Durable change in glycaemic control following intensive management of type 2 diabetes in the ACCORD clinical trial

Aims/hypothesis We aimed to determine the persistence of glycaemic control 1 year after a limited period of intensive glycaemic management of type 2 diabetes. Methods 4119 ACCORD Trial participants randomised to target HbA(1c) <6.0% (42 mmol/mol) for 4.0 +/- 1.2 years were systematically transitioned to target HbA(1c) 7.0-7.9% (53-63 mmol/mol) and followed for an additional 1.1 +/- 0.2 years. Characteristics of participants with HbA(1c) <6.5% (48 mmol/mol) or >= 6.5% at transition were compared. Changes in BMI and glucose-lowering medications were compared between those ending with HbA(1c) <6.5% vs >= 6.5%. Poisson models were used to assess the independent effect of attaining HbA(1c) <6.5% before transition on ending with HbA(1c) <6.5%. Results Participants with pre-transition HbA(1c) <6.5% were older with shorter duration diabetes and took less insulin but more non-insulin glucose-lowering agents than those with higher HbA(1c). A total of 823 participants achieved a final HbA(1c) <6.5%, and had greater post-transition reductions in BMI, insulin dose and secretagogue and acarbose use than those with higher HbA(1c) (p<0.0001). HbA(1c) <6.5% at transition predicted final HbA(1c) <6.5% (crude RR 4.9 [95% CI 4.0, 5.9]; RR 3.9 [95% CI 3.2, 4.8] adjusted for demographics, co-interventions, pre-intervention HbA(1c), BMI and glucose-lowering medication, and post-transition change in both BMI and glucose-lowering medication). Progressively lower pre-transition HbA(1c) levels were associated with a greater likelihood of maintaining a final HbA(1c) of <6.5%. Follow-up duration was not associated with post-transition rise in HbA(1c). Conclusions/interpretation Time-limited intensive glycaemic management using a combination of agents that achieves HbA(1c) levels below 6.5% in established diabetes is associated with glycaemic control more than 1 year after therapy is relaxed.