To the Editor.
—I read with concern the report of the McSPI investigators on AF after cardiac surgery.1 There is a clear incongruity in their conclusions about the effects of β-blockers on AF. The authors demonstrate that, given preoperatively, β-blockers reduce the incidence of AF, but, given postoperatively, β-blockers increase the incidence of AF. Not only does this finding not make sense, it is contrary to the results of multiple randomized, placebo-controlled trials that have consistently demonstrated a reduction in the incidence of postoperative AF with prophylactic administration of β-blockers after surgery.2The answer may lie in the fact that the McSPI investigators have no information on whether β-blockers were started prophylactically or as treatment for AF. Many cardiologists now use βblockers3 in preference to digoxin4 for rate control of AF. If patients who developed AF were begun on β-blockers as treatment, the authors' results would be explained as