Introduction
Insomnia is a prevalent issue among cancer survivors, with studies revealing that nearly 60% of them experience insomnia during the perioperative phase, and this prevalence persists even beyond 1 year.1 Nearly 15% of them had a first episode of insomnia, and 19.5% experienced a relapse of insomnia before or after surgery.2 Perioperative insomnia, along with clustered symptoms such as anxiety, depression and pain, has been found to be significantly associated with poorer sleep quality, overall health status, psychological distress and prognosis.3–5 Therefore, it is imperative for healthcare providers to effectively address insomnia during the perioperative period to improve their life quality and reduce adverse post-surgery outcomes for cancer survivors.
Although pharmacotherapy has traditionally been the main insomnia treatment, it can cause various side effects, including withdrawal effects, drug addiction, rebound insomnia, depression and anxiety, cognitive impairment and an increased risk of overall mortality.6 Preliminary findings suggested that certain non-pharmaceutical interventions including cognitive behavioural therapy for Insomnia (CBT-I), exercise, various forms of complementary medicine (CAM), such as hypnosis, acupuncture, massage, among others, may demonstrate potential efficacy in addressing perioperative insomnia for patients with cancer.7–13 However, it is essential to note that while some studies show promising results, the evidence regarding the effectiveness of these non-pharmaceutical interventions modalities in the perioperative setting remains varied and inconclusive.7–13
CBT-I has been recommended as the first-line intervention for cancer-related insomnia. The application during the perioperative period has shown improvements in sleep quality for patients with laryngeal carcinoma.7 As for preoperative exercise training on sleep outcomes in patients with cancer, the result showed that exercise can provide better sleep duration and efficiency compared with usual care measured by the accelerometer.8 However, some studies using subjective measurements of insomnia did not find significant benefits from preoperative exercise interventions.9 10 Besides, in comparison to standard care, preoperative electroacupuncture (EA) treatment, both alone and in combination with intraoperative EA treatment, has shown promising results, including quicker recovery times, improved sleep efficiency, longer light sleep and increased total sleep time.11 Similarly, studies involving hypnosis have demonstrated its effectiveness in reducing insomnia and pain during and after surgery.12 In addition, massage also has been found to enhance sleep quality and reduce anxiety levels in patients undergoing colorectal surgery.13 It can be seen that the existing research on the efficacy of a certain treatment method is inconsistent, such as the impact of exercise on sleep among patients with cancer. Additionally, although many studies have explored the effects of CBT-I, EA, exercise, massage and hypnosis on sleep in patients with cancer during the perioperative period, no studies have compared the efficacy between them.
In this scenario, it is crucial to determine the comparative efficacy of different non-pharmaceutical interventions. However, there is currently limited evidence available, with only one systematic review (SR) focusing on exercise management for patients with cancer with insomnia during the perioperative period.14 Moreover, it is difficult to use simply individual randomised controlled trials or even pairwise meta-analysis to address this issue.
Network meta-analysis (NMA), also known as multiple treatments comparison meta-analysis, offers a valuable approach for synthesising information from various clinical trials, incorporating both direct and indirect data when head-to-head treatment comparisons are lacking. NMA provides a higher degree of precision through rank probability, which can have significant implications for treatment decisions in clinical practice.15 Thus, NMA can facilitate the identification of the most effective perioperative non-pharmaceutical interventions for cancer survivors struggling with insomnia.
Accordingly, this investigation may be the first SR and NMA examining the effectiveness and safety of perioperative non-pharmaceutical interventions, primarily using the Pittsburgh Sleep Quality Index (PSQI) or Insomnia Severity Index (ISI), for cancer insomnia survivors. The findings of this study will compare the therapeutic effects of various non-pharmaceutical therapies and establish a hierarchy of efficacy to serve as a foundation for informed treatment decisions.
Methods and analysis
Study registration
This protocol was registered on PROSPERO and conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Protocols statement guidelines (online supplemental file 1).
Supplemental material
Inclusion criteria
Types of studies
Our database will encompass randomised and/or controlled clinical trials, and/or quasi-experimental (clinical studies without a control and/or comparator group). Full-text journal articles and unpublished clinical trials with accessible results online will all be included. To ensure the inclusion of all relevant reference materials, a thorough examination of previous SRs and meta-analyses will be conducted.
Participants
This study will include adult patients aged 18 years and older, irrespective of cancer type or stage, who are candidates for surgical intervention (only surgical resections), either during or after neoadjuvant treatment, as well as those scheduling directly to surgery without prior systemic treatment. Eligible participants will have self-reported insomnia at baseline prior to surgery. There will be no restrictions based on race, nationality or educational background.
Types of interventions
Non-pharmaceutical interventions employed during the perioperative period will be limited to the following three categories, encompassing various aspects: (1) Psychological interventions: CBT-I, mindfulness, meditation and hypnosis; (2) physical interventions: massage, exercise (including aerobic exercise, resistance exercise and walking); (3) mind-body interventions: acupuncture (including acupuncture, EA and auricular acupuncture), yoga, tai chi and qi gong. For inclusion in the study, eligible studies must use one or more of these interventions. Besides, patients should receive the intervention either prior to surgery or across both preoperative and postoperative phases.
Types of comparator(s)/control
The comparison group included usual care, sham intervention, waiting control, placebo and sleep education.
Primary outcomes
The change of PSQI or ISI from baseline will be used as the primary outcome indicator in this study.
The PSQI instrument generates seven component scores that assess various aspects of sleep, including daytime dysfunction, habitual sleep efficiency, latency, duration, sleep disturbances and sleep quality. Scores on the PSQI range from 0 to 21, with higher scores indicating poorer sleep quality and more sleep disruption.16
The ISI is used to evaluate the types and manifestations of sleep problems. It encompasses the severity of symptoms, the individuals’ satisfaction with their sleep patterns, the impact of insomnia on their daily functioning, their awareness of the influence of insomnia on their well-being and the level of frustration resulting from their insomnia. Scores on the ISI range from 0 to 28, with higher scores indicating more severe insomnia.17
Secondary outcomes
-
The change of sleep efficiency (SE) from baseline.
SE is obtained through various methods, including polysomnography data, sleep diaries or actigraphy.18–20
-
The change of Richards-Campbell Sleep Questionnaire (RCSQ) from baseline.
The RCSQ is a five-item questionnaire designed to evaluate perceived sleep depth, sleep latency, frequency of awakenings, SE and sleep quality. Each item of the RCSQ is scored on a 100 mm Visual Analogue Scale, with higher scores indicating better sleep. The average score of these five items, referred to as the ‘total score’, represents the overall perception of sleep.21
-
The change of European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 (EORTC QLQ-C30) from baseline.
The EORTC QLQ-C30 includes an item specifically assessing sleep disturbances: ‘During the past week, have you had trouble sleeping?’ Participants rated this item on a scale from 1 to 4, with ‘not at all’ coded as 1 and ‘very much’ coded as 4. Higher scores indicate greater problems with sleep disturbance.22
-
Quantitative statistics of adverse events will be considered as a secondary indicator in the analysis.
Furthermore, in studies with pre-surgery interventions, outcome assessments followed treatment and preceded surgery. For interventions both pre-surgery and post-surgery, assessments occurred after pre-surgery treatment and subsequently after post-surgery treatment.
Exclusion criteria
First, any duplicate studies will be excluded. Second, studies involving patients with comorbid conditions known to affect insomnia (including but not limited to major depressive disorder, anxiety disorders, post-traumatic stress disorder and substance use disorders) will not be considered. Third, abstracts, editorials, clinical observations, case studies, cohort studies, case–control studies and surveys that are not experimental, including cross-sectional and retrospective research, will be excluded. Additionally, we did not include studies with no full-text available. In conclusion, any study meeting one or more of the aforementioned criteria will be excluded from the analysis.
Search methods for identification of studies
We will search for all publications from the inception of the database to 1 December 2023. The literature databases include PubMed, MEDLINE, Embase, Web of Science and Cochrane Central Register of Controlled Trials and the Chinese literature databases include CNKI, VIP, Wanfang. The search strategy will focus on specific keywords related to the study topic such as ‘perioperative’, ‘cancer’, ‘insomnia’, ‘non-pharmaceutical’, ‘CAM’, ‘complementary and integrative medicine (CIM)’, ‘acupuncture’, ‘electroacupuncture’, ‘auricular acupuncture’, ‘cognitive behavioral therapy’, ‘mind-body therapies’, ‘mindfulness’, ‘meditation’, ‘massage’, ‘yoga’, ‘tai chi’, ‘qigong’, ‘exercise’, ‘resistance training’, ‘walking’ and ‘hypnosis’. The final search formula will be derived by incorporating subject words, subordinate words and free words to ensure comprehensive coverage. Additionally, the reference lists of the selected articles will be meticulously examined to identify any additional relevant studies that might have been missed during the initial search process. An example search strategy for Web of Science is shown in table 1. The search strategies of all the databases are shown in online supplemental file 2.
Supplemental material
Search strategy for the Web of Science
Data collection and analysis
Selection of studies
Two independent researchers (LL and QJ) will conduct the literature screening, duplicates elimination and data extraction, ensuring the accuracy and reliability of the findings with EndNote V.X9. The screening process will commence by reviewing the article titles, the abstracts and full texts of the remaining articles to determine their eligibility for inclusion. In the event of any discrepancies, a third party (YC) will be involved. The entire selection procedure will be presented transparently in a PRISMA flow diagram, adhering to the recommended guidelines for reporting meta-analyses (figure 1).
Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram of the study selection process.
Data extraction and management
Data extraction will be done using a predetermined form (QJ and LL). The extracted data will cover various aspects of the study, including the first author’s name and publication year for basic information. The research methods, encompassing study design, intervention and control measures and sample size, will also be recorded. Additionally, relevant outcome indicators and their corresponding measurement data will be included, along with any available follow-up information.
In the event of missing data, direct communication with the authors of the original studies will be made to gather any important information. In case of failure, missing data will be tackled systematically: using available data sources, employing appropriate imputation methods and performing sensitivity analyses. For the imputation methods, the maximum SD from a similar study with the same measurement will be used when variability measures are missing. The correlation method will be applied when results are available for baseline and follow-up but not for mean change.23 Following the completion of data extraction, the two researchers will cross-check the extracted data for accuracy and consistency. Should any disagreements arise, they will be resolved through team discussions or by seeking consultation with a third researcher (YC).
Quality assessment
The Cochrane tool for assessing the risk of bias in randomised trials will be used by LL and PY.24 For non-randomised studies, the risk of bias will be evaluated by Risk of Bias in Non-randomised Studies-of Interventions.25 Justification for the inclusion, total number of studies in the network, method used for incorporating of non-randomised studies will be provided in the result. The quality of each study will be evaluated based on five key areas: bias in the randomisation process, deviation from established interventions, bias in missing outcome data, bias in outcome measurement and selective reporting of results. To grade the quality of evidence for the main outcomes, LL and PY will use the Grading of Recommendations Assessment, Development and Evaluation (GRADE) system. The evidence quality will be categorised as ‘high’, ‘moderate’, ‘low’ or ‘very low’ according to the GRADE rating standards. The rating process involves presenting effect estimates and rating the quality from direct, indirect and NMA (combined direct+indirect) evidence. Besides, the GRADE rating will be regulated according to the risk of bias, inconsistency, indirectness and imprecision of the trials.26
Meta-analysis
Meta-analysis will be conducted using Stata V.15.0 software. Conventional pairwise meta-analyses will be performed using a random-effects model, Hartung-Knapp-Sidik-Jonkman approach. Continuous variables will be assessed using either standardised mean differences or weighted mean differences, accompanied by 95% CI. For dichotomous variables, the pooled OR with a corresponding 95% CI will be calculated. The presence of heterogeneity among trials will be evaluated using the I² statistics and p value. According to the Cochrane Handbook for Systematic Reviews of Interventions Interpretation of I²: 0%–40% (might not be important); 30%–60% (moderate heterogeneity); 50%–90% (substantial heterogeneity); 75%–100% (considerable heterogeneity). P value<0.05 indicates statistically significant heterogeneity. For potential unexplained heterogeneity, we will calculate the prediction intervals.27 In order to investigate potential sources of heterogeneity, subgroup analysis and meta-regression will be employed. The NMA will be conducted to compare the effectiveness of different interventions and controls within a Bayesian framework, with a p value of <0.05 indicating a statistically significant difference between direct and indirect multiple treatment comparisons. Within the Bayesian framework, we will review prior research on the impact of non-pharmaceutical therapies on insomnia in patients with cancer during the perioperative period to understand the approximate range of effects and establish prior specifications. Besides, by comparing the credible intervals of outcome measures such as the PSQI and ISI across different non-pharmacological therapies, we can assess the differences and uncertainties in their effects. If the credible intervals of two treatments do not overlap, or if the overlap is minimal, it indicates a significant difference in their effects. A network graph will be employed to enhance the visual representation of the interconnectedness among outcomes and facilitate the analysis of both direct and indirect comparisons. To assess publication bias, Egger’s test and funnel plots will be used if the number of studies exceeds a certain threshold. Local inconsistency at the network level will be examined using a node-splitting model. Based on the results, a consistency or inconsistency model will be selected. We will choose the consistency model if global and local tests indicate no significant inconsistency, fit statistics support the consistency model and heterogeneity is low to moderate. Besides, we will select the inconsistency model if tests reveal significant inconsistency. Convergence of the results will be evaluated by analysing the potential scale reduction factor (PSRF), where a PSRF value close to 1 signifies successful convergence.28
Subgroup analysis, meta-regression analysis and sensitivity analysis
Subgroup analysis and meta-regression analysis will be employed to investigate potential sources of heterogeneity and inconsistency. Subgroup analysis will be conducted based on various types of non-pharmaceutical interventions, the time point and quantity of treatments received by patients, SE measured by polysomnography, sleep diaries or actigraphy and different cancer diagnoses. Additionally, a meta-regression analysis will be performed, considering surgical candidates who underwent or were treated after neoadjuvant therapy or scheduled directly to surgery as covariates. Furthermore, sensitivity analysis will be carried out using exclusion methods to assess the robustness of the findings. Studies with a high risk of bias, poor quality, the presence of outliers or lack of randomisation will be excluded. We will restrict the primary analysis to results of studies only at low risk of bias.
Publication bias assessment
To assess the presence of publication bias, a comparison-adjusted funnel plot will be generated if the number of included studies exceeds 10. This plot will aid in detecting any potential reporting bias and provide a comprehensive evaluation of the study outcomes. Additionally, Egger’s test will also be employed to assess 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
Ethical approval is not required for this study since it involves the analysis of existing studies. The anticipated results will be disseminated through publication in a peer-reviewed journal.
Discussion
Surgical resection is universally acknowledged as the gold-standard curative procedure for early-stage cancer, offering the highest likelihood of optimal recovery and treatment success. However, the presence of surgery-related sleep disruption significantly affects patients’ overall well-being.5 Particularly, the manifestation of perioperative insomnia has been consistently linked to adverse consequences, including protracted recovery periods and diminished life quality.4
In recent years, non-pharmaceutical interventions have emerged as promising modalities for managing cancer-related insomnia due to their advantages of fewer side effects. Notably, interventions such as CBT-I, acupuncture, exercise and yoga have gained recognition for their potential efficacy.29–32 Although insomnia is common before and after surgeries, there are limited studies on this topic and no guidelines to manage this condition. Apart from a single SR investigating the impact of preoperative exercise training on sleep disturbances in patients with cancer, no other intervention has been subjected to a comprehensive review.14 Consequently, clinicians and patients encounter significant challenges when assessing treatment options.
To our knowledge, this study will be the first SR and NMA to synthesise the full body of high-level evidence and examine the effectiveness of several non-pharmaceutical interventions during the perioperative period for cancer-related insomnia. The presenting results for all outcomes of clinical importance will assist clinicians, policymakers, researchers and patients with cancer with insomnia in decision-making on how to manage insomnia and its associated symptoms in the best way.
This post was originally published on https://bmjopen.bmj.com