Right-sided heart failure in patients with left ventricular assist devices

Right heart failure (RHF) following the implantation of a left ventricular assist device (LVAD) is a significant and prevalent complication which increases morbidity and mortality. The pathophysiology of RHF post-LVAD implantation is complex and multifactorial which makes the prediction of RHF in LVAD patients challenging. There is currently no single index that can accurately predict the occurrence of RHF following LVAD implantation, despite the numerous suggested risk factors. As a result, various risk-scoring systems have been proposed to combine different variables and enhance prognostic accuracy. Additionally, managing RHF is crucial even before LVAD implantation. This can include the use of diuretics, inotropes such as levosimendan, and even temporary mechanical circulatory support to prepare the right ventricle for the increased demands after implantation. Temporary right ventricular assist device (t-RVAD) is a potential treatment option for selected patients with severe right ventricular dysfunction as a bridge-to-recovery or as a permanent solution. The objective of this thesis is to investigate the predictors, therapeutic options, and outcomes of patients with RHF following LVAD implantation. In chapter 1, we briefly introduce RHF following LVAD implantation, covering its prevalence, pathophysiology, prediction, and management. In chapter 2, we investigated the prediction of RHF post-HeartMate 3 implantation in the population of the European registry for patients with mechanical circulatory support (EUROMACS) using machine learning (ML) algorithms. We found that using the random forest model, tricuspid annular plane systolic excursion (TAPSE) emerged as the most critical predictor of RHF, demonstrating the highest variable importance score followed by serum creatinine while using the unadjusted Bayesian model, right ventricular ejection fraction (by visual scoring) was associated with increased odds of the RHF, with an OR of 4.45, indicating a substantial risk factor. On the other hand, using the adjusted regression model after ML variable selection, showed that severe tricuspid regurgitation has the highest odds ratio, with an OR of 2.59, signifying a strong association with RHF. A risk score was calculated using the variable importance from the random forest model and the results of the adjusted Bayesian regression coefficients with a cut-point of 2.297 is optimal for categorizing patients into low and high-risk groups (AUC = 0.702, sensitivity=63%, specificity=70%). In chapter 3, we explored the evidence related to the use of levosimendan for the prevention and/or mitigation of RHF in patients undergoing LVAD implantation. Pre-operative optimization of heart failure therapy could help in the prevention, and/or mitigation of RHF. Levosimendan is a non-conventional inotropic agent that works by amplifying the calcium sensitivity of troponin C in cardiac myocytes, without increasing the intra-cellular calcium or exacerbating ischemia. In this chapter, we sought to conduct a systematic review to synthesize evidence related to the use of levosimendan for the mitigation and/or prevention of RHF in patients undergoing LVAD implantation. We found two studies only met the predefined criteria. Although limited, these data on levosimendan use in patients undergoing LVAD are encouraging and suggest that there are at least hemodynamic improvements alongside improved organ perfusion. So far, no survival benefits have been shown for the use of levosimendan in LVAD patients. In chapter 4, we explored the effectiveness of levosimendan on the occurrence of postoperative RHF, the need for RVAD implantation, length of ICU stays, and 30-day and 1-year mortality in patients undergoing LVAD implantation compared to propensity score-matched controls using the EUROMACS database. We found that 3661 patients received mainstream LVAD, of which 399 (11%) were treated with pre-LVAD levosimendan. In the largest European LVAD registry, we found that patients who received levosimendan had significantly higher risk of developing RHF (based on the EUROMACS RHF score), received more RVAD in the early postoperative period, and stayed more days in the ICU. However, there was no significant difference regarding the occurrence of post-LVAD early severe RHF or mortality (at 30-day and 1-year). Moreover, in a propensity score-matched cohort we found no evidence for an association between the preoperative use of levosimendan and the occurrence of post-LVAD RHF, the need for RVAD implantation in the early postoperative period, the length of ICU stay, or mortality (at 30-day and 1-year) following LVAD implantation. In chapter 5, we investigated the evidence about the safety and efficacy of using the temporary right ventricular assist device (t-RVAD) as a treatment option in patients with severe right ventricular failure. We conducted a systematic review and found Thirty-one studies met the inclusion criteria, from which data were extracted. Successful t-RVAD weaning ranged between 23% and 100%. Moreover, 30-day survival post temporary RAVD implantation ranged from 46% to 100%. Bleeding, acute kidney injury, stroke, and device malfunction were the most frequently reported complications. Notwithstanding, t-RVAD is a lifesaving option for patients with severe RHF, but the evidence stems from small non-randomized heterogeneous studies utilizing a variety of devices. In chapter 6, we discussed the design of the EEVAD study, which is a prospective, multicenter, non-randomized, investigator-initiated clinical trial recruiting consecutive patients treated with HeartMate 3 LVAD. The EEVAD study aims to characterize the evolution of right heart function and geometry post-LVAD implantation and to develop a validated risk stratification tool for post-LVAD RHF. The co-primary endpoints include early RHF (within 30 days) and late RHF (>30-day and up to one year) following LVAD implantation. Imaging analyses are conducted by a Core Lab using echocardiography, CT, and MRI. Secondary endpoints encompass all-cause mortality, ICU/hospital stay length, stroke, bleeding, infections, valvular dysfunction, arrhythmias, rehospitalizations, LVAD explant, cardiac transplantation, and recovery. Patients are followed for two years following LVAD. Predictors of post-LVAD RHF and other outcomes are identified using conventional statistics and machine learning, including neural networks. EEVAD Extended will further explore RHF predictors via biomarker discovery. In conclusion, this thesis provides an overview of right-sided heart failure post-LVAD including how to predicate RHF post-HeartMate 3 implantation in the population of the EUROMACS using machine learning algorithms. It also explored the evidence related to the use of levosimendan for the prevention and/or mitigation of RHF in patients undergoing LVAD implantation in the current literature. Moreover, we explored the effectiveness of levosimendan on the occurrence of postoperative RHF, need for RVAD implantation, length of ICU stays, and 30-day and 1-year mortality in patients undergoing LVAD implantation compared to propensity score-matched controls using the EUROMACS database. Additionally, it investigated the evidence about the safety and efficacy of using the temporary RVAD as a treatment option in patients with severe right ventricular failure.

Publisher

University of Galway

Rights

Attribution-NonCommercial-NoDerivatives 4.0 International

cbn

Collections

University of Galway Theses (PhD Theses)