Simulated airplane headache: a proxy towards identification of underlying mechanisms

This AH-simulated study aimed at identification of mechanisms that might contribute in development of AH under a well-controlled environment. A pressure chamber (Training Chamber Low Pressure High Altitude Six Man Capacity, Model ATV-183, Aero-Test Equipment Co. Inc., Dallas, Texas, USA) was used as a proxy to simulate AH. The pressure chamber has been expanded to accommodate 7 participants instead of 6; hence, the number of recruited participants for each group was limited to 7. Two groups of AH-group and non-AH-group participated in this study. Participants were examined by the researchers and a medical authority in accordance to the AH criteria’s defined by IHS [1], in order to verify an accurate diagnosis of AH. The average age was 24 ± 1.6 years, where 4 of the participants were males and the remaining 10 were females. The study protocol was approved by the local ethics committee of Region Nordjylland (approval number N-20150026) and was conducted at the Center for Flight- and Naval Medicine, Skalstrup Airbase, Denmark. Written informed consent was obtained from all participants prior the study and travel expenses were compensated.

Those who were enrolled in the study were all adults above 18 years either healthy (non-AH-group) or with AH (based on the diagnostic criteria for AH according to IHS’ definition of AH [1]).

Individuals were excluded during screening, if they were pregnant, frequent users of any medications including paracetamol, triptans, having an addictive or previously addictive drug abuse, suffering from epilepsy, asthma, other primary headaches, migraine or any psychological and cardiovascular diseases. The applied criteria were maintained to ensure a high safety level for the participants and a controlled simulation of AH headache if occurred in the pressure chamber.

No participant was allowed in stepping the pressure chamber if a headache occurred in the trial day due to any reason. Participants were also asked not to eat or drink coffee, chili, nicotine, or chocolate and to avoid brushing teeth or using mouthwash prior to the study (see saliva assessments below). Participants were also asked to avoid demanding exercise or energy drinks that might had an influence on the study outcomes.

Participants were asked to rate their pain intensity on a scale of 1–10, where 1–3 was considered as mild headache, 4–7 reflected moderate headache, and 8–10 was considered as severe headache.

Participants were instructed to sit in a pressure chamber, which was located in a military airbase, Skalstrup Airbase, Denmark. Two safety observers from the airbase were present during the entire procedure, ensuring that every steps proceeds safely. The chamber contained 7 seats. Participants seated in the pressure chamber for approximately one hour to experience a simulated airplane flight by altering the pressure and air composition similar to what occurs during a real flight [26]. The pressure switched from the atmospheric pressure observed at ground level, to the pressure value present at an altitude of 2438 m, and returned to the ground level when the time was over. The ascending and descending speed were set to 457.2 m/min ≈ 0.055 bar/min. The ascending and descending periods lasted for 5–6 min each, while the cruising phase lasted for almost 47 min, in an altitude of 2438 m to simulate closely what occurs during a one-hour flight in the real-world condition. No service was provided during simulation period and ear protection was provided for all participants to filter the noise. Eyes were kept open and no participant slept during the trial.

A selection of assessments was used at 3 allocated time points: before (pre-simulated flight), during (during simulated flight) and after (post-simulated flight) the exposure to the pressure chamber. Baseline assessments were conducted within 30 min prior to entrance to the pressure chamber. Second assessments were conducted in the middle of the simulated flight, when the participants had been present in the chamber for 30 min, where the maximal altitude and pressure had been achieved. The third and final assessments were conducted within 15 min after participants left the chamber. Saliva samples were collected in each allocated time point for ELISA-analysis of PGE₂ and cortisol. Each sample consisted 1.5 ml natural saliva without any stimulation. The participants delivered the samples by continuously salivating in a small cup, hereafter the saliva samples were transferred to plastic centrifuge tubes by single use plastic pipettes. The total amount of collected saliva for each participant was 4.5 ml. All samples were placed in −80 °C freezer and were kept frozen until analysis.

Blood oxygen, pulse and blood pressure were assessed at the allocated time points. This allowed measurements of the participants’ pulse frequency, systolic and diastolic blood pressure as well as the SPO-concentration in their blood. Blood pressure and pulse were assessed by using a pressure cuff (BP-102 M, Hangzhou Sejoy Electronics & Instruments Co, Hanzhou, China), while the oxygen saturation was assessed by using a finger-based pulse oximeter (CMS50d, Contec Medical Systems Co, Qinhuangdao, China).

A portable compact thermo-camera (FLIR systems E60 thermal imager, FLIR Systems, Wilsonville Oregon, USA) was used to capture thermo-images of participants’ faces to record the average facial skin temperature at each dedicated time point. The thermo-camera was calibrated before use. The sensitivity of the device was <0.05 °C. Temperatures in the waiting room and the pressure chamber were recorded and environmental humidity was also monitored. All images were transported on a USB pen and stored in Excel 2010 (Microsoft Corp., Seattle, WA, USA) for further analysis.

All participants were asked to report the possible occurrence of headache and its characteristics before, during, and after their stay in the pressure chamber. They were all provided with a specific questionnaire for each time point, which aimed at documenting the findings, and ensuring the safety of our participants. The participants were asked to fill out details of their current and recent health situation, and were requested to note any changes that might occur during their stay. While in the chamber, participants were asked to continuously note and rate possible symptoms related to AH, which included pain intensity rating, duration, location of headache, type of headache as well as stress level and general wellbeing. These findings were then collected and compared with the international guidelines for AH in order to ensure matching the criteria set by IHS.

Fourteen participants completed this study. Seven were admitted to the non-AH-group, as they did not have any experience with bouts of AH, while the remaining seven participants were placed in the AH-group, as they had frequently experienced bouts of AH while travelling by flight.

No participants in the non-AH-group developed a headache or any accompanying symptoms, and none were reported stress or anxiety while in the pressure chamber.

All seven participants of the AH-group experienced a headache with AH characteristics in accordance with IHS’ criteria. The majority of AH participants (43%) described the location of their headaches as fronto-orbital. The quality of the headache was described mainly as stabbing (43%). The pain intensity was described as severe (58%) or moderate (29%) (see Table 1).

Between group comparison showed that cortisol level was not statistically different in the non-AH-group (1.94 ± 3.15 ng/ml) compared to the AH-group (4.62 ± 2.29 ng/ml) at pre-simulated flight (p = 0.062). However, cortisol was significantly higher in AH-group (5.94 ± 2.03 ng/ml) compared with non-AH-group (1.52 ± 1.16 ng/ml) during the simulated flight (p < 0.001). No significant difference was found for cortisol in the non-AH-group (1.19 ± 0.50 ng/ml) compared to the AH-group (2.02 ± 0.92 ng/ml) at post-simulated flight (p = 0.053) (see Fig. 1a).

Within group comparison in the non-AH-group demonstrated that cortisol did not show any alterations when it was compared between the time points: pre-simulated flight and during simulated flight (p = 1.00), during simulated flight and post-simulated flight (p = 0.93) or pre-simulated flight and post-simulated flight (p = 1.00) (see Fig. 1a).

Within group comparison in the AH-group, cortisol stayed unaltered between pre-simulated flight and during simulated flight (p = 0.29) but it showed a significant difference between during simulated flight and post-simulated flight (p < 0.001) and between pre-simulated flight and post-simulated flight (p = 0.02) (see Fig. 1a).

Between group comparison showed that PGE₂ was not statistically different in the non-AH-group (41.43 ± 6.03 pg/mL) compared to the AH-group (38.93 ± 25.55 pg/mL) at pre-simulated flight (p = 0.82). Non-AH-group (42.13 ± 14.63 pg/mL) compared to the AH-group (38.16 ± 9.85 pg/mL) during the simulated flight did not show any significant difference either (p = 0.52). However, PGE₂ was statistically different in the non-AH-group (52.28 ± 9.02 pg/mL) compared to the AH-group (73.32 ± 16.87 pg/mL) at post-simulated flight (p = 0.01) (see Fig. 1b).

Within group comparison in the non-AH-group, demonstrated that PGE₂ did not show any remarkable alterations when it was compared between the time points of pre-simulated flight and during simulated flight (p = 1.00), or during simulated flight and post-simulated flight (p = 0.41). However, a significant difference was found between pre-simulated flight and post-simulated flight (p = 0.01) (see Fig. 1b).

Within group comparison in the AH-group, showed no statistically significant difference in PGE₂ levels between pre-simulated flight and during simulated flight (p = 1.00). However, a significant difference was found during simulated flight and post-simulated flight (p = 0.01) and between pre-simulated flight and post-simulated flight (p = 0.04) (see Fig. 1b).

Between group comparison showed that SPO was not statistically different in the non-AH-group (97.43 ± 1.51%) compared to the AH-group (98.00 ± 1.00%) at pre-simulated flight (p = 0.49). However, SPO was significantly different in the non-AH-group (95.57 ± 1.81%) compared to the AH-group (91.85 ± 2.48%) during the simulated flight (p < 0.001). At post-simulated flight, SPO was not statistically different between the non-AH-group (97.57 ± 1.13%) and the AH-group (98.14 ± 1.06%) (p = 0.095) (see Fig. 1c).

Within group comparison in the non-AH-group, indicated that SPO did not show any difference when it was compared between the time points of pre-simulated flight and during simulated flight (p = 0.06), during simulated flight and post-simulated flight (p = 0.07) or pre-simulated flight and post-simulated flight (p = 1.00) (see Fig. 1c).

Within group comparison in the AH-group, SPO was significantly different between pre-simulated flight and during simulated flight (p < 0.001), during simulated flight and post-simulated flight (p < 0.001), but did not show a difference between pre-simulated flight and post-simulated flight (p = 1.00) (see Fig. 1c).

Between group comparison showed that the pulse did not show any significant difference in the non-AH-group (70.43 ± 17.85 bpm) compared to the AH-group (74.14 ± 10.51 bpm) at pre-simulated flight (p = 0.62). No difference was found in the non-AH-group (74.43 ± 20.64 bpm) compared to the AH-group (75.00 ± 8.91 bpm) during the simulated flight, (p = 0.96), or post-simulated flight (p = 0.20) in the non-AH-group (72.86 ± 6.23 bpm) compared to the AH-group (67.29 ± 6.40 bpm) (see Fig. 2a).

Within group comparison in the non-AH-group, indicated that the pulse did not show any significant variations between pre-simulated flight and during simulated flight (p = 0.74), during simulated flight and post-simulated flight (p = 1.00), or pre-simulated flight and post-simulated flight (p = 1.00) (see Fig. 2a).

Within group comparison in the AH-group, presented a similar observation, where pulse did not show a significant alteration through assessment time-points: between pre-simulated flight and during simulated flight (p = 1.00), during simulated flight and post-simulated flight (p = 1.00), or pre-simulated flight and post-simulated flight (p = 0.93) (see Fig. 2a).

Between group comparison showed that SBP was not significantly different in the non-AH-group (133.14 ± 18.69 mmHg) compared to the AH-group (121.14 ± 11.57 mmHg) at pre-simulated flight (p = 0.26). This was also the case when SBP was compared between the non-AH-group (125.71 ± 12.58 mmHg) and the AH-group (123.43 ± 14.27 mmHg) during the simulated flight (p = 0.77), and also in the non-AH-group (126.57 ± 11.19 mmHg) compared to the AH-group (117.00 ± 10.60 mmHg) at post-simulated flight (p = 0.13) (see Fig. 2b).

Within group comparison in the non-AH-group, indicated that SBP did not show any significant variations when the comparisons were performed between pre-simulated flight and during simulated flight (p = 0.78), during simulated flight and post-simulated flight (p = 1.00), or pre-simulated flight and post-simulated flight (p = 1.00) (see Fig. 2b).

Within group comparison in the AH-group, also indicated similar findings. SBP did not vary significantly when comparisons performed between pre-simulated flight and during simulated flight (p = 0.50), during simulated flight and post-simulated flight (p = 1.00), or pre-simulated flight and post-simulated flight (p = 1.00) (see Fig. 2b).

Between group comparison showed that DBP was not different in the non-AH-group (81.14 ± 3.72 mmHg) compared to the AH-group (73.00 ± 6.98 mmHg) at pre-simulated flight (p = 0.69), nor was it different in the non-AH-group (77.00 ± 10.86 mmHg) compared to the AH-group (78.00 ± 8.66 mmHg) during the simulated flight (p = 0.86), or in the non-AH-group (77.43 ± 9.40 mmHg) compared to the AH-group (72.71 ± 9.39 mmHg) at post-simulated flight (p = 0.34) (see Fig. 2c).

Within group comparison in the non-AH-group, presented that DBP did not show a variation between the assessments’ time-points: pre-simulated flight versus during simulated flight (p = 1.00), during simulated flight and post-simulated flight (p = 1.00), or pre-simulated flight and post-simulated flight (p = 1.00) (see Fig. 2c).

Within group comparison in the AH-group showed no significant change in DBP when comparisons were performed between pre-simulated flight and during simulated flight (p = 0.15), during simulated flight and post-simulated flight (p = 0.12), or pre-simulated flight and post-simulated flight (p = 1.00) (see Fig. 2c).

Between group comparison showed that the average facial skin temperature was not statistically different between the non-AH-group (34.97 ± 0.24 °C) and the AH-group (35.31 ± 0.30 °C) at pre-simulated flight (p = 0.12). The facial skin temperature did not show any significant difference between the non-AH-group (35.10 ± 0.14 °C) and the AH-group (35.37 ± 0.52 °C) during the simulated flight (p = 0.19), or at post-simulated flight (p = 0.33), where non-AH-group (34.36 ± 0.53 °C) compared with the AH-group (35.30 ± 0.42 °C), (see Fig. 2d).

Within group comparison in the non-AH-group, indicated that the facial skin temperature did not show any significant difference between pre-simulated flight and during simulated flight (p = 1.00), during simulated flight and post-simulated flight (p = 1.00), or pre-simulated flight and post-simulated flight (p = 1.00) (see Fig. 2d).

Within group comparison in the AH-group, also presented that no significant alteration in the facial skin temperature occurred when it was compared between pre-simulated flight and during simulated flight (p = 1.00), during simulated flight and post-simulated flight (p = 1.00), or pre-simulated flight and post-simulated flight (p = 1.00) (see Fig. 2d).

The concentration of cortisol was elevated during the simulated flight for the AH-group, whereas it dropped for all members of the non-AH-group during the stay in the pressure chamber. This is somewhat in accordance to the results from the study by Simeoni et al. [27], who claimed that the salivary cortisol level elevated significantly in their healthy participants in a simulated flight. The increase in the cortisol level for these participants was particularly high, which is believed to be due to the fact that the simulated flight reached an altitude of 7620 m, which induced hypoxia. The participants were divided into two groups. One group had the oxygen masks on during the simulated flight, whereas the other group did not wear any oxygen masks. Considering that in our simulated flight that only reached an altitude of 2438 m, and no one had any mask on, it could be explained as to why the cortisol levels in our healthy participants are not severely affected. It has previously been shown that certain passengers that suffers from AH, have developed anxiety at the time of flight traveling [6]. It has also been proven that people suffering from anxiety have elevated levels of cortisol [28]. Our simulated flight was assisted by safety observers who could immediately stop the simulated flight in the pressure chamber if it was necessary and this was a source of confidence for all participants. However, our participants suffering from AH were aware of the fact that they were about to experience a simulated flight, and thus might have a risk of developing a bout of AH in the chamber, which was reflected on elevated levels of cortisol. In this group, cortisol levels are significantly higher before and during the simulated flight. We do note, however, that cortisol levels in both groups are almost equal after the simulated flight, which might suggest, that both groups are relieved to see that the simulated flight has come to an end. Simeoni et al. [27] has shown that the group wearing oxygen masks in their study only experienced a slight raise in cortisol levels during the simulated flight. It has been proven that headache can be triggered by extended durations of hypoxia. It could be of great interest to see if the use of oxygen masks by passengers suffering from AH would decrease cortisol levels, or prevent bouts of AH during flight.

The PGE₂ levels for both the AH- and the non-AH-groups were almost identical both before and during the simulated flight. However, the PGE₂ level was found elevated in both groups after the flight, where the increase was higher in the AH-group. This is in accordance with Benedetti et al. [29] who showed that the PGE₂ level increased significantly, when participants were moved to an altitude of 3500 m [29], where they were exposed to hypoxia at this altitude. All participants in this study were healthy volunteers, who experienced HAH, while situated in an elevated position. HAH can be a risk factor for AH as it was reported in a recent study by Bui et al. [2] . Even though HAH and AH are two separate headaches, hypoxia might be one of the main players in both headaches as a consequence of atmospheric pressure alteration. We therefore speculate that there might be a possible relation between HAH and AH [2]. Our data show that the PGE₂ level is increased significantly after the simulated flight for the AH-group, which might show a potential link with development of AH. PGE₂ is known for causing vasodilation in the cerebral arteries and it is thought to play a major role in development of certain headaches [14‐18]. Ipekdal et al. [10] demonstrated that triptans cause vasospasm in cerebral arteries and thereby prevent AH if the drug is administered 30 min before the flight travel. It is well-known that vasodilation of the cerebral arteries occurs during a migraine attack and that triptians are effective in acute migraine [30, 31]. At least in part, based on effectiveness of triptans, vasodilation of the cerebral arteries might be one of the potential causes of AH [2, 10]. An interesting finding by Benedetti et al. [29], was that the use of oxygen masks decreased the PGE₂ level as well as the intensity of the HAH. Due to the low atmospheric pressure in the airplane, the body would absorb less oxygen, where hypoxia can lead to headache [2, 32, 33]. It remains uncertain whether hypoxia is associated with AH or not, but it would be interesting to examine if an increase in cabin oxygen levels or use of oxygen masks can decrease the PGE₂ level and consequently prevent AH. Another possible way of preventing AH, might be by aid of EP₄ receptor antagonist, BGC20-1531. It has been shown that elevated levels of PGE₂ up-regulate EP₄-receptors, which has been suggested as a target for pain treatment [34]. It might be of great interest to see if BGC20-1531 can be used in the future to prevent AH.

The average SPO for the healthy participants was found decreased in both groups during the simulated flight in the pressure chamber. A previous study has shown that the SPO decreases during flight [28]. A total of 84 healthy participants were examined in that study. The SPO decreased on an average from 97% to 95% in 46% of the population during a flight lasting at least for an hour. Despite the fact that our population only consists of 14 participants, our data indicate that the SPO is different for healthy participants compared to participants suffering from AH. The SPO for our healthy participants decreased gradually from 97.57 ± 1.13% to 95.57 ± 1.81%, which is in accordance with the results by Humphreys et al. [28]. The drop was significant for the AH-group, where the saturation decreased from 98.00 ± 1.00% to 91.85 ± 2.48%. The decreased SPO for both groups might be the result of the low atmospheric pressure inside the pressure chamber. It has been shown that the cabin pressure decreases with 8 hPa for every increasing 300 m, until it stabilizes at 846 hPa for the remaining duration of the flight at an altitude of 2500 m [35]. It has been proposed that lower atmospheric pressure can lead to a headache due to hypoxia [11]. As compensation, human body would absorb a lesser amount of oxygen as pressure drops. Our data also indicate that the average SPO was elevated after the participants left the pressure chamber. It remains unclear whether the AH-group is more sensitive to the changes in the atmospheric pressure, and whether or not the low saturation alone can trigger a headache with AH features in this group. If this is the case, it would be interesting to see if an increased oxygen concentration in the cabin could alter the outcome.

Our data indicate a gradual increase in pulse for both groups of non-AH-group and AH-group during the simulation trial in the pressure chamber, even though the increase did not reach to a significant change statistically. This finding is in accordance with the study by Gardiner et al. [36] who examined the effect of altitude on blood pressure and pulse in healthy participants. The participants were elevated to an altitude of 1828 m, which caused a decrease in blood pressure and an increase in pulse. Another study by Bian et al. [37] has also measured pulse and blood pressure for participants with HAH during a flight at the altitude of 3500 m, where they showed an increased blood pressure and pulse. However, the increased blood pressure was not significantly correlated with HAH. An increased pulse is a normal physiological compensation in human body, when the atmospheric pressure is low. SBP and BDP for the non-AH-group indicated a decrease during the flight, which to some extent correlates with the results by Gardiner et al. [36]. The SBP and BDP for the AH-group increased gradually during the simulated flight, which was similar to the study by Bian et al. [37]. It is known that healthy participants can compensate changing pressure by lowering the arterial blood pressure [36]. It remains unclear however, if the gradual elevation of blood pressure and pulse frequency would play a role in the development of HAH, and thereby in AH, which requires further studies. The study by Bian et al. [37] indicates that a significant elevated pulse can trigger a bout of HAH. A non-significant increased pulse was observed in our AH-group during the simulated flight, but this gradual increasing has potentially been insufficient to evoke an attack or might be just indicating that elevated pulse might not be a trigger for AH. Furthermore, it has been shown that one patient with AH had a normal pulse and blood pressure, but it is unknown if recordings were performed during the flight [38]. Pulse and blood pressure might be influenced by multiple confounding factors and this point should definitely be considered when interpreting data.

We measured facial skin temperature around the nose, sinuses and the temporal areas, to record if any alteration would occur. Studies have shown that the body temperature is slightly elevated in higher altitudes [39]. We expected to see skin temperature alterations in the pressure chamber. An insignificant tendency was seen, that the average facial skin temperature was higher in the AH-group compared to the non-AH-group during simulation in the chamber. The study by Drummond et al. [40] has examined the facial temperature among migraine patients and healthy participants. Their results showed that the facial temperature in the orbital region was significantly higher among migraine patients during a migraine attack compared to healthy participants. Vasodilation of the cerebral arteries might be considered as a potential cause of AH and migraine where both headaches respond to triptans [2, 10]. Gradual elevation of facial skin temperature in the AH-group, compared with the non-AH-group in the chamber, could have been triggered by an internal occurrence of a potential vasodilation process. This needs further investigation to identify whether it can be systemic, or regional, peripheral or central or both and to what extent. It should be noted that the participants wore ear protection during their stay in the chamber, which interfered with the proper measurement of the average temperature on facial region. Had the participants been able to remove the earmuffs (not possible due to need for communication/extreme noise), then the average temperature would presumably have been recorded accurately. However, the average temperature between the two groups and between the three time-slots did not show any significant alterations. It is plausible that temperature changes could be relevant. A super sensitive technique or proper preparations might yield a different outcome.

In summary, several mechanisms have been postulated to explain the alterations seen in AH. Fig. 3 presents proposed mechanisms underlying development of AH as described here based on present findings and other cited studies.