BACKGROUND Patients with high anxiety sensitivity (AS) become extremely anxious with heart rate increases, palpitations, and symptoms of psychological arousal. AS predicts panic attacks. In atrial fibrillation (AF), AS correlates with symptom preoccupation and reduced quality of life. We assessed whether AS is associated with outcomes of rhythm-control versus rate-control in congestive heart failure (CHF) patients with AF. METHODS AND RESULTS Before random assignment, 933 participants (172 women) in the Atrial Fibrillation and Congestive Heart Failure Trial completed the Anxiety Sensitivity Inventory (ASI). Cox proportional hazards models showed no main effects of treatment (P=0.61) or AS (P=0.72) for time to cardiovascular death, but these factors interacted significantly (P=0.020). High AS patients (upper quartile, ASI ≥33) randomly assigned to rhythm-control had significantly lower cardiovascular mortality than those receiving rate-control (hazard ratio, 0.54; 95% confidence interval, 0.32-0.93; P=0.022). With lower ASI scores (<33), treatments did not differ (hazard ratio, 1.12; 95% confidence interval, 0.83-1.51; P=0.46). The interaction between treatment and dichotomized ASI scores remained significant (P=0.009) after adjustment for covariates including age, sex, hypertension, diabetes, creatinine, ejection fraction, time since first diagnosis of AF, New York Heart Association functional class, depression symptoms, marital status, and baseline β-blockers, angiotensin-converting enzyme inhibitors, oral anticoagulants, and implantable cardioverter-defibrillators. CONCLUSIONS Atrial fibrillation and congestive heart failure patients with high AS had better long-term prognosis with rhythm- than rate-control. If replicated, AS should be considered in treatment selection. Research is also needed concerning mechanisms and possible joint AS-AF treatments. CLINICAL TRIAL REGISTRATION URL: http://www.clinicaltrials.gov. Unique identifier: NCT88597077.