Regional vessel wall distensibility was determined by measuring luminal area and pressure, using intravascular ultrasound (Sonicath; Boston-Scientific, Watertown, MA, USA; 3.5 Fr, 30 MHz) in 45 left coronary sites from 40 patients. Luminal area in diastole (A) and in systole was measured at the diseased sites. With the ratio of luminal area changes (dA) to coronary pressure changes (dP) during a cardiac cycle, the total distensibility index was calculated by the formula: [(dA/A)/dP] x 10(3). At sites with non-circumferential disease, perimeters in diastole (L) and in systole were measured at the normal and diseased portions, and the changes in perimeters (dL) during a cardiac cycle were calculated. The regional distensibility index was obtained by the formula: [(dL/L)/dP] x 10(3). In 22 sites with circumferential disease, the total distensibility index was 1.03 +/- 0.61/mmHg, significantly lower than that for 23 sites with non-circumferential disease (1.45 +/- 0.89/mmHg; P < 0.05). In non-circumferential disease, the regional distensibility index at the diseased portion was significantly lower (0.33 +/- 0.47/mmHg) than that at the normal portion (1.11 +/- 0.75/mmHg; P < 0.01). Coronary sites with residual non-circumferential disease after angioplasty also exhibited heterogeneity of regional distensibility (0.73 +/- 0.76 at disease sites versus 1.58 +/- 0.95/mmHg at normal sites, n = 10, P < 0.05). These results indicate that heterogeneous regional wall distensibility exists at sites with non-circumferential disease where the total vessel distensibility is preserved by the presence of compliant normal portion. This heterogeneity of regional wall distensibility may represent a biomechanical factor that explains the mechanism of plaque rupture that occurs mainly at the shoulder of the non-circumferential disease.