From 1978 to 1986 a total of 189 pregnant diabetic women gave birth at our hospital. In this randomized prospective study the influence of maternal diabetes treatment in normoglycemic patients, continuous subcutaneous insulin infusion (n = 48) versus intensified conventional treatment (n = 41), is evaluated. These two groups of patients are further compared to patients (n = 28) who underwent conventional diabetes treatment during pregnancy. It can be shown from our data that the rate of complications such as preeclampsia, intrauterine growth retardation, premature labor and premature delivery can be reduced by intensified conventional and insulin pump treatment as compared to conventionally treated patients with late onset of pregnancy care. As expected, in the groups of CSIIand ICT patients no difference in the rate of pregnancy complications nor in fetal outcome could be demonstrated. Among CSII pregnancies 12/48 were complicated, in the ICT population the respective figure was 13/41 (CT: 20/28). The mean gestational age at the time of delivery ranged between 38 and 40 weeks, depending on the severity of maternal diabetes. CT patients were delivered earlier in all White classes. Fetal morbidity was nearly equal in CSII and ICT children, in CT patients it was greatly enhanced. Also the mortality (perinatal and neonatal) was considerably larger in CT patients (6/28), again, in the CSII and ICT population the mortality was nearly identical (2/48 and 3/41). We conclude, from our prospective information, that insulin pump therapy during pregnancy is indicated if intensified conventional treatment does no lead to normoglycemia. This may be the case if maternal diabetes is complicated by dysregulated insulin resorption, prolonged duration of diabetes, or inconsistant working or living conditions. It could be demonstrated that the higher primary costs of ICT as well as CSII therapy are very effective as compared to conventional treatment.