The ability of birth defects monitoring to detect new human teratogenic and mutagenic agents may be limited if only isolated defects are monitored. Surveillance for "new" multiple congenital anomalies (MCA) may improve the detection of environmental agents associated with new defect patterns. To evaluate the feasibility of such monitoring, we examined data from two programs: 1) the Metropolitan Atlanta Congenital Defects Program (MACDP), which ascertains infants with serious defects diagnosed in the first year of life, and, 2) the Italian Multicenter Register for Congenital Malformations (IPIMC), which ascertains newborn infants with birth defects from many hospitals in Italy. We focused on 24 relatively serious defects and defect groups. For a baseline period (MACDP: 1968-1988, 581,000 births; IPIMC: 1986-1989, 448,000 births), we identified all possible combinations of defects occurring in the same baby. For a test period (MACDP: 1989-1990, 77,000 births; IPIMC: 1990, 91,500 births), we identified babies with "new" MCA (i.e., combinations of defects not observed before in the system). During this period in MACDP, of the 85 babies with two or more defects, 9 babies had new MCAs. In IPIMC, of the 54 babies with two or more defects, 10 babies had new MCAs. A review of the records of infants with new MCAs in MACDP and IPIMC did not identify commonalities in maternal characteristics. This analysis illustrates the feasibility of monitoring for new MCAs in surveillance systems. This approach, complemented by an evaluation of exposures, may be a powerful additional tool in searching for human teratogens and mutagens.