Sleep during oncological treatment – a systematic review and meta-analysis of associations with treatment response, time to progression and survival

Abstract

Introduction: Disrupted sleep and sleep-wake activity are frequently observed in cancer patients undergoing oncological treatment. These disruptions are often associated with aggravated symptom burden and diminished health-related quality of life that in turn may compromise treatment adherence and, thus, effectiveness. In addition, disrupted sleep has been linked to carcinogenic processes, which ultimately could result in worse prognostic outcomes. Aims: Our aim was to systematically review and conduct a meta-analysis of studies examining the associations between sleep and sleep-wake activity and prognostic outcomes in cancer patients undergoing oncological treatment. Methods: A comprehensive systematic search of English language papers was undertaken in June 2020 using PubMed, The Cochrane Library, and CINAHL. Two reviewers independently screened 4,879 abstracts. A total of 26 papers were included in the narrative review. Thirteen papers reporting hazard ratios reflecting associations between a dichotomized predictor variable (sleep) and prognostic outcomes were subjected to meta-analysis. Results: Nineteen of the 26 eligible studies on a total of 7,092 cancer patients reported associations between poorer sleep and poorer response to treatment, shorter time to progression, and/or reduced overall survival, but were highly heterogeneous with respect to the sleep and outcome parameters investigated. Meta-analysis revealed statistically significant associations between poor self-reported sleep and reduced overall survival (HR = 1.33 [95% CI 1.09-1.62], k = 11), and shorter time to progression (HR = 1.40 [95% CI 1.23-1.59], k = 3) and between poor objectively assessed sleep and reduced overall survival (HR = 1.74 [95% CI 1.05-2.88], k = 4). Conclusion: The current findings indicate that disturbed sleep during treatment may be a relevant behavioral marker of poor cancer prognosis. The limited number of studies, the common use of single item sleep measures, and potential publication bias highlight the need for further high quality and longitudinal studies.