originally posted at http://blogs.warwick.ac.uk/simongates/entry/the_captivate_trial/ on 24 january 2014
…published a couple of years ago in American Journal of Respiratory and Critical Care Medicine (Wunderink et al Am J Resp Crit Care Med 2011; 183(11): 1561-1568). It’s a trial of recombinant tissue factor pathway inhibitor (tifacogin) for patients with severe community acquired pneumonia, and randomised people to tifacogin 0.025mg/kg/h, 0.075 mg/kg/h, or placebo. The rationale for it was that tifacogin seemed to be beneficial in the subgroup with severe community acquired pneumonia in a previous trial of patients with sepsis (which rings alam bells with me, but that’s another issue). The trial was international, involving 188 centres, and randomised 238 patients, so a major undertaking.
The interesting point about it was that they performed an interim analysis, as a result of which they stopped randomisation to the higher dose of drug due to lack of efficacy (futility) but continued to randomise to the lower dose. This seems extraordinary; if the high dose isn’t doing anything, it seems pretty unlikely that the low dose would. I could understand it if the high dose was stopped because of toxicity or increase in adverse outcomes, like death, but that doesn’t seem to have been the case.
Unsurprisingly, the final trial results showed no difference in mortality between tifacogin (18%) and placebo (17.9%). Has there ever been a case where a promising-looking subgroup result was shown in a subsequent trial to be correct?