Can cavitation bubbles generated by mechanical heart valves be detected by transcranial Doppler?

Increasingly, transcranial Doppler (TCD) systems are being considered as potential tools to determine the thrombogenicity of herat valve prostheses and/or to serve as early warnings of impending thromboembolic strokes. TCD is thought to detect the presence of intracranial arterial microemboli based on high intensity signals (HITS). While the exact cause of these HITS is not known, it has been theorized that they are due in part to mechanical valve induced cavitation. To determine if this is correct, in vitro experiments were conducted to find if TCD could detect mechanical valve induced cavitation bubbles. An in vitro flow system was developed in which 29mm Medtronic Hall, St. Jude Medical and Medtronic Intact valves, one each, were simultaneously subjected to validated non-cavitation and cavitation generating conditions. Polystyrene particles as well as air bubbles were infused into the flow system to provide controls for comparison. The flow field proximal to each valve was interrogated using TCD sample volumes located 2.5 cm proximal to, 1.3 cm proximal to, and at the valvular inflow surface. The TCD system accurately detected both infused air bubbles and polystyrene particles at all sample volume locations. Under known cavitation generating conditions without infused air bubbles, the TCD failed to register HITS at any sample volume location. Even though cavitation at the valve surface might have been masked by a "system generated artifact", cavitation formed bubbles must persist and pass through the two proximal sample volumes if they were to be captured intracranially. These data suggest that the TCD is unable to detect mechanical valve induced cavitation bubbles either because the bubbles are too small, too few, or too short lived. It is therefore probable that clinically recorded HITS associated with mechanical valves are from origins other than valve induced cavitation bubbles.