To noninvasively study positional effects on superior vena caval configuration in humans, endoscopic ultrasonography was performed in 34 subjects including 20 with lung cancer, 5 with esophageal cancer and 9 with other diseases. None of the these subjects had cardiovascular involvement or respiratory dysfunction. A fiberoptic esophagoscope equipped with a 7.5 MHz linear array ultrasonic transducer at its tip (EPE-703, Toshiba-Machida) was used for the study. The actual movement of the superior vena cava (SVC) was clearly observed at the hilar level in all cases. During the cardiac cycle the anteroposterior diameter of the SVC was observed to reach a maximum at the atrial systole and reached a minimum at the late ventricular systole. With respiration, the SVC increased in diameter during inspiration and decreased during expiration. Moreover M and B mode figures of the SVC wall were recorded in left (LLD) and right decubitus (RLD) and supine position (SUP) in 34 subjects. On quiet ventilation of FRC level the diameter of the SVC was unchanged. Both the maximal and minimal diameters of the SVC, which were corrected for body surface area (BSA), were 11.3 +/- 0.3 (mean +/- SEM) mm/m2 and 9.8 +/- 0.3 mm/m2 in right lateral decubitus position, 9.4 +/- 0.3 and 7.9 +/- 0.3 in the supine position, 8.5 +/- 0.3 and 7.1 +/- 0.3 in the left lateral decubitus position, respectively. The size of the SVC was the greatest in the right lateral decubitus position and was the smallest in left lateral decubitus position (p less than 0.01, multiple comparison). It was suggested that the geometry of the SVC is influenced by thoracic pressure and gravity and that it behaves very similarly to pulmonary vascular vessels as a collapsible tube.