Continuous thermodilution, but not indirect Fick-derived cardiac index, independently predicts mortality in 1045 cardiac ICU patients

Background

Cardiac index (CI) guides haemodynamic management in critically ill patients with cardiogenic shock or advanced cardiac failure. Continuous thermodilution and indirect Fick-derived CI are frequently used interchangeably, although their relative prognostic value is uncertain.

Purpose

To compare the prognostic performance of continuous thermodilution and indirect Fick-derived CI for ITU mortality.

Methods

We retrospectively analysed 1,045 consecutive critically ill cardiac patients admitted to a quaternary ITU undergoing invasive hemodynamic monitoring (mean age 58.9±16.2 years; 67% male; 65% on inotropes; mortality 24.9%). CI was assessed simultaneously by continuous thermodilution and indirect Fick method using estimated oxygen consumption (VO2). Multivariable models included both CI measurements and were adjusted for age, sex, inotrope use, mechanical ventilation, and renal ultrafiltration.

Results

Agreement analysis showed minimal mean bias (0.11 L/min/m²) but wide limits (−1.83 to +2.07) and percent error of 64%, well above accepted interchangeability thresholds. Both CI measurements demonstrated strong physiological coherence, showing robust associations with SvO2 (R²=0.60, p<0.001).

In univariable analysis, continuous thermodilution predicted ITU mortality (OR 0.63, 95% CI 0.52–0.76, p<0.001), whereas indirect Fick-derived CI did not (OR 1.13, p=0.156). Continuous thermodilution showed moderate discrimination (AUC 0.71), while indirect Fick-derived CI showed poor discrimination (AUC <0.6).

In multivariable analysis including both measurements and clinical confounders, only continuous thermodilution retained independent prognostic significance. Prognostic value was stronger in patients receiving inotropes (OR 0.59). A CI threshold of 2.35 L/min/m² identified high-risk patients for mortality with high specificity (81.5%).

Conclusions

Continuous thermodilution and indirect Fick-derived CI are not clinically interchangeable. Although both methods are positively associated with SvO2, only continuous thermodilution provides independent prognostic information for mortality, supporting its preferential use for invasive risk stratification in high-risk critically ill cardiac ITU patients.