Volumetric preload measurement by thermodilution: a comparison with transoesophageal echocardiography

End-diastolic volume indices determined by transpulmonary thermodilution and pulmonary artery thermodilution may give a better estimate of left ventricular preload than pulmonary capillary wedge pressure monitoring. The aim of this study was to compare volume preload monitoring using the two different thermodilution techniques with left ventricular preload assessment by transoesophageal echocardiography (TOE).

Twenty patients undergoing elective cardiac surgery with preserved left-right ventricular function were studied after induction of anaesthesia. Conventional haemodynamic variables, global end-diastolic volume index using the pulse contour cardiac output (PiCCO) system (GEDVI(PiCCO)), continuous end-diastolic volume index (CEDVI(PAC)) measured by a modified pulmonary artery catheter (PAC), left ventricular end-diastolic area index (LVEDAI) using TOE and stroke volume indices (SVI) were recorded before and 20 and 40 min after fluid replacement therapy. Analysis of variance (Bonferroni-Dunn), Bland-Altman analysis and linear regression were performed.

GEDVI(PiCCO), CEDVI(PAC), LVEDAI and SVI(PiCCO/PAC) increased significantly after fluid load (P<0.05). An increase >10% for GEDVI(PiCCO) and LVEDAI was observed in 85% and 90% of the patients compared with 45% for CEDVI(PAC). Mean bias (2 SD) between percentage changes (delta) in GEDVI(PiCCO) and deltaLVEDAI was -3.2 (17.6)% and between deltaCEDVI(PAC) and deltaLVEDAI -8.7 (30.0)%. The correlation coefficient (r2) for deltaGEDVI(PiCCO) vs deltaLVEDAI was 0.658 and for deltaCEDVI(PAC) vs deltaLVEDAI 0.161. The relationship between deltaGEDVI(PiCCO) and deltaSVI(PiCCO) was stronger (r2=0.576) than that between deltaCEDVI(PAC) and deltaSVI(PAC) (r2=0.267).

GEDVI assessed by the PiCCO system gives a better reflection of echocardiographic changes in left ventricular preload, in response to fluid replacement therapy, than CEDVI measured by a modified PAC.