Background: A recently halted clinical trial showed that intensive treatment of type 2 diabetes mellitus was associated with increased mortality. Given the phenotypic heterogeneity of diabetes, therapy targeted at insulin status may maximize benefits and minimize harm.
Methods: In this longitudinal cohort study, we followed 503 patients with type 2 diabetes who were free of cardiovascular disease from 1996 until data on mortality and cardiovascular outcomes were censored in 2005. Phenotype-targeted therapy was defined as use of insulin therapy in patients with a fasting plasma C peptide level of 0.2 nmol/L or less and no insulin therapy in patients with higher C peptide levels.
Results: The mean age of the cohort was 54.4 (standard deviation 13.1) years, and 56% were women. The mean duration of diabetes was 4.6 years (range 0-35.9 years). Of the 503 patients, 110 (21.9%) had a low C peptide level and 111 (22.1%) were given insulin. Based on their C peptide status, 338 patients (67.2%) received phenotype-targeted therapy (non-insulin-treated, high C peptide level [n = 310] or insulin-treated, low C peptide level [n = 28]), and 165 patients (32.8%) received non-phenotype-targeted therapy (non-insulin-treated, low C peptide level [n = 82] or insulin-treated, high C peptide level [n = 83]). Compared with the insulin-treated, low-C-peptide referent group, the insulin-treated, high-C-peptide group was at a significantly higher risk of cardiovascular events (hazard ratio [HR] 2.85, p = 0.049) and death (HR 3.43, p = 0.043); the risk was not significantly higher in the other 2 groups. These differences were no longer significant after adjusting for age, sex and diabetes duration.
Interpretation: Patients with low C peptide levels who received insulin had the best clinical outcomes. Patients with normal to high C peptide levels who received insulin had the worst clinical outcomes. The results suggest that phenotype-targeted insulin therapy may be important in treating diabetes.