To evaluate the accuracy of flow measurements in aortocoronary bypass grafts with the ultrasound transit-time method, an in vitro and in vivo comparison was carried out. The in vitro comparison with evaluation against both true flow and the ultrasound Doppler method, was carried out with a fresh saphenous vein mounted in a pulsatile flow rig. The two flow probes were placed on the graft 4-5 cm apart to avoid acoustic interference, and blood was pumped through the system at different flow rates. The comparison between the methods showed excellent agreement with a linear correlation coefficient of 0.996, and a mean error of -2.9 ml/min with limits of agreement +/- 13.1 ml/min (+/- 2 SD = 95% of measured differences between these limits). However, against true flow, both methods overestimated flow slightly with mean error 4.4 and 7.3 ml/min for the transit-time and Doppler, respectively. Both methods showed excellent correlation with true flow (correlation coefficient 0.998 for the transit-time and 0.997 for the Doppler method). The in vivo accuracy was evaluated by comparison of the two methods in 9 patients. The two probes were placed on the same saphenous vein grafts 4-5 cm apart, and a total of 34 measurements in 17 grafts were carried out including measurements at baseline and after papaverine injection. The correlation coefficient was 0.990 and linear regression analysis gave the equation: Transit-time flow = 1.00 x Doppler flow + 1.3. In terms of flow, the mean error was 1.5 ml with limits of agreement +/- 17.2 ml.(ABSTRACT TRUNCATED AT 250 WORDS)