Reliability of pulse contour method-based hemodynamic measurements assessed by different operators.

INTRODUCTION. Pulse Contour Methods (PCMs) provide stroke volume (SV) and cardiac output (CO) from the analysis of the arterial waveform. Artifacts affecting arterial wave morphology may reduce the reliability of PCMs in estimating such haemodynamic variables. The new operator who is facing a PCM for the first time should need a training period to gain ability in recognizing artifacts and avoiding mistakes in CO assessment. OBJECTIVES. The aim of this study is to evaluate the potential bias rising when com- paring CO values obtained by transthoracic echocardiography (TTE-CO) and Most-Care (MC-CO) system recorded by trained (T) and not-trained operators (NT). METHODS. 23 consecutive patients (10 male, mean age 60 ± 15) admitted to general ICU for heterogeneous pathologies were enrolled. Inclusion criteria were: the presence of a radial artery catheter for invasive blood pressure monitoring, need for a transthoracic echocardi- ography evaluation, age [18 years old. Exclusion criteria were: cardiac arrhythmias, aortic valve pathologies, ascending aortic pathologies, pregnancy, arteriovenous fistulas, and significant artery obstruction. For each patient, during haemodynamic stability (i.e., mean arterial pressure changes\10 %), CO was obtained with Most-Care (Vytech Health, Padua, Italy) by T (i.e., who has been using Most-Care for at least 6 months) and NT operator (who had only read the user manual of MostCare). TTE-CO was performed with MyLabTM 70 Xvision (Esaote, Italy) by the same operator during CO measurements obtained by T and NT (T-CO, NT-CO). TTE-CO was calculated as the product of heart rate by SV averaged over five consecutive measurements obtained during both NT-CO and T-CO Most-Care estimations. Bland–Altman analysis was used. RESULTS. 46 paired CO values were obtained. TTE-CO values ranged from 3.0 to 8.9 l/ min, T-CO ranged from 3.2 to 9.0 l/min, and NT-CO from 4.5 to 16.0 l/min. The mean bias between T-CO and TTE-CO was 0.04 l/min (limits of agreement, LoA -0.5 to 0.6 l/min, percentage error, PE = 11 %), and between NT-CO and TTE-CO -2.8 l/min (LoA -9.8 to 4.5 l/min, PE [ 30 %). The fast flush test, to improve the quality of the arterial pressure waveform, was performed in 100 % of cases by T and in 48 % of cases by NT. Resonance over-shoot eliminator (R.O.S.E., Becton–Dickinson, Becton Drive, NJ) was used in 32 % of cases by T and in 9 % of cases by NT. CONCLUSIONS. The comparison between T-CO and NT-CO showed that bias and per- centage of errors were very relevant for NT. A period of training is needed for the new user to assess correctly CO values with such PCM. This would avoid misinterpretation of arterial pressure waveform-derived variables and could help the clinical staff to get reliable hemodynamic data for daily clinical practice.