Abstract
Background:
The associations among cortisol levels, body water status, and acute mountain sickness (AMS) remain unclear. We investigated associations between AMS prevalence and severity with resting saliva cortisol levels at low altitude (LA) and high altitude (HA) and with fluid balance during a HA stay.
Methods:
Twenty-two physically fit and healthy participants (12 women, 10 men) were transported to HA (Testa Grigia, 3480 m). In the late afternoon at LA, on the next day 3–4 hours after arrival at HA and in the morning after an overnight stay, heart rate, oxygen saturation, and systolic and diastolic blood pressures were measured in a sitting position after 10 minutes of rest; cortisol levels were quantified in saliva samples taken pre-ascent and 3–4 hours after arrival at HA. AMS was scored with the 1993 Lake Louise Score (LLS, cut-off ≥3). Urine volume and fluid and food intake were recorded during the altitude stay.
Results:
Pre-ascent cortisol levels were associated with fluid retention during the altitude stay (r2 = 0.33, p < 0.05) and both were positively related to the LLS (r2 = 0.49 and r2 = 0.26, p < 0.05, respectively).
Conclusions:
In conclusion, resting LA cortisol levels and fluid retention upon rapid exposure to altitude seem to be associated with AMS. This suggests a potential link among cortisol homeostasis, fluid balance, and AMS risk.
Introduction
Acute mountain sickness (AMS) may develop in nonacclimatized persons rapidly ascending to altitudes above 2500 m. AMS is characterized by nonspecific symptoms such as headache, dizziness, nausea, vomiting, loss of appetite, fatigue, and insomnia (Hackett and Roach, 2001; Burtscher et al., 2011). Symptoms of AMS typically appear 6–12 hours after arrival at high altitude (HA) and usually subside over a period of one or more days (Basnyat and Murdoch, 2003).
Despite the progress in discovering mechanisms potentially linked to AMS development, there is still ambiguity about the exact etiology (Luks et al., 2017). Higher trait anxiety and higher levels of anxiety before a mountain ascent, for instance were reported in climbers susceptible to AMS (Missoum et al., 1992). In addition, higher trait anxiety at low altitude (LA) was found predictive for severe AMS at HA, (Boos et al., 2018) hence, suggesting an influence of the individual stress regulation on AMS development. Since trait anxiety relates to HPA-axis regulation (Taylor et al., 2008), it is possible that cortisol levels relate to the risk of AMS. Findings on the influence of cortisol responses on AMS development, however, are conflicting. Cortisol levels have been found to be unchanged (Smith et al., 2011; Woods et al., 2012) or increased when ascending to HA (Sutton et al., 1977; Humpeler et al., 1980; Richalet et al., 1989; Ermolao et al., 2009; Woods et al., 2012). In addition, associations between elevated cortisol levels and AMS have been described in some reports (Sutton et al., 1977; Richalet et al., 1989), but not in others (Woods et al., 2012). Paradoxically, synthetic corticosteroids such as dexamethasone and prednisolone, acting on vascular permeability, inflammation, the sympathetic nervous system, oxygen saturation, and redox state (Swenson, 2016), have successfully been used to prevent and treat AMS (Ferrazzini et al., 1987; Levine et al., 1989; Woods et al., 2012; Swenson, 2016). Thus, on the one hand, high cortisol levels may indicate high stress levels possibly negatively impacting on AMS development, and on the other hand, cortisol-like substances can prevent or treat AMS. Besides being linked to stress regulation, cortisol may act as a weak mineralocorticoid leading to fluid retention (Whitworth et al., 2005). Fluid retention was found to be linked to AMS severity (Loeppky et al., 2005), yet, the relationships among cortisol response, fluid retention, and AMS severity are not yet fully explored.
The present study aimed to test the hypothesis that saliva cortisol levels measured under resting conditions before ascending to HA and fluid homeostasis during HA exposure are related to AMS.
Materials and Methods
Participants and settings
Twenty-two physically fit and healthy participants gave their written informed consent for participation in this study. Baseline characteristics are shown in Table 1. The data for this report were collected during an intervention study investigating the effects of low-dose acetazolamide pretreatment on AMS development, which led to some methodological constraints as outlined in detail below. Participants were randomly assigned to receive acetazolamide (n = 9) or placebo (n = 13) before exposure to HA. They ingested 125 mg acetazolamide or placebo, once 10 hours and once 1 hour before arrival at HA. Participants reported no acute diseases before ascending to altitude and none was on any medication other than that of the intervention. They were transported by bus and cable car from LA to HA (Testa Grigia, 3480 m; Aosta valley, Italy). The study was approved by the Ethical Commission of the Medical University of Innsbruck and carried out in conformity with the ethical standards of the most recent version of the Declaration of Helsinki.
Baseline Characteristics of the Participants
Measurements
In the morning after the overnight stay at HA and in the afternoon, before performing the measurements, AMS symptoms were recorded by an experienced physician using the 1993 Lake-Louise-Scoring questionnaire (Roach et al., 1993). The highest score at any time point was used for data analysis. Participants were considered to suffer from AMS when they reported headache and had a score ≥3. Measurements were performed after a 10-minutes rest in a sitting position in the late afternoon at LA (574 m) and after the first overnight stay at HA (3480 m, also in the late afternoon). Assessments included body mass, heart rate (HR), peripheral oxygen saturation by pulse oximetry (SpO2, Onyx 9500 finger pulse oximeter; Nonin Medical, Inc.), systemic systolic and diastolic blood pressure (BP; Omron R3, Japan), and saliva sampling. Body mass was determined to the nearest 0.1 kg after emptying the bladder with participants wearing shorts and a t-shirt (∼5 hours after the last meal). On the day of the travel to the hut and during the first day at altitude, 24-hour fluid and food intake (amount and composition) were recorded in logbooks, and urine was collected in containers. Participants had access to the same limited number of known food products. Net water balance was calculated as fluid intake (drinking volume + soup volume, which was a major component of the food intake on the first day at altitude) minus urinary output neglecting insensible water loss. Saliva samples were obtained with Salivettes (Sarstedt, Germany) to quantify saliva cortisol levels, which have been shown to reflect changes in unbound serum cortisol (El-Farhan et al., 2017). Samples were taken pre-ascent and 3–4 hours after arrival at HA. Participants were not allowed to drink coffee or alcohol before saliva collection. During saliva sampling, the participants sat comfortably in a quiet room, placed a cotton swab in the mouth and chewed it for about 60 seconds to stimulate salivation. The samples were stored in a cool bag (4°C–8°C) and transported within 48 hours to the laboratory of G.B. (University of Salzburg, Austria) where cortisol levels were determined by radioimmunoassay as outlined in detail elsewhere (Carr et al., 1977).
Statistics
Data were analyzed using IBM SPSS Statistics 21. According to Shapiro–Wilk testing, all data were normally distributed. Changes of parameters from LA to HA were analyzed by paired Student's t-tests. To calculate differences between men and women and between the acetazolamide and the control groups, unpaired Student's t-tests or Mann–Whitney (for AMS score) tests were used. ANOVA with repeated measurement design was used to evaluate differences between the acetazolamide and the placebo groups and between men and women. Spearman correlation analysis was used to calculate associations between AMS scores and resting cortisol levels, HA fluid balance, HA body mass changes, and HA SpO2 levels. Multiple linear regression analysis with stepwise backward variable selection was used to identify predictors of the AMS score with pre-ascent cortisol level, sex, medication (acetazolamide or placebo), SpO2 at HA, fluid intake minus output during the first day at HA (or drinking volume), LA to HA changes in resting HR, and systolic BP as independent variables. We chose to include fluid balance and drinking volume in separate regression analyses because water content of food, metabolic water production, and insensible water loss, all of which might have influenced fluid balance, were not assessed. In addition, we decided not to include changes in body weight in the regression analysis, as measurements performed under nonstandardized conditions and in the late afternoon do not adequately reflect fluid balance (Armstrong, 2005). Data are presented as means ± SD or ranges. Due to missing data, sample size varies slightly (n in Tables 2 and 3). The level of significance was set at p ≤ 0.05.
Differences Between the Acetazolamide and Placebo Group in the Course of the Study Protocol
All parameters of the first part of the table were measured in the late afternoon. The AMS score represents the highest recorded score on the first day after an overnight stay at altitude. Fluid intake—output represents fluid balance on the first day at altitude.
A trend toward differences between the acetazolamide and the placebo group.
AMS, acute mountain sickness; ANOVA, analysis of variance; BP, blood pressure; ES, effect size (partial eta squared); HA, high altitude; HR, heart rate; LA, low altitude; SpO2, oxygen saturation; m/w, men/women.
Changes in the Parameters from Low Altitude to the High-Altitude Conditions
All parameters were measured in the late afternoon.
A trend toward significant changes.
Significant changes.
Results
The average highest AMS score for the entire group was 2.5 ± 1.9 (range 0–7) with the women having an average score of 2.8 ± 2.0 (range 0–7) and the men of 2.2 ± 1.8 (range 0–5) (p > 0.05). Five out of the 22 participants developed AMS, that is, headache and a score ≥3, 2 in the acetazolamide group (n = 9), and 3 in the placebo group (n = 13) (p > 0.05). SpO2 tended to be less decreased (−6.5% vs. −9.0%, p = 0.060) in the acetazolamide group; ΔBP tended toward increased values in the placebo group (9.6 ± 9.5 mmHg vs. −0.3 ± 13.3 mmHg, p = 0.062) (Table 2). Changes from LA to HA are shown in Table 3. In the entire group, SpO2 decreased from 97.0% ± 1.0% to 89.0% ± 2.9%, whereas resting HR (67.7 ± 7.9 to 85.8 ± 7.2 b/min) and systolic and diastolic BP (115.7 ± 15.0 to 121.6 ± 12.5 and 70.3 ± 11.3 to 78.9 ± 7.7, respectively) increased (all p < 0.05). BP values of women were unchanged, while for the entire group, they increased (Table 3). The AMS score was positively correlated with both LA cortisol levels and fluid balance during the first day at HA (r2 = 0.494 and r2 = 0.255, p < 0.05, respectively; Fig. 1). LA cortisol levels were associated with fluid balance during the altitude stay (r2 = 0.333, p < 0.05). SpO2 at HA showed a significant association with the AMS score (r2 = −0.438, p < 0.05). When applying stepwise multiple regression analysis, all variables were removed from the model except LA cortisol levels, which related to the AMS score (beta = 0.724, p = 0.001; adjusted r2 = 0.492, p = 0.001). In this data set, sex and medication (placebo vs. acetazolamide) were not related to the AMS score, alone or in multiple regression analysis.
Relationship of the highest AMS score at HA with the LA cortisol level and fluid balance on the first day at altitude as well as the association between the latter and the LA cortisol level (r2 and p-values represent Spearman's test outcomes). AMS, acute mountain sickness; HA, high altitude; LA, low altitude.
Discussion
The main findings of the present study were as follows: (1) LA pre-ascent saliva cortisol levels were associated with the AMS score; (2) during the altitude stay fluid retention and lower SpO2 levels were related to higher AMS scores; and (3) pre-ascent cortisol levels were associated with increased fluid retention at altitude. These findings suggest a link among baseline HPA-axis regulation, fluid balance, and AMS symptoms.
AMS etiology
AMS is the most common altitude illness, which predominantly develops in nonacclimatized persons rapidly ascending to HA (Hackett and Roach, 2001). Its etiology is not yet fully understood, and various mechanisms have been proposed. These include an increased sympathetic activity (Loeppky et al., 2003; Burtscher et al., 2008; Swenson, 2016), a blunted hypoxic ventilatory response (Burtscher et al., 2008), an altered redox status (Bailey et al., 2009; Lafuente et al., 2016; Irarrázaval et al., 2017), fluid retention (Bärtsch et al., 1991; Westerterp et al., 1996; Loeppky et al., 2005; Gatterer et al., 2013), elevated intracranial pressure (ICP) (Lawley et al., 2016), disruption of the blood–brain barrier (BBB) (Bailey et al., 2009), and activation of the trigeminovascular system (Bailey et al., 2009; Lawley et al., 2016). An interaction between these factors is likely. For example, chronic or transient elevations in ICP could activate trigeminal afferents (Lawley et al., 2016), free radicals may cause direct structural damage to the BBB (Chan et al., 1984), increased sympathetic tone may affect renal function and elevate vasopressin (Wortsman, 2002; Loeppky et al., 2003; Burtscher et al., 2008; Swenson and Olsen, 2014), excess cortisol may further contribute to fluid retention (Woods et al., 2012), while increased extracellular water levels may influence BBB permeability and ICP (Roach and Hackett, 2001; Gatterer et al., 2013). Furthermore, reduced diuresis might limit bicarbonate excretion, which would slow ventilatory acclimatization keeping arterial oxygen saturation low, which in turn could further impact AMS development (Burtscher et al., 2004).
AMS and cortisol
Several studies reported increased cortisol levels when going to HA (Sutton et al., 1977; Humpeler et al., 1980; Richalet et al., 1989; Sawhney et al., 1991; Ermolao et al., 2009; Woods et al., 2012), but reports are inconsistent (Smith et al., 2011) and only a few studies reported associations with AMS symptoms (Sutton et al., 1977; Richalet et al., 1989; Woods et al., 2012). Sutton et al. found the highest plasma cortisol levels in the most ill during an altitude stay (Sutton et al., 1977). In contrast, Woods et al. (2012) found no such relationship and reported decreased cortisol levels up to an altitude of 4270 m and then increases at higher altitudes (i.e., 5150 m). These divergent findings may to some extent be explained by differences in the type of altitude exposure, differing time points, and conditions of the cortisol sampling, as well as by inter- and intraindividual variability in cortisol responses to stressors (Woods et al., 2012). Circulating cortisol concentration follows a circadian rhythm (Torres-Ruiz et al., 2017) possibly impacting its association with AMS (Richalet et al., 1989). In addition, stressors such as exercise, bad weather, or unfamiliar circumstances may increase cortisol levels independently of altitude exposure (Woods et al., 2012), again potentially influencing relationships with AMS.
AMS and stress regulation
In the present study, late afternoon pre-ascent resting cortisol levels did not significantly increase upon ascent to altitude, but were associated with the later AMS score. Assuming that higher resting cortisol levels indicate higher general stress levels or a more stressful dealing with unfamiliar situations, we speculate that there is a link among an individual's stress response, cortisol, and AMS. In support of this contention, higher trait anxiety and higher levels of anxiety before a mountain ascent were reported in climbers susceptible to AMS (Missoum et al., 1992). In young healthy males at LA, Taylor et al. (2008) reported a relationship between trait anxiety and HPA-axis regulation. Boos et al.'s finding of higher trait anxiety at LA predictive for severe AMS at HA also supports such a link (Boos et al., 2018).
AMS, cortisol, and fluid balance
The relationship between cortisol levels and fluid balance further suggests that cortisol might contribute to fluid retention, which relates to AMS severity (Bärtsch et al., 1991; Westerterp et al., 1996; Loeppky et al., 2005; Gatterer et al., 2013). In line with this, cortisol acts as a weak mineralocorticoid inducing fluid retention by increasing renal tubular reabsorption of sodium (Garrod, 1958; Whitworth et al., 2000, 2005). Fluid retention at altitude may be further exacerbated by sympathetic activation and its action on renal function (Swenson and Olsen, 2014). A link among excess fluid intake, water retention, and AMS remains speculative, but increased extracellular water, together with other factors such as changed endothelial permeability and upregulation of inducible nitric oxide synthase, could influence BBB permeability and lead to brain tissue swelling, eventually increasing ICP and causing symptoms of severe AMS (Roach and Hackett, 2001). Also, fluid retention might limit the excretion of bicarbonate, necessary for the renal correction of the respiratory alkalosis from the hypoxia-induced hyperventilation (Cerretelli and Samaja, 2003). This would then slow ventilatory acclimatization, keeping SpO2 at relatively lower levels, prolonging the hypoxic stress, possibly impacting on AMS development (Burtscher et al., 2004).
Limitations
Some important limitations have to be acknowledged. First, the sample size was small, limiting the power for subgroup analyses, which were further limited due to missing data. Second, the data were collected in the course of a study, designed to investigate the effects of low-dose acetazolamide pretreatment on AMS development. Therefore, some constraints in the methodological approach to test the hypothesis of a relationship among saliva cortisol levels, fluid balance, and AMS score had to be accepted. Also, it might be argued that drug administration could have influenced outcomes. On the background of the low baseline AMS risk of our study, illustrated by the low prevalence of AMS, the lack of significant effects of acetazolamide and the limited sample size do not allow to conclude absence of effects. Acetazolamide, by inducing mild diuresis and enhancing bicarbonate loss, is expected to limit fluid retention and aid ventilatory acclimatization (Swenson, 2016). We therefore checked the acetazolamide effect on AMS and fluid homeostasis which turned out to be small to medium (Cohen's d: 0.3, p > 0.25, Table 2), illustrating its limited efficacy and numbers needed to treat at the used low dosage (Kayser et al., 2012). In addition, the regression analysis failed to recognize the intervention as a protective factor. To control for any effects of acetazolamide, by analyzing the placebo group separately, we still found significant relationships between AMS score and LA cortisol levels (r2 = 0.799, p < 0.001), between fluid balance and LA cortisol levels (r2 = 0.537, p = 0.010), and a trend toward significance was found between AMS score and SpO2 (r2 = 0.301, p = 0.051). This is another indication that any acetazolamide effects were minor. Also, worth mentioning is that acetazolamide intake was found not to change plasma cortisol levels in subjects not exposed to altitude (Frayser et al., 1975) and that, at least in rats, it had no effect on salivary secretory rate (Young et al., 1987). Third, the physiological measurements were taken in the late afternoon. Even though circadian rhythms of parameters such as cortisol levels were controlled for in the present investigation as LA and HA measurements were performed at the same time of day, it should be considered when comparing the present results with results from literature. Fourth, body mass was measured in the late afternoon and changes might not adequately represent fluid balance changes (Armstrong, 2005). Accordingly, no significant relationship among fluid balance, fluid intake, or fluid loss with body weight changes was found. Therefore, body mass changes were not included in the regression analysis. Concerning fluid balance measurements, it should be mentioned that the water content of the (limited) solid food intake was not considered. Finally, we do not have any information on trait anxiety scores in our participants.
Conclusions
This study suggests that resting LA cortisol levels, which are associated with fluid retention at HA, may be connected to AMS risk. This suggests a link among individual stress regulation, cortisol homeostasis, water balance, and AMS risk that merits to be further investigated.
Footnotes
Acknowledgments
We thank Prof. Prim. Dr. Galvan and the Salzburger Landeskrankenanstalten, Abteilung für Nuklearmedizin for their support and the analyses of the cortisol levels.
Author Disclosure Statement
No competing financial interests exist.