Correlation between measured energy expenditure and clinically obtained variables in trauma and sepsis patients.

Background: Indirect calorimetry is the preferred method for determining caloric requirements of patients, but availability of the device is limited by high cost. A study was therefore conducted to determine whether clinically obtainable variables could be used to predict metabolic rate.
Methods: Patients with severe trauma or sepsis who required mechanical ventilation were measured by an open-circuit indirect calorimeter. Several clinical variables were obtained simultaneously. Measurements were repeated every 12 hours for up to 10 days.
Results: Twenty-six trauma and 30 sepsis patients were measured 423 times. Mean resting energy expenditure was 36 +/- 7 kcal/kg (trauma) vs 45 +/- 8 kcal/kg (sepsis) (p < .0001). The single strongest correlate with resting energy expenditure was minute ventilation (R2 = 0.61, p < .0001). Doses of dopamine, dobutamine, morphine, fentanyl, and neuromuscular blocking agents each correlated positively with resting energy expenditure. In the case of the inotropics and neuromuscular blockers, there was a probable covariance with severity of illness. A multiple regression equation was developed using minute ventilation, predicted basal energy expenditure, and the presence or absence of sepsis: resting energy expenditure = -11000 + minute ventilation (100) + basal energy expenditure (1.5) + dobutamine dose (40) + body temperature (250) + diagnosis of sepsis (300) (R2 = 0.77, p < .0001).
Conclusion: Severe trauma and sepsis patients are hypermetabolic, but energy expenditure is predictable from clinical data. The regression equations probably apply only to severe trauma and sepsis. Other studies should be conducted to predict energy expenditure in other patient types.