BACKGROUND: Childhood adversity predicts adolescent suicidal ideation but there are few studies examining whether the risk of childhood adversity extends to suicidal ideation in midlife. We hypothesized that childhood adversity predicts midlife suicidal ideation and this is partially mediated by adolescent internalizing disorders, externalizing disorders and adult exposure to life events and interpersonal difficulties. METHOD: At 45 years, 9377 women and men from the UK 1958 British Birth Cohort Study participated in a clinical survey. Childhood adversity was prospectively assessed at the ages of 7, 11 and 16 years. Suicidal ideation at midlife was assessed by the depressive ideas subscale of the Revised Clinical Interview Schedule. Internalizing and externalizing disorders were measured by the Rutter scales at 16 years. Life events, periods of unemployment, partnership separations and alcohol dependence were measured through adulthood. RESULTS: Illness in the household, paternal absence, institutional care, parental divorce and retrospective reports of parental physical and sexual abuse predicted suicidal ideation at 45 years. Three or more childhood adversities were associated with suicidal ideation at 45 years [odds ratio (OR) 4.31, 95% confidence interval (CI) 2.67-6.94]. Psychological distress at 16 years partially mediated the associations of physical abuse (OR 3.41, 95% CI 2.29-5.75), sexual abuse (OR 4.99, 95% CI 2.90-11.16) with suicidal ideation. Adult life events partially mediated the association of parental divorce (OR 6.34, 95% CI -7.16 to 36.75) and physical (OR 9.59, 95% CI 4.97-27.88) and sexual abuse (OR 6.59, 95% CI 2.40-38.36) with suicidal ideation at 45 years. CONCLUSIONS: Adversity in childhood predicts suicidal ideation in midlife, partially mediated by adolescent internalizing and externalizing disorders, adult life events and interpersonal difficulties. Understanding the pathways from adversity to suicidal ideation can inform suicide prevention and the targeting of preventive interventions.