OBJECTIVES: To study the onset and compare risk factors for pure depression (DEP), pure anxiety (ANX), and comorbid anxiety-depression (ANXDEP) in the aging population. DESIGN: Prospective study with 3-year intervals over a 9-year period. SETTING: Data of the Longitudinal Aging Study Amsterdam were used, which is a population-based study among older adults (55-85 years at baseline). PARTICIPANTS: Older adults free of depression and anxiety at baseline (N = 1,712). MEASUREMENTS: Clinically relevant levels of depression and anxiety were measured with the Center for Epidemiologic Studies Depression scale ≥16 and Hospital Anxiety and Depression Scale ≥7, respectively. A broad range of potential sociodemographic, health, and psychosocial risk factors for anxiety and/or depression were examined by using polytomous logistic regression analyses. RESULTS: Within 9 years, 184 subjects (10.8%) developed DEP, 93 (5.4%) ANX, and 103 (6.0%) ANXDEP. Concerning sociodemographics, higher age and lower educational level were predictors for DEP. Health indicators were predictive for DEP and ANXDEP but not for ANX. Depressive symptoms at baseline were predictive for DEP, whereas initial anxiety symptoms were predictive for ANX and ANXDEP. Neuroticism increased the risk of DEP and ANXDEP. Mixed effects of psychosocial variables were found: DEP was associated with recent widowhood, whereas ANX and ANXDEP were associated with other life events such as having an ill partner. CONCLUSION: Although onset of ANXDEP demonstrated communality in risk factors, comparing risk factors associateds with DEP and ANX revealed more differences than similarities. This underlines the need to distinguish anxiety from depression in preventive strategies. © 2009 American Association for Geriatric Psychiatry.