Will we ever learn? Probably not, because we are too confident: Heart failure drug doesn’t prevent heart failure, and sepsis drug doesn’t decrease mortality from sepsis.
12 Nov, 11 | by Dr Richard Saitz, Editor of Evidence-Based Medicine
They all sound good at first. A decade later they don’t sound so good.
Human recombinant activated protein C (HR APC; drotrecogin alfa; Xigris) addresses the pathophysiology of sepsis. It is low in sepsis and associated with death. Replacing it should work. So a randomized trial was done in 2001 (http://bit.ly/vQ0f5B) showing that it decreased mortality (from 31% to 25% at 28 days) with a small increase in serious bleeding (2 to 3.5%, p=0.06). In the US, the FDA approved it. Subsequently a retrospective study finds substantial bleeding risk. Then 2011 brings a similarly sized trial (n=1696) finds no mortality benefit and it is withdrawn from the market (http://1.usa.gov/vwQXtc).
Nesiritide is a vasodilator that addresses the pathophysiology of heart failure. It should work. So a randomized trial was done showing improvements in physiological parameters and dyspnea at 3 hours (http://bit.ly/sl1yRi). Another trial found similar improvements in symptoms (http://1.usa.gov/vjYReu). In the US, the FDA approved it in 2001. In 2011, a much larger trial (n=7141) found no effect on dyspnea or heart failure rehospitalization, and a doubling of hypotension (http://bit.ly/sW5oRV).
So even if they don’t work at least they cost money and cause harm…
What happened? In both cases these were “pivotal” trials—single studies after which governmental approval followed and changed clinical practice. In the case of APC, the early trial was stopped prematurely for benefit. We know this risks overestimation of benefits. In the case of nesiritide, an early (small) trial did not find (or look for) important clinical outcomes. The second one did.
In retrospect the drugs should not have been approved or used widely. The evidence was insufficient. Maybe these were just mistakes, and now the lessons regarding appropriate interpretation of evidence have been learned. But I don’t think so.
I am not sure we will ever learn. Why? Because of how we are wired and how we understand things. First, we want to help. We want treatments that work, particularly for bad common diseases like sepsis and heart failure. Second, we find new treatments that can fix the physiology that has gone awry. We believe they will work and we take any supporting evidence we can get. Once we have it, we are not easily dissuaded (for examples see last blog re breast and prostate cancer screening (http://bit.ly/tk4i1j), or consider the evidence that shows intensive glucose control has little if any benefit but plenty of risk (http://bit.ly/w3nqx6), and evidence that finds the risk of pre-hypertension cannot be identified yet treatment is recommended (http://bit.ly/va9w9w).
We become confident that we know what works, and when facts conflict with our expectations we ignore them (see a superb discussion re: the Surety of Fools by Daniel Kahneman who coined the term “the illusion of validity” (http://nyti.ms/t6qihi).
Lets see how the latest information on bisphosphonates, drugs that decrease fracture but are now known to also increase fractures is handled…(see EBM online http://bit.ly/rGi5cg)