Pregnancy is believed to exacerbate diabetes complications, although the degree to which this occurs, and the advice that should be given to women contemplating pregnancy is unclear. We examined 776 nulliparous and 582 parous women with Type 1 diabetes from a cross-sectional study performed in 31 European centres. Glycaemic control was better in parous women. Age and duration adjusted prevalence of microalbuminuria was similar in parous and nulliparous women, but macroalbuminuria was lower in parous women (6% versus 10%, p < 0.0001). Prevalence of all retinopathy was lower in parous women (34% in women who had two or more pregnancies, 45% in women who had one), compared with 48% in nulliparous women (chi 2 for trend = 47.1, p < 0.0001). Proliferative retinopathy was lower in parous (8% and 7%, respectively) compared with nulliparous women (16%, chi 2 for trend = 52.2, p < 0.0001). These differences persisted when adjusted for glycaemic control. Excluding referrals for pregnancy, parous women were more likely to have been referred to the diabetes clinic with complications than nulliparous women (p = 0.001). It is unlikely that our findings can be explained by women with complications being advised against pregnancy, or by the better glycaemic control in parous women. Equivalent levels of microalbuminuria and background retinopathy in parous and nulliparous women suggests that pregnancy may not exacerbate these early complications.