Risk factors for acute mountain sickness Predicting who will develop AMS Tight-fit hypothesis and increased intracranial pressure Other potential contributors to development of AMS Following the earliest descriptions of altitude illness by Acosta, described in detail in Chapter 1, the first modern account of acute altitude illness was provided in 1913 by Ravenhill (1913), who was serving as a medical officer for a mining company whose mines were located at 4700 m in Chile. Noting that fatigue, cold, and lack of food complicated previous descriptions by explorers and mountain climbers, he had the opportunity to observe individuals suffering from the effects of altitude alone because the mines were served by a railway, which allowed rapid ascent to high elevation. Drawing on the local Bolivian term for altitude illness, he referred to the illness as puna of the “normal” type, and described it in the following manner: He also went on to describe puna of the nervous and cardiac types, which correspond in our present nomenclature to the other forms of acute altitude illness, high altitude cerebral edema (HACE) and high altitude pulmonary edema (HAPE), respectively. After his descriptions, altitude illness was well recognized, but the distinction between and importance of the two more severe forms seem to have been lost, until Lizárraga (1955) described the reentry form of HAPE in Peru in 1955, Houston (1960) and Hultgren and Spickard (1960) described the form that affects unacclimatized lowlanders in 1960, and Fitch (1964) subsequently described HACE in 1964. This chapter marks the transition in this book from a detailed review of high altitude physiology to consideration of these illnesses and other clinical problems that affect people traveling at high altitude and focuses in detail on the first and most common of the problems described by Ravenhill, acute mountain sickness (AMS). HACE and HAPE will be discussed further in Chapters 21 and 22. After considering the scope of the problem, risk factors for its development, and the current understanding of AMS pathophysiology, the chapter provides a detailed review of the clinical aspects of AMS including the clinical presentation, diagnostic approach, and strategies for prevention and treatment. A fuller account of the historical features of AMS and the other forms of acute altitude illness is provided in a detailed monograph on the topic by West (1998). Of the main forms of acute altitude illness, AMS is by far the most common, but the documented incidence varies from as low as 25% to as high as 84% due to a variety of factors. The reported incidence rises, for example, as measurements are made at steadily higher elevations (Meier et al. 2017). For example, Honigman et al. (1993) documented an incidence of 25% in tourists traveling to 1900–2400 m in Colorado compared to incidence rates of 43% among trekkers at 4343 m in Pheriche, Nepal (Hackett et al. 1976), and rates as high as 75–77% on Kilimanjaro (5895 m) (Davies et al. 2009; Karinen et al. 2008). Another excellent demonstration of the effect of altitude on the reported incidence was a study of climbers at European alpine huts by Maggiorini et al. (1990) in which they documented rates of 9% at 2850, 13% at 3050 m, and 34% at 3650 m. Even at a given elevation, incidence may differ, however, based on the timing of the assessment. Assessment of AMS symptoms after only a few hours of exposure may lead to lower incidence than if subjects are assessed after longer periods of time as symptoms are typically most prominent after the first night at a given elevation. Another factor with a significant effect on the reported incidence rate at a given elevation is the mode by which those assessed in a given study ascend to high altitude. In their study of trekkers arriving in Pheriche, for example, Hackett et al. (1976) found that trekkers who flew into Lukla had an incidence of 49%, compared to 31% in those trekkers who walked the entire distance from Jiri. Similarly, Murdoch (1995) studied guests at the Everest View Hotel in the same region of Nepal and found an incidence of 84% among those who flew to directly to an airstrip near the hotel at 3740 m compared to 61% among those who trekked to the hotel from altitudes under 3000 m over a few days. Finally, Basnyat et al. (2000) noted a very high incidence of AMS (68%) among religious pilgrims, who traveled by bus to 2000 m and then rapidly ascended on foot to the festival site at Gosainkunda Lake (4300 m). It is important to remember, however, that even among people trekking rather than driving or flying to high elevation, the reported incidence of altitude illness can still be quite high depending on the altitude attained and the overall rate of ascent. This is best demonstrated by the very high incidence rates reported on Kilimanjaro, noted above, which likely reflect the notoriously fast ascent profiles on many of the routes used to climb that mountain. An under recognized factor affecting reported incidence rates is the proximity of a given region to the equator. West et al. (1983) have shown that barometric pressure at a given elevation is latitude-dependent due to differences in air mass over the equatorial and polar regions. As a result, ambient oxygen tensions and AMS risk at 4000 m near the equator, for example, may be different than those observed at the same absolute elevation at other latitudes. This issue is nicely demonstrated by studies reporting incidence rates of 52–55% among workers transported by air to the South Pole (2835 m) (Anderson et al. 2011; Anderson et al. 2015), far higher than that reported for people driving to similar elevations at lower latitudes (Honigman et al. 1993). This may be attributable, in part, to the fact that the barometric pressure at that elevation results in physiologic altitudes of about 3400 m. Finally, different scoring systems, including the Lake Louise Acute Mountain Sickness Score (Roach et al. 1993; Roach et al. 2018), the Environmental Symptoms Questionnaire (Sampson et al. 1983), and a scoring system used in some Chinese studies (Ren et al. 2010), are used to diagnose AMS in research studies and different numerical thresholds can be used within those systems (described later in this chapter). Depending on which system and numerical threshold are used, different incidence rates may be reported for a particular location. One question worth considering is whether increasing awareness of acute altitude illness and educational campaigns over time by groups such as the Himalayan Rescue Association in Nepal have decreased the occurrence of AMS in certain regions. Gaillard et al. (2004), for example, compared the prevalence of AMS in separate cohorts of trekkers in the Annapurna region of Nepal in 1986 and 1998 and found an increase in altitude illness awareness (80% to 95%) and a decrease in AMS prevalence (43% to 29%) as well as slower ascent profiles in the later cohort. More recently, McDevitt et al. (2014) surveyed English-speaking trekkers at 3500 m on the Annapurna circuit and found lower rates of AMS than documented in previous studies, which they attributed to slower ascent profiles and increased use of acetazolamide. The positive effects of these educational efforts may be balanced out, however, by the increase in road construction in certain mountain regions that allows much more rapid access to higher elevations for large numbers of people who may not be aware of issues related to acute altitude illness (Reisman et al. 2017). By far the most important reason individuals develop AMS is that they ascend too high, too fast. An individual, for example, who travels to and sleeps at 4500 m over only two days is more likely to develop symptoms than someone who spends five days reaching the same sleeping elevation. Perhaps the best demonstration of this issue are the studies by Hackett et al. (1976) and Murdoch (1995), noted earlier, as well as a study by Schneider et al. (2002), in which AMS was far more common among those who flew to high elevation compared to those who trekked the entire way. Beyond this important factor, however, there are many variables that have been investigated as risk factors for AMS. For any given altitude attained or ascent profile, there is marked interindividual variability in the development of AMS. Based on this observation, there has long been suspicion that genetic differences account for this variability. While Kriemler et al. (2014) have shown evidence of familial clustering of AMS following fast ascent to 3500 m, no specific genetic polymorphisms have been convincingly identified as playing a role. Studies have examined the role, for example, of the bradykinin receptor B2 gene (Wang et al. 2010) and ACE gene polymorphisms (Dehnert et al. 2002; Kalson et al. 2009) but no definitive links were found with AMS susceptibility. MacInnis and Koehle (2016) reviewed the evidence for a genetic predisposition to acute altitude illness and concluded that whereas genomic studies have identified genes that likely influence the risk of HAPE, no convincing associations have been established regarding either AMS or HACE. Following publication of this study, Yu et al. (2016) reported evidence that a specific polymorphism of the EPAS1 gene may be protective against AMS, but further support for this or other associations is lacking. In the case of AMS, a link has been difficult to establish given the unclear pathophysiology of the disease as well as the subjective nature of the diagnostic criteria (discussed below). While the phenotype of HACE is better defined and brain imaging permits more accurate diagnosis, the incidence of the disease is so low as to preclude adequate genomic studies (MacInnis and Koehle 2016). A common misperception is that good physical fitness is protective against the various forms of altitude illness. This is not the case, as demonstrated by Milledge et al. (1991), who found no correlation between exercise capacity, as measured by Despite the fact that good physical fitness is not protective against AMS, physical training prior to a high altitude sojourn is still important and should still be advised in those who do not exercise on a regular basis. The empirical observation that good physical fitness improves exercise tolerance in a hypoxic environment, prevents physical exhaustion, which can mimic AMS, and provides an increased margin of safety on prolonged trekking and mountaineering expeditions is supported by recent data showing that individuals with greater sea-level fitness experienced a lower sense of effort and less fatigue when trekking at high altitude compared to less fit individuals (Rossetti et al. 2017a). High altitude travelers are often counseled to avoid heavy exercise immediately following ascent, but the evidence for this recommendation is mixed. One study often cited to support the recommendation is that by Roach et al. (2000) in which they exposed seven subjects to approximately 4800 m altitude in a chamber for 10 hours on two separate occasions, one spent entirely at rest and the other with intermittent exercise and found significantly higher AMS scores during the exposure with exercise. While several other studies have also shown a relationship between long duration exercise and AMS (DiPasquale et al. 2015; Kammerer et al. 2018), others have failed to establish that exercise worsens AMS symptoms. For instance, Schommer et al. (2012) found no difference in AMS severity when subjects were exposed to normobaric hypoxia equivalent to 4500 m on separate occasions, one at rest and the other involving three 45-minute bouts of exercise at 50% of the subject’s altitude specific maximum exercise capacity. Similarly, Rupp et al. (2013) compared AMS symptoms in 12 healthy men under three conditions—(1) three 80-minute bouts of cycling at 45% of the altitude specific maximum power output spread over an 11-hour exposure to an FIO2 of 0.12; (2) a similar exercise protocol in normoxia; and (3) at rest during an 11-hour exposure to the same degree of hypoxia—and found AMS scores were no different between the hypoxia-exercise and hypoxia-rest conditions but were significantly lower during normoxia-exercise. A particular strength of this study is that by including a session of exercise in normoxia, they could separate out the effects of exercise on well-being from the effects of hypoxia. One important difference between these studies is that, whereas the subjects in the studies by Schommer et al. (2012) and Rupp et al. (2013) resided at sea level, all of the subjects in the study by Roach et al. (2000) resided between 1600 m and 1800 m in elevation, which may have affected their responses to the hypoxic exposure. One other potential limitation in the research in this area is that this question has generally been addressed with chamber experiments rather than through exercise studies in the field environment. Conflicting data have been reported regarding the effects of age. While many studies report no difference (Horiuchi et al. 2016a; Kayser 1991; Stokes et al. 2010; Ziaee et al. 2003), other studies have reported either a decreased (Gonggalanzi et al. 2016; Hackett et al. 1976; McDevitt et al. 2014; Roach et al. 1995; Salazar et al. 2012) or increased risk (Bian et al. 2015). Wu et al. (2018) attempted to synthesize the data on this issue through a meta-analysis of 17 observational studies and found no clear relationship between age and AMS risk. Graham and Potyk (2005) suggest that some reasons for the variability between studies may be due to methodological problems in many of the observational studies, as well as the fact that nonspecific symptoms of AMS may be attributed to comorbid conditions that increase in prevalence with age, thereby leading to underreporting of AMS. Sex also does not have a clear effect on AMS risk. While Kayser (1991) found females to have a higher rate of AMS than males (69% versus 57%) among trekkers going over the Thorong pass in Nepal (5400 m), Honigman et al. (1993) and other more recent observational studies (Horiuchi et al. 2016a; Lawrence and Reid 2016; Schneider et al. 2002) have found no difference in this regard. Similarly, in their study deriving a model for assessing the risk of severe high altitude illness, Canouï-Poitrine et al. (2014) examined the role of sex but did not find any statistically significant effect on the odds ratio for severe altitude illness. Data regarding the effects of body habitus are more consistent as multiple studies suggest that obese individuals are at greater risk for AMS than the nonobese (Hirata et al. 1989; Honigman et al. 1993; Kayser 1991; Ri-Li et al. 2003; Yang et al. 2015). The degree of obesity necessary to increase risk is not clear from these studies, nor is the underlying mechanism for the association. One potential explanation is that altered lung mechanics and metabolic derangements in the severely obese may impair ventilation, thereby leading to exaggerated degrees of hypoxemia that trigger acute altitude illness (San Martin et al. 2017). Historically, there had been an unsubstantiated impression among some mountaineers and Sherpa that smokers have less AMS than nonsmokers. Although evidence is lacking, some habituation to a modest level of carboxyhemoglobinemia from smoking conceivably could result in some degree of preacclimatization. While several prospective cohort studies (Sanchez-Mascunano et al. 2017; Song et al. 2014) found evidence of a protective effect of smoking, other studies suggest smoking does the opposite and increases the risk of AMS (McDevitt et al. 2014; Vinnikov et al. 2015). Several recent meta-analyses have not provided any clarity on this issue as they have demonstrated either a protective role (Xu et al. 2016) or no effect at all (Masuet-Aumatell et al. 2017; Vinnikov et al. 2016). Fortunately, there are multiple clear adverse effects of smoking on health and, as a result, practitioners have ample reasons to recommend smoking cessation rather than ongoing tobacco use to prevent AMS. Several studies have shown that a history of migraine or other headaches at low altitude is a risk factor for headache or AMS following ascent (Bian et al. 2013; Burtscher et al. 2011; Davis et al. 2016), although not all studies provided information on how they distinguished migraine from other types of headache. These studies are limited by the fact that the effect of migraine or other headaches was not evaluated in multivariate models that included well-established risk factors for AMS such as a prior history of AMS, rate of acclimatization, and degree of preacclimatization. Studies using multivariate models that include such risk factors have shown conflicting results. Richalet et al. (2012) and Canoui-Poitrine et al. (2014), for example, demonstrated that, along with other variables, migraine headache is a risk factor for severe altitude illness in a model derived from and validated in very large cohorts of high altitude travelers. In contrast, Schneider and Bartsch (2018) performed a secondary analysis of data obtained from 1320 mountaineers staying at 4559 m and found that a history of migraine headache did not contribute to the development of AMS when multivariate analysis was conducted using the well-established risk factors noted above. From a theoretical perspective, it would make sense that impaired hypoxic ventilatory responsiveness (HVR) would predispose to AMS. Individuals with a blunted response would have lower alveolar and arterial PO2 values at any given altitude and might, therefore, become more symptomatic than those with stronger responses and higher oxygen tensions. Several reports have provided support for this concept, demonstrating blunted ventilatory responses to hypoxia in individuals who had developed AMS or HAPE (Lakshminarayan and Pierson 1975; Matsuzawa et al. 1989). However, a number of other field and chamber studies have failed to find a relationship between HVR measured prior to high altitude travel and subsequent development of AMS (Milledge et al. 1988; Milledge et al. 1991; Richard et al. 2014; Savourey et al. 1995). Hohenhaus et al. (1995), for example, found that compared with healthy individuals, HVR was significantly lower in subjects who developed HAPE but not in subjects with AMS, while Bärtsch et al. (2002) found no relationship between HVR, measured at sea level, and subsequent AMS upon ascent to 4559 m. In the latter study, however, they did document that failure to increase HVR on the first day at high altitude was associated with a higher incidence of AMS and more hypoxemia. While these studies generally measured the ventilatory response to hypoxia at rest, in a very large prospective cohort study, Richalet et al. (2012) demonstrated that a blunted ventilatory response to hypoxia during exercise was one of several factors associated with increased risk of developing severe acute altitude illness. A large number of studies have attempted to determine whether an individual’s oxygen saturation at high altitude affects the likelihood of developing AMS, with some studies reporting a link between hypoxemia and development of AMS (Basnyat et al. 1999; Burtscher et al. 2004; Karinen et al. 2010; Roach et al. 1998; Tannheimer et al. 2002) and others reporting no relationship (Leichtfried et al. 2016; O’Connor et al. 2004; Roach et al. 1995). In considering these studies, however, it is important to note that only a few use a prospective approach and examine whether the presence of hypoxemia early in the trip predicts the development of AMS later in the stay at high altitude (Karinen et al. 2010; Roach et al. 1998; Tannheimer et al. 2002). Roach et al. (1998), for example, measured oxygen saturation (SpO2) in 102 climbers at 4200 m on Denali, questioned them about AMS symptoms on their return from their summit bids, and found that the SpO2 measured before climbing from base camp had a modest correlation (r = 0.48) with subsequent AMS scores. Many of the other the studies on this question measured SpO2 at the same time the subjects were assessed for AMS (Basnyat et al. 1999; Koehle et al. 2010; O’Connor et al. 2004) and, as a result, do not inform the discussion about whether a low SpO2 upon arrival at high altitude predisposes the individual to develop AMS at a later time. Regardless of the particular study design, however, one of the challenges of using oxygen saturation to predict risk of developing AMS is the large interindividual variability in SpO2 among healthy at a given elevation (Luks and Swenson 2011). In the study by Roach et al. (1998), for example, even though the average saturation was lower in those with AMS scores ≥3 than among those with lower scores, there was large overlap in SpO2 between the two groups; many of the subjects in the group with AMS scores of zero had SpO2 values as low as those seen among those with the highest AMS scores. Another marker of the adequacy of gas exchange besides the oxygen saturation is the diffusion capacity for carbon monoxide (DLCO). This parameter has also not been shown to have a clear relationship to the risk of developing AMS. Ge et al. (1997), for example, measured DLCO in 32 subjects at 2260 m and following ascent to 4700 m and found those individuals who developed AMS saw smaller increases in their DLCO following ascent than those without AMS, a finding they attributed to the development of subclinical pulmonary edema. Dehnert et al. (2010) investigated this same question and found no difference in the DLCO (adjusted for alveolar volume) between individuals with and without AMS both upon arrival and following two days at 4559 m. The risk of AMS generally increases following rapid ascent to >2300 m from altitudes below 1000 m, although cases have been identified at elevations as low as 1900 m. Importantly, the altitude at which symptoms start varies significantly between individuals, with very susceptible people developing symptoms below 2300 m and others not developing symptoms until much higher elevations, if at all. Symptoms typically come on over a period of hours at a given elevation rather than being present immediately upon ascent. In their study of a general tourist population traveling to 1900–2950 m in Colorado, for example, Honigman et al. (1993) found that 65% developed symptoms within 12 hours of arrival at altitude whereas in another 34% onset occurred between 12 and 36 hours following arrival. Onset after 36 hours is uncommon and, as a result, symptoms developing after this time frame warrant a careful search for alternative etiologies. In the majority of individuals, symptoms are self-limited and resolve without treatment over a period of one to four days as the body acclimatizes to hypobaric hypoxia, although in rare cases of individuals with very poor altitude tolerance, symptoms may persist for longer periods of time. Even though symptoms resolve in the majority of cases, they may recur with subsequent large gains in elevation if the individual does not take appropriate precautions. AMS is marked by a constellation of symptoms, including headache, anorexia, nausea, vomiting, fatigue, and persistent light-headedness. Headache is the cardinal feature of AMS and is typically bilateral, mild-moderate in intensity, throbbing or pulsatile in character, frontal, temporal, or diffuse in location, and aggravated by movement, bending over, coughing, or Valsalva maneuver (Silber et al. 2003). Disturbed sleep was previously considered one of the main symptoms of AMS, but it has been increasingly recognized that poor sleep is a common problem at high altitude, even for those who are not sick and therefore is no longer seen as one of the primary symptoms (Roach et al. 2018). Pulmonary symptoms such as cough and dyspnea out of proportion to that normally experienced at high altitude are not features of AMS. Similarly, diarrhea is not one of the gastrointestinal manifestations of AMS, and its presence should prompt consideration of other etiologies. There are no physical examination findings specific to AMS. Consistent changes in blood pressure and heart rate have not been found, although Singh noted the presence of bradycardia in two-thirds of soldiers with AMS between 3350 m and 5490 m (Singh et al. 1969). Crackles on auscultation and peripheral edema have been reported but may also be seen in individuals without AMS (Hackett and Rennie 1979). Variable changes in body temperature have also been reported with some studies reporting a mild increase of between 0.5°C and 1.2°C depending on the level of AMS severity (Maggiorini et al. 1997) and others reporting a decrease of 1.7°C with no change in metabolic rate (Loeppky et al. 2003). Lower SpO2 values have also been reported in those with AMS compared to those who remain healthy (Basnyat et al. 1999), but AMS can still be seen in those with relatively preserved gas exchange. Importantly, individuals with AMS should have no neurologic symptoms or signs, aside from headache and persistent dizziness. Findings such as altered mental status, ataxia, or focal neurologic deficits should prompt consideration of HACE and other diagnoses. Although the general principles are the same regardless of the setting, the diagnosis of AMS varies to some extent depending on whether diagnosis is being made in a clinical or research setting. Individuals are deemed to have AMS if they have headache and one or two other symptoms such as nausea, vomiting, fatigue, or persistent lightheadedness and dizziness upon standing following a recent ascent to over 2300 m. The diagnosis is made solely on the basis of the reported symptoms, as there are no characteristic physical examination findings or diagnostic laboratory tests. The individual should have a normal neurologic examination and mental status. If not, consideration must be given to other diagnoses, HACE in particular. If an individual ascended rapidly by car, plane, or cable car, a period of several hours should have passed before the onset of symptoms, while those who ascended more gradually by climbing or trekking can have symptoms immediately upon arrival as they may have already spent time above their altitude threshold for developing symptoms. Individuals acclimatized to a given elevation can still develop AMS with an abrupt, large gain in elevation above their current location. Traditionally, the presence of headache has been required to diagnose AMS but a large Chinese study on altitude illness in military recruits in Tibet used a set of novel criteria that did not require headache (Ren et al. 2010), which prompted debate as to whether headache is necessary for the diagnosis (Roach et al. 2011; West 2011). In a recent update of the diagnostic criteria for AMS, however, headache was still a required symptom (Roach et al. 2018). Diagnosis can be difficult in preverbal children less than 4 years of age who cannot adequately report symptoms to their parents. To address this problem, Yaron et al. (1998) created the Children’s Lake Louise score, which uses a fussiness score in lieu of the headache score in the standard Lake Louise criteria and a pediatric symptom score to assess appetite, vomiting, playfulness, and ability to sleep. They showed that parents were capable of using this scoring system in an appropriate manner. In general, the safest course is to assume that behavior such as irritability, tearfulness, and refusal of food in a child who has gained altitude in the previous hours or days indicates AMS until proven otherwise (Yaron and Niermeyer 2008). Although AMS is the most likely cause of symptoms shortly following a large acute gain in altitude, it is important to remember that the symptoms are highly nonspecific and, as a result, may be caused by other disorders such as exercise-associated hyponatremia (Ayus and Moritz 2008), dehydration, viral syndrome, migraines, or carbon monoxide poisoning (Table 20.1). The last entity may occur in individuals using cooking stoves in poorly ventilated tents, although a study from Denali did not reveal any correlation between carbon monoxide levels and AMS symptoms (Roscoe et al. 2008). The prudent approach is to attribute symptoms that develop following ascent to acute altitude illness but to consider alternatives when clinical features are suggestive or the patient does not respond to appropriate therapeutic interventions. Acute psychosis Acute toxic encephalopathy due to substance ingestion Alcohol hangover Carbon monoxide intoxication Cerebrovascular accident Dehydration Encephalitis Hypoglycemia Hyponatremia Hypothermia Meningitis Migraine headache Physical exhaustion Ruptured intracranial aneurysm or arteriovenous malformation Research studies require a more systematic approach to diagnosis and, in particular, that researchers use the same system to facilitate comparison of results between studies. The oldest and most complicated scoring system is the Environmental Symptom Questionnaire (ESQ) (Sampson et al. 1983), consisting of 67 questions, graded from 0 (absent) to 5 (extremely severe). An observer can administer the questions or the subject can complete a self-assessment, as the two methods give similar results. Factorial analysis was used to identify nine distinct symptom groups, two of which refer to acute altitude illness: the AMS-C score, which emphasizes cerebral symptoms, and the AMS-R score, which focuses primarily on respiratory symptoms. Beyond the fact that a 67-item questionnaire can be cumbersome to administer, particularly on a repeated basis, it was subsequently determined that there was high correlation between the AMS-C and the other subscores at high altitude, making it difficult to distinguish illness due to hypoxic exposure from other environmental stresses (Bartsch et al. 1993b). Because the 11-item AMS-C (Table 20.2) score can identify AMS with moderate sensitivity and high specificity (Bartsch et al. 1993b; Sampson et al. 1983), it is now relied upon to diagnose AMS in the research setting rather than the full ESQ. Dim vision Dizzy Faint Feel hungover Feel sick Feel weak Headache Lack of coordination Lightheaded Loss of appetite Sick to stomach Another widely used, simple approach, referred to as the Lake Louise Scoring System (LLSS), was created at the International Hypoxia Symposium in 1991 and later modified at the following symposium in 1993 (Roach et al. 1993). The score was based on self-reported answers to questions about five symptoms, including headache, gastrointestinal upset, fatigue/weakness, dizziness/lightheadedness, and sleep disturbance, graded on a scale from 0 (absent) to 3 (severe). The score could be supplemented with a separate score assessing physical examination findings (ataxia, mental status, peripheral edema) and a functional score assessing the reduction in activity due to reported symptoms. When the symptom score alone was used, an individual was deemed to have AMS based on the presence of headache and a total score ≥3. This original LLSS included sleep disturbances as part of the diagnostic criteria, but it was increasingly recognized over time that central sleep apnea and disturbed sleep were not always present in people with AMS and, importantly, occurred in those who were otherwise well. In addition, in many cases, the questionnaire was being administered shortly following arrival at high altitude, making it impossible to assess sleep disturbances in those who had not tried to sleep to that point. Based on several well-done studies that examined this issue in greater detail and confirmed that disturbed sleep correlated poorly with other symptoms of AMS (Hall et al. 2014; Macinnis et al. 2013), a consensus panel subsequently revised the diagnostic criteria in 2018, removing the sleep question and basing the AMS score on questions about four symptoms including headache, gastrointestinal symptoms, fatigue and/or weakness, and dizziness/lightheadedness (Table 20.3) (Roach et al. 2018). Under this new system, AMS is still defined as a Lake Louise Acute Mountain Sickness score ≥3 points following a recent gain in altitude, including at least one point for headache. A Clinical Functional Score is also included in the updated system but is not taken into account when computing the AMS score. Component Score Description Headache 0 None at all 1 Mild headache 2 Moderate headache 3 Severe, incapacitating headache Gastrointestinal symptoms 0 Good appetite 1 Poor appetite or nausea 2 Moderate nausea or vomiting 3 Severe, incapacitating nausea and vomiting Fatigue and/or weakness 0 Not tired or weak 1 Mild fatigue/weakness 2 Moderate fatigue/weakness 3 Severe, incapacitating fatigue/weakness Dizziness/light-headedness 0 No dizziness/light-headedness 1 Mild dizziness/light-headedness 2 Moderate dizziness/light-headedness 3 Severe, incapacitating dizziness/light-headedness AMS Clinical Functional Score Overall, if you had AMS symptoms, how did they affect your activities? 0 Not at all 1 Symptoms present but did not force any change in activity or itinerary 2 My symptoms forced me to stop the ascent or to go down on my own power 3 Had to be evacuated to a lower altitude As an alternative to the ESQ and LLSS, Wagner et al. (2007) proposed the use of a visual analog scale (VAS) to diagnose and grade the severity of AMS. Kayser et al. (2010), however, showed that the utility of this approach is limited and comparisons between studies done using either approach may be difficult. Even though the VAS and LLSS scores had a reasonable correlation (r = 0.84), the respective scores do not scale linearly and instead show a threshold effect; with low LLSS scores, the VAS scores all fell below the identity line, while in the higher range of LLSS scores, VAS scores fell above the line of identity. Another set of criteria was developed in China in 1996 that did not actually require headache for the diagnosis of AMS. Because the article laying out these criteria was published in Chinese, it remained largely unknown until a large study on the incidence of AMS in military recruits was published using these diagnostic criteria and noted a very high incidence of AMS (Ren et al. 2010; West 2011). This prompted a search for the criteria and subsequent publication of an English language translation (West 2010). No studies have adequately compared these criteria to the AMS-C or LLSS, which remain the primary means of diagnosing AMS in the majority of published studies. In considering the AMS-C and LLSS, several issues warrant further attention. One challenge with assessing the performance of the various scoring systems is the lack of a diagnostic gold standard for AMS. Using a Lake Louise Score ≥5 from the original scoring system as the reference measure, Meier et al. (2017) conducted a systematic analysis of 91 studies comprising 66,944 study participants and found that the VAS, AMS-C score, and clinical functional score performed similarly to the original LLSS, with sensitivity for the diagnosis varying between 67% and 82% and specificity ranging from 67% to 92%. One challenge in this study, however, is the significant heterogeneity between the studies included in the analysis. When the same investigators use the same methods across separate studies, the sensitivity and specificity of a scoring system are increased (Macholz et al. 2018). How well the AMS-C and other scoring systems compare to the most recent version of the LLSS is unclear at this time. Using the AMS-C system, an individual is deemed to have AMS with a score ≥0.7. This threshold, which was chosen to increase specificity at the expense of lower sensitivity (Sampson et al. 1983), is roughly equivalent to a Lake Louise Acute Mountain Sickness score of either ≥4 (Maggiorini et al. 1998) or ≥5 (Meier et al. 2017). No studies have yet compared the new LLSS to the AMS-C, but because the sleep component, which contributed to the score in many ill people under the old system, has now been eliminated, one can expect that an AMS-C ≥0.7 will roughly correspond to an updated LLSS ≥3. The updated guidelines recommend assessing the AMS score after six hours following ascent to avoid misclassification of symptoms related to travel or exercise as AMS. The importance of delaying assessment in this manner to increase specificity is nicely highlighted by a study by Muhm et al. (2007) in which they monitored 502 individuals at simulated altitudes of 198 m, 1220 m, 1830 m, 2134 m, or 2440 m, assessing for AMS every two hours for up to 20 hours using the ESQ. While the prevalence of AMS did increase over time and was highest after 10–12 hours, some individuals met criteria for AMS after only two hours in the chamber. Further highlighting the nonspecific nature of the main symptoms of AMS, individuals in this study reported AMS-compatible symptoms, such as fatigue and malaise even at the lowest two simulated altitudes, far below those at which people develop true AMS. Both the LLSS and the ESQ are English language questionnaires. Although many studies in international settings translate the instruments into other languages, translation is not always performed and some non-English speakers may be using the English language questionnaire. This issue may be significant for comparing results between studies, as Dellasanta et al. (2007) have shown that use of English questionnaires affected measured outcomes in some nationalities. The old LLSS could be used in children aged 4–11 years, although a different diagnostic threshold may be necessary for this group, as children in this age range tend to over-report symptoms (Southard et al. 2007). Whether the new system can be used in children in this age range has not been established, but given the similarity with the old system, there is no reason to suspect application of the new guidelines to children should cause new problems. Although the majority of studies that include AMS as an outcome measure use either or both the ESQ and LLSS, there remains considerable variability between studies in terms of defining and reporting basic parameters, a problem that makes comparison of results between studies and conduct of meta-analyses challenging. To address this problem, a group of experts in clinical high altitude research used a Delphi method to identify 42 key parameters in five categories (setting, individual factors, AMS and headache, HACE, HAPE, and treatment) that should be reported in all clinical high altitude research studies (Brodmann Maeder et al. 2018). It remains too early to determine the extent to which researchers will adhere to these guidelines. Given the high rate of AMS among travelers to certain elevations, identifying those susceptible to the disorder would be useful as it could assist in development of preventive strategies for these individuals. Unfortunately, identifying susceptible individuals among altitude naive travelers has proven challenging. In what is clearly the strongest and most comprehensive effort to date on this issue, Richalet et al. (2012) proposed a multivariate predictive model to identify individuals susceptible to severe high altitude illness (severe AMS, HACE, or HAPE) with the key factors being a prior history of severe altitude illness, a history of migraine headaches, decreased ventilatory responses during hypoxic exercise, and desaturation >22% during hypoxic exercise. This model, which does not address susceptibility to less severe forms of AMS, was based on a very large derivation cohort and was then validated in an equally large prospective cohort. In a subsequent refinement of this model, Canouï-Poitrine et al. (2014) developed separate prediction scores depending on whether or not the individual had previous experience at high altitude. One problem with some of these proposed decision tools, however, is that they involve testing strategies, such as a hypobaric chamber or a cardiopulmonary exercise testing with hypoxic exposure, that might not be available to all providers trying to assess risk for a given individual. Canouï-Poitrine et al. (2014) did examine the performance of the models with and without inclusion of the hypoxic exercise test and found that discrimination was affected to some extent by removal of this parameter. The area under the curve decreased from 0.8 to 0.72 when the exercise data were excluded in people with no prior experience at high altitude and from >0.9 to 0.84 in those with prior experience. In the end, whether the expense and time necessary to complete these tests are warranted in light of the self-limited and mild nature of the symptoms that many individuals will develop at high altitude and the ease of simple, less costly preventive strategies (discussed later) is also an open question. Whether such laboratory-based prediction strategies are superior to simply evaluating the risk of a proposed ascent profile (Luks et al. 2019) also remains unclear. In contrast to HAPE, where the mechanism underlying its development is relatively well understood (Chapter 22), the underlying pathophysiology of AMS remains unclear despite substantial research in this area. Given that headache is the primary symptom of AMS, the discussion of AMS pathophysiology must begin by considering the mechanisms behind the development of headache, in general, and how these mechanisms may relate to events at high altitude. Headache consists of pain above the orbitomeatal line and is thus easily identifiable and is specific to the head. However, the brain parenchyma itself is insensate (Ray and Wolff 1940) and thus does not convey painful stimuli. Instead, sensory feedback is produced by the stimulation of the trigeminal nerve, predominantly via the first ophthalmic nerve, to intra- and extracranial structures (the trigeminovascular network, Figure 20.1) (Goadsby et al. 2009). In a landmark set of experiments, Ray and Wolff (1940) mapped referred head pain to the stimulation of specific cerebral structures. These observations documented the striking somatotopical selectivity of trigeminal nerve innervation; that is, the location of head pain often neighbors the site of trigeminal activation. For example, patients reported pain isolated to the left temporal region when the left middle cerebral artery was distended. Conversely, pain isolated to the lower back of the head was reported following stimulation of the lower part of the occipital sinus (Figure 20.2). These experiments established that many headaches are characterized and diagnosed through a distinct and reproducible set of self-reportable pain characteristics (ICHD-2 2004). Figure 20.1Trigeminal vascular network. (A) Systematic representation of the ascending and descending components of the trigeminal vascular system highlighting that trigeminal activation maybe due to mean distension (e.g., dilation), prolonged low-grade arterial venous stretch (e.g., increase arterial or venous pulsatility), or chemical activation (e.g., inflammation). (B) Schematic representation of a possible scenario explaining high altitude headache. In the proposed model, arterial vasodilation initially causes trigeminal activation and mild headache pain within the first few hours. Thereafter, while the cerebral arteries remain dilated, headache pain increases due to trigeminal sensitization, which uncouples the original stimulus (vasodilation) from the reported pain. Figure 20.2Somatotopical selectivity of the trigeminal nerve innervation to the site of head pain. (A) The dura and the meningeal arteries are shown on the left and the cerebral and tributary veins are shown on the right. Area of pain following stimulation of (1) the margin of the superior sagittal sinus; (2) the middle meningeal artery; and (3) the tributary veins to the superior sagittal sinus. (B) The dura and venous sinuses of the posterior fossa. Area of pain following stimulation of (1) the superior sagittal sinus; (2) the sigmoid sinus; (3) the transverse sinus; (4) the upper part of the occipital sinus; (5) the lower part of the occipital sinus; and (6) the dura of the floor of the posterior fossa. ⚪ Indicates the point of stimulation without pain; ⦁ indicates the point of stimulation causing pain. (Source: Ray and Wolff 1940.) Trigeminovascular input unites on the trigeminal ganglion and subsequently converges on second-order neurons within the brain stem (trigeminal cervical complex) (Goadsby et al. 2009). Within this trigeminal cervical complex, several trigeminal neurons synapse onto the superior salivatory nucleus to form a trigeminal-parasympathetic reflex, responsible for releasing vasodilator molecules into the central nervous system (Bergerot et al. 2006). The fact that autonomic symptoms are associated with migraine attacks and ipsilateral vasodilation corresponding to the site of headache pain (Iversen et al. 1990) highlights the involvement of the parasympathetic system (Gupta and Bhatia 2007). However, whether cerebral vasodilation is the inciting cause of trigeminal afferent feedback or is a result of reflex parasympathetic outflow or other vasodilator pathways is unknown. Although subject to debate, activation of the terminal vascular system via the dilation of extra- and intracranial arteries has been proposed as the source of headache pain in what is referred to as the “vascular hypothesis.” Infusion of various substances, including nitroglycerin, glyceryl trinitrate, calcitonin gene-related peptide, and pituitary adenylate cyclase-activating polypeptide-38 (PACAP38), in healthy individuals and those who suffer from migraines produces cerebral vasodilation and a reproducible acute headache with both an immediate and delayed response (Amin et al. 2012; Ashina et al. 2000; Christiansen et al. 2008; Lassen et al. 2002). Alternative explanations for trigeminal activation, especially in migraine, include sterile neurogenic inflammation of the dura mater (Markowitz et al. 1987), or feedforward activation from brainstem sites such as the periaqueductal gray matter and dorsolateral pons (brainstem generator) due to dysregulation of sensory processing (Goadsby et al. 2017). A further important feature of cerebral pain processing is the concept of peripheral (Olesen et al. 2009) and central (Burstein 2001) sensitization of trigeminal nociceptive afferents and central trigeminovascular neurons. An enhanced sensitivity and widening of the spatial nociceptive field may help explain the protracted nature of head pain, exacerbation of pain by previously nonpainful factors such as standing and exercise, hypersensitivity to stimuli such as light and sound, and related neurological symptoms including nausea and vomiting, fatigue and altered sleep, and cognitive function. Moreover, the temporal nature of nociceptive activation is important in the processing of pain. For example, whereas extended dilation and stimulation may cause pain, short duration stimulation of small and large arteries may have little effect. A variety of signaling pathways interact with each other and affect the perception of headache. Interactions between the trigeminal system, ventroposterior thalamus and cortex provide information regarding pain localization and motor defense, whereas the ventral trigeminal, periaqueductal gray, and hypothalamus modulate the antinociceptive and autonomic components of pain. Furthermore, other areas such as the amygdala, hippocampus, anterior thalamus, basal ganglia, and cerebellum are involved in the emotional, arousal, attentional and motor preparation of pain sensation (Sanchez del Rio and Alvarez Linera 2004). These multiple intersecting pathways highlight that the perception of pain is a highly complex process involving multiple areas of the brain, which may explain why it has proven difficult to identify a single factor responsible for the development of headache in AMS. The previous discussion assumes that, similar to the situation with migraine, the integrative neurological mechanisms responsible for headache are also responsible for the other symptoms seen in AMS: fatigue, nausea, vomiting, and dizziness/weakness. However, independent mechanisms have also been proposed for these symptoms including hypoxia-mediated dysfunction of central neurotransmitters (Hackett 1999) and ischemia and/or hypoxia of vomiting centers located in the brain stem around the fourth ventricle (Jones 2002). A long-standing concept in AMS pathophysiology is that symptom onset is related to the development of raised intracranial pressure (ICP). The tight-fit hypothesis, first proposed by Ross in 1985 suggests that individuals with a higher ratio of CSF to brain and blood volume have a greater ability to compensate for cerebral edema or increased blood volume by displacing CSF from the intracranial compartment and, as a result, are protected against the development of AMS (Ross 1985). The hypothesis has not been tested extensively but several studies provide indirect support. Wilson and Milledge (2008) analyzed data from Brian Cummins’s 1985 Kishtwar expedition and found an inverse correlation between ventricular size and headache scores, although this was based on analysis of only a small number of subjects. Wilson et al. (2013) also found that ventricle volumes and the pericerebellar CSF volume were predictive of the intensity of high altitude headache. Given that ICP is elevated in persons with high altitude cerebral edema (Houston and Dickinson 1975; Singh et al. 1969; Wilson 1973) as well as those with severe AMS (Singh et al. 1969), it is clear that under certain circumstances, hypoxia causes pathological events that elevate ICP. The questions that remain, however, are whether ICP is elevated under hypoxic conditions associated with mild to moderate AMS and whether a rise in ICP causes AMS symptoms. Several studies provide data on ICP changes in acute hypoxia. Schaltenbrand (1933) performed continuous measurements of ICP by lumbar puncture during acute reductions in atmospheric pressure inside a hypobaric chamber. Although the threshold altitude varied between patients, CSF pressure rose consistently, especially above 3000 m. Importantly, administration of oxygen either aborted the rise in ICP or restored it to normal, in most cases making it unlikely that the change in ICP was due to hypobaria rather than hypoxia. Hartig and Hacket (1992) measured ICP via lumbar catheter in three subjects during acute hypoxic gas exposure (FIO2 0.11 for 10 min) and during a six-hour hypobaric chamber exposure (5000 m) and found increased ICP in two of the three participants during both exposures but no relationship between the changes in ICP and incidence of headache. Bailey et al. performed lumbar punctures before and after a 16-hour exposure to normobaric hypoxia (FIO2 0.12) in subjects allowed to take acetaminophen to reduce headache symptom severity and found no difference in mean ICP (Bailey et al. 2006). In a study referenced earlier, Brian Cummins implanted invasive telemetric monitoring devices in three individuals and found ICP was normal in one individual and rose by 7 mmHg and 5 mmHg in the other two when measured in the supine position at 5030 m (Wilson and Milledge 2008). The significance of these findings is unclear. While several of the studies show that ICP increases in acute hypoxia, the increases in ICP are mild, do not coincide with symptoms, and resolve over time at the same altitude even though symptoms persist. An important limitation of many of the studies noted above, however, is that they utilize spot measurements of ICP or continuous data reported as mean values, an approach that does not reflect the dynamic nature of ICP (Torbey et al. 2004) and its implication for AMS pathophysiology. An exception is the study by Hartig and Hackett (1992), which used continuous direct recordings of CSF pressure and demonstrated a remarkable threefold increase in CSF pressure from 10 mmHg to 30 mmHg in phase with the nadir of the oscillating oxygen saturation during periodic breathing. Further evidence along such lines can be found in unpublished case studies by Lawley and colleagues of two individuals with an Ommaya reservoir (a catheter placed from the lateral ventricle to a reservoir under the scalp for prophylactic delivery of chemotherapy for treatment of a hematologic malignancy), in which ICP was measured continuously in normoxia and over 24 hours in a hypobaric chamber at 4500 m. One individual demonstrated large fluctuations in ICP in phase with the oscillation of oxygen saturation within one hour of hypobaric hypoxia that continued for 8 hours and was aborted with supplemental oxygen (FIO2 of 1.0) before the individual had to be removed from the chamber. The second individual had normal ICP and no symptoms while awake, but significant increases in ICP with sleep and changes in position during sleep that were accompanied by severe symptoms upon awakening which resolved through the day (Figure 20.3). This latter case suggested that the position of the head and neck during sleep may have affected cerebral venous drainage and, as a result, ICP, while the two cases viewed together suggest that ICP changes may follow varying time courses between individuals. Figure 20.3Changes in intracranial pressure over 24 hours in a single individual exposed to hypobaric hypoxia. Continuous measurements of intracranial pressure, blood pressure, and oxygen saturation over 24 hours at high altitude. Intracranial pressure is normal after eight hours at altitude, but when the individual fell asleep, intracranial pressure increased substantially (around 10 mmHg) and even more so (around 19 mmHg) when sleeping on their left side. The individual woke up with a moderate frontal temporal headache. After 24 hours, the participant began to hyperventilate, causing an increase in oxygen saturation and a decrease in intracranial pressure below normal, leading to relief of headache pain. Figure courtesy of Justin Lawley. (Derived from unpublished data.) Increased ICP may be due to volume changes within the craniospinal compartment. As cells within the brain or spinal cord are unlikely to change their volume per se, changes in blood volume, water volume (edema), and cerebral spinal fluid volume warrant consideration. MRI studies suggest that all individuals experience a generalized increase in brain volume, especially within the gray matter, with acute hypoxic exposures ranging from 20 minutes to 22 hours (Bailey et al. 2006; Dubowitz et al. 2009; Lawley et al. 2014; Sagoo et al. 2017). Under these conditions, the increase in brain water (edema) appears minor, but fluid shifts interpreted from changes in diffusion indices within the white matter are observed quite consistently (Kallenberg et al. 2007; Lawley et al. 2013; Schoonman et al. 2008), suggesting the increase in brain volume is likely due to arterial and venous dilation and additional blood volume. CSF volume decreases by a similar amount in order to maintain a constant intracranial volume, albeit with the likely consequence of reduced intracranial compliance and the development of ICP waves. However, it is unlikely that the observed changes in intracranial compartment volumes are responsible for a static elevation in ICP, as ample CSF volume remains within the hypoxic brain. An alternative explanation may lie within the cerebral venous circulation. In the face of increased cerebral inflow, individuals with restricted venous outflow may be at risk for sagittal sinus hypertension and thus raised ICP (Wilson and Imray 2016). Moreover, the combined or independent increase in arterial blood volume may also compress venous outflow pathways (Lawley et al. 2014), causing venous hypertension. The exact mechanism(s) linking changes in ICP to AMS are unknown and, as such, not established; however, continuous or transient compression of trigeminal-sensitive structures may play a role, as headache, nausea/vomiting, dizziness, and fatigue are common complaints in clinical patients with even mildly raised ICP. A variety of other factors have been considered in the development of AMS. Given that hypoxia is a potent cerebral vasodilator (Severinghaus et al. 1966), differences in the cerebral blood flow (CBF) response may account for the development of high altitude headache and AMS. Using transcranial Doppler ultrasound, a number of investigators have shown greater increases in the middle cerebral artery velocity in subjects with AMS compared to asymptomatic controls (Baumgartner et al. 1994; Jansen et al. 1999), although other studies have not confirmed this finding (Ainslie and Subudhi 2014). Similarly, a variety of studies using transcranial color-coded Doppler ultrasound or magnetic resonance imaging have demonstrated dilation of both intracranial and extracranial arteries (Arngrim et al. 2016; Imray et al. 2014; Lawley et al. 2014) (Figure 20.4), and intracranial venous structures (Lawley et al. 2014; Sagoo et al. 2017; Wilson et al. 2013) in response to acute hypoxia, which, based on classic experiments by Ray and Wolff (1940), could be responsible for the common frontal headache experienced in hypoxia. It is important to note that while these data provide insight into vascular dilation and changes in the intracranial blood volume in acute hypoxia, they do not establish a causal link between specific extracranial or intracranial cerebral arterial dilation and AMS nor do they identify the specific vascular culprits in development of AMS. Further work along these lines is necessary to establish whether the vascular hypothesis adequately explains the development of headache at high altitude. Figure 20.4Widespread dilation of intracranial and extracranial arteries after 190 minutes of normobaric hypoxia (SpO2 70–75%.) Left panel. Magnetic resonance 3D time-of-flight angiography showing location of each arterial segment marked in red with the inclusion of four intra- and four extracranial arteries. Middle cerebral artery (MCA); cerebral part of the internal carotid artery (ICAcerebral); basilar artery (BA); cavernous part of the internal carotid artery (ICAcavernous); middle meningeal artery (MMA); external carotid artery (ECA); superficial temporal artery; cervical part of the internal carotid artery (ICAcervical); superficial temporal artery (STA). Right panel. Circumference change of intra- and extracranial arteries. Percent changes (95% CI) from baseline in mean circumference of intracranial (A) and extracranial (B) arteries. *P-values < 0.05 for comparison of absolute changes during hypoxia compared to sham in patients and between patients and controls during hypoxia. Note that widespread vasodilation occurs in controls and migraine patients, although the patient group experienced severe headache pain. (Source: Arngrim et al. 2016.) Through the mechanism of cerebrovascular autoregulation, constant CBF is maintained over a range of mean arterial pressures. Evidence suggests that impaired autoregulation following ascent may contribute to the development of AMS, as several studies have demonstrated that a low cerebral autoregulation index, a marker of impaired autoregulation was associated with lower SpO2 and higher AMS scores following exposure to acute hypoxia (Bailey et al. 2009b; Van Osta et al. 2005), although other studies have not found such relationship (Subudhi et al. 2011; Subudhi et al. 2014; Subudhi et al. 2015). It has been hypothesized that impaired cerebral autoregulation contributes to AMS via breakdown of the blood-brain barrier and edema formation. Alternatively, it is possible that impaired buffering of arterial blood pressure (involving a greater dynamic stretch of intracranial arteries) may have more subtle effects on the progression or severity of high altitude headache and/or AMS. Whether impaired autoregulation contributes to development of AMS or is merely an epiphenomenon of hypoxic stress is not clear. In the studies noted earlier (Bailey et al. 2009b; Van Osta et al. 2005), individuals with AMS were generally more hypoxemic. Given that there is a graded reduction in cerebral autoregulation with progressive hypoxia (Horiuchi et al. 2016b), it is possible that impaired autoregulation is an epiphenomenon of hypoxic stress rather than a factor in the causal pathway of AMS. It has been hypothesized that systemic inflammatory mediators may cross into the brain via a leaky blood-brain barrier or at the circumventricular organs and directly activate the trigeminovascular system to cause pain or sensitize the system such that afferent feedback occurs at a lower stimulus threshold. To date, however, the evidence supporting such a link has been limited. Although several studies have reported elevated eicosanoids and cytokine concentrations paralleling the development of AMS (Lundeberg et al. 2018; Richalet et al. 1991), other studies have failed to show increases in inflammatory markers with exposure to high altitude (Swenson et al. 1997) or a link between isolated inflammatory pathways and AMS (Grissom et al. 2005; Luks et al. 2007; Muza et al. 2004). More recently, Julian et al. (2011) found that resistance to AMS was associated with downregulation of the inflammatory and permeability responses, whereas AMS-susceptible individuals showed no evidence of an exaggerated inflammatory response. Although some studies have suggested a role for nonsteroidal anti-inflammatory agents in AMS prevention, and thereby a potential link between systemic inflammation and AMS (Burns et al. 2018; Gertsch et al. 2012; Lipman et al. 2012), they failed to include an analysis of systemic inflammatory mediators and how these might have been affected by the medications. One other challenge with much of the work on inflammatory pathways is that most investigations and hypotheses have considered systemic inflammation rather than looking specifically at inflammation in the central nervous system where AMS symptoms originate. Indeed, nonsteroidal anti-inflammatory medications effectively cross the blood-brain barrier (Parepally et al. 2006) and may therefore exert a neuroprotective effect independent of the systemic circulation. It has also been proposed that changes in vascular endothelial growth factor (VEGF) expression may alter blood-brain barrier integrity and contribute to development of AMS. Studies in mice, for example, have demonstrated a hypoxia-induced increase in VEGF and evidence of cerebral edema that was prevented by VEGF inhibition (Schoch et al. 2002). Data from humans have been mixed, however, as some have demonstrated higher levels of free plasma VEGF in those with AMS at 4300 m (Tissot van Patot et al. 2005), while others found not relationship between AMS and VEGF concentrations in cerebrospinal fluid or blood (Bailey et al. 2006; Maloney et al. 2000; Walter et al. 2001). In the end, given that MRI studies have failed to demonstrate frank cerebral edema in AMS (Kallenberg et al. 2007; Lawley et al. 2013; Sagoo et al. 2017; Schoonman et al. 2008), the involvement of VEGF and other factors affecting blood-brain barrier permeability is likely limited to, at best, subtle disruption of the blood-brain barrier. Considerable attention has also been focused on the role of reactive oxygen species (ROS). Several studies have shown, for example, that with acute normobaric hypoxia, subjects with AMS typically had greater increases in free radical production or lipid peroxidation than AMS-negative individuals (Bailey et al. 2009a; Bailey et al. 2009b). Analysis of this finding is limited by the same problem discussed regarding impaired autoregulation. Because the individuals with AMS were also more hypoxemic, it is possible that the higher levels of ROS were a marker of hypoxic stress rather than a factor in the causal pathway for AMS. In fact, in other studies that control for the degree of hypoxemia, free radical formation is similar in all individuals (Bailey et al. 2006; Bailey et al. 2011). Another argument against a causal role for ROS is the fact that attempts to pharmacologically attenuate the rise in ROS in human volunteers have not been shown to prevent AMS (Bailey and Davies 2001; Baillie et al. 2009). Subjects with AMS may develop a state of expanded plasma and extracellular fluid volume similar to that seen in subjects starting day-long exercise at low altitude. Evidence of such fluid retention is provided by the clinical observation of lower urine output in soldiers with AMS compared to those without symptoms (Singh et al. 1969) and the finding that trekkers with AMS gained weight, while trekkers without AMS lost weight by the time they reached 4243 m (Hackett et al. 1982). More recently, Loeppky et al. (2005) exposed healthy men and women to hypobaric hypoxia (426 mmHg, ∼4880 m) for 8–12 hours and noted significant fluid retention among the 16 individuals with the highest AMS scores, while Gatterer et al. (2019) noted an association between fluid retention following arrival at 3480 m and the Lake Louise AMS score. The mechanism for the observed antidiuresis remains unclear. In the study noted above, Loeppky et al. (2005) found that antidiuretic hormone levels fell in those who remained free of AMS but increased and continued to rise in those with AMS, and closely correlated with symptom severity and fluid retention. Other studies have implicated alterations in the renin-aldosterone axis, but the evidence has been inconsistent, with some studies finding that AMS scores correlate with aldosterone levels (Bartsch et al. 1988; Milledge et al. 1989) and others showing no such relationship (Hogan et al. 1973; Loeppky et al. 2005). Similar variability has been seen in studies of atrial natriuretic peptide (Bartsch et al. 1988; Milledge et al. 1989). In the study by Gatterer et al. (2019) noted previously, preascent saliva cortisol levels were associated with fluid retention following ascent, suggesting that alterations in cortisol homeostasis may affect fluid balance and, therefore, AMS risk, but a more precise link between cortisol metabolism and fluid balance was not elucidated in this study. As with many of the other factors discussed above, the causal link between the alterations in fluid balance and development of AMS is not clear. Hypervolemia could potentially lead to vascular distention and, as a result, activation of the trigeminovascular system, but it remains possible that alterations in fluid balance are a marker of impaired acclimatization rather than playing a part in AMS pathophysiology. There are several challenges in interpreting the extensive literature on the pathophysiology of AMS. First, as noted above, many of the studies demonstrating a relationship between AMS and a particular factor, such as impaired cerebral autoregulation, do not establish whether this relationship is, in fact, causal. In fact, the variables in question may be epiphenomenon of acute hypoxic exposure and a consequence of impaired acclimatization rather than being a cause of AMS. Second, the timing at which AMS is assessed in a given study design makes a big difference in observed outcomes. This is best demonstrated by the second case of unpublished data on ICP changes in acute hypoxia from Lawley and colleagues noted previously. If the assessment for AMS was made after only 12 hours in hypoxia, this individual would have been AMS-negative, whereas they would have been AMS-positive with an assessment at 16 hours and AMS-negative again at 24 hours. To the extent that different studies assess for the presence of AMS at varying time points, it is hard to sort through discrepant results and tease out a true relationship and a causal pathway. Finally, there are significant challenges in assessing headache, the cardinal symptom of AMS and likely key to understanding AMS pathophysiology. Headache is a complex, multidimensional subjective experience typically described by clinical interviews or questionnaires and is difficult to measure in an objective manner. While assessments of headache can be performed on single or multiple occasions, it may be more representative to characterize headache burden as a composite score of multiple measurements of pain intensity over time (Wilson and Imray 2016). Doing so may permit more accurate and reliable identification of headache burden and diminish the impact of individuals with transient headaches on study outcomes but may preclude the identification of any time dependent stimulus response relationships or the onset of trigeminal sensitization. While various treatment strategies are available for those with AMS, the best approach is to prevent its development in the first place. In the majority of cases, AMS can be prevented through nonpharmacologic measures and, in particular, slow ascent, while in other cases, pharmacologic prophylaxis is warranted. In considering these measures, it is important to remember that physiologic responses to high altitude and susceptibility to altitude illness vary significantly between individuals. As a result, a strategy that works for one individual may not be suitable for another. The primary goal of nonpharmacologic measures is to facilitate acclimatization to hypoxia during or, in some cases, ahead of the planned ascent. Given that overly rapid ascent is one of the primary risk factors for developing acute altitude illness (Hackett et al. 1976), the best means to prevent AMS, as well as HACE and HAPE, is to slow the rate of ascent. Rather than focusing on the speed at which someone walks or climbs, the key issue is the rate at which they increase their sleeping elevation. Consensus guidelines from the Wilderness Medical Society (Luks et al. 2019), for example, recommend that once an individual travels above 3000 m, they should not increase the sleeping elevation by more than 500 m day−1 and should include a rest day every three to four days, during which they sleep at the same elevation for another night. This is very similar to the approach recommended in multiple leading reviews on the topic (Bartsch and Swenson 2013; Basnyat and Murdoch 2003; Hackett and Roach 2001), albeit with some slight differences in the specified altitude changes. Despite the widespread recommendations for this practice, there are little in the way of data to support it. The recommendations likely originated following an epidemiological study of trekkers on the route to Everest base camp (Hackett et al. 1976) in which trekkers who walked from Jiri rather than flying to Lukla or who spent an extra night in Pheriche had a lower rate of AMS than trekkers who followed faster ascent profiles. While subsequent survey data (Basnyat et al. 1999) provided further evidence in support of this approach, there has been only one randomized, controlled study on the effect of ascent rate on AMS incidence. Bloch et al. (2009) randomized climbers on Muztagh Ata (7546 m) to one of two ascent profiles and found that those individuals randomized to a slower ascent had lower symptom severity, incidence of AMS and greater chance of summit success than those randomized to the faster profile. In one of the few other prospective studies, Purkayastha et al. (1995) compared the incidence of AMS in Indian soldiers traveling to 3500 m and found higher rates in those who traveled by plane (84%) than those who traveled by truck (51%). While slow ascent is clearly important, it is necessary to remember that there is significant interindividual variability in susceptibility and, as a result, people who adhere to these principles may still get sick while some who ascend faster than recommended do not. This variability was demonstrated nicely in a study by Murdoch (1999) in which he surveyed 283 trekkers in the Everest region of Nepal asking about AMS symptoms and speed of ascent. Half the trekkers ascending at the very low mean rate of 100–200 m day−1 became sick while almost half the trekkers ascending at 500–600 m day−1 remained free of AMS. The challenge for the first-time high altitude traveler is that they do not know where they fall within that spectrum of susceptibility. It is these individuals that should adhere to the published guidelines and perhaps even ascend at slower than recommended rates. As individuals make repeated trips to high altitude, they will learn their personal tolerances and can adapt the general rules as necessary. One of the challenges with applying the guideline recommendations is the fact that on many major climbing routes or trekking circuits, the established sleeping locations are not necessarily laid in at 300–500 m intervals and logistical factors often mandate faster ascents. When overly large gains in elevation are necessary for such reasons, rest days should be used before and after such large gains to maintain an adequately slow rate as averaged over the entire trip (Luks 2012). Recent interest has focused on other methods to facilitate acclimatization, decrease the risk of altitude illness and, perhaps, even allow faster than recommended ascent rates. In staged ascent, individuals spend a number of days at a moderate elevation before ascending to the target elevation. Beidleman et al. (2009), for example, demonstrated that six days of staging at 2200 m decreased the incidence and severity of AMS following ascent to terrestrial 4300 m when compared to a rapid ascent to a simulated altitude of 4300 m in a hypobaric chamber. More recent data suggest even two days of staging may be sufficient for this purpose (Beidleman et al. 2018). Preacclimatization involves repeated, time-limited exposures to hypoxia in the days to weeks prior to the planned ascent through either actual ascents to high altitude or chamber-based exposures to either normobaric or hypobaric hypoxia. The data on this approach have been mixed with some studies showing benefit in terms of preventing AMS (Beidleman et al. 2004; Wille et al. 2012) and others showing no effect at all (Faulhaber et al. 2016; Fulco et al. 2011; Schommer et al. 2010). One of the challenges of interpreting the literature on this topic is the marked variability between study protocols with regard to the dose, frequency and duration of the hypoxic exposures and their proximity to the eventual ascent to high altitude. For example, Wille et al. (2012) exposed subjects to seven one-hour sessions of normobaric hypoxia equivalent to 4500 m then performed assessments at 5300 m two days following cessation of the protocol, while Schommer et al. had subjects perform 70 minutes of normobaric hypoxic exercise three times per week for three weeks, with the degree of exposure increasing from 2500 m to 3500 m over the three-week period, before performing assessments at 4559 m five days after the last training session. Another issue is that some of the studies fail to provide objective evidence of physiologic changes in the subjects that could theoretically assist acclimatization and prevent problems during a subsequent ascent (Fulco et al. 2013). Given these issues, it is hard to recommend a specific approach that works best. From a general standpoint, longer, more frequent, and greater degrees of hypoxic exposure, as well as closer proximity to the ultimate ascent, will likely be of greater utility, but the precise recipe for doing this remains unknown. Another concern with staging and preacclimatization is whether individuals actually have the time to engage in these methods of preparation. One alternative to these time-intensive strategies is to “bring the mountains to the home” or other places where people spend a lot of time as part of their daily lives. In particular, systems are available for erecting hypoxic tents around the bed in which someone sleeps or delivering hypoxic gas mixtures through a tight-fitting mask while engaging in exercise on a stationary bicycle or treadmill. Although multiple systems are on the market and anecdotal reports point to their increasing use, the data supporting this practice are very limited. In the lone study to date, Dehnert et al. (2014) randomized 76 healthy individuals to sleep in tents while breathing either air or an FIO2 of 0.14–0.15 for 14 consecutive nights and then assessed the incidence of AMS four days later during a 20-hour exposure to a simulated altitude of 4500 m. Due to technical problems with the nitrogen generators in the tent systems, only 21 of the 37 subjects assigned to the hypoxia group slept in conditions sufficient for acclimatization (>2200 m). In data analysis limited to this subgroup, sleeping in normobaric hypobaric hypoxia decreased symptoms and incidence of AMS, but the logistical problems highlight a potential problem with these systems: If the intended degree of hypoxia cannot be adequately maintained in a controlled research setting, how effectively can it be applied when used in everyday practice? Other strategies, such as avoiding alcohol or caffeine, are often recommended but lack supporting evidence. In the case of caffeine, which, among other things, acts as an adenosine inhibitor with various influences on the central nervous system, abrupt cessation of intake may even provoke withdrawal symptoms that resemble AMS (Hackett 2010). Another common-sense recommendation with good physiologic rationale but little supporting evidence is to avoid opiate pain medications, particularly at night, as the respiratory depressant effects could worsen oxygenation and possibly provoke AMS. Patrician et al. (2019) reported that use of an expiratory resistance device with a small amount of deadspace during sleep in hypoxia reduced the apnea-hypopnea index by 50% and improved headache scores without an effect on overall Lake Louise AMS scores, while Johnson et al. (2010) showed that use of noninvasive positive pressure ventilation (NIPPV) during sleep improves nocturnal oxygen saturation and decreases AMS symptoms. The expiratory resistance mask may have some utility in the field environment given that it is lightweight, portable, and does not require a power source, whereas NIPPV or continuous positive airway pressure devices are likely to benefit those individuals sleeping in a lodge or other high resource setting where weight, bulk, and access to electrical power are less of an issue. Several other nonpharmacologic strategies are often suggested as having benefit but have proven to be ineffective. Because the balance of carbohydrates, protein, and fat intake affects the production of carbon dioxide and, as a result, ventilation, questions have been raised as to whether particular dietary regimens decrease the risk of AMS. Lawless et al. (1999) did find that ingestion of an oral carbohydrate beverage 2.5 hours into exposure to a simulated altitude of 4600 m caused small but statistically significant increases in PaO2 and oxygen saturation but found no effect on development of AMS. In a chamber study of 19 subjects given either a high (68%) or normal (45%) carbohydrate diet for four days prior to eight hours of exposure to 10% normobaric oxygen, Swenson et al. (1997) found no difference in the AMS scores between individuals taking the two diets. Given the role nitric oxide plays in physiologic responses to hypoxia, questions have been raised recently about whether dietary nitrate supplementation in the form of beetroot juice affects various physiologic parameters and susceptibility to AMS. Multiple studies using a variety of protocols, however, have found no significant effect on various markers of oxygenation or hemodynamic status and no prophylactic effect against AMS (Cumpstey et al. 2017; Hennis et al. 2016; Masschelein et al. 2012; Rossetti et al. 2017b). Trekkers and climbers are often urged to drink plenty of fluids to prevent AMS, but there is little sound evidence to support this recommendation. Basnyat et al. (1999) surveyed trekkers at Pheriche (4243 m) and found that individuals with higher fluid intake (up to 5 L day−1 in some cases) had a lower incidence of AMS (odds ratio, 1.54). This study is limited, however, by the retrospective nature of the data collection, the likelihood of recall bias and difficulty establishing causality, as only those individuals who were feeling well may have been able to achieve higher fluid intake. Other, more systematic studies have not found any support for a relationship between hydration status and AMS incidence or severity. Aoki and Robinson (1971), for example, found no differences in AMS scores during two days of exposure to a simulated altitude of 4270 m between individuals made hypovolemic with furosemide diuresis and those receiving placebo or vasopressin to maintain euvolemia. More recently, Castellani et al. (2010) showed hypohydration degraded aerobic performance at simulated high altitude, but had no effect on development of AMS. Richardson et al. (2009) did show that dehydration was associated with an increase in AMS scores with exposure to hypobaric hypoxia but also noted increased symptoms in the setting of hyperhydration. The question remains as to why the recommendation about maintaining hydration persists despite the lack of supporting evidence. What likely accounts for the perception of benefit is the fact that individuals maintaining adequate volumes of fluid intake are preventing dehydration, a problem for which individuals are at higher risk at high altitude due to the lower humidity and whose symptoms can mimic those of AMS. While preventing dehydration is an important end in and of itself, too much fluid intake may also be problematic. In addition to the study by Richardson et al. (2009), noted previously, Gatterer et al. (2013) showed that AMS was more common in those with a positive net water balance during a 12-hour exposure to a simulated altitude of 4500 m with ad libitum fluid intake. For most individuals, adequately slow ascent is sufficient to prevent AMS. In many circumstances, however, pharmacologic measures may be necessary to further decrease the risk. The decision to initiate pharmacologic prophylaxis should be based on the patient’s prior history at high altitude as well as several features of their planned ascent such as the altitude gain on the first day, the average increase in sleeping elevation over the course of the trip, and the number of rest days (Table 20.4) A fuller description of this risk assessment is provided in published guidelines (Luks et al. 2019). Individuals deemed to have low risk ascent profiles can forego pharmacologic prophylaxis while those with moderate to high risk profiles should strongly consider use of medications. Prior history of acute altitude illness Planned sleeping elevation on first night at high altitude Rate of ascent once above 2500 m Prior history of HACE or HAPE Number of planned rest days Use of preacclimatization strategies
Introduction
The Scope of the Problem
Risk Factors for Acute Mountain Sickness
Genetic predisposition
Exercise capacity
Physical activity following arrival
Age and sex
Body habitus
Smoking
History of migraine headaches
Ventilatory responses to hypoxia
Abnormal gas exchange
Clinical Features of AMS
Altitude of onset
Time course
Symptoms and signs
Diagnosing AMS
Clinical setting
Research setting
How Well Do the Systems Compare to Each Other
What is the Appropriate Scoring Threshold to Diagnose AMS?
What is the Appropriate Timing for Assessment for AMS?
Language Issues
Use in Children
Predicting Who Will Develop AMS
Pathophysiology of AMS
Pathophysiology of headache
Initiation of Head Pain
Parasympathetic Reflex
Potential Mechanisms for Trigeminal Activation
Central Modulation and Perception of Headache
Relationship between Other Symptoms of AMS
Tight-fit hypothesis and increased intracranial pressure
Evidence for Increased Intracranial Pressure
Mechanisms for Changes in Intracranial Pressure in Hypoxia
The Link between Increased ICP and AMS
Other potential contributors to development of AMS
Changes in Cerebral Blood Flow
Autoregulation and Dynamic Pulsatility of Cerebral Vasculature
Inflammatory Mediators
Vascular Endothelial Growth Factor
Reactive Oxygen Species
Alterations in Fluid Balance
Challenges in understanding the pathophysiology of AMS
Prevention of AMS
Nonpharmacologic measures
Rate of Ascent
Staged Ascent and Preacclimatization
Hypoxic Tents
Other Measures That May Be Effective
Measures Shown Not to Be of Benefit
Dietary intake
Preventing dehydration
Pharmacologic prophylaxis
Who Warrants Prophylaxis?
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
