Predicted versus measured resting energy expenditure in patients requiring home parenteral nutrition

Guidelines from the European Society for Clinical Nutrition and Metabolism (ESPEN) recommend between 20 and 35 kcal/kg daily for patients requiring home parenteral nutrition (PN). Other guidelines use predictive equations. However, these equations have not been validated. Indirect calorimetry is recommended as the gold standard for determining resting energy expenditure (REE). The aim of this study was to compare the frequently used equations with measured REE.

Seventy-six hospitalized patients suffering from intestinal failure (ages 21-85 y) were enrolled between January 2012 and May 2014. They were eligible for implementation of home parenteral nutrition (HPN) due to short bowel syndrome (54%), intestinal fistulae (24%), cancer obstruction (16%), and radiation-induced intestinal injury (6%). REE measurements were compared with predictive equations by Harris and Benedict (HB), Owen, Ireton-Jones, and Mifflin, as well as recommendations from ESPEN.

In all, 152 calorimetry measurements (two per patient) were performed in 76 patients, after total PN administrations. An average result of REE measurement by indirect calorimetry was 1181 ± 322 kcal/d. Variability in momentary energy expenditure (MEE) from one measurement to the other was 8% ± 7%. Bland-Altman analysis showed a mean bias of -192 ± 300 kcal/d between MEE and estimated energy expenditure using the HB equation, which means that the equation increased the score on average by 192 ± 300 kcal/d. Limits of agreement (LoA) between the two methods was -780 to +396 kcal/d. Estimation energy expenditure using the Ireton-Jones equation gave a mean bias of -359 ± 335 kcal/d. LoA between the two methods was -1015 to +297 kcal/d. For Owen equation, Bland-Altman analysis showed a mean bias of -208 ± 313 kcal/d and the LoA between the two methods was -822 to +406 kcal/d. Using the Mifflin equation, estimation energy expenditure gave a mean bias of -172 ± 312 kcal/d and the LoA between the two methods was -784 to +439 kcal/d. Using the ESPEN range (20-35 kcal/kg daily) analysis showed mean bias of -13 ± 326 kcal/d and the LoA was -652 to +626 kcal/d for 20 kcal/kg daily and mean bias of -909 ± 436 kcal/d with the LoA between the two methods -1764 to -54 kcal/d for 35 kcal/kg daily.

If REE cannot be measured by indirect calorimetry in patients qualified for HPN, the Ireton-Jones equation and the 20 kcal/kg/d ESPEN recommendation seem to be the most appropriate ones as it provides results that constitute the best approximation of calorimetric examination results.