PREDICTIVE EQUATIONS ARE INACCURATE TO ASSESS CALORIC NEEDS IN NON-CAUCASIAN ADULTS FROM CHILE

Abstract Background Predictive equations are frequently used to estimate resting energy expenditure (REE) since indirect calorimetry is not always available and is expensive. Aim To determine the concordance between the estimation of resting energy expenditure using predictive equations and its measurement by indirect calorimetry. Material and methods Analysis of the registry of indirect calorimetry performed in non-hospitalized participants. Harris-Benedict, FAO/OMS/ONU, Mifflin St Jeor, and ESPEN equations were used to estimate REE in these individuals. The concordance between measured and estimated REE using real, ideal and adjusted weight was calculated using the concordance coefficient analysis of Lin and Bland & Altman plots in all participants and in subgroups separated according to their body mass index. Results We retrieved 680 measurements and discarded 247 that did not comply with the inclusion criteria. Thus, we studied 433 participants aged 36 (29-48) years (341 females, 79%) with a body mass index (BMI) of 30 (26.7-33.1) kg/m2.All predictive equations had concordance values of ​​less than 0.90. The proportion of participants in which the difference was greater than 10% ranged from 36% to 87%.The ESPEN equation had the greater proportion of erroneous estimations of REE in all participants and BMI subgroups when the real weight was used. Conclusions We observed a low level of concordance between REE estimated using predictive equations and measured by indirect calorimetry. These results should alert clinicians about the inaccuracy of predictive equations.