PD466 - ACCURACY OF PREDICTIVE EQUATIONS COMPARED WITH INDIRECT CALORIMETRY IN ESTIMATING RESTING ENERGY EXPENDITURE IN PATIENTS WITH TYPE 2 DIABETES

Linked sessions

PD466

ACCURACY OF PREDICTIVE EQUATIONS COMPARED WITH INDIRECT CALORIMETRY IN ESTIMATING RESTING ENERGY EXPENDITURE IN PATIENTS WITH TYPE 2 DIABETES

 

P. S. Barcellos1,2,3,*, M. Lemos-Araujo4, D. M. Linhares-Sousa2, E. Lima3, N. Borges1, D. P. M. Torres1

1FCNAUP, Porto, Portugal, 2PPGN, 3CCNUT, 4PPGEF, UFMA, São Luís, Brazil

 

Rationale: This study aimed to assess resting energy expenditure (REE) measured by indirect calorimetry (IC) and compare it with estimates from predictive equations to evaluate their accuracy.

Methods: T2DM patients (26) from Hospital São João, Portugal. REE was measured by IC and estimated using the Harris–Benedict, Ireton-Jones, Barcellos, Huang et al., Martin et al., and Ikeda et al. equations. Statistical analysis was performed using IBM SPSS, and results were expressed as mean ± standard deviation. Comparisons were conducted using the Mann–Whitney U test, with a significance level set at p ≤ 0.05.

 

Results: Table 1:  Evaluation of different equations used to estimate Resting Energy Expenditure in the validation cohort (n = 26).

 

Equations

REE

(kcal)

REE

(kcal/kg)

Accurate prediction¹ (%)

Underestimation² (%)

Overestimation³ (%)

Bias⁴ (%)

ICC⁵ [95% CI]

REE-CI

1767.06 ± 664.10

24.20 ± 8.71

 

 

 

 

 

Harris–Benedict

1436.50 ± 228.77

19.54 ± 1.62

15.4

57.7

26.9

-9.69 ± 29.52

0.140

Ireton–Jones

1473.28 ± 406.05

20.04 ± 4.89

26.9

46.2

26.9

-9.04 ± 30.33

0.712

Barcellos I

1915.37 ± 437.51

26.20 ± 5.70

30.8

23.1

46.2

16.42 ± 41.65

0.055

Barcellos II

1884.63 ± 465.67

25.62 ± 5.25

30.8

23.1

46.2

19.79 ± 43.68

0.140

Huang

1613.64 ± 248.67

22.01 ± 2.42

23.1

46.2

30.8

1.53 ± 34.48

0.658

Martin

1640.48 ± 48

21.5 ± 5.04

32.0

36.0

32.0

3.74 ± 36.74

0.460

Ikeda

1366.15 ± 183.58

18.66 ± 1.60

15.4

65.4

19.2

-13.54 ± 29.34

<0.001

¹% predicted  ±10% of the measured; ² % <−10% of the measured; ³ %>+10% of the measured; ⁴Bias: mean % error between estimated and measured REE; ⁵Intraclass correlation coefficient.

Conclusion: None of the predictive equations showed accuracy in estimating REE in patients with T2DM. Underestimation was predominant, with the Harris–Benedict and Ikeda equations. Although Huang and Martin performed better, variability among methods highlights the need for caution, especially given the metabolic heterogeneity in this population.

Disclosure of Interest: None declared