Background. The diagnostic criteria of gestational diabetes mellitus (GDM) have been based on the risk of future maternal diabetes rather than the short-term risk of mother and infant. Our aim was to illustrate the relation between various adverse pregnancy outcomes and maternal glucose levels in women with mild glucose intolerance using a graphical approach. Methods. Observational study of 2,885 pregnant women examined with a 2-h, 75-g oral glucose tolerance test (OGTT) based on risk indicators. Only women with 2-h capillary blood glucose <9.0 mmol/l were included, as women with 2-h values ≥9.0 mmol/l were treated for GDM. Empirical frequencies of adverse outcomes were related to 2-h values by linear and quadratic logistic models. Adjustments for well-known confounders were performed by a multiple logistic model. Results. Linear trends were demonstrated for the outcomes: shoulder dystocia, caesarean section rate (univariate analysis only), spontaneous preterm delivery, and macrosomia (large-for-gestational age infants). None of the outcomes deviated significantly from linearity. No significant trend was found for hypertension or neonatal hypoglycaemia and jaundice. Conclusions. A gradually increasing risk for a number of adverse pregnancy outcomes was found with increasing glucose levels. No obvious threshold value for GDM was demonstrated for 2-h values up to 9.0 mmol/l.