Quick Links

Recommendations Summary

CI: Determination of Resting Metabolic Rate (RMR) 2006

Click here to see the explanation of recommendation ratings (Strong, Fair, Weak, Consensus, Insufficient Evidence) and labels (Imperative or Conditional). To see more detail on the evidence from which the following recommendations were drawn, use the hyperlinks in the Supporting Evidence Section below.


  • Recommendation(s)

    CI: Indirect Calorimetry to determine RMR

    Indirect calorimetry is the standard for determination of RMR in critically ill patients since RMR based on measurement is more accurate than estimation using predictive equations.

    Rating: Strong
    Imperative

    CI: RMR Predictive equations for non-obese patients

    If predictive equations are needed in non-obese, critically ill patients, consider using one of the following, as they have the best prediction accuracy of equations studied (listed in order of accuracy): Penn State, 2003a (79%), Swinamer (55%) and Ireton-Jones, 1992 (52%). In some individuals, errors between predicted and actual energy needs will result in under- or over-feeding.

    Rating: Fair
    Conditional

    CI: Inappropriate RMR Predictive equations for this population

    The Harris-Benedict (with or without activity and stress factors), the Ireton-Jones, 1997 and the Fick equation should not be considered for use in RMR determination in critically ill patients, as these equations do not have adequate prediction accuracy. In addition, the Mifflin-St. Jeor equation should not be considered for use in critically ill patients, as it was developed for healthy people and has not been well researched in the critically ill population.

    Rating: Strong
    Imperative

    CI: RMR Predictive Equations for obese patients

    If predictive equations are needed for critically-ill, mechanically-ventilated individuals who are obese, consider using Ireton-Jones, 1992 or Penn State, 1998, as they have the best prediction accuracy of equations studied. In some individuals, errors between predicted and actual energy needs will result in under- or over-feeding.

    Rating: Fair
    Conditional

    • Risks/Harms of Implementing This Recommendation

      • Anxiety may be caused by indirect calorimetry procedures employing a face mask or canopy
      • In some individuals, estimation of RMR with predictive equations will lead to under- or over-feeding.

    • Conditions of Application

      Certain predictive equations were designed for application in mechanically-ventilated patients.

      The AARC Clinical Practice Guidelines (1994) recommend that measurements may be indicated in patients with the following conditions:

      • Neuro trauma
      • Paralysis
      • COPD
      • Acute pancreatitis
      • Cancer with residual tumor
      • Multiple trauma
      • Amputations
      • Patients with no accurate height or weight
      • Long term acute care (ventilator units)
      • Severe sepsis
      • Extreme obesity
      • Severely hypermetabolic or hypometabolic patients
      • Failure to wean.

      The AARC Clinical Practice Guidelines (1994) also provide recommendations for hazards and complications, limitations of the procedures and infection control.

      Hazards and Complications

      • Short-term disconnection of patient from ventilator for connection to an indirect calorimetry machine may result in hypoxemia, bradycardia and patient discomfort
      • Inappropriate calibration or system setup may result in erroneous results, causing incorrect patient management
      • Isolation valves in calorimeters may increase circuit resistance and cause increased work of breathing or dynamic hyperinflation
      • Inspiratory reservoirs may cause reduction in alveolar ventilation, due to increased compressible volume of the breathing circuit
      • Manipulation of the vent circuit may cause leaks that may lower alveolar ventilation.

      Limitations of the Procedure

      • Leaks in ventilator circuit, endotracheal tube cuffs or uncuffed tubes, through chest tubes or bronchopleural fistula
      • Peritoneal and hemo-dialysis procedures remove CO2 during the treatment and require a few hours after the treatment for acid-base to stabilize. Patients should not be measured during or for four hours after these dialysis treatments.
      • Inaccurate measures may be caused by:
        • Unstable O2 delivery, due to vent blender or mixing characteristics
        • FIO2 above 60%
        • Inability to separate inspired from expired gases, due to bias flow with intermittent mandatory ventilation systems
        • Anesthetic gases other than O2, CO2 and nitrogen in the system
        • Water vapor presence
        • Inappropriate calibration
        • Total circuit flow exceeding internal gas flow of calorimeter
        • Leaks within the calorimeter
        • Inadequate measurement length.

      Measures should be done by personnel trained in and with demonstrated and documented ability to calibrate, operate and maintain the calorimeter, having a general understanding of how mechanical ventilation works and recognizing calorimeter values within the normal physiologic range.

      More frequent measures may be needed in patients with rapidly changing clinical course, as recognized by hemodynamic instability, spiking fevers, immediate postoperative status and ventilator weaning.

      Infection Control

      • Use standard precautions for contamination of blood and bodily fluids
      • Appropriate use of barriers and handwashing
      • Tubing to connect expired air from ventilator to indirect calorimetry should be disposed of or cleaned between patients
      • Connections in the inspiratory limb of the circuit should be wiped clean between patients and equipment distal to the humidifier should be disposed of
      • Bacteria filters may be used to protect equipment in inspire and expired lines.

    • Potential Costs Associated with Application

      • Cost of equipment, supplies and staff needs to be addressed in all indirect calorimetry measurements
        • For patients who require mechanical ventilation, the cost of portable indirect calorimeters may be up to $35, 000 and the cost of tubing used to connect with the ventilator for gas collection varies
        • The cost of trained staff to run the tests and maintain equipment can be considerable, since each test may require one hour of staff time
        • Insurance companies may pay a technician fee for running the test and a professional fee if a licensed medical professional (e.g., MD) interprets the test.
      • A calculator is required for equation calculation.

    • Recommendation Narrative

      Recommendation: Indirect Calorimetry for Determination of RMR

      • Indirect calorimetry is the standard for determination of RMR in critically ill patients
      • When indirect calorimetry cannot be performed, predictive formulas may be necessary
        • 14 positive-quality cross-sectional studies: Casati et al, 1996; Donaldson-Andersen et al, 1998; Ogawa et al, 1998; Brandi et al, 1999; Flancbaum et al, 1999; Epstein et al, 2000; Cheng et al, 2002; Dickerson et al, 2002; Faisy et al, 2003; MacDonald and Hildebrandt, 2003; Marson et al, 2003; Alexander et al, 2004; Frankenfield et al, 2004; O'Leary-Kelley et al, 2005
        • Eight neutral-quality cross-sectional studies: Cutts et al, 1997; Ireton-Jones, 1997; Ahmad et al, 1999; Brandi et al, 1999; Glynn et al, 1999; Ireton-Jones and Jones, 2002; Jansen et al, 2002; Campbell et al, 2005
        • Two neutral-quality cohort studies: Barak et al, 2002; Compher et al, 2004
        • Grade I.

      Recommendation: Predictive Equations with Best Prediction Accuracy

      • Seven cross-sectional studies comparing RMR and the Ireton-Jones, 1992, equations report similar mean values, however, for an individual, energy predictions may be different by as much as 500 kcal (52% of non-obese subjects predicted within 10% of RMR)
        • Five of positive quality: Flancbaum et al, 1999; Cheng et al, 2002; Dickerson et al, 2002; MacDonald and Hildebrandt, 2003; Frankenfield et al, 2004
        • Two of neutral quality: Ireton-Jones, 1997; Campbell et al, 2005.
        • Further research regarding the accuracy of the Ireton-Jones, 1992 equation is warranted
        • Grade III.
      • Two positive quality cross-sectional studies comparing RMR and the Penn State equation report adequate precision (79% of non-obese subjects predicted within 10% of RMR): MacDonald and Hildebrandt, 2003; Frankenfield et al, 2004
        • Further research in the critically ill population is needed regarding the Penn State equation
        • Grade III.
      • One positive-quality cross-sectional study comparing RMR and the Swinamer equation, reported that 55% of non-obese subjects were predicted within 10% of RMR: MacDonald and Hildebrandt, 2003
        • Further research in the critically ill population is needed regarding the Swinamer equation
        • Grade III.

      Recommendation: Predictive Equations with Inadequate Prediction Accuracy

      • 13 studies comparing RMR and the Harris-Benedict equation (without adjustments) generally report an underestimation of energy needs in the critically ill population by as much as 1, 000kcal or more
        • Nine positive-quality cross-sectional studies: Donaldson-Andersen et al, 1998; Brandi, Santini et al, 1999; Flancbaum et al, 1999; Cheng et al, 2002; Dickerson et al, 2002; Faisy et al, 2003; MacDonald and Hildebrandt, 2003; Alexander et al, 2004; Frankenfield et al, 2004
        • Three neutral-quality cross-sectional studies: Ireton-Jones, 1997; Ahmad et al, 1999; Campbell et al, 2005
        • One neutral-quality cohort study: Compher et al, 2004
        • Grade I.
      • 13 studies comparing RMR and the Harris-Benedict equation with stress and activity factors ranging from 1.1 to 1.6 may be biased or imprecise by 900kcal or more, depending on the factors used
        • Cross-sectional studies include nine of positive quality: Casati et al, 1996; Donaldson-Andersen et al, 1998; Brandi and Santini et al, 1999; Cheng et al, 2002; Dickerson et al, 2002; Faisy et al, 2003; MacDonald and Hildebrandt, 2003; Alexander et al, 2004; O'Leary-Kelley et al, 2005
        • Four cross-sectional studies of neutral quality: Cutts et al, 1997; Ireton-Jones and Jones, 2002; Jansen et al, 2002
        • One neutral-quality retrospective cohort study: Barak et al, 2002
        • Grade I.
      • Five cross-sectional studies comparing RMR and the Fick equation generally report little agreement between methods
        • Four of positive quality: Ogawa et al, 1998; Flancbaum et al, 1999; Epstein et al, 2000; Marson et al, 2003; one neutral quality: Brandi, Santini et al, 1999
        • Grade I.
      • Three cross-sectional studies comparing RMR and the updated Ireton-Jones, 1997, equations report similar mean values. However, only 36% of subjects were predicted within 10% of RMR
        • Two of positive quality: Alexander et al, 2004; Frankenfield et al, 2004
        • One of neutral quality: Ireton-Jones and Jones, 2002
        • Further research in the critically ill population is needed regarding the Ireton-Jones, 1997, equations
        • Grade II.
      • At the current time, the Mifflin-St. Jeor equation has not been adequately researched in the critically ill population, according to one positive-quality cross-sectional study (Frankenfield et al, 2004) for descriptive purposes only
        • Further research in the critically ill population is needed regarding the Mifflin-St. Jeor equation
        • Grade V.

      Recommendation: Predictive Equations for Critically Ill Individuals with Obesity

      • Studies including critically ill individuals with obesity studied the application of several predictive equations and the use of actual or adjusted weight in six studies
        • One positive-quality cross-sectional study: Frankenfield et al, 2004
        • Four neutral-quality cross-sectional studies: Cutts et al, 1997; Ireton-Jones, 1997; Glynn et al, 1999; Ireton-Jones and Jones, 2002
        • One positive-quality retrospective cohort study: Barak et al, 2002.
      • One study reported that the Harris-Benedict equation, using actual weight multiplied by a factor of 1.2 (60% of subjects predicted within 10% of RMR) or using an adjusted weight multiplied by a factor of 1.3 (67% of subjects predicted within 10% of RMR), resulted in the most accurate predictions: Glynn et al, 1999
      • A second study reports that the Penn State, 2003a, equation predicts within 10% of RMR in 61% of subjects and the Ireton-Jones, 1992, equations predict within 10% of RMR in 72% of subjects: Frankenfield et al, 2004
        • Further research is needed in critically ill individuals with obesity
        • Grade III.

    • Recommendation Strength Rationale

      • Recommendation: Indirect Calorimetry for Determination of RMR
        • A number of studies were identified
        • Conclusion statement was Grade I.
      • Recommendation: Predictive Equations with Best Prediction Accuracy
        • A number of studies were identified, but the research was limited or inconclusive
        • Conclusion statement for Ireton-Jones, 1992, was Grade III
        • Conclusion statement for Penn State was Grade III
        • Conclusion statement for Swinamer was Grade III. 
      • Recommendation: Predictive Equations with Inadequate Prediction Accuracy
        • A number of studies were identified, but the research was limited or inconclusive
        • Conclusion statement for Fick was Grade I
        • Conclusion statement for Harris-Benedict was Grade I
        • Conclusion statement for Ireton-Jones, 1997, was Grade II
        • Conclusion statement for Mifflin-St. Jeor was Grade V.
      • Recommendation: Predictive Equations with Inadequate Prediction Accuracy
        • A number of studies were identified, but the research was limited or inconclusive
        • Conclusion statement was Grade III.

      Click here to link to the page that lists the Predictive Equations for Determining Resting Metabolic Rate.