Potential effect of suboptimal racial and ethnic categorization on benefits of precision medicine

Pharmacogenetics studies have sought to use race and ethnicity as categorical variables to stratify patients; however, the specificity of racial categories used varies across different papers. We examined the diversity of racial and ethnic categories used to describe populations classified as ‘White/Caucasian’ according to National Institutes of Health (NIH) guidelines. Seventy total pharmacogenetics papers involving the CYP2C9, CYP2C19, and CYP2D6 pharmacogenes were examined and the race and ethnic categorization data for each paper was extracted. In total, we found 66 unique racial and ethnic categories used to classify ‘Caucasian/White’ populations. We then compared the CYP450 allele frequency data for these racial and ethnic categories as described in the papers to the global allele frequency for ‘Caucasian’ populations as compiled by the Clinical Pharmacogenetics Implementation Consortium. In this analysis, we found several racial and ethnic groups whose pharmacogene allele frequencies differed from the allele frequency for ‘Caucasians’ as accepted in the literature. Overall, we found that racial and ethnic classification varies greatly between different papers and that the broad categorization of these populations as ‘White/Caucasian’ may not accurately capture the genetic heterogeneity between these populations.