The middle cerebral artery blood velocity response to acute normobaric hypoxia occurs independently of changes in ventilation in humans

What is the central question of this study? Does the ventilatory response to moderate acute hypoxia increase cerebral perfusion independently of changes in arterial oxygen tension in humans? What is the main finding and its importance? The ventilatory response does not increase middle cerebral artery mean blood velocity during moderate isocapnic acute hypoxia beyond that elicited by reduced oxygen saturation.


New Findings
• What is the central question of this study?
Does the ventilatory response to moderate acute hypoxia increase cerebral perfusion independently of changes in arterial oxygen tension in humans?
• What is the main finding and its importance?
The ventilatory response does not increase middle cerebral artery mean blood velocity during moderate isocapnic acute hypoxia beyond that elicited by reduced oxygen saturation.

Participants
A total of 11 male volunteers (23 ± 3 years, 173 ± 7 cm, 73 ± 9 kg: mean ± SD) completed this study. Prior to inclusion they completed a screening (Medical Health Questionnaire) to ensure that they were healthy and not taking any prescription or over-the-counter medications. Instructions were provided to participants to refrain from alcohol, caffeine and heavy exercise for >24 h before the study.

Experimental protocol
At an initial familiarisation session participants breathed through a mouthpiece for a 10-min period to establish normal breathing values.
They were then familiarised to the Normoxia, Hypoxia, Hyperpnoea and Hypoxia + RB conditions (described below) for 15 min each.
At the start of the experimental session, participants again breathed through the mouthpiece for 10 min in order to establish normal values for P ETCO 2 and partial pressure of end-tidal oxygen (P ETO 2 ).
The four experimental conditions were then undertaken with P ETCO 2 and P ETO 2 controlled using a dynamic end-tidal forcing system to manipulate inspired (humidified) gases on a breath-by-breath basis (Robbins et al., 1982;Prodel et al., 2016). In all conditions, P ETCO 2 was held at baseline +1 mmHg. A 10-min recovery period was undertaken between conditions.
Conditions 1 and 2 were always performed before conditions 3 and 4, since the target R f and V T in conditions 3 and 4 were based on values obtained during conditions 1 and 2. However, a coin toss was used to randomize the order of conditions 1 and 2 (Normoxia and Hypoxia) and conditions 3 and 4 (Hyperpnoea and Hypoxia + RB). R f was guided using a metronome, while V T was guided using an oscilloscope.
During conditions 1 and 2, participants were able to view a screen showing an innocuous documentary to shift their attention away from breathing.

Experimental measures
All participants rested on a chair in an upright position throughout the study. Heart rate (HR) was monitored by electrocardiography (lead II, ECG) and beat-to-beat arterial blood pressure assessed using finger was used to measure the right MCA V mean through the temporal window (Willie et al. 2011). The probe was fixed using a modifiable head kit that locked the angle of insonation at the optimum position to ensure signal stability, and once the signal was acquired no further adjustments were made during the protocol.

Data analysis
All measurements recorded were converted from analog to digital data at 1 kHz (PowerLab, 16/30; ADInstruments, Dunedin, New Zealand) and stored for offline analysis (LabChart Pro; ADInstruments).
Respiratory data were extracted on a breath-by-breath basis, while cardiovascular and cerebrovascular data were extracted on a beat-bybeat basis. The last 5 min of each condition was used for analysis and averaged for data representation.
Mean arterial blood pressure (MAP) was calculated as: Cardiac output (CO) was calculated as SV × HR and total peripheral resistance (TPR) calculated as MAP∕CO. Cerebrovascular conductance index (CVCi) was calculated as MCA V mean /MAP (Flück et al., 2017).

Statistical analysis
A two-way repeated measures analysis of variance (ANOVA) was used to examine the main effects of oxygenation (hypoxia, normoxia), breathing (spontaneous, controlled) and their interaction (oxygenation × breathing). Significant interactions were explored post hoc using the Student-Newman-Keuls test. Statistical analysis was performed using SigmaPlot (version 14.0, Systat Software Inc., San Jose, CA, USA). Data are displayed as mean ± SD, unless otherwise indicated. Differences were considered significant if P < 0.05.

DISCUSSION
We sought to determine the contribution of the increased ventilation to cerebrovascular responses to acute normobaric isocapnic hypoxia in humans. Our major novel finding is that MCA V mean was increased to the same extent by Hypoxia and Hypoxia + RB, and not increased Cerebral blood flow increases when P ETO 2 falls below ∼58 mmHg and/or oxygen saturation (S pO 2 ) falls below ∼90% (Gupta et al., 1997).
This response may be governed by local vasodilatory mechanisms F I G U R E 3 MAP, TPR, CO and HR during the Normoxia, Hypoxia, Hyperpnoea and Hypoxia + RB conditions. *P < 0.05. Horizontal bars show mean and SD that enhanced ventilation independently increases MCA V mean during acute isocapnic hypoxia. Contrary to this hypothesis, MCA V mean did not increase during Hyperpnoea and the increases in MCA V mean during Hypoxia and Hypoxia + RB were similar. Interestingly, when participants were asked to wilfully restrain increases in ventilation during poikilocapnic hypoxia, internal carotid artery and MCA V mean increased to a greater extent than during poikilocapnic hypoxia with unrestricted breathing (Ogoh et al., 2014). However, P ETCO 2 was higher when ventilation was wilfully restricted, which could have explained the increased cerebral blood flow; therefore, to discern the effects of arterial CO 2 and ventilation, we matched ventilation during hypoxia to normoxia while controlling P ETCO 2 .
We observed that Hypoxia and Hypoxia + RB evoked similar increases in MCA V mean . Previously, Bilo et al. (2012) demonstrated that 15 min of breathing at a reduced rate of 6 breaths/min while at high altitude (4559-5400 m) enhanced S pO 2 (+6-9%) and reduced pulmonary artery pressure (−4 mmHg). Such changes in systemic oxygenation and central haemodynamics might be expected to modify cerebral blood flow, although this was not measured by Bilo et al. (2012). In contrast to Bilo et al. (2012) and a contribution to the elevated MCA V mean in these conditions is possible.
Dyspnoea arises from the activation of multiple sensory afferent mechanisms (e.g. pulmonary vagal afferents, respiratory muscle mechanoreceptors) and its perception involves several cortical sites (von Leupoldt & Dahme, 2005). These mechanisms could be another pathway by which the ventilatory response could increase cerebral blood flow during hypoxia. Indeed, respiratory discomfort induced with loaded breathing increases blood flow in the right anterior insula, cerebellar vermis and medial pons (Peiffer et al., 2001), while active inspiration and expiration increases cortical blood flow in discrete regions when compared with a passive condition (i.e., mechanical ventilation) (Ramsay et al., 1993). Functional magnetic resonance imaging has been used to identify the central pattern of activation induced by hypoxia (Critchley et al., 2015), but the extent to which this is explicitly related to changes in ventilation has not been explored.
Herein, we observed increases in MCA V mean during Hypoxia, but not during either Hypoxia + RB or Hyperpnoea, suggesting that changes in breathing rate and depth do not contribute. However, it is possible that neither the hypoxic stimulus nor the ventilatory response in our study evoked a sufficiently large activation of higher brain centres perfused by the middle cerebral artery.

Methodological considerations
There are several methodological issues that should be considered.
MCA V mean was used to assess cerebral perfusion, and the relative strengths and weaknesses of this technique have been discussed (Ainslie & Hoiland, 2014;Willie et al., 2011). In brief, in the absence of MCA diameter measurements, we can only assume that the MCA V mean data are representative of MCA flow. Given that acute hypoxia can cause cerebral vasodilatation (Wilson et al., 2011), it is possible that the magnitude of the increases in MCA V mean underestimate the true increases in cerebral blood flow under these conditions. However, as this would presumably have affected the Hypoxia and Hypoxia + RB conditions equally, it is not likely that it could affect the findings of this study. Further studies are warranted using advanced brain imaging technologies (e.g., magnetic resonance imaging, arterial spin labelling).
Finally, care should be taken when considering our findings gathered under conditions of acute normobaric hypoxia in the context of high altitude conditions of chronic hypobaric hypoxia (Coppel et al., 2015).
In conclusion, the results of this study indicate that during moderate isocapnic acute hypoxia the hypoxic ventilatory response does not increase cerebral perfusion beyond that elicited by reduced oxygen saturation.