Validation of cardiac output estimates by digital video subtraction angiography in dogs. Correlation with thermodilution values.

To determine the validity of cardiac output estimates obtained by digital video subtraction angiography (DVSA), they were compared with thermodilution values in 6 anesthetized dogs. Multiple levels of cardiac output were produced in each dog by graded openings of peripheral arteriovenous fistulas. Both direct contrast and mask-mode images were evaluated. There was good correlation between thermodilution and direct contrast values (r = 0.88) as well as between thermodilution and mask-mode values (r = 0.89), though the mask-mode outputs were 12 +/- 1.1% less than the direct contrast values (p less than 0.05). Direct contrast and mask-mode values correlated well (n = 6; r greater than or equal to 0.95). Digital outputs exhibited little variation, indicating excellent reproducibility. These results demonstrate the validity of area-length estimates of cardiac output by DVSA.