Quantitative Structure-Activity Relationship Models to Predict Cardiac Adverse Effects.

Area of Science:

• Pharmacology
• Toxicology
• Computational Chemistry

Background:

• Drug-induced cardiotoxicity is a significant factor in drug candidate attrition during development.
• Early and accurate cardiac toxicity assessment is crucial for regulatory approval.

Purpose of the Study:

• To develop predictive models for various drug-induced cardiac adverse events.
• To enhance the safety assessment of drug candidates in preclinical and clinical stages.

Main Methods:

• Extracted postmarket adverse event data from the FDA Adverse Event Reporting System for 2002 drugs.
• Utilized 243 cardiac toxicity-related preferred terms, grouped into 12 clinically relevant categories.
• Employed two commercial Quantitative Structure-Activity Relationship (QSAR) platforms to build predictive models.

Main Results:

• Developed 12 models covering diverse cardiac endpoints like cardiac arrest, heart failure, and arrhythmias.
• Achieved up to 80% sensitivity and 80% negative predictivity in cross-validation.
• Integrated multiple data sources including drug labeling, literature, and clinical data.

Conclusions:

• The developed QSAR models offer fast, reliable, and comprehensive predictions of cardiotoxicity.
• These models can significantly aid in drug discovery and regulatory safety assessment processes.
• Facilitates identification of potential cardiotoxic compounds early in development.