In type 2 diabetes (T2D) patients, an association between poor sleep and poor glycemic control has been previously established. However, optimal sleep may be defined in a number of ways. Authors for a new study sought to identify which sleep characteristics are most significantly associated with glycated hemoglobin A1c (HbA1c). Sleep was evaluated in 172 T2D patients through seven-day wrist-actigraphy and questionnaires pertaining to sleep. Sleep measures included total sleep duration, variability in sleep duration, midsleep time, variability in midsleep time, sleep efficiency, subjective sleep quality, and subjective insomnia symptoms. The authors employed linear regression to measure the association between the different sleep measures and HbA1c. The sleep measures most significantly correlated with HbA1c were variability in sleep duration (β=0.239; P=0.002; R2=4.9%), total sleep duration (U-shaped: β=1.161/β2=1.044; P=0.017/0.032; R2=4.3%), subjective sleep quality (β=0.191; P=0.012; R2=3.6%), variability in midsleep time (β=0.184; P=0.016; R2=3.4%), and sleep efficiency (β=−0.150; R2=2.3%). HbA1c was not influenced by midsleep time and subjective insomnia symptoms. A significant correlation was observed between variability in sleep duration, total sleep duration, and subjective sleep quality in combination and HbA1c; this accounted for 10.3% of the HbA1c variance. In adjusted analyses, the outcomes were similar but associations were not as strong; total sleep duration and subjective sleep quality were the most significant factors associated with HbA1c and constituted 6.0% of the variance in HbA1c.