OBJECTIVES: Type 2 diabetes mellitus (T2DM) is a known risk factor for atrial fibrillation (AF). The association between glycemic control and developing AF is not clear. This study identifies the association between glycemic control level during the previous 12 months and AF diagnosis in patients with T2DM. METHODS: This case control study used incidence density sampling to identify patients of age ≥18 years with T2DM in the US Veteran Affairs (VA) national EMR database from 2000-2014. Patients were required to have ≥2 hemoglobin A1c (HbA1c) measurements in the year prior to index date. Index date of cases was first AF diagnosis date. Patients with all HbA1c <7% were considered to have good glycemic control; those with 1 or more HbA1c ≥7% were poor control. Cases had an AF or atrial flutter diagnosis ≥12 months after T2DM diagnosis, identified by ICD 9 codes between February 1, 2002 and December 1, 2014. Controls were selected from the same population but without AF diagnosis on or before index date. Cases and controls were matched 1:4 based on diabetes duration and T2DM diagnosis calendar year. Logistic regression was used to assess the association between glycemic control and AF controlling for age and sex. RESULTS: Of 1,752,409 patients with T2DM, 38,094 cases were matched to 152,342 controls. The mean age (SD) of cases and controls was 71.2 (9.7) and 66.2 (10.3) years, respectively, (p<0.001) and 98.2% (cases) and 96.1% (controls) were male. After adjusting for age and gender, patients with AF were 48% more likely to have poor glycemic control compared to patients without AF [OR 1.48 (95%CI 1.45-1.52)] (p-value <0.001). CONCLUSIONS: Patients with AF had poor glycemic control in previous year of AF diagnosis, compared to patients without AF. However, analyses were not controlled for all the potential confounders, which can affect estimate.