Study design and setting

A randomized, crossover trial was performed to evaluate the effect of short-term descent to 590 m in healthy residents at moderate altitude (median (IQR) altitude of 1230 m (1130; 1560)). The study was carried out from July 2023 to March 2024. Evaluations at lowland took place during two consecutive nights at the Cantonal Hospital of Grisons (590 m), Switzerland. Moderate altitude assessments were conducted at the participants’ homes, located > 1000 m. Assessments between low and moderate altitudes were separated by a 2-week washout phase at > 1000 m. The trial has been registered at clinicaltrials.gov (NCT05826808).

Participants

Participants were recruited via flyer and social media but also included a few colleagues and family members. Healthy males and females living > 1000 m (27 born at this altitude, 17 living for at least a year at this altitude), aged between 18 and 70, with a body mass index between 18.5 and 30 kg/m2 were included in the study. People with active diseases or health-related conditions that require treatment or were on regular medication, current smokers, regular alcohol consumers, or drug users were excluded. The study has been approved by the Cantonal Ethics Committee of Zurich (05–2023), and participants gave their written informed consent.

Randomization and interventions

Participants were randomized to groups A and B. Randomization was achieved with a computer-generated randomization sequence, using the python3 library random. Participants assigned to group A were first examined at 590 m, followed by examinations at their home elevations (> 1000 m). Participants assigned to group B were first examined at > 1000 m and then at 590 m. Participants immediately descended to 590 m before undergoing measurements at the sleep laboratory and spent the day between the two consecutive nights in the sleep laboratory at 590 m. They were instructed not to spend nights at < 1000 m in the previous 4 weeks before their study participation.

Measurements Sleep studies

All sleep examinations were performed using portable polysomnography (SOMNOmedics GmbH, Randersacker, Germany) for 8 h each night. The polysomnographic recordings included electroencephalography (C3A2, C4A1, F1A2, F2A1, O1A2, O2A1), electrooculography, electromyography (submental and anterior tibialis), electrocardiography, pulse oximetry, nasal thermal airflow and pressure, thoraco-abdominal movements, and transcutaneous capnography (Sentec AG, Therwil, Switzerland). Sleep studies were analyzed in 30-s epochs by two evaluators in joint review according to the American Academy of Sleep Medicine criteria [4]. Briefly, apneas were defined as a reduction in airflow of more than 90% for at least 10 s. Obstructive apneas were differentiated from central apneas by rib cage and abdominal asynchrony. Mixed apneas were scored as obstructive events. Hypopneas were defined as an airflow reduction of 30% for at least 10 s and an oxygen desaturation of 3% or an arousal [5,6,7,8].

Clinical examination and questionnaires

Evening and morning evaluations included measurement of blood pressure, heart rate, and lung auscultation. Sleepiness was evaluated by the Karolinska Sleepiness Scale [9], Epworth Sleepiness Scale [10], and Stanford Sleepiness Scale [11]. Subjective sleep quality was rated on a visual analogue scale, marked with “very bad” at 0 mm and “excellent” at 100 mm. Subjective insomnia was evaluated by three questions (“How long did it take you to fall asleep?”, “How often did you wake up in the night?”, and “How long were you awake during the night?”).

Outcomes

The primary outcome was time spent under 90% oxygen saturation (T90). T90 has been shown to be a relevant parameter because it provides information about the duration and degree of hypoxia during the whole sleep. A previous study showed that a cut-off value of 12 min of T90 is an independent predictor of cardiovascular mortality in older males [12]. Important secondary outcomes were nocturnal oxygen saturation (SpO2), apnea–hypopnea index (AHI), and oxygen desaturation index (ODI). Additional outcomes were slow wave sleep (SWS), arousal index (AI), transcutaneous pCO2 (TcCO2), and wakefulness after sleep onset (WASO).

Sample size estimation

By assuming a clinically relevant difference in T90 of 12 min, a standard deviation of 24 min, a power of 80%, and a two-sided alpha level of 0.05, 34 participants are required. To account for dropouts, the recruiting aim was at least 40 participants (20 females, 20 males), who were then randomized into the two aforementioned groups.

Data analysis and statistics

An investigator blinded to the altitude allocation performed all statistical analyses. The data are summarized as means ± SD. The primary outcome was analyzed by intention-to-treat, with missing data replaced by multiple imputations (n = 20) using chained regression equations. Secondary outcomes were analyzed in the per-protocol population, which is defined as all participants completing all four nights. Mean values, SEs, mean differences, and 95% CIs were computed using mixed linear regression models with outcomes as dependent variables, altitude and sex fixed effects, and participants as random effects. To investigate predictors for T90 and ODI, exploratory mixed linear regression analyses were performed, using randomization sequence, age, BMI, sex, and altitude as predictors. All other analyses were performed unadjusted for baseline confounders. Statistical significance was assumed when the 95% CI of mean differences did not overlap with the value zero. Statistical analysis was performed by STATA version 15.1 (StataCorp).