Rick Schreurs

98 Chapter 5 standard echocardiographic optimization [41]. More recently the RESPOND-CRT trial, a prospective, multicenter, randomized, double-blind study in 1039 patients with systolic HF and wide QRS complex, showed that RA SonR1 based optimization of CRT is non-inferior to standard echocardiographic optimization in terms of responder rate and results in a 35% risk reduction for heart failure hospitalization [42, 43]. The current animal study provides support for the use of the SonR system and provides a better understanding of the pathophysiology of the signal. Although the present study is small, several findings may be worthwhile for further investigation in order to further improve the SonR algorithm: • Our study has shown that the SonR1 signal is suitable for AV-delay optimization during LV-only pacing. This indicates that the SonR1 algorithm might be used for optimizing LV fusion pacing, a pacing mode that has been shown superior to biventricular pacing when using the adaptive CRT algorithm [44, 45]. • The present study shows some variability in the SonR data, as evidenced from the AV-scans shown in figure 2 . This variability hampers accurate determination of the inflection point. This may be circumvented by using a longer sampling time. Alternatively, the data show that the exact point is not important, as long as SonR1 does not increase exponentially, as is the case during the short AV-delays. So using an AV-delay that is just several tens of milliseconds longer than the range where the exponential increase is seen seems to be sufficient. Future perspectives The SonR1 peak-to-peak amplitude is just one simple characteristic of the acquired SonR signal. It may well be that there is complementary information in the SonR signals, like the width of SonR1, the time delay between the M1 and T1 components as well as the time between onset of QRS complex and S1 (elsewhere referred to as electromechanical activation time or EMAT [46]). In addition, also the SonR equivalent of the second heart sound, S2, may contain useful information such as the amount of splitting of S2 (timing difference between the closure of the aortic (A2) and pulmonary (P2) valves. Finally, optimal diastolic filling (leading to enhanced systolic function), can be calculated from SonR1 and SonR2 components since they appear to correlate to mitral and aortic valve closure [47]. Therefore, it seems worthwhile to design future studies, searching for even more comprehensive algorithms for hemodynamic optimization, such as multiparameter analysis of the accelerometer signals. Limitations Main limitation of this study is the small number of animals and the variability in conduction abnormalities. On the other hand, the fact that a relation was found between SonR1 on the one hand and contractility and AV-delay on the other, despite this variability, indicates that this study provides better understanding of the origin of the SonR1 signal. Further support for the ideas unfolded here require additional experiments. Results from animal