Measured versus estimated energy needs in critically ill patients with a suspected infection: A feasibility study.

BACKGROUND

Energy needs of critically ill patients fluctuate throughout their intensive care unit (ICU) admission. The relationship between energy expenditure and infection and sepsis is unclear. Indirect calorimetry (IC) guides energy delivery and may help understand disease states and correlations between energy requirements, infection and clinical markers.

METHODS

We compared measured energy expenditure (mEE) using once daily IC with usual predictive equations in critically ill patients with suspected infection. Correlations between mEE and markers of infection and illness severity (C-reactive protein (CRP), lactate, vasopressors, and Sequential Organ Failure Assessment (SOFA) score) were described. Regression models were used for analysis.

RESULTS

Seventy valid IC tests from 46 patients showed 20 kcal/kg and Penn State University (PSU) equation as the most accurate (1520 kcal and 1749 kcal versus median mEE 1626 kcal), although no predictive equations were within 10% of mEE. No significant differences in median mEE were found over time (1711 kcal in the early acute, 1542 kcal in the late acute, and 1700 kcal in the chronic late phases of critical illness). Moderate correlations were found between mEE and lactate level (early acute) (rho -0.572, p = 0.013) and vasopressor dose (late acute) (rho -0.428, p = 0.004). A strong correlation was observed with CRP level (chronic late) (rho -0.851, p = 0.007).

CONCLUSION

Energy targets are best determined with IC due to wide variation between predictive estimates and mEE. Further research with a larger longitudinal sample is needed to better understand energy changes with markers of infection and illness severity, and optimal timing and frequency of IC.