The International Clearinghouse for Birth Defects Surveillance and Research conducted a study on very rare defects (VRDs) to test methodologies in their population surveillance and to increase the knowledge of their epidemiology. Eight VRDs: acardia (AC), amelia (AM), bladder exstrophy (BE), cloaca exstrophy (CE), conjoined twins (CT), cyclopia (CY), "true" phocomelia (PH), and sirenomelia (SI) were selected, all of whom showed prevalences in the order of 1/100,000 births, except for BE: 1/48,000 births. Materials in this investigation from 25 million pregnancy outcomes, were provided by 22 Clearinghouse-member programs. The study protocol provided a working definition, a summary of the phenotypic characteristic, and a list of ICD-9 and ICD-10 codes for each VRDs. Learned lessons include: (1) The suspected associations of decreasing risk with advancing maternal age in AM and SI, and increasing risk in BE, and increasing frequency of twins in SI, were confirmed. (2) Morphologically similar defects showed dissimilar epidemiological characteristics, namely, AM and PH, and BE and CE. (3) Heterogeneity in total prevalences for most VRDs among different surveillance programs were attributed to operational reasons, except for SI and CT in which Amerindian ethnicity seems to be associated with higher prevalence. (4) Verbatim description is essential and must be stored in electronic files. In addition to codes. (5) Dysmorphologists or clinical geneticists are an essential part of the coordinating team of the surveillance program. (6) ICD coding system is insufficient. (7) Surveillance programs should be a valuable source of information on exposures to risk factors during pregnancy.