OBJECTIVE: There is no consensus on the importance of visit-to-visit glycemic variability in diabetes. Therefore, we assessed the effects of visit-to-visit variability (VVV) in HbA1c and fasting glucose on major outcomes in the ADVANCE (Action in Diabetes and Vascular Disease: Preterax and Diamicron MR Controlled Evaluation) trial. RESEARCH DESIGN AND METHODS: ADVANCE was a factorial randomized controlled trial of intensive glucose control and blood pressure lowering in patients with type 2 diabetes. VVV in the intensive glucose treatment group was defined using the SD of five measurements of HbA1c and glucose taken 3-24 months after randomization. Outcomes were combined macro- and microvascular events and all-cause mortality occurring post 24 months. Sensitivity analyses were performed using other indices of variability and in the standard glucose treatment group. RESULTS: Among 4,399 patients in the intensive group, an increase in VVV of HbA1c was associated with an increased risk of vascular events (P = 0.01) and with mortality (P < 0.001): highest versus lowest tenth hazard ratio (95% CI) 1.64 (1.05-2.55) and 3.31 (1.57-6.98), respectively, after multivariable adjustment. A clear association was also observed between VVV of fasting glucose and increased risk of vascular events (P < 0.001; 2.70 [1.65-4.42]). HbA1c variability was positively associated with the risk of macrovascular events (P = 0.02 for trend), whereas glucose variability was associated with both macro- and microvascular events (P = 0.005 and P < 0.001 for trend, respectively). Sensitivity analyses using other indices, and patients in the standard glucose treatment group, were broadly consistent with these results.