Search

Background Right heart failure (RHF) remains a major source of morbidity and mortality after left ventricular assist device (LVAD) implantation, yet efforts to predict RHF have proved only modestly successful. Acute pharmacologic unloading of the LV to assess ventricular and hemodynamic reserve may better predict post-implant RHF. Objectives To compare pre-implant baseline and post-vasodilator hemodynamic predictors of INTERMACS defined severe or severe-acute RHF after LVAD implantation. Methods After IRB approval, we performed a retrospective analysis of all subjects implanted with a durable LVAD at our institution between February 2017 and April 2020 (n=108). Only those who underwent vasodilator testing with nitroprusside during their pre-implant right heart catheterization were ultimately included (n=29). Continuous variables were compared with t-test and Levene's test of equality of variances; categorical via Fisher's Exact test. Logistic regression was used to determine predictors of severe or severe-acute RHF. Results Ten subjects met criteria for severe or severe-acute RHF (34%) after implant. There were no significant differences in demographics, HF etiology, INTERMACS profile, echo assessment of LV/RV function, or pre-implant baseline hemodynamics including RAP:PAWP ratio and PAPi between groups (TABLE). Both groups saw a similar reduction in systemic mean arterial pressure with nitroprusside, which led to significant and similar reductions in PAWP and pulmonary pressures. Compared to the group with severe or severe-acute RHF, those without significant RHF achieved higher peak cardiac index (CI) with nitroprusside infusion, suggesting better cardiac reserve. Peak CI was a significant predictor of RHF (p= 0.022), even after controlling for baseline values (p=0.048). No subject achieving a peak CI > 2.66 L/min/m2 developed severe or severe-acute RHF, and every 0.25 L/min/m2 decrease in peak CI was associated with 2.3 fold increased risk. Mortality was higher and hospital length of stay longer in the severe or severe-acute RHF group. Conclusion Peak CI with nitroprusside infusion predicts post-implant LVAD RHF better than resting hemodynamics in this small cohort. Hemodynamic reserve measurements may prove to be an important risk stratification tool when assessing LVAD candidacy.

Introduction Respirophasic hemodynamic patterns are prognostic in the setting of heart failure (e.g. Kussmaul's sign), though the implications of hemodynamic respirophasic variation patterns in patients supported with a left ventricular assist device (LVAD) are unknown. Many LVAD patients remain poorly hemodynamically optimized, and the ability to optimize these patients with ramp speed increase is a positive prognostic marker. However, the response to ramp speed increase may be unpredictable. Alterations in hemodynamics with respiration may predict ramp speed response by reproducing pre-load alterations during respiration similar to LVAD speed manipulation. Hypothesis Absence of pulmonary artery wedge pressure (PAWP) respiratory variation in patients with LVADs is associated with an inability to reduce the PAWP with increasing pump speed. Absence of right atrial pressure (RAP) respiratory variation is a marker of right ventricular (RV) dysfunction. Methods We retrospectively collected data from 16 separate, consecutive hemodynamic ramp studies on 12 patients with HeartMate II LVADs (average 880 ± 483 days of LVAD support). All PAWP and RAP at inspiration and expiration were independently measured in a blinded fashion. Patient demographic and clinical data was obtained from chart review. No patient had clinically significant pulmonary disease. Results The degree of respiratory variation of PAWP correlated with PAWP reduction with speed increase (R = 0.81). PAWP respiratory variation also correlated with the LV transmural pressure (PAWP - RAP) (R = 0.87) (Figure) . Patients admitted for RV failure (n=6) within 60 days of the ramp study did not have significantly different RAP respiratory variation at baseline compared to those not admitted for RV failure (n=6) (27% versus 40%, p = 0.09). However, there was a difference in RAP respiratory variation at the peak speed of the ramp study (36% RAP respiratory variation in RV failure patients compared to 61% for non-RV failure patients, p = 0.045). Conclusions The degree of PAWP respiratory variation is correlated with the ability to unload the LV by increasing LVAD speed. PAWP respiratory variation is a marker of a favorable hemodynamic response to speed changes during ramp study. In addition, patients with reduced RAP respiratory variation were more likely to have RV failure requiring hospital admission. Right and left-sided filling pressure respiratory variation patterns can provide valuable clinical information to the treating physician regarding the LVAD patient's underlying biventricular physiology.