A multiple pregnancy of high rank may occur in a couple at risk for a Mendelian disorder. Prenatal diagnosis is hampered by the difficulty of (1) obtaining chorionic villi from each zygote and (2) unequivocally relating each sample to the corresponding embryo. The calculation of the genetic risk according to the number of zygotes led us to propose a diagnostic strategy based on embryo reduction, a technique initially designed to improve the perinatal outcome of multiple pregnancies with normal embryos. We report a case in which this approach allowed rational use of first-trimester chorionic villus sampling in a quintuplet pregnancy at risk for non-ketotic hyperglycinaemia, resulting in the selective birth of unaffected twins.