An international study has pointed to low-density lipoprotein (LDL) cholesterol as a possible marker to anticipate arrhythmias in people with type 2 diabetes, thanks to the advanced use of data analysis and artificial intelligence.
The finding comes from the "GATEKEEPER" project, led by a team from the University of Ioannina (Greece) in cooperation with specialists from the Polytechnic University of Madrid (UPM), who have applied data science techniques to explore this relationship.
"This research analyzes the complex relationship between type 2 diabetes mellitus (DM2) and cardiac arrhythmias, employing a data-driven approach," explained Laura López, one of the UPM researchers who participated in this work.
For the research, clinical data, questionnaires, and health measurements obtained at the patients' homes using portable devices were collected over a period of six months. With this information base, the team examined the simultaneous variations in cardiological and glycemic parameters using machine learning algorithms.
"The main objective is to identify the characteristics that have statistically significant relationships with the occurrence of arrhythmic events in patients with DM2. In this way, it seeks to enrich the understanding of the relationships between both pathologies to help generate more focused treatment strategies and personalized arrhythmia management plans for this vulnerable population," explained the researcher from the Life Supporting Technologies group at UPM.
To achieve this, the experts resorted to classification models and predictive algorithms such as "Random Forest" and "Gradient Boosting," framed within ensemble methods. This approach combines multiple models instead of relying on a single one, in order to achieve more robust and precise decisions.
After processing the data with these tools, the study determined that high levels of LDL cholesterol constitute the most determining risk factor for distinguishing the occurrence of arrhythmic episodes in patients with type 2 diabetes. In addition, the EuroQoL-5D-5L quality of life questionnaire scores and the waist-to-hip ratio were identified as statistically significant variables.
According to researchers from UPM, the relevance of this work lies in that it highlights the importance of lipid control to reduce cardiovascular complications in people with diabetes. "Clinically, these findings reinforce the need to integrate LDL monitoring into the routine care workflows of DM2.
"LDL measurements could be incorporated into patient-specific arrhythmic risk calculators or used as thresholds to initiate aggressive lipid-lowering therapies, which has enormous potential to translate into truly personalized treatment pathways," details the UPM researcher.