Development and validation of a clinical-pharmacogenomic prediction model for clopidogrel-related high on-treatment platelet reactivity in patients with coronary artery disease.

CYP2C19

Clopidogrel bioactivation depends on the enzyme encoded by CYP2C19; loss of function leaves high platelet reactivity on treatment, compounded by co-prescription of a proton pump inhibitor.

This work builds a prediction model for high on-treatment platelet reactivity under clopidogrel, intended for use before platelet function testing results are available, to identify patients for priority testing or therapy adjustment. Three hundred and eighty patients with coronary artery disease were enrolled retrospectively and split into a training set (n = 266) and a validation set (n = 114), high reactivity being defined as ADP-induced platelet aggregation of 40% or more after about seven days of dual antiplatelet therapy. Seven independent variables emerged after LASSO selection and multivariable logistic regression: carrier status for a non-functional CYP2C19 allele (OR 4.341; p < 0.001), proton pump inhibitor use (OR 2.630; p = 0.007), acute coronary syndrome (OR 2.445; p = 0.014), age, platelet count, left ventricular ejection fraction and haemoglobin. The logistic model reached an area under the curve of 0.866 (95% CI 0.816-0.916) in training and 0.870 (95% CI 0.793-0.946) in validation, genotype being the strongest contributor. The authors state that external validation in multicentre prospective cohorts is required before clinical implementation.

Synthesis written by Geno'X. For the full original abstract, please refer to the source publication.

The value of this model is not to replace genotyping but to show that CYP2C19 loss of function and proton pump inhibitor-driven phenoconversion act additively, while practice still treats them as two separate questions. Its weakness is the endpoint: high platelet reactivity is a laboratory marker, not an ischaemic event, and predicting the former well does not guarantee predicting the latter. With strictly internal validation in 114 patients from a single centre, an area under the curve of 0.87 is an optimistic figure to be treated as provisional.

Analysis by Dr Thibaut Benquey