We evaluated the performance of 3 different left ventricular leads (LV) for resynchronization therapy: bipolar (BL), quadripolar (QL) and active fixation leads (AFL). We enrolled 290 consecutive CRTD candidates implanted with BL (n = 136) or QL (n = 97) or AFL (n = 57). Over a minimum 10 months follow-up, we assessed: (a) composite technical endpoint (TE) (phrenic nerve stimulation at 8 V@0.4 ms, safety margin between myocardial and phrenic threshold <2V, LV dislodgement and failure to achieve the target pacing site), (b) composite clinical endpoint (CE) (death, hospitalization for heart failure, heart transplantation, lead extraction for infection), (c) reverse remodeling (RR) (reduction of end systolic volume >15%). Baseline characteristics of the 3 groups were similar. At follow-up the incidence of TE was 36.3%, 14.3% and 19.9% in BL, AFL and QL, respectively (p < 0.01). Moreover, the incidence of RR was 56%, 64% and 68% in BL, AFL and QL respectively (p = 0.02). There were no significant differences in CE (p = 0.380). On a multivariable analysis, "non-BL leads" was the single predictor of an improved clinical outcome. QL and AFL are superior to conventional BL by enhancing pacing of the target site: AFL through prevention of lead dislodgement while QL through improved management of phrenic nerve stimulation.