NICOM vs LiDCO™ plus during changes in cardiac output in critically ill patients

Introduction

NICOM (Cheetah) is a new non-invasive cardiac output (CO) monitor based on bioreactance. We tested the agreement of NICOM vs LiDCO™ plus in terms of absolute CO values and of the ability to track changes after therapeutic intervention in critically ill patients.

Methods

Patients requiring haemodynamic monitoring were monitored with LiDCO™ plus and NICOM. Three measurements of CO by LiDCO were used to measure CO at baseline and to calibrate PulseCO. After fluid challenges (FC = 250 ml colloid/5 minutes) or inotropic therapy alterations, the change in stroke volume (SV) for PulseCO and NICOM CO was recorded. Patients able to increase SV ≥10% as measured by PulseCO were considered responders (R). Bland-Altman analysis was performed for NICOM vs LiDCO CO at baseline. The coefficient of variation (CV) and percentage error (PE) were calculated. ΔSV of NICOM vs PulseCO was analysed with correlation analysis. In patients receiving FC, a ROC analysis was performed to detect the sensitivity and specificity of NICOM to track ΔSV as measured by PulseCO CO.

Results

Baseline haemodynamics in 30 patients enrolled: BA analysis for NICOM vs LiDCO CO showed a mean bias of -0.18 l/minute (LOA -2.5 to 2.16). Mean CO was 5.8 l/minute, PE was 41%. Mean LiDCO CO CV was 6.8%. A total of 81 pairs of data were collected after FC or inotrope dose change. Correlation analysis showed r² of 0.55 (P < 0.0001) for changes post FC and r² of 0.54 (P < 0.006) post inotrope dose changes. ROC analysis for NICOM CO in R and NR after a FC showed an area under the curve (AUC) of 0.8 (P < 0.0001; Figure 1). An increase in NICOM SV >8% showed a sensitivity of 76% and a specificity of 75% to predict PulseCO changes >10%.

Figure 1

ROC analysis for NICOM CO in R and NR.

Conclusions

NICOM demonstrated a moderate agreement with LiDCO but showed excellent agreement with PulseCO in tracking CO changes following therapeutic interventions.