Roquilly et ses amis francais randomized 150 critically ill victims of severe French trauma to hydrocortisone or placebo for 7 days. The treated group had an absolute 16-19% lower risk for hospital-acquired pneumonia, the primary endpoint (36% vs. 51-54% depending on which intention-to-treat analysis you prefer, p=0.007 to 0.01). (n=150). JAMA 2011;305:1201-1209.