) sensitivity to OSS at 3,800 m. The conflicting findings may be due in part to the influence of hypoxia on baseline shear stress. For example, likely due to a more modest severity of hypoxia at 3,800 m (32) compared with the present study, Tremblay et al. (60) did not observe changes in resting mean shear rate or shear rate pattern. Thus, whether the endothelium remains naïve to disturbed shear stress at baseline may determine the influence of hypoxia on acute OSS exposure. This hypothesis requires further exploration as other factors related to the magnitude and duration of hypoxia exposure may also be involved.

Mechanisms Underlying the Impact of Hypoxia on the Baseline Shear Stress Pattern

We observed an increase in retrograde shear stress and OSI during acute and, to a lesser extent, in sustained hypoxia compared with normoxia. One possible explanation for the influence of hypoxia on shear patterns may be the dose-dependent chemoreflex-mediated sympathoexcitation. Increases in muscle sympathetic nerve activity manifest rapidly (30, 44, 45, 51) and persist with sustained and chronic hypoxia (15, 22, 34). It has been established that α-adrenergic receptor blockade increases acute hypoxic forearm blood flow (63, 64), providing support that hypoxia-mediated sympathetic vasoconstriction may influence upstream conduit artery hemodynamics; however, acute hypoxia is associated with local vasodilation that offsets the sympathoexcitation (13). The balance of local vasodilation and sympathetic vasoconstriction alters arteriolar tone and hence the critical pressure that must be overcome by arterial pressure to prevent the collapse of downstream vessels (48, 49). Retrograde flow has been hypothesized to manifest when arterial pressure falls below the critical pressure (21). For example, sympathoexcitatory maneuvers can decrease mean shear rate and increase retrograde shear rate and OSI (40, 52), although sufficient increases in MAP (perfusion pressure) may mitigate these changes (40). In acute hypoxia, MAP was nonsignificantly (P = 0.240) lower versus normoxia, and nonsignificant decreases in MAP alongside a maintained FVR may explain the increase in retrograde shear stress. The lower retrograde shear stress observed in sustained hypoxia compared with acute hypoxia may be due to the increased MAP in conjunction with (and “overcoming”) the increased FVR. Thus, the degree of retrograde shear stress in the acute and sustained hypoxia conditions may be the result of different combined alterations in MAP and FVR, which involve hypoxia induced changes in sympathoexcitation and compensatory vasodilation (8, 35, 64).

Recently, a separate investigation from the same expedition reported that short, mild sympathoexcitation and sympathoinhibition (via lower body negative and positive pressure, respectively) did not alter shear rate magnitude or pattern or FMD at sea level or after 11–14 days at high altitude (62). Whether greater levels of sympathoinhibition (i.e., local pharmacological receptor blockade) ameliorates shear stress and FMD under hypoxic conditions is unknown and requires future investigation.

Impact of Hypoxia on FMD

As indicated above, we observed a significant decline in FMD in acute and sustained hypoxia compared with normoxia; the mean relative decreases in FMD in acute (−29 ± 48%) and sustained hypoxia (−25 ± 31%) were broadly similar to those observed on a similar research expedition to the Ev-K2-CNR Pyramid Laboratory in 2012 [acute hypoxia ($F{I}_{O_2}$ = 0.11): −28%, sustained hypoxia: −8%] (31). Indeed, the change in baseline antegrade and mean shear stress correlated with the change in preintervention FMD, suggesting that low shear stress is associated with reductions in FMD during acute and sustained hypoxia. Thus, the influence of hypoxia on FMD may be mediated indirectly by its effect on shear stress pattern/magnitude. In addition, changes in FMD from normoxia to acute and sustained hypoxia correlated with one another (Fig. 4C), suggesting that individual vulnerability to hypoxia associated endothelial dysfunction is somewhat consistent over time with prolonged exposure. Variability of FMD in response to hypoxia in lowlanders is common, with some studies reporting reductions (5, 17, 18, 31, 32, 62), and others reporting no change in FMD (10, 23, 43, 60, 61). Factors including cardiovascular risk (17, 18), duration and severity of hypoxia (10, 32), prophylaxis (5), active versus passive ascent (10, 42), isocapnic versus poikilocapnic hypoxia (31, 32), FMD protocol (18, 23), and differences in FMD stimulus (5, 31, 60) likely contribute to the nonuniform findings regarding FMD in hypoxia. To assist in addressing the latter point, we measured blood viscosity to calculate shear stress, as opposed to providing shear rate as a surrogate. Sustained hypoxia reduces plasma volume and increases red blood cell volume (50), eliciting increases in hematocrit and subsequently whole blood viscosity. Indeed, without accounting for viscosity, the SSAUC during sustained hypoxia would have been underestimated by ~25%, highlighting the importance of measuring blood viscosity during sustained hypoxia. Future investigations should seek to elucidate the relationship between shear stress and FMD under hypoxic conditions via controlled, stepwise alterations in shear stress [i.e., progressive handgrip exercise (66) and/or skin heating (7)].

Methodological Considerations

We did not assess endothelium-independent vasodilation (i.e., sublingual nitroglycerine). Endothelium-independent vasodilation is not influenced by acute perturbations in shear stress (56); however, a decrease in endothelium-independent vasodilation has been observed in acute (32) and sustained hypoxia (31). Thus, it is possible that a decrease in vascular smooth muscle function contributed to the reduction in FMD with hypoxia. Another limitation was the lack of a time-control condition. However, studies using a similar intervention have found no change in FMD in the contralateral (i.e., control) arm after 30 min of rest (28, 54, 57). Furthermore, although the presence of retrograde shear rate at the vessel wall has been observed under resting conditions in the brachial artery (67), in the present study shear stress was derived from the whole vessel cross section and it is unclear whether the observed alterations in shear stress during the OSS intervention are reflective of the wall shear stress. The complexity of flow patterns that manifest at the arterial wall during this intervention merits future characterization [i.e., using multigate Doppler (1)]. Finally, while these findings demonstrate hypoxia-associated reductions in FMD, the study was conducted with young, healthy men. Therefore, the results cannot be generalized beyond the tested population. Future work is required to investigate the influence of acute and sustained hypoxia exposure on women, older and/or clinical populations, and high-altitude natives.

Conclusions

A 30-min exposure to augmented OSS caused brachial artery endothelial dysfunction in normoxia but not during acute or sustained hypoxia. The decline in FMD in acute and sustained hypoxia, which may in part be due to altered baseline shear stress magnitude and pattern, may supersede any further impairment by cuff-induced OSS. Uncovering the mechanisms responsible for the altered shear stress in acute and sustained hypoxia may provide insight on the indirect actions of hypoxia on endothelial function.

GRANTS

This work was supported by the Natural Sciences and Engineering Research Council of Canada (to P. N. Ainslie) and the Canadian Foundation for Innovation and a Canada Research Chair (to P. N. Ainslie). J. C. Tremblay was supported by an Alexander Graham Bell Doctoral Canada Graduate Scholarships (Natural Sciences and Engineering Research Council of Canada) and a Doctoral Field Research Grant (Queen’s University).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.C.T. and K.E.P. conceived and designed research; J.C.T. and C.A.H. performed experiments; J.C.T. analyzed data; J.C.T., P.N.A., and K.E.P. interpreted results of experiments; J.C.T. prepared figures; J.C.T. drafted manuscript; J.C.T., C.A.H., P.N.A., and K.E.P. edited and revised manuscript; J.C.T., C.A.H., P.N.A., and K.E.P. approved final version of manuscript.