Purpose Effectiveness of cardiac resynchronization therapy (CRT) in patients without left bundle branch block (non-LBBB) QRS morphology is limited. Additional selection criteria are needed to identify these patients. Methods Seven hundred ninety consecutive patients with non-LBBB morphology, who received a CRT-device in 3 university centers in the Netherlands, were selected. Pre-implantation 12-lead ECGs were evaluated on morphology, duration, and area of the QRS complex, as well as on PR interval, left ventricular activation time (LVAT), and the presence of fragmented QRS (fQRS). Association of these ECG features with the primary endpoint: a combination of left ventricular assist device (LVAD) implantation, cardiac transplantation and all-cause mortality, and secondary endpoint-echocardiographic reduction of left ventricular end-systolic volume (LVESV)-were evaluated. Results The primary endpoint occurred more often in non-LBBB patients with with PR interval >= 230ms, QRS area < 109 mu Vs, and with fQRS. Multivariable regression analysis showed independent associations of QRS area (HR 2.33 [1.44, 3.77],p= 0.001) and PR interval (HR 2.03 [1.51, 2.74],p< 0.001) only. Mean LVESV reduction was significantly lower in patients with baseline RBBB, QRS duration < 150 ms, PR interval >= 230 ms, and in QRS area < 109 mu Vs. Multivariable regression analyses only showed significant associations between QRS area >= 109 mu Vs (OR 2.00 [1.09, 3.66]p= 0.025) and probability of echocardiographic response to CRT. Conclusions In the heterogeneous non-LBBB patient population, QRS area and PR prolongation rather than traditional QRS duration and morphology are associated to both clinical and echocardiographic outcomes of CRT.