Rick Schreurs

120 Chapter 6 Study Limitations In this study the pacing experiments were performed in relatively healthy canine hearts. Previous work from our group showed that chronic total AV-block leads to structural changes (hypertrophy) and electrical remodeling (QT-time prolongation), but normal contractility [41]. On the other hand, patients treated with CRT have different levels of myocardial remodeling and heart failure, which may affect the response to pacing [42]. Results of this study might differ from patient data since long-term structural remodeling was not included in both the animal and computational experiments. On the other hand, differences in RV and LV response observed in this study could mean that different pacing settings might affect the positive and/or negative remodeling of the LV and RV. A major difference between this animal experiment and patients in day-to-day life is that the animals were anesthetized. To allow comparison between the experiments and simulations, the model’s LV afterload in the baseline situation was adapted to fit the measured mean arterial pressure in the dogs. After the change in activation delays regulation was disabled, which is likely similar to the anesthetic condition were regulation is slow. Caution should however be taken when translating results of this study towards the clinical setting considering that loading conditions potentially affect the effect of pacing delay changes [43]. Future studies are needed to investigate how load-dependent the observed effects of pacing delay optimization are and how homeostatic regulation interacts with changes in LV and RV contractility. A final limitation of the CircAdapt (and most other computer models in this field) is that changes in function of the autonomic nervous system are not taken into account. Clinical perspective The results of this study raise the question what outcome measure is best to use for optimization of pacing delay. Current clinical practice focusses almost exclusively on the LV, using parameters like LV dP/dt max , aortic outflow integral, and LV systolic (or aortic) pressure. This study demonstrates that improving LV function can reduce RV function. Furthermore, cardiac output is not necessarily increasing when LV dP/dt max increases. Hence, an exclusive focus on the LV might not lead to the best overall outcome. Therefore, future studies on optimization of therapy should not exclusively focus on the LV but also include measures of RV and/or whole heart function. In our analysis of cardiac output we also demonstrated that the moment of measurement affects what physiological phenomenon is actually observed. Contractile function alone might better be observed in the first beats while longer lasting measurements, that allow reaching a steady state, will provide more information about the loading of the heart and its interaction with homeostatic regulation. Insights acquired in the present opens the way for designing better optimization protocols, possibly even including computer modeling.