Background: The benefits of intensive glycemic control in patients with type 2 diabetes are not well quantified. It is therefore not clear which patients will benefit most from aggressive glycemic control.
Objective: To evaluate the efficacy of glycemic control in type 2 diabetes.
Design: Markov decision model.
Patients: Diabetic patients at a health maintenance organization.
Main outcome measures: Risks for developing blindness and end-stage renal disease; number of patients and patient-years needed to treat to prevent complications.
Results: For a patient in whom diabetes developed before 50 years of age, reducing hemoglobin A1c levels from 9% to 7% results in an estimated 2.3-percentage point decrease (from 2.6% to 0.3%) in lifetime risk for blindness due to retinopathy. The same change in a patient with diabetes onset at 65 years of age would be expected to decrease the risk for blindness by 0.5 percentage points (from 0.5% to < 0.1%). However, the Markov model predicts substantially greater benefit when moving from poor to moderate glycemic control than when moving from moderate to almost-normal glycemic control. Targeting less than 20% of the patients at one health maintenance organization for intensified therapy may prevent more than 80% of the preventable patient-time spent blind. The risks for end-stage renal disease and the risk reduction provided by improved glycemic control are lower than those for blindness.
Conclusions: This probability model, based on extrapolation from the experience with type 1 diabetes, suggests that patients with early onset of type 2 diabetes will accrue substantial benefit from almost-normal glycemic control. In patients with later onset, moderate glycemic control prevents most end-stage complications caused by microvascular disease. These results have important implications for informing patients and allocating health care resources.