Introduction: The Nexfin device estimates arterial pressure by the volume clamp method through a finger pneumatic cuff. It also allows to estimate cardiac index (CInoninv) by pulse contour analysis of the non-invasive arterial pressure curve. We evaluated the ability of the device to track changes in cardiac index induced by a fluid challenge.Methods: We included 45 patients for whom a volume expansion (500 mL of saline infused over 30 min) was planned. The volume expansion-induced changes in cardiac index measured by transpulmonary thermodilution (CIinv, PiCCO device) and in CInoninv were recorded.Results: In seven patients, the Nexfin could not record the arterial curve due to finger hypoperfusion. Considering both the values obtained before and after volume expansion (n = 76 pairs of measurements), the bias (lower and upper limits of agreement) between CIinv and CInoninv was 0.2 (-1.8 to 2.2) L/min/m2. The mean change in CInoninv was 10 ± 11%. The percentage error of CInoninv was 57%. The correlation between the changes in CIinv and CInoninv observed during volume expansion was significant (P = 0.0002) with an r2 = 0.31.Conclusions: The estimation of CI by the Nexfin device in critically ill patients is not reliable, neither for estimating absolute values of CI nor for tracking its changes during volume expansion.