Psychological and behavioural interventions in bipolar disorder that target sleep and circadian rhythms: A systematic review of randomised controlled trials

•Sleep and circadian rhythm disruptions are prominent symptoms of bipolar disorder and promising treatment targets.•Previous reviews have focused on non-randomised uncontrolled trials or trials that largely recruited patients with unipolar depression.•Our review suggests that the evidence base for such interventions is limited. All studies assessed mood, but more than half did not assess sleep or circadian rhythms.•The effect of bright light therapy on depression was the only domain with enough data to be meta-analysed (Nc = 6; g=-0.74 [95 % CI=-1.05 to -0.42], p < 0.001).•There is a need for adequately powered trials that incorporate comprehensive measures of sleep, circadian rhythms, mood, and functioning. Sleep and circadian disruptions are prominent symptoms of bipolar disorder (BD) and potential targets for adjunctive interventions. The aim of this review was to appraise the effectiveness of psychological and behavioural interventions in BD that target sleep and circadian rhythms, as reported by randomised controlled trials. Nineteen studies met the inclusion/exclusion criteria. They were summarised via narrative synthesis and meta-analysis wherever appropriate. Six studies delivered bright light therapy, five interpersonal and social rhythm therapy, two blue-light blocking glasses, one cognitive behavioural therapy for insomnia, one total sleep deprivation, and four combination treatments. More than half of the studies (N = 10, 52 %) did not measure sleep or circadian rhythms despite being the principal target of the intervention. Overall, the evidence base for the effectiveness of these interventions was limited. There was a small number of studies for each intervention, and a lack of consistency in protocols and outcomes. Meta-analysis was possible for the effect of bright light therapy on depression, revealing a medium-to-large post-treatment effect (Nc = 6; g=-0.74 [95 % CI=-1.05 to -0.42], p < 0.001).