Get started
Work

Can sleep predict relapse in Bipolar Disorder?

Year: 2025
Type: dissertation
Abstract: Bipolar disorder is a long-term, serious mental illness affecting around 2% of the global population, and carries a high burden of disability, especially among young people. Symptoms vary in severity and duration, but most patients spend about 50% of their time in a mood state - either depression, hypomania, or mania. Episode lengths differ, but the recurrence rate is approximately 50%, rising with more severe symptoms. Sleep patterns change depending on the episode. In depression, patients may experience hypersomnia or insomnia, but generally feel more tired than usual, whereas a reduced need for sleep characterises mania. In the general population, shorter sleep duration or poorer sleep quality is associated with a less positive mood the following day. This connection also appears prevalent between sleep and mood in bipolar disorder. Changes in sleep duration are among the most common warning signs reported by patients before a manic episode, whereas specific changes in sleep are less evident before depression. Early and effective intervention improves clinical outcomes and decreases recurrence risk. Identifying reliable early warning signs for both mood changes and episode recurrence would improve treatment outcomes and enhance patients' quality of life. Questions and hypotheses: The central research question is: Can daily collection of sleep variables be used to predict mood changes in bipolar disorder? I set three objectives to answer this question: 1) To determine whether there are detectable changes in sleep linked to changes in mood; 2) To investigate whether a relationship exists between sleep patterns and clinically assessed mood episodes in individuals with BD; 3) To examine whether there are daily correlations between fluctuations in sleep and mood in individuals with BD, based on self-reports. My hypotheses were based on published literature, suggesting links between sleep and mood in bipolar disorder, as well as the common sleep patterns reported during depressive and manic episodes. However, the variable findings suggest that the links might not be universal for all patients or for both mania and depression, and that further analyses, such as examining standard deviations, could reveal the most common relationships. Material and methods: Three separate methodologies were employed to address the three research objectives: first, a systematic literature review was conducted (PROSPERO: CRD42023405950) in accordance with PRISMA guidelines, selecting studies of bipolar disorder investigating correlations between sleep and mood data collected for at least 3 weeks. Then, two secondary data analyses were performed using the Biposense study dataset. The study monitored 29 bipolar outpatients in remission for one year, requiring participants to record daily their sleep over the previous 24 hours and their mood from 0-100 (depressed to manic). Additionally, patients underwent clinical assessments of mood states every two weeks. Two approaches were used to examine the dataset: 1) Multi-level logit models applying person-centred sleep data were used to investigate the likelihood of patients being in the prodromal phase or in a new episode compared to being in an euthymic state; 2) Cross-correlation functioning and time series analyses were used to determine how many patients showed a significant relationship between changes in their sleep patterns and mood. Results: Systematic review: 17 publications (13 studies) were included, out of 7159 records identified. Nine categories were created to summarise the relationship between sleep and mood. Overall, changes towards longer sleep duration, earlier time falling asleep, and later time waking up were mostly followed by depressive mood, whereas the opposite was found before manic mood. However, when looking at the correlation between sleep and depressive mood, few findings were statistically significant. Multi-level modelling: Compared to the euthymic state, in the week(s) leading up to a depressive episode, patients were more likely to wake up later than usual and to show greater variability in total sleep duration, total time spent in bed, and time of waking up. In the week(s) leading up to a manic episode, patients were more likely to have less variability in the time they fell asleep and woke up. Cross-Correlation functioning: 41% of participants showed a significant negative cross-correlation between the time spent in bed and self-reported mood the following day. Furthermore, one-fourth of participants showed a significant negative cross-correlation between average weekly total sleep duration and subsequent mood changes. In contrast, few participants showed relationships between 7-day variability in sleep behaviour and mood changes. Conclusion: Specific sleep patterns characterise the period before changes in either mood symptoms or episode recurrence. Overall, longer sleep duration, more time spent in bed, and variable awakening times indicate a shift towards depressive mood or the onset of a depressive episode. Conversely, shorter sleep, less time in bed, and reduced variability of wake-up time signal a shift towards a more manic mood or episode. Using a linear 0-100 mood scale appears helpful for identifying correlations with sleep duration. The absence of statistically significant findings in the systematic review on sleep duration and depressive mood may be explained by the Biposense results, which found that only one-third of participants showed clear correlations between sleep duration and mood. Patients with bipolar disorder can differ in their sleep patterns during depression: some experience hypersomnia, others insomnia. These differences in individual clinical presentation may explain why the sleep-mood correlation is less evident on a group level, and why the correlations for depression were less clear than those for mania. Future research should focus on identifying correlations between sleep and mood, stratified by baseline sleep characteristics, such as insomnia or hypersomnia. Furthermore, studies that implement changes and early intervention based on identifying early warning signs, including changes in sleep patterns, are necessary to determine if tracking these symptoms truly benefits patients and improves clinical outcomes. My research identifies sleep patterns that suggest an imminent change in mood symptoms or relapse risk. My findings may aid clinical care in identifying reliable early warning signs of worsening symptomology.:Table of Contents Abbreviations 1 Figures list of content 3 Tables list of content 4 Chapter 1: Introduction 5 1.1 Research rationale 5 1.2 Epidemiology and classification 5 1.3 Clinical symptoms and diagnostic criteria 7 1.4 Natural course of BD 9 1.5 Historical background 11 1.6 Treatment and care 12 1.7 Sleep and circadian disturbances in BD 15 1.8 Sleep research: Tools and objectives 21 1.9 Sleep and mood correlations 24 1.10 Research objectives and hypotheses 25 Chapter 2: Methods 26 2.1 Research design 26 2.2 Systematic review (Chapter 5) 27 2.3 The BipoSense study (Chapters 7 and 9) 29 2.4 Ethical considerations and transparency 34 Chapter 3: Statement of contributions to publications 35 3.1 Publication 1 contribution 35 3.2 Publication 2 contribution 35 3.3 Publication 3 contribution 36 Chapter 4: Publication 1 Abstract 38 Chapter 5: Publication 1 PDF 38 Chapter 6: Publication 2 Abstract 39 Chapter 7: Publication 2 PDF 39 Chapter 8: Publication 3 Abstract 40 Chapter 9: Publication 3 PDF 41 Chapter 10: Discussion 41 10.1 Overview 41 10.2 Key findings 41 10.3 Methodological considerations and limitations 42 10.4 Integration and synthesis 46 Chapter 11: Conclusion 49 Chapter 12: Summary 50 12.1 English 50 12.2 German 52 Chapter 13: References 56 Chapter 14: Appendices 78 14.1 Declaration of autonomy, Anlage 1 78 14.2 Declaration of compliance, Anlage 2 81 14.3 Thesis supplementary materials 82
Language: English
Field: Medicine
Open Access status: green
OpenAlex ID: W7169635411
Cites:
Cited by:
Related to: 10