Pharmacogenomics of antibacterial and antiviral therapies: Clinical actionability, evidence gaps, and future directions.

Abstract

Anti-infective drugs have profoundly transformed the history of medicine. Yet, with the presence of approximately 4.1-5 million interindividual genomic variants in human genome, patients are expected not to respond equally to the same anti-infective drug. This genetic variability, together with nongenetic factors, influences therapeutic outcomes and contributes to drug-induced adverse events in predisposed individuals. Historically, the identification of HLA-B∗57:01 as a predictor of abacavir hypersensitivity in patients with HIV represented the first successful clinical application of pharmacogenomics (PGx) in infectious diseases. Since then, the field has continued to evolve, as evidenced by the discovery of multiple clinically relevant gene-drug pairs, primarily related to immune responses, drug metabolism, and drug transport pathways. The evidence accumulated to date has established a number of mandatory (HLA-B∗57:01-abacavir) and actionable (MT-RNR1-aminoglycosides, CYP2B6-efavirenz, G6PD-nitrofurantoin, G6PD-nalidixic acid, and G6PD-dapsone) gene-drug pairs, whereas most other associations remain informative or exploratory without current guideline-based prescribing recommendations. Despite this progress, robust PGx evidence remains predominantly focused on antiretrovirals, anti-hepatitis C virus, and selected antimicrobial drug classes, such as β-lactams, aminoglycosides, sulfonamides, and antituberculosis drugs. For many other anti-infective agents, current evidence suggests that host genetic variation may have a limited impact on drug efficacy or safety, or that existing studies remain insufficiently powered or replicated to support clinical translation. The narrow ancestral diversity in PGx studies and clinical trials has also restricted the breadth of the knowledge gained and, consequently, the development of inclusive guidelines. This review summarizes the current PGx landscape of antibacterial and antiviral drugs and highlights key challenges and opportunities to improve clinical actionability. Greater inclusion of previously underrepresented populations, coupled with integrative multiomics approaches powered by artificial intelligence and machine learning, could accelerate PGx biomarker identification, validation, and integration into personalized patient care. SIGNIFICANCE STATEMENT: Rapid and effective deployment of anti-infective drugs requires incorporating knowledge of host genetic determinants of drug effectiveness or adverse events, the latter of which is a leading cause of death. The evolution of data science, driven by available genomic data, represents an unprecedented opportunity to accelerate pharmacogenomics discovery in infectious diseases. If acquired at a population level and integrated into medical records, pharmacogenomics data can guide the prescription and/or dosing of anti-infective drugs, shifting infectious disease management toward personalized care.