Among the oldest diabetes drugs still in wide use, and a clean example of a real pharmacogenomic effect that guidelines deliberately decline to act on — which is more useful to know than one they do act on.
A real metabolic effect that guidelines still say not to act on. The reduced-function CYP2C9 alleles slow the clearance of gliclazide, glimepiride and other sulfonylureas, so blood levels run higher. The Dutch Pharmacogenetics Working Group reviewed this and recommends no dose adjustment: a meta-analysis of 2,769 patients found no significant increase in sulfonylurea-induced hypoglycaemia for CYP2C9*3, and where levels do run higher the added effectiveness offsets the risk. Worth knowing precisely because the gap between "measurable in the blood" and "change what you prescribe" is where most pharmacogenomic overclaiming happens.
Dutch Pharmacogenetics Working Group (DPWG) guidance on CYP2C9 and sulfonylureas; meta-analysis of CYP2C9*3 and sulfonylurea-induced hypoglycaemia, 2,769 patients