LOGISTIC-REGRESSION MODEL TO ESTIMATE THE RISK OF UNBALANCED OFFSPRING IN RECIPROCAL TRANSLOCATIONS

The aim of this study was to estimate the risk of viable unbalanced of fspring for a parental carrier of reciprocal translocation. On a large computerized database of reciprocal translocations we used logistic r egression to model this risk. The status of the progeny is the outcome variable. Explanatory covariates are cytogenetic characteristics of t he translocation, age and sex of the parental carrier, and potential v iability of the gametes. The results obtained by the logistic model de monstrate the important role of certain variables such as the sex of t he parental carrier and the R band length of the translocated segments . Within the group of lower risk (risk of viable unbalanced offspring less than 5%), 97% of the individuals are correctly classified with th is model. For this group, the choice prenatal diagnosis can be best di scussed by considering both the risk for viable unbalanced offspring a nd the risk of induced abortion following prenatal diagnosis.