Thirty-two consecutive infants with transposition of the great arteries (TGA) and ventricular septal defect underwent subxiphoid 2-dimensional echocardiography (2-D echo). Two independent observers prospectively evaluated each echocardiogram for the presence or absence of left ventricular (LV) outflow tract obstruction, whether outflow obstruction was dynamic or fixed, or both, and the precise anatomic type of fixed obstruction. Compared with the LV-to-pulmonary artery gradient determined at cardiac catheterization, 2-D echo yielded low false-negative (7 to 13%) and false-positive (0 to 6%) rates for diagnosing the presence or absence of LV outflow tract obstruction. Moreover, the false-negative cases were only minor errors, because the measured LV-pulmonary artery gradients proved to be less than 25 mm Hg. Compared with the long-axial oblique LV angiogram, 2-D echo yielded no false-negative results in detection of outflow tract obstruction, which was at least partly fixed. Compared with autopsy/surgical observation, 2-D echo made no significant errors in delineating the exact anatomic type of fixed obstruction. The diagnostic accuracy of 2-D echo in detecting and characterizing LV outflow tract obstruction limits the need for "routine" cardiac catheterization before repair in infants with TGA and intact ventricular septum. Furthermore, because certain types of fixed LV outflow tract obstruction are difficult for the surgeon to visualize and alleviate, precise knowledge of the anatomic type of fixed obstruction influences the choice among Rastelli, intraatrial baffle and arterial switch procedures in patients with TGA and ventricular septal defect.