TABLE 7.

Comparison of predicting total energy expenditure from age, weight, height, physical activity, and resting metabolic rate either measured or predicted from published equations

Prediction	Males (n = 638)								Females (n = 1019)
	Regression1								Regression1
	TEE	rc	Bias2	LOA3	Intercept	Slope	rs	MAPE	TEE	rc	Bias2	LOA3	Intercept	Slope	rs	MAPE
	kJ/d		kJ/d	kJ/d (%)	kJ/d			(%)	kJ/d		kJ/d	kJ/d (%)	kJ/d			(%)
Reference value	10,192 ± 2081								8059 ± 1676
Measured RMR	10,128 ± 1963	0.94	−64	−1280 to 1406 (−12.9 to 13.4)	381	0.96	0.94	4.5	8058 ± 11573	0.94	−1	−1131 to 1132 (−14.7 to 13.8)	443	0.94	0.93	4.6
Livingston et al. [16]	10,235 ± 1896	0.86	43	−2091 to 2006 (−23.0 to 20.1)	921	0.92	0.86	5.8	7961 ± 1670	0.84	98	−1758 to 1954 (−22.9 to 24.2)	117	1.0	0.82	7.7
Porter et al. [18]	10,124 ± 1858	0.85	−68	−2067 to 2204 (−23.1 to 22.3)	913	0.90	0.84	6.4	8052 ± 1681	0.83	−7	−1887 to 1900 (−24.7 to 23.6)	−78	1.01	0.81	8.2
Ikeda et al. [15]	9983 ± 1809	0.85	−209	−1874 to 2293 (−20.8 to 22.7)	1052	0.88	0.85	6.7	8120 ± 1631	0.83	61	−1927 to 1804 (−25.4 to 22.1)	263	0.98	0.81	8.0

LOA, limit of agreement; MAPE, median absolute percentage error; RMR, resting metabolic rate; TEE, total energy expenditure.

¹Regression, measured TEE versus predicted; r_s, Spearman correlation coefficient; r_c , Lin’s concordance correlation coefficient.

²Measured − predicted; TEE = mean ± SD.

³LOA = 1.96 SD limits of agreement.