Long-term trends in the prescription of antidiabetic drugs: real-world evidence from the Diabetes Registry Tyrol 2012-2018.

INTRODUCTION: Prescription patterns of antidiabetic drugs in the period from 2012 to 2018 were investigated based on the Diabetes Registry Tyrol. To validate the findings, we compared the numbers with trends of different national registries conducted in a comparable period of time. RESEARCH DESIGN AND METHODS: Medication data, prescription patterns, age groups, antidiabetic therapies and quality parameters (hemoglobin A1c, body mass index, complications) of 10 875 patients with type 2 diabetes from 2012 to 2018 were retrospectively assessed and descriptively analyzed. The changes were assessed using a time series analysis with linear regression and prescription trends were plotted over time. RESULTS: Sodium/glucose cotransporter 2 inhibitors (SGLT-2i) showed a significant increase in prescription from 2012 to 2018 (p<0.001), as well as metformin (p=0.002), gliptins (p=0.013) and glucagon-like peptide-1 agonists (GLP-1a) (p=0.017). Significant reduction in sulfonylurea prescriptions (p<0.001) was observed. Metformin was the most frequently prescribed antidiabetic drug (51.3%), followed by insulin/analogs (34.6%), gliptins (28.2%), SGLT-2i (11.7%), sulfonylurea (9.1%), glitazones (3.7%), GLP-1a (2.8%) and glucosidase inhibitors (0.4%). CONCLUSIONS: In this long-term, real-world study on prescription changes in the Diabetes Registry Tyrol, we observed significant increase in SGLT-2i, metformin, gliptins and GLP-1a prescriptions. In contrast prescriptions for sulfonylureas declined significantly. Changes were consistent over the years 2012-2018. Changes in prescription patterns occurred even before the publication of international and national guidelines. Thus, physicians change their prescription practice not only based on published guidelines, but even earlier on publication of cardiovascular outcome trials.

Major risk stratification models do not predict perioperative outcome after coronary artery bypass grafting in patients with previous percutaneous intervention.

OBJECTIVE: To investigate whether common risk stratification models in cardiac surgery predict perioperative outcome of coronary artery bypass grafting (CABG) in patients with previous percutaneous coronary interventions (PCIs). METHODS: We retrospectively analyzed the perioperative mortality and morbidity of 367 patients with prior elective PCI versus 2361 patients without prior PCI, who underwent first-time isolated CABG between 2001 and 2009 at our institution. Receiver operating characteristics (ROC) were used to describe the performance and accuracy of the European System for Cardiac Operative Risk Evaluation (EuroSCORE) and the Society of Thoracic Surgeons (STS) risk model in predicting mortality and morbidity. RESULTS: Both groups were comparable concerning preoperative logistic EuroSCORE (PCI: 4.9 ± 6.57, non-PCI: 4.60 ± 5.45, p=0.51). Patients with previous elective PCI had increased perioperative mortality (PCI: 3.8% vs non-PCI: 2.1%, p=0.01) and higher rates of major adverse cardiac events (8.4% vs 4.5% respectively, p=0.003). Discriminatory power for 30-day mortality was higher in the non-PCI group (EuroSCORE area under the curve (AUC): 0.875 vs 0.552 in the PCI group). Logistic EuroSCORE predicted 30-day mortality in the non-PCI group (confidence interval (CI)=0.806-0.934, p=0.0004) but not in the PCI group (CI=0.301-0.765, p=0.8). Discriminatory power for morbidity or mortality (M&M) was lower in the PCI group (AUC: 0.980 vs 0.713 for the non-PCI group). The STS risk model had a lower discriminatory power for predicting M&M in PCI patients (AUC: 0.611 vs 0.686 for the non-PCI group, p<0.001). CONCLUSIONS: The EuroSCORE and the STS risk model were inaccurate in predicting perioperative mortality after CABG in patients with history of elective PCI. There is a need for modification of risk models to improve risk assessment for surgical candidates with prior PCI.