In-vivo prediction of cavitation near a Medtronic Hall valve.

BACKGROUND AND AIM OF THE STUDY: Quantification of cavitation near mechanical heart valves in vivo is currently based on measurements of high-frequency pressure fluctuations (HFPF). Until now, mechanical resonance components have been removed using a high-pass filter. However, isolating cavitation and resonance signal components by separating the deterministic and non-deterministic parts of the HFPF signal has recently been proposed as a new method. It was hypothesized that the non-deterministic signal energy, Enon-det, of the HFPF signal correlates with previous parameters indicating cavitation in vivo, specifically the root mean square pressure after appropriate high-pass filtering (P(RMS)) and the rate of change in transvalvular pressure with respect to time (dP/dt).

METHODS: Medtronic Hall 29 mm mitral valves were implanted in five pigs (body weight 80 kg). The hemodynamic condition was varied by infusion of dobutamine and blood volume regulation to achieve a wide range of dP/dt (100 to 4,700 mmHg/s). According to previous studies, P(RMS) was deduced using the HP-filtering cut-off frequency of 50 kHz for this particular valve. The transvalvular dP/dt was derived as the average slope 5 ms prior to valve closure.

RESULTS: Power relationships were found between Enon-det and P(RMS) (P(RMS) = 1.27 x Enon-det0.49), and between Enon-det and dP/dt. The correlation between Enon-det and P(RMS) was very strong (r = 0.98), whereas the correlation between Enon-det and dP/dt was notably weaker (r = 0.68).

CONCLUSION: Enon-det was an excellent predictor of P(RMS), obtained after appropriate filtering. The present data suggested that the signal energy above 50 kHz was almost completely non-deterministic. It can therefore be proposed that Enon-det is a more direct and easily accessible parameter than those previously suggested for the purpose of cavitation quantification. This parameter is able to quantify the pressure fluctuations thought to occur from cavitation in vivo, and may be applicable for non-invasive recordings.