I n this edition of Circulation: Arrhythmia and Electrophysiology, Elayi and colleagues 1 examined mortality in patients from the Atrial Fibrillation Follow-Up Investigation of Rhythm Management (AFFIRM) study who received external (transthoracic) cardioversion shocks for atrial fibrillation (AF) compared with those who had not received cardioversion shocks. The AFFIRM trial studied 4060 patients with a history of AF and compared a strategy of rhythm control versus rate control. 2 The strategy of using antiarrhythmic drug therapy to maintain sinus rhythm did not improve all-cause mortality when compared with rate control. In this substudy analysis of AFFIRM, the association of cardioversion shocks and subsequent risk of death was determined using Cox proportional hazards modeling. The primary finding was that patients who received cardioversion shocks were not at higher risk of death compared with the patients who did not receive cardioversion shocks. Other than an association of cardioversion with cardiac hospitalizations, these neutral findings were independent of ejection fraction, numbers of cardioversion episodes, or history of permanent versus paroxysmal AF.
Article see p 465It is important to emphasize that a Cox proportional hazards model examines association of variables and does not establish "cause and effect." 3 The finding in the study by Elayi et al that there was an association of having had a cardioversion episode and having been hospitalized for cardiac reasons illustrates this principal. For instance, the hospitalization could have been for the purpose of medical management of the patient's AF, during which AF was treated with cardioversion, or for exacerbation of a chronic cardiac condition that triggered AF and then required cardioversion. The data do not elucidate the specific reasons for the observed association.The findings from this substudy of AFFIRM are relevant to the much broader debate about the association of implantable cardioverter-defibrillator (ICD) shocks with subsequent mortality. Post hoc analyses of data from the Multicenter Automatic Defibrillator Implantation Trial-II (MADIT II) and the Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT) have raised concerns about the safety of shock therapy. 4 -7 In both of these large, prospective, randomized trials, the mortality risk as assessed by the Cox proportional hazards method was higher in patients who had ICD shock therapies than in those who did not. Over the study follow-up time period, the mortality risk for patients having shocks for ventricular fibrillation (VF) or ventricular tachycardia (VT) was 3-to 6-fold that of ICD patients not shocked for VT or VF. In addition, there was a 2-fold increased mortality risk associated with inappropriate ICD shocks. The majority of inappropriate shocks were due to either AF or another supraventricular arrhythmia detected in the high-rate ICD therapy zones.The association of ICD shocks with an increase in subsequent mortality has led to a resurrection of animal research from the 198...