Our study has several important limitations: (1) the data are based on death certificates, which are known to be variably complete and sometimes inaccurate; (2) Down syndrome cases may be incompletely ascertained. We observed about half the number of Down syndrome-associated deaths that would be expected in this population, based on estimates of Down syndrome prevalence at birth. However, given the increasing median age at death during this time period, it is difficult to assess the level of underascertainment in our study; (3) possible differential reporting of Down syndrome by diagnosis may have occurred. For example, a physician may be more likely to note Down syndrome on the death certificate if the person died of a cause well-recognized to be related to Down syndrome (such as congenital heart disease), than if the a person dies of a cause not known to be related to Down syndrome (such as breast cancer).