Beyond CYP2C19: clinical phenotype drives events and P2Y12 escalation in a real-world Andean population
M Tinoco Lara, B Pozo Hidalgo, K Lopez Patino, V Arguero Rodriguez, M Camafort Babkowski, L Cardenas AldazAbstract
Background
Clopidogrel response is highly variable due to both CYP2C19 loss-of-function (LOF) alleles and acquired mechanisms related to comorbidity and drug–drug interactions. Where platelet function testing (PFT) is not routinely available, genotyping alone may miss residual risk. We evaluated clopidogrel non-responsiveness using an integrated strategy combining CYP2C19 genotyping with clinical variables and risk tools (ABCD-GENE and PARIS).
Purpose
To determine profiles and outcomes in LOF carriers versus non-LOF metabolizers and to assess phenotype-driven escalation of P2Y12 inhibition.
Objective
To compare the clinical profiles of patients with CYP2C19 LOF-related genetic resistance and non-LOF metabolizers and to assess a clinically driven precision strategy for identifying non-responsiveness and guiding treatment decisions, including revascularization, switching to ticagrelor, and/or prolonged dual antiplatelet therapy (DAPT).
Methods
Observational cohort study of consecutive coronary artery disease patients undergoing CYP2C19 genotyping (N=117). Participants were grouped as LOF carriers (intermediate/poor metabolizers) or non-LOF (normal metabolizers). We recorded baseline demographics, risk factors, presentation, and revascularization (PCI/CABG) and followed patients for >12 months for adjudicated new atherothrombotic events and antithrombotic treatment changes (clopidogrel, ticagrelor, and/or oral anticoagulation). Haemoglobin and renal function were collected when available to derive PARIS and ABCD-GENE scores.
Results
Genetic resistance (CYP2C19 LOF) was present in 26.5% (31/117), indicating that 1 in 4 patients had a baseline genetic predisposition to reduced clopidogrel responsiveness. A “genetic blind spot” emerged: among non-LOF (normal) metabolizers, 51.8% experienced a new atherosclerotic event during follow-up. Genotype did not discriminate event occurrence (p=0.957), suggesting that residual risk was largely driven by acquired factors. The strongest predictors of new events were proton pump inhibitor (PPI) use (OR 3.81; p = 0.094), statin therapy (OR 1.87), reflecting higher baseline risk, renal dysfunction (eGFR < 60 mL/min/1.73 m²; OR 1.73), and diabetes (OR 1.45), supporting an acquired resistance phenotype. Risk scores added clinical value: ABCD-GENE may identify high-risk patients despite a “normal” genotype, while PARIS aligned with real-world decisions—among those classified as high ischemic risk (particularly patients with renal dysfunction and diabetes), 74.4% were switched to ticagrelor, suggesting phenotype-driven escalation of antiplatelet therapy.
Conclusion
CYP2C19 genotyping alone did not predict ischemic events. An integrated, phenotype-driven approach—highlighting proton pump inhibitor exposure and cardiometabolic/renal comorbidity—may better identify patients who require escalation to more potent P2Y12 inhibitors when PFT is unavailable.