Performance of 2019 ESC risk classification and the Steno type 1 risk engine in predicting cardiovascular events in adults with type 1 diabetes: A retrospective study

Abstract

Aims: The study compares the performance of the European Society of Cardiology (ESC) risk criteria and the Steno Type 1 Risk Engine (ST1RE) in the prediction of cardiovascular (CV) events.

Methods: 456 adults with type 1 diabetes (T1D) were retrospectively studied. During 8.5 ± 5.5 years of observation, twenty-four patients (5.2%) experienced a CV event. The predictive performance of the two risk models was evaluated by classical metrics and the event-free survival analysis.

Results: The ESC criteria show excellent sensitivity (91.7%) and suboptimal specificity (64.4 %) in predicting CV events in the very high CV risk group, but a poor performance in the high/moderate risk groups. The ST1RE algorithm shows a good predictive performance in all CV risk categories. Using ESC classification, the event-free survival analysis shows a significantly higher event rate in the very high CV risk group compared to the high/moderate risk group (p < 0.0019). Using the ST1RE algorithm, a significant difference in the event-free survival curve was found between the three CV risk categories (p < 0.0001).

Conclusions: In T1D the ESC classification has a good performance in predicting CV events only in those at very high CV risk, whereas the ST1RE algorithm has a good performance in all risk categories.

Keywords: 2019 ESC guidelines; Cardiovascular disease; STENO Type 1 risk engine; Type 1 diabetes.