Despite improvements in drug and device therapy for heart failure, hospitalization rates and mortality have changed little in the past decade. Randomized clinical trials using gene transfer to improve function of the failing heart are the focus of this review. Four randomized clinical trials of gene transfer in heart failure with reduced ejection fraction (HFrEF) have been published. Each enrolled patients with stable symptomatic HFrEF and used either intracoronary delivery of a virus vector or endocardial injection of a plasmid. The initial CUPID trial randomized 14 subjects to placebo and 25 subjects to escalating doses of adeno-associated virus type 1 encoding sarcoplasmic reticulum calcium ATPase (AAV1.SERCA2a). AAV1.SERCA2a was well tolerated, and the high-dose group met a 6 month composite endpoint. In the subsequent CUPID-2 study, 243 subjects received either placebo or the high dose of AAV1.SERCA2a. AAV1.SERCA2a administration, while safe, failed to meet the primary or any secondary endpoints. STOP-HF used plasmid endocardial injection of stromal cell-derived factor-1 to promote stem-cell recruitment. In a 93-subject trial of patients with ischemic etiology heart failure, the primary endpoint (symptoms and 6 min walk distance) failed, but subgroup analyses showed improvements in subjects with the lowest ejection fractions. A fourth trial randomized 14 subjects to placebo and 42 subjects to escalating doses of adenovirus-5 encoding adenylyl cyclase 6 (Ad5.hAC6). There were no safety concerns, and patients in the two highest dose groups (combined) showed improvements in left ventricular function (left ventricular ejection fraction and -dP/dt). The safety data from four randomized clinical trials of gene transfer in patients with symptomatic HFrEF suggest that this approach can be conducted with acceptable risk, despite invasive delivery techniques in a high-risk population. Additional trials are necessary before the approach can be endorsed for clinical practice.