Early intervention approaches are built on the premise of preventing disability, burden, and cognitive sequelae caused by bipolar disorder. The objective of this systematic review was to characterise the effectiveness of all the available psychological and pharmacological treatments for early intervention in people at high risk of developing bipolar disorder. The study was registered with PROSPERO (CRD42019133420). We did a systematic search to identify studies published in ten databases up to March 27, 2020. Randomised controlled trials and cohort studies that assessed the effect of pharmacological or psychological interventions in people at high risk of developing bipolar disorder were included. Studies of first episodes of mania were excluded. Eligible papers were assessed for quality and data were extracted. The primary outcomes were change in manic and depressive symptoms from baseline to endpoint. Of the 2856 citations retrieved by our search, 16 studies were included; five evaluated pharmacotherapeutic strategies (three randomised controlled trials and two open-label studies), ten assessed psychotherapeutic strategies (four randomised controlled trials and six open-label studies), and one randomised controlled trial assessed combination therapy; these 16 trials included a total of 755 participants at high risk of developing bipolar disorder. Quality assessment indicated fair to good quality for open-label studies, and a high risk of bias in four randomised controlled trials. Among the pharmacotherapeutic interventions, there is preliminary support for the efficacy of aripiprazole in reducing mood symptoms in people at high risk of developing bipolar disorder. Psychological interventions were effective for various outcomes. There was substantial methodological heterogeneity across studies. This systematic review underscores the need for multicentre, prospective, methodologically homogeneous studies evaluating conversion to bipolar disorder as an outcome measure.Copyright © 2020 Elsevier Ltd. All rights reserved.