The purpose of this study was twofold:1) to determine whether at high altitude cerebral blood flow (CBF) as assessed during CO2 inhalation and during hyperventilation in subjects with acute mountain sickness (AMS) was different from that in subjects without AMS and2) to compare the CBF as assessed under similar conditions in Sherpas at high altitude and in subjects at sea level. Resting control values of blood flow velocity in the middle cerebral artery (V MCA), pulse oxygen saturation ( ), and transcutaneous were measured at 4,243 m in 43 subjects without AMS, 17 subjects with AMS, 20 Sherpas, and 13 subjects at sea level. Responses of CO2 inhalation and hyperventilation onV MCA, , and transcutaneous were measured, and the cerebral vasomotor reactivity (VMR = ΔV MCA/ ) was calculated as the fractional change ofV MCA per Torr change of , yielding a hypercapnic VMR and a hypocapnic VMR. AMS subjects showed a significantly higher resting controlV MCA than did no-AMS subjects (74 ± 22 and 56 ± 14 cm/s, respectively;P < 0.001), and was significantly lower (80 ± 8 and 88 ± 3%, respectively; P < 0.001). Resting controlV MCA values in the sea-level group (60 ± 15 cm/s), in the no-AMS group, and in Sherpas (59 ± 13 cm/s) were not different. Hypercapnic VMR values in AMS subjects were 4.0 ± 4.4, in no-AMS subjects were 5.5 ± 4.3, in Sherpas were 5.6 ± 4.1, and in sea-level subjects were 5.6 ± 2.5 (not significant). Hypocapnic VMR values were significantly higher in AMS subjects (5.9 ± 1.5) compared with no-AMS subjects (4.8 ± 1.4; P < 0.005) but were not significantly different between Sherpas (3.8 ± 1.1) and the sea-level group (2.8 ± 0.7). We conclude that AMS subjects have greater cerebral hemodynamic responses to hyperventilation, higherV MCAresting control values, and lower compared with no-AMS subjects. Sherpas showed a cerebral hemodynamic pattern similar to that of normal subjects at sea level.
- mountain sickness
- transcranial Doppler
- transcutaneous carbon dioxide
acute mountain sickness (AMS) is usually a benign and self-limiting illness that strikes healthy individuals who ascend to altitudes above 2,500 m without sufficient acclimatization (10). Failure to acclimatize may lead to AMS, which may develop into high-altitude pulmonary edema (HAPE) and high-altitude cerebral edema (HACE), both being lethal complications of AMS. The process of acclimatization involves the onset of progressive hyperventilation secondary to the hypoxia from reduced ambient air pressure (23). It results in hypocapnia, which causes cerebral vasoconstriction and a decrease of cerebral blood flow (CBF). This effect of hypocapnia on CBF is counterbalanced by the effect of hypoxia, resulting ultimately in a net increase of CBF at high altitude (4, 16, 27, 28). In an attempt to treat subjects with AMS, the addition of 3% CO2 to the inspired air showed in one study an increase of CBF in half of the subjects (11), whereas in another study an increase of CBF velocity (V CBF), measured with transcranial Doppler, could not be demonstrated (3). Otis and co-workers (22) showed that in subjects inhaling 5% CO2, first at sea level and later at high altitude where some of the subjects developed AMS,V CBF increased more at sea level than at high altitude. In contrast, it has been demonstrated that in healthy subjects the increase ofV CBF during hypercapnic hyperoxia at altitude was greater than the increase of V CBFat sea level (15).
The data from these previous studies might suggest that CO2 inhalation produced a smaller increase of CBF in subjects with AMS than in healthy subjects. However, responses to CO2 variations on CBF have not been studied in subjects with and without AMS at altitude. The hypothesis of this study was that the cerebral hemodynamic response to CO2 inhalation and to hyperventilation, expressed as vasomotor reactivity (VMR = fractional change of V CBFper Torr change of ), was different in subjects with AMS compared with subjects without AMS. Accordingly, the purpose of the present study was to determine VMR in an attempt to identify a mechanism for altered CBF at altitude. Therefore, we assessed at 4,243-m altitude in newcomers with and without AMS the response of the blood flow velocity of the middle cerebral artery (MCA) (V MCA), measured with transcranial Doppler (TCD) sonography, to CO2 inhalation and to voluntary hyperventilation. In addition, because there are inadequate data on CBF responses to changes of CO2 in high-altitude residents, theV MCA was assessed under similar conditions of high and low CO2 in a group of healthy Sherpas and was compared with a control group at sea level.
Subjects and AMS score.
The study was approved by the Institutional Human Investigation Committee of the Academic Medical Centre. It was carried out at the health post run by volunteer doctors of the Himalayan Rescue Association in Pheriche, a village situated at 4,243 m in the Khumbu valley in Nepal. The investigation took place during the months of October and November. Trekkers from many parts of the world who come to climb Kala Patar (5,545 m), a mountain adjacent to Mount Everest, normally stay for 1 or 2 nights in Pheriche. The trekkers who during their stay in Pheriche volunteered to participate in the study were previously healthy, smoking and nonsmoking, male and female subjects of varying background. They all resided normally at <600 m above sea level. Before they were studied, they stayed above 2,800 m (Lukla) at least for 4 nights, of which 3 nights were above 3,440 m (Namche Bazar). All volunteers were accepted for participation in the study, whether they came to Pheriche over land from Kathmandu, or after flying to Lukla, or by reaching Pheriche after crossing the Chola pass (5,420 m). Some of the trekkers with serious signs and symptoms of AMS were referred by the doctors of the health post.
AMS was assessed by using the Lake Louise AMS scoring system, which is a standard scoring system involving a self-reporting questionnaire to assess AMS (29). This self-report questionnaire includes five main symptoms of AMS: headache, gastrointestinal symptoms (anorexia, nausea, or vomiting), fatigue and/or weakness, dizziness and/or lightheadedness, and difficulty in sleeping. Each symptom is rated on a scale ranging from zero to three points. The AMS self-report score is the sum of responses to these five symptoms. In this study, a total score of three or more points in the presence of at least a mild headache was considered as AMS (AMS group) (29). From zero to two points was considered as no AMS (no-AMS group). Subjects with three or more points but no headache were considered as not having AMS. However, three subjects who had three or more points and no headache, but who showed signs of HAPE and HACE, were included in the AMS group. The questionnaire was completed by each trekker before the measurements were performed. Data of 43 trekkers without AMS (mean age 35 ± 12 yr; 37 men and 6 women) and of 17 trekkers with AMS (mean age 38 ± 12 yr; 13 men and 4 women) were used for the study. A group of 20 healthy Sherpas (mean age 37 ± 12 yr; 12 men and 8 women), who are Himalayan natives of Tibetan ancestry, were born in the Khumbu area and were living in and around Pheriche, were also studied, as well as a group of 13 healthy Dutch hospital workers (mean age 31 ± 6 yr; 9 men and 4 women) who were studied at sea level and who served as control group. In total, data of 93 subjects, divided into four different groups (no AMS, AMS, Sherpa, sea level), were analyzed.
Pulse oxygen saturation ( ) and transcutaneous were measured by using the Fastrac Respiratory Status Monitor (model 765500-101, Sensormedics). The transcutaneous sensor was heated to a temperature of 42°C and then was attached to the skin below the clavicle with a double-sided adhesive ring. The barometric pressure of the Fastrac monitor was set at 447 Torr, corresponding to the altitude of 4,243 m in Pheriche.
V MCA was measured by using a 2-MHz pulsed TCD probe (T2-64B EME, Ueberlingen). TCD insonation of the stem of the MCA was performed according to the methods described by Aaslid (1). The meanV MCA is calculated and displayed automatically by this TCD instrument. It is computed as the time mean of the peak velocity envelope, being the trace of the peak flow velocities as a function of time, and is expressed in centimeters per second. Doppler signals from the left MCA were identified through the temporal window with a hand-held probe and measured at a depth of 45–55 mm, corresponding to the proximal segment of the MCA. In each subject, a constant depth-range and angle of insonation were kept throughout the study. When the Doppler signal was inadequate at the left side, the right MCA was insonated.V MCA values, as calculated by the TCD apparature over four to five heart cycles, were registered, and, for every individualV MCA measurement, four consecutive registrations were averaged.
Subjects rested in the supine position for ∼15 min, after which , trancutaneous , andV MCA were measured at three different levels of .1) First, the values were measured with the subjects at rest while breathing ambient air (resting control value). 2) Then, the values were measured during hypercapnia. Hypercapnia was produced by CO2 inhalation. CO2 (0.6–1 l/min) was added to the ambient air and inhaled through a nonreturning valve connected to a mouthpiece. In this study, the inspiratory CO2 concentration was not controlled exactly. However, with a mean minute ventilation of 12 l/min as has been measured in subjects after 3–5 days at 4,300 m, the inspiratory CO2 concentration is calculated to be at least 5% (12). Measurements were made when transcutaneous and remained stable for ≥5 min.3) Finally, the values were measured during hypocapnia. Hypocapnia was induced by forced voluntary hyperventilation with a fixed respiratory frequency of ±25/min, and the same set of measurements was performed when transcutaneous and remained stable during ≥5 min. VMR was calculated for the hypercapnic state as well as for the hypocapnic state, yielding a hypercapnic VMR and a hypocapnic VMR.
Comparisons between resting control values and between effects of CO2 challenges were assessed by using ANOVA, and, when significance was found, a post hoc test (Bonferroni-Dunn) was performed to delineate where differences lay. The relationships between variables were subjected to linear correlation analysis by using Pearson’s correlation coefficient. Comparisons and correlations were considered significant whenP < 0.05. Data are expressed as means ± SD.
Resting control measurements.
The age did not differ among the four groups. Resting control values ofV MCA, , and transcutaneous of the four groups are presented in Table 1. Regression analysis showed a weak negative correlation between the resting control values ofV MCA and in the no-AMS group (r = −0.43,P = 0.0038) as well as in the AMS group (r = −0.58,P = 0.0138). In the Sherpa group, no correlation could be demonstrated (r = 0.07, P = 0.77). No correlations were found between the resting control values ofV MCA and transcutaneous in any of the four groups.
VMR during CO2 inhalation.
Responses of CO2 inhalation onV MCA, , and transcutaneous are shown in Fig.1. The absolute increases as well as the fractional increases ofV MCA were not significantly different among any of the four groups. The increased ventilation from the CO2stimulation produced an increase in in all three groups at altitude, and this increase was significantly greater in the AMS group compared with the no-AMS group (10 ± 6 and 5 ± 3% respectively; P < 0.0001). Increases of were not significantly different among any of the four groups. Hypercapnic VMR values, as shown in Table 1, were not significantly different between any of the four groups.
VMR during hyperventilation.
Responses of hyperventilation onV MCA, , and trancutaneous are shown in Fig. 1. The absolute decrease ofV MCA in the AMS group was significantly greater compared with the other three groups. However, the relative decrease ofV MCA was not significantly different between the AMS and the no-AMS group. The increase of during hyperventilation was significantly higher in the AMS group compared with the no-AMS group (13 ± 9 and 6 ± 6%, respectively;P < 0.0001). During active hyperventilation, the decreases of were significantly different between the sea-level group (12 ± 2 Torr) on the one hand and the no-AMS (9 ± 2 Torr) and the AMS groups (9 ± 3 Torr) on the other hand. Hypocapnic VMR values of the four groups are shown in Table 1. VMR values in the AMS group were significantly higher compared with the no-AMS group (P < 0.005). There was a significant negative correlation between the hypocapnic VMR and the decrease of in the no-AMS group as well as in the AMS group (Fig. 2). This correlation was weak in the Sherpa group and could not be demonstrated in the sea-level group.
Responses in Sherpas.
Responses to CO2 challenges in Sherpas are shown in Table 1 and Fig. 1. During CO2 inhalation, the Sherpa group showed no significant differences with any of the three groups concerning the absolute and relative increase ofV MCA, the increase of , and the VMR. However, the Sherpas had a significantly smaller increase in compared with the AMS group (6 ± 3 and 10 ± 6%, respectively;P < 0.05). During hyperventilation, the absolute and relative decreases ofV MCA in Sherpas were significantly lower compared with the AMS group but were not significantly different between the Sherpa group and the sea-level group. The increase of in the Sherpa group showed no significant difference with the no-AMS group and the AMS group. There was also no significant difference among any of the three groups concerning the decrease of . The hypocapnic VMR values in the Sherpa group were not significantly different compared with the sea-level group but were significantly lower compared with the no-AMS (P < 0.005) and the AMS group (P < 0.0001).
In this cross-sectional study, subjects at altitude with AMS showed an increased cerebral hemodynamic response to hyperventilation compared with healthy subjects, but the response to CO2 inhalation was not different. The study demonstrated that Sherpas had resting controlV CBF and cerebral VMR values to CO2 inhalation and to hyperventilation that were not different from those of native sea-level dwellers. The study also confirmed earlier findings that subjects with AMS had significantly higher resting control V CBF and lower values compared with those of healthy subjects (4).
TCD was used as a noninvasive device to measure CBF. Although TCD does not provide a direct measure of CBF, relative changes in V MCAaccurately reflect relative changes in CBF, provided that the diameter of the insonated vessel remains constant (5, 6, 19). There is now ample direct and indirect evidence to support the contention that the diameter of the MCA does not change significantly with changes in (6, 14). Jensen et al. (15, 16) showed in two studies during isocapnic hypoxia with levels to 60–70%, a good agreement between TCD data and CBF measurements. This may suggest that the diameter of the stem of the MCA did not dilate significantly at the moderate degrees of hypoxia as were measured in our study.
Normally, end-expiratory and transcutaneous give both good correlations with arterial . End-expiratory is normally 3–5 Torr below arterial in patients without lung disease. In one study, where the Fastrac monitor was used with the sensor heated to 42°C, mean transcutaneous was 1.7 Torr below arterial (17). In our study, we measured transcutaneous in subjects with AMS, who may have had some degree of pulmonary edema, possibly leading to a larger arterial-to-alveolar difference than in no-AMS subjects. In these AMS subjects, end-expiratory would underestimate arterial compared with no-AMS subjects.
Resting control measurements at rest.
V MCA was similar in the three groups who were healthy, despite different environmental (no AMS, sea level) and adaptive (Sherpa) backgrounds. The normal value of V MCA at sea level varies from 41 to 94 cm/s, with an average of 60 cm/s during the awake and resting state (2). This is in accordance with the meanV MCA of 60 cm/s of the sea-level group in our study.
Previous studies have shown an initial increase of CBF (16, 27, 28) and of V CBF (4, 13, 22) on arrival of the subjects at altitude. These increases returned to sea-level values after 3–5 days (13, 16, 27, 28). In the present study, all volunteers, with and without AMS, stayed at least 4 nights above 2,800 m, of which 3 nights were above 3,440 m. ThusV MCA in these volunteers could have been expected to have returned to within preascent values, because the measuredV MCA in the healthy subjects at altitude was similar to that of the sea-level group.
Several authors established a relationship between the increases in CBF (27) or V MCA (9,22) and the occurence of AMS. But in two other studies this relationship was not evident (16, 24). Baumgartner and co-workers (4), however, demonstrated a significant correlation between the increase ofV MCA and the presence of symptoms of AMS, probably from a lower arterial than in healthy subjects. The results of our study are in agreement with those of Baumgartner et al. We demonstrated a significantly higher resting controlV MCA and lower resting control in the AMS group compared with the no-AMS group. There was also a significant inverse relationship between the resting control values of andV MCA in the no-AMS group and in the AMS group. These findings suggest that the lower in the AMS group, compared with the no-AMS group, may be one of the mechanisms that could explain the higher V MCA.
VMR was calculated as the relative change ofV MCA per Torr change of (5, 19). Correspondingly, the sea-level group showed a relative increase ofV MCA of 5.6 ± 2.5% per Torr change at the hypercapnic level, and a relative decrease of theV MCA of 2.8 ± 0.7% per Torr change at the hypocapnic level. This corresponds to a CBF increase of 6% and a CBF decrease of 2% per Torr change, respectively, as found in normotensive healthy men (30), indicating that the changes in TCD values we measured can be taken to reflect changes in CBF (6).
In this study, no significant difference in the hypercapnic VMR values could be demonstrated among any of the four groups (Table 1). However, hypocapnic VMR values were significantly different between each of the four groups, except between the Sherpa group and the sea-level group. Normally, at sea level, VMR values are unchanged in the same subject when remains unchanged during CO2 inhalation and during hyperventilation, because the VMR is independent of the change of within the range of 25–50 Torr (18, 25). However, during CO2 inhalation and during hyperventilation at altitude, the increase of the ventilation produced substantial improvements of arterial oxygenation, which were under both circumstances larger in the AMS group (10 and 13% , respectively) than in the no-AMS group (5 and 6% , respectively). These effects of changes of as well as of onV MCA are opposite at the hypercapnic level and additive at the hypocapnic level. However, it was not the aim of this study to define the relative contributions of the CO2 and hypoxia effects onV MCA.
Measurement of VMR at altitude is an attempt to evaluate whether other mechanisms than the direct effect of changes contribute in the regulation of CBF at altitude. Besides the effect of hypoxia on theV CBF (15), there are two other factors that might explain the difference in VMR between the no-AMS and the AMS groups in our study. First, the gradual declines of and cerebrospinal fluid bicarbonate during the process of acclimatization at altitude result in larger pH changes per Torr change and consequently in largerV MCA changes per Torr change. Second, the respiratory alkalosis from the acute and active hyperventilation in our subjects might have produced an increased affinity of oxygen to hemoglobin (Bohr effect). This Bohr effect results in less oxygen availability and causes theV CBF to respond to hypoxia (31). However, with the resting control values and the changes during hyperventilation being similar in the no-AMS group and the AMS group, it is possible that these two factors do not play an important role in explaining the difference in VMR values between these two groups.
During hyperventilation, a significant negative correlation between VMR and the decrease of could be detected in the AMS and the no-AMS group (Fig. 2), but not in the sea-level group, since VMR is at sea level relatively independent of the change of within the range that we measured. The subjects with small decreases of produced large relative decreases ofV MCA. A possible explanation might be that the contribution of the increase onV MCA has a great effect on V MCAwhen the changes of are small. On the other hand, with large changes of , the same increase of contributes relatively little to the change ofV MCA.
Ringelstein and co-workers (25) measured in humans at sea level total vasomotor reactivity (TVMR), defined as the sum of the fractionalV MCA increase (52.5%) during maximal dilatation of the cerebral vessels from 5 to 6% CO2 inhalation and the fractional V MCAdecrease (35.5%) during maximal vasoconstriction from hyperventilation. By using this method to calculate TVMR in our groups (Fig. 3), we found in the sea-level group, the no-AMS group, and the AMS group a pattern of resetting of TVMR in which the hypercapnic section becomes smaller in favor of a larger hypocapnic section. Otis and co-workers (22) found in subjects at sea level during 5% CO2 inhalation a hypercapnic section of TVMR of 45%. This is similar to the 39% observed in our sea-level group. After ascent to 4,100-m altitude, the hypercapnic section decreased significantly to 23% in their group, compared with 24% in our AMS group. They attributed this blunted hypercapnic response to CO2inhalation at altitude to the exhausted potential of cerebral resistance vessels to further dilate. The explanation for the pattern of resetting of TVMR in our study may be the increased vasoconstriction of the brain resistance vessels caused by the improvements in these groups. An additional contributing factor to this pattern of resetting of TVMR could be the increased sensitivity of theV CBF to hypoxia, as reported by Jensen and co-workers (15) in newcomers after 5 days at altitude.
Resting controlV MCA of the Sherpa group was similar to that of the sea-level group. To our knowledge, no studies are available in which CBF has been measured in Himalayan highlanders. However, several studies performed in Andean high-altitude natives indicate that their CBF is significantly below normal sea-level values (20, 21, 29). A possible factor that might contribute to this discrepancy between the Andean high-altitude natives and the Sherpas is the hematocrit, which is an important marker of blood viscosity and which has a marked influence on CBF at sea level and at altitude (21). Winslow and co-workers (31) found that the hematocrit in Sherpas in Nepal was significantly lower than in high-altitude Andean natives (48.4 ± 4.5 and 52.2 ± 4.6%, respectively; P < 0.003). This difference in hematocrit might contribute to the higherV MCA in the Sherpas compared with the Andean natives.
Interestingly, the Sherpas exhibited similar VMR values as the sea-level control group during CO2inhalation as well as during hyperventilation. However, in contrast to the sea-level group, they showed increases of 6% during CO2 inhalation and of 9% during hyperventilation. Milledge and Sørensen (21) demonstrated that the cerebral vasodilatory response to hypoxia was preserved in Andean high-altitude residents, even after lifelong exposure to hypoxia. Despite this preserved vasodilatory response to hypoxia, the hypercapnic and hypocapnic VMR values in the Sherpa group and the sea-level group were similar. An explanation for the similar VMR values is that a higher blood viscosity in the Sherpas, due to a higher hematocrit, has a slowing-down effect on the changes ofV MCA during the alterations of arterial oxygen and CO2tensions.
The Sherpas, who are Himalayan natives of Tibetan ancestry, and who have lived in the Himalaya far longer than the high-altitude natives in the Andes, show a cerebral hemodynamic pattern that is normal by sea-level standards. Similar patterns of adaptation to lifelong hypoxia on the Tibetan plateau have recently been shown in other studies (7,8).
In the present study we showed that in subjects with AMS the cerebral hemodynamic response to acute hyperventilation is increased compared with subjects without AMS. We confirmed that in AMS subjects at rest the CBF velocity is higher and is lower. We also found that Sherpas, studied at high altitude, have the same V MCA and essentially the same VMR pattern as do native sea-level dwellers.
The authors acknowledge the subjects for their cooperation and the doctors of the Himalayan Rescue Association, Fiona McPherson and Caroline MacKenzie, for their enthusiasm and motivation of their patients to participate in the study. The authors thank Dr. Joseph Odoom for critically reviewing the manuscript and Prof. Can Ince for constructive criticism.
Address for reprint requests: G. F. A. Jansen, Dept. of Anesthesiology, H-1-Z, Academic Medical Centre, Univ. of Amsterdam, PO Box 22600, 1100 DE Amsterdam, The Netherlands.
- Copyright © 1999 the American Physiological Society