- Click here for explanation of classification scheme.

1. Review the subjects and methods used to formulate the Harris-Benedict equations
2. Illustrate how closely predicted BEE agrees with measured BEE
3. Examine the rational for and limitations of the factors used in the equations
4. Compare subsequent studies of BEE with those of Harris and Benedict

The 1919, 1928, 1932, and 1935 study samples of Harris and Benedict

None reported.

Definitions:

“basal energy expenditure (BEE)” – subjects traveled to test site, arriving ~8 a.m., fasted during previous 12 hr; laid still on a bed or couch for ~30 mins; Subjects in the 1935 series were studied in their homes immediately on awakening in the morning before rising from bed (truly basal conditions).

Study Protocol:

The study methods, design and subjects studied were compared.

The study techniques of completing indirect calorimetry were described in detail.

Harris and Benedict took a then common shortcut in converting gas exchange values to energy expenditure by using conversion constants for the nonprotein respiratory quotient.

Carbon dioxide production and oxygen consumption were measured gravimetrically.  From the 1932 experiments, a calorimeter that had a helmet (constructed from a metal bucket) for gas collection was used; whereas in 1935 series a protable field unit made up of a mouthpiece and nose clips was used.  In both types, CO2 was measured gravimetrically; oxygen consumption was measure volumetrically (spirometry in 1932 series; by a timed, metered pump in the 1935 series).

1. Demographic characteristics such as height, weight, age, oxygen consumption (mL/min), basal energy expenditure (kcal/d) and BMI
2. The information was re-analyzed by the 1919 series and the entire series (1919-1935) by gender.  Correlation coefficients (r ) for weight and age, weight and height, weight and energy expenditure, height and energy expenditure.  Partial  correlation for weight and energy expenditure with height controlled; heainght and energy expenditure and weight controlled, and weight and energy expenditure and age controlled were reported;
3. Energy differences from measured basal energy expenditure minus predicted (disregarding sign) using the original and combined series equations
4. Measured basal energy expenditure in the obese and nonobese subjects, controlling for weight, height and age

Number of articles included : Included 8 primary research studies by Benedict; And used an additional 18 articles for a total of 26 articles

The sample sizes of studies performed by Harris and Benedict original and supplemental series ranged from n=6 to n=136. 

 

Characteristics of participants:

ANTHROPOMETRIC

1919 Men (n=136)

	Mean±SD	Range
Age	27± 9	16-63
Wt, kg	64±8	33-109
Ht, cm	173±10	148-198
BMI	21.1±2.8	15.2-32.5

1919 Men (Supplemental n=31)

	Mean±SD	Range
Age	24±4	19-44
Wt, kg	65±7	51-80
Ht, cm	173±6	161-190
BMI	21.6±2.4	16.8-25.5

1928 Men (n=27)

	Mean±SD	Range
Age	34±8	21-89
Wt, kg	72±20	55-157
Ht, cm	174±8	154-188
BMI	23.8±7.2	17.3-55.5

1932-1935 Men (n=6)

	Mean±SD	Range
Age	83±7	74-91
Wt, kg	65±10	52-79
Ht, cm	171±4	161-176
BMI	22.2±2.7	18.3-25.4

1919 Women (n=103)

	Mean±SD	Range
Age, y	31±14	15-74
Wt, kg	56±11	36-94
Ht, cm	162±5	151-176
BMI	21.5±4.1	12.3-34.6

1928 Women (n=33)

	Mean±SD	Range
Age, y	31±14	15-74
Wt, kg	64±12	45-99
Ht, cm	161±6	151-180
BMI	24.4±4.7	16.1-38.3

1932-1935 Women (n=33)

	Mean±SD	Range
Age, y	76±6	66-88
Wt, kg	56±12	32-81
Ht, cm	154±7	138-168
BMI	23.4±4.8	16.2-40.0

Results:   

RESTING ENERGY EXPENDITURE

	Men	Women
	Mean±SD	Mean±SD
1919	1632± 205	1349±156
1919 suppl	1653±156
1928	1648±297	1352±117
1932-1935	1293±280	1108±179

Harris-Benedict completed partial correlation analysis which revealed the equations used today with the r² value of 0.75, P<.001 for men and r²=0.53, p<.0001 for women.

Normal BEE was considered to be within 10% of the value predicted by the equation; although a notable number of subjects fall outside the 10% limit.  9% of the original men and 20% of the original women, had actual BEE outside the ± range.

Benedict argued that the prediction for women was 5% too high.

Benedict expressed that lean body mass rather than total body weight was the quantity most predictive of BEE because of a strong covariance with lean body mass; Covariance between total body weight, lean body mass, and BEE has since been well-established.  (Segal, 1985, 1987, 1989).

Age and gender:

The gender effect in the original data was reconfirmed by reanalysis using a general linear model, using only the measured BEE and controlling for covariance with wt, height and age.  A statistically significant difference (p<0.01) remained between male (a558±123 kcal/day) and female subjects (1446±127 kcal/day) (8% on average).

When the covariance between measured lean body mass and BEE was controlled, young men had a higher BE than elderly men by 6%; and men had a higher BEE than women (by 4%).

Other studies:

Confirm about a 5% difference between measured BEE and the Harris-Benedict standards

OBESE AND NONOBESE SUBJECTS

Obese persons have reduced energy expenditure per kg total body weight but normal energy expenditure per kg lean body mass (Segal, 1985, 1987, 1989)

HB 1919 data included 7 obese men and probably 22 women.  In a reanalysis, when covariance of body weight was controlled for, obese subjects had a significantly lower BEE than non-obese subjects (P<0.01).  When height and age were added into the equation, Bee became nearly identical between the obese and nonobese subjects.  This indicates that HB equations accurately predicted BEE in subjects with some degree of obesity.

Feurer (1983) studied a more more obese population (BMI 37-78) than HB (BMI 28-40); use of total body weight in the HB equations in Feurer’s morbidly obese study population, did, in fact, overestimate true BEE by a variable degree (15% on average)

As stated by the author in body of report:

“Error is inherent in the Harris-Benedict equations. Only between 50-75% of the variability in BEE is explained by these equations, but subsequent equations have not general improved on this level of random error.

Harris-Benedct equations systematically overestimate BEE by at least 5%.”

HB data appear to apply up to a body mass index of 35-40.”

University/Hospital:

Pennsylvania State University

Strengths:

• Researcher did an excellent job describing subject sample differences and measuring protocol back in early 1900’s
• Reanalysis important information for practicing RDs

Limitations:

• Publication standards in 1998 didn’t have to include search strategy and methods used to select studies, explicitly.
• Did not discuss rationale for including 1935 Series (visits to subjects’ homes and measured in beds before arising (BEE), i.e., 5 men, 10 women) with other data where subjects drove to testing site (REE).