The left ventricular ultrasound videointensity of an intravenous injection of sonicated albumin is improved if the agent is diluted with dextrose prior to sonication. The objective of this study was to determine the mechanism for improved left ventricular ultrasound contrast with intravenous sonicated multi-fold dilutions of albumin with dextrose compared to sonicated albumin alone. Epicardial short axis images of the left ventricle were obtained in 11 mongrel dogs and incremental one part sonicated dilutions (up to 10-fold) of albumin with 5 or 50% dextrose were given intravenously to determine which dilution and dextrose concentration produced optimal left ventricular videointensity. Microbubble size and concentration of these dilutions were measured. The one to seven-fold sonicated dilutions resulted in a slight, but significantly larger microbubble size when compared to sonicated albumin alone (SA), but no difference in concentration. All dilutions produced significantly higher end-diastolic peak videointensity (PVI) in the left ventricle than SA (range 160-569% of SA PVI; p < 0.001) with the three to five-fold dilution producing maximal PVI. Five percent dextrose dilutions produced the same videointensity as 50% dilutions. End-systolic videointensity of both 5 and 50% dextrose dilutions were also over 250% higher than SA (p < 0.001). This resulted in good or excellent end-systolic endocardial border definition in the majority of injections. Therefore, the mechanism for improved left ventricular chamber opacification with multifold sonicated dilutions of albumin with dextrose appears to be due to a small increase in microbubble size and not increased viscosity or microbubble concentration.