OBJECTIVE To examine the association of elevated early pregnancy hemoglobin A1c (HbA1c) levels with adverse pregnancy outcomes in women with preexisting diabetes mellitus. METHODS Retrospective cohort study of 330 women with preexisting diabetes enrolled in a Diabetes in Pregnancy Program at an academic institution between 2003 and 2011 who had an early HbA1c determination. The frequencies of composite maternal adverse pregnancy outcomes (birth at<37 weeks, preeclampsia, and medically indicated birth <39 weeks), and composite fetal adverse pregnancy outcomes [shoulder dystocia, Apgar scores<7 at 5 minutes, small for gestational age (SGA), large for gestational age (LGA), and stillbirth] were compared between HbA1c categories (<6.5, 6.5-7.4, 7.5-8.4 and ≥ 8.5%). RESULTS There was no statistically significant difference between composite adverse maternal pregnancy outcomes and composite adverse fetal pregnancy outcomes as well as other individual outcomes between different HbA1c categories. Of the vaginally delivered women in our cohort, the 37 patients with HbA1c levels of ≥ 8.5% had a significantly higher frequency of fetal shoulder dystocia than the 62 with HbA1c levels of < 8.5% (24.2 vs. 1.6%, P = 0.002). Neonates of patients with HbA1c ≥ 8.5% were more likely to have low five minute Apgar scores than neonates of patients with HbA1c < 8.5%, but this was of borderline statistical significance (7.4% vs. 0.5%, P = 0.05). CONCLUSIONS In patients with preexisting diabetes mellitus, HbA1c levels of ≥ 8.5% during early pregnancy are not useful in predicting most adverse outcomes, although there may be an increased risk for shoulder dystocia.