RESUMEN
INTRODUCTION: The aims of this study were to characterize insulin-treated individuals aged ≥75 years with type 2 diabetes using basal insulin analogs (BIA) or regular insulins (human insulin (HI)/neutral protamine Hagedorn (NPH)) and to compare the benefits and risks. RESEARCH DESIGN AND METHODS: The analysis was based on data from the DPV (Diabetes-Patienten-Verlaufsdokumentation) and DIVE (DIabetes Versorgungs-Evaluation) registries. To balance for confounders, propensity score matching for age, sex, diabetes duration, body mass index and hemoglobin A1c (HbA1c) as covariates was performed. RESULTS: Among 167 300 patients aged ≥75 years with type 2 diabetes (mean age, 80.3 years), 9601 subjects used insulin regimens with basal insulin (HI/NPH or BIA). Of these 8022 propensity score-matched subjects were identified. The mean diabetes duration was ~12 years and half of the patients were male. At the time of switch, patients provided with BIA experienced more dyslipidemia (89.3% vs 85.9%; p=0.002) and took a greater number of medications (4.3 vs 3.7; p<0.001) and depression was more prevalent (8.4% vs 6.5%; p=0.01). Aggregated to the most actual treatment year, BIA was associated with a higher percentage of patients using basal-supported oral therapy (42.6% vs 14.4%) and intensified conventional insulin therapy (44.3% vs 29.4%) and lower total daily insulin doses (0.24 IU/kg/day vs 0.30 IU/kg/day; p<0.001). The study did not reveal significant differences in efficacy (HbA1c 7.4% vs 7.3%; p=0.06), hospitalizations (0.7 vs 0.8 per patient-year (PY); p=0.15), length of stay (16.3 vs 16.1 days per PY; p=0.53), or rates of severe hypoglycemia (4.07 vs 4.40 per 100 PY; p=0.88), hypoglycemia with coma (3.64 vs 3.26 per 100 PY; p=0.88) and diabetic ketoacidosis (0.01 vs 0.03 per 100 PY; p=0.36). CONCLUSION: BIA were used in more individually and patient-centered therapy regimens compared with HI/NPH in patients with a mean age of 80 years. Both groups were slightly overtreated with mean HbA1c <7.5%. The risk of severe hypoglycemia was low and independent of insulin type. Further analyses of elderly patients with type 2 diabetes are needed to provide evidence for best practice approaches in this age group.