Increasing incidence of Down's syndrome with advancing paternal age for given maternal age has been demonstrated. Comparisons are made between an almost complete Down's syndrome sample from the Copenhagen Metropolitan Area and a randomly selected sample of births from the same area and the same time period. Men above 55 years have a significantly increased risk of getting children with Down's syndrome.

The aim of the study was to obtain incidence figures for Down's syndrome throughout a period where a considerable change in the age distribution of child-bearing mothers has taken place and to study if the expected fall in incidence has occurred. In parts of the Copenhagen Metropolitan area 235 liveborn patients with Down's syndrome were ascertained in the period 1960 to 1971 in a population of 1-2 million with a total of 204771 births. All patients available were examined cytogenetically (75%). In 160 (90-4%) a regular trisomy 21 was observed. In 6-2% of the cases translocations and in 2-3% of the cases mosaics were found. Two double trisomies and a double trisomy mosaic were observed. Throughout the period 1960-71 the percentage of women over 30 years delivering children decreased from 23-4% in the beginning of the period to 16-2% at the end of the period. In the first part of the period 52-6% of the cases were born to mothers over 30, at the end of the period 40% of Down's syndrome mothers were of that age. However, the incidence was unchanged throughout the whole period, about 1-15 per 1000 births. For some age groups a steady rise in incidence of trisomy 21 cases was found throughout the whole period. These findings may be explained by better ascertainment of patients at the end of the period; however, environmental factors may also play a role.

From prenatal diagnosis data obtained on mothers aged 35 years and above in the Federal Republic of Germany (DFG data), older fathers are demonstrated to have an increased risk of having trisomy 21 offspring. For paternal ages of 41 years upward, the age effect is quite strong. The risk for a fetus to have any de novo chromosomal aberration increases more with advancing paternal age for older mothers than for younger ones. Thus the ages of both parents have to be taken into account as an indication for prenatal diagnosis. Risk figures for trisomy 21 and for any de novo chromosomal aberration are given, together with preliminary recommendations for prenatal diagnosis for different combinations of parental ages.

Based on 2890 prenatal diagnoses from 12 European countries the risk for a chromosomally abnormal fetus at amniocentesis after the birth of a child with a chromosome abnormality has been estimated to be 1.3 per cent when the mother's age is 34 years or less at amniocentesis and 1.8 per cent if the mother is older. This risk does not depend on paternal age, and it is independent of the type of the chromosome abnormality of the index child. Some geographical heterogeneities were detected. Therefore, the overall risk has to be considered as a rough estimate. The chromosome constitution of the abnormal fetus differed from that of the index patient in 21 of 41 cases. Several explanations for the higher risk have been discussed. If the index child had trisomy 18, 13 or a sex chromosome abnormality, the fetus tended to be a female. If the index child was a trisomy 21, the fetal sex ratio was normal.

This paper reanalyzes the data from the Hook and Cross (1982) paper in this journal concerning the association between Down's syndrome and paternal age. The New York State (NYS) data are compared with a large European collaborative study by Ferguson-Smith and Yates (1984). The maternal-age-dependent risks in the NYS data were found to be significantly higher than in the European data. When the NYS data was divided into three groups by means of the paternal age, a marked two-peaked distribution was found. The maternal-age-dependent risk was high when the fathers were up to 33 years old, low when the fathers' ages were 34-39 years and high again when the fathers were at least 40 years old. The differences were significant. The results speak in favour of the existence of temporal, geographic, or environmental variations in the risk for de novo trisomy 21, as well as of a paternal age effect. The existence of a "paternal age effect" in at least some populations is confirmed. If the results of this paper are confirmed in other investigations, it will be necessary to revise present genetic counselling rules towards far more individually specified considerations.