Key points are not available for this paper at this time.
Background and Objectives: This retrospective study evaluated patients with diabetes classified according to cluster-derived diabetes phenotype groups and compared their baseline metabolic characteristics, documented complication profiles, and HbA1c course during follow-up. Materials and Methods: A total of 158 patients with type 1 or type 2 diabetes followed at Zonguldak Bülent Ecevit University Endocrinology Outpatient Clinic were included. Phenotype assignment was based on baseline domains corresponding to the Ahlqvist framework: age at diagnosis, BMI category, HbA1c, beta-cell function, insulin resistance, and autoantibody status. Complications were not used for phenotype assignment. Baseline characteristics, binary complication status, exploratory phenotype-contrast logistic regression, and longitudinal HbA1c data were evaluated. Results: The groups showed significant differences in age, age at diagnosis, HbA1c, obesity status, fasting glucose, C-peptide, fasting insulin, HOMA1-%B, HDL cholesterol, ALT, and eGFR. Hepatic steatosis/suspected NAFLD, retinopathy, nephropathy, polyneuropathy, and ketosis/ketoacidosis differed among groups when complications were analyzed as ever-positive versus negative. In exploratory phenotype-contrast logistic regression, Clusters 3–4 were associated with hepatic steatosis/suspected NAFLD, Cluster 2 with retinopathy and polyneuropathy, Cluster 3 with nephropathy, and Clusters 1–2 with ketosis/ketoacidosis; the contrast for coronary artery disease did not reach statistical significance. In a linear mixed-effects model for repeated HbA1c measurements, phenotype group was associated with HbA1c levels, whereas the time effect and group-by-time interaction were not statistically significant. Conclusions: Cluster-derived diabetes phenotype groups showed distinct baseline metabolic characteristics and different documented complication profiles. These findings should be interpreted as exploratory because of the retrospective design and incomplete follow-up data.
Aslan et al. (Sun,) studied this question.