Understanding 30-Day Mortality After First STEMI Through DAGs: Unravelling Epidemiological Cause-Effect Links

Area of Science:

• Cardiology
• Epidemiology
• Data Science

Background:

• Traditional statistical tests struggle with causal inference.
• Directed acyclic graphs (DAGs) provide a robust framework for analyzing causality.
• Understanding epidemiological factors impacting ST-elevation myocardial infarction (STEMI) mortality is crucial.

Purpose of the Study:

• To employ DAGs for exploring causal relationships between epidemiological factors and 30-day mortality in first-time STEMI patients.
• To identify key determinants of mortality following STEMI.

Main Methods:

• Utilized data from the North India (NORIN)-STEMI study registry (3,192 patients).
• Applied the Non-combinatorial Optimization via Trace Exponential and Augmented Lagrangian for Structure Learning (NOTEARS) method for DAG structure learning.
• Implemented a permutation testing framework for statistical validation of DAG links.

Main Results:

• Analyzed data from 2,946 first-time STEMI patients; 7.7% experienced 30-day mortality.
• Higher mean age (59.90±13.89 years) was observed in the mortality group (p<0.0001).
• Significant associations found between mortality and HbA1c, triglycerides, smoking, sex, education, occupation, socioeconomic status, physical activity, stress, and hypertension.

Conclusions:

• The developed DAG identified causal pathways and confounding variables influencing STEMI mortality.
• Sex emerged as a significant factor, impacting mortality directly and indirectly via age, alcohol consumption, stress, hypertension, and socioeconomic status.
• Sex acts as a confounding factor, with its influence on mortality being modified by other variables.