Ten study participants were recruited as part of a 21-day research expedition to Everest base camp in the Nepal Himalaya. The expedition was considered a low-risk ascent profile (Zuckerman, 2012; see Figure 1 for ascent profile). All participants had no prior-history of cardiovascular, respiratory or cerebrovascular disease. Nine subjects were residents of Calgary (1045 m). One subject was an Irish national (0 m; sea-level) who had been living in Calgary for 5 months prior to the expedition. Several over-the-counter medications were taken during ascent, predominantly as a method of curtailing symptoms associated with exposure to HA (headache, diarrhea, cough, flu-like symptoms). The medications used were: Imodium, Ibuprofen, Neocitran, Advil, Pepto bismal and DM expectant syrup. Several of the female participants were taking oral contraception (Yasmin and Tricylen-lo) during the expedition. Diamox (acetazolamide) was not taken by any participant at any time during the ascent.

This study abided by the Canadian Government Tri-Council policy on research ethics with human participants (TCPS2) and conformed with the standards set by the latest revision of the Declaration of Helsinki, except for registration in a database. Ethical approval was received in advance through the Mount Royal University Human Research Ethics Board (Protocols 100012 and 100922) and was harmonized with the Nepal Health Research Council (Protocol 109-2017). All participants were recruited via verbal communication and all provided written and informed consent prior to voluntary participation in the study. Although this study took place in the context of a large research expedition to altitude, the specific study design, research question and data collection were planned a priori.

Posterior cerebral artery blood velocity (PCAv) and NVC response were assessed and later compared between Calgary (1045 m), Namche (3440 m), Deboche (3820 m), and Pheriche (4240 m). Similarly, arterial blood gasses (ABGs; PaO₂, SaO₂, PaCO_2, arterial bicarbonate [HCO₃^-], base excess and arterial pH) were collected at the same pre-determined locations as PCAv and NVC. Most baseline ABGs were collected at 1045 m. However, two participants had their baseline ABGs taken at Kathmandu (1400 m) due to logistical issues. Therefore, when describing ABGs within the results section we have referred to 1045 m/1400 m as baseline.

A total of 3 days were spent in Kathmandu (1400 m) prior to beginning the expedition. In addition to NVC, PCAv, and ABGs, several daily measures were obtained each morning, beginning in Kathmandu and continued for the duration of the expedition. The measures collected each morning were: HR, arterial blood pressure (Systolic; SBP, Diastolic; DBP), pressure of end-tidal CO₂ [(P_ET)CO₂], peripheral oxyhemoglobin saturation (SpO₂), Hb and hematocrit (hct) (see Table 1).

During ascent, participants spent a total of 2 days at each rest day altitude (3440, 3820, and 4240 m). The 1st day consisted of arrival having trekked from the preceding lower altitude, with recovery from the trek and one night’s sleep at the new altitude. The 2nd day consisted of data collection and an overnight stay before further ascent beginning the following morning. PCAv, NVC and ABGs were measured on the 2nd day at each altitude. This was implemented to reduce the possible effects of recent exercise on each variable. Participants were also tested sequentially in an identical fashion for PCAv and NVC at each location to account for the possible effects of circadian rhythm on cerebral hemodynamics.

Analysis revealed no significant effect of altitude on weight, BMI and HR (P > 0.05, see Table 1). Further analysis revealed a significant effect of altitude on SBP (F_1.891,17.017 = 3.232, P = 0.067 N_P² = 0.264), where SBP was significantly higher at 4240 m, compared with 1400 m (119 ± 11 vs. 113 ± 11 mmHg, mean ± SD, respectively; P < 0.001; see Table 1). A significant effect of altitude was observed for DBP (F_3,27 = 19.39, P < 0.001 N_P² = 0.683), with DBP being significantly elevated at 3440, 3820, and 4240 m compared with 1400 m (87 ± 8, 90 ± 10 and 87 ± 9 vs. 78 ± 11 mmHg, respectively; P < 0.001, P = 0.002, and P < 0.001, respectively; see Table 1). Similarly, a significant effect of altitude was observed in MAP (F_3,27 = 13.613, P < 0.001 N_P² = 0.602). MAP was elevated at 3440, 3820, and 4240 m compared with 1400 m (96 ± 7, 100 ± 10, and 98 ± 9 vs. 89 ± 11 mmHg, respectively; P = 0.026, P = 0.006, and P < 0.001, respectively; see Table 1). A significant effect of altitude was observed in (P_ET)CO₂ (F_1.246,11.214 = 25.012, P < 0.001 N_P² = 0.735). (P_ET)CO₂ was significantly lower at 3440, 3820, and 4240 m compared with 1400 m (26.5 ± 3.0, 25.3 ± 2.7, and 21.1 ± 1.9 vs. 32.6 ± 6.8 Torr, respectively; P = 0.043, P = 0.018, and P < 0.001, respectively; see Table 1). Furthermore, (P_ET)CO₂ was significantly lower at 4240 m compared with both 3440 and 3820 m (P < 0.001 and P < 0.001, respectively). Finally, a significant effect of altitude was observed for SpO₂ [X²(3) = 25.906, P < 0.001]. SpO₂ was significantly lower at 3440, 3820, and 4240 m compared with 1400 m (91.80 ± 2.49, 91.40 ± 2.84, and 87.60 ± 2.01 vs. 96.40 ± 0.97%, respectively; P = 0.005, P = 0.005, and P = 0.005, respectively; see Table 1). Furthermore, SpO₂ was significantly lower at 4240 m compared with 3440 m (P = 0.008).

Analysis revealed a statistically significant effect of altitude on PaO₂ (F_3,18 = 138.270, P < 0.001 N_p² = 0.958). PaO₂ was significantly lower at 3440, 3820, and 4240 m compared with 1045/1400 m (48.5 ± 8.1, 54.6 ± 6.2, 48.4 ± 5.0 vs. 84.8 ± 5.4 mmHg, respectively; P < 0.001, P < 0.001, and P < 0.001, respectively; see Figure 3A). A statistically significant effect of altitude on SaO₂ (F_3,18 = 17.313, P < 0.001 N_p² = .743). SaO₂ was significantly lower at 3440, 3820, and 4240 m compared with 1045/1400 m (84.9 ± 7.9, 88.8 ± 3.7, 85.5 ± 3.7 vs. 96.8 ± 0.8%, respectively; P = 0.032, P = 0.003, and P < 0.001, respectively; see Figure 3B). A statistically significant effect of altitude on PaCO₂ (F_3,18 = 23.123, P < 0.001 N_p² = 0.794). PaCO₂ was significantly lower at 3440, 3820, and 4240 m compared with 1045/1400 m (30.2 ± 4.9, 29.6 ± 3.8, 29.1 ± 3.7 vs. 34.8 ± 4.6 mmHg, respectively; P = 0.015, P = 0.016, and P < 0.001, respectively; see Figure 3C). A statistically significant effect of altitude on [HCO₃^-] (F_3,18= 11.342, P < 0.001, N_p² = 0.654). [HCO₃^-] was significantly lower at 3820 and 4240 m compared with 1045/1400 m (19.5 ± 2.0, 19.7 ± 2.8 vs. 23.2 ± 2.7 mmol/L, respectively; P = 0.043 and P = 0.021, respectively; see Figure 3D). No significant differences were found between 1045/1400 m and 3440 m for [HCO₃^-] (23.2 ± 2.7 vs. 21.0 ± 3.1 mmol/L, respectively; P = 0.417), but a significant difference for [HCO₃^-] was found between 3440 and 4240 m (P = 0.006). A statistically significant effect of altitude on base excess (F_3,18 = 7.132, P = 0.002, N_p² = 0.543). Base excess was significantly lower at 4240 m compared with 3440 m (-4.143 ± 1.18 vs. -2.286 ± 1.19 mmol/L, respectively; P = 0.022; see Figure 3E). Analysis observed no statistically significant effect of altitude on arterial pH when compared with baseline (F_3,18 = 2.788, P = 0.070 N_p² = 0.317; see Figure 3F).

There were no significant differences for baseline PCAv or baseline PCAv tAUC between locations [X²(3) = 5.640, P = 0.131 and X²(3) = 6.60, P = 0.086, respectively; see Figures 4A,B, respectively].

We first sought to assert that the NVC response was evoked at each location. We therefore compared the absolute mean and peak PCAv achieved during each individual VS to baseline PCAv. We replicated this analysis for PCAv tAUC, thus covering all three NVC parameters.

This study investigated the influence of incremental ascent to 4240 m over 7 days on NVC. Our results demonstrate: (1) the NVC response was evoked consistently at each altitude, and (2) there was no significant difference in the magnitude of the NVC response at the level of the PCA between altitudes in healthy acclimatized individuals.

Exposure to HA and the associated drop in total barometric pressure induced a state of hypoxemia. This is evident from the significant decreases in PaO₂ and SaO₂ (see Figures 3A,B). During prolonged exposure to HA, hypoxia increases activation of the peripheral chemoreceptors, enhancing the HVR (Ainslie and Ogoh, 2010; Willie et al., 2015; Hoiland et al., 2016). The HVR increases ventilatory drive partially restoring PaO₂ and SaO₂ levels toward resting values (Bernardi et al., 2006; Smith et al., 2014). Comparing values recorded at 3440m with those obtained at higher altitudes, with further decreases in barometric pressure, we observed no further reductions in PaO₂ and SaO₂, revealing attenuation of a further decline in oxygenation associated with continued ascent.

Secondly, we observed profound hypocapnia, as demonstrated by significant decreases in PaCO₂, indicative of hyperventilation (see Figure 3C). Hypocapnia serves to blunt both central and peripheral chemoreceptor activation and subsequently reduces ventilatory drive. Combined, these data reveal the presence of two antagonistic vasoactive stressors: hypoxia and hypocapnia. Hypoxia, below a threshold (<40–45 mmHg), has a vasodilatory effect in the systemic circuit (Ainslie and Ogoh, 2010). In contrast, hypocapnia is a potent vasoconstrictor (Brugniaux et al., 2007). The cerebrovasculature is particularly sensitive to changes in ABGs, especially PaCO₂ (Brugniaux et al., 2007; Ainslie and Ogoh, 2010). Although prior research has shown that the presence of hypocapnia decreases the NVC response (Szabo et al., 2011), this study was laboratory-based and did not include the hypoxic vasodilatory stress associated with HA. Furthermore, the presence of sustained hypocapnia presents further issues by inducing tissue alkalosis (Krapf et al., 1991).

Hyperventilation-induced hypocapnia triggers an acid-base adjustment in the form of renal compensation. To counteract respiratory alkalosis, and maintain pH within normal limits, a series of compensatory steps are required through increased acid retention coupled with an increase in HCO₃^- excretion via the renal system (Zouboules et al., 2018). The end-product of a relative metabolic acidosis helps to decrease arterial pH toward normal values, thereby preserving acid-base homeostasis (Goldfarb-Rumyantzev and Alper, 2014). The sophisticated and refined nature of this compensatory adaptation is crucial as the vasoconstrictive effects of hypocapnia/alkalosis can decrease CBF (Szabo et al., 2011). Furthermore, a blunted HVR response has been shown to be a strong predictor of AMS (Richalet et al., 2012). We observed evidence of renal compensation by way of significant decreases in arterial [HCO₃^-] (see Figure 3D) in response to hypocapnia, such that arterial pH was maintained constant throughout ascent compared with BL.

We found no significant differences in baseline PCAv and baseline PCAv tAUC during ascent (Figure 4). In terms of NVC, we observed that for each of our designated NVC parameters, a significant response was evoked during VS (see Figures 5–7). Our results demonstrate the remarkable reproducibility of the within-group NVC response. For between-altitude comparisons, our findings demonstrate that NVC remained intact at the level of the PCA, as demonstrated by no significant differences in each of the three NVC response parameters between altitudes (Figure 8), although we observed both inter-subject variability of the NVC response and intra-subject variability in the NVC response between altitudes (Figure 2). These findings are consistent with a previous study (Caldwell et al., 2017), where NVC testing was performed temporally in response to acclimatization at a single altitude. These studies differed in terms of ascent-profile, altitude reached, analysis and stimulus to evoke NVC. However, the finding that NVC remains intact during ascent to HA is consistent.

The utilization of TCD for the assessment of CBF has limitations. Although TCD provides excellent beat-by-beat measurement of the intracranial arteries arising from the circle of Willis, it does not provide any indication of pre-existing diameter of the insonated vessel. This is relevant at HA where there are multiple and competing stressors, which influence vessel tone (hypoxic vasodilation and hypocapnic vasoconstriction). There is also recent literature to suggest that there are stimulus-evoked changes in diameter of the PCA as a function of distance from the primary visual cortex (Bizeau et al., 2017). There is therefore an inherent risk of misinterpreting the magnitude of the CBF change (Caldwell et al., 2017). Furthermore, there is a risk of technical error given the sensitivity of transcranial Doppler technique. To account for this, we ensured the same sonographers were present during each testing session. We also strictly abided by a set of guidelines for insonating the PCA (see Willie et al., 2011). We also acknowledge that TCD provides measurements of velocity rather than flow. However, recent literature has demonstrated CBF velocity to be an appropriate and reliable surrogate for CBF (Secher et al., 2008; Ainslie and Ogoh, 2010). Although our findings demonstrate an intact NVC response during incremental ascent to HA, further research is required to fully explore this issue. The cerebrovasculature can be subdivided into three categories: large pial arteries which run along the surface of the cortex (i.e., PCA), pial arterioles which penetrate the cortex and finally parenchymal arterioles and capillaries (Filosa et al., 2016; Phillips et al., 2016). It has been postulated that NVC-mediated responses are initiated within the parenchymal arterioles and capillaries (Filosa et al., 2016). Whether the NVC response within the microvasculature is affected during incremental ascent to HA remains unknown. Although we did not directly measure the NVC response at the microvascular level, the application of TCD to conduit arteries for assessment of NVC in human volunteers is a well-established technique (Willie et al., 2011; Phillips et al., 2016). If the diameter of the PCA was increased during VS, this would have evoked a blunted hyperemic response, that is the increase in PCA diameter would have caused a reduction in PCAv (Phillips et al., 2016). As this was not observed in our study, we assume a constant PCA diameter during VS, with the increase in PCAv most likely due to downstream vasodilation at the level of the neurovascular unit.