DICTrank: The largest reference list of 1318 human drugs ranked by risk of drug-induced cardiotoxicity using FDA labeling

Abstract

AbstractDrug-induced cardiotoxicity (DICT) is one of the leading causes of drug attrition in clinical trials or withdrawal from the market. Many studies have been conducted to detect DICT in the early stage of drug development and clinical diagnosis, but the success is limited, as evident by the high attrition rate at all clinical phases due to DICT. Most of these efforts, if not all, have focused on specific adverse events and/or mechanisms associated with DICT, such as QT prolongation and hERG-related cardiotoxicity. However, given the broad spectrum of cardiotoxicity, it is necessary to develop a reference drug list with a systematic annotation of DICT potential across a large number of drugs and drug classes. Such a list is essential for developing effective DICT diagnostic biomarkers and early prediction strategies using new approach methods (NAMs), including artificial intelligence (AI). By utilizing labeling documents for FDA (U.S. Food and Drug Administration)-approved drugs, we developed a drug classification scheme to prioritize drugs based on their DICT potential. This resulted in DICTrank, which is the largest dataset of drugs annotated with ranked DICT risk in humans. DICTrank categorizes drugs into four categories of DICT concerns by integrating DICT severity and labeling content. The dataset consists of 1318 drugs, classified as follows: Most-DICT-Concern (341), Less-DICT-Concern (528), No-DICT-Concern (343), and Ambiguous-DICT-Concern (106; lacking sufficient information in the labeling document to determine cardiotoxicity potential). DICTrank covers a wide range of drug therapeutic categories. Using this extensive DICT dataset, we discovered that several therapeutic categories were significantly enriched with drugs of Most-DICT-Concern as well as the association of daily dose with drug class. These categories include antineoplastic agents, sex hormones and modulators of the genital system, anti-inflammatory and antirheumatic products, beta-blocking agents, and cardiac therapy. DICTrank represents the largest drug list for DICT to date, and it could contribute to the development of NAMs and AI models for the early identification of DICT risk during drug development and beyond.