Comparison between invasive cardiac output and left ventricular assist device flow parameter

Abstract

OBJECTIVES Evaluate the correlation between left ventricular assist device flow parameter and invasive cardiac output measurements. METHODS We retrospectively evaluated right heart catheterization examinations performed in left ventricular assist device patients from 2 tertiary medical centers. We evaluated the correlation between cardiac output measurement methods (indirect Fick and thermodilution) and pump flow parameter using linear regression, agreement was graphically displayed using Bland-Altman plot technique. Clinical, echocardiographic, pump and haemodynamic parameters were compared between patients with and without discordance, defined as at least 20% difference between measurements. RESULTS The study population consisted of 102 patients (median age 58 [51–64], 86% males, 17 ± 12 months post left ventricular assist device implantation) with a total of 544 measurement compared. Discordance between measurements were present in 102 of 226 (45%) comparisons between indirect Fick and pump flow and in 72 of 161 (48%) between thermodilution and pump flow. A comparison of indirect Fick and left ventricular assist device exhibited a statistical correlation of R = 0.751, and that of thermodilution and left ventricular assist device of R = 0.789. Parameters associated with the presence of discordance between cardiac output measurements included a higher rate of aortic valve opening, lower indirect Fick and higher thermodilution cardiac output. After excluding the lowest tertile of indirect Fick cardiac output values, the correlation between measurements improved (thermodilution: R = 0.879 and indirect Fick: R = 0.843, p < 0.001). CONCLUSIONS The current left ventricular assist device flow parameter provides an estimation of cardiac output that correlates well with indirect Fick and exhibits the strongest correlation with thermodilution. This correlation was stronger after excluding lower cardiac output values.