The aim was to investigate whether personality traits predict onset of the first depressive or manic episode (the vulnerability hypothesis) and whether personality might be altered by the mood disorder (the scar hypothesis). A systematic review of population-based and high-risk studies concerning personality traits and affective disorder in adults was conducted. Nine cross-sectional high-risk studies, seven longitudinal high-risk studies and nine longitudinal population-based studies were found. Most studies support the vulnerability hypothesis and there is evidence that neuroticism is a premorbid risk factor for developing depressive disorder. The evidence for the scar hypothesis is sparse, but the studies with the strongest design showed evidence for both hypotheses. Only few studies of bipolar disorder were found and the association between personality traits and bipolar disorder is unclear. Neuroticism seem to be a risk factor by which vulnerable individuals can be identified, thus preventing the development of depressive disorder. A connection between personality traits and development of bipolar disorder, and evidence of a personality-changing effect of affective episodes need to be further investigated.