Studies of clinical cohorts and retrospective reports have identified psychiatric disorders as paramount risk factors for suicide. Much less is known about how self-reported mental health is related to completed suicide. To study the relation between self-reported mental health and risk of completed suicide, the authors prospectively followed a population-based Norwegian cohort of 61,588 men and 69,774 women aged 39-44 years for an average of 10.4 years between 1994 and 2007. Self-reported mental health was measured using an instrument based on the Hopkins Symptom Checklist and the General Health Questionnaire. Completed suicides were registered in the official Norwegian Cause of Death Registry. Females reported higher levels of mental distress than males. In comparison with persons reporting the fewest mental health symptoms, the adjusted hazard ratio for suicide increased from 1.8 (95% confidence interval (CI): 1.1, 2.9) in the moderately depressed group to 8.9 (95% CI: 4.4, 18.2) in the most depressed group. The risk difference was greatest in males. At each level of the mental health index, males had double the risk of suicide of females (hazard ratio = 2.3, 95% CI: 1.5, 3.3). This study shows a dose-response effect of self-reported mental health problems on completed suicide and replicates the gender paradox observed in the general population with prospective data.