Predictive Energy Equations: Which One to Use?

Dietetic interns challenge us to keep our practice current. The question of what predictive energy equation to use in clinical practice was a recent example of how our roles as educators linked to our need to keep current. Fraser Health (FH) interns recognized their preceptors had different approaches to calculating estimated energy needs for their patients and wondered how something so core to dietitians’ daily activities could be so varied in practice. A poll of Canadian dietitians and a literature review provided perspectives on the use of predictive energy equations.

Indirect calorimetry (IC) is the ‘gold standard’ for calculating energy requirements. Several barriers prevent routine measurements on patients including lack of a metabolic cart or personnel to operate equipment, and the impracticality of completing repeated measures on every patient who requires an energy calculation. Consequently, dietitians rely on predictive energy equations to estimate patients’ energy needs.

Over 200 predictive energy equations have been published in efforts to accurately predict energy needs (MacDonald and Hildebrandt, 2003; Ireton- Jones, 2005). No equation has consistently proven to be accurate in hospitalized individuals in acute or critical care. Inaccuracies are introduced due to controversy over which weight to use, the limited number of validation studies done, and the homogeneity of reference populations used when establishing the formulae, amongst other variables (Frankenfield et al., 2005; Fabiono et al., 2009; Walker and Heuberger, 2009).

An e-mail poll in October 2009 to members of the Dietitians of Canada (DC) Clinical Nutrition Managers’ Network yielded a list of the equations used by at least one dietitian at each of the 16 responding sites: 13 sites used Harris Benedict (HB); eight used the American College of Chest Physicians’ equation (25kcal/kg); six used Mifflin-St Jeor; five used one of the Ireton-Jones’ equations; and two used one of the Penn State equations. These results were consistent with practices worldwide (HB remains the most frequently-used formula (Fabiono et al., 2009)), and the varied practices that the FH dietetic interns had noticed.

The e-mail poll also yielded information on which weight (actual, ideal, or adjusted) is used in energy calculations. Four sites used actual weight only and 12 sites used either actual or adjusted weight depending on the individual case. Most sites adjusted body weight using a 25% factor (based on a 1984 American Dietetic Association publication as referenced in Krenitsky, 2005). Other sources recommended a 50% factor (Barack et al., 2002; Krenitsky, 2005). The use of adjusted body weight is not well supported (Barack et al., 2002; Ireton-Jones, 2005; Krenitsky, 2005).

The following points summarize the support for the predictive energy equations with graded evidence-for-use in clinical practice:

Harris-Benedict (Harris and Benedict, 1919): Not sufficiently accurate to be used in the critically ill (Grade I evidence) (Frankenfield et al., 2007) Use actual body weight (Walker and Heuberger, 2004; Krenitsky, 2005).

American College of Chest Physicians (Cerra et al., 1997): Use 25-30 kcal/kg actual body weight for the nonobese critically ill (Grade IV and V evidence). Use 11-14 kcal/kg actual body weight or 22-25 kcal/kg ideal body weight for the obese critically ill (Grade III evidence) (McClave et al., 2009).

Mifflin-St Jeor (Mifflin et al., 1990): Best predictor of energy needs in healthy nonobese and obese adults (Frankenfield et al., 2005). Use actual body weight (McClave et al., 2009). Not recommended in the critically ill (Grade V evidence) (Frankenfield et al., 2007).

Ireton Jones et al. (1992): Most accurate in young and obese adults compared to other populations. Insufficient data to reject the equation (Grade III evidence supporting its use) (Frankenfield et al., 2007; Walker and Heuberger, 2009)

Penn State 1998 and 2003 (Frankenfield et al., 2004): May use 1998 equation for obese, critically ill ventilated patients (Frankenfield et al., 2005). May use 2003 equation for nonobese critically ill patients (Walker and Heuberger, 2009) (Grade III evidence) (Frankenfield, et al., 2007; Walker and Heuberger, 2009).

Swinamer (Swinamer et al., 1987): May use for nonobese critically ill patients (Grade III evidence) (Frankenfield, et al, 2007; Walker and Heuberger, 2009).

Even with an accurate predictive equation, obtaining goal intakes is often hindered by delays in establishing feeding access, enteral feed intolerances, and feedings that are held for tests. Our efforts should go toward achieving goal tubefeeding rates and maximizing oral intake regardless of what equation is used. Further, monitoring markers of feeding adequacy including laboratory measures, weight, ventilator weaning, and wound healing allow us to adjust energy goals as appropriate.

Based on the results of our survey and literature review, we can have confidence in reporting to our interns that there is more to learn before setting standards for estimating energy needs in hospital inpatients. We use these teaching moments as opportunities to discuss the concept of evidence-based practice using best available research, clinical expertise, and a patient-focused approach to guide decision-making to provide the best possible patient care.

A summary chart of the origins and evidence on predictive energy equations is available on request.

Contact

Tamar Kafka, RD, MSc
Dietetic Internship Coordinator/Research Dietitian
Fraser Health
New Westminster BC
E: Tamar.Kafka@fraserhealth.ca