The folks at the Temple U SOP are doing some interesting stuff in one of their pharmacy labs with a focus on Coumadin:

“Prescribing this medicine is like trial and error in finding the right dosage that works best for you,” says Krynetskiy. “Five milligrams is a typical dose, but a little less or a little more could have dramatic consequences or no benefit at all.”

Doctors call this optimal dosage the therapeutic window, and Krynetskiy is trying to find it through pharmacogenomics, the study of a person’s response to drugs based on their genetic makeup. It’s a collaboration that crosses campuses and includes Krynetskiy and fellow clinical faculty at the School of Pharmacy, clinicians at Temple University Hospital and Jeannes Hospital. The researchers are studying why people process the same drug differently. In this case, they’re trying to find the correlation between genotypes, or a person’s inner code of DNA, and the correct dosage of Warfarin. By collecting saliva samples and extracting DNA from 77 participants already on the drug, the researchers can look for variances, genetic clues, which make people metabolize the same drug in very different ways.

Sounds more like a fun lab experiment than something that’ll be clinically valuable for something as cheap as warfarin. This might be more interesting in terms of cost-benefit by choosing a drug that’s both expensive and has a narrow therapeutic index. Aminoglycosides, some cancer drugs, and then there’s always the iatrogenic narrowing of therapeutic windows — especially via the P450 isoenzyme — that might benefit from this kind of relatively blunt pharmacogenomic hashing. At the very least, some interesting and possibly useful trends might be established.

Warfarin, as cheap as it is, probably isn’t a bad place to start. At the very least, I bet it makes for an awesome lab — we never did anything nearly as cool when I was in school…

Update from Eric:

It’s not the cost of the drug – it’s the cost of the 29% of Warfarin users that are hospitalized in the first year due to a drug-related adverse event.

If this is indeed the case, then preventing just one hospitalization could pay for dozens, and possibly hundreds of these tests, not to mention the impact on human and opportunity costs associated with hospitalization and ADEs.