What Are We Supposed to Do With Niacin and Warfarin?
After much hullabaloo, results from the AIM-HIGH (N. Engl. J. Med. 2011; 365:2255-67) and WARCEF (N. Engl. J. Med. 2012;366:1859-69) studies have left us with exactly the uncertainty regarding niacin and warfarin that they were supposed to have resolved. A stark reminder of this occurred on the same clinic day earlier this week when I saw two patients who would have been prime candidates for participation in these trials. One has had a low HDL level but nearly normal LDL levels for years; should I have added niacin? The second has a low left-ventricular ejection fraction (0.15) but no history of atrial fibrillation or LV thrombus. Warfarin or aspirin or neither?
For the first encounter, I left it largely up to the patient after I described the dilemma. We could raise his HDL, but might not impact his clinical course. He opted to stay away from more medication. For the second patient, I selected warfarin, largely remembering the lessons from a SAVE subanalysis that suggested an increase of 18% in stroke risk for every 5 percent decrease in LVEF (N. Engl. J. Med. 1997;336:251-7). It’s also hard to completely trust WARCEF; the trial took a remarkably long time (8 years) to successfully enroll patients, perhaps because of investigator bias in favor of warfarin. In addition, we need more information about the relationship between treatment arm and ejection fraction.
Good lessons here: we honor clinical trials with "Late Breaker" sessions and prominent publications, but the answers we seek are often not forthcoming. Indeed, the niacin and warfarin stories are just two examples; I could have cited many others. Therefore, I suggest we embrace the "gray" and recognize that large, well-designed, multicenter, multiyear, randomized clinical trials leave huge gaps of uncertainty – even when they are positive.
Just remember this uncertainty the next time you see an 85-year-old with almost any form of acute or chronic heart disease. The inclusion and exclusion criteria of the "definitive" trial for an intervention in that disease would have almost certainly excluded your patient!