Abstract Study question Is the use of metformin or insulin associated with an increased risk of large or small for gestational age birth in comparison to drug-naïve GDM? Summary answer No difference in SGA / LGA odds with drug-naïve mothers versus metformin-exposed, but higher odds of SGA and lover odds of LGA compared to insulin-exposed. What is known already In Europe, the originator brand of metformin has received a label update allowing the use during pregnancy. Use of metformin during pregnancy is associated with a reduced risk of LGA, but increased risk of SGA birth compared to insulin. In CLUE, metformin only exposure demonstrated a reduced odds of LGA birth (weighted OR (wOR) 0.82, 95% CI 0.67 to 0.99), but increased risk for SGA (wOR 1.65, 95% CI 1.16 to 2.34). Insulin use during a hyperglycemic pregnancy is discussed to increase birth weight. Study design, size, duration This was a follow-up study performed on our previously published population-based register data cohort study CLUE with maternal exposure to metformin, insulin or diagnosed with GDM but drug-naïve until birth. The follow-up study included 3,964 subjects in the metformin only cohort, 5,273 in the insulin only cohort and 82,675 women diagnosed with GDM but without drug treatment during the respective pregnancy (naïve cohort). Participants/materials, setting, methods The study population included singleton children born to 18-45 year old women from Finland during 2004–2016. Exclusion criteria were maternal type 1 diabetes, exposure to glucocorticoids and antidiabetic medications except metformin and insulin during the respective pregnancy. Analyses of SGA and LGA used logistic regression to estimate ORs with 95%CIs, with naïve cohort as the reference. Inverse probability of treatment weighting (IPTW) with stabilized weights based on propensity scores (PS) was used to control for confounding. Main results and the role of chance The odds of SGA birth were significantly higher in the naïve cohort (wOR 1.41, 95% CI 1.11 to 1.79) compared to those exposed to insulin, and the odds for LGA birth were significantly lower (wOR 0.72, 95% CI 0.64 to 0.82). No difference was found in the odds for SGA or LGA when the metformin cohort was compared with naïve cohort (wOR 0.97, 95% CI 0.73 to 1.27 and wOR 0.91, 95% CI 0.75 to 1.11, respectively). Limitations, reasons for caution Drug exposure was assessed on the dispensed drug only. and not intake, also prohibiting dose and treatment duration stratification. Although IPTW methods were used to account for many characteristics, the prevalence of underlying diseases differed across treatment cohorts, the naïve cohort which was defined by indication and absence of treatment. Wider implications of the findings This analysis adds to the growing evidence that insulin exposure in pregnancy may increase the risk of LGA, which is associated with increased adult obesity. Metformin did not increase the odds of SGA or LGA. Metformin may be considered a reasonable treatment during pregnancy, e.g. at risk for LGA birth. Trial registration number not applicable