Estimates of the prevalence of congenital heart defects (CHD) have been published over many years and from many regions. As they are based on different definitions of which cases to include in the CHD prevalence, published prevalence estimates vary substantially. With the increasing use of echocardiography in neonatal intensive care, a patent ductus arteriosus (PDA) or flow over the atrial septum will often be visible. These findings may be coded as CHD at discharge and in this way falsely increase the CHD prevalence in the population. There are several purposes for which population-based data on CHD may be used: etiology, planning of treatment, or obtain information on outcome, including mortality. For etiology studies, it is important to include terminations of pregnancy as well as all births with CHD. For mortality studies in live births, inclusion of preterm born infants with PDA will increase overall mortality of CHD. The Danish Register of Congenital Heart Disease is based on hospital discharge diagnoses and diagnoses from outpatient visits. To increase the validity of these data, extensive data cleaning has been carried out based on record review and knowledge on the discharge coding practice. We include PDA and atrial septal defects as CHD cases if these defects are still open 2 months after birth. International consensus on how to define CHD would improve the validity and comparability of epidemiological studies on CHD.