Predictive equations and energy expenditure: Is a guess good enough?

Abstract

DITURE: IS A GUESS GOOD ENOUGH? Suzie Ferrie Critical Care Dietitian, Royal Prince Alfred Hospital. Lacking access to indirect calorimetry, many clinicians need to be able to predict a patient’s energy requirements; equations are available that provide a quick and non-invasive way to do so, and these can be a valuable addition to the clinician’s tool box. However, as with any tool, the results will be poor if the user fails to understand how the tool works, or lacks skill in using it properly. The equations do have limitations! Predictive equations cannot transform incorrect or inaccurate data into a useful result, they cannot replace other essential components of patient assessment (such as physical examination), and they are no substitute for ongoing close monitoring of the nutrition support patient. Understanding the origins of the predictive equations can help in using them more appropriately. The number and type of subjects used in developing the equations may indicate the most appropriate population in which they can be used, and suggests some weaknesses of the equations. For example, the Harris-Benedict equations were developed from a group of young (average age 29) and lean (average BMI 21) Americans at the beginning of the 20 century when lifestyle and diet were very different from ours (1). The Schofield equations used more than 7000 people from 23 different countries, with a wide range of ages and weights but with a preponderance of men; a significant number (more than 10% of the total) were very fit military personnel (2,3). The Mifflin-StJeor equations were based on a group of contemporary Americans (50% were obese!) (4) The Ireton-Jones equations, notably, were developed using sick hospital patients, but may be quite rigid in their application to some diagnostic groups (5). It is important to note the assumptions underlying each equation, and the reasons for their format (the existence of different equations for men and women, for example, or separating by age group). It may be appropriate to use an age value other than the patient’s actual chronological age, or to use an adjusted weight value in the equation. The use of adjustment factors (such as ‘stress’ or ‘activity’ factors) (6,7) may allow a hospital patient’s needs to be predicted from an equation developed in healthy people – but there are many pitfalls here, and an enormous potential for increasing the errors inherent in this method. There are other considerations too: to avoid overfeeding, it is important to ensure that the predicted requirement is viewed as the patient’s total needs (rather than distinguishing between protein and non-protein energy). Additionally, expressing the result as a range (rather than a fixed value) makes it clear that it is an estimate and avoids implying an unrealistic level of accuracy, which can undermine the credibility of the clinician. References