A systematic evaluation of the accuracy of continuous wave echo Doppler measurements across prosthetic valve leakages and regurgitant lesions has not been performed. Continuous echo Doppler velocity measurements in an in vitro, steady flow model, across the leaks of 12 intact mechanical prostheses and across six circular nozzles (area, 0.5-20 mm2) at pressure drops between 30 and 105 mm Hg were analyzed and compared to the velocities predicted by the modified Bernoulli equation. Laser Doppler anemometry of flow velocities through the nozzles was performed in addition. Despite excellent correlation, there was substantial overestimation of "Bernoulli predicted"-velocities by echo Doppler in the prosthetic leaks (mean +12.3 +/- 9.4%; range, 90.3-143.4%). Also in nozzles < or = 10 mm2, but not in those > 20 mm2, an overestimation of the "Bernoulli predicted"-velocities was observed (mean +6.2 +/- 2%). Laser Doppler anemometry of flow velocities through the nozzles showed slightly lower values than predicted by the Bernoulli equation. This effect apparently is due to transit time effects leading to spectral broadening and should be taken into account when using echo Doppler measurements in very small (< 10 mm2) orifices, such as mild to moderate regurgitant lesions and prosthetic valve leakage.