Abstract
Background
Given the high prevalence and clinical impact of high-altitude headache (HAH), a better understanding of risk factors and headache characteristics may give new insights into the understanding of hypoxia being a trigger for HAH or even migraine attacks.
Methods
In this prospective trial, we simulated high altitude (4500 m) by controlled normobaric hypoxia (FiO2 = 12.6%) to investigate acute mountain sickness (AMS) and headache characteristics. Clinical symptoms of AMS according to the Lake Louise Scoring system (LLS) were recorded before and after six and 12 hours in hypoxia. O2 saturation was measured using pulse oximetry at the respective time points. History of primary headache, especially episodic or chronic migraine, was a strict exclusion criterion.
Findings
In total 77 volunteers (43 (55.8%) males, 34 (44.2%) females) were enrolled in this study. Sixty-three (81.18%) and 40 (71.4%) participants developed headache at six or 12 hours, respectively, with height and SpO2 being significantly different between headache groups at six hours (p < 0.05). Binary logistic regression model revealed a significant association of SpO2 and headache development (p < 0.05, OR 1.123, 95% CI 1.001–1.259). In a subgroup of participants, with history of migraine being a strict exclusion criterion, hypoxia triggered migraine-like headache according to the International Classification of Headache Disorders (ICHD-3 beta) in n = 5 (8%) or n = 6 (15%), at six and 12 hours, respectively.
Interpretation
Normobaric hypoxia is a trigger for HAH and migraine-like headache attacks even in healthy volunteers without any history of migraine. Our study confirms the pivotal role of hypoxia in the development of AMS and beyond that suggests hypoxia may be involved in migraine pathophysiology.
Introduction
With increasing numbers of people who experience high altitude comes a demand for an awareness of the presentations and knowledge of the syndromes including treatment options for altitude-related illness. High-altitude headache (HAH) is the most prominent symptom of acute mountain sickness (AMS) and has been defined by the International Headache Society as headache, usually bilateral, aggravated by exertion and caused by ascent above 2500 meters (1,2). Nevertheless, the underlying pathophysiological mechanisms and even more the potential risk factors are still poorly understood. Current pathophysiological theories include the elevation of intracranial pressure, alteration of brain oxygen consumption, impaired glucose metabolism, free radical production and increased cerebral blood flow. Each of these factors is discussed as a proposed mechanism underlying HAH (3–8).
Most data on the clinical presentations, features and predictors of HAH derive from observational studies in the field (9–11). The interpretation of the pathophysiological processes behind HAH is limited because of the complexity of confounding factors such as great exertion, inadequate fluid and food intake, sleep deprivation, environmental factors (i.e. temperature, humidity), comorbidities and history of primary headaches.
Thus, assessment of headache features and other important pathophysiological markers might help to clarify the pathophysiological relevance of hypoxia on HAH development. These markers could include arterial oxygen saturation (SpO2) that should, however, be measured during strictly standardized, passive normobaric hypoxia exposure to avoid confounding factors.
Recent studies have shown that a migraine attack can be triggered or become more pronounced by high altitude; however, only in patients with history of migraine (4,10). It is possible that migraine-like headache could be triggered by profound hypoxia simulating high altitude. However, thus far this has never been investigated prospectively in a large cohort of migraine-naïve individuals. On the other hand, recent papers speculate on hypoxia triggering cortical spreading depression secondarily promoting migraine headache attacks (12,13).
This prospective study investigates the influence of controlled normobaric hypoxia (i.e. simulated high altitude of 4500 m) on the development of AMS and headache characteristics in a large cohort of healthy volunteers. We hypothesized that hypoxia is the major trigger for AMS and HAH but also may induce a migraine-like headache in a sub-proportion, even in volunteers without history of primary headaches.
Materials and methods
Participants
A detailed description of the study protocol has been published recently (14).
In brief, participants were recruited by advertisements on the homepage of the Austrian Alpine Association and by information via the mailing list of the University of Innsbruck. Exclusion criteria included cardiovascular, respiratory, neurological and psychiatric diseases, episodic or chronic migraine and chronic headache (defined as any headache occurring on more than 15 days per month). In addition, smoking, pregnancy, permanent residence at altitudes exceeding 1000 m, an overnight stay at altitudes >2500 m in the previous month, or exposure above 2500 m two weeks prior to the 12-hour hypoxic exposure were also exclusion criteria. At enrollment all individuals had a face-to-face interview with a headache specialist excluding all those with history of migraine.
Participants were instructed to abstain from all anti-inflammatory medication and nutritional supplements for two weeks prior to the exposure and from alcohol starting the day before the experiment. Caffeine was not allowed on the day of the exposure. All participants gave their written informed consent prior to the participation in the study. The study was carried out in conformity with the ethical standards laid down in the 2008 Declaration of Helsinki and was approved by the ethics committee of the Medical University of Innsbruck.
Procedures
Participants were passively exposed at a fraction of inspired oxygen (FiO2) of 12.6% in our hypoxic chamber in Innsbruck at 600 m (corresponding to a simulated altitude hypoxia of 4500 m, partial pressure of inspired oxygen (PiO2) = 83.9 mmHg) for 12 hours. Room temperature and humidity were kept constant at 22–24 ℃ and 23–27%, respectively. Prior to entering the hypoxic chamber all participants were examined, including a medical routine check. A pregnancy test was performed in all women prior to the 12-hour hypoxic exposure. During the simulated hypoxia exposure, the same food (i.e. brown bread, cheese, boiled ham, cucumber, banana, apple, cookies and chocolate) and drinks (water and apple juice) were provided and could be consumed as the patient desired. Most of the time participants were seated but some activities (e.g. standing, walking and stretching) were also performed. Recumbent position or sleeping was not allowed. Measurements, described in detail in the following sections, were performed before, during and/or after the session.
Measurements and instruments
The Lake Louise scoring system (LLS) was used to assess the incidence and severity of AMS (2). It is a self-assessment questionnaire including five symptom complexes (headache; gastrointestinal symptoms like anorexia, nausea or vomiting; fatigue and/or weakness, dizziness and/or light headedness; and difficulty sleeping); scores range from 0 to 3. The participants self-rated their status: 0 for no discomfort, 1 for mild, 2 for moderate and 3 for severe symptoms. The symptom complex “difficulty sleeping” was not taken into account because participants did not stay overnight in the hypoxic chamber. AMS was diagnosed whenever the symptom headache was present in combination with at least one other symptom. A total score of ≥ 3 signified either mild disease or AMS at an early stage, while a score ≥ 4 was used to recognize a more severe disease state (5). The LLS at the point when leaving the chamber was taken to distinguish between AMS+ and AMS–. SpO2 was measured before hypoxia by pulse oximetry (Onyx II 9550, Nonin, Plymouth, MI, USA) (i.e. before entering the high-altitude chamber) and at 6 (i.e. in the high-altitude chamber) and 12 (i.e. in the high-altitude chamber) hours, respectively.
Detailed headache characteristics were evaluated before hypoxia (i.e. before entering the high-altitude chamber) and at six and 12 hours (i.e. in the high-altitude chamber) respectively, using a standardized questionnaire evaluating all migraine and HAH features according to the International Classification of Headache Disorders, third edition beta (ICHD-3 beta) (1). Additionally, we asked for headache characteristics (pulsating, pressing, burning, stabbing, dull or other), neck stiffness, factors improving (fluids, stand up, resting, avoiding light, avoiding noise or other) or worsening headache (movement, exertion, stand up, stretching, coughing, nausea, cold, light, noise or other).
Statistical analysis
This study tested the hypothesis that simulated exposure to hypoxia triggers headache. Metric variables (age, height, SpO2 at six hours, SpO2 at 12 hours, visual analog scale (VAS 0–100) at six hours, VAS 0–100 at 12 hours, onset of headache) were compared using Mann-Whitney U test.
A univariate analysis between headache groups (headache versus non-headache) including demographic (age, height, sex) and clinical parameters (SpO2) was conducted in a first step. A binary logistic regression analysis with dichotomized outcome (i.e. headache yes versus no) at both outcome endpoints (i.e. six hours and 12 hours) was applied to analyze predictors for headache. Only variables with significant difference in the univariate approach (i.e. height and SpO2) were included in this model.
The accuracy of SpO2 levels to differentiate between headache yes versus no was evaluated by the receiver operating characteristic (ROC) analysis.
Data analyses were performed with the use of the SPSS statistical software package (SPSS version 22, IBM) and Prism 6 (GraphPad Inc, 2014). A p value of less than 0.05 (two tailed) was considered to indicate statistical significance. Results are expressed as mean ± standard deviation (SD), 95% confidence interval (CI), odds ratio (OR) or median + interquartile range (IQR).
Results
Distribution of baseline characteristics and arterial oxygen saturation using pulse oximetry (SpO2) at simulated high altitude (4500 m).
Data are presented as median and IQR, statistical differences were compared using Mann-Whitney U test. AMS: acute mountain sickness; LL: Lake Louise scale; IQR: interquartile range; VAS: visual analog scale ranging from 0 to 100.
Clinical features and characteristics of headache at simulated high altitude (4500 m).
Data are presented as absolute numbers (N) and percentage of respective headache groups.
No statistical significant differences in baseline characteristics (sex, age, height, SpO2 at baseline) or under hypoxia (SpO2, headache intensity and onset of headache) could be detected in a univariate model between females with or without use of oral contraceptives.
None of the participants reported an aura symptom
Results of the 21 individuals who terminated early at six hours were included only in the six-hour analysis. Baseline characteristics were not significantly different (age, height) with the exception of sex (p = 0.02). In this subpopulation start of headache symptoms was significantly earlier (138 versus 210 minutes, p = 0.002), headache intensity more severe (45 versus 27, p = 0.01) and SpO2 lower (81 versus 84, p = 0.048) than in those who continued until 12 hours. Only one of those 21 participants fulfilled migraine-like headache criteria.
SpO2
No baseline differences were detected comparing SpO2 at baseline between the headache groups. SpO2 was significantly lower in the headache group at six hours (p < 0.05) (Figure 1). The binary logistic regression model adjusted for height revealed SpO2 as a significant predictor for headache at six hours (OR 1.123, 95% CI 1.001–1.259, p < 0.05). The ROC showed an area under the curve (AUC) of 0.698 (95% CI 0.555–0.840) for SpO2 at six hours.
Distribution of arterial oxygen saturation between headache groups (SpO2).
Discussion
The main results of this prospective study were that simulated high altitude by hypoxia (FiO2 = 12.6%, ≅4500 m) triggers profound headache in more than 80 % of healthy volunteers without a history of primary headaches. Furthermore, we found that SpO2 predicts headache. Additionally, a proportion of headache patients fulfilled migraine attack criteria under induced hypoxia.
HAH is defined as headache, usually bilateral and aggravated by exertion, caused by ascent above 2500 m (1). Approximately 10%–25% of un-acclimatized individuals who ascend to 2500 m will develop HAH (2). With an 80% headache rate in our study, the incidence is surprisingly high, but comparable to results published for altitudes of 4500 m and above (15,16). Interestingly, of the known risk factors for the development of HAH (rapid ascent, young age, exertion, history of altitude illness, migraine and genetic predisposition) only young age, history of altitude illness and rapid ascent are likely to explain the high HAH rate in our study population, considering history of primary headache, especially episodic or chronic migraine were an exclusion criterion (4,10,11). Furthermore, during simulated high altitude, individuals were not exposed to exertion and fluid and food intake was not limited. Keeping this in mind, our headache rate of greater than 80% is even more impressive, underlining the pivotal role of hypoxia as the most potent trigger for HAH.
The cause of high-altitude headache is still elusive. In our prospective study we minimized potential triggers and risk factors to rapid ascent to 4500 m by induced hypoxia (FiO2 of 12.6%) (4). Pulse oximetric measurements of SpO2 were normal in all participants before exposure but were evaluated to be significantly lower in HAH individuals at six hours. Furthermore, in the binary logistic regression model lower SpO2 was associated with headache (p < 0.05). ROC analysis for SpO2 and occurrence of headache revealed an AUC of 0.698. Taken together our results propose a predominant role of hypoxia in the pathophysiology of HAH.
(a) Migraine-like headache features at six hours using modified ICHD-3 beta (1).
ICHD-3 beta: International Classification of Headache Disorders, third edition beta.
Not surprisingly, those 21 participants who terminated early after the six-hour examination suffered from more severe headache with early onset and showed significantly lower SpO2 values at six hours. The exclusion of those participants from the 12-hour analysis may explain the loss of significance regarding SpO2 at this time point. Interestingly, only one person from this subgroup fulfilled migraine-like headache features.
To the best of our knowledge, this is the first report on this observation in a prospective study, possibly adding to our pathophysiological concept of migraine and HAH in a large sample. However, future studies will have to reproduce and further elaborate on this, while also including biomarker sampling and prospective therapeutic intervention.
Limitations
Our study was conducted prospectively but without a control group. However, we do not believe that including a control group would add to the understanding of HAH, since being exposed only to a room without hypoxia would not trigger headache, nor would blinding of volunteers be feasible. Given the young mean age of our study population, we might have included individuals who will develop episodic migraine in future years. Another important limitation is that caffeine intake was not allowed on the days of the experiment. Therefore, we cannot fully rule out that at least a small proportion of patients may have suffered from caffeine-withdrawal headache.
Conclusion
In conclusion, the present findings confirm that hypoxia is substantially involved in the development of AMS. Furthermore our data suggest that hypoxia may also trigger migraine-like headache attacks in healthy volunteers without any history of episodic or chronic migraine. Thus, we believe that elements we provided should be implemented in new official HAH diagnosis criteria.
Furthermore, the role of hypoxia in migraine pathophysiology must be investigated in future trials.
Key findings
The present findings confirm that hypoxia is substantially involved in the development of acute mountain sickness (AMS). Our data suggest that hypoxia may also trigger migraine-like headache attacks in healthy volunteers without any history of episodic or chronic migraine. Furthermore, the role of hypoxia in migraine pathophysiology must be investigated in future trials.
Footnotes
Acknowledgments
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
This work was supported by the Medical University of Innsbruck and Department of Sport Science, University of Innsbruck, Austria, and the Austrian National Bank (OENB) No. 14287.