CrossTalk proposal: Blood flow pulsatility in left ventricular assist device patients is essential to maintain normal brain physiology

Article published in The Journal of Physiology, available at https://doi.org/10.1113/JP276729


Edited by: Francisco Sepúlveda and Emma Hart
For the first time in history, some humans live without a palpable pulse (Purohit et al. 2018). This remarkable physiology is the consequence of surgical implantation of a continuous-flow left ventricular assist device (CF-LVAD) in patients with end-stage heart failure. Blood flow produced by CF-LVADs has a low oscillatory profile in the aorta that results in significantly reduced pulsatility in all arterial compartments (Castagna et al. 2017;Fig. 1). Despite remarkable gains in quality of life and longevity, complications that affect not only morbidity, such as gastrointestinal bleeding, but also mortality, such as strokes, are still prevalent in CF-LVAD patients. Low pulsatility has been proposed as a major culprit in contributing to these adverse events (Mancini & Colombo, 2015;Goldstein et al. 2018). In this CrossTalk proposal, we present the current arguments Eric J. Stöhr trained in exercise science in Germany and obtained his PhD in 2011 in human cardiovascular physiology in the UK. After postdoctoral studies and appointment to faculty, he was awarded a Marie Skłodowska-Curie Fellowship and joined Columbia University Irving Medical Center in 2016 where he studies advanced heart failure patients. His research aims at understanding the interaction between the heart muscle dynamics and arterial function in health and disease. Joshua Z. Willey is a vascular neurologist with a research interest in cerebrovascular physiology and disease with mechanical circulatory support. He completed his MD, neurology training and stroke/epidemiology fellowships all at Columbia University Medical Center where he is now an Assistant Professor of Neurology. in favour of maintaining an appropriate amount of arterial pulsatility, in particular in the cerebral circulation, to lower risk in these patients.

Cerebral microcirculation and O 2 kinetics
A macro-circulatory link between cardiac output, aortic stiffness and arterial pulsatility with the brain is well-established (Mitchell et al. 2011;Jefferson et al. 2015). At the level of the microcirculation, it is thought that the healthy circulation already presents with absence of pulse pressure (O'Rourke & Hashimoto, 2007), and hence CF-LVADs would not create a different environment for gas exchange from normal physiology. However, even in healthy individuals, measurements of arteriolar haemodynamics have revealed pulsatile patterns (Rappaport et al. 1959;Shore, 2000). An important implication is that a pulsatile velocity profile entails that cerebral transit time slows in the diastolic phase and facilitates the oxygen gradient for gas exchange. In CF-LVAD patients, the increased diastolic blood velocity may result in an overall elevated mean blood velocity (Brassard et al. 2011;Castagna et al. 2017;Fig. 1B), thereby impairing oxygen kinetics (Wardlaw et al. 2002). However, data on absolute blood velocities are scarce, or their interpretation currently lacks confidence because the assessment of cerebral blood velocities, even in the pre-arteriolar circulation, has typically not been performed with the necessary angle correction of the Doppler signal. Whatever the real O 2 kinetics in CF-LVAD, it is known that cerebral blood flow is also regulated for reasons other than O 2 requirements (Mintun et al. 2001). Thus, the low pulsatile, diastolic-dominant haemodynamics of CF-LVAD impact on cerebral artery properties beyond gas exchange, as discussed in the following paragraphs.

Cerebral autoregulation
Cerebral autoregulation has been proposed to take effect across a more narrow range of perfusion pressure than previously thought (Willie et al. 2014). Consequently, the low systolic blood pressure and low-to-normal mean arterial pressure coupled with a normal cardiac output mean that CF-LVAD patients may find themselves on an unusual point of the perfusion-cerebral blood flow (CBF) curve, with high flow into a low-resistance cerebral circulation (Cornwell et al. 2014). The high-flow low-resistance is directly caused by the low-pulsatile haemodynamics of CF-LVAD. Notwithstanding, cerebral autoregulation may be preserved in CF-LVAD patients (Ono et al. 2012;Cornwell et al. 2014), independent of end-tidal CO 2 concentrations (Cornwell et al. 2014). However, some remaining differences to normal brain physiology can be noted. For instance, the variance in CBF was most similar between healthy individuals and CF-LVAD patients, while patients with pulsatile devices responded significantly differently to a sit-to-stand challenge (Cornwell et al. 2014). These intriguing findings may indicate a meaningful role of added pulsatility in the context of LVAD and justify a more detailed investigation into the dynamics of perfusion pressure (i.e. pulse pressure) and cerebral autoregulation in the setting of low absolute pressures (Ono et al. 2017). Rather than being disturbed itself, the maintained cerebral autoregulation in CF-LVAD may cause a reduction in pulsatility since the total flow is already high.

Endothelial function, bleeding and aortic stiffness
Pulsatility of flow causes cyclical stretch of the arterial wall that is a critical contributor to endothelial production of J Physiol 597.2 nitric oxide and cardiovascular health (Hahn & Schwartz, 2009). The high occurrence of bleeding events, such as GI bleeding and haemorrhagic strokes, indicates a primary problem with endothelial integrity. A recent study confirms elegantly that a staggering proportion of LVAD patients have cortical microbleeds in a pattern similar to cerebral amyloid angiopathy, a condition with high rates of arteriolar fragility (Yoshioka et al. 2017). Furthermore, reduced pulsatility appears responsible for the marked reduction in endothelial nitric oxide bioavailability in CF-LVAD patients when compared to those on support with pulsatile device (Witman et al. 2015), although this may be more relevant in the systemic circulation than in the brain (Zhang et al. 2004). While shear rate has not been measured in the cerebral circulation of CF-LVAD patients, it is conceivable that it would be higher than normal in the diastolic phase of the cardiac cycle, a circumstance that, when present in the carotid artery, has been associated with adverse cerebral events in non-LVAD populations (Mutsaerts et al. 2011). In addition, the high diastolic flow likely contributes to increased arterial stiffness observed in CF-LVAD patients by markedly attenuating the normal systolic-diastolic stretch and recoil cycle (Ambardekar et al. 2015;Patel et al. 2017). It is important to underline that in pulsatile circulations, aortic stiffness increases the transmission of pulsatility to the periphery, and, if exceeding normal pulsatility, is detrimental to the brain and other end-organs (Webb et al. 2012). Paradoxically, this means that the reduced Windkessel effect in CF-LVAD patients because of the larger diastolic flow and increased aortic stiffness might be beneficial in some individuals via a mild augmentation of pulsatile dynamics transmitted to the periphery, which would otherwise be harmful to end-organs. Finally, elegant insight into bleeding-associated complications in CF-LVAD -which may include blood-brain barrier disruption and cortical microbleeds -has been provided by Vincent et al. (2018). These authors showed that the loss of von Willebrand factor from the high shear forces within the mechanical device was, at least in part, offset by increased arterial pulsatility, which promoted new von Willebrand factor release from the endothelium. Hence, mild increases in arterial pulsatility may mitigate bleeding risk in CF-LVAD patients.

Additional considerations
Two common misconceptions related to CF-LVAD physiology, and specifically pulsatility, deserve attention. First, it is commonly assumed that CF-LVADs should produce perfectly continuous flow if the aortic valve does not open (Floras et al. 2015). This assumption overlooks the role of fluctuations of the intra-ventricular pressure within each cardiac cycle. The resulting changes in pressure gradient between LVAD inflow and aortic outflow graft creates variability in pump flow between systole and diastole and thereby generates arterial pulsatility (Khalil et al. 2008;Pagani, 2008).
Second, the absolute blood volume in relation to the pulsatility is often ignored. Although pulsatility is typically reduced with a higher LVAD speed, the concomitant increase in cardiac output may have significant effects beyond that of reduced pulsatility. Acutely, a larger flow into the cerebral circulation will result in increased resistance and possibly higher pressure. In any case, it is important to consider cardiac output in relation to the local peripheral vasodilatation and vasoconstriction. Studies examining the effects of pulsatile cardiopulmonary bypass reported that the number of perfused vessels in the microcirculation was increased compared with a continuous-flow circulation (O'Neil et al. 2012;Inamori et al. 2013). Importantly, the authors also reported, 'pulsatility resulted in a reduction in the prevalence of pathologic hyper-dynamically perfused vessels' (O'Neil et al. 2012). This observation strongly supports a role of pulsatility independent of blood volume since the latter was not significantly different between pulsatile and continuous-flow bypass. One final comment relates to the newest generation of CF-LVADs. Whether the recent improvements in outcomes, including the reduced incidence of stroke in HeartMate 3 patients (Mehra et al. 2018), can be attributed to the added pulsatility and the greater load-sensitivity of the device itself -and hence greater intrinsic pulsatile oscillation within one cardiac cycle (Pagani, 2008) -remains to be confirmed. Collectively, the presented evidence suggests that CF-LVAD patients are currently not exposed to a normal brain physiology and that mild increases in arterial pulsatility may be beneficial.

Call for comments
Readers are invited to give their views on this and the accompanying CrossTalk articles in this issue by submitting a brief (250 word) comment. Comments may be submitted up to 6 weeks after publication of the article, at which point the discussion will close and the CrossTalk authors will be invited to submit a 'LastWord' . Please email your comment, including a title and a declaration of interest, to jphysiol@physoc.org. Comments will be moderated and accepted comments will be published online only as 'supporting information' to the original debate articles once discussion has closed.