Background: Late-life depression is associated with premature mortality, however, little is known whether excess mortality rates of depression have changed over time. This study aims to identify and explain secular trends in excess mortality of major depressive disorder (MDD) and subthreshold depression (SUBD). Methods: Cohort-sequential-longitudinal study of 4084 community-dwelling older adults in the Netherlands based on data from the Longitudinal Aging Study Amsterdam (LASA). Six measurement cycles were included from 1992/93 until 2008/09, each linked to the overall 5-year mortality, covering a 16-year time span. MDD and SUBD were identified using a two-stage screening procedure with the Center for Epidemiological Studies Depression Scale and the Diagnostic Interview Schedule. Age and sex were covariates. Education, health and lifestyle factors, and use of antidepressants were included as putative explanatory factors. Generalized Estimating Equations was used to investigate the association between the interaction ‘Depression × Time’ and 5-year mortality, and to find explanatory factors for the trend. Results: A downward trend in excess mortality of MDD was found (OR =.92, 95%-CI:.85–.99, P =.04), adjusted for age and sex, which could not be explained by education, health and lifestyle factors, nor antidepressants use. Sex differences in the trend were not found (P =.77). No trend in excess mortality of SUBD was found (OR = 1.01, 95%-CI:.97–1.04, P =.65). Limitations: The findings do not imply a similar trend for other countries. Conclusions: The results indicate a favorable development in excess mortality of community-dwelling older adults with MDD, while those with SUBD do not show a clear trend in excess mortality.