Objective: To determine whether hyperglycaemia in the gestational diabetes (GDM) range independently predicts adverse pregnancy outcomes in Uganda. Design: Prospective observational cohort study. Setting:Five major hospitals in urban/semi-urban central Uganda. Sample:237 women with gestational diabetes, 2,641 normoglycaemic controls. Methods:Women were screened with oral glucose tolerance test (OGTT) at 24-28 weeks of gestation. Cases of GDM were identified (WHO 2013 diagnostic criteria) and received standard care. Data was collected on maternal demographics, anthropometrics, prenatal management, umbilical cord c-peptide levels, and pregnancy outcomes. Participants with diabetes in pregnancy (DIP) were excluded from the analysis. Outcomes:Primary outcomes: Birthweight large for gestational age (LGA; >90th centile) and perinatal death. Secondary outcomes: Caesarean delivery, preterm birth <37 weeks, umbilical cord c-peptide concentration >90th centile (>1.35 mcg/L), and neonatal admission. Results:Women with GDM had a median of only two glucose measurements recorded in third trimester, and only one fifth received therapeutic management (mostly metformin, one participant received insulin). GDM was not independently associated with LGA (adjusted odds ratio, aOR 1.12; 95% CI 0.81-1.56) or perinatal death (aOR 0.66; 95% CI 0.26-1.66), but increased the risk of Caesarean delivery. Mid-gestational BMI of >30kg/m2 was strongly associated with LGA, and mean arterial pressure >90 mmHg was the strongest predictor of perinatal death. Conclusions:Even without active management, GDM was not associated with large birthweight or perinatal death in this population. Interventions that target blood pressure and obesity are likely to be more beneficial in improving LGA and perinatal mortality, than management of GDM. Funding:Medical Research Council Keywords:Gestational diabetes, Africa