Genotype, phenotype, and medication recommendation agreement among commercial pharmacogenetic-based decision support tools.

Abstract

The degree of agreement between four commercial pharmacogenetic-based decision support tools (DSTs) was examined in five outpatients with major depressive disorder and at least two previous antidepressant failures. Comparisons were made across seven pharmacokinetic (CYP1A2, CYP2B6, CYP2C19, CYP2C9, CYP2D6, CYP3A4, and UGT2B15) and seven pharmacodynamic (BDNF, COMT, HLA-A, HTR2A, HTR2C, OPRM1, and SLC6A4) genes that were included on ≥2 of the four DST testing panels. Among these overlapping genes, genotype (33-100%) and predicted phenotype (20-100%) agreement varied substantially. Medication recommendation agreement was the greatest for mood stabilizers (84%), followed by antidepressants (56%), anxiolytics/hypnotics (56%), and antipsychotics (55%). Approximately one-quarter (26%) of all medication recommendations were jointly flagged by two or more DSTs as "actionable" but 19% of these recommendations provided conflicting advice (e.g., dosing) for the same medication.The level of disagreement in medication recommendations across the pharmacogenetic DSTs indicates that these tests cannot be assumed to be equivalent or interchangeable. Additional efforts to standardize genetic-based phenotyping and to develop medication guidelines are warranted.