AI Identifies Sex-Specific Adverse Events With High Precision

Researchers have developed an algorithm that is able to accurately predict sex differences in drug response through the implementation of pharmacogenomic data.

The study, published in Cell, presented AwareDX, a resource that uses machine learning to understand sex-specific adverse drug effects (AEs) with the potential for use in drug discovery, repositioning, and pharmacogenetic studies, as well as for further analyses of electronic health records (EHRs) and clinical trials, according to investigators.

Women are twice as likely as men to develop adverse drug reactions (ADRs) to a drug, partly due to differences in pharmacokinetics and pharmacodynamics that induce increased drug bioavailability and sensitivity to medication, according to the study investigators. However, many population-specific ADRs are underexplored, and “women [remain] severely underrepresented in clinical trials,” investigators wrote.

Related: New Study Shows Sex Biases in Drug Dosage Trials Lead to Overmedicated Women​

The investigators also pointed out that although the FDA’s Adverse Event Reporting System (FAERS) offers the potential to “systematically quantify sex-specific risks of drugs,” it also is subject to considerable biases due to differential prescription and the sex-specific nature of some diseases and drugs.

AwareDX (Analyzing Women At Risk for Experiencing Drug toXicity) “mitigates sex biases in the data by using a machine learning adaptation of propensity score matching,” study authors wrote. A random forest (RF) model predicts the likelihood of being female given confounding factors, and matches drug-exposed females to drug-exposed males, effectively mitigating 79% of underlying sex biases, according to the study.

The approach successfully flagged significant sex-specific AEs, including many that were not identified prior to the use of AwareDX.

Investigators additionally confirmed the accuracy of the AwareDX algorithm by incorporating not only clinical literature, but also metabolic and genetic bases of sex differences in drug response.

The resource includes 20,817 ADEs that it predicted as posing sex-specific risks, in which 62.7% of significant hits posed increased risks to women.

“We believe that AwareDX could vastly advance the incorporation of sex in considerations of drug safety and efficacy. Knowledge of sex differences during drug prescription has the potential to significantly reduce adverse events, making AwareDX a valuable tool for the advancement of precision medicine,” study authors wrote.

Reference:

1. Chandak P, Tatonetti NP. Using machine learning to identify adverse drug effects posing increased risk to women. Cell. September 22, 2020. doi: https://doi.org/10.1016/j.patter.2020.100108.