Assessment of Left Ventricular Contractility and Loading Conditions Under Continuous-flow Left Ventricular Assist Device Support

Abstract

Continuous-flow left ventricular assist devices (cf-LVADs) now form a cornerstone of the treatment of advanced heart failure. With their increasing use, improving our understanding of the mechanical, histopathological and physiological interactions between these devices and the patients in whom they are implanted is paramount. Initially, through non-invasive assessment of the impact of dynamic manoeuvres on the pump flow waveform and left ventricular dimensions, I demonstrate that changes in afterload pressure, posture and intrathoracic pressure have significant and highly variable effects on pump flow. The relationship between the intrathoracic pressure changes, loading conditions and pump flow is then assessed invasively, with low preload, low arterial resistance and increased ventricular-arterial coupling predictive of pulsatility loss and suction events during modified Valsalva manoeuvre. Using a pulsatile mock loop circulation, I assess the contribution of the outflow conduit to pump afterload. Haemodynamically significant gradients can be generated across an unobstructed HVAD outflow graft, and their magnitude predicted using an empirically derived model incorporating conduit diameter, mean pump flow, systolic dQdt and conduit length. I then demonstrate in vivo that some degree of tissue ingrowth into the cf-LVAD outflow graft due to acute inflammatory, chronic inflammatory, fibrotic or neointimal reaction is a near-universal phenomenon and is associated with a small but measurable decrease in pump flow and flow pulsatility over time. In order to enable an integrated assessment of left ventricular contractility, energetics and loading conditions, I describe a method to derive pressure-volume loops using non-invasive inputs that are readily available in the clinic setting. This method is validated invasively by assessing its ability to detect predictable pharmacodynamic effects of intravenous Milrinone. Finally, I utilise this pressure-volume derivation in order to assess the haemodynamic effects of exercise, revealing increased left ventricular contraction that is likely driven by increased preload, and profoundly diminished unloading effect of increased pump speed during exercise. Overall, this thesis sheds light on the complex interplay between ventricular contractility, loading conditions and cf-LVAD performance under ‘real-world’ conditions, with significant implications for both clinical practice and future research in this area.