We aimed to compare pregnancy outcomes in 4665 women according to the following types of hyperglycaemia in pregnancy sub-types: (i) normoglycaemia, (ii) gestational diabetes mellitus (GDM), (iii) diabetes in pregnancy (DIP), (iv) early-diagnosed (i.e., <22 weeks of gestation) GDM (eGDM), and (v) early-diagnosed DIP (eDIP). The prevalence of normoglycaemia, eGDM, eDIP, GDM, and DIP was 76.4%, 10.8%, 0.6%, 11.7%, and 0.6%, respectively. With regard to pregnancy outcomes, gestational weight gain (11.5 ± 5.5, 9.0 ± 5.4, 8.3 ± 4.7, 10.4 ± 5.3, and 10.1 ± 5.0 kg, p < 0.0001) and insulin requirement (none, 46.0%, 88.5%, 25.5%, and 51.7%; p < 0.001) differed according to the glycaemic sub-types. eGDM and eDIP were associated with higher rates of infant malformation. After adjustment for confounders, with normoglycaemia as the reference, only GDM was associated with large-for-gestational-age infant (odds ratio 1.34 (95% interval confidence 1.01–1.78) and only DIP was associated with hypertensive disorders (OR 3.48 (1.26–9.57)). To conclude, early-diagnosed hyperglycaemia was associated with an increased risk of malformation, suggesting that it was sometimes present at conception. Women with GDM, but not those with eGDM, had an increased risk of having a large-for-gestational-age infant, possibly because those with eGDM were treated early and therefore had less gestational weight gain. Women with DIP might benefit from specific surveillance for hypertensive disorders.