How High-Pressure Oxygen Therapy May Improve Sleep Quality in Different Patients

Sleep is a fundamental physiological process characterized by altered consciousness, reduced responsiveness to external stimuli, and restorative neural activity. Adequate sleep is crucial for maintaining both physiological and psychological functioning, especially regarding brain performance and homeostasis (1, 2). Poor sleep can contribute to, and result from, impaired brain function (3).

Sleep disturbances, including fragmented sleep, REM sleep behavior disorder, and excessive daytime sleepiness, are common in neurodegenerative diseases such as Parkinson’s disease and are driven by underlying neurodegenerative changes (4, 5). Sleep disturbances are also prevalent in conditions like PTSD (6) and Long COVID (7), where they are associated with underlying neuroinflammation (8) and dysregulated autonomic function.

Sleep quality naturally declines with age (9), marked by shorter sleep duration, increased fragmentation, reduced slow-wave sleep, and alterations in circadian rhythms and hormone secretion patterns. Impaired sleep in older adults is associated with diminished cognitive function (10), increased cardiovascular and metabolic risk (11, 12), weakened immune responses, and heightened susceptibility to chronic illnesses.

Hyperbaric oxygen therapy (HBOT), involving the inhalation of pure oxygen at supra-atmospheric pressures, has emerged as a neuromodulatory intervention. Dedicated HBOT protocols, employing the hyperoxic-hypoxic paradox (HHP), through cyclic oxygen fluctuations, activate neurobiological pathways including hypoxia-inducible factors (HIF), vascular endothelial growth factor (VEGF), stem cell proliferation, angiogenesis, and mitochondrial biogenesis, alongside reductions in neuroinflammation (13–15). These mechanisms underlie the neuroplasticity and clinical improvements observed in stroke, traumatic brain injury, post-COVID syndrome, PTSD, and age-related functional decline (14, 16–23).

Several studies suggest potential beneficial effects of HBOT on sleep quality across distinct clinical populations (24–26). For example, HBOT has been associated with improved sleep duration and reduced sleep latency in children with cerebral palsy (25). In Parkinson’s disease, a meta analysis and a systematic review reports association with enhanced sleep efficiency, increased total sleep duration, reduced sleep latency, and fewer awakenings following HBOT (27, 28). Similarly, patients with PTSD experience reduced nightmares and improved sleep quality following HBOT (29).

Although emerging evidence suggests that HBOT may influence sleep-related outcomes in specific clinical contexts, existing studies are limited by small sample sizes, heterogeneous populations, and varying outcome measures. This retrospective study aimed to assess HBOT’s impact on sleep quality among patients treated for various clinical indications by using subjective reports in sleep quality questionnaires.

The study is a retrospective, longitudinal analysis evaluating changes in sleep quality, assessed by the Pittsburgh Sleep Quality Index (PSQI), among patients completing a 60-session HBOT protocol at the Sagol Center for Hyperbaric Medicine and Research.

Between January 2017 and January 2025, a total of 395 patients completed a 60-session hyperbaric oxygen therapy (HBOT) protocol and filled out the Pittsburgh Sleep Quality Index (PSQI) questionnaire both at baseline (prior to the first session) and at the end of the treatment course.

The mean age at baseline of the cohort was 57.9 ± 14.6 years, and 31% (n = 125) were female. Of the total, 180 patients received HBOT for healthy aging, 92 for long COVID symptoms, and 123 for post-traumatic stress disorder (PTSD). Baseline demographic and sleep quality characteristics are summarized in Table 1.

The overall PSQI score at baseline for the entire cohort was 9.0 ± 4.0. When stratified by indication, baseline scores were 7.1 ± 3.3 for healthy aging, 8.4 ± 3.6 for long COVID, and 12.1 ± 3.5 for PTSD. Following treatment, the global PSQI score improved significantly in all groups: to 6.5 ± 3.0 in the healthy aging group, 7.7 ± 3.3 in the long COVID group, and 10.9 ± 3.6 in the PTSD group (all p < 0.001). In the full cohort, the mean change in PSQI score was 0.8 points (SD = 0.88), with a 95% confidence interval ranging from 0.71 to 0.89, indicating a small but precisely estimated average improvement in sleep quality. Effect sizes for the global PSQI score ranged from moderate to large across study populations (Cohen’s d = 0.37–0.91), indicating a clinically meaningful treatment effect.

Significant improvements were also observed across most individual PSQI components (Figure 1). Following treatment, sleep disturbances improved significantly (Z = −13.29, p < 0.001), as did subjective sleep quality (Z = −7.8, p < 0.001), sleep latency (Z = −5.37, p < 0.001), and daytime dysfunction (Z = −9.12, p < 0.001). Although the mean sleep duration score remained numerically unchanged (1.6 vs. 1.6), the Wilcoxon signed-rank test indicated a modest shift in the paired distribution of scores (Z = −2.21, p = 0.027). No significant changes were observed in sleep efficiency (Z = −0.39, p = 0.73), Medication use, as reported in the questionnaire, did not change significantly (Z = −1.24, p = 0.24). After Bonferroni correction for multiple comparisons, improvements in sleep disturbances, sleep latency, and subjective sleep quality remained statistically significant across groups, whereas findings with p-values close to 0.05 should be interpreted cautiously.

To address concerns regarding regression to the mean and baseline severity differences between diagnostic groups, we performed a multivariable linear regression with post-treatment PSQI as the dependent variable and baseline PSQI, age, sex, and treatment indication as predictors. Baseline PSQI emerged as the strongest independent predictor of post-treatment sleep quality (B = 0.492, p < 0.001). Neither age, sex, nor treatment indication independently predicted post-treatment PSQI after adjustment. To further examine the impact of baseline sleep severity, participants with normal sleep at baseline (PSQI ≤ 5) were excluded from this subgroup analysis (n = 63) and the effect of HBOT on disturbed sleep (PSQI > 5) was evaluated. Among participants with elevated baseline PSQI scores (n = 332), significant improvements were observed for overall score, as well as across most sleep parameters as assessed using paired t-test for total PSQI and Wilcoxon for sleep parameters. Overall score decreased from 10.76 ± 3.46 to 8.82 ± 3.67 (p < 0.001). Sleep duration improved significantly (Z = −3.000, p = 0.003). Sleep disturbances showed a large and highly significant improvement (Z = −13.081, p < 0.001). Daytime dysfunction improved significantly (Z = −4.466, p < 0.001), as did subjective sleep quality (Z = −7.849, p < 0.001). Sleep latency also improved significantly (Z = −6.414, p < 0.001), reflecting reduced time to fall asleep. No statistically significant changes were observed in sleep efficiency (Z = −1.291, p = 0.197) or use of sleep medications (Z = −1.321, p = 0.186).

These findings suggest that among individuals with disturbed sleep at baseline, HBOT led to clinically and statistically significant improvements in multiple aspects of sleep, particularly sleep disturbances, latency, and subjective sleep quality.

The current study evaluated the effects of hyperbaric oxygen therapy (HBOT) on sleep parameters, as assessed through a validated sleep assessment questionnaire. Findings demonstrated a significant and robust improvement in sleep quality across diverse clinical populations, suggesting that HBOT was associated with improved sleep quality and supporting the need for prospective controlled trials to validate efficacy, identify patients most likely to benefit, and elucidate underlying mechanisms.” One possible explanation for the observed improvements in sleep is the amelioration of the primary condition for which HBOT was administered. For example, sleep disturbances are a core symptom of PTSD. As HBOT has been shown to reduce the overall symptom burden in PTSD (16, 17), this may secondarily lead to reductions in nightmares, improved sleep initiation and maintenance, and enhanced subjective sleep quality.

A second, broader potential explanation is that HBOT improves sleep through generalized enhancement of brain function, including modulation of autonomic balance (32), mitochondrial activity (15, 33, 34), inflammation (35), and neuroplasticity (15, 36). HBOT induces angiogenesis VEGF and HIF pathways, which may directly influence neuronal perfusion and energy metabolism (15, 37, 38), thereby promoting overall neuronal health and sleep quality. Previous studies have linked disrupted mitochondrial function to sleep disorders (39), supporting the notion that enhanced mitochondrial biogenesis may play a role in the observed improvements. Additionally, HBOT’s anti-inflammatory effects, both systemic and neuroinflammatory, are relevant given the established links between inflammation, circadian disruption, and impaired sleep. Enhanced neuroplasticity and upregulation of neurotrophic factors such as brain-derived neurotrophic factor (BDNF) following HBOT may also contribute to improvements in sleep architecture and quality (40, 41).

A third potential mechanism involves direct effects of HBOT on sleep-regulatory pathways, such as adenosine accumulation and melatonin signaling. Adenosine functions as a key neuromodulator promoting sleep (42), and increased tissue oxygen availability may influence adenosine metabolism and receptor signaling (1). Preliminary findings also suggest that HBOT affects melatonin receptor activity, potentially enhancing circadian alignment and sleep onset (43).

An important factor to consider in interpreting these results is treatment dosage. All patients in this study completed 60 daily HBOT sessions. Evidence from other populations indicates that a sufficient cumulative oxygen dose is critical to induce durable neuroplastic changes and long-term clinical effects (44, 45). For example, prior studies in individuals with long COVID and in children with cerebral palsy using only 10 HBOT sessions did not demonstrate improvements in sleep parameters. However, in the CP study, 20 sessions did result in measurable benefits (25). Same with relation to patients who are suffering from long COVID, unlike 40 HBOT sessions found to be effective (19) 10 sessions protocol was not (46). This suggests a dose–response relationship between cumulative oxygen exposure (measured in atmosphere-minutes) and symptomatic improvement, including for sleep.

The study’s limitations are primarily related to its retrospective design and the absence of a control group, which preclude causal inference regarding the effects of HBOT on sleep outcomes. Consequently, the observed improvements in PSQI scores should be interpreted cautiously and considered as associations rather than definitive evidence of therapeutic efficacy. Changes in patient-reported sleep quality may reflect regression to the mean, natural fluctuations in sleep symptoms over time, expectancy or placebo effects, or other non-specific aspects of participation in a structured treatment program involving daily clinical interaction and monitoring. HBOT in this cohort was delivered as part of routine clinical care rather than as an isolated experimental intervention. Accordingly, the observed changes in sleep quality may reflect both direct physiological effects of HBOT and indirect effects related to participation in a structured clinical environment. In addition, sleep outcomes in this study were assessed exclusively using the self-reported PSQI. Although the PSQI is a well-validated and widely used instrument for evaluating perceived sleep quality, it reflects subjective sleep experience rather than objective sleep physiology. Subjective assessments of sleep quality may diverge from objectively measured sleep parameters such as sleep architecture, sleep efficiency, or fragmentation measured using instrumental methods including polysomnography or sleep polygraphy. Several studies have demonstrated that patient-reported sleep disturbances do not always correspond to objective sleep metrics in the same individuals, highlighting the complex relationship between perceived and physiological sleep quality (47, 48). This discrepancy may arise from multiple factors, including cognitive and emotional influences on sleep perception, underlying neuropsychiatric conditions, and thus, recent studies have emphasized that subjective and objective sleep measures capture partially distinct aspects of sleep health and should ideally be interpreted as complementary rather than interchangeable assessments. In addition, given the psychological and contextual influence on subjective sleep assessments, patient-reported sleep quality may be affected by mood, stress levels, expectations regarding treatment efficacy, and the broader therapeutic environment in which treatment occurs. In the context of an uncontrolled retrospective cohort, improvements in PSQI scores may therefore reflect not only physiological changes in sleep but also expectancy effects or placebo-related influences associated with participation in an intensive treatment program.

Accordingly, reliance solely on questionnaire-based evaluation introduces the possibility of measurement bias and limits the ability to determine whether the observed improvements reflect changes in sleep physiology, sleep perception, or both. Future prospective, controlled studies investigating the effects of HBOT on sleep should incorporate objective sleep assessments such as actigraphy, sleep polygraphy, or polysomnography alongside validated questionnaires in order to provide a more comprehensive characterization of sleep outcomes.

Potential confounding by baseline comorbidities and pharmacotherapy should also be considered. The study population demonstrated variability in medical conditions and medication use, including antidepressants, benzodiazepines, and cardiovascular medications, which may independently influence sleep quality. Although the multivariable regression analysis adjusted for major demographic and clinical variables including baseline PSQI and treatment indication, residual confounding cannot be fully excluded in a retrospective observational design.

An additional limitation is the lack of detailed longitudinal pharmacotherapy data. While overall sleep medication use was captured via the PSQI questionnaire, information regarding specific medications, dosages, and changes in pharmacological treatment during the HBOT course was not available; therefore, potential medication-related effects on sleep outcomes cannot be fully excluded.

This retrospective analysis suggests that hyperbaric oxygen therapy was associated with improvements in patient-reported sleep quality across several clinical populations. However, due to the observational design and the absence of a control group, causal conclusions cannot be drawn. These findings support the need for prospective randomized controlled trials incorporating objective sleep assessments to better characterize the potential role of HBOT in modulating sleep and to identify patient populations that may benefit the most.

The author(s) declared that financial support was not received for this work and/or its publication.

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request. Requests to access these datasets should be directed to KD-B. Kerendoenyas@gmail.com

The studies involving humans were approved by Shamir Medical Center's Institutional Review Board. The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required from the participants or the participants' legal guardians/next of kin in accordance with the national legislation and institutional requirements.

KD-B: Investigation, Writing – original draft. KE: Investigation, Writing – review & editing. EL: Writing – review & editing. SF: Writing – review & editing. VE: Writing – review & editing. RS: Writing – review & editing. SE: Writing – review & editing.

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The author(s) declared that Generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.