Statistical methods for detecting a moderate paternal age effect on incidence of disorder when a maternal one is present

Abstract

A statistical method is developed for detecting a moderate effect on the incidence of a disorder caused by advancing age of one parent when it is known that advancing age of the other parent is of great aetiological importance. The method is a conditional test procedure given the parental ages, therefore no assumptions about the parental age distributions have to be made. The method is applied on Danish material on Down's syndrome, for which a paternal age effect is demonstrated. Methods used in some well-known previous investigations have been discussed. Several of them, e.g. the classical one of Penrose (1933), could hardly detect any paternal age effect in Down's syndrome on the available data, because these methods are heavily affected by certain fertility patterns not recognized previously.