Genotype and phenotype concordance for pharmacogenetic tests through proficiency survey testing

Abstract

Context: Pharmacogenetic testing is becoming increasingly important for drugs such as irinotecan, warfarin, clopidogrel, codeine, and the tricyclic antidepressant drugs. Laboratories certified to perform pharmacogenetic testing must demonstrate successful performance on proficiency testing surveys.

Objectives: To examine the performance of laboratories subscribing to pharmacogenetic proficiency testing surveys for genes that encode CYP2C9, VKORC1, UGT1A1, CYP2C19, and CYP2D6.

Data sources: College of American Pathologists Pharmacogenetic Proficiency Survey (PGX), 2007-2011.

Conclusions: For all 5 genes challenged, there was good performance among participating laboratories for reporting wild types (95.4%-99.1% correct). For CYP2C9, VKORC1, and UGT1A1, there was no statistical difference in the percentage correctly detected for variant alleles compared to wild type. For CYP2C19, participating laboratories were greater than 90% successful in detecting variant genotypes of *1/*2, *1/*8, and *2/*3. However, several laboratories failed to detect *2/*2, *2/*4 and *1/*17 variant genotypes. For CYP2D6, laboratories were successful in detecting the most important variant genotypes (*1/*4, *1/*2, *4/*10, *10/*10, *2/*4, *2/*10) but not some of the less frequently encountered variant genotypes (*1/*10, *2/*3, *1/*35, *4/*5, and *1/*1XN). The interpretation of phenotypes by participating laboratories was largely consistent with the genotypes reported, with errors in genotyping leading to errors in phenotype assignment. Improvements in genotyping accuracy are needed for some pharmacogenetics laboratories with reference to CYP2C19 and CYP2D6.