For example, we know that CRT is efficacious in selected patients with heart failure who meet criteria for dyssynchrony, but up to 30% of patients in earlier trials of CRT have been nonresponders (11 - 12). The European Society of Cardiology recently published its guideline for use of CRT in mild heart failure, outlining the evidence-based characteristics of patients most likely to respond. The guideline recommends that patients who are eligible for CRT have some heart failure symptoms—NYHA class greater than I, an ejection fraction of 0.35 or less, a QRS duration of 150 ms or greater, and sinus rhythm—despite maintenance of optimal medical therapy (13). This conservative approach is appropriate if we consider data beyond the inclusion criteria of REVERSE (REsynchronization reVErses Remodeling in Systolic left vEntricular dysfunction) (6), MADIT-CRT (Multicenter Automatic Defibrillator Implantation Trial with Cardiac Resynchronization Therapy) (8), and RAFT (9). In each of these trials, although patients with much shorter QRS intervals could have been enrolled, the mean QRS in patients who were randomly assigned was greater than 150 ms. In REVERSE, patients with an ejection fraction as high as 0.40 were included, but the mean ejection fraction was only 0.27. The patients who benefitted most demonstrably from CRT in these trials are probably not the ones near the inclusion cutoffs, an observation that has been described in the data on implantable cardioverter-defibrillator use for primary prevention of sudden cardiac death (14). Moreover, even the patients considered to have NYHA class I symptoms in REVERSE and MADIT-CRT were allowed to have had symptoms at some point before enrollment; RAFT did not include patients with NYHA class I symptoms.