Purpose: To better define the reliability of left ventricular ejection fraction (LVEF) and left ventricular filling, as determined by either hand-carried ultrasound (HCU) or formal transthoracic echocardiography (TTE), in the critically ill surgical patient.
Materials and methods: Prospective cross-sectional study of 80 surgical intensive care unit patients with concomitant (<30 minutes apart) formal TTE and clinician-performed cardiac HCU. Visual estimates of LVEF and left ventricular filling ("underfilled" vs "normally filled") were recorded, both by clinicians performing HCU and fellowship-trained echocardiographers.
Results: Bland-Altman plot analysis of LVEF estimates revealed good interobserver agreement between HCU and formal TTE (% LVEF mean bias, -2.2; with 95% limits of agreement, +/-22.1). This was similar to agreement between independent echocardiography observers (% LVEF mean bias, 1.3; with 95% limits of agreement, +/-21.0). However, assessments of left ventricular filling demonstrated only fair to moderate interobserver agreement (kappa = 0.22-0.40). Of note, a greater percentage of the 5 standard acoustic windows were obtainable using formal TTE (72% vs 56%).
Conclusions: Formal TTE offers no advantage over HCU for determination of LVEF in critically ill surgical patients, even though the former allows for a more complete examination. However, estimations of left ventricular filling only demonstrate fair to moderate interrater agreement and thus should be interpreted with care when used as markers of volume responsiveness.