To Compress or Not to Compress: Revisiting the Use of Chest Compressions in LVAD Patients

Abstract

Purpose: Repeated invasive cardiac output testing in patients supported by a ventricular assist device (VAD) with right heart catheterization carries an increased risk of pump thrombosis during times of reduced anticoagulation, and come at considerable healthcare costs. There are inherent flaws in measuring cardiac output by fick and thermal method. Insufficient data exists on non-invasive assessment of cardiac output in these patients. We hypothesize that in clinically stable patients, flow on axial VAD consoles and Doppler derived cardiac outputs (measured through right ventricular outflow tract on transthoracic echocardiograph) can be used to estimate the cardiac output. Methods: We retrospectively assessed right heart catheterization results to obtain fick and thermal cardiac outputs of stable patients supported by the axial continuous-flow VAD (HeartMate II, Thoratec Inc) between April 2011 and September 2014. At the time of their right heart catheterization, flow estimations were obtained from their VAD console (VF). Correlations coefficients (r) were obtained to assess relationships between VF and cardiac outputs by fick and thermal method. Additionally, echocardiograms at the time of right heart catheterization were assessed, and Doppler derived cardiac outputs measured through the right ventricular outflow tract (rCO) and assessed for correlation with fick and thermal cardiac outputs. Results: Forty-two subjects were assessed with mean VAD speed 9081+/-181 rpm. Fick cardiac output (5.13+/-1.4 L/min) and VF (5.31+/-0.9 L/min) had a correlated well; r of 0.43 (p= 0.0048). In 35 subjects, thermal cardiac outputs (5.19+/-1.5 L/min) were available and also correlated with VF (5.42+/-0.96 L/min, r= 0.43, p= 0.0097). No correlation was found between 27 assessments of rCO (3.83+/-1.6 L/min, r= 0.20, p= 0.303) and fick cardiac output (5.28+/-1.4 L/min). Likewise, there was no significant correlation between 22 values of rCO (4.10+/-1.6 L/min, r= 0.13, p= 0.564) and thermal cardiac output (5.28+/-1.3 L/min). Conclusion: Flow estimates from the HeartMate II console correlate with cardiac outputs derived invasively by the fick and thermal methods in stable patients without concern for pump thrombosis. Doppler derived cardiac outputs from the right ventricular outflow tract were inaccurate in patients on VAD support.