Improving adaptation in male high-altitude workers through pressurization: a randomized controlled trial on sleep architecture

Abstract

Background

Sleep disruption at high altitude, driven by hypobaric hypoxia, is prevalent among occupational cohorts and adversely affects health, safety, and cognitive performance. While oxygen supplementation improves nocturnal hypoxia, it does not replicate the integrated physiological stimulus of true altitude descent. This study investigated the effects of simulated, stepwise altitude descent via nocturnal pressurization on sleep in acclimatized workers residing at high altitude.

Methods

In a 5-week randomized controlled trial, 24 male workers residing at 3500 m were allocated to a Pressurization group (n = 12) or a Control group (n = 12). The Pressurization group slept in a pressurized environmental simulator at sequentially lower simulated altitudes (3000 m, 2500 m, then 2000 m) over three weeks, while both groups worked at 3500 m during the day. The Control group remained at ambient 3500 m throughout. Sleep architecture and nocturnal oxygen saturation (SpO₂) were monitored weekly. Linear mixed-effects models assessed the effects of Group, Time, and their interaction.

Results

A significant Group × Time interaction was observed for nocturnal oxygenation (Mean SpO₂: F(4,88) = 28.36, p < 0.001; Minimum SpO₂: F(4,88) = 7.16, p < 0.001). The Pressurization group showed dose-dependent increases in Mean and Minimum SpO₂, peaking at the 2000 m simulated altitude. Concurrently, the Pressurization group exhibited a significant increase in deep sleep percentage at 2000 m (Group × Time effect: 3.25%, p = 0.010) and a marked reduction in the arousal index (ArI) at 2500 m and 2000 m (− 5.42 and − 7.01 events/h, respectively, p < 0.001). Epworth Sleepiness Scale scores also decreased significantly in the Pressurization group during descent (p < 0.01). No significant changes were observed in total sleep time, Rapid Eye Movement sleep, or wake time percentage.

Conclusion

Nocturnal pressurization simulating stepwise descent to lower altitudes significantly improves nocturnal oxygenation, consolidates deep sleep, reduces sleep fragmentation, and alleviates daytime sleepiness in high-altitude workers. This intervention, which mimics the combined barometric and hypoxic stimulus of true descent, presents a distinct and promising non-pharmacological strategy for managing sleep disturbances in occupational settings at high altitude.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s00420-026-02229-z.

Introduction

Sleep impairment upon ascent to high altitude is a nearly universal phenomenon, with subjective complaints of insomnia, frequent awakenings, and unrefreshing sleep emerging from altitudes as low as 2500 m (Venkat et al. 2021; Patrician et al. 2024). Accumulating evidence from systematic reviews indicates that acute exposure to hypoxia significantly alters sleep architecture, manifested as increased N1 percentage, decreased N3 percentage, decreased REM sleep, and increased nocturnal arousals (Aquino-Lemos et al. 2016; Guo et al. 2024). These disturbances are rooted in profound physiological alterations driven by hypobaric hypoxia, which destabilizes respiratory control, leading to periodic breathing (PB), reduces nocturnal oxygen saturation, and fragments sleep architecture—notably through a marked reduction in N3 percentage and REM sleep (Wickramasinghe and Anholm 1999; Weil 2004). Such sleep disruption is not merely a transient inconvenience; it is closely associated with anxiety (Dong et al. 2013; Bian et al. 2019), impairs cognitive function (Frost et al. 2021), and negatively impacts the long-term physical and mental health of high-altitude residents (Guan et al. 2015; Tan et al. 2024). With approximately 70% of individuals experiencing Sleep-disordered Breathing (SDB) at high altitude (Yang et al. 2025), effective interventions are a critical unmet need.

Current strategies to mitigate high-altitude sleep disruption include pharmacotherapy (e.g., acetazolamide) and diffuse oxygen supplementation (Weil 2004; Rojas-Córdova et al. 2023; Patrician et al. 2024). Oxygen enrichment, typically administered via nasal cannula, has been shown to improve nocturnal oxygenation and reduce central apneas in miners at 4200 m (Moraga et al. 2014). Pressurization can similarly elevate arterial oxygenation by increasing ambient barometric pressure, but through a different mechanism—raising partial pressure (PiO₂) via Dalton’s law rather than directly increasing Fraction of inspired oxygen (FiO₂) (Mengjiang Ding 2016). At 3500 m, the direct impact of hypobaria on sleep is minor compared to the disruptive effects of hypoxia itself (Heinzer et al. 2016). Moreover, while oxygen supplementation is relatively low-cost and easy to implement, pressurization technologies face substantial practical barriers, including high equipment costs, limited portability, and operational complexity. Importantly, oxygen supplementation does not alter ambient barometric pressure, whereas pressurization can simulate stepwise descent by increasing both PiO₂ and barometric pressure—a combination that may more holistically influence respiratory drive and acid–base balance (Heinzer et al. 2016; Moraga et al. 2018). The application of pressurization to simulate stepwise descent as a repeated nocturnal intervention for acclimatized residents remains underexplored, particularly in direct comparison to the outcomes expected from oxygen therapy alone.

In this study, we employed a Randomized Controlled Trial (RCT) with Linear Mixed-effects Modeling (LMM) to evaluate the effects of different simulated altitude descent via nocturnal pressurization on a comprehensive set of sleep parameters in healthy male workers residing at 3500 m. A control group maintained at a constant altitude allowed us to isolate the pressurization effect. We hypothesized that this intervention would significantly improve nocturnal oxygenation, reduce sleep fragmentation, and increase slow-wave sleep, with a pattern of change distinct from that reported for diffuse oxygen supplementation.

Methods

Study design and participants

This RCT enrolled 24 male workers originally from low-altitude regions who had been residing at 3500 m for ≥ 3 months. Participants were randomly allocated using a random number generator to either the Pressurization group (n = 12), residing in a pressurization chamber (Developed by High-Altitude Human Habitat Environment Engineering Research Center, China Construction Advanced Technology Research Institute; Manufactured by Yunju Technology Co., Ltd, China Construction Third Engineering Bureau, Figure S1), or the Control group (n = 12). Key inclusion criteria were age 18–45 years, and normotensive blood pressure. Exclusion criteria included cardiopulmonary disease, endocrine disorders, history of severe altitude illness, or use of medications affecting catecholamine or Renin–Angiotensin–Aldosterone System (RAAS) pathways. The study protocol was approved by the Affiliated Hospital of Chengdu University Ethics Committee [Grant No. PJ2022-045–02], and all participants provided written informed consent. Due to the perceptible nature of the environmental intervention, blinding of participants was not feasible. Sleep assessors were blinded to group allocation. Sample size (n = 12 per group) was selected based on practical recruitment feasibility at the worksite and consistency with previous high-altitude intervention trials (Frost et al. 2021).

Experimental protocol and environmental control

The study comprised a 5-week protocol. Following a baseline week (1st week) under natural conditions at 3500 m, the Pressurization group received a 3-week intervention during which they slept in the environmental simulator at sequentially lower simulated altitudes (2nd week, at 3000 m; 3rd week, at 2500 m; then 4th week, at 2000 m), while maintaining daytime work alongside the Control group at the natural 3500 m worksite (Fig. 1). The Control group remained under consistent natural 3500 m conditions throughout. All participants completed a final recovery week (5th week) at 3500 m. The specific environmental parameters within the simulator for each simulated altitude are detailed in Table 1.

Fig. 1.

Study Flow Diagram

Table 1.

Controlled Environmental Parameters in the Pressurization Chamber

Simulated Altitude	Pressure (kPa)	O₂ Partial Pressure (kPa)	Temperature (°C)	Humidity (%)
3000 m	70	14.21	16.61	57.69
2500 m	75	15.20	18.74	59.77
2000 m	80	16.24	21.50	61.93

Measurements and procedures

Sleep architecture and nocturnal physiology were monitored using a forehead-mounted sleep monitor (UMindSleep Forehead Sleep Monitor, Figure S1, S1). This device utilizes electroencephalographic (EEG) signals for precise sleep staging and generates a comprehensive report including sleep architecture and peripheral oxygen saturation (SpO₂) data. The device uses a built‑in automated algorithm that outputs four states per 30‑second epoch: wake, light sleep, deep sleep, and Rapid Eye Movement (REM) sleep. The algorithm has been validated against polysomnography (Cohen ‘s κ = 0.76 for deep sleep, 0.69 for light sleep) (Chen et al. 2023).

All participants adhered to a fixed sleep‑wake schedule during each monitoring night: lights off at 22:00 and wake‑up at 07:00. On each designated monitoring night (at the end of weeks 1–5), all 24 participants were monitored simultaneously using 24 identical devices (one per participant). Participants were instructed to wear the monitor before bedtime at the end of each experimental week. Recordings were conducted for a minimum of 7 h per night to ensure data capture across a full sleep period. All devices were pre calibrated by the manufacturer and cross checked by the research team before each use. Data quality was verified by a trained researcher who visually inspected each recording for signal integrity and completeness; recordings with > 10% artifact were excluded (none met this threshold). The following key sleep parameters were as follows:

1. Total Sleep Time (TST): The total duration of sleep (in minutes) during the recording period.

2. Sleep Efficiency: The percentage of total recording time spent asleep, calculated as (TST / total recording time) × 100%.

3. Wake Time Percent: The percentage of total sleep time spent in the awake state.

4. Deep Sleep Percent: The percentage of TST spent in deep sleep.

5. Light Sleep Percent: The percentage of TST spent in light sleep.

6. REM Sleep Percent: The percentage of TST spent in REM sleep.

7. Arousal Index (ArI): The number of micro-arousals (abrupt shifts to faster EEG frequency) per hour of sleep.

8. Mean Nocturnal Oxygen Saturation (Mean SpO₂): The average peripheral oxygen saturation (%) across the total sleep period.

9. Minimal Nocturnal Oxygen Saturation (Min SpO₂): The minimal peripheral oxygen saturation (%) across the total sleep period.

Daytime sleepiness was evaluated subjectively using the Epworth Sleepiness Scale (ESS). The ESS was administered at a fixed time (09:00–10:00) on the morning after each weekly monitoring session, under the supervision of a trained research assistant who checked each questionnaire for completeness on‑site. The ESS assesses an individual's likelihood of dozing off or falling asleep in eight common daily situations (e.g., sitting and reading, watching TV). Each item is scored from 0 (would never doze) to 3 (high chance of dozing), yielding a total score ranging from 0 to 24. According to standard interpretation, an ESS score > 6 indicates a tendency toward sleepiness, > 11 indicates significant sleepiness, and > 16 indicates severe sleepiness.

All data were anonymized using participant identification codes and stored securely for subsequent statistical analysis.

Statistical analysis

Data were analyzed using R software (version 4.3.2). LMMs with random intercepts for participants were fitted to evaluate the effects of Group, Time, and their interaction on each outcome. Time was treated as a categorical factor with baseline (3500 m) as the reference. Model fit was assessed via AIC/BIC. The significance of fixed effects was examined using ANOVA. A p-value < 0.05 was considered statistically significant.

Results

Baseline participant characteristics

The study included 24 male workers, randomly assigned to the Pressurization group (n = 12) or Control (n = 12) group. There were no significant differences between groups in key baseline demographics, including age, body mass index, and high-altitude work history (all p > 0.05; refer to Table 2 for statistical details).

Table 2.

Comparison of Baseline Characteristics

	n	age (years)	BMI(kg/m2)	High-Altitude Work Duration (Months)
Control	12	29.92 ± 5.12	21.79 ± 1.66	12.42 ± 2.11
Test	12	29.25 ± 7.24	21.90 ± 1.76	12.50 ± 1.93
t		0.260	− 0.149	− 0.101
P		0.792	0.883	0.921

Nocturnal oxygenation

Mean SpO₂: There were significant main effects of group and time. A significant Group × Time interaction was identified (F(4, 88) = 28.36, p < 0.001). The Control group’s Mean SpO₂ remained stable across all time points (range: 87.92 ± 0.53% to 89.08 ± 0.54%). In contrast, the Pressurization group showed a progressive increase during simulated descent, peaking at 2000 m (96.33 ± 0.36%) before returning near baseline during recovery (89.75 ± 0.37%). The model estimates confirmed significantly greater increases in the Pressurization group at 3000 m (3.08%, p = 0.001), 2500 m (6.25%, p < 0.001), and 2000 m (7.67%, p < 0.001). See Table 3, 4 and Fig. 2.

Table 3.

Changes in Indicators at Different Altitude

	Control Group					Pressurization Group
	1st week	2nd week	3rd week	4th week	5th week	1st week2	2nd week3	3rd week4	4th week5	5th week6
Mean SpO₂ (%)	88.92 ± 0.57	89.08 ± 0.54	87.92 ± 0.53	88.50 ± 0.69	88.83 ± 0.61	89.08 ± 0.45	92.33 ± 0.31	94.33 ± 0.26	96.33 ± 0.36	89.75 ± 0.37
Min SpO2 (%)	80.25 ± 0.77	78.58 ± 1.24	79.58 ± 1.10	78.92 ± 1.19	78.25 ± 1.10	79.25 ± 1.10	82.17 ± 1.06	83.25 ± 0.59	85.17 ± 0.84	78.08 ± 0.93
ESS Score	10.25 ± 0.69	10.17 ± 0.92	10.92 ± 0.75	9.75 ± 0.74	9.83 ± 0.66	10.42 ± 0.71	9.25 ± 0.71	7.25 ± 0.70	6.33 ± 0.54	9.50 ± 0.60
TST (min)	406.54 ± 13.37	411.21 ± 13.71	415.12 ± 16.25	403.83 ± 12.37	416.42 ± 14.77	409.33 ± 12.48	407.42 ± 11.27	410.83 ± 11.49	427.96 ± 11.76	414.71 ± 9.07
Sleep Efficiency (%)	82.75 ± 1.01	85.75 ± 1.56	84.92 ± 1.77	85.33 ± 0.99	83.92 ± 1.57	83.17 ± 1.59	84.17 ± 1.13	85.58 ± 0.93	88.83 ± 1.59	84.33 ± 1.36
Wake Time Percent (%)	9.00 ± 0.96	8.83 ± 1.15	8.75 ± 1.39	8.67 ± 0.87	9.25 ± 1.05	8.92 ± 1.17	8.33 ± 1.05	7.67 ± 0.83	7.83 ± 1.02	8.42 ± 1.11
REM Percent (%)	23.33 ± 1.72	21.83 ± 1.49	22.75 ± 1.36	23.50 ± 1.46	22.50 ± 1.45	24.67 ± 1.03	23.00 ± 1.37	23.25 ± 1.05	23.17 ± 1.13	24.17 ± 1.04
Light Sleep Percent (%)	50.83 ± 1.51	52.75 ± 1.25	50.42 ± 1.75	49.42 ± 1.09	51.25 ± 1.39	49.50 ± 1.14	51.58 ± 2.05	49.00 ± 0.64	47.58 ± 1.92	49.75 ± 0.99
Deep Sleep Percent (%)	16.83 ± 1.27	16.58 ± 1.22	18.92 ± 1.54	18.42 ± 1.62	17.08 ± 1.28	16.92 ± 1.03	17.08 ± 1.07	20.08 ± 1.65	21.75 ± 1.48	17.67 ± 1.30
ArI (events/h)	13.98 ± 1.59	14.77 ± 1.50	15.18 ± 1.75	15.02 ± 1.32	15.87 ± 1.16	14.08 ± 1.63	13.81 ± 1.61	9.87 ± 1.16	8.12 ± 0.98	14.96 ± 1.79

Data are presented as mean ± SD. ESS: Epworth Sleepiness Scale; TST: Total Sleep Time; REM: Rapid Eye Movement; ArI: Arousal Index

Table 4.

LMMs for Nocturnal Oxygenation and ESS Score

	Mean SpO₂	MinSpO₂	ESS Score
Intercept	88.92 (0.49)***	80.25 (1.01)***	10.25 (0.71)***
Group	0.17 (0.69)	− 1.00 (1.43)	0.17 (1.00)
Time (3000 m)	0.17 (0.63)	− 1.67 (1.12)	− 0.08 (0.94)
Time (2500 m)	− 1.00 (0.63)	− 0.67 (1.12)	0.67 (0.94)
Time (2000 m)	− 0.42 (0.63)	− 1.33 (1.12)	− 0.50 (0.94)
Time (Recovery)	− 0.08 (0.63)	− 2.00 (1.12)	− 0.42 (0.94)
Group × Time (3000 m)	3.08 (0.89)***	4.58 (1.58)**	− 1.08 (1.33)
Group × Time (2500 m)	6.25 (0.89)***	4.67 (1.58)**	− 3.83 (1.33)**
Group × Time (2000 m)	7.67 (0.89)***	7.25 (1.58)***	− 3.58 (1.33)**
Group × Time (Recovery)	0.75 (0.89)	0.83 (1.58)	− 0.50 (1.33)
ANOVA (Group)	86.35***	5.82*	9.05**
ANOVA (Time)	19.75***	7.28***	3.36*
ANOVA (Interaction)	28.36***	7.16***	3.60**
AIC	471.64	614.44	555.89
BIC	504.05	646.85	588.29

Values represent estimates (standard errors). Significance: ***p < 0.001, **p < 0.01, *p < 0.05. All models included random intercepts for participants. Baseline refers to 3500 m altitude

Reference: Control group, Baseline 3500 m

Fig. 2.

Longitudinal trends in Nocturnal Oxygenation and ESS Score

Minimum SpO₂: There were significant main effect of group and time. A significant Group × Time interaction was also observed (F(4, 88) = 7.16, p < 0.001). The Control group’s Minimum SpO₂ showed minor fluctuations (range: 78.25 ± 1.10% to 80.25 ± 0.77%). The Pressurization group exhibited clear improvement at lower altitudes, with the highest value at 2000 m (85.17 ± 0.84%). The interaction was significant at 3000 m (4.58%, p = 0.005), 2500 m (4.67%, p = 0.004), and 2000 m (7.25%, p < 0.001).

Subjective sleepiness (ESS score)

A significant Group × Time interaction was found for ESS scores (F(4, 88) = 3.60, p = 0.009). The Control group’s scores were stable (range: 9.75 ± 0.74 to 10.92 ± 0.75), while the Pressurization group’s scores decreased during simulated descent, reaching a nadir at 2000 m (6.33 ± 0.54). The model confirmed significantly greater reductions in the Pressurization group at 2500 m (− 3.83, p = 0.005) and 2000 m (− 3.58, p = 0.009) (Table 3, 4 and Fig. 2).

Sleep architecture

Deep Sleep Percent: A significant main effect of Time was observed (F(4, 88) = 12.41, p < 0.001). While both groups showed some variability, the Pressurization group’s mean increased to 21.75 ± 1.48% at 2000 m, compared to 17.08 ± 1.28% for the Control group at the same time point. The model indicated a significant Group × Time effect specifically at 2000 m (3.25%, p = 0.010) (Table 3, 5 and Fig. 3).

Table 5.

LMMs for Sleep Architecture

	Wake Time Percent	REM Percent	Light Sleep Percent	Deep Sleep Percent
Intercept	9.00 (1.07)***	23.33 (1.33)***	50.83 (1.43)***	16.83 (1.36)***
Group	− 0.08 (1.51)	1.33 (1.88)	− 1.33 (2.03)	0.08 (1.93)
Time (3000 m)	− 0.17 (1.26)	− 1.50 (1.08)	1.92 (1.61)	− 0.25 (0.88)
Time (2500 m)	− 0.25 (1.26)	− 0.58 (1.08)	− 0.42 (1.61)	2.08 (0.88)*
Time (2000 m)	− 0.33 (1.26)	0.17 (1.08)	− 1.42 (1.61)	1.58 (0.88)
Time (Recovery)	0.25 (1.26)	− 0.83 (1.08)	0.42 (1.61)	0.25 (0.88)
Group × Time (3000 m)	− 0.42 (1.78)	− 0.17 (1.52)	0.17 (2.28)	0.42 (1.24)
Group × Time (2500 m)	− 1.00 (1.78)	− 0.83 (1.52)	− 0.08 (2.28)	1.08 (1.24)
Group × Time (2000 m)	− 0.75 (1.78)	− 1.67 (1.52)	− 0.50 (2.28)	3.25 (1.24)**
Group × Time (Recovery)	− 0.75 (1.78)	0.33 (1.52)	− 0.17 (2.28)	0.50 (1.24)
ANOVA (Group)	0.43	0.29	1.04	0.41
ANOVA (Time)	0.29	1.15	2.72*	12.41***
ANOVA (Interaction)	0.09	0.54	0.02	2.17
AIC	634.36	627.57	693.4	595.04
BIC	666.76	659.97	725.81	627.45

Values represent estimates (standard errors). Significance: ***p < 0.001, **p < 0.01, *p < 0.05. All models included random intercepts for participants. Baseline refers to 3500 m altitude

Reference: Control group, Baseline 3500 m

Fig. 3.

Longitudinal trends in Sleep Architecture

Light Sleep Percent: A significant main effect of Time was found (F(4, 88) = 2.72, p = 0.035), but no significant Group × Time interaction was detected (F(4, 88) = 0.023, p = 0.999).

REM Sleep Percent: No significant main effects or Group × Time interaction were observed (Interaction F(4, 88) = 0.54, p = 0.704). Mean values were similar between groups throughout the study.

Wake Time Percent: No significant main effects or Group × Time interaction were found (Interaction F(4, 88) = 0.09, p = 0.984). Both groups maintained similar, stable levels of wakefulness after sleep onset.

Sleep continuity and global measures

Arousal Index (ArI): A significant Group × Time interaction was identified (F(4, 88) = 10.72, p < 0.001). The Control group’s ArI remained elevated (range: 13.98 ± 1.59 to 15.87 ± 1.16 events/h). In contrast, the Pressurization group’s ArI markedly decreased at 2500 m (9.87 ± 1.16 events/h) and 2000 m (8.12 ± 0.98 events/h). Model estimates confirmed substantially greater reductions in the Pressurization group at 2500 m (− 5.42 events/h, p < 0.001) and 2000 m (− 7.01 events/h, p < 0.001) (Table 3, 6 and Fig. 4).

Table 6.

LMMs for Sleep Continuity and Global Measures

	ArI	TST	Sleep Efficiency
Intercept	13.98 (1.47)***	406.54 (12.80)***	82.75 (1.38)***
Group	0.10 (2.08)	2.79 (18.10)	0.42 (1.95)
Time (3000 m)	0.79 (0.95)	4.67 (13.33)	3.00 (1.50)*
Time (2500 m)	1.20 (0.95)	8.58 (13.33)	2.17 (1.50)
Time (2000 m)	1.04 (0.95)	− 2.71 (13.33)	2.58 (1.50)
Time (Recovery)	1.89 (0.95)*	9.88 (13.33)	1.17 (1.50)
Group × Time (3000 m)	− 1.06 (1.34)	− 6.58 (18.85)	− 2.00 (2.12)
Group × Time (2500 m)	− 5.42 (1.34)***	− 7.08 (18.85)	0.25 (2.12)
Group × Time (2000 m)	− 7.01 (1.34)***	21.33 (18.85)	3.08 (2.12)
Group × Time (Recovery)	− 1.01 (1.34)	− 4.50 (18.85)	0.00 (2.12)
ANOVA (Group)	2.16	0.06	0.23
ANOVA (Time)	10.29***	0.29	4.14**
ANOVA (Interaction)	10.72***	0.8	1.48
AIC	612.36	1164.24	680.03
BIC	644.77	1196.64	712.44

Values represent estimates (standard errors). Significance: ***p < 0.001, **p < 0.01, *p < 0.05. All models included random intercepts for participants. Baseline refers to 3500 m altitude

Reference: Control group, Baseline 3500 m

Fig. 4.

Longitudinal trends in Sleep Continuity and Global Measures

Sleep Efficiency: A significant main effect of Time was observed (F(4, 88) = 4.14, p = 0.004), but no significant Group × Time interaction was found (F(4, 88) = 1.48, p = 0.217). Both groups showed similar patterns, with the highest mean efficiency in the Pressurization group at 2000 m (88.83 ± 1.59%).

Total Sleep Time (TST): No significant main effects or Group × Time interaction were detected (Interaction F(4, 88) = 0.80, p = 0.530).

Discussion

This study demonstrates that nocturnal pressurization to simulate stepwise altitude descent produces a unique pattern of sleep improvement in high-altitude residents. The most salient finding is the selective and significant enhancement of deep sleep (slow-wave sleep) alongside markedly improved oxygenation and reduced cortical arousals, without a concurrent rebound in REM sleep. This profile contrasts with the more generalized sleep architecture improvements (affecting both SWS and REM sleep) often reported with diffuse oxygen supplementation (Nussbaumer-Ochsner et al. 2012; Moraga et al. 2014), underscoring a distinct physiological mechanism of action for pressurization.

Integrated stabilization of respiration

The primary and most robust effect of pressurization was a marked, dose-dependent improvement in both mean and minimum SpO₂, which underpins all subsequent benefits. At high altitude, hypobaric hypoxia induces PB with Cheyne-Stokes-like oscillations that fragment sleep via repetitive cortical arousals (Weil 2004; Nussbaumer-Ochsner et al. 2012; Tang 2014). Both pressurization and supplemental oxygen improve SpO₂, but through distinct pathways. Supplemental oxygen elevates arterial PO₂, suppressing peripheral chemoreflex and the consequent hypocapnia that triggers central apnea (Moraga et al. 2014). Pressurization, by simulating descent, increases ambient barometric pressure and thus PiO₂, which may more effectively mitigate chronic hypocapnic alkalosis and stabilize the respiratory control system’s set-point (Heinzer et al. 2016).

The mechanistic distinction becomes clearer when considering loop gain (LG). Hypobaric hypoxia increases chemoreflex sensitivity (i.e., LG) during ventilatory acclimatization, and an LG > 1.0 directly correlates with central sleep apnea severity (r = 0.76 for Apnea Hyponea Index) (White et al. 1987; Bird et al. 2020). Supplemental oxygen acutely suppresses peripheral chemoreflex output but does not reset the hypocapnic alkalotic state that sustains elevated LG. In contrast, pressurization simultaneously corrects the hypoxic chemoreflex stimulus and raises the apneic threshold toward eucapnia, thereby reducing LG‑driven instability from both sides of the respiratory feedback loop (Orr et al. 2018). This integrative resetting explains why the observed reductions in arousal index at 2500 m and 2000 m (− 5.42 to − 7.01 events/h) are notably larger than those typically reported in oxygen trials (Nussbaumer-Ochsner et al. 2012; Moraga et al. 2014), despite comparable SpO₂ improvements. Thus, oxygen supplementation suppresses the output of the chemoreflex loop, whereas pressurization alters the loop’s operating set‑point—a distinction with direct therapeutic implications for altitude‑induced sleep fragmentation.

Prioritized consolidation of deep sleep

A key and distinctive finding was the significant increase in deep sleep (SWS) percentage in the intervention group, peaking at 2000 m, without concomitant significant changes in REM sleep. This pattern contrasts with studies of diffuse oxygen, which often report concurrent increases in both SWS and REM sleep (Nussbaumer-Ochsner et al. 2012). SWS is highly vulnerable to disruption by respiratory events and arousals (Wickramasinghe and Anholm 1999). The stabilization of breathing and reduction in arousals evidenced by our data likely created the sustained, uninterrupted sleep periods necessary for deep sleep consolidation. The lack of significant REM sleep rebound may indicate that the inherent ventilatory instability and reduced muscle tone of REM sleep are less amenable to correction via intermittent, nocturnal-only pressurization, or that a longer intervention is required for REM architecture to normalize (Plywaczewski et al. 2003; Johnson et al. 2010). This selective effect suggests pressurization may prioritize the restoration of the most fragile and physiologically restorative sleep stage first.

Subjective and systemic benefits

The physiological improvements translated directly into a meaningful clinical benefit, as reflected in the significantly reduced ESS scores during simulated descent. Improved and stable oxygenation likely attenuates the sympathetic nervous system overdrive characteristic of chronic hypoxia and raises the cortical arousal threshold (Ainslie et al. 2007). This shift toward autonomic balance and the objective reduction in sleep fragmentation (lower ArI) are the probable mechanisms behind the enhanced subjective sleep quality (Szymczak et al. 2009; Bian et al. 2019). The intervention did not alter total sleep time, indicating its primary benefit lies in profoundly improving the quality and restorative depth of sleep, not its mere duration.

Comparative efficacy and implications

While both pressurization and oxygen supplementation target nocturnal hypoxia, pressurization more closely mimics the integrative physiology of true altitude descent. It resets the barometric context for respiratory control, potentially offering a stimulus that parallels partial acclimatization (Moraga, López et al. 2018). For high-altitude residents or workers, this represents a novel non-pharmacological strategy that not only alleviates sleep disruption but may also confer broader adaptive benefits. Our findings, particularly the potent effect on arousal index and deep sleep, suggest that for improving sleep continuity and restorative depth, pressurization may offer advantages over low-flow oxygen alone (Moraga et al. 2014; Pun et al. 2025).

Limitations and future directions

This study was conducted in a cohort of acclimatized male workers, so findings may differ in unacclimatized individuals or other populations. Several limitations should be noted. First, the sample size was modest (n = 12 per group), limiting statistical power to detect smaller effects and increasing the risk of type II error for secondary outcomes; replication in larger cohorts is needed. Second, the exclusion of females limits generalizability, as sex differences in hypoxic ventilatory response and sleep architecture are well documented; future studies should include both sexes. Third, the 5‑week protocol with only one recovery week does not address long‑term sustainability or potential rebound effects; the return of SpO₂ and deep sleep to near baseline during recovery suggests benefits are largely reversible. Longer‑term follow‑up is required to determine whether repeated pressurization induces lasting adaptation. Additionally, the intermittent nature and specific altitude steps of our protocol define the intervention's scope. Future mechanistic studies should directly measure loop gain, PaCO₂ oscillations, and sympathetic activity during pressurization, and direct comparative trials against oxygen therapy are needed to elucidate relative efficacy and optimal applications.

Conclusion

Nocturnal pressurization simulating stepwise descent effectively improves nocturnal oxygenation, consolidates deep sleep, and alleviates daytime sleepiness in high-altitude workers. By addressing the dual drivers of hypobaric hypoxia—namely reduced barometric pressure and oxygen availability—this intervention engages physiological pathways distinct from conventional nocturnal oxygen therapy. It represents a promising non-pharmacological strategy that not only ameliorates sleep disruption but may also partially mimic the restorative acclimatization benefits of true altitude descent. Further research is warranted to assess its long-term applicability and impact on occupational performance and health in high-altitude settings.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributors

X Shan conceived, designed the study and collected the data, finalized the analysis, wrote the drafts of the manuscript. S Li supervised the study and helped revise drafts of the manuscript. Q Deng, J Wen, Y Wang, and S Chen designed the study, wrote the drafts of the manuscript and interpreted the findings. M Lin, X Liu, X Miao and H Huang collected the data and analyzed the data. All authors read and approved the final manuscript.

Funding

This work was supported by the National Key Research and Development Program of China [Grant No. 2021YFB2301703-03] and Chengdu Science and Technology Bureau Achievement Transformation Demonstration Project [Grant No. 2026-YF09-00029-SN].

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare that they have no competing interests.

Ethics approval and consent to participate

The study was approved by the ethics committee of Affiliated Hospital of Chengdu University [Grant No. PJ2022-045–02]. All participants provided written informed consent, and the consent forms were securely archived by the research team.

Consent for publication

Not applicable.