Energy expenditure in mechanically ventilated patients: The weight of body weight!

(1)

Article

Reference

Energy expenditure in mechanically ventilated patients: The weight of body weight!

GRAF, Séverine, et al.

Abstract

Optimal nutritional care for intensive care unit (ICU) patients requires precise determination of energy expenditure (EE) to avoid deleterious under- or overfeeding. The reference method, indirect calorimetry (IC), is rarely accessible and inconstantly feasible. Various equations for predicting EE based on body weight (BW) are available. This study aims at determining the best prediction strategy unless IC is available.

GRAF, Séverine, et al . Energy expenditure in mechanically ventilated patients: The weight of body weight! Clinical Nutrition , 2017, vol. 36, no. 1, p. 224-228

DOI : 10.1016/j.clnu.2015.11.007 PMID : 26653566

Available at:

http://archive-ouverte.unige.ch/unige:87455

Disclaimer: layout of this document may differ from the published version.

(2)

Original article

Energy expenditure in mechanically ventilated patients: The weight of body weight!

S everine Graf

^a^,^b

, Claude Pichard

^a

, Laurence Genton

^a

, Taku Oshima

^a

, Claudia Paula Heidegger

^b^,^*

aClinical Nutrition Unit, Geneva University Hospitals, Switzerland

bDivision of Intensive Care, Geneva University Hospitals, Switzerland

a r t i c l e i n f o

Article history:

Received 21 September 2015 Accepted 6 November 2015

Keywords:

Indirect calorimetry Predictive equations Energy expenditure Body weight Intensive care unit

s u m m a r y

Background & aims:Optimal nutritional care for intensive care unit (ICU) patients requires precise determination of energy expenditure (EE) to avoid deleterious under- or overfeeding. The reference method, indirect calorimetry (IC), is rarely accessible and inconstantly feasible. Various equations for predicting EE based on body weight (BW) are available. This study aims at determining the best prediction strategy unless IC is available.

Methods: Mechanically ventilated patients staying72 h in the ICU were included, except those with contraindications for IC measurements. IC and BW measurements were routinely performed. EE was predicted by the ESPEN formula and other predictive equations using BW (i.e. anamnestic (AN), measured (MES), adjusted for cumulated water balance (ADJ), calculated for a body mass index (BMI) of 22.5). Comparisons were made using Pearson correlation and Bland&Altman plots.

Results: 85 patients (57±19 y, 61 men, SAPS II 43±16) were included. Correlations between IC and predicted EE using the ESPEN formula with different BW (BWAN, BWMES, BWADJ, and BWBMI22.5) were 0.44, 0.40, 0.36, and 0.47, respectively. Bland&Altman plots showed wide and inconsistent variations.

Predictive equations including body temperature and minute ventilation showed the best correlations, but when using various BWs, differences in predicted EE were observed.

Conclusion: No EE predictive equation, regardless of the BW used, gives statistically identical results to IC.

If IC cannot be performed, predictive equations including minute ventilation and body temperature should be preferred. BW has a signiﬁcant impact on estimated EE and the use of measured BWMESor BW BMI 22.5is associated with the best EE prediction.

Clinical trial registration number on ClinicalTrial.gov: NCT02552446.

Ethical committee number: CE-14-070.

1. Introduction

Optimal nutritional care for intensive care unit (ICU) patients requires precise determination of the energy expenditure (EE) to avoid under- or over-feeding, both being deleterious[1,2]. Although indirect calorimetry (IC) is considered the reference method to

determine EE, it is rarely available and suffers from some limitations in its feasibility. Various equations for predicting EE based on body weight (BW) have been proposed. This study aims at determining the best EE prediction strategy when IC is not available.

EE can be determined by three methods in ICU patients. Firstly EE can be computed from pulmonary artery catheter measurements, using the Fick equation; however this method is highly invasive as it requires a right heart catheter. Secondly EE can also be measured by IC, which is the recommended method[3,4]. How- ever, IC requires a trained staff, an expensive and validated calo- rimeter for accurate measurements[5,6]and is contraindicated in a number of clinical situations in the critically ill patient (e.g. air leaks in the circuit, ventilation with inspired fraction of oxygen (FiO2) > 60%, end-expiratory pressure (PEEP) > 9 cm H2O, gas

*Corresponding author. Intensive Care Unit, Geneva University Hospital, Rue Gabrielle-Perret-Gentil 4, 1211 Geneva 14, Switzerland. Tel.:þ41 22 372 74 40;

fax:þ41 22 382 74 70.

E-mail addresses: severine.graf@hcuge.ch(S. Graf), claude.pichard@hcuge.ch (C. Pichard), laurence.genton@hcuge.ch (L. Genton), oshima.taku@hcuge.ch (T. Oshima), Claudia.Heidegger@hcuge.ch, claudia-paula.heidegger@hcuge.ch (C.P. Heidegger).

Contents lists available atScienceDirect

Clinical Nutrition

j o u r n a l h o m e p a g e : h t t p : / / w w w . e l s e v i e r . c o m / l o c a t e / c l n u

http://dx.doi.org/10.1016/j.clnu.2015.11.007

(3)

exchange therapy or inhaled nitrogen gas). Only about 70% of the critically ill ICU patients can be measured by this method [7].

Finally, EE can be estimated by using predictive equations. Of note is that most of these equations were developed for healthy people.

For ICU patients, a number of correction factors have been suggested taking account of different stress levels, type of treatments, and body compositions such as severe obesity or extreme leanness (i.e. sex, age, height, weight, body temperature, minute ventilation, pathology, severity of disease…). All predictive equations include patients' body weight (BW). However, the use of BW in ICU patients is highly controversial because it largely varies according toﬂuid retention and edema related to the inﬂammation process and resuscitation therapies[8].

This quality of care study is part of a global effort to improve and harmonize our clinical practices. It aims at identifying the best EE prediction strategy, with regard to the choice of the BW and of the predictive equation, when IC is not available.

2. Materials and methods 2.1. Patients

Patients mechanically ventilated, staying72 h in the ICU were included. Due to our infection control policy, patients with respi- ratory multi-drug resistant bacteria were excluded as well as those with unstable pH, FiO2>60%, PEEP>9 cm H2O, and pulmonary ﬁstula.

The Ethical Committee of the Geneva University Hospital approved this study. They waived the need for a written consent because the study was part of a continuous quality of care improvement in our ICU, and because the study did not generate any additional risks for the patient as the measurements are routinely performed in our ICU.

2.2. Methods

The clinical data were prospectively collected from our elec- tronic medical records. These data included gender, body temperature, details of mechanical ventilation (type of ventilator, spontaneous or controlled ventilation, minute ventilation, etc), hemodynamic parameters (heart rate, arterial pressure, etc), labo- ratory data, treatments received the day of IC measurement (morphine, sedatives etc.), anthropometric data (height, anamnestic BW, age), ICU admission diagnosis and severity scores (APACHE II and SAPS II) during theﬁrst 48 h after ICU admission and at the time of the IC measurement.

❖Anthropometric data and predictive equations

Anamnestic body weight (BW_AN) was obtained from the elec- tronic medicalﬁle or from the family members. Actual body weight (BWMES) was measured using the built-in bed scale. Adjusted body weight (BW_ADJ) was determined by correcting the measured body weight (BWMES) according to the cumulativeﬂuid balance calculated from the ICU admission to the day of the IC measurement.

Ideal BW (IBW) was calculated for a reference body mass index (BMI) at 22.5 kg/m²(IBW_22,5), and at 25 kg/m²(IBW₂₅) or calculated using the Metropolitan Life Insurance tables (IBWMLI), the Lorentz equation (IBWLO) (height (cm)100((height (cm)150)/2.5 for women and/4 for men)) and the Broca equation (IBW_BRO) (height (cm)100). We then used the following equations to calculate EE with the different BWs listed above: ESPEN formula [9], Har- riseBenedict, Black et al.[10], Faisy et al.[7], Frankenﬁeld et al.[11], Brandi et al.[12], Ireton-Jones et al.[13], Penn State et al.[14]and Swinamer et al.[15](web Appendix Table 4).

❖ Indirect Calorimetry

IC was performed using the Deltatrac II^®(Datex, Finland) as previously described[5]. Each measurement was done by trained professionals for at least 25 min. Data of theﬁrst 5 min were dis- carded. Patients were measured in standardized conditions[16]

and on-going feeding was not interrupted.

2.3. Statistical analysis

Data are presented as mean±standard deviation (SD), or median and range, as appropriate. Parametric analyses were used after conﬁrming the normal data distribution using Skewness and Kur- tosis tests.

We compared the different BWs by one-way ANOVA followed, in case of signiﬁcance, by a Bonferroni post-hoc test. T-tests were used to assess the differences of mean EE obtained from IC measurements with binary variables such as gender, fever, surgical versus medical patients, fed versus fasting patients, and spontaneous versus controlled ventilation modes.

We correlated the predictive equations with different BWs and IC by Pearson's correlation tests. The distribution of patients with predicted EE below 90% of measured EE by IC (underestimation), between 90 and 110% (acceptable), and over 110% (overestimation) were calculated in percentage. These percentages were considered as a measure of the precision of the predictive strategies. Bland&

Altman plots were performed to determine the variability between the IC data (considered as the reference) and the predictive equations with the various BWs. Mean differences with limits of agreement (±2 SD) were recorded as the reﬂection of the accuracy.

STATA version 13.0 (StataCorp, Texas, USA) was used for the analyses and the level of signiﬁcance was set at 0.05.

3. Results

Eighty-five patients were included (Table 1andweb Appendix Fig. 1), of which 26 patients were on controlled and 59 on spontaneous ventilation. The different BWs are presented inTable 1and show significant differences with measured BW (p<0.001) (web Appendix Table 1). The influence of the different BWs used in the EE predictive equations is significant for the Black equation (p¼0.0001), the Brandi equation (p¼0.0017), the ESPEN formula (p < 0.0001), the Frankenfield equation (p ¼ 0.0018), the Har- riseBenedict equation (p<0.0001) and the Penn State equation (p ¼ 0.0001), but has no impact for the Ireton-Jones equation (p¼0.7744), the Faisy equation (p¼0.0950) and the Swinamer equation (p¼0.7900).

Measured EE was obtained by IC performed on the median of the 4th day after admission (range 1e15). No statistical difference was observed between the results of medical and surgical patients (p¼0.6935), fed and fasting patients (p¼0.3967), and spontaneous and controlled ventilation (p¼ 0.1126). On the contrary, gender and temperature37C were associated with signiﬁcant difference in EEs (p¼0.0004 and 0.0051, respectively).

Calculated EEs with predictive equations differed from EE measured by IC (Table 2; more details onweb Appendix Table 2).

Best Pearson correlations with measured EE are observed with results from the Frankenﬁeld or the Penn State predictive equations, whereas the ESPEN formula, whatever the BW selected, presented poor correlation (0.36).

The occurrence of underestimation of EE when calculated, compared to the measured EE is least frequent with the Penn State equation using measured BW, whereas overestimation of EE is the least frequent with the Frankenﬁeld equation using the ideal BW calculated for a BMI at 22.5 kg/m²(Fig. 1,web Appendix Table 3).

S. Graf et al. / Clinical Nutrition xxx (2015) 1e5 2

(4)

The plot distribution shows no tendency for systematic over- or underfeeding when using the different predictive equations.

Similar results are found when using the Penn State equation with the measured BW.

Finally, none of the equations are accurate enough to replace IC as their variability (±2 SD) is clinically signiﬁcant (±652 kcal using Frankenﬁeld equation with ideal BW for a BMI at 22.5) (Fig. 2).

4. Discussion

This study shows for thefirst time that the choice of the BW used in predictive equations has a significant impact on the evaluation of the EE in ICU patients (web Appendix Table 1). It also reconfirms that no predictive equation accurately matches the EE measured by IC.

4.1. Body weight measurement is inaccurate

Ireton-Jones et al.[17]suggested that the measured BW is more suitable for use in EE predictive equations compared to the ideal Table 1

Patients' anthropometric and clinical characteristics at calorimetry measurement.

All (n¼85) Men (n¼61) Women (n¼24)

Age (years) 57±19 56±20 60±16

Height (m) 1.71±0.09 1.75±0.07 1.62±0.07

BWAN 76.1±14.8 78.9±12.7 69.0±17.4

BWBMI22.5 66.1±7.1 68.9±5.8 59.1±4.6

BWMES 79.5±15.3 82.3±14.7 72.2±14.9

BWADJ 75.9±15.5 78.7±14.5 68.8±16.2

BMI with BWMES 27.1±4.9 27.0±4.8 27.5±5.2

BMI with BWADJ 25.9±4.9 25.8±4.6 26.2±5.7

SAPS II (point) 43±16 44±16 40±17

APACHE II (point) 23±6 23±6 22±7

Inspired fraction of oxygen (%) 28±7 27±6 29±8

PEEP (cm H2O) 6±1 6±1 6±1

Minute ventilation (l/min) 9±3 10±3 9±3

Body temperature (C) 37±1 37±1 37±1

C-reactive protein (mg/l) 108±79 95±85 119±62

Heart rate (pulse) 83±21 82±23 88±17

Mitral arterial pressure (mmHg) 85±15 85±15 86±17

Data are presented in mean±standard deviation.

BW: body weight; AN: anamnestic; MES: measured; ADJ: adjusted for cumulated water balance; SAPS II: severity score“Simpliﬁed Acute Physiology Score”; APACHE II: severity score“Acute Physiology And Chronic Health Evaluation”; PEEP: positive end-expiratory pressure.

Table 2

Comparison between indirect calorimetry and predictive equations with the most relevant body weights.

Pearson correlation Mean±DS Mean^b Limits of agreement (±2 SD)

Indirect calorimetry e 1879±532 e e

ESPEN* BWANa 0.44 2181±490 271 (1272; 229)

BWADJa 0.40 2178±516 265 (1339; 271)

BWMESa 0.36 2281±518 368 (1477; 186)

BWBMI22.5 0.47 1895±309 15 (812; 428)

Frankenﬁeld* BWAN 0.71 1780±316 130 (518; 777)

BWADJ 0.70 1802±312 111 (549; 770)

BWMES 0.70 1824±307 88 (572; 749)

BWBMI22.5a 0.71 1677±280 232 (419; 558)

Faisy BWAN 0.63 1951±276 42 (753; 670)

BWADJ 0.63 1954±268 42 (746; 714)

BWMES 0.62 1982±263 70 (785; 699)

BWBMI22.5 0.63 1871±244 38 (675; 752)

Penn State* BWAN 0.70 1871±363 38 (619; 695)

BWADJ 0.69 1899±361 14 (658; 350)

BWMES 0.69 1927±355 15 (687; 322)

BWBMI22.5a 0.72 1738±310 171 (464; 807)

*ANOVA p-value<0.05.

aBonferroni post-hoc p-value<0.05.

b Mean differences.

14 17 14 21 38 33 29 52

19 19 19 24 29 20 19 19 27 28 21

50 44 49 51

36

42 46 41 47 44 49 46 68 59 55 65

29 18 18 20 12

39 35 40 29 27 31 35 13

0 10 20 30 40 50 60 70 80 90 100

Amnestic Adjusted Measured BMI 22.5 Amnestic Adjusted Measured BMI 22.5 Amnestic Adjusted Measured BMI 22.5 Amnestic Adjusted Measured BMI 22.5

ESPEN Frankenfield Faisy Penn State

Percentage of patients

ESPEN formula and predictive equations with various body weights

over 110%

90-110%

under 90%

Fig. 1.Under- or overestimation of energy expenditure by predictive equations compared to indirect calorimetry.

(5)

BW obtained from the Metropolitan life Insurance Company stan- dards. Cutts et al.[8]reported that adjusted BW is preferable to measured BW for overweight patients, in order to avoid overfeeding.

Our results are consistent with those of others studies demonstrating that EE predictive equations poorly predict EE measured by IC[18e20]. This observation is at least partly due to the inaccuracy of BW determination/evaluation. First, the so- called actual BW is often estimated or anamnestic, which can create major errors. Second, measured BW can be a poor indicator of the metabolic requirements in the presence offluid retention related to metabolic stress, kidney or heart failures,fluid administration or diuretics administration [8]. However, it seems less worse when BW is estimated. Third, the BW poorly reflects the metabolically active lean body mass (i.e. mostly muscle mass) which deeply influences EE[21]. Indeed, even with the same BW, a sarcopenic obese elderly patient typically features a reduced lean body mass and low EE, whereas a young athletic patient would have a minimal fat body mass, and a high EE. As the metabolically active lean body mass is the major determinant of the resting EE, the development of predictive formulas including this factor in replacement of body weight seems a reasonable approach. However measurements of lean body mass are difficult

in the critical care settings, mainly because of the technical limits of the different methods [22]. The gold standard for measuring body composition is dual-energy X-ray absorptiometry (DXA).

However, DXA requires complex equipment and trained manpower and cannot be performed at the bedside for routine use in all ICU patients. Bioelectrical impedance analysis (BIA) allows to measure body composition but results are strongly inﬂuenced by ﬂuid and electrolytic imbalances frequently observed in the critically ill patients (edema, ascites, renal failure, etc.), making the measurements unreliable. Muscle sonography is a promising alternative to assess muscle mass, and may become useful in the near future[23]. Measurements of lean body mass derived from routine CT-Scan obtained at the level of the third lumbar vertebra (L3), allows to calculate the total skeletal muscles area at this body level, but does not measure whole body composition, and it's capacity to predict the total body lean mass remains to be proven. Finally, BW does not take into account the impact on EE of preexisting metabolism alterations[24]as well as metabolic alterations secondary to the injury response[25].

4.2. ICU-speciﬁc equations perform better

The Frankenﬁeld equation[10]with the ideal BW for a BMI at 22.5 kg/m²and the Penn State equation[13]with the measured BW, show the best correlation with the measured EE. However, the relative risks of under- or overestimation of EE should be taken into account when using these two equations.

The worst results were obtained with the ESPEN formula using the measured BW (mean difference368±555 kcal) as shown in Table 2andweb Appendix Table 2. Aﬁxed amount of calories per kg of BW, is recommended by the ESPEN Society for ICU patients (i.e.

20e25 kcal/kg acute phase, 25e30 kcal/kg/d later on)[4]. Although this concept is pragmatic, it is probably too simple toﬁt for a highly heterogenous population of ICU patients as previously suggested [26]. Compared to other equations, the Faisy[7], Frankenﬁeld[11]

and Penn State [14] take into account specific criteria such as minute ventilation or body temperature. In spite of this characteristics, their accuracy compared to IC remains limited probably because of other factors influencing EE such as medications, treatments and significant EE variations occurring during the course of the critical illness[27e29].

In absence of IC, our results suggest that the use of measured BW or adjusted BW for cumulated water balance is an acceptable op- tion when used in conjunction with ICU patient specific EE predictive equations (i.e. Frankenfield[10], Penn State[13]), although this method is associated with overestimation of EE in nearly 20% of cases. The use of ideal BW based on the body height is the easiest method, independent of anyfluid overload, but does not take into account the variations related with the metabolic disturbances or the lean body mass mentioned above, and therefore leads to the largest errors of EE prediction. This method is associated with a high risk of underestimation of EE even when used with ICU- specific EE predictive equation.

4.3. Future developments

As the IC is the recommended method to know precisely EE, an on-going international initiative to develop a new IC is supported by the European Society of Clinical Nutrition (ESPEN) and the Eu- ropean Society of Intensive Care Medicine (ESCIM)[2]. This device has been conceived to be accurate, affordable and without the actual limitations of clinical use. The clinical tests are currently performed and the availability of the device is expected for 2016.

A-Frankenfield Equation with ideal BW for a BMI at 22.5 vs. Indirect Calorimetry

B-ESPEN equation with measured BW vs. Indirect Calorimetry -1000

-500 0 500 1000 883

-419 232

Indirect Calorimetry - Frankenfield BW22.5

1000 1500 2000 2500 3000

(Indirect Calorimetry + Frankenfield BW BMI22.5)/2'

-1500 -1000 -500 0 500 1000 1500

-1477 741

-368

Indirect Calorimetry - ESPENmes

1000 1500 2000 2500 3000

(Indirect Calorimetry + ESPENmes) / 2

Fig. 2.Bland&Altman plots comparing energy expenditure between indirect calorimetry and predictive equations using different body weights.

S. Graf et al. / Clinical Nutrition xxx (2015) 1e5 4

(6)

4.4. Study limitations

This prospective unblinded study has a number of limitations such as the great heterogeneity of the ICU patient population and its metabolic variations during the different phases of critical illness which impact on the results of EE measurements. Indeed, timing of IC measurements was made according to our routine measurements of EE and probably does not reflect the changes of EE during the rest of the day in the ICU. Also, our study population may not reflect all ICU patients as patients with contraindications to IC, representing about 30% of all ICU patients, were not included and so didn't take part in the comparisons of predictive equations with various BWs and IC. In spite of these limitations, we believe that our findings remain clinically relevant for most of the ICU patients.

5. Conclusion

No predictive equation of EE, regardless of the BW used, gives statistically identical results to IC. If IC cannot be performed, EE predictive equations which include body temperature and minute ventilation should be preferred. BW has a signiﬁcant impact on estimated EE, the use of measured BW or the ideal BW for BMI at 22.5 is associated with the best EE estimation.

Conﬂict of interest

Financial support came from the APSI-ICU quality funds of the Geneva University Hospital. All authors declare that they have no conﬂict of interest, except Claude Pichard who has received grants from the European Society of Clinical Nutrition (ESPEN) and the European Society of Intensive Care Medicine (ESCIM) to lead an international initiative to develop a new IC.

Statement of authorship

Severine Graf conceived the study, participated in its design, recruited patients, collected, analyzed and interpreted the data, and drafted the manuscript. Claudia-Paula Heidegger participated in the study design, obtained funding and drafted the manuscript.

Laurence Genton participated in the interpretation of the data and drafts the manuscript. Taku Oshima participated in the collection and interpretation of the data and drafts the manuscript. Claude Pichard participated in the study design, interpretation of the data, and drafted the manuscript.

Appendix A. Supplementary data

Supplementary data related to this article can be found athttp://

dx.doi.org/10.1016/j.clnu.2015.11.007. References

[1] Heidegger CP, Berger MM, Graf S, Zingg W, Darmon P, Costanza MC, et al.

Optimisation of energy provision with supplemental parenteral nutrition in critically ill patients: a randomised controlled clinical trial. Lancet 2013 Feb 2;381(9864):385e93.

[2] Guttormsen AB, Pichard C. Determining energy requirements in the ICU. Curr Opin Clin Nutr Metab Care 2014 Mar;17(2):171e6.

[3] McClave SA, Martindale RG, Kiraly L. The use of indirect calorimetry in the intensive care unit. Curr Opin Clin Nutr Metab Care 2013 Mar;16(2):202e8.

[4] Singer P, Berger MM, Van den Berghe G, Biolo G, Calder P, Forbes A, et al.

ESPEN guidelines on parenteral nutrition: intensive care. Clin Nutr 2009 Aug;28(4):387e400.

[5] Graf S, Karsegard VL, Viatte V, Heidegger CP, Fleury Y, Pichard C, et al. Eval- uation of three indirect calorimetry devices in mechanically ventilated patients: which device compares best with the Deltatrac II((R))? A prospective observational study. Clin Nutr 2015 Feb;34(1):60e5.

[6] Sundstrom M, Tjader I, Rooyackers O, Wernerman J. Indirect calorimetry in mechanically ventilated patients. A systematic comparison of three in- struments. Clin Nutr 2013 Feb;32(1):118e21.

[7] Faisy C, Guerot E, Diehl JL, Labrousse J, Fagon JY. Assessment of resting energy expenditure in mechanically ventilated patients. Am J Clin Nutr 2003 Aug;78(2):241e9.

[8] Cutts ME, Dowdy RP, Ellersieck MR, Edes TE. Predicting energy needs in ventilator-dependent critically ill patients: effect of adjusting weight for edema or adiposity. Am J Clin Nutr 1997 Nov;66(5):1250e6.

[9] Kreymann KG, Berger MM, Deutz NE, Hiesmayr M, Jolliet P, Kazandjiev G, et al.

ESPEN guidelines on enteral nutrition: intensive care. Clin Nutr 2006 Apr;25(2):210e23.

[10] Black AE, Coward WA, Cole TJ, Prentice AM. Human energy expenditure in afﬂuent societies: an analysis of 574 doubly-labelled water measurements.

Eur J Clin Nutr 1996 Feb;50(2):72e92.

[11] Frankenﬁeld DC, Omert LA, Badellino MM, Wiles 3rd CE, Bagley SM, Goodarzi S, et al. Correlation between measured energy expenditure and clinically obtained variables in trauma and sepsis patients. J Parenter Enteral Nutr 1994 SepeOct;18(5):398e403.

[12] Brandi LS, Santini L, Bertolini R, Malacarne P, Casagli S, Baraglia AM. Energy expenditure and severity of injury and illness indices in multiple trauma patients. Crit Care Med 1999 Dec;27(12):2684e9.

[13] Ireton-Jones CS, Turner Jr WW, Liepa GU, Baxter CR. Equations for the estimation of energy expenditures in patients with burns with special reference to ventilatory status. J Burn Care Rehabil 1992 MayeJun;13(3):330e3.

[14] Frankenﬁeld D, Smith JS, Cooney RN. Validation of 2 approaches to predicting resting metabolic rate in critically ill patients. J Parenter Enteral Nutr 2004 JuleAug;28(4):259e64.

[15] Swinamer DL, Grace MG, Hamilton SM, Jones RL, Roberts P, King EG. Predic- tive equation for assessing energy expenditure in mechanically ventilated critically ill patients. Crit Care Med 1990 Jun;18(6):657e61.

[16] Compher C, Frankenﬁeld D, Keim N, Roth-Yousey L. Best practice methods to apply to measurement of resting metabolic rate in adults: a systematic re- view. J Am Diet Assoc 2006 Jun;106(6):881e903.

[17] Ireton-Jones CS, Turner Jr WW. Actual or ideal body weight: which should be used to predict energy expenditure? J Am Diet Assoc 1991 Feb;91(2):193e5.

[18] Anderegg BA, Worrall C, Barbour E, Simpson KN, Delegge M. Comparison of resting energy expenditure prediction methods with measured resting energy expenditure in obese, hospitalized adults. J Parenter Enteral Nutr 2009 MareApr;33(2):168e75.

[19] Boullata J, Williams J, Cottrell F, Hudson L, Compher C. Accurate determination of energy needs in hospitalized patients. J Am Diet Assoc 2007 Mar;107(3):

393e401.

[20] Kross EK, Sena M, Schmidt K, Stapleton RD. A comparison of predictive equations of energy expenditure and measured energy expenditure in critically ill patients. J Crit Care 2012 Jun;27(3):321. e5e12.

[21] Watson LP, Raymond-Barker P, Moran C, Schoenmakers N, Mitchell C, Bluck L, et al. An approach to quantifying abnormalities in energy expenditure and lean mass in metabolic disease. Eur J Clin Nutr 2014 Feb;68(2):234e40.

[22] Thibault R, Genton L, Pichard C. Body composition: why, when and for who?

Clin Nutr 2012 Aug;31(4):435e47.

[23] Mourtzakis M, Wischmeyer P. Bedside ultrasound measurement of skeletal muscle. Curr Opin Clin Nutr Metab Care 2014 Sep;17(5):389e95.

[24] Leibel RL, Rosenbaum M, Hirsch J. Changes in energy expenditure resulting from altered body weight. N Engl J Med 1995 Mar 9;332(10):621e8.

[25] Jeevanandam M, Young DH, Schiller WR. Obesity and the metabolic response to severe multiple trauma in man. J Clin Investig 1991 Jan;87(1):262e9.

[26] McClave SA, Lowen CC, Kleber MJ, Nicholson JF, Jimmerson SC, McConnell JW, et al. Are patients fed appropriately according to their caloric requirements?

J Parenter Enteral Nutr 1998 NoveDec;22(6):375e81.

[27] Chiolero R, Revelly JP, Tappy L. Energy metabolism in sepsis and injury.

Nutrition 1997 Sep;13(9 Suppl.):45Se51S.

[28] Doig GS, Simpson F, Sweetman EA, Finfer SR, Cooper DJ, Heighes PT, et al.

Early parenteral nutrition in critically ill patients with short-term relative contraindications to early enteral nutrition: a randomized controlled trial.

JAMA 2013 May 22;309(20):2130e8.

[29] Plank LD, Connolly AB, Hill GL. Sequential changes in the metabolic response in severely septic patients during theﬁrst 23 days after the onset of perito- nitis. Ann Surg 1998 Aug;228(2):146e58.