BackgroundThe longitudinal changes in maternal serum ferritin (SF) levels across gestation, which indirectly reflect iron supplementation, have not been extensively investigated in relation to gestational diabetes mellitus (GDM).MethodsWe conducted a retrospective cohort study at a tertiary maternal hospital in Shanghai. Women with SF concentration measurements at 8.0–13.6 weeks' gestation (GW), 29.0–31.6 GW, and an oral glucose tolerance test (OGTT) at 24–28 GW were included. We utilized logistic regression analysis to assess GDM association with maternal SF levels and longitudinal changes.ResultsThe study included 17 560 women, with 2160 (12.3%) participants diagnosed with GDM. Adjusted odds ratios (ORs) (95% confidence intervals [CIs]) for GDM across increasing quartiles of SF concentrations at 8.0–13.6 GW were 1.00 (reference), 1.139 (95% CI: 1.012–1.283), 1.093 (95% CI: 0.969–1.233), and 1.248 (95% CI: 1.111–1.403). Similarly, at 29.0–31.6 GW, increasing quartiles of SF concentrations were associated with higher adjusted ORs for GDM: 1.00 (reference), 1.165 (95% CI: 1.029–1.320), 1.335 (95% CI: 1.184–1.505), and 1.428 (95% CI: 1.268–1.607). Pregnant women with higher SF levels (upper 25th percentile) at 8.0–13.6 GW had a reduced GDM risk if their SF levels decreased to the lower 25th percentile at 29.0–31.6 GW. Conversely, the subgroup with higher SF levels (upper 25th percentile) at both time points had the highest incidence rate of GDM (15.3%, 1.235 [95% CI: 1.087–1.404]).ConclusionsMaternal SF levels independently and positively associated with GDM risk during early and late gestational stages. Considering the increased GDM risk, routine iron supplementation for iron‐replete women is questionable.image