“Advance decision making” (ADM) refers to people planning for a future when they may lose the capacity to make decisions about treatment (decision making capacity for treatment or DMC-T). This can occur in a variety of physical and mental health scenarios. Statutory provision for ADM is likely to be introduced to mental health legislation in England and Wales, which will support planning for mental health crises. Conceptually, it may have particular utility for people with Bipolar Affective Disorder (bipolar) due to the pattern of rapid loss and then recovery of DMC-T during episodes of illness. Furthermore, ADM is recommended by clinical experts in bipolar. However, the empirical evidence base for ADM in bipolar is unclear. Therefore, a systematic review is required to collate available evidence and define future research directions. A PRISMA concordant systematic review of empirical literature on the use of ADM in bipolar. We found 13 eligible articles which reported on 11 studies. Of the eligible studies 2 used a mixed methods design, 8 were quantitative descriptive studies and 1 was a randomised controlled trial. Outcomes of included studies fell into 4 categories: Interest in ADM, type of ADM preferred, barriers to completing ADM and impact of ADM. The available evidence suggests that people with bipolar are interested in engaging with ADM which is supported, collaborative and allows them to state treatment requests and refusals. Evidence in this area is limited. Clinicians should be aware that service users with bipolar are likely to value their support in creating ADM documents. In addition, it seems that people with bipolar may face fewer barriers and achieve greater success with ADM compared to those with other severe mental illnesses. Given the greater focus and likely demand for ADM following upcoming legal reform, further research is urgently needed to ensure available resources are most effectively targeted to achieve the best outcomes from ADM activities. This research should focus on clarifying: causal assumptions around ADM, the outcomes which are valued by key stakeholders, barriers to achieving these outcomes, stakeholder opinions on supporting ‘self-binding’ and the development and evaluation of models of ADM which are tailored for fluctuating DMC-T.