PREDICTING

ENERGY

REQUIREMENTS

Frances Phillips APD

Chief Dietitian

Royal Perth Hospital Nov 2010

“…..Estimated energy requirementsare only a starting point, and that allclinicians should regularly reviewtheir patients to ensure they aremeeting their nutritional goals and toevaluate the effectiveness ofnutritional support”

Weekes Proc Nutr Soc 2007

Survey

Survey:

Use Energy Equations Non Critical

Care Adult Patients Stand Up

Harris Benedict EquationHands at Side

Schofield …. Assume eg = SF 10% and AF 15%

(a) If Multiply BMR x [1.0 + (SF + AF)] (SF +AF = 25%)

(Non Cumulative approach)

Hands on Shoulders(b) If Multiply BMR x (SF or AF), and then multiply other factor (Cumulative approach)

Hands on Head Neither HB or Schofield

Wave Hands

Basil Metabolic Rate (BMR)

• Largest part of an individual’s total energy expenditure(TEE)

• Energy for internal mechanical activities and maintenanceof body temperature

• Measure:

Post fast At physical and mental rest 22-290C environment No artificial stimulants (eg. tea, coffee, nicotine) No physical activity previous day

• Very difficult to control in clinical practice

Weekes Proc Nutr Soc 2007, Garrel NCP 1996, Levine Pub Health Nutr 2005

BMR

Women:

Highest between ages 23-35Increases more slowly > 65 kgMore rapid drop > 50 years than for Men

Men:

Highest between ages 18-20Increases more slowly > 75 kg

Horgan Eur J Clin Nutr 2003

Resting Metabolic Rate (RMR)

65-70% of TEE

Little day to day variations

Declines with age Loss Fat Free Mass & Gain of Fat Mass

Subject to less strict conditions than BMR• Client Specific (Fasting, Exercise, Resting, Comfort)

• Machine Specific (Calibration, Steady State,

Adequate Test Time)

Rest of TEE

Thermogenic response to food (10%)

Expenditure physical activity (15 - 30%)Garrel NCP 1996, Pochlman Med Sci Sports Exerc 1989 , Roberts Pub Health Nutr 2005

Energy Expenditure Equations

Approx 138 Formulae (or more)....

Published by 40 Authors• Many involve critical care patients

Most common Australian Energy Expenditure Equations (Non Critical Care Pts)

• Harris Benedict Equation

• Schofield Equations

Others:• Energy Estimates (per kg)

• Others………..

Reeves Eur J Clin Nutr 2003, Reeves Nutr Rev 2003

Energy Equations - Limitations

All equations open to criticism for multiple reasons

Equations Based on Predictions for “Groups” Poor predictive value for “Individuals”

Equations do not explain 20% of variations between individuals

(eg. tissue variations, disease, genetics

trauma including time since trauma)

Require some clinical judgement

Reeves Eur J Clin Nutr 2003, Schoeller JADA 2007, Weekes Proc Nutr Soc 2007

Energy Equations – Limitations

(cont) Many based on measurements taken over 50

years with some data closer to 100 years• Early data was manually “ cleaned up”

• Some population groups “over” or “under” represented

• Lack of validation studies

• Gender % variations (eg. more males)

Current population More elderly

Increased height and weight

More overweight and obese

Early puberty

Equations focused on “healthy” – not designed tobe applied for disease or injury Reeves Nutr Rev 2003

Harris – Benedict Equation

Background Developed between 1909 and 1917 and based on indirect

calorimetry. Published in 1919 (J. Arthur Harris & Francis G.Benedict)

Predominantly “normal” weight healthy white subjects(Boston) (N =239)

136 men aged 16-63 years (13 underweight)

103 women aged 15-74 years (21 underweight)

Younger (Mean Age F 27 +/- 9 Years , M 31 +/- 14 Years)

Leaner (Mean BMI F 21 +/-3 Years , M 22 +/- 4 Years)

More active than current population

Resting Metabolic Rate Frankenfield JADA 2005, Reeves Nutr Rev 2003

Harris – Benedict Equation

Males:

RMR = 278 + (57.5 x W) + (20.9 x H) – (28.3 x A)

Females:

RMR = 2741 + (40 x W) + (7.7 x H) – (19.6 x A)

Key:

RMR Resting Metabolic Rate (kJ/day)

W Weight (kg)

H Height (cm)

A Age (years)

Reeves Nutr Rev 2003

Adjustments to the HB Equation

HB Equation x AF X IF (Long et al 1979)

• Activity factor• Resting 1.1

• Confined to bed 1.2

• Out of bed 1.3

• Injury factor (N = 20-39 pts)• Minor Op 1.2

• Skeletal Trauma 1.35

• Cancer Cachexia 1.3-1.5

• Major Sepsis 1.6

• Severe Thermal Injury 2.1

• Febrile 1 +0.09 per 0.5 0 c > NReeves Nutr Rev 2003, Long JPEN 1979

? HB = BEE or REE

Early works report HB = BEE

Later considered conditions = REE

Seale 1995:

Factor of 1 to 1.5 added to the BEE of HB Equation = RMR

Thomas 2007, Seale Am J Clin Nutr 1995

Harris – Benedict Equation

Validation

1950’s Measured RMR Within 5%

More recent studies (Indirect Calorimetry)

HB overestimates RMR by 6-15%

More predictive for Men

Large variation between persons

Reeves Nutr Rev 2003, Garrel NCP 1996

Harris – Benedict Equation

Interperson Variation Study

Published 1996 Garrel NCP 1996

RMR (HB Equation vs Indirect Calorimetry)

Subjects (39 Men and 28 Women)

• Healthy, Normal Weight, Sedentary

Results:

• Mean overestimate 11.7%

• RMR Quartile Variations

• Predictive values women

less accurate than men

20.6% 11.5% 9.5% 4%

Lowest Highest

Systematic Review HB 2005

25 Studies …..Focused on individuals not

groups

HB as a Prediction of RMR

• Healthy Non Obese Individual Adults

45 - 80% Individuals

Overestimates > Underestimates

• Healthy Obese Individual Adults

38 - 64% Individuals Overestimates > Underestimates

Adjusted weight ↓ Risk of overestimating RMR

↑ ↑ Maximum underestimation error

• Older Adults

Men: Maximum underest 19%, overest 9% RMR

Women: Maximum underest 27%, overest 12% RMRFrankenfield JADA 2005

Schofield Equations First published in 1985

commonly used in Europe and Australia

Developed from the statistical screening of data inthe literature from 1914 to 1980

Assumed linear relationship between BMR andWeight

Formed the basis for FAO/WHO/UNU Report 1985.

Recommended for use in Australians (1990 , 2005)Schofield Hum Nutr Clin Nutr 1985, Ramirez-Zea Pub Health Nutr 2005, Reeves Nutr Rev 2003, FAO/WHO/UNU 1985,

Hayter Eur J Clin Nutr 1994, Truswell 1990, Aus Govt 2005

Schofield Equations Meta-analysis 114 studies (7173 data points BMR)

inc HB

4809 Men and 2364 Women

Infants to Adults (0-100 Yrs)

↑ Europeans and North American &

↓ Developing countries

Italians (esp military)

47% of the Schofield database >18 Years

Higher proportion of Males, esp Italians 57%

Very active, ↑ BMI's than other caucasians

Few elderly (Only 1-2% > 60 years)

Studies used “Closed-Circuit” Methods – ↑BMR’s

? Stress/anxiety breathing from closed chamberRamirez-Zea Pub Health Nutr 2005, Ferro-Luzzi Pub Health Nutr 2005

Schofield Equations

Italians • Young, physically active

(inc labourers, miners)

• Higher BMR/kg than others

• Most divergent subject group

Indians and Chinese • 10% Lower BMR than Europeans +

Americans

Age, Sex & Weight matchedHayter Eur J Clin Nutr 1994

Schofield Equations

Age (yrs) RMR (Males) RMR (Females)

0-3 (0.249 x W) – 0.127 (0.244 x W) – 0.130

3-10 (0.095 x W) + 2.110 (0.085 x W) + 2.033

10-18 (0.074 x W) + 2.754 (0.056 x W) + 2.898

18-30 (0.063 x W) + 2.896 (0.062 x W) + 2.036

30-60 (0.048 x W) + 3.653 (0.034 x W) + 3.538

>60 (0.049 x W) + 2.459 (0.038 x W) + 2.755

Key:

RMR = Resting Metabolic Rate (MJ/day)

W = Weight (kg)

TEE = BMR x (1.0 + %IF + %AF) or BMR x IF (Elia) x AF

Schofield Hum Nutr Clin Nutr 1985, Reeves Nutr Rev 2003

FAO/WHO/UNU (1985)

Age (yrs) RMR (Males) RMR (Females)

0-3 (0.255 x W) – 0.226 (0.255 x W) – 0.214

3-10 (0.0949 x W) + 2.07 (0.0941 x W) + 2.09

10-18 (0.0732 x W) + 2.72 (0.051 x W) + 3.12

18-30 (0.064 x W) + 2.84 (0.0615 x W) + 2.08

30-60 (0.0485 x W) + 3.67 (0.0364 x W) + 3.47

>60 (0.0565 x W) + 2.04 (0.0439 x W) + 2.49

Key:

RMR = Resting Metabolic Rate (MJ/day)

W = Weight (kg)

Curtin University 2006

British Use Of Schofield Equations

British Dietetic Association 2007

Calculate BMR eg. using Schofield Equation Adjust for Stress Factors (SF)

Elia Nomogram (1990) or

Todorovic and Micklewright (2004) Add a combined factor for activity and dietary induced

thermogenesis (AF)

Bedbound, immobile + 10%

Bedbound, mobile or sitting + 15-20%

Mobile, on ward + 25%

BDA Note: Adding SF and AF is not cumulative Eg If 10% SF and 15% AF = BMR ↑ 25% total

(not ↑ BMR by 10%, then 15%)

Thomas 2007, Elia Med Int 1990

Elia Nomogram

Estimate energy requirements in disease

Used in conjunction with Schofield equations (British Dietetic Assoc)

? Expert opinion

No reference to research or studies or how it is derived

Reeves Nutr Rev 2003

Elia NomogramElia Pub Health Nutr 2005

Stress Factors (% BMR)Todorovic & Micklewright 2004

Brain Injury 0-50%

Cerebral Haem 30%

CVA 5%

COPD 15-20%

Infection 25-45%

IBD 0-10%

ICU 0-60%

Leukaemia 25-34%

Lymphoma 0-25%

Pancreatitis 3-10%

Sepsis / Abcess 20%

Solid Tumours 0-20%

Transplantation 20%

Surgery 5-40%

AF : Schofield vs HB Equations

Bed Rest 1.2

Very Sedentary 1.3

Sedentary 1.4

Light 1.5

Moderate 1.8-1.7

Heavy 2.1-1.8

Very Heavy 2.3-2

Resting 1.1

Confined to Bed 1.2

Light 1.3

Moderate 1.4-1.5

Heavy 1.75

Griffith University 2007

Schofield Equations

Height not required

Other equations including height weredeveloped, however including height did notimprove accuracy.

Mean BMI (>16 years) is not known.

14.6% BMI > 25

4.5% BMI > 30

Lower % Overweight / Obese than today

Reeves Nutr Rev 2003, Horgan Eur J Clin Nutr 2003, Schofield Hum Nutr Clin Nutr 1985

Schofield Equations

General: Predictive value 36-53%

Men: Overestimate BMR in men 3.7%

Especially 18 - 29.9 Years

Greater overestimate BMR when overweight and obese included

Women: Equations not accurate in women

Reeves Nutr Rev 2003, Ramirez-Zea Pub Health Nutr 2005

Schofield Equations

Poor predictive value for adults ( > 18years )

Overestimates BMR current/most populations

7-10%

especially if overweight or obese included

Not accurate for the Elderly

Not accurate for young Australian Men/Women

(18-30 years)

Hayter Eur J Clin Nutr 1994, Piers Eur J Clin Nutr 1997, Reeves 2003, Henry Pub Health Nutr 2005

General Estimates of Energy

Sedentary Adults 25 to 30 kcal (=105to126 kJ) /kg body weight (BW)

25 = Unstressed pts ranging to 35 = Pyrexia or Extreme Stress

19 to 21 = Obesity

Slightly Hypermetabolic Patients

For Weight Gain 30 to 35 kcal (=126to146 kJ) / kg BW

Anabolic Patients

Hypermetabolic patients

Severely Stressed patients 35 kcal (=146 kJ) / kg BW

Patients with Malabsorption

Reeves Nutr Rev 2003, Weekes Proc Nutr Soc 2007

General Estimates of Energy

Does not take into account Age, Gender or Energy Expenditure

Use Actual or Ideal Weight?

? Developed for ventilated critically ill patients

Underestimates true requirements

In reality, amount of Energy per kg progressively decreases with increase in body size

Does not take into account body composition

What determines which value in the range should be used?

Reeves Nutr Rev 2003

Case Studies – Equation Comparisons

Green and Smith and Whelan Eur J Clin Nutr 2008

How estimate BMR?Which stress factor for 37 clinical conditions?Given Standard Scale of Stress Factors (Elia 1990)

40 year old maleNo PMHBMI 23Referred for Oral Nutrition Support

Reeves and Capra Nutr Rev 2003 ****TEE: 50 year old multiple skeletal traumas (Wt 68kg, Ht 175 cms)

Reeves and Capra Eur J Clin Nutr 2003Energy Requirement:50 year old post laryngectomy (exc Larynx Ca).Wt 54kg, Ht 170 cms, RIB, Medically stableWeight loss 7 kg over 6 months secondary to DysphagiaWeight stable & intake adequate (2 wks prior to surgery)

“New” Equations

(eg. Oxford)

Why better predictive value?

• Excluded all Italian subjects

• Included larger number from Tropics

• Sample Size

11,000 measurements

Garrel NCP 1996, Henry Pub Health Nutr 2005

A Few Closing Quotes

“…Clinical judgment must be used todetermine what level of nutrition careshould be based on these estimates”.

Frankenfield JADA 2005

“ …For many patient groups, however ,there is a general lack of studies fromwhich to make any usefulrecommendations”

Weekes Proc Nutr Soc 2007

The Last Words …….

“Nutrition is not an exact science.The most valuable tool in caring forthese patients is a clinical team thatuses energy expenditure calculationsas a starting point but relies onclinical experience andindividualization of patientmanagement to enhance outcome”

Ireton-Jones JPEN 2004

www.health.qld.gov.au/nutrition/resources/est_rqts.pdf

Queensland Government

NEMO Nutrition Support Group

“Estimating Energy & Protein Requirements for Adult Clinical

Conditions”

March 2009 (4 Pages)

REFERENCES

Australian Government & NHMRC (2005) Nutrient Reference Valuesfor Australia and New Zealand.

Curtin University of Technology (2006) Dietitians’ Pocketbook. Elia, M. (1990) Artificial nutritional support. Med. Int 82, 3392-6. Elia, M. (2005) Insights into energy requirements in disease. Pub

Health Nutr. 8(7A) 1037-52. FAO/WHO/UNU(1985) Energy and protein requirements. Report of

a joint FAO/WHO/UNU expert consultation. Technical Report SeriesNo. 724.

Ferro-Luzzi, A. (2005) The conceptual framework for estimatingfood energy requirement. Pub Health Nutr. 8(7A) 940-952.

Frankenfield, D. et al (2005) Comparison of predictive equations forresting metabolic rate in healthy non-obese and obese adults: Asystematic review. JADA. 105: 775-89.

Garrel, D et al (1996) Should we still use the Harris and BenedictEquations? NCP 11: 99-103.

Green, AJ et al (2008) Estimating resting energy expenditure inpatients requiring nutritional support: a survey of dietetic practice.Eur J Clin Nutr. 62: 150-3.

Griffith University (2007) Handbook of Nutrition and Dietetics.

REFERENCES (continued) Hayter, JE. et al (1994) A re-examination of basal metabolic rate

predictive equations: the importance of geographic origin ofsubjects in sample selection. Eur J Clin Nutr. 48: 702-7.

Henry, CJK. (2005) Basal metabolic rate studies in humans:measurement and development of new equations. Public HealthNutrition. 8(7A) 1133-52.

Horgan, GW et al (2003) Predicting basal metabolic rate in theobese is difficult. 57: 335-40.

Ireton-Jones, C (2004) Clinical dilemma: Which energy expenditureequation to use? JPEN. 28(4) 282-3.

Levine, JA (2005) Measurement of Energy Expenditure. PublicHealth Nutrition. 8(7) 1123-32.

Long, CL et al (1979) Metabolic response to injury and illness:Estimation of Energy and Protein needs from Indirect Calorimetryand nitrogen Balance. JPEN 3(6) 452-6.

Piers, LS et al (1997) The validity of predicting the basal metabolicrate of young Australian men and women. Eur. J Clin. Nutr. 51:333-7.

Pochlman, ET (1989) A review: exercise and its influence on restingenergy metabolism in man. Med Sci Sports Exerc. 21: 515-25.

Ramirez-Zea, M. (2005) Validation of three predictive equations forbasal metabolic rate in adults. Public Health Nutrition. 8(7) 1213-28.

REFERENCES (continued) Reeves, MM et al (2003) Predicting energy requirements in the clinical

setting: Are current methods evidenced based? Nutr Rev 61(4) 143-51.

Reeves, MM et al (2003) Variations in the application of methods usedfor predicting energy requirements in acutely ill adult patients: asurvey of practice. Eur J Clin Nutr 57, 1530-5.

Roberts, SB et al (1985) Energy requirements and aging. Pub HealthNutr 8(7) 1028-36.

Schofield, WN et al (1985) Predicting basal metabolic rate, newstandards ad review of previous work. Hum Nutr Clin Nutr 39C,Suppl1, 5-41.

Schoeller, DA. (2007) Making indirect calorimetry a gold standard forpredicting energy requirements for institutionalized patients. JADA.March. 390-2.

Seale, JL (1995) Energy expenditure measurements in relation toenergy requirements. Am J Clin Nutr. 62 (Suppl) 1042S-1046S.

Thomas, B. et al (2007) Manual of Dietetic Practice (BDA) 4th Edition Todorovic, VE et al (2004) A pocket guide to clinical nutrition. BDA Walker, PM (1990) Chapter 24. Predicting food energy requirements

from estimates of energy expenditure. In Recommended NutrientIntakes – Australian Papers. Ed Truswell, A.S.

Weekes, C. (2007) Controversies in the determination of energyrequirements. Proc Nutr Soc 66: 367-77.