This study evaluated the role of interelectrode spacing, pulse widths greater than 10 ms, and depth of esophageal insertion on minimizing transesophageal atrial pacing threshold in 30 patients aged 1 day to 77 years. Interelectrode spacings of 15, 22 and 28 mm were evaluated by establishing strength-duration curves in 2 or more serial studies in 12 patients; electrode spacing had no effect on pacing threshold. In 23 patients studied with 22-mm electrode spacing, pulse widths of 15 and 20 ms had no significant effect on current threshold requirements compared with 10-ms pulse widths. In 20 patients, pacing threshold and esophageal electrograms were obtained at 1.0- to 2.5-cm intervals with a 22-mm lead using a pulse width of 10 ms. Average minimal pacing threshold was 10.2 mA (range 4.5 to 20). The site of minimal pacing threshold was highly correlated with patient height (r = 0.987), and occurred within 1.1 cm (0 to 2.5 cm) of the site of the maximal bipolar atrial electrogram amplitude and 0.95 cm (0 to 3 cm) of the site of the maximal unipolar atrial electrogram. Bipolar electrode spacing of 15, 22 or 28 mm has little effect on transesophageal pacing threshold. In most patients, pulse widths greater than 10 ms do not significantly decrease pacing threshold. Correct catheter insertion depth is critical to minimize pacing threshold and may be predicted by either the site of the maximal atrial electrogram amplitude or patient height.