Edinburgh experience is described of 329 consecutive deliveries in diabetic women, 90% of whom required treatment with diet and insulin. The high perinatal fetal mortality of 20·9% in this unselected series includes a high proportion of deaths which occurred before the start of labour, sometimes in women who had failed to seek antenatal care. 98% of all survivors have been examined at intervals for up to 18 years. A necropsy rate of 94% was achieved by two paediatric pathologists. Lethal malformations were present in 1·5% of all births compared with 2·1% in Copenhagen and 2·2% in Boston. These affected the central nervous system in Edinburgh, but the heart was much more affected in the other two centres. The incidence of significant malformations was 7·9% in the whole series and 8·8% in all who survived the first month. This figure may be higher than in a normal population but further comparison with a simultaneous control series was impossible. Not only the incidence but perhaps the type of malformation may vary in different communities. This may result in part from genetic factors, but other possible causes exist. Maternal diabetic vasculopathy is thought to correlate with the occurrence of congenital malformations in Copenhagen. Pregnancy associated with diabetic degenerative disease is rare in Edinburgh, and this could account for differences in incidence and quality. Such a correlation, however, does not exist in Boston, and the only Edinburgh mother who produced a child with sacral agenesis was still free from obvious degenerative disease 16 years later. Symptomatic hypoglycemia during early pregnancy does not correlate with the production of malformed babies in Copenhagen, but it does exemplify the kind of fetal insult which could be commoner in one centre than in another. No conclusive evidence exists to implicate this or any other iatrogenic factor. The present weights and heights of most surviving children fall within the normal distribution of weight and height for age and sex. There are, however, more children than expected whose heights fall below the 3rd centile. A weight-to-height index has been calculated for each child, and an index of 1·25 or more is taken to represent excessive weight for height. High indices are common and low indices are uncommon. The incidence of high indices increases progressively with age and is greater in girls than in boys. The incidence so far of juvenile clinical diabetes among survivors is 0·77 per 100, which is already 22 times greater than in the general population under 16 years of age. Neither of the two cases recorded has a high weight-to-height index: one is short and thin, the other is of average build.
