DETERMINATION OF CALORIES BY NIR

Authors: Kays, Sandra; Barton Ii, Franklin

Submitted to: United States Japan Natural Resources Protein Panel
Publication Type: Proceedings
Publication Acceptance Date: 12/3/2002
Publication Date: 12/31/2002
Citation: KAYS, S.E., BARTON II, F.E. DETERMINATION OF CALORIES BY NIR. PROCEEDINGS OF THE 31ST UNITED STATES JAPAN NATURAL RESOURCES PROTEIN PANEL. 2002. P. EE 1-8.

Interpretive Summary: An expanding epidemic of obesity in the United States, with its short- and long-term health effects has made food energy content a critical component in diet selection. In the USA, statement of the energy content per serving and energy from fat per serving is compulsory for food nutrition labels. This report describes work conducted at the USDA-ARS Russell Research Center on the evaluation of energy content in a broad range of cereal products using near-infrared (NIR) reflectance spectroscopy. NIR reflectance spectroscopy, which measures the amount of light energy reflected by a substance, is a very rapid and accurate method of measuring constituents of materials without requiring extensive sample preparation, or creating chemical waste. Previous reports from this laboratory have demonstrated the potential of NIR for the simultaneous prediction of dietary fiber, protein and fat in cereal food products. It would be advantageous to be able to simultaneously determine all macronutrients required on the nutrition label using NIR, however, very little is known about the potential of NIR spectroscopy to determine the available energy in foods for human consumption. The objective of this project was to determine the potential of NIR spectroscopy for the prediction of energy in diverse cereal food products. It was found that models developed to predict utilizable energy (energy adjusted for unutilized protein and insoluble dietary fiber) gave results that were well within the accuracy required for nutrition labeling purposes. NIR spectroscopy, thus, provides a rapid and acceptable method of predicting energy in diverse cereal products.

Technical Abstract: Although NIR spectroscopy has been used to determine energy content of feeds, little is known about the capability of near-infrared (NIR) spectroscopy for the determination of available energy in foods for human consumption. Recent studies have been conducted to determine the utilizable energy in processed cereal food products by NIR spectroscopy. NIR reflectance spectra of ground cereal products were obtained with an NIRSystems 6500 spectrometer. Gross energy for each product was determined by oxygen bomb calorimetry and values for available energy were calculated by adjusting gross energy to allow for unutilized protein and undigested insoluble dietary fiber. Modified partial least squares models, developed to predict gross energy and utilizable energy of diverse cereal products, gave standard errors of cross validation for NIR prediction of gross energy (range 4.05-5.49 kcal/g), energy of samples after adjustment for unutilized protein (range 3.99-5.38 kcal/g) and energy of samples after adjustment for unutilized protein and insoluble dietary fiber (range 2.42-5.35 kcal/g) of 0.053, 0.053 and 0.088 kcal/g, respectively, with multiple coefficients of determination of 0.96. Use of the models with independent validation samples (n=58) gave predicted energy values within the accuracy required for USA nutrition labeling legislation. Thus, NIRS provides a rapid and acceptable method for predicting energy of diverse cereal foods.