This study was undertaken to find a strict, unbiased epidemiological delineation of the VATER non-random association of congenital malformations and, based on registry information, to identify a group of probable VATER association infants suitable for etiological analyses. Information on 5,260 infants with multiple malformations was collected from four large registers of congenital malformations. Data were analyzed using a statistical method in which various putative confounders were controlled for. Our results indicate the existence of a distinct group of malformations corresponding to the VATER association: esophageal atresia, anal atresia, upper preaxial limb reduction defects, and costo-vertebral malformations. A subdivision into an upper and a lower group of VATER association was indicated, with heart malformations associated with the upper group and kidney malformations associated with the lower group. Restricting the inclusion criteria for VATER association to the above mentioned core malformations, few infants seem to belong to the VATER association, thus limiting the possibilities of carrying out etiological analyses. A relatively large number of infants may belong to a family of related conditions among which VATER association is a subgroup. In the search for risk factors, a strict definition of the VATER association is needed in order to not dilute the study material with irrelevant cases. The present study provides such strict inclusion criteria. © 2001 Wiley-Liss, Inc.