International Commission for Protection against Environmental Mutagens and Carcinogens. ICPEMC Working Paper 5/6. Perspectives in mutation epidemiology, 6. A 1983 view of sentinel phenotypes.

A sentinel phenotype is a clinical disorder or syndrome that (1) occurs sporadically as a consequence of a single, highly penetrant mutant gene, (2) is a dominant or X-linked trait of considerable frequency and low fitness, and (3) is uniformly expressed and accurately diagnosable with minimal effort at or near birth. Although 1828 autosomal dominant traits are known in human beings, 36 can be considered as candidate sentinel phenotypes, along with 5 X-linked disorders. Based on surveys of malformations in infants and children, 16 additional traits are proposed beyond previous lists. In Hungary, the 24 syndromes or defects with reliable manifestations in newborn infants occur with a frequency of 2.5-3.3 per 10 000 live births. As markers of human mutations, sentinel phenotypes have the advantage of representing germinal mutations that result in significant health problems. There are severe disadvantages that have, to date, prevented the launching of a field demonstration of the value of these phenotypes in mutation epidemiology. Agreement on a list of phenotypes has been delayed by continued recognition of two or more distinct genetic diseases within what was once thought to be a single disorder. For the same reason, most of the candidate sentinel phenotypes have not been assigned unique codes in the International Classification of Diseases. Each of the disorders is so rare and has features that overlap with so many other syndromes that highly trained clinical dysmorphologist and pediatric ophthalmologists would have to be engaged in any study. The sentinel phenotype approach, like other strategies in mutation epidemiology, would encounter problems with linkage among files of data, privacy, and access to sufficiently large populations. In contrast with the approach using multiple protein variants (as in the study of blood from offspring of survivors of the atomic bombs in Hiroshima and Nagasaki), the sentinel phenotype approach would likely be much less expensive and would encounter far fewer false attributions of paternity, but also would require a much larger study population. The best option for the present, in our opinion, is to broaden and sustain critical discussion of the approach. Perhaps the goal should be to plan a field demonstration by involving appropriate clinicians, epidemiologists, and public health officials. A pilot effort underway in Hungary may well give insight to applying the approach in a significantly larger population.