There is substantial evidence that genotype influences drug safety and effectiveness.  Over 300 drugs have pharmacogenetic information in their Food and Drug Administration-approved labeling, and guidelines are available by the Clinical Pharmacogenetics Implementation Consortium (CPIC) to guide the incorporation of pharmacogenetic information into drug prescribing.^1,2 However, there remains limited uptake of pharmacogenetic testing in clinical practice. This is in part because of the demand for evidence demonstrating improved patient outcomes with genotype-guided therapy. While randomized controlled trials are the gold standard for establishing the efficacy of an intervention, they are costly and challenging to conduct for pharmacogenetic interventions since only a subset of the population (i.e., those with a variant genotype) are likely to benefit from the intervention. Alternative approaches to evidence generation include use of observational data from patients who received pharmacogenetic testing and pragmatic clinical trials. An example of each is provided below.

Observational data

CYP2C19 genotyping to predict response to clopidogrel and guide antiplatelet therapy after a percutaneous coronary intervention (PCI) is one of the most common pharmacogenetic implementations.³ Clopidogrel is a prodrug that relies on the CYP2C19 enzyme for bioactivation. Clopidogrel is less effective in patients with a non-functional CYP2C19 allele, who cannot generate sufficient concentrations of the active clopidogrel metabolite to effectively inhibit platelet aggregation.⁴ On behalf of the NIH-funded Implementing GeNomics In pracTicE (IGNITE) Network, investigators from seven institutions where CYP2C19 testing had been integrated into clinical practice pooled data for over 1,800 patients who had undergone percutaneous coronary intervention to examine outcomes with CYP2C19-guided antiplatelet therapy.⁵ At each site, alternative therapy (e.g., prasugrel or ticagrelor) was recommended in patients with a non-functional allele in whom clopidogrel was predicted to be ineffective, but the ultimate prescribing decision was left to the provider. The investigators found a significantly lower occurrence of adverse cardiovascular events (i.e., death, myocardial infarction, and stroke) in patients with a non-functional allele treated with alternative therapy versus clopidogrel. These data were consistent with more recent clinical trial data, demonstrating the value of real-world data.⁶

Pragmatic clinical trial data

Tramadol, hydrocodone, and codeine are dependent on the CYP2D6 enzyme for formation of more potent opioid metabolites. CYP2D6 intermediate and poor metabolizers with very little to no enzyme activity have low concentrations of the more potent opioid metabolites and may not attain sufficient pain relief with these drugs.⁷ In contrast, ultra-rapid metabolizers, with increased enzyme activity, may generate toxic concentrations of the more active metabolites and are at increased risk for respiratory depression.⁷ Investigators at the University of Florida conducted a hybrid implementation-effectiveness trial of CYP2D6-guided versus usual post-operative pain management.⁸ For patients randomized to the genotype-guided arm, recommendations were to avoid tramadol, hydrocodone, and codeine in those with a high risk phenotype (i.e., poor, intermediate, and ultra-rapid metabolizers) and to consider an opioid not metabolized by CYP2D6 (e.g., hydromorphone or morphine). Given the pragmatic nature of the trial, the prescribing decision was left to the provider. The majority of patients (72%) with a high-risk phenotype in the genotype-guided arm, but none of those in the usual care arm, received an opioid other than tramadol, hydrocodone, or codeine.⁸ There was similar post-operative pain intensity in the genotype-guided and usual care arm, but lower opioid consumption in the genotype-guided arm. A large, multi-site pragmatic trial is currently on-going to further examine outcomes with CYP2D6-guided post-surgical pain management.⁹