Distinction of cardiometabolic profiles among people ≥75 years with type 2 diabetes: a latent profile analysis

• Main Authors: Antoine CHRISTIAENS, Michel P. HERMANS, Benoit BOLAND, Séverine HENRARD
• Format: Article
• Language: English
• Published: BMC 2019-08-01
• Series: BMC Endocrine Disorders
• Subjects: Cardiometabolic profile; Homeostasis model assessment; Older patients; Type 2 diabetes; Type 2 diabetes classification; Type 2 diabetes management
• Online Access: http://link.springer.com/article/10.1186/s12902-019-0411-2

Abstract Background Older patients with type 2 diabetes mellitus represent a heterogeneous group in terms of metabolic profile. It makes glucose-lowering-therapy (GLT) complex to manage, as it needs to be individualised according to the patient profile. This study aimed to identify and characterize subgroups existing among older patients with diabetes. Methods Retrospective observational cohort study of outpatients followed in a Belgian diabetes clinic. Included participants were all aged ≥75 years, diagnosed with type 2 diabetes, Caucasian, and had a Homeostasis Model Assessment (HOMA2). A latent profile analysis was conducted to classify patients using the age at diabetes diagnosis and HOMA2 variables, i.e. insulin sensitivity (HOMA2%-S), beta-cell-function (HOMA2%-β), and the product between both (HOMA2%-βxS; as a measure of residual beta-cell function). GLT was expressed in defined daily dose (DDD). Results In total, 147 patients were included (median age: 80 years; 37.4% women; median age at diabetes diagnostic: 62 years). The resulting model classified patients into 6 distinct cardiometabolic profiles. Patients in profiles 1 and 2 had an older age at diabetes diagnosis (median: 68 years) and a lesser decrease in HOMA2%-S, as compared to other profiles. They also presented with the highest HOMA2%-βxS values. Patients in profiles 3, 4 and 5 had a moderate decrease in HOMA2%-βxS. Patients in profile 6 had the largest decrease in HOMA2%-β and HOMA2%-βxS. This classification was associated with significant differences in terms of HbA1c values and GLT total DDD between profiles. Thus, patients in profiles 1 and 2 presented with the lowest HbA1c values (median: 6.5%) though they received the lightest GLT (median GLT DDD: 0.75). Patients in profiles 3 to 5 presented with intermediate values of HbA1c (median: 7.3% and GLT DDD (median: 1.31). Finally, patients in profile 6 had the highest HbA1c values (median: 8.4%) despite receiving the highest GLT DDD (median: 2.28). Other metabolic differences were found between profiles. Conclusions This study identified 6 groups among patients ≥75 years with type 2 diabetes by latent profile analysis, based on age at diabetes diagnosis, insulin sensitivity, absolute and residual β-cell function. Intensity and choice of GLT should be adapted on this basis in addition to other existing recommendations for treatment individualisation.