This from the great NYU Clinical Correlations website.  Enjoy

Hello again Primecutters and welcome to another edition of your favorite weekly blog. Our fearless editor has allowed me a repeat performance of last week and I am honored to provide the followers of this column with another scintillating tale of the week’s medical knowledge.

Ok, here's evidence of a benefit, at least, of genetic testing for warfarin.

The study was conducted by Medco, and presented at the ACC cardiology meeting this week.

Podcast file: 

You are missing some Flash content that should appear here! Perhaps your browser cannot display it, or maybe it did not initialize correctly.

2-11-10.mp3

FDA seems to have acted to encourage use of genetic testing in warfarin dosing with a lack of evidence of any clinical benefit.

Clinical outcomes is what really drives mainstream medicine, and we as clinicians and patients should demand that decisionmakers in Washington base their guidelines on science, not on surrogate markers.

The changes in 2007 suggesting pharmacogenomic testing didn't go far enough, evidently, so now, with pretty minimal evidence, FDA has gone ahead and upped the ante with specific recommendations for starting doses depending on the VKORC polymorphism profile. They cite "multiple studies" to justify a dosing table (!), which is news to me.

In the future we will know how to dose drugs on an individual basis. Right now we use population-based information to make dosing decisions and this can't be right knowing what we know about differences between individual metabolism. For example we already individualize dosing for warfarin a blood thinner. Patients have a check of their INR at regular intervals to see if their warfarin is at the right dose. The dose is modified based on the INR results.

Unexpected development: a small study reports good efficacy of a cage-like device implanted in the hearts of patients with atrial fibrillation. Right now we give people a medicine coumadin to prevent strokes from atrial fibrillation (a common heart arrhythmia). A company has introduced a device that evidently accomplishes the same goal. Complications were up around 7% which is more toxic than coumadin which has about a 1% per year serious bleed risk (about the same as aspirin). The Reuter's article calls coumadin "notoriously difficult to manage " as if it were some sort of villain.

Yesterday the FDA posted a note expressing concern about the bleeding risk of rivaroxaban a new type of blood thinner expected to eventually supplant coumadin. Millions of Americans take coumadin for stroke prevention and prevention and treatment of blood clots. http://money.cnn.com/news/newsfeeds/articles/djf500/200903171127DOWJONES... The problem with coumadin is that it has to be monitored at least monthly. It has a narrow "therapeutic index " which means it can cause bleeding if the level gets too high and repeat blood clots if the level gets too low.

Did the world really need another clinical trial looking at vitamin K reversal of coumadin? Well it's got one in no less a journal than Annals of Internal Medicine. I thought this question must have been answered by now by a handful of studies: if your patient has an excessive INR from coumadin you can quickly correct the number with vitamin K. The problem as the new article reports is that bleeding is not necessarily improved with correction of the INR. Wow.

Today I got a sign-it sheet from Medco asking me to authorize pharmacogenetic testing for a patient on warfarin. Now I know the Medco CEO wants to eat doctors and wants to tell us how to practice but this is crossing the line I think. Does Mr. Snow realize that pharmacogenetic testing for warfarin though trendy at the moment is not cost effective? How about $200 000 per QALY! Usually we have to be less than $50 000 per year to be cost effective like mammography Pap smears etc. It wasn't even that hard to find an analysis--try Google!

"The use of a pharmacogenetic algorithm for estimating the appropriate initial dose of warfarin produces recommendations that are significantly closer to the required stable therapeutic dose than those derived from a clinical algorithm or a fixed-dose approach." Yes but does it lead to less bleeding or less thrombosis? Isn't that what we care about? http://content.nejm.org/cgi/content/short/360/8/753?rss=1&query=current