Article Text

Hyperbaric oxygen therapy for poststroke insomnia: a systematic review and meta-analysis protocol
  1. Rui Shi1,
  2. Wenyi Meng2,
  3. Zhaozheng Liu2,
  4. Liping Chang2,
  5. Ruozhu Lu1,
  6. Xingyu Chen1,
  7. Wen Xue1,
  8. Yue Deng1
  1. 1Changchun University of Chinese Medicine, Changchun, Jilin, China
  2. 2Affiliated Hospital, Changchun University of Chinese Medicine, Changchun, Jilin, China
  1. Correspondence to Professor Yue Deng; dyuesr7138{at}


Introduction Insomnia stands as a frequent consequence of a cerebrovascular event, afflicting a substantial fraction of those who endure the aftermath of stroke. The ramifications of insomnia following a stroke can further exacerbate cognitive and behavioural anomalies while hindering the process of neurological convalescence. While several randomised controlled trials (RCTs) have scrutinised the effects of hyperbaric oxygen therapy (HBOT) on poststroke insomnia, the advantages and drawbacks persist in a state of ambiguity. We advocate for a systematic review and meta-analysis of randomised clinical trials to comprehensively evaluate the effectiveness and safety of HBOT in the context of poststroke insomnia.

Methods and analysis A systematic search will be conducted from nine major databases (PubMed, Web of Science, EMBASE, VIP Information Database, Cochrane Central Register of Controlled Trials, China National Knowledge Infrastructure, China Biomedical Literature Database and Wanfang Database, Physiotherapy Evidence Database (PEDro)) for HBOT for poststroke insomnia of RCTs. The search procedures will adhere to a rigorous approach, commencing from the inception date of each database and continuing until 1 November 2023, with inquiries conducted exclusively in English and Chinese. The primary outcome will focus on the alteration in the quality of sleep while secondary outcomes will encompass the evaluation of adverse events and the rate of reoccurrence. The process of selecting studies, extracting data and evaluating the quality of research will be carried out independently by two reviewers. The quality of the included literature will be assessed using the tools of the Cochrane Collaboration. Meta-analysis will be performed by using RevMan V.5.4 and STATA V.16.0.b software. Finally, the quality of evidence will be assessed using the Grading of Recommendations, Assessment, Development and Evaluation method.

Ethics and dissemination As all data are derived from published investigations via databases without direct patient contact, ethical approval is obviated in this study. The scientific studies will be published in professional academic publications.

PROSPERO registration number CRD42023468442.

  • Stroke medicine
  • Systematic Review
  • Meta-Analysis

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • This systematic review protocol will meticulously adhere to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis Protocols guidelines, ensuring utmost transparency and equity throughout the review process.

  • This study is a systematic assessment of the efficacy and safety of hyperbaric oxygen for the treatment of poststroke insomnia, which can comprehensively synthesise the available evidence and provide evidence-based results.

  • The literature search will conduct a comprehensive analysis of randomised controlled trials selected from multiple databases encompassing a broad scope of sources to minimise selection bias.

  • Variations in hyperbaric oxygen therapy protocols, such as pressure level, number of treatments, timing of initiation and duration, among other factors, may contribute to the introduction of heterogeneity across studies.

  • The study’s limitations may stem from the relatively small number of well-designed randomised controlled trials, potentially impacting the overall robustness of the findings.


Stroke constitutes the second-leading global cause of mortality, accounting for approximately 11% of all deaths, with rising year-on-year incidence, representing a priority public health concern requiring worldwide attention and redress.1–4 Beyond major neurological deficits, stroke survivors frequently encounter numerous secondary complications including impaired communication, mobility and cognition.5–7 Poststroke insomnia, one of the most prevalent psychiatric disorders, shows a prevalence of 38%–40%.8 9 Prior investigations indicate poststroke insomnia may stem from variable degrees of neuronal injury from the stroke itself coupled with compression of neighbouring brain tissue by cerebral oedema. The severe damage to discrete brain areas governing sleep regulation blocks conduction in the reticulo-activating system, disrupting normal sleep architecture.10 11 Studies indicate strokes impacting dorsal/covered brainstem regions, parietal or lateral thalamus, or subcortical areas confer higher risk for insomnia.12 Stroke-associated insomnia not only impairs patient quality of life but also heightens susceptibility to psychiatric disorders including depression and anxiety, further adversely influencing neurological and physiological recovery.13 14 A preceding meta-analysis, encompassing nine studies, has illuminated the protracted diminution of both total sleep duration and sleep efficiency in individuals afflicted by strokes during the antecedent period to the cerebrovascular event. This decline is discernible in the electroencephalogram patterns across diverse sleep stages.15 Furthermore, it is noteworthy that insomnia may endure for an extended duration following a stroke. The antecedent meta-analysis findings also manifest that the incidence of insomnia remains relatively constant among stroke survivors during the acute, subacute, or chronic stages of stroke (40.7%, 42.6% and 35.9%, respectively).16 Insomnia, far from being inconsequential, begets cognitive, psychological and behavioural perturbations,17 and regrettably, it engenders escalated mortality rates and a heightened propensity for suicide among affected patients.18–20

The administration of psychotropic pharmaceuticals and soporific agents persists as a prevailing therapeutic modality for insomnia in the aftermath of cerebrovascular events. Conventional pharmacotherapeutic agents encompass benzodiazepines, non-benzodiazepines, sedative tricyclic antidepressants and melatonin receptor agonists, exemplified by alprazolam, eszopiclone, zaleplon, zolpidem, diazepam, zopiclone and agomelatine.21 While these pharmaceutical agents do exhibit a modicum of efficacy in ameliorating sleep, their concomitant side effects loom conspicuously, encompassing concerns regarding drug dependency, cognitive degradation, somnolence, vertigo, dyskinesia and rebound insomnia on cessation.22–25 A precedent investigation demonstrated that prolonged benzodiazepine utilisation culminated in an augmented incidence of cerebrovascular events attributed to dose escalation.26 An extensive literature scrutiny, encompassing several clinical investigations, divulged that benzodiazepines and the so-called class Z medications (comprising zolpidem and zopiclone) precipitated an elevated proclivity for falls due to orthostatic hypotension, vertigo, disequilibrium, sedation, muscular enfeeblement and ataxia, among other factors. Importantly, these fall events directly contributed to an amplified mortality risk. Consequently, the study ardently advocated the contemplation of non-pharmacological substitutes and proposed the imposition of stringent restrictions on the minimum prescribed dosages and the abbreviated temporal span of drug utilisation.27

Concerning non-pharmacological interventions, cognitive behavioural therapy for insomnia (CBT-I) stands as one of the more prevalent modalities.28 Cognitive–behavioural therapy represents a multifaceted therapeutic regimen encompassing sleep education, behavioural interventions, cognitive restructuring and relaxation therapies.29 CBT-I, in particular, exhibits marked efficacy in ameliorating self-appraised insomnia severity.30 31 Nevertheless, the therapeutic approach’s technical adaptability may diverge contingent on the educational attainment of the recipient, potentially influencing treatment efficacy.32 The therapy also contends with challenges such as a limited cadre of practitioners possessing requisite technical credentials, an extended treatment duration and a gradual manifestation of effects, with insufficient corroboration of its application thus far.33 34 Furthermore, as alluded to in antecedent investigations, patients and healthcare providers opine that postcerebrovascular injury insomnia ameliorates concomitantly with comorbidity management, or they perceive cognitive–behavioural interventions as ineffective for insomnia.35 These collective factors contribute to a certain constraint on the widespread adoption of cognitive–behavioural therapy. Consequently, it becomes imperative to investigate alternative modalities for poststroke insomnia characterised by sustained efficacy and safety. As a non-invasive and non-pharmacological treatment, hyperbaric oxygen therapy (HBOT), has received increasing attention in recent years.

HBOT involves administering 100% oxygen under hyperbaric conditions with absolute atmospheric pressure exceeding 1.4 ATA.36 Chamber pressure and frequency are tailored to each patient, commonly 2.0–3.0 absolute atmosphere (ATA) for 2–3 daily sessions in adults. Treatment duration depends on the condition, typically 1.5–2 hours.37 With over 300 years of development, HBOT has proven efficacy against various diseases including soft tissue infections, gas gangrene, vasculitis, encephalitis sequelae, deafness and carbon monoxide poisoning.38 Most side effects during HBOT are tolerable and reversible, chiefly middle ear barotrauma from pressure imbalance between the inner and outer ear.39 HBOT represents a non-invasive physical therapy to facilitate recovery after neuronal damage, effectively enhancing brain oxygenation, cerebral circulation and metabolism while attenuating oxidative stress and inflammation, improving mitochondrial function, and promoting neuroregeneration through high-dose oxygen inhalation.40–42 Animal experimentation and initial clinical investigations have substantiated that HBOT can markedly mitigate brain injury resultant from a spectrum of neurological maladies including ischaemia-reperfusion injury, hypoxic-ischaemic encephalopathy, traumatic brain injury, intraventricular haemorrhage and more.37 43–45 Previous studies have suggested that hyperbaric oxygen is more advantageous for stroke treatment in the early stages (eg, within 1 hour of stroke onset46 and within 6 hours after reperfusion therapy47 48). However, subsequent research has validated that HBOT has equally effective therapeutic outcomes even when repeated treatments are postponed, particularly in enhancing executive function and histologic outcomes,49 as well as promoting neurological recovery.50 At present, HBOT has been extensively employed in the rehabilitation of various neurological disorders, including traumatic brain injury, cerebral palsy and various poststroke sequelae.51–54 HBOT has been demonstrated to ameliorate cognitive and language impairments55 as well as mental and sleep disorders.56

At present, some clinical trials have preliminarily suggested that HBOT could be an alternative treatment option for poststroke insomnia. However, to date, there has been no systematic review of the efficacy and safety of HBOT for the treatment of poststroke insomnia, and there is a lack of comprehensive evaluation of the available evidence. To elucidate the therapeutic effects of hyperbaric oxygen on poststroke insomnia, we propose to actualise a systematic evaluation and meta-analysis to provide the optimal clinical recommendations on the efficacy and safety of hyperbaric oxygen for poststroke insomnia.

Materials and methods


Our research protocol has been duly registered with the International Platform of Registered Systematic Review and Meta-Analysis Protocols (PROSPERO). The code was CRD42023468442. We shall cleave to the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA-P) and stringently hew to the benchmarks and deportment set forth therein.57 The particulars of the PRISMA-P checklist are delineated in online supplemental appendix 1.

Eligibility criteria

The eligibility criteria were guided by the Population-Intervention-Comparators-Outcomes-Study design framework.

Type of studies

This analysis will exclusively incorporate randomised controlled trials (RCTs), with eligibility confined to studies published in English or Chinese. To ensure optimal evidentiary standards, trials exhibiting suboptimal design, duplicate publication, inaccessibility of salient details, flaws in randomisation or statistical methods, incomplete documentation (both manual and electronic), data aberrations, statistical inconsistencies in baseline data or incomplete data will be stringently barred from our analyses.


Subjects fulfilling diagnostic criteria for poststroke insomnia will be incorporated. The study will encompass patients diagnosed with poststroke insomnia employing standard neuroimaging modalities (eg, brain MRI, MR angiography and CT). The insomnia diagnosis shall satisfy the standardised diagnostic guidelines stipulated in the Diagnostic and Statistical Manual of Mental Disorders, International Classification of Sleep Disorders, Chinese Classification of Mental Disorders and International Statistical Classification of Diseases and Related Health Problems. The diagnosis of insomnia cannot be explained by causes other than stroke. There exist no constraints predicated on race, age, gender, occupation, educational attainment or type of stroke. Patients with other severe psychiatric disorders, such as mania and schizophrenia, will be excluded from the study, as will patients with other significant medical conditions (eg, cancer, blood disorders, acute infectious diseases, liver disease, kidney disease). Studies that do not provide a conclusive diagnosis or a means of validation will also be excluded.

Types of intervention

The experimental group received HBOT alone or in conjunction with other conventional treatments, without limitations on the frequency, pressure or duration of interventions. The control group could be conventional treatment, no intervention, sham control, placebo and no HBOT under the same treatment regimen.

Type of outcomes

The primary outcome was the application of the Pittsburgh Sleep Quality Index (PSQI), which has been demonstrated to be a valid and reliable measure of sleep quality.58 Secondary outcomes included insomnia severity index (ISI), total effectiveness rate (TER), relapse rate after follow-up, quality of life and HBOT-related adverse events (such as barotrauma, oxygen toxicity, abdominal pain, fatigue, claustrophobia, ocular complications, cataract, hyperoxia myopia, hypoglycaemia).

Searching strategy

Electronic searches

A comprehensive and meticulous systematic inquiry will be executed to ascertain pertinent studies for this research endeavour. The pursuit of relevant literature shall extend to the databases, namely, PubMed, Web of Science, EMBASE, Physiotherapy Evidence Database (PEDro), Chinese Biomedical Literature Database (CBM), VIP Information Database, Cochrane Central Register of Controlled Trials, Chinese National Knowledge Infrastructure (CNKI) and Wanfang Database. The quest for relevant articles will encompass the entire publication history of these databases until 1 November 2023. To ensure a comprehensive investigation, a harmonious amalgamation of medical subject headings and unconstrained textual expressions will be employed. The search strategy shall encompass the terms ‘stroke,’ ‘insomnia,’ ‘hyperbaric oxygen,’ ‘RCT,’ and their corresponding synonyms and associated terminology. Various search algorithms shall be employed per the specifications of each database to optimise retrieval of pertinent studies.

Extra resources and search techniques

To identify additional references, a rigorous and comprehensive search strategy will be employed. First, a meticulous perusal of the citation indices of the primary papers and pertinent reviews will be undertaken to ascertain any conceivably relevant research. Furthermore, the database and the China Clinical Trials Registry will be searched to locate unpublished or ongoing studies. The search strategy will be developed and executed by two reviewers (LC and WM), ensuring thoroughness and accuracy. A third reviewer (XC) will be involved to ensure completeness and address any discrepancies. Table1 provides an illustrative example of the search strategy using the PubMed database.

Table 1

The search strategy for PubMed database

Data collection and analysis

Study selection

EndNote V.X9 software can be used to effectively manage literature. Initially, the software will identify and eliminate any duplicated articles. Subsequently, two impartial reviewers (WM and RL) will examine the titles, abstracts and keywords to exclude publications unrelated to the research subject. Any disagreements or comments will be resolved through discussion between the reviewers. Relevant publications matching predefined qualification measures or of uncertain applicability will be retrieved and examined comprehensively. If critical data information is missing, the original authors of the study will be contacted by phone or email to obtain the missing or insufficient data from the preliminary research.59 If the missing data prove unattainable, the classifications of ‘missing at random’ or ‘missing not at random’ will signify the distinct scenarios. In the case of ‘randomly missing’ data, only the available data will undergo analysis. For ‘non-randomly missing’ data, the missing values will be substituted with estimates (eg, replacing with the mean/median or using values predicted via a regression model), and sensitivity analyses will be performed to assess result consistency. Trials excluded due to non-compliance with inclusion criteria or substantial unanalysable missing data will be meticulously documented, elucidating the grounds for their exclusion.

Figure 1 shows a flow chart that follows the recommendations for PRISMA to give a clear visual depiction of the research selection process.

Figure 1

The flow chart of the study selection process. RCTs, randomised controlled trials.

Data extraction

Two impartial reviewers (WM and RL) will use a standardised data collection template to garner fundamental information. The extracted data will encompass crucial details pertinent to the study’s inclusion: (1) Basic information: title, publication type, year and journal of publication, authors, country, study purpose, trial enrolment and source of study funding. (2) Study characteristics: study design, randomisation method, blinding, allocation, concealment and completeness of outcome data. (3) Participant characteristics: age, sex, eligibility criteria, sample size, duration of illness and prior treatment. (4) Intervention treatment information: frequency of treatment, type of intervention, duration of treatment, duration of each intervention, treatment pressure and dose, etc and (5) Study outcome indicators: including PSQI, ISI, TER, AEs and relapse rate after follow-up. The two reviewers will engage in deliberation and resolution of any discord that may surface in the course of study curation and data retrieval. In the event of persistent discord, a third reviewer (XC) will be solicited for ultimate adjudication.

Risk of bias assessment

The Cochrane Risk of Bias Tool version 2, developed by the Cochrane Collaboration for evaluating bias risk in randomised trials, will be used to determine the level of bias present in the included literature. Seven crucial areas are covered by this instrument: random sequence generation, allocation concealment, participant and staff blinding, blinding of outcome assessment, inadequate outcome data, selective reporting, and other possible sources of bias. There will be three categories for each domain: ‘low risk,’ ‘high risk’ and ‘unclear.’

Two impartial reviewers (RL and LC) will rigorously examine the risk of bias in the selected research. In the event of disagreement, a third reviewer (XC) will be appointed to achieve consensus. When differences in the risk of bias between research are discovered, we will analyse their potential impact on the overall findings. To visually depict the assessment of bias, RevMan V.5.4 will be used to generate graphs and comprehensive summaries that demonstrate the risk of bias.

Synthesis of data and evaluation of heterogeneity

RevMan V.5.4 software will be used for the statistical analysis. The ORs will be used to produce a succinct estimate accompanied by a 95% CI for dichotomous data. For continuous data, a brief estimation of the standardised mean difference will be provided, along with a 95% CI. A p<0.05 will be used as the statistical significance threshold.

The presence of heterogeneity will be assessed, and the decision to do a meta-analysis and choose a model will be based on the results. If no significant heterogeneity is discovered (I2≤50%), the fixed-effect model shall be used. However, if I2 is greater than 50%, indicating heterogeneity, we will prioritise examining clinical, methodological or statistical heterogeneity to find plausible sources. To investigate the causes of clinical or methodological heterogeneity, we will use subgroup analyses or meta-regression analyses with STATA V.16.0.b software. To synthesise the data in the face of statistical heterogeneity, a random-effects model will be employed. Should significant heterogeneity persist across trials and remain unexplained, descriptive analyses will be conducted as an alternative to meta-analyses.

Subgroup analyses

Subgroup analyses will be performed on the results of RCTs to investigate potential sources of heterogeneity. These analyses will be performed if adequate data are available to detect variations in the following features, for example, date of publication, level of risk of bias for included RCTs, characteristics of insomnia, participant characteristics, sample size, follow-up time points, frequency of treatment and total duration of treatment.

Sensitivity analyses

Sensitivity analyses will be done, if appropriate, to assess the robustness and reliability of the review’s results about key outcomes. To evaluate the influence of individual studies on the overall meta-analysis, we will iteratively reanalyse the data, one research at a time, systematically removing studies with small sample sizes, a high risk of bias or incomplete findings. Following that, we will reassess the extent of the influence. This approach will allow us to identify any inconsistencies in the results. Subsequently, we will thoroughly discuss these discrepancies and exercise caution when formulating our conclusions.

Quality of evidence

The Grading of Recommendations, Assessment, Development and Evaluation methodology will be used to gauge the evidence quality. This approach categorises evidence quality into four categories: high, moderate, low and extremely low. Risk of bias, inconsistency, publication bias, indirectness and imprecision are among the assessment characteristics used to determine evidence quality.

Publication bias

Inverted funnel plots will be generated using STATA V.16.0.b to evaluate the probability of publication bias. The inverted funnel plots will be used to evaluate the probability of publication bias. Furthermore, Egger’s test will be conducted on meta-analyses comprising more than 10 trials to examine the presence of publication bias.

Patient and public involvement

Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

Ethics and dissemination

As all data are derived from published investigations via databases without direct patient contact, ethical approval is obviated in this study. The scientific studies will be published in professional academic publications.

Ethics statements

Patient consent for publication


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Contributors RS and ZL were involved in the initial conceptualisation of this study. RS and YD collaborated on the design of the protocol. RL, WM, LC and XC were responsible for conducting the data search and investigation. The original draft of the manuscript was written by RS and later reviewed and approved by YD. RS, WX and ZL provided methodological guidance. All authors have read and given their approval to the final version of this article.

  • Funding This work was supported by TCM Cardiovascular Clinical Medicine Research Center of Jilin Province grant number YDZJ202202CXJD046

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.