Nocturnal Hypoxia Exposure With Simulated Altitude For 14 Days Does Not Significantly Alter Working Memory or Vigilance in Humans

Abstract

Study Objectives:

To assess the effect of 2 weeks of nocturnal hypoxia exposure using simulated altitude on attention and working memory in healthy adult humans.

Design:

Prospective experimental physiological assessment.

Setting:

General Clinical Research Center.

Participants:

Eleven healthy, nonsmoking, subjects (7 men, 4 women). The subjects had a mean age of 27 ± 1.5 years and body mass index of 23 ± 0.9 kg/m2

Interventions:

Subjects were exposed to 9 hours of continuous hypoxia from 2200 to 0700 hours in an altitude tent. Acclimatization was accomplished by graded increases in “altitude” over 3 nights (7700, 10,000 and 13,000 feet), followed by 13,000 feet for 13 consecutive days (FIO₂ 0.13).

Measurements and Results:

Polysomnography that included airflow measurements with a nasal cannula were done at baseline and during 3 time points across the protocol (nights 3, 7, and 14). Attention (10-minute Psychomotor Vigilance Task) and working memory (10-minute verbal 2-back) were assessed at baseline and on day 4, 8, 9, and 15. Nocturnal hypoxia was documented using endpoints of minimum oxygen saturation, oxygen desaturation index, and percentage of total sleep time under 90% and 80%. Total sleep time was reduced, stage 1 sleep was increased, and both obstructive and nonobstructive respiratory events were induced by altitude exposure. There was no difference in subjective mood, attention, or working memory.

Conclusions:

Two weeks of nocturnal continuous hypoxia in an altitude tent did not induce subjective sleepiness or impair objective vigilance and working memory. Caution is recommended in the extrapolation to humans the effects of hypoxia in animal models.

Citation:

Thomas RJ; Ramisier R; Boucher J; Kotlar Y; Vigneault K; Weiss JW; Gilmartin G. Nocternal hypoxia exposure with simulated altitude for 14 days does not significantly alter working memory or vigilance in humans. SLEEP 2007;30(9):1195-1203.

SLEEP-DISORDERED BREATHING IS ASSOCIATED WITH AN ADVERSE IMPACT ON ATTENTION, MOOD, SUBJECTIVE SLEEPINESS, AND EXECUTIVE FUNCTION IN adults and children.^1–4 Well-described consequences of disordered respiration during sleep include recurrent oxygen desaturations, chronic partial sleep deprivation, sleep fragmentation from repetitive arousals, hypoxia, and cytokine dysregulation.^{5, 6} All of these consequences have been implicated as possible contributors to impaired vigilance and executive dysfunction. The effects of intermittent nocturnal hypoxia in mediating sympathoexcitation and daytime hypertension are well established in both patients with sleep-disordered breathing and normal volunteers undergoing hypoxic exposure.⁷ The relative contributions of sleep fragmentation and nocturnal hypoxia in mediating cognitive dysfunction remain to be fully determined. While correlative analysis suggests that both are important, statistical analysis cannot truly dissociate the effects of hypoxia because increasingly severe sleep-disordered breathing is associated with more severe oxygen desaturations as well as greater degrees of sleep fragmentation.

Animal models of intermittent nocturnal hypoxia (rats, mice) have shown convincing evidence of executive dysfunction and excessive sleepiness.⁸ Sleep quality was not measured during hypoxia exposure in most such studies, and, given the risk for periodic breathing during sleep or direct sleep-disruptive effects of hypoxia itself, there remains a high likelihood that some degree of sleep fragmentation was present in these models. Given this limitation of the current studies to dissect out the independent contribution of intermittent nocturnal hypoxia, effects on cognition independent of sleep fragmentation remain to be completely defined. Contrary to the findings in animal models, studies in patients with sleep apnea, however, do suggest that the role of sleep fragmentation may overwhelm the effects of intermittent hypoxia.⁹ Studies in humans evaluating changes in cognitive function in response to exposure to nocturnal hypoxia with minimal sleep fragmentation simply have not been completed.

The primary purpose of this study was to evaluate the impact of nocturnal hypoxia exposure on daytime vigilance and working memory in normal human volunteers. We hypothesized that recurrent nocturnal hypoxic exposure would result in a significant decline in daytime vigilance and working memory performance. Our exposure utilized a normobaric hypoxic exposure through an “altitude tent” (FiO₂ = 0.13) during sleep on 14 consecutive nights. We assessed sleep quality and respiration during sleep utilizing attended 16-channel polysomnography. Attention and working memory were evaluated in the preexposure baseline condition, at the midpoint during, and after 14 nights of hypoxic exposure. Contrary to our expectations, we found no significant change in cognition and subjective sleepiness that, in clinical equivalents of sleep hypoxia in patients with sleep apnea, would reasonably be expected to adversely impact cognition.

Methods

Study Subjects

Eleven healthy, nonsmoking, subjects (7 men, 4 women) completed the study. Screening procedures included a detailed sleep history and a specific enquiry regarding prior exposure to altitude. The subjects were selected to not have delayed or advanced habitual sleep times (2200–2300 hours to 0600–0700 hours.), unrefreshing sleep, habitual loud snoring, daytime napping, restless legs, anxiety, depression, past or current drug abuse, or active medical conditions such as diabetes and hypertension. All women began exposure during the week following menses, and all tested negative for pregnancy (urinary B-HCG test). The subjects had a mean age of 27 ± 1.5 years a and body mass index of 23 ± 0.9 kg/m². All subjects underwent a history and physical examination to exclude cardiopulmonary or neurologic disease prior to giving written informed consent. This protocol was reviewed and approved by the Institutional Review Board at the Beth Israel Deaconess Medical Center, Boston, Massachusetts.

Hypoxic Exposure

Subjects were exposed to 9 hours of continuous mildly hypocapnic hypoxia from 2200 to 0700 hours for 14 consecutive nights. Subjects underwent acclimatization to the hypoxic exposure with graded increases in “altitude” over 3 nights. Altitude levels started at sea level, followed by 1 night at 7700 feet, 1 night at 10,000 feet, and then 14 consecutive days at 13,000 feet. Exposure was considered to begin (night 1) on the first night at 13,000 feet. Independent evaluation of the FiO₂ at this altitude setting using an oxygen sensor showed the FiO₂ to be 0.13 with the tent system set at 13,000 feet (Gasman Personal Gas Detector, Crowcon, Charleston, South Carolina, USA). During the day, subjects were at sea level.

The hypoxic exposure was achieved using a commercially available normobaric “altitude tent” (Colorado Altitude Training, Colorado Springs). Subjects slept in a standard hospital bed while in the tent, which measures 9 × 7 × 6 feet. Altitude was set and continuously monitored using a central controller with real-time output. Altitude and CO₂ levels within the tent were monitored continuously throughout the exposure. CO₂ was removed using soda-phosphate crystals with a fan-driven system within the tent to allow continuous passage of tent gas across the system to allow stable CO₂ levels to be maintained. Independent verification of CO₂ levels, as performed by an automatic CO₂ monitoring system (Realterm, Colorado Altitude Training, Colorado Springs), yielded 0.4% mm Hg as an average value during the night (Range: 0.1%–0.52%). Oxygen saturation was monitored continuously overnight, and mean O₂ saturations were 84% during the exposure.

Polysomnography

Standard polysomnography using an Embla (Embla, Denver, Colorado, USA) system included recording the electroencephalogram, electrooculogram, chin and tibialis electromyogram, thermistor, nasal pressure, thoracic and abdominal effort, electrocardiogram, and finger oxymetry at baseline and 3 points during the course of the study (night 3, night 7, and night 14). Standard stages (rapid eye movements [REM] and stage 1 to 4 non-REM [NREM] sleep) and 3-second electroencephalogram alpha arousals (American Academy of Sleep Medicine arousals)¹⁰ were scored.

Respiratory Event Scoring

Modified standard criteria¹¹ were used to score abnormal respiration due to the following reasons: (1) during the hypoxic exposure, desaturations were no longer as useful a method to determine the physiologic significance of a respiratory abnormality, (2) respiratory abnormalities were expected to be predominantly nonobstructive in nature, and (3) standard arousals were not always associated with hypopnea termination due to altitude-induced periodic breathing.

Obstructive apnea was scored when there was an absence of airflow for greater than 10 seconds on the nasal cannula and thermistor with continued respiratory effort. Central apnea was scored when there was an absence of airflow on the nasal cannula and thermistor for greater than 10 seconds with no evidence of respiratory effort. Hypopneas with flow-limitation were identified when there was a sequence of progressive flow-limited breaths, the entire abnormality lasting at least 10 seconds, which terminated in an abrupt sinusoidal recovery breath. Hypopneas without flow-limitation were identified when there was no progressive flow limitation but a clearly evident (approximately 30%) reduction in airflow and respiratory effort followed by a recovery in amplitude of both signals. Hypopneas were thus scored in the following circumstances: when there was an associated 3% oxygen desaturation, a 3-second American Academy of Sleep Medicine electroencephalogram alpha/beta arousal, or a major (30%-50%) reduction in signal amplitude followed by sinusoidal recovery breath.

Periodic breathing time was scored separately as a measure of altitude-induced respiratory change. The conventional approach to score periodic breathing in research sleep studies requires 10 minutes of continuous waxing and waning respiration.¹¹ We relaxed this duration requirement to 6 cycles or 2 continuous minutes to capture shorter periods of periodic breathing, expressed as a percentage of total sleep time. Central apneas occurring within periodic breathing sequences or continuous periods of central apneas were identically designated—for estimating the duration of periodic breathing, sequences with mixed periodic breathing and central apneas were considered “periodic breathing time.”

An intermittent oxygen desaturation index (ODI, dips per hour of sleep) was calculated using a 3% criterion. The desaturation criterion required a 3% drop from the preceding 20-second stable baseline—thus a drop from 88% to 85% would be scored but not an 88% to 86% change. Thus, this measure was different from the traditional ODI, which uses a at least a 4% decrease in SaO₂ (at nadir) with an increase to within 1% of the prehypopnea baseline value,¹² in that (1) the threshold was different and (2) progressive and severe desaturations may not be identified if the preevent baseline is itself abnormal (as can be at altitude). As global measures of nocturnal hypoxic exposure, the following were computed: total sleep time at less than 90% saturation and less than 80% saturation.

All scoring was done manually by 1 registered polysomnographic technician (JB) with more than 30 years of experience in sleep medicine. The scoring rules were developed by 1 of the authors (RJT), who manually checked the scoring, epoch by epoch. Rescoring was done, as necessary, by RJT.

Sleepiness, Attention and Working Memory Assessments

Subjects practiced the tests until performance (reaction time, percentage correct) were stable, usually about 20 minutes for the working memory task and 1 practice run of the Psychomotor Vigilance Task (PVT). Further practice sessions were not permitted. Assessments were performed at baseline and repeated at 4 time points during the protocol (day 4, 8, 9, 15). To minimize circadian effects on performance, testing was done between 0800 and 1000 hours. A 10-minute PVT was used as a measure of attention. This hand-held button-box unit provides a number of variables, such as mean and slowest reaction times, lapses, and false responses. The working memory assessment used a 10-minute verbal 2-back task. In this task, the subject has to keep 2 letters of the alphabet in short-term memory at any given point in time, update this representation, compare it with the next letters of the alphabet, and make the required target or nontarget response. For example, in the stream g-b-j-g-j-x, j is a 2-back. One alphabet was presented to the subject every 4 seconds, with a requirement for rapid responses while maintaining accuracy. Percentage correct and mean response times were computed. Two 10-minute sets were presented consecutively with a 1-minute break. This provided 300 responses per testing point, and mean value of all responses provided a single performance speed for tabulation and computation. An integrated assessment of subjective sleepiness and mood was assessed on a scale of 0 (very sleepy compared to usual state, depressed mood) to 10 (no change from usual state, normal mood state). This measure was obtained from the Performance Vigilance Task device, immediately before and following the 10-minute PVT, and averaged.

Statistical Methods

To calculate sample size, we used the number of lapses on the PVT (reaction times ≥ 500 milliseconds) as the primary variable of interest. Lapses are a sensitive biomarker of increased homeostatic sleep drive in the context of conditions causing excessive sleepiness, including sleep deprivation or sleep fragmentation, and are quite rare in rested individuals.¹³ Our sample size was powered (0.9) to detect a change in mean lapse frequency of 2 ± 2 at baseline to 6 ± 6 after exposure.

Means and standard deviations were computed for polysomnographic and cognitive variables. Analysis of variance was used to assess sleep and cognition, with Tukey multiple comparisons to assess significant group differences. Respiratory variables were analyzed using the Kruskal-Wallis test because the data were not normally distributed. Comparison between protocol time points used the Wilcoxon rank sum test, e.g., midpoint versus end of protocol. The intraclass correlation coefficient, (ICC) was used to assess contribution from interindividual differences in susceptibility to altitude-induced periodic breathing and attention/working memory performance. STATA SE8 (StataCorp LP, College Station, Texas, USA) was used for analysis.

Results

Sleep Quality During Sleep Hypoxia

Total sleep time was significantly decreased during exposure. There was a mean reduction of 45 minutes by the end of the early acclimatization period, and this was the only value (change from baseline) that was significant after adjusting for multiple comparisons. This initial decrease was followed by a small mean increase (but not statistically significant) by the midpoint and end of the exposure but remained below baseline values by 32 minutes. The other significant differences across the protocol were in stage 1 sleep (baseline to early acclimatization only) and the American Academy of Sleep Medicine arousal index (significantly different only from baseline to early acclimatization). There was no significant difference in sleep efficiency, slow-wave sleep, or REM sleep measures across the study duration (Table 1).

Table 1.

Polysomnography at Simulated Altitude

Variable	Preexposure	Early- acclimatized	Midexposure	Postexposure	Statistical significance
					F(3,40)	P
TSTa	452.9±20.8	407.5±58.1	420.5±38.4	420.9±20	3.05	0.039
SE	93.6±3.6	86.1±9.1	89.5± 8.5	89.3±3.8	2.42	0.079
Stage 1b	10.7±3.2	16.2±6.2	12.3±4.3	14.3± 4.6	3.05	0.039
Stage 2	59.6 ± 6	55.5 ± 7.3	57.4 ± 5.3	56.8 ± 5.6	0.97	0.42
Stage 3	5.1±3.7	5.4±3.5	5±4.1	4.6±2.6	0.09	0.97
Stage 4	2.2±4.5	0.4±0.9	0.2±0.5	0.12±0.2	1.92	0.14
REMS	23.1±4.4	22.8±6.1	25.1±4.4	22.2±6.4	0.66	0.58
REML	81.5±31.4	88.5±35.2	83±48.7	66.3±17	0.76	0.53
ArIb	22.9±7	36.7±11.8	32.8±12.1	31.7±13.4	3.22	0.032
PLMI	0.08±0.2	0.05±0.12	0.05±0.12	0.1±0.19	0.22	0.89

Data are expressed as mean ± SD. TST refers to total sleep time; SE, sleep efficiency; REMS, percentage of rapid eye movement (REM) sleep; REML, REM sleep-onset latency, ArI, American Academy of Sleep Medicine arousal index; PLMI, periodic leg movement index.

^aSignificant baseline vs other time points.

^bSignificant only for baseline vs early acclimatized stage.

Sleep Hypoxia

The model successfully induced nocturnal hypoxia, as summarized in Table 2. Superimposed intermittent saturation fluctuations occurred throughout the night in unstable/light stage 2 NREM sleep. There were no significant differences at the 3 altitude exposure time points, averaging 24 to 29 episodes of 3% desaturations per hour of sleep.

Table 2.

Respiration at Simulated Altitude

Variable	Preexposure	Early-acclimatized	Midexposure	Postexposure	Kruskal-Wallis		ICCa
					χ2 3 df	P
CAI	0.4±0.6	15.7±22.5	11.7±15.4	15.2±20.6	20.96	< 0.001	0.53
OAI	0.1±0.2	0.2±0.3	0.1±0.1	0.2±0.3	3.021	0.39	0
HI	6.3±6.7	26.8±17.7	25.2±15.9	22.8±21	13.53	0.004	0.76
HFLI	6.3±6.7	15.4±12.1	14.1±9.9	12.4±14.4	19.57	< 0.001	0.27
RDI	7.1±7.7	42.7±32.7	36.9±25.9	38.3±34.7	14.50	0.002	0.67
PBT % TST	0	29±22	26.8±20.2	28±24.9	27.6	< 0.001	0.66
ODI	1.2±0.8	26.4±29.3	24.3±19.6	29.3±27.3	25.58	< 0.001	0.57
Min SaO2	90.6±1.5	73.5±3.4	75.8±2.9	76.7±2.7	30.08	< 0.001	0
% TST < Sa 90	0.1±0.4	87.4±24.7	92.4±8.7	95.4±5.2	28.10	< 0.001	0
% TST < Sa 80	0	13.7±16.5b	4.6±7.4	2.2±2.9	20.46	< 0.001	0

Data are expressed as mean ± SD. df refers to degrees of freedom; ICC, intraclass correlation coefficients the CAI central apnea index; OA, obstructive apnea index, HI: hypopnea index without flow limitation; HFL, hypopnea index with flow limitation; RDI, respiratory disturbance index; PBT, periodic breathing time as a percentage of total sleep time (TST); ODI = oxygen 3% desaturation index; MinSaO2, minimum nocturnal saturation; % TST Sa, percentage of TST with saturations less than designated threshold.

^aAltitude exposure only.

^bSignificant differences from other hypoxia exposure timepoints. For the baseline study, separate scoring of with / without flow limitation was not done because all events were assumed to be obstructive.

Respiration During Sleep Hypoxia

Altitude exposure resulted in a sustained increase in central apneas, periodic breathing, and hypopneas with and without flow limitation (Table 3). Although a trend of reduced severity may be suggested across the course of the experiment, these did not reach statistical significance. Considering all the 3 timepoints obtained during hypoxia, the respiratory disturbance index (RDI, apneas + hypopneas with and without flow limitation, per hour of sleep) was significantly greater than the ODI (39.37 ± 30 vs. 26.48 ± 24.94, paired t test t = 5.072, P < 0.001) but not the arousal index (39.37 ± 30 vs. 33.86 ± 12.20, paired t test t = 1.47, P = 0.15). The arousal index was significantly higher than the ODI (33.86 ± 12.20 vs. 26.48 ± 24.94, paired t test t = 2.27, P = 0.03). There were no significant differences in respiratory variables under hypoxia, using pairs of timepoints and the rank sum test. Nearly 90% of total sleep time during exposure was spent in saturations less than 90%. Time spent under 80% was significantly increased only at the early acclimatization point (13.7% ± 16.5 %).

Table 3.

Attention and Working Memory at Simulated Altitude

Variable	Pre-hypoxia	Day 3	Day 8	Day 9	Day 15	Statistical significance		Blocked ANOVA(subject)		ICC
						F4,50	P	F4,40	P
Mean RT	259±33	256±45	255±41	250±33	254±35	0.10	0.98	0.62	0.65	0.85
Median RT	246±29	242±37	242±37	236±30	243±33	0.13	0.97	0.97	0.43	0.87
Slowest RT	391±83	388±131	377±91	389±85	373±77	0.08	0.99	0.25	0.91	0.70
False starts	0.6±0.9	1.4±1.4	0.6±1.3	0.7±1.1	0.4±0.7	1.22	0.32	1.37	0.26	0.10
Lapses	0.8±0.98	1.9±4.1	1.3±1.5	1.1±1.2	0.7±1.4	0.51	0.73	0.96	0.43	0.47
Mood	8.2±1.4	7.1±2	7±2.1	6.7±1.7	8.1±1.5	1.66	0.17	3.53	0.02	0.47
2-back mean RT	975±266	936±216	965±274	896±264	894±245	0.24	0.91	2.63	0.05	0.90
2-back slowest RT	1462±582	1265±439	1668±963	1528±945	1364±702	0.46	0.77	1.83	0.14	0.73
2-back fastest RT	797±222	725±170	780±268	745±237	722±218	0.24	0.92	1.45	0.24	0.83
2-back % correct	92.4±2.2	94.1±3.2	94.2±1.5	93.1±1.9	93.9±1.6	1.23	0.31	1.38	0.26	0.10

ANOVA refers to analysis of variance; ICC, intraclass correlation coefficient; RT, mean reaction time in milliseconds; Mood, visual analogue scale composite of subjective alertness and mood (average of 2 assessments immediately before and after the 10-minute Psychomotor Vigilance Task).

ICC: Intra class coefficient

Subjective Sleepiness and Cognitive Performance

There were no significant differences from the baseline across the period of exposure in any of the following variables: subjective assessment of state (a composite of mood and sleepiness), 2-back mean reaction times, percentage of correct and slowest and fastest 10% of responses, and the following variables from the PVT: mean, median, and 10% slowest reaction times, false starts, or lapses (Table 4).

Individual Differences in Altitude-Induced Periodic Breathing

There was evidence of individual differences in susceptibility to periodic breathing at altitude. The ICCs were high for central apnea index (0.53), hypopnea index (0.76), RDI (0.67), periodic breathing time (0.66), and ODI (0.57). The ICC was not significant for the percentage of total sleep time below 90%. Using an RDI cut-off of 20 at the end of the exposure, we were able to categorize those who were sensitive (n = 6) or relatively resistant (n = 5) to altitude-induced sleep-disordered breathing. At the 2-week point, the resistant versus sensitive groups were clearly separated in terms of central apnea index (1.8 ± 1.2 vs. 28.6 ± 22.5, P = 0.03), the periodic breathing time (percentage of total sleep time, 7% ± 3.1% vs. 49% ± 7.6%, P = 0.001), the RDI (9.3 ± 6.9 vs. 67.1 ± 24.2, P = 0.001), ODI (6.6 ± 3.3 vs. 51.9 ± 19.6, P < 0.001), and the arousal index (21.4 ± 4.2 vs. 42.1 ± 11, P = 0.004). The percentage of total sleep time spent below 90% oxygen saturation was not different (94.5% ± 5% vs. 96.3% ± 5.7%, P = 0.6). Sleep architecture by standard scoring remained unchanged (other than the arousal index, above)—total sleep time (422.4 ± 27.8 vs. 419.4 ± 10.9 minutes, P = 0.8), stage 1 as a percentage of total sleep time (12.6 ± 2.6 vs. 16 ± 5.8, P = 0.26), slow-wave sleep as a percentage of total sleep time (4.2 ± 3.5 vs. 5.2 ± 1.6, P = 0.6), REM sleep latency (63.4 ± 16.5 vs. 69.2 ± 18.9 minutes, P = 0.62), and REM sleep as a percentage of total sleep time (20.8 ± 7.9 vs. 23.6 ± 4.8, P = 0.52).

Individual Differences in Working Memory and Attention

The ICC was high for most measures of cognition (Table 3) other than false starts and 2-back percentage correct. The former were rare, and the latter was minimized by training and encouraging accuracy first and then speed. To explore if individual differences concealed performance decrements, a randomized-block analysis of variance was also done with subject as a blocking factor—in this analysis (Table 3), significant differences were noted in midexposure subjective sleepiness-mood (reduced) and 2-back mean reaction times (faster on the final testing). There was no relationship of working memory performance to the amount of periodic breathing.

Discussion

The major findings from this study are (1) normal human subjects exposed to continuous nocturnal hypoxia in an altitude tent showed no significant changes in alertness, objective vigilance, or working memory; (2) hypoxic exposure of this severity induced significant new obstructive respiratory events in addition to the centrally mediated periodic breathing that is well characterized in the setting of hypoxia; and (3) individual differences in attention/working memory and susceptibility to periodic breathing were also demonstrated.

Hypoxia has well-described adverse effects on cognitive brain function.^14,15 Our subjects, following hypoxia exposure in an altitude tent, demonstrated no subjective change in alertness or objective vigilance or working memory impairment. During the period of exposure, sleep and ventilatory acclimatization to hypoxia occurred, but hypoxia remained significant and in the range typically associated with severe clinical symptoms when induced primarily by sleep-disordered breathing. This was not the result expected at the start of the study and raises important questions regarding the role of nocturnal hypoxia in the cognitive neurobiology of sleep-disordered breathing. Altitude exposure, as expected, induced a marked increase in respiratory abnormalities. Use of the nasal cannula-pressure transducer system allowed the scoring of a full spectrum of induced respiratory abnormality. Pure central apneas were slightly less than a third of all induced respiratory events. About a third of the new hypopneas demonstrated prominent flow-limitation—we speculate that this may reflect a role for periodic breathing mechanisms or hypoxemia in the pathogenesis of obstructive upper airway disease, but this study was not intended to directly address these mechanisms. The ODI was less than the RDI, as respiratory control instability resulted in short cycles of periodic breathing that did not always cause a further 3% reduction in oxygen saturation.

Strong individual differences were seen in the severity of altitude-induced period breathing (high ICCs for central apneas, RDI, and flow-limitation events). Variable sensitivity to altitude-induced periodic breathing or susceptibility to periodic breathing using proportional assist ventilation in healthy individuals is well known.^16–18 Thus, the ICC was high for central and total respiratory events, the time in periodic breathing, but not for the percentage of total sleep time with saturations less than 90%. Moreover, there was no significant difference in total sleep time, sleep stages, REM latency, or hypoxia exposure between more and less sensitive individuals, whereas respiratory indices and arousals were significantly different. The difference in arousal indices was less striking that that in respiratory indices. In other words, relative sensitivity or resistance to altitude-induced periodic breathing was expressed predominantly as periodic breathing related features rather than sleep quality, other than a doubling of the arousal index.

The evidence that nocturnal hypoxia causes cognitive deficits in humans is largely correlative.^19,20 Sleep fragmentation and hypoxia indices can be statistically linked to impaired attention, affective symptoms, and executive dysfunction.²¹ In the clinical model, these are necessarily linked and often inseparable. Patients with severe hypoxia typically have severe sleep fragmentation, except those who have a dominant component of hypoventilation, in whom desaturation may exceed sleep fragmentation. Hypoxia induces periodic breathing and thus amplifies the respiratory abnormality. However, hypoxia is not necessary to induce cognitive dysfunction, as in the experimental models of auditory sleep fragmentation and in those with non-hypoxic obstructive sleep-disordered breathing.^2,22,23 In those with mixed obstructive and periodic breathing, desaturations are relatively mild to moderate in the presence of severe sleepiness. Limitations of the method of scoring respiratory abnormalities, typically event counts, may not adequately reflect the impact of abnormal breathing on sleep and further limit the accuracy of associations. This is especially problematic in the pediatric age group in which thermistors are still in wide use, and discrete events and typical adult type arousals are less common, forcing a fall back on desaturation indices.

The best human clinical evidence that hypoxia may be directly linked to cognitive dysfunction comes from the effects of supplemental oxygen in chronic obstructive lung disease, but hypoxia in this condition is not exclusively sleep related, and there may be independent effects of hypercarbia.^24–26 There is an extensive literature describing the effects of simulated and real altitude on cognition and brain function (reviewed in¹⁵). Effects are seen on psychomotor performance, perception, learning, memory, language especially verbal fluency, and cognitive flexibility. An increase in reaction time and latency of P300 are observed. Reduced thresholds of touch, smell, pain, and taste, together with somesthetic illusions and visual hallucinations, have been reported at extreme exposures. Impairment in short-term memory is especially noticeable above 6000 meters. Alterations in accuracy and motor speed are seen at lower altitudes. Alterations in cerebral perfusion induced by hypocapnia may play a role in this instance.²⁷ Sleep fragmentation induced by sleep-disordered breathing may also be a mediator, as those with the greatest cognitive deficits also have increased hypoxia responsiveness, a risk factor for periodic breathing at altitude.²⁸ One important difference in our model that more closely approximates sleep apnea is that subjects had hypoxic exposure only during the sleep period. This may have allowed recovery processes to be activated if necessary during the wake period, preventing cumulative effects.

A murine model of intermittent hypoxia during sleep has demonstrated executive dysfunction, excessive sleepiness, and oxidative injury to basal forebrain structures.^8,29 Prior models of intermittent hypoxia have demonstrated executive and learning dysfunction and hippocampal injury.^13,30,31 The murine model may not be a pure form of sleep hypoxia, as sleep fragmentation can be induced by hypoxia in rats.²³ However, the model used in this manuscript was not a model of intermittent hypoxia but, rather of chronic exposure with acclimatization, which is quite different, and the presented results in relation to murine experiments are not directly comparable. Species differences and developmental susceptibility cannot be excluded, and it is possible that humans are more resistant to hypoxia-induced functional or structural brain injury. The results using brain morphometric techniques are mixed, with both extensive and limited hippocampal signal reductions in hypoxic sleep-disordered breathing.³² No information on similarly symptomatic nonhypoxic human disease is yet available.

Use of the nasal cannula-pressure transducer technique³³ and scoring respiratory events with and without evidence of flow limitation have resulted in potentially interesting insights into respiratory control and its interactions with the upper airway. To our best knowledge, this study is the first to use nasal cannula-based scoring for measuring altitude/hypoxia-induced respiratory abnormality and showed that nearly half of the new abnormality was associated with overt flow limitation. These events vastly outnumbered central apneas at all points of the protocol. The role of periodic breathing and respiratory-control dysfunction in the induction or amplification of obstructive sleep-disordered breathing has been intermittently proposed over the years.^34,35 Our results support this role—that of hypoxia-induced periodic breathing directly causing upper airway obstruction. Hypocapnia has an important role in the pathogenesis of altitude-induced respiratory dysfunction¹⁷; we did not directly measure end-tidal or transcutaneous CO₂ in our study. Some of the previously described statistical associations of hypoxia with cognitive dysfunction in sleep apnea may have been mediated through worsening of sleep fragmentation through this mechanism. None of the previously published studies associating cognitive outcomes with nocturnal hypoxia have used the nasal cannula–pressure transducer technique to assess airflow, which may have resulted in underestimates of respiratory abnormality.

Our model has some limitations. Our sample size was small and does not allow exploration of effects based on age, sex, periodic breathing sensitivity, and baseline attention/working memory performance characteristics. The small sample size also necessitates the presented results to be considered preliminary. We did not perform a Multiple Sleep Latency Test. The tests we used may not have been sensitive enough. Subjective assessments of sleepiness and mood following sleep deprivation are progressively inaccurate over time even if performance is deteriorating, but the complete absence of subjective change argues against this explanation. The PVT used has been shown in numerous publications to be sensitive to the effects of excessive sleepiness.¹³ We used the 2-back working memory task to assess the effects of a single night of sleep deprivation and auditory sleep fragmentation and cognitive function in sleep apnea and narcolepsy, with consistently slowing of performance.^36,37 Standard deviations on the 2-back task were wide, but this is explained by stable baseline differences between individuals, a known characteristic of working-memory performance. The ICCs for the 2-back variables were all greater than 0.8. When using subjects as a blocking factor in a randomized block analysis of variance, borderline significant changes in mean reaction time was noted, but in the direction of faster responses. It is thus unlikely that a major component of sleepiness-induced brain function was not sampled by our methods. More subtle effects however cannot be excluded.

The degree and duration of hypoxia may have been insufficient. Although patients with severe sleep-disordered breathing can have greater severities of nocturnal hypoxia, our exposure was not insignificant and takes it to the limit that we believe may be safely induced in healthy individuals. The majority of patients seen in clinical practice today with symptomatic disease would fall within these limits. Because sleep restriction and a single night of sleep deprivation or fragmentation induce sleepiness,^13,38,39 the complete absence of effect after 2 weeks of hypoxia exposure is certainly not definitive but is suggestive of a lack of hypoxia effects in the range and mode tested. We now have exposed 8 individuals to 4 weeks of intermittent hypoxia (20 desaturations per hour, from greater than 90% to 83%-85%), with no detectable change on a 20-minute PVT, 20-minute 2- and 3-back working memory task, the Rey Auditory Verbal Learning Test, and the Multiple Sleep Latency Test (unpublished observations). These results challenge the role of moderate degrees of hypoxia as the primary mediator of cognitive dysfunction in those without extremely severe hypoxic sleep apnea syndromes.

Our model did not exactly mimic that experienced by patients with sleep-disordered breathing, which is dominated by intermittent hypoxia. Experimental models have shown distinct differences in the biologic effects of continuous and intermittent hypoxia. The altitude tent produced a mix of continuous and intermittent hypoxia, without full reoxygenation. Sleeping at altitude induces a baseline level of hypoxia on which is superimposed further intermittent drops during periods of periodic breathing. Acclimatization effects also occur at altitude, with a reduction in periodic breathing, although in our sample the difference between the first and last night of full hypoxia were not significant. The effect of the increase in inspired CO₂ that likely reduced sleep hypocapnia (this was not directly measured) is not known, nor is the change in 2, 3 DPG. Subjectively, our subjects slept relatively well in the tent, contrary to the usual experience at altitude.¹⁷ The sleep architecture under hypoxia showed some expected differences, with an increase in stage 1 sleep and reduction in total sleep times but no reduction in REM sleep and no change in REM sleep latency or slow-wave sleep. The arousal index was increased to a lesser extent than the respiratory abnormality, providing evidence of some dissociation of respiratory and sleep effects of hypoxia. Our model thus had less sleep fragmentation and was a purer model of nocturnal hypoxia, compared with actual altitude exposure, but the reduced total sleep time and increased stage 1 sleep show that a better model is needed. In the altitude tent and also under hypoxic conditions in general, REM sleep is relatively undisturbed and most of the scored RDI is contributed by intermittent periods of unstable NREM sleep. These periods of relatively undisturbed sleep may have allowed adequate homeostatic sleep functions. The tent environment itself may have contributed to this unplanned benefit, as the average CO₂ level in the tent, maintained by the gas scrubber, was 0.3% to 0.4%. We did not measure end-tidal CO₂ in our subjects, but hypocapnia and periodic breathing was probably less than would have occurred if the CO₂ levels were 0.03% to 0.04%.

In conclusion, we report the absence of significant subjective sleepiness or objective vigilance and working-memory impairment following hypoxia exposure at simulated altitude. Although not a perfect replication of what occurs in patients with sleep-disordered breathing, the results do not support a primary role of sleep hypoxia in inducing cognitive dysfunction in the average clinic patient. In clinical practice, nocturnal desaturation is likely still important, as hypoxia-induced periodic breathing contributes to sleep fragmentation. Caution is recommended in the extrapolation of the effects of intermittent hypoxia in animal models to human sleep apnea. The dominant clinical focus on hypoxia (e.g., the American Academy of Sleep Medicine hypopnea criteria that have been adopted by the field and health care policies) may not be biologically optimized, and comparative assessments of sleep fragmentation and hypoxia deserve further study. Models of nocturnal hypoxia free of periodic breathing are required to assess the effects of sleep hypoxia on human cognitive function.

Figure 1.

Hypoxia with periodic breathing, unstable non-rapid eye movement (NREM) sleep. A 25-year old woman after 2 weeks of exposure to 13,000 feet altitude for 2 weeks shows short cycle periodic breathing (20–25 seconds) in this 120-second snapshot from Stage 2 sleep. Saturations fluctuate between 89% and 81%. Small cardiogenic oscillations can be seen on the flow and effort traces, suggesting that the airway was open during the apnea. The traces from the top of the figure are electroencephalogram (EEG) channels C4-A1, C3-A2, O2-A1, O1-A2 (50 μV between lighter horizontal lines), right and left electrooculogram (EOG), mentalis electromyogram (EMG), nasal pressure (Pres), computed flow from the pressure signal (Flow), thoracic (Thor) and abdominal (Abd) effort with piezo-bands, electrocardiogram (ECG), joined (right and left) tibialis anterior electromyogram (Tib), and finger pulse oximetry (Ox).

Figure 2.

Hypoxia without periodic breathing, stable non-rapid eye movement (NREM) sleep. The same subject, traces, and screen compression as in Figure 1 during stable stage 2 NREM sleep, showing an absence of respiratory events. There is minimal flow limitation and no arousals; oxygen saturation is 84%-83%. The traces from the top of the figure are electroencephalogram (EEG) channels C4-A1, C3-A2, O2-A1, O1-A2 (50 μV between lighter horizontal lines), right and left electrooculogram (EOG), mentalis electromyogram (EMG), nasal pressure (Pres), computed flow from the pressure signal (Flow), thoracic (Thor) and abdominal (Abd) effort with piezo-bands, electrocardiogram (ECG), joined (right and left) tibialis anterior electromyogram (Tib), and finger pulse oximetry (Ox).

Figure 3.

Hypoxia without periodic breathing, rapid eye movement (REM) sleep. The same subject, traces, and screen compression as in Figure 1 during REM sleep, showing an absence of respiratory events. Altitude-induced periodic breathing is minimal or mildest during REM sleep. There is minimal flow limitation, especially during phasic REM sleep and no arousals; oxygen saturation fluctuates from 87% to 83%, probably secondary to tidal volume changes typical of REM sleep. The traces from the top of the figure are electroencephalogram (EEG) channels C4-A1, C3-A2, O2-A1, O1-A2 (50 μV between lighter horizontal lines), right and left electrooculogram (EOG), mentalis electromyogram (EMG), nasal pressure (Pres), computed flow from the pressure signal (Flow), thoracic (Thor) and abdominal (Abd) effort with piezo-bands, electrocardiogram (ECG), joined (right and left) tibialis anterior electromyogram (Tib), and finger pulse oximetry (Ox).