Feasibility of Delivering 5-Day Normobaric Hypoxia Breathing in a Hospital Setting

Abstract

BACKGROUND:

Beneficial effects of breathing at ${\mathbf{\text{F}}}_{{\mathbf{\text{IO}}}_{\mathbf{\text{2}}}}$ < 0.21 on disease outcomes have been reported in previous preclinical and clinical studies. However, the safety and intra-hospital feasibility of breathing hypoxic gas for 5 d have not been established. In this study, we examined the physiologic effects of breathing a gas mixture with ${\mathbf{\text{F}}}_{{\mathbf{\text{IO}}}_{\mathbf{\text{2}}}}$ as low as 0.11 in 5 healthy volunteers.

METHODS:

All 5 subjects completed the study, spending 5 consecutive days in a hypoxic tent, where the ambient oxygen level was lowered in a stepwise manner over 5 d, from ${\mathbf{\text{F}}}_{{\mathbf{\text{IO}}}_{\mathbf{\text{2}}}}$ of 0.16 on the first day to ${\mathbf{\text{F}}}_{{\mathbf{\text{IO}}}_{\mathbf{\text{2}}}}$ of 0.11 on the fifth day of the study. All the subjects returned to an environment at room air on the sixth day. The subjects' ${\mathbf{\text{S}}}_{{\mathbf{\text{pO}}}_{\mathbf{\text{2}}}}$, heart rate, and breathing frequency were continuously recorded, along with daily blood sampling, neurologic evaluations, transthoracic echocardiography, and mental status assessments.

RESULTS:

Breathing hypoxia concentration dependently caused profound physiologic changes, including decreased ${\mathbf{\text{S}}}_{{\mathbf{\text{pO}}}_{\mathbf{\text{2}}}}$ and increased heart rate. At ${\mathbf{\text{F}}}_{{\mathbf{\text{IO}}}_{\mathbf{\text{2}}}}$ of 0.14, the mean ${\mathbf{\text{S}}}_{{\mathbf{\text{pO}}}_{\mathbf{\text{2}}}}$ was 92%; at ${\mathbf{\text{F}}}_{{\mathbf{\text{IO}}}_{\mathbf{\text{2}}}}$ of 0.13, the mean ${\mathbf{\text{S}}}_{{\mathbf{\text{pO}}}_{\mathbf{\text{2}}}}$ was 93%; at ${\mathbf{\text{F}}}_{{\mathbf{\text{IO}}}_{\mathbf{\text{2}}}}$ of 0.12, the mean ${\mathbf{\text{S}}}_{{\mathbf{\text{pO}}}_{\mathbf{\text{2}}}}$ was 88%; at ${\mathbf{\text{F}}}_{{\mathbf{\text{IO}}}_{\mathbf{\text{2}}}}$ of 0.11, the mean ${\mathbf{\text{S}}}_{{\mathbf{\text{pO}}}_{\mathbf{\text{2}}}}$ was 85%; and, finally, at an ${\mathbf{\text{F}}}_{{\mathbf{\text{IO}}}_{\mathbf{\text{2}}}}$ of 0.21, the mean ${\mathbf{\text{S}}}_{{\mathbf{\text{pO}}}_{\mathbf{\text{2}}}}$ was 98%. These changes were accompanied by increased erythropoietin levels and reticulocyte counts in blood. All 5 subjects concluded the study with no adverse events. No subjects exhibited signs of mental status changes or pulmonary hypertension.

CONCLUSIONS:

Results of the current physiologic study suggests that, within a hospital setting, delivering ${\mathbf{\text{F}}}_{{\mathbf{\text{IO}}}_{\mathbf{\text{2}}}}$ as low as 0.11 is feasible and safe in healthy subjects, and provides the foundation for future studies in which therapeutic effects of hypoxia breathing are tested.

Introduction

Potential benefits of breathing ${\text{F}}_{{\text{IO}}_{\text{2}}}$ < 0.21 on certain diseases have been suggested in clinical and preclinical studies. One of the first studies to demonstrate such a benefit evaluated the effect of hypoxia in a mouse model of Leigh syndrome, the most common pediatric mitochondrial disease.¹ The mice were placed in a hypoxemic environment with an ${\text{F}}_{{\text{IO}}_{\text{2}}}$ of 0.11 over a period of 21 – 28 d on average. This initial study showed that breathing ${\text{F}}_{{\text{IO}}_{\text{2}}}$ of 0.11 prevented the progression and development of Leigh syndrome, whereas supplemental oxygen ( ${\text{F}}_{{\text{IO}}_{\text{2}}}$ 0.55) accelerated the demise of the animals.¹ A subsequent study showed that advanced neurologic disease in Leigh syndrome could even be reversed.² Chronic continuous hypoxia has now been extended to other models, and recent studies have shown that it benefits neurologic disease in a mouse model of Friedreich ataxia and multiple sclerosis as well as in mouse models of aging.^3-6 Although multiple studies have tested hypoxia as a potential treatment in murine models, there are no feasibility studies in humans in a hospital setting. One recent study tested daily intermittent exposure to an ${\text{F}}_{{\text{IO}}_{\text{2}}}$ of 0.09 in subjects with chronic incomplete spinal cord injury and showed an improvement in walking speed and distance walked.^7,8

Furthermore, a variety of exposures to both normobaric and hypobaric hypoxic environments and training equipment (eg, portable nitrogen generators with masks) have become popular training techniques used by elite endurance athletes.⁹ Hypoxemia exposure in athletes has been shown to increase aerobic capacity, thereby improving endurance during prolonged exercise.^9,10 Prolonged therapeutic hypoxia in humans is challenging due to cardiopulmonary and neurologic safety concerns, and has yet to be studied conclusively. Severe pulmonary hypertension with or without pulmonary edema has been reported in healthy subjects breathing ${\text{F}}_{{\text{IO}}_{\text{2}}}$ of 0.10 to 0.12 at sea level and at high altitude.^11-15 Neurologic symptoms, including confusion, headache, fatigue, and difficulty sleeping, are common in individuals with hypoxemia and in high-altitude mountain climbers.^16,17 To investigate the therapeutic potential of hypoxic breathing, it is important to establish whether it is feasible and safe to deliver prolonged hypoxia in a hospital environment.

Based on our recent study of hypoxia treatment of a mouse model deficient for complex I of the mitochondrial respiratory chain, we targeted a ${\text{P}}_{{\text{aO}}_{\text{2}}}$ of ∼45 – 55 mm Hg, which corresponds to ${\text{S}}_{{\text{pO}}_{\text{2}}}$ between 80 and 85%.¹ As shown in earlier aviation studies in civilians, our hypothesis was that prolonged moderate hypoxemia exposure is safe and feasible in healthy subjects.¹⁸ Our aim was to build a safe intra-hospital model for exposing humans to prolonged normobaric hypoxic air while closely monitoring cardiopulmonary and neurologic function. We aimed to use equipment already widely used in the sporting industry to achieve normobaric hypoxia. In contrast to previous trials, to minimize risks, we sought to mimic the gradual acclimatization process used by mountain climbers before embarking on high-altitude climbs.¹⁹ With this gradual acclimatization method, we attempted to reduce the effects of altered mental status associated with an acute decrease in ${\text{S}}_{{\text{pO}}_{\text{2}}}$ seen in climbers who are rapidly exposed to high altitude. Recent studies with residents of high-altitude regions (eg, Tibetans, Andeans, Ethiopians) have shown that the neurologic effects of acute exposure to a high-altitude–related decreased ${\text{S}}_{{\text{pO}}_{\text{2}}}$ can partially recover with acclimatization.^20,21 To achieve the benefits of acclimatization and reduce acute oxygen desaturation, we reduced the continuous ${\text{F}}_{{\text{IO}}_{\text{2}}}$ experienced by our healthy volunteers from 0.16 to 0.11 over a 5-d period in a hospital setting.

QUICK LOOK.

Current Knowledge

The use of prolonged hypoxia has been demonstrated to be safe and feasible, and has been shown to be a possible treatment for various disorders and diseases in murine models. Intermittent hypoxia breathing in humans has been shown to increase walking speed and distance in patients after chronic spinal cord injury. Intermittent hypoxia breathing has been readily studied in humans and has been shown to be safe and feasible.

What This Paper Contributes to Our Knowledge

In healthy volunteers, breathing normobaric hypoxia was shown to be safe and feasible, with no physiologic or biochemical adverse events. This study was able to show that building a normobaric hypoxic environment in a hospital setting is feasible.

Methods

This study was performed in a clinical research suite on the 12th floor of the Massachusetts General Hospital, Boston, Massachusetts, in a dedicated hospital room at 35 m above sea level with an approximate barometric pressure of 754 mm Hg. Throughout this article, we refer to this barometric pressure as normobaric. The study was approved by the Massachusetts General Hospital investigation review board (2016P001802) and published as a clinical trial on clinicaltrials.gov (NCT02860975). After obtaining informed consent, 5 healthy volunteers between 18 and 40 years of age were enrolled in the study (Fig. 1A). Our primary end point was to assess the safety and feasibility of continuous exposure (5 d) to humidified normobaric hypoxic gas with ${\text{F}}_{{\text{IO}}_{\text{2}}}$ as low as 0.11 in a hospital environment. The secondary end points of this study included descriptions of physiologic and biochemical response to prolonged normobaric hypoxia as well as changed on return to room air.

Fig. 1.

Demographic data (A) and schema of changes of concentrations of ${\text{F}}_{{\text{IO}}_{\text{2}}}$ during the study period (B). Target ${\text{F}}_{{\text{IO}}_{\text{2}}}$ (0.11) is shown in red.

The Hospital Hypoxic Environment

A hypoxic environment was created by using a commercially available tent normally used for athletic performance enhancement (Hypoxico Inc, New York, NY). The tent construction and materials minimize entry of outside air. We used a small, room-sized tent (at-home cubicle altitude tent [Hypoxico]) that housed a hospital bed, a chair, and a small table (Fig. S1 [see the supplementary materials at http://www.rcjournal.com]). Nitrogen-enriched humidified gas was added to the tent by using nitrogen generators (Everest Summit II Altitude Generato [Hypoxico]). The air flow from these generators was high enough to create a slight positive pressure effect, which limited entry of outside air and oxygen into the tent. CO₂ scavenging was performed by adding a CO₂ absorber (soda lime) to the nitrogen generator inlet to maintain a goal-inspired CO₂ concentration < 450 ppm (0.045%). Relative humidity was maintained to range between 30 and 40%.

Subjects’ caloric need was calculated on the day before study initiation by indirect calorimetry (VMAX Encore Metabolic Cart, Care Fusion, Yorba Linda, California). All the subjects received a similar composition of macronutrients during the 7-d period to standardize their diet. To perform hygienic needs (eg, use of the toilet, washing, bathing) or during daily ultrasound assessments, the subjects were able to exit the tent, and a hypoxic gas mixture was continuously delivered via high-flow nasal cannula at 60 – 70 L/min by adding nitrogen (medical grade nitrogen gas, Airgas, Radnor, PA) to medical air to obtain the desired ${\text{F}}_{{\text{IO}}_{\text{2}}\mathrm{.}}$ To keep ${\text{F}}_{{\text{IO}}_{\text{2}}}$ constant when subjects exited the tent, they were instructed to breathe through their nose while keeping their mouths closed. ${\text{F}}_{{\text{IO}}_{\text{2}}}$ was gradually decreased to 0.11 over a period of 5 d to target an ${\text{S}}_{{\text{pO}}_{\text{2}}}$ between 80 and 85% (with a corresponding ${\text{P}}_{{\text{aO}}_{\text{2}}}$ of 45 – 55 mm Hg). Overnight, except for study day 5, ${\text{F}}_{{\text{IO}}_{\text{2}}}$ was increased by 0.02 to allow for adaption to sleep at low oxygen (Fig. 1B). ${\text{F}}_{{\text{IO}}_{\text{2}}}$ both in the tent or when delivered by high-flow nasal cannula was continuously monitored by using a portable oxygen percentage analyzer (MAXO2+A [Hypoxico]).

Physiologic, Neurologic, and Laboratory Testing

Study subjects underwent continuous heart rate and ${\text{S}}_{{\text{pO}}_{\text{2}}}$ monitoring (RAD57 Pulse Oximeter, Masimo, Irvine, California) with the remaining vital signs (breathing frequency, blood pressure, and temperature) obtained every 4 h. Blood pressure was measured noninvasively by using a portable monitor (CARESCAPE, GE Healthcare, Boston, Massachusetts). Withdrawal rules included an ${\text{S}}_{{\text{pO}}_{\text{2}}}$ of ≤70% for >1 min (Table S1 [see the supplementary materials at http://www.rcjournal.com]). Each morning of the study, transthoracic echocardiography was performed noninvasively to monitor for signs of pulmonary hypertension. By using transthoracic echocardiography, tricuspid valve regurgitation was measured and used to approximate pulmonary pressures. If tricuspid regurgitation was absent, then tricuspid acceleration time was measured and used.²² In addition, lung ultrasound and chest auscultation were performed twice daily to monitor for evidence of pulmonary edema. From a neurologic standpoint, the optic nerve sheath diameter was measured via ultrasound each morning and used as a surrogate for intracranial pressure; daily neurologic examinations were also performed (ie, twice daily assessments of reflexes, cranial nerve function, ataxia, strength). In addition, a full review of systems was performed twice daily as well as completion of a questionnaire to assess for symptoms of acute mountain sickness by using the Lake Louis Acute Mountain Sickness Score (Table S2 [see the supplementary materials at http://www.rcjournal.com]).¹⁶

Blood samples were collected as follows for biochemical analysis. All the subjects had blood collections taken before breakfast before initiating hypoxia (day 0: baseline, fasting), after initiation of hypoxia (day 1), and each morning before breakfast on subsequent days: days 2, 3, 4, 5, and on day 6 before and after cessation of hypoxia. In addition, on day 6, an arterial blood gas sampling was collected immediately before exiting the hypoxic tent. Finally, blood samples were collected before breakfast on day 7. As a surrogate marker of hypoxia-inducible factor pathway activation, reticulocyte count, and erythropoietin levels were also measured.²³ Absolute concentrations for lactate and β-hydroxybutyrate were determined in plasma samples from all the subjects by using a previously described stable-isotope dilution method.²⁴

Statistical Analysis

We used the Kendall τ on individual subject data to test for a monotonic trend, analyzing ${\text{S}}_{{\text{pO}}_{\text{2}}}$, breathing frequency, and heart rate (all both day and night) over the exposure days (and ${\text{F}}_{{\text{IO}}_{\text{2}}}$). Other explored continuous variables were described as individual data (ie, Lake Louis Acute Mountain Sickness Score, right and left optic nerve sheath diameters, daily kcals, daily urine and total output and daily fluid balance, pulmonary artery pressure, average systolic [mm Hg], and diastolic [mm Hg] blood pressures) or by mean ± SD (ie, reticulocyte count, erythropoietin levels, body mass index, serum creatinine, serum bicarbonate, urine sodium and chloride levels, lactic acid levels, and β-hydroxybutyrate). In terms of reticulocyte count and erythropoietin levels, the Kruskal-Wallis test was used to test for significance over time.

Results

Subject demographics are described in Figure 1A. Each subject completed the study in its entirety without adverse events. Targeted ${\text{F}}_{{\text{IO}}_{\text{2}}}$ was always reached and maintained successfully in all the subjects, including when the subjects went outside the hypoxia tent (by using high-flow nasal cannula).

Tolerability and Safety

All the subjects completed the study. The tolerability was assessed by using the subjective portion of the Lake Louis Acute Mountain Sickness Score system, administered twice daily on each day of the study. The highest Lake Louis Acute Mountain Sickness Score recorded at any one point during the study was 2, and, thus, no subjects met subjective criteria for acute mountain sickness. The most common symptoms were mild headache and sleeping difficulties (Table S3 [see the supplementary materials at http://www.rcjournal.com]).

With respect to the safety profile of the study, we monitored for evidence of hypoxia-related organ dysfunction. From a neurologic standpoint, optic nerve sheath ultrasounds were normal and remained unchanged for all the subjects throughout the study protocol, which suggests normal intracranial pressures and lack of cerebral edema (Table S4 [see the supplementary materials at http://www.rcjournal.com]). As mentioned above, no subjects met criteria for acute mountain sickness based on the subjective portion of the Lake Louis Acute Mountain Sickness Score. In addition, no subject had any abnormal objective neurologic examination findings or any abnormal mental status examinations, also assessed twice daily, to complete and to supplement the Lake Louis Acute Mountain Sickness Score.

From a cardiovascular standpoint, although heart rates increased with continued exposure to hypoxia (Fig. 2), daily electrocardiogram monitoring revealed sinus tachycardia, and no arrhythmias were identified. Tachycardia resolved immediately on return to normoxia. With respect to pulmonary function, we found no evidence for pulmonary edema both by lung ultrasonography and auscultation. The maximum number of Kerley B lines observed was 4 in subject no. 3 on day 1, with no clinical symptoms of pulmonary edema observed in our subjects (Table S5). Breathing frequency remained unaltered when compared with baseline throughout the study (Fig. S2 [see the supplementary materials at http://www.rcjournal.com]). None of the subjects had desaturations < 70% ${\text{S}}_{{\text{pO}}_{\text{2}}}$ while awake. None of the subjects had desaturations < 70% while asleep that lasted > 1 min, and the lowest single brief desaturation recorded during the study through continuous pulse oximetry was 58% (Fig. 3). Renal function and serum electrolyte status were monitored daily. There were no instances of acute kidney injury during the study period based on serum creatinine levels and no abnormalities in serum electrolytes were detected (Fig. S3 [see the supplementary materials at http://www.rcjournal.com]).

Fig. 2.

Heart rate (beats/min) increases with decreasing ${\text{F}}_{{\text{IO}}_{\text{2}}}$ and recovers to baseline with return to normoxia. A: During the day, and B: during the night.

Fig. 3.

A: ${\text{S}}_{{\text{pO}}_{\text{2}}}$ over time during the day. B: ${\text{S}}_{{\text{pO}}_{\text{2}}}$ over time during the night. C: Arterial blood gas levels at ${\text{F}}_{{\text{IO}}_{\text{2}}}$ 0.11.

In all the subjects, body weight decreased throughout the hypoxic portion of the study and increased after return to normoxia (Fig. S4). Caloric intake throughout the hypoxic days of the study did not differ from caloric intake on return to normoxia, which suggests that weight loss during prolonged hypoxia was not secondary to decreased caloric intake but rather due to sensible fluid loss (Table S6). In fact, in subject no. 5, we closely followed daily fluid balance and observed a net volume deficit during the hypoxic portion of the study (Table S7 [see the supplementary materials at http://www.rcjournal.com]), with an impressive peak urine output of ∼5 L.

Response to Hypoxia

The subjects demonstrated a progressive increase in heart rate and decease in ${\text{S}}_{{\text{pO}}_{\text{2}}}$ as ${\text{F}}_{{\text{IO}}_{\text{2}}}$ decreased (Figs. 2 and 3, respectively). As mentioned above, major changes in breathing frequencies as a response to decreasing ${\text{F}}_{{\text{IO}}_{\text{2}}}$ were not detected during the study (see the supplementary materials at http://www.rcjournal.com). Noninvasive means of tidal volume measurements were not performed. Arterial blood gas sampling obtained at ${\text{F}}_{{\text{IO}}_{\text{2}}}$ of 0.11 just before exit from hypoxic gas demonstrated hypoxemia and a hypocarbic respiratory alkalosis (pH range, 7.46 – 7.56) (Fig. 3C). Echocardiography revealed that estimated pulmonary artery pressures increased with decreasing ${\text{F}}_{{\text{IO}}_{\text{2}}}$_, but no values met transthoracic echocardiography criteria for pulmonary arterial hypertension (Table S8 [see the supplementary materials at http://www.rcjournal.com]). Furthermore, right- and left-ventricular function remained normal for all the subjects during the study. Systemic blood pressures did not change with breathing hypoxia throughout the duration of the study (Table S9 [see the supplementary materials at http://www.rcjournal.com]).

There was a decrease in serum bicarbonate just before exiting the hypoxic tent as measured both by blood gas (range, 16 – 22 mEq/L) as well as by serum chemistry (level of 20 mEq/L for all 3 subjects sampled at this time point; the other 2 subjects’ samples were not processed at this time point) (Fig. S5 [see the supplementary materials at http://www.rcjournal.com]). Although urine sodium and chloride levels were measured and decreased early in the time course of hypoxic gas exposure, urine bicarbonate was not measured in our study. Urine sodium and chloride levels remained lower than baseline, even after return to normoxia on days 6 and 7 (Fig. S6 [see the supplementary materials at http://www.rcjournal.com]). Lactic acid levels did not change in a statistically significant way during the study protocol, suggestive against detrimental end-organ hypoxia (Fig. S7 [see the supplementary materials at http://www.rcjournal.com]). β-Hydroxybutyrate levels, however, did trend upward, with a peak just after returning to normoxia. Levels normalized on recheck 5 h later and on the final sampling on day 7 (Fig. S8 [see the supplementary materials at http://www.rcjournal.com]).

Daily levels of erythropoietin and the reticulocyte count percentage were measured throughout the study period (Fig. 4). There was a gradual increase in erythropoietin levels that correlated with the degree of hypoxia, with the most marked increase when the subjects were breathing ${\text{F}}_{{\text{IO}}_{\text{2}}}$ of 0.11 (Table S10 [see the supplementary materials at http://www.rcjournal.com]). The reticulocyte count increased as well with decreasing ${\text{F}}_{{\text{IO}}_{\text{2}}}$ . Although erythropoietin levels precipitously decreased on return to normoxia to levels below pre-hypoxia baseline, the reticulocyte count percentage continued to increase after the return to normoxia, which suggests that hypoxia-inducible factor activation was abruptly interrupted by return from ${\text{F}}_{{\text{IO}}_{\text{2}}}$ of 0.11 to room air (Fig. 4; Table S11 [see the supplementary materials at http://www.rcjournal.com]).

Fig. 4.

Daily erythropoietin concentration (A) and reticulocyte count. (B). Data are shown as mean, with whiskers depicting SD. * P < .05.

Discussion

Detrimental roles of excess tissue oxygen levels have been suggested in certain diseases, including Leigh syndrome and Friedreich ataxia.^1,3 Although beneficial effects of chronic hypoxia breathing have been reported in preclinical models of these and other diseases, the safety and intra-hospital feasibility of prolonged hypoxia as a treatment has not been examined.^1,3 In this study, we examined physiologic and biochemical effects of prolonged normobaric hypoxia in healthy volunteers. We found that when breathing an hypoxic gas mixture, ${\text{F}}_{{\text{IO}}_{\text{2}}}$ as low as 0.11, is safe, feasible, and tolerable in healthy volunteers. Breathing ${\text{F}}_{{\text{IO}}_{\text{2}}}$ of 0.11 decreased ${\text{S}}_{{\text{pO}}_{\text{2}}}$ to ∼80% and a ${\text{P}}_{{\text{aO}}_{\text{2}}}$ of ∼45 mm Hg, accompanied by mild hypocapnia and respiratory alkalosis.¹ Although hypoxia induced a mild, dose-dependent tachycardia, blood pressure was not affected, and pulmonary hypertension or edema was not observed. None of the subjects showed mental status changes or signs of cerebral edema. On biochemical examination, we observed increased levels of erythropoietin and reticulocytosis, which suggests an activation of hypoxia-inducible factor signaling. These observations revealed a physiologic and biochemical response to breathing moderate hypoxia in healthy humans and to provide a foundation for future therapeutic application of hypoxia. With regard to pulmonary function, the maximum number of Kerley B lines found in any of our subjects at a given time was 4. There was not an increase of Kerley B lines in this subject or any other subject over time, and our subjects showed no symptoms of pulmonary edema.²⁵

In contrast to our results, subjects from 2 previous studies exposed to normobaric ${\text{F}}_{{\text{IO}}_{\text{2}}}$ of 0.12 or to hypobaric high altitude at 4,559 m experienced high-altitude sickness symptoms, pulmonary hypertension, pulmonary edema, and neurologic symptoms of acute mountain sickness in previous studies.^11,13 In comparing the hypoxia exposure protocols, an important difference in our study design that likely contributed to avoidance of adverse effects of hypoxia is the gradual decrease of ${\text{F}}_{{\text{IO}}_{\text{2}}}$ over time as well as the small increase in ${\text{F}}_{{\text{IO}}_{\text{2}}}$ overnight while carefully monitoring each subject. None of our 5 subjects experienced severe high-altitude sickness, pulmonary hypertension, or pulmonary edema. Thus, there was no need for therapeutic treatments or a return to normobaric normoxia to resolve any adverse effects of hypoxemia at any time.

Physiologic Findings

In a recent human study completed at the German Aerospace Center, 2 healthy volunteers were exposed to prolonged normobaric hypoxia gradually induced in a stepwise manner for a period of 3 weeks, followed by a period of 2 weeks with stable normobaric hypoxia that corresponds to an altitude of ∼7000 m ( ${\text{F}}_{{\text{IO}}_{\text{2}}}$ of 0.085 during the daytime).¹⁵ Both volunteers experienced severe pulmonary hypertension and high-altitude sickness. However, the normobaric hypoxia was well tolerated, without progression to cardiac decompensation and the pulmonary hypertension was resolved over time.¹⁵ Pulmonary hypertension due to hypoxic pulmonary vasoconstriction is a well-known phenomenon that has been described previously in the literature.^11,26 In 1993, Frostell et al¹¹ showed that pulmonary hypertension induced by acute exposure to ${\text{F}}_{{\text{IO}}_{\text{2}}}$ of 0.11 can be reversed in humans by inhalation of 40 ppm of nitric oxide gas, a potent selective pulmonary vasodilator. Similarly, Sherrer et al¹³ reversed pulmonary vasoconstriction induced by hypobaric hypoxemia with inhaled nitric oxide in mountain climbers at high altitude, > 4,000 m.

Our subjects did not demonstrate pulmonary hypertension based on transthoracic echocardiography. The lack of pulmonary hypertension in our subjects might be explained by the following 3 conditions that we implemented in our physiologic study. First, a stepwise progression to hypoxemia allows a better adaptation to hypoxemia, as used by mountain climbers. Second, we targeted a relatively higher ${\text{F}}_{{\text{IO}}_{\text{2}}}$ compared with other studies cited above (0.11 vs 0.085).¹⁵ Third, all our subjects were young and healthy, thereby decreasing the risk of pulmonary hypertension. As shown in our study, the daily erythropoietin levels of the volunteers increased as ${\text{F}}_{{\text{IO}}_{\text{2}}}$ was decreased. Similar findings were reported in previous literature that investigated the physiologic effects of intermittent hypoxia exposure. Erythropoietin production induced by intermittent hypoxia has shown to have both psychiatric antidepressant effects and protection against ischemia-reperfusion injury in both rats and humans.^27-29

Safety

From a safety standpoint, there was no evidence of major hypoxia-related organ dysfunction in our study; we observed no signs of cerebral or pulmonary edema, no evidence of pulmonary hypertension, and no changes in serum lactate values. Although the subjects demonstrated notable weight loss, most likely secondary to increased urine output, no subjects had any evidence of acute kidney injury or serum electrolyte disturbances. As mentioned above, one of our subjects had a peak urine output of 5 L. As described by Hildebrandt et al³⁰ acute hypoxic diuretic response typically occurs when breathing ${\text{F}}_{{\text{IO}}_{\text{2}}}$ between 0.16 and 0.10.³¹ Both the decreased peripheral oxygen saturation and increased heart rate observed with progressive hypoxia immediately resolved on return to normoxia. From a tolerability standpoint, no subjects had symptoms that would qualify for acute mountain sickness and none requested early termination of the protocol.

Several elite endurance athletes use hypobaric hypoxemia training to enhance their performance while competing at normal barometric pressures. A recent meta-analysis has shown that training at high altitudes is not only safe but also has been shown to improve aerobic capacity of these athletes.¹⁰ The use of intermittent hypoxia in rat models has been extensively investigated; studies^7,8 have shown that daily intermittent hypoxia is a safe and effective therapy to induce motor function recovery. Lovett-Barr et al⁷ described that the use of daily intermittent hypoxia had no effect on breathing; they found no difference between tidal volume, breathing frequency, minute ventilation, and mean inspiratory flow between normoxia-treated and daily intermittent hypoxia treated rats.

Clinical Benefits

Several studies have reported that intermittent hypoxia restored motor function in various clinical disorders.²⁹ In a rat model, daily intermittent hypoxia has shown to be an effective therapy for improvement of multiple motor systems after chronic spinal injuries.⁷ In the same study, daily intermittent hypoxia improved forelimb mobility in rats that experienced incomplete cervical spinal injuries.⁷ In a human study, daily intermittent hypoxia exposure improved walking speed and endurance in individuals with chronic incomplete spinal cord injury.⁸ This study found that, with exercise (walking), the benefits of daily intermittent hypoxia exposure is enhanced and showed greater functional benefits in patients with spinal cord injuries.⁸ Whether the use of prolonged hypoxia for spinal cord injury is more efficacious to regain motor function needs further investigation.

Study Limitations

There are some limitations associated with our study design and subject population. The limited sample size (N = 5) limits the generalizability of the present results. It is also important to note that only healthy volunteers were enrolled in this study.

Conclusions

The present study sets a foundation for an eventual trial of hypoxia as a therapy for various diseases and disorders. We have shown safety and intra-hospital feasibility of breathing hypoxic gas through a gradual protocol in healthy volunteers; however, further studies are needed in patients to ascertain safety and utility, and also to determine whether it is beneficial in disease.